PB9 3-174 787
EPA/600/R-93/058c
April 1993
Im9m9ul
SOLAR
WORLD
CONGRESS
Proceedings of the Biennial Congress of
the International Solar Energy Society,
Denver, Colorado, USA, 19-23 August 1991
VOLUME 2. PART I
M.E. ARDEN
SUSAN M.A. BURLEY
MARTHA COLEMAN
REPRODUCED BY
U.S. DEPARTMENT OF COMMERCE
NATIONAL TECHNICAL INFORMATION SERVICE
SPRINGFIELD, VA. 22161
PERGAMON PRESS
-------
-------
INTERNATIONAL SOLAR ENERGY SOCIETY
The International Solar Energy Society is a worldwide nonprofit organization dedicated to the
advancement of the utilization of solar energy. Its interests embrace all aspects of solar energy, including
characteristics, effects and methods of use, and it provides a common meeting ground for all those
concerned with the nature and utilization of this renewable non-polluting resource.
Founded in 1954, the Society has expanded over the years into a truly international organization with
members in more than 90 of the world's countries. It has been accepted by the United Nations as a
nongovernmental organization in consultative status, and it is widely regarded as the premier body of it'
type operating in the solar energy field.
The Society is interdisciplinary in nature and numbers among its members most of the world's leading
figures in solar energy research and development, as well as many with an interest in renewable energy and
its practical use. High academic attainments are not a prerequisite for membership, only a special interest
in this particular field.
Organization
Day-to-day administration is provided by the Society's headquarters office, which since 1970 has been
located in Australia. The headquarters house the Secretary-Treasurer and the Administrative Secretary,
together with members of their supporting staff.
In countries and regions in which sufficient interest exists, Sections of the Society have been established.
These Sections, which are largely autonomous, organize meetings and other local activities and in some
cases produce their own publications. All Society members are eligible to belong to their respective
national or regional Sections, although in some cases this may involve the payment of an additional
Sectional fee. In recent years the number of Sections has increased slowly but steadily.
Activities of the Society are:
1. Publications of Solar Energy, a monthly scientific journal of an archival nature, containing scientific
and technical papers on solar energy and its utilization, reviews, technical notes and other items of
interest to those working in the field of solar energy.
2. Publication of a less technical magazine, SunWorld.
3. Publication of a newsletter for members, ISES News.
4. Organization of major International Congresses on solar energy at which numerous scientific and
technical papers are presented and discussed. These Congresses are held every two years in different
countries, normally in conjunction with equipment exhibitions, and are widely attended.
5. Publication of the Proceedings of each International Congress. Whereas copies of the Society's three
periodicals (items 1-3 above) are supplied to all members as part of their membership, copies of
Congress Proceedings are available (from the publisher) only on special order and at an additional cost.
Special pre-publication prices are normally available to Society members.
6. More recently ISES has become increasingly involved with other major Non-Governmental
Organizations in matters relating to the application of renewable energy and other global environmental
problems, and is currently preparing its contribution for presentation at the United National Conference
on Environment and Development (UNCED - or popularly referred to as ECO 92).
Headquarters:
International Solar Energy Society Telephone: 61 3 571 7557
PO Box 124 Fax: 61 3 563 5173
Caulfied East, Vic. 3145 Telex: AA 154 087 CITVIC
AUSTRALIA
v
-------
AMERICAN SOLAR ENERGY SOCIETY
The American Solar Energy Society (ASES) is the United States Section of ISES and presently
has over 4,000 members.
ASES seeks to promote the widespread near-term and long-term use of solar energy. To achieve
that goal, ASES:
• Fosters the use of science and technology in the application of solar energy;
• Encourages basic and applied research and development in solar energy;
• Promotes education in fields related to solar energy; and
• Provides information relating to all aspects of solar energy.
Activities:
• ASES conducts the National Solar Energy Conference as a annual forum for exchange of
information about advances in solar energy technologies, programs, and concepts. The
conference features speakers who are national leaders in their technical and professional fields.
Workshops, exhibits and tours of solar applications highlight this annual event, which is attended
by more than 450 solar energy enthusiasts from throughout the country.
• ASES publishes Solar Today, a bi-monthly magazine. Each issue highlights practical
applications of solar energy, presents the latest results of solar energy research, covers
developments in the nation's solar energy industry, and includes member discussion of solar-
related issues.
• Each year, ASES sponsors a Roundtable in Washington, DC, bringing together energy decision-
makers in a highly visible public forum. Each Roundtable addresses an issue of critical
importance to ASES members and the nation.
• To ensure worldwide dissemination of information about solar energy developments, ASES
annually publishes Advances in Solar Energy. This compendium of the latest R&D
developments is authored by ASES members who are nationally recognized experts on their
respective topics.
• Technical, regulatory and educational issues are addressed in periodic White Papers, which
present critical analyses of important solar energy topics.
• ASES educates the public and energy decision-makers on the benefits of solar energy through a
public relations campaign and information materials.
• ASES has 16 state and regional chapters, which are independendy incorporated organizations
providing services to their members appropriate to the local areas. Typical activities include
newsletters, technical meetings, public outreach activities, and government relations.
Headquarters
2400 Central Avenue, Suite B-l
Boulder, CO 80301
Telephone : 303-443-3130
Fax : 303-443-3212
vi
-------
Table of Contents
Volume 1: Solar Electricity, Biofuels, Renewable Resources
1.1 Photovoltaic Thin-Film Materials and Devices 1
1.2 Photovoltaic High Efficiency and Applications 39
1.3 Photovoltaic Modeling and Batteries 73
1.4 Large-Scale Photovoltaic Applications 1 117
1.5 Large-Scale Photovoltaic Applications II 155
1.6 Small-Scale Photovoltaic Applications 1 193
1.7 Small-Scale Photovoltaic Applications II 231
1.8 Photovoltaic Applications 269
1.9 Photovoltaic Utility Issues 307
1.10 Photovoltaic Issues 343
1.11 Posters: Photovoltaics 393
1.12 Posters: Photovoltaic Systems 465
1.13 Solar Thermal Electric 531
1.14 Wind Energy Experiences 563
1.15 Wind Energy Systems Performance 599
1.16 Wind Systems Applications and Hydropower 639
1.17 Utility and Regulatory Issues 677
1.18 Solar Hydrogen Technologies 715
1.19 Biotechnology 757
1.20 Bio-Chemical Conversion 797
1.21 Biofuels 835
1.22 Radiation Instruments, Measurements 891
1.23 Radiation Models 931
1.24 Radiation Models, Simulation 969
1.25 Radiation Resources 999
1.26 Use of Radiation Data 1037
1.27 Renewable Resource Posters 1075
Volume 2: Active Solar and Solar Heat
2.1 Collectors I: Selective Surfaces 1139
2.2 Collectors II 1183
2.3 Collectors IE 1219
2.4 Collectors IV 1257
2.5 Solar Domestic Hot Water 1 1297
2.6 Solar Domestic Hot Water II 1337
2.7 Solar Domestic Hot Water III 1373
2.8 Passive Domestic Hot Water 1409
2.9 Solar Water Heaters 1449
2.10 Active Heating I: Seasonal Storage 1485
2.11 Active Heating II: Heating System Performance 1523
2.12 Active Heating III 1563
2.13 Active Cooling I: Advances in Open Cycle Absorption 1599
2.14 Active Cooling II 1635
2.15 Active Solar 1673
vii
-------
2.16 Posters: Active Solar I 1709
2.17 Posters: Active II 1781
2.18 Concentrating Collectors 1853
2.19 High Flux : 1891
2.20 Solar Heat Storage 1929
2.21 Central Receivers 1967
2.22 Dish Collectors 2005
2.23 Line-Focus Collectors 2043
2.24 Detoxification and Materials 2083
2.25 Solar Heat 2121
2.26 Solar Heat Posters 2159
2.27 Desalination 2251
2.28 Solar Ponds 2289
2.29 Desalination and Solar Ponds Posters 2327
2.30 Solar Drying 1 2401
2.31 Solar Drying H 2439
Volume 3: Passive Solar, Socio-Economic, Education
3.1 Solar Building Designs 2497
3.2 Zero-Energy Building Designs 2529
3.3 Emerging Architecture 2565
3.4 Vernacular Architecture 1 2599
3.5 Vernacular Architecture II 2635
3.6 Passive Commerical Buildings 2665
3.7 Daylighting I 2701
3.8 Daylighting n 2739
3.9 Atriums 2779
3.10 Passive Strategies and Materials 1 2817
3.11 Passive Strategies and Materials II 2851
3.12 Passive Strategies and Materials m 2887
3.13 Transparent Insulation I 2925
3.14 Transparent Insulation II 2957
3.15 Convection and Mass 2989
3.16 Comfort 3027
3.17 Passive Cooling 1 3065
3.18 Passive Cooling n 3101
3.19 Passive Computer Analysis 1 3137
3.20 Passive Computer Analysis II 3173
3.21 Passive Computer Analysis in 3209
3.22 Monitored Passive Modules 3247
3.23 Extended Passive Monitoring 3285
3.24 Passive Non-Computer Design Tools 3325
3.25 Sustainability 1 3363
3.26 Sustainability n 3399
3.27 Passive Posters 1 3437
3.28 Environmental Effects 3521
3.29 National Solar Programs 3559
3.30 Developing Country Applications I 3597
vlii
-------
3.31 Developing Country Applications II 3633
3.32 Technology Transfer 3667
3.33 Socio-Economic Posters 3705
3.34 Education 3757
3.35 Education Posters 3803
Volume 4: Plenaries, State-of-the-Art, Farrington Daniels Lecture
ix
-------
Contents of Volume 2
2.1 Collectors I: Selective Surfaces
G. Brouwer 1141
Performance Criteria For Spectral Selective Coatings On Solar Absorbers
T. K. Lee, S. H. Chon, Y. H. Chea and P. C. Auh 1147
The Effects of the Pulse Current Electrolysis on the Electrocystallization of a Black Chrome
Solar Selective Coating
L. Cindrella, C. E. Sooriamoorthi and S. John 1154
Improved Nickel-Based Alloy Coatings for Solar Thermal Applications
Z. Yu-ying, S. Gang and W. De-rong 1161
The New Method To Determine Technic Conditions of Sputtered Base Layer of Spectral
Selective Surface
A. G. Monger and D.R. Mills 1165
Radiation Limitation of Stagnation Temperature in High Temperature Selective Absorbers
Qi-Chu Zhang, D.R. Mills and J.C. Kelly 1171
High Absorptance Selective Surface for High Temperature Solar Thermal Collectors
A. Roos, E. Wackelgard, G. Chinyama 1177
Tin Oxide Coated Selective Absorber Surfaces with Extreme Thermal, Chemical and
Mechanical Stability
2.2 Collectors II
S. Svendsen and F. Kristiansen 1183
Temperature Distribution in the Glazing on Solar Collectors
A. Nordgaard and W. Beckman 1189
Investigation of Flat-Plate Monolithic Silica Aerogel Collectors
G. Angermeier, S. Harrison, A. Richter and H. Soltau 1194
Developing a General Model of the Wind Dependent Heat Loss of a Flat Solar Receiver
Y. Yiqin, K. G. T. Hollands and A. P. Brunger 1200
Measured Top Heat Loss Coefficients for Flat Plate Collectors with Inner Teflon Covers
Y. B. Safdari, D. Witek, and A. Fakheri 1206
Two Dimensional Transient Analysis For The Solar Heating of A Fluid By A Partially
Radiation Absorbing Medium
A. Goetzberger, J. Dengler, M. Rommel and V. Wittwer 1212
The Bifacial-Absorber Collector: A New Highly Efficient Flat Plate Collector
2.3 Collectors III
B. Perers and H. Walletun 1221
Dynamic Collector Models for 1 HrTxme Step Derived from Measured Outdoor Data
L. Broman and G. Datta 1227
Solar Collector Geometries At High Latitudes
D. R. Mills, A. G. Monger and G.L. Morrison 1233
Optimisation of Fixed Reflector Design to Minimise Both Backup Requirements and Solar
Collector Receiver Area
J. K. Nayak, S.V. Bopshetty and S.P. Sukhatme 1239
Thermal Performance of Solar Concrete Collector
C. F. Kutscher and C. Christensen 1245
Unglazed Transpired Solar Collectors: An Analytical Model and Test Results
X
-------
J. C. Baltazar, E. Torres 1251
Thermal Analysis of a Refrigerant-Filled Solar Collector
2.4 Collectors IV
T. Mizuno, H. Shimizu, S. Asai 1259
Efficiency of a Solar Collector used in a Typical Transparent Honeycomb Insulation
S. Kothari, N. K. Bansal, N. S. Rathoce 1265
Use of Transparent Insulation as Cover in an Integrated Collector cum Storage System
W. J. Platzer, V. Wittwer 1272
Future Development in the Field of Transparent Insulation Systems
T. Kobayashi, H. Shimizu, S. Asai 1278
Numerical and Experimental Analysis of Convection Suppresionfor Charged Gas Solar
Collector
N. Benz, W. Scholkopf and R. Sizmann 1284
Thermal Performance of Improved Evacuated Flat Plate Collectors
K. R. Schreitmiiller, Martina Niemann and Gunther Rockendorf 1290
Detailed Experimental Investigations on Evacuated and Other High Performance Collectors
2.5 Solar Domestic Hot Water I
F. de Winter, A. Arata and M. Perlman 1299
Design and Performance of Small Solar Water Heating Systems
A. Arata and F. de Winter 1306
Design and Performance of Large Solar Water Heating Systems
W. S. Duff and E.C. Boardman 1313
Developing Performance Models of Solar Energy Systems by Daily Energy Input/Output
Curves
W. T. Carlson. J. H. Davidson and W.S. Duff 1319
Comparison of Experimental and TRNSYS Ratings of Generic Drain-Back Solar Water
Heater
W. S. Duff and M. Chandrashekar 1325
Model To Model Testing of Six Solar Energy Design Programs
E. Mannik, J. Atkinson and G. Morrison 1331
A Process for Making Simplified Methods for Systems with Storage
2.6 Solar Domestic Hot Water H
P. J. Schaefer, W. Beckman and S. A. Klein 1339
Comparison Between Experimental and Simulated DHW Sort-Term Test Results
J. H. Davidson, W.T. Carlson and W.S. Duff 1345
Impact of Component Selection on SRCC Rating of Drain-Back Solar Water Heaters
L.A.M. Ramaekers and C.J. van der Leun 1351
Integrated Collector Storage DHW System Numerical Simulation of Heat Transfer and Fluid
Flow
G.A.H. van Amerongen, H. Visser and A.C. de Geus 1357
Investigation on Solar Domestic Hot Water Systems Combined With Auxiliary Heaters
C. W. J. van Koppen 1363
Low Flow or Single Pass: The Heart of the Matter
T. Esbensen 1369
Low-Flow Solar Hot Water System
xi
-------
2.7 Solar Domestic Hot Water III
S.Furbo 1375
Low Flow Solar Heating Systems - Theory and Practice
C. Arkar, S. Medved and P. Novak 1383
Dynamic Method For Solar System Testing-Measurement Results And Long-Term
Performance Prediction of Different SDHW Systems
A.C. de Geus, H. Visser, G.A.H. van Amerongen 1386
Optimization of the Primary Circuit ofSDWH Systems
W. Spirkl and J. Muschaweck 1392
Using a Plug Flow Model for Short Term Testing of Solar Domestic Hot Water Systems
E. M. Kleinbach, W. Beckman, and S. A. Klein 1397
Performance of One-Dimensional Models for Stratified Thermal Storage Tanks
J. van Berkel, W. B. Veltkamp, and A. B. Schaap 1403
Thermal Behavior of a Heat Exchanger Coil in a Stratified Storage
2.8 Passive Domestic Hot Water
G. R. Swindler 1411
Rooftop Greenhouse/Solar Water Heater: An Urban Retrofit for Energy Self-Reliance
Ee-Tong Pak 1417
Thermally Stratified Hot Water Storage
H. A. Walker and J. H. Davidson 1424
Design Optimization of a Two-Phase Solar Water Heater Operating in Fort Collins, Colorado
with R-123
E. Haines, D. Boleyn, C. Dallas 1430
Testing the Copper Cricket (Tm) Solar Water Heater by an Electric Utility
D. Galor 1436
The Contribution of Energy-Conscious Building Regulations to Creative Design Proposals
H. Visser and A. C. de Geus 1442
Integrated Collector Storage: Model Development For Performance Calculations and Test
Evaluation
2.9 Solar Water Heaters
V. Balasubramanian, S. Jayaraman, K. Perumal, G. Sankarasubramanian 1451
In Situ Solar Steam Generation With CPC Modules
J. I. Stewart 1456
Development and Enhancement of a Solar Boosted Domestic Hot Water System
W. Qing 1462
ICC Solar Water Techno-Economic Analysis
M. Rommel and V. Wittwer 1468
Transparently Insulated Solar Pond
D. Gudino, M. P. Salas and J. J. Hermosillo 1474
Computer Programs for Designing and Simulating of a Swimming Pool Solar Heater
L. Weide, W. Liting and H. Zhichen 1480
Flow Pattern and Heat Extraction from a Horizontally Disposed Single Ended Evacuated All
Glass Solar Absorber Tube
2.10 Active Heating I: Seasonal Storage
S. Svendsen, P. Berg and K. Duer 1487
Heat Transfer in Boreholes for Seasonal Solar Storage in the Ground
xii
-------
M. Unsal and B. Yumrutas 1493
Prediction of the Long Term Performance of Seasonal Spherical Underground Solar Thermal
Energy Stores
M. Beck, M. Reuss, H. Schulz, J. Spannig and B. Wagner 1499
Longterm Storage of Low Temperature Solar Heat in the Ground
R. Kiibler, N. Fisch, F. Miiller and E. Hahne 1505
Central Solar Heating Plants with Seasonal Storage-Results of a Site Specific Feasibility Study
in Germany
W. W. Wisniewski 1511
Annual Cycle Solar Heat Storage
J-O. Dalenback 1517
The Sard Project-Solar Heating Plant with Seasonal Storage
2.11 Active Heating II: Heating System Performance
Y. Shao 1525
Entropy Analysis of a Solar Space Heating System
C.C. Smith 1531
Comparison of Large Sized Solar Air and Liquid Systems for Industrial Space Heating
S. Carpenter and J. Kokko 1537
Performance of Solar Preheated Ventilation Air Systems
B. Daniotti, V. K. Sharma 1543
Experimental Studies on Large Sized Conventional Solar Air Heater for Agricultural Use
M. Yarosh, J. Huggins, T. Tiedemann 1549
Technical Inspections of a Large Number of Commercial Solar Water Heating Systems on
Florida Schools
B. W. Zingano, H.M. Kanjere and D. Kaduya 1555
Performance of Solar Water Heaters in Five Hospitals in Malawi . ~
2.12 Active Heating III
K. G. T. Hollands, A. Karagiozis, A. P. Brunger and G. Brouwer 1565
Explaining Selective Surface Degradation Effects on Solar Heating Performance
S. S. Peltola 1571
Potential of Solar District Heating in Finland
D. S. Breger, J E. Sunderland and H. Elhasnaoui 1577
Preliminary Design Development of a Central Solar Heating Plant with Seasonal Storage at
the University of Massachusetts, Amherst
0. Guisan, B. Lachal, A. Mermoud, D. Pahud 1583
Study of a 20,000 rr? Seasonal Heat Storage Fed by Solar Collectors
D. Chwieduk 1589
A Simulation Study of a Dual Source Solar Assisted Heat Pump System for a Family House
Heating in Poland
S. Takama 1595
10 Years Operation of a Large Scale Solar System, Doho Park Gymnasium
2.13 Active Cooling I: Advances in Open Cycle Absorption
K. Speidel, H.P. Kleinmeier 1601
Solar Cooling and Air Conditioning Processes Using Chemical Reactions
1. Haim, G. Grossman and S. Shavit 1607
Simulation and Analysis of Open Cycle Absorption Systems for Solar Cooling
xiii
-------
T. G. Lenz and S. V. Potnis 1613
Energy Balance and Mass Transfer Studies for a Liquid Desiccant Based Solar Cooling System
M. N. A. Hawlader, A. P. Stack and B. D. Wood 1619
Open-Cycle Absorption Solar Cooling: Glazed and Unglazed Open Flow Liquid Absorbent
Solar Collector/Regenerator
R. Yang, W.D. Chang and C. J. Peng 1625
Experimental Study for a Glazed Solar Collector/Regenerator of the Open Cycle Absorption
Solar Cooling System
Y. Saito 1630
Regeneration Performance of an Adsorbent in the Desiccant/Regenerator and Improvement of
Regeneration Efficiency
2.14 Active Cooling n
P. Kumar, T. K. Chaudhuri, and A. Dasgupta 1637
Exergy Analysis of a Solar Absorption Refrigeration System for Efficient Utilization of Solar
Energy
H. Zhi-cheng, X. Wen-hui and M. Wei-bin 1643
A 2-Stage LiBr Absorption Chiller for Solar Cooling
J. G. Cervantes, E. Torres and J. C. Baltazar 1649
Performance Testing of a Solar Assisted Heat Pump
S. Ito and N. Miura 1655
A Comparison of Heat Pump Systems Using Different Types of Direct Expansion Solar
Collectors
L.C. Chen, F.K. Kao and T. Tang 1661
An Experimental Study of Lithium Chloride Solar Liquid Desiccant System
D. R. Neill, M. Bean, T. Ho, and L. Huang 1667
Renewable Energy Space Cooling
2.15 Active Solar
K. Knappmiller and W. Duff 1675
Computing Incidence Angle Modifiers for Advanced Solar Collectors
C. Armenta-Deu and B, Lukac 1680
A Method to Compute the Incidence Angle Modifier and to Estimate Its Incidence on
Collectible Solar Energy
K.R. Schreitmuller, M. Mack and G. Hesse 1685
ISFH-An Adaptive Simulation Model
C. Bankston 1691
A Simplified Technique for Estimating the Economic Optimum Temperature Swing of Thermal
Energy Storage in Solar Heating Systems
V. L. Fara, A. Bucur 1696
Computer Simulation of a Physico-Mathematical Model of Solar Energy Thermal Storage
A. C.de Geus and H. Visser 1702
Validation of a Paramenter Identification Method within The IEA Dynamic Systems Testing
Group
2.16 Posters: Active Solar I
M. Unsal 1711
Optimal Collector Side Mass Flow Rate in Double-Loop Solar Water Heating Systems
-------
Y. Wei 1715
A Performance Prediction of a Solar Active Heating System and the Practice in North-West
China
B. J. Huang, S. C. Du and R.H. Yen 1720
Stability Analysis of Solar Thermosyphon Water Heaters
J. Nicolas, Ph. Andre, J-F. Rivez, V. Debbaut 1726
Use of an Expert System for the Real Time Control of a Solar Building
F. Kristiansen and S. Svendsen 1732
Cover System for Large Roof-Placed Collectors
R. Alim 1738
A Combined Dual Liquid-air Fluid Transport Medium In An Optimum Flat Plate Solar
Collector Energy System
M. Reuss and H. Schulz 1742
Performance Measurements on two Selfbuilt Domestic Hot Water Systems
D. L.Loveday 1749
The Use of Profiled Metal Cladding as a Solar Collector in an Air-Source Heat Pump System
A. C. Gillet 1755
Summer Overheating Problems in a Solar TV Complex Building in Brussels, Belgium - One
Year Experience - Records and Remedies
G. Galli, L. Laurenti and A. Ponticiello 1761
Comparison of the Performance of Solar Heating Systems Assisted by Heat Transformers and
by Compression or Absorption Heat Pumps
S. Fischer, A. Hauer, S. Hoist, W. Schoelkopf 1769
Thermochemical Energy Storage With Low Temperature Heat For Space Heating
A. B. Schapp, W. B. Veltkamp and J. van Berkel 1774
Evacuated Dish Flat Plate Collector
2.17 Posters: Active II
M. Adj, Y. Sfaxi, A. Girardey, and M. Grignon 1783
A System for Recuperating Solar Energy Falling on a Horizontal Slab
A. F. Burke 1789
A Superinsulated Passive/Active Solar House Design, Performance, and Economics
M. Y. Othman and K. Sopian 1794
Solar Assisted ADS and RSS Rubber Drying
W. J. Yan and L.C. Chen 1800
An Experimental Study of Basin-Type Integral Solar RegeneratorID ehumidifier
K. Abdullah, G. Brouwer 1806
Experiments and Simulation Results on the Thermal Performance of a Solar Tea Dryer in
Indonesia
M. J. Carvalho, M. Collares-Pereira, J. Cruz Costa, J. Oliveira 1812
Monitoring of Ten Water Heating Solar Systems
S. U. Chaudry and L. F. Jesch 1818
Computer Control and Monitoring of a Solar Space Heating System
J. Xiao-huan, Z. Shu-xia and H. Wen-xu 1823
Study in the Selective Transparent Materials of Solar Spectrum
W. Taixin ,Z. Quinhua, C. Hong and C. Xiaoxi 1828
The Preparing Process and Properties Analysis of Solar Selective Absorbing Surfaces On Mild
Steel
xv
-------
B. J. van den Brink and H.Buijs 1834
Active Hot Water Supply in Combination With The Seasonal Storage For Heating
I. Memon and B.M. Gibbs 1840
Aspects of Low Temperature Latent Heat Storage By Using Sodium Sulphate Derahydrate Salt
Mixture
D. P. Rao, S. Karmakar and T.C. Thulasidas 1846
Thermal Analysis of Solar Cooker
2.18 Concentrating Collectors
M. Collares-Pereira, F. Mendes, O. Brost, M. Groll, S. Roesler... «... 1855
Optimized Heat Pipe for Application in Integrated CPCs
K.G.T. Hollands, A.P. Brunger and I.D.Monison 1860
Evaluating Improvements To A Low-Concentration-Ratio, Non-Evacuated, Non-Imaging Solar
Collector
J. J. O'Gallagher, R. Winston and W. Duff 1866
The Integrated CPC: Solar Thermal Energy for the Nineties
T. Kotajima, H. Konno, A. Suzuki, S. Yamashiro and T. Fujii 1872
Evaluation of Heat-Collection Performance for a Stationary Mid-Temperature CPC Solar-
Collector
P. C. Eames and B. Norton . , 1878
A Unified Model for Optics and Heat Transfer in Line-Axis Concentrating Solar Energy
Collectors
P.C. Eames and B. Norton 1884
The Effect of Sky Conditions on the Partition of Incident Solar Energy Between the
Components of a CPC Solar Energy Collector
2.19 High Flux
U. Schoffel and R. Sizmann .1893
Optimization of Two-Stage Concentrating Systems for High Temperature and High Photon
Flux Density Applications
M. Schubnell and H. Ries 1899
Influence of Sunshape on the Flux Distribution in Concentrators
R. Winston, D. Cooke, P. Gleckman, H. Krebs, J. O'Gallagher and D. Sagie 1905
Brighter Than the Sun
J. O'Gallagher, R. Winston, C. Zmola, L. Benedict, D. Sagie and A. Lewandowski 1910
Ultra High Flux Concentration Using a CPC Secondary and the Long Focal Length SERI
Solar Furnace
J. A. Del Arco, J. Rodriquez a 1916
Performance and Characterization of the PSA Solar Furnace
A. Lewandowski and C. Bingham 1922
The SERI High Flux Solar Furnace
2.20 Solar Heat Storage
M.A. Rosen , 1931
On the Importance of Temperature In Performance Evaluations for Sensible Thermal Energy
Storage Systems
E. Hahne, U. Taut and U. Gross 1937
Salt Ceramic Thermal Energy Storage for Solar Thermal Central Receiver Plants
xvi
-------
A. Gluck, R. Tamme, H. Kalfa, C. Streuber and T. Weichert 1943
Development and Testing of Advance TES Materials for Solar Thermal Central Receiver
Plants
A. Bashir and B. M. Gibbs 1949
Heat Transfer of a Gas-Solid Counter-Current Thermal Storage System
A. Bashir and B.M. Gibbs 1955
A Gas-Solid Counter-Current Thermal Storage System—Hydrodynamic Characteristics
M.A. Rosen and F.C. Hooper 1961
Evaluating the Energy and Exergy Contents of Stratified Thermal Energy Storages for Selected
Storage-Fluid Temperature Distributions
2.21 Central Receivers
D. Feuermann and J. M. Gordon 1969
Analysis of A Two-Stage, Linear Fresnel Reflector Solar Concentrator
M. Castro, J. Carpio, F. Yeves, J. Berndrdez, J. Peire 1975
Solar Central Receiver Project Results
C. Winter, W. Meinecke, A. Neumann 1981
Solar Thermal Power Plants: No Need for Energy Raw Materials - Only Conversion
Technologies Pose Environmental Questions
P. Arbogast 1987
Mini Power Towers for Solar Thermal Applications
B. Gupta 1993
Solar Thermal Technology —Is It Ready For The Market
R. Pitz-Paal, J. Morhenne, M. Fiebig 1999
Optimization of the Surface Geometry of a Volumetric Foil Receiver
2.22 Dish Collectors
I. Mayer 2007
Solar Flux Distribution on Axisymmetric Receivers at the Focus of Multi-Facetted Dish
Concentrators
S. Kaneff 2013
Viable Arrays of Large Paraboloidal Dish Solar Thermal Collectors
Youssef A.M. Elgendy 2019
Analysis of a Piano-Concave Heat Pipe Receiver
L. Leon, F. Verduzco, V. Toledo, E. Villanueva, P. Quinto and A. Sanchez 2025
Solar Plant Study of Power Generation with Gas Turbine
G. Jorgensen 2031
Comparison of Predicted Optical Performance with Measured Results for Dish Concentrators
E. Soubassakis and Jose G. Martin 2037
Solar Thermal Energy Utilization
2.23 Line-Focus Collectors
M. Collares-Pereira, M. Gordon, A. Rabl, R. Winston 2045
A New High-Concentration Two-Stage Optical Design for Line Focus Systems
R. Almanza, R. Soriano and M. Mazari ...2051
Second Generation of Aluminum First Surface Mirrors for Solar Energy Applications
B. Rohle, M. Lazarov and R. Sizmann 2057
Low Emissive TiNxOy-Cu-Solar Selective Absobersfor High Temperature Applications
xvii
-------
O. Gutfleisch and L.K. Thomas 2063
Influence of the Substrate-Temperature of Sputtered Au-MgO-Films During Their Fabrication
on Their Spectral Reflectance and Thermal Stability
J. I. Ajona and E. Zarza 2070
Parabolic Trough Optical Efficiency Versus Assembly Tolerances. Analytical and
Experimental Approaches.
P. Schissel, G. Jorgensen and R. Pitts 2076
Application Experience and Field Performance of Silvered Polymer Reflectors
2.24 Detoxification and Materials
J. Kleinwachter, M. Mitzel, M. Wierse, R. Werner, M. Groll, B. Bogdanovic, B. Spliethoff,
A. Ritter 2085
Solar Powerstation with Thermochemical Storage System
A. Imhof, C. Suter and A. Steinfeld 2091
Solar Thermal Decomposition ofCC03 on an Atmospheric-Open Cyclone Reactor
J. Blanco, S. Malato, M. S&ichez, A. Vidal, B. Sanchez 2097
PSA's Work in Solar Photocatalytic Water Detoxification
G. Jorgensen, T. Wendelin and M. Carasso 2103
Determination of Accuracy of Measurements by SERI's Scanning Hartmann Optical Test
Instrument
M. R. Nimlos, T. A. Milne and J.T. McKinnon 2109
Gas-Phase Solar Detoxification of Hazardous Wastes: Laboratory Studies
C. S. Turchi and H.F. Link 2115
Relative Cost of Photons From Solar or Electric Sources for Photocatalytic Water
Detoxification
2.25 Solar Heat
M. B6hmer, M. Becker and M. Sanchez 2123
Development of Volumetric Air Receivers
J. Oman and P. Novak 2129
Prediction of Radiative Properties of Coal Particles Suitable for Absorption of Solar Energy in
Gas
A. Venkatesh 2135
Solar Thermal Water Pump
M. R. Amor, J. K. Raine and A. S. Tucker 2141
Double Diaphragm Solar Powered Water Pump
P. C. Lobo and M.P. Martins 2147
Performance of a Liquid Piston Engine for Low Head Pumping
V. L. Fara and R. Grigorescu 2153
Flourescent Solar Concentrators: Experimental Results
2.26 Solar Heat Posters
T. Yucheng, Z. Ruipei and Y. Tiengzhu 2161
Solar Drying for Candy Fruits
A. Hadji Saghati 2167
A New Concept on Design of Solar Water Pumping Through Open Heat Pipe
E. Azad and J. E. Mahallati 2171
Solar Continuous Corn Dryer in Fluidized Bed
xviii
-------
A. Brew-Hammond, J. Roullier and D. Appeagyei-Kissi 2177
Computer-Aided Thermodynamic Analysis of a Minimum Maintenance Solar Pump
L. Imre, G. Hecker and L. Fabri 2183
Solar Assisted High Performance Agrucultural Dryer
M. Fournier and Y. Maurissen 2189
Modelization of a Solar Fish Dryer Working by Natural Convection
M.A. Mohamad and Salah M. Khalil 2194
Solar Drying of Sand for Egyptian Glass Industry
M. Lazarov,T. Eisenhammer and R. Sizmann 2199
New Absorber Geometries for Wavelength and Angular Solar Selctive Absorber Cover
Combinations
R. Corval£n, R. Romdn and I. Napoleoni 2205
Development Of A Solar Drier-Greenhouse Prototype For Agricultural Produce
V. Heinzel, J. Holzinger, S. Petersen 2212
Boiling Water Collectors- Low Cost Tubular Collectors without Convective Heat Losses
B. Bandarsyah, N. Suharta, D. Schneller and H. Notzold 2217
Nine Years Experience with Solar Thermal Pumps in Indonesia
G. Tengesdal 2220
Solar Assisted Crop Drying A Competitive Alternative Also in Cool Humid Climates
K. Miiller and M. Reuss 2226
Solar Drying—A Survey of Different Technologies and their Influence on Product Quality
H. Dai and L. Zonghan 2232
Fully Developed Laminar Flow and Heat Transfer in The Passages of V-Corrugated Solar Air
Heater
Y. Tripanagnostopoulos and P. Yianoulis 2238
Double Pass Solar Concentrating Collector
C. A. Estrada Gasca, G. Alvarez-Garcia and R.E. Cabanillas 2243
Mathematical Model of an All-Glass Evacuated Tubular Solar Collector
2.27 Desalination
J. Hannekan and W. Rice 2253
Design Study of a Novel Solar or Waste Heat Powered Desalination System
T. Baumgartner, D. Jung, F. Kossinger and R. Sizmann 2259
Multi-Effect Ambient Pressure Desalination with Free Circulation of Air
M.A.C. Chendo and S.U. Egariewe 2264
Effects of Pebbles and Wick on the Performance of a Shallow-Basin Solar Still
E. Zarza, J.I. Ajona, J. Leon, K. Genthner, A. Gregorzewsky 2270
Solar Thermal Desalination Project at the Plataforma Solar de Almeria
C. Schwarzer, M. Wiedenfeld, N. K. Bansal and K.-H. Ert 2276
A Novel Solar Sterilization Water and Distillation System: Experiment and Thermodynamic
Analysis
E. Mitwally 2282
Feasibility of Sea Water Desalination Using Wind Energy Resources in the Middle East
2.28 Solar Ponds
W.W.S. Charters, R.W.G. MacDonald, B. P. Marett, and D.R.Kaye 2291
Micro Filtration Techniques for Effective Use of Bitterns in Salt Gradient Solar Ponds
L. Shensheng and L. Zengan 2297
Experimental Study of Mini-Saturated Solar Ponds Under Natural Conditions
xix
-------
J. Srinivasan 2303
Periodic Solar Pond For A Rural Community
G.L. Cler, P. Hartstirn, T.A. Newell 2309
Solar Ponds for Grain Drying: A Feasibility Study for Seed Corn Drying
A. Joyce, P. Raposo and M. Collares Pereira 2315
Conductivity Probe for Measuring Small Variations in Salinity Gradient of a Solar Pond
R.N. Abdel-messih and T. A. Newell 2321
Wind Erosion Modeling of Solar Pond Gradient Zones
2.29 Desalination and Solar Ponds Posters
G.L. Morrison, Sudjito, A.G. Fane and P. Hogan 2329
Solar Heated Membrane Distillation
L. Shensheng, M. Zhen and L. Weide 2335
Computer Modelling of Solar Pond With Latent Heat Storage
R. Almanza and R. Castaneda 2341
Can Any Ca-Clay be Used as a Liner for NACL Solar Ponds?
E. Sartori 2347
Evaporation From a Free Water Surface with Salt Concentration
W. G. Chun, H.Y. Kwak, T.K. Lee, S.H. Cho, and P.C. Auh 2352
Thermal Analysis on a Small Cylindrical KIER Pond
K. R. Woods and R. R. Dedolph 2358
Fertilizer Salt Charged Solar Pond Integrated with "Programmed Agriculture"
P.T. Tsilingiris 2364
The Development of a Low-Cost Flat Response Detector for Use Underwater
M. A. Javed and L. F. Jesch 2370
Monitoring An Experimental Solar Pond Fitted With Transparent Insulation
N.S. Bishena 2374
Salt Water Desalination, Using Solar Energy
S.W. Ali 2379
Comprehensive Application of Solar Stills in Pakistan
R. Prasad and D. P. Rao 2384
Enhancement of Energy Storage In The Ground Beneath Solar Ponds
J. M. Huacuz V. and M. Silis C 2390
Solar-Pond Process For The Production of Sodium Sulphate From Astrakanite
J. J. Hermosillo.... 2395
Experimental Study About Convective Solar Stills
2.30 Solar Drying I
W.W.S. Charters and S. Theerakulpisut 2403
Heat Pumps for Agricultural <6 Industrial Drying Processes
H. Farzaad 2409
A Large Scale Fish Dryer
L. Zongnan and Chen Chaoxiong .2415
The Rational Selection of Aperture Ar.eafor Mixed Mode Solar Dryer
C. Sanchez, E. L6pez, N. Arias, A. Arias 2421
Mathematical, Technic, and Economical Analysis of a Solar Dryer for Tobacco in Colombia
W. Radajewski and D. Gaydon 2527
In-Storage Solar Grain Dryer Using Rest-Periods
XX
-------
P. C. Lobo and J. W. Ribeiro 2433
Drying Data for Carnauba Leaves
2.31 Solar Drying II
M. Mahr and J. Blumenberg 2441
Modelling of Free Convection Air Flow in a Solar Dryer
O.V. Ekechukwu, B. Norton 2447
Effects of Seasonal Weather Variations on the Measured Performance of a Natural-Circulation
Solar-Energy Tropical Crop Dryer
F. Parrini, S. Vitale, A. Biondo and R. Lo Cicero 2453
ARCEL Simulation Model to Perform the Thermal Behaviour of a Storage in the Soil Below a
Greenhous
H.P. Garg, J. Prakash, D.S. Hrishikesan 2458
Theoretical Analysis of a Solar Timber Drying System
U. Luboschik, P. Schalajda, 2464
Design, Construction, and Monitoring Results of Two Medium Sized Solar Dryers in Germany
and Spain Using Natural Convection
K.S. Rao, D. Singh and S. Swaroop 2470
Thermal Design and Performance Comparison of Solar Timber Seasoning Kilns
xx i
-------
-------
2.1 Collectors I: Selective Surfaces
-------
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
-------
1141
PERFORMANCE CRITERIA FOR SPECTRAL SELECTIVE COATINGS ON SOLAR
ABSORBERS
G.Brouwer
Van Heugten Consulting Engineers
P.O. Box 305
6500 AH Nijmegen, the Netherlands
ABSTRACT
In Task X of the IEA Solar Heating and Cooling Programme, titled Solar Materials Research and
Development, the focal point of concern is to investigate and to improve material efficiency with respect
to energy performance and durability of solar buildings and solar systems. The scope of the cooperative
research described here is the establishment of performance criteria for new materials,in particular
"spectral selective coatings on solar absorbers". The research provides firstly a means of material
selection for various solar applications and locations and secondly a methodology to estimate the energy
efficiency. A guideline for material selection was developed using climate data, operating and boundary
conditions, simulation programs and related material properties. In addition,a methodology is
presented, based on a case study from the University of Waterloo, Ca,on the energy benefits and
degradation effects of selective absorber surfaces.
A database of material properties directed to the application field ofsriar energy was performed as a
support to designers and manufacturers for selecting materials.
The results, the ways and means to improve system performance and material durability respond to
requests to design and to operate efficiently with respect to environmental quality and material
conservation in the future.
KEYWORDS
Material selection; spectral selective coatings; solar hot water systems; simulation; operating conditions;
material properties; energy benefit; thermal performance.
INTRODUCTION
The excessive growth of productivity and technology and the prodigious use of materials and energy
they yielded continuously attack our environmental resources. Management is an important means to
balance the system Material, Energy and Environment, when meeting the requirements of the consumer
and the producer.
In the research of IEA Task X the following objectives were stated :
- Investigate how new materials or assemblies of materials can improve system efficiency and expand
the application of solar systems to a wider variety of needs.
- Estimate and evaluate the energy benefits of using new materials.
- Determine the necessary and quantit ative for the properties of advanced materials which
possess greater system and component performance.
Taking this into account, the study on performance criteria for new solar materials provides firstly a
means of materials selection for various locations and solar applications and secondly a methodology to
estimate the energy benefits of,in particular spectral selective coatings on solar absorbers in Solar Hot
Water Systems.
-------
1142
LOCATIONS AND CLIMATE FACTORS
The criteria to classify locations or climate types of research on solar energy applications have
concentrated on the ambient temperature and the solar radiation. They primarily affect the solar collector
performance. Because of the considered application category in this part of IEA research, viz : Solar
Hot Water and,in relation to the participating countries in IEA research,the following climate regions
from Trewartha's classification (1961) were considered : Subtropical, Temperate Continental (=
Temperate Semi-Arid), Temperate Oceanic and Boreal. Climate factors affect on the one hand heat gain
from solar systems and on the other hand they influence degradation and consequently decrease the
heat gain. The use of solar radiation in an optimal way for typical applications is the principal feature of
such a system. The climate factors are divided in two groups : the macro-climatic and the micro-climatic
factors. On behalf of performance and degradation aspects the concerned climatic factors are : solar
radiation, temperature, air pollution, Ultra Violet radiation and precipitation (rain, snow, hail).
SOLAR ENERGY APPLICATIONS AND SYSTEMS
The realisation of solar energy systems exists within a context of requirements that increase in scope
with the application (e.g. building) for which the system is designed, to the site, to the region and its
climate, up to the scale of national energy goals. Regional climate conditions may exert a strong
influence on thermal performance, designs, components and systems, but also on material selection.
The thermal performance and the economic feasibility determine whether realisation in a specific
application field is justified or not. E.g. considerable absence of direct solar radiation during summer
does not justify the installation of a solar system for cooling. The annual local climate conditions (solar
radiation), the specific annual heat or cooling demand and the governmental rules and permissions
(health, safety, fire resistance) restrict a specific solar energy application and are fundamental first
evaluation criteria before selecting components and materials in a more technical way. The selection on
technical grounds takes place by considering the operating conditions, the boundary conditions with
respect to the engineering properties and the thermal performance during life time. They were estimated
from a parameter study and a cost/gain optimisation process.
Environment and application
Energy conservation links the environment (climate) directly with the energy demand, which is in turn
related to the application. Two main aspects are of importance in relation to application and solar
system: the amount and the simultaneous occurrence of energy demand and ambient energy load over a
year. E.g. Solar Drying performs very well by using solar energy directly. Direct use of solar energy is
also applicable to Solar Water Heating in all considered climate types and during the whole year. In
general, the different climate types and application fields with their specific system (material)
temperatures during the year require an adaption of the material, component, device or solar system.
Solar systems
From the analysis of active solar systems results a general scheme, where all current systems were
incorporated ; e.g. compact systems, air- and liquid collector systems, direct and indirect heating,
gravity and forced circulation etc. It shows the big number of possibilities that components and
materials have to withstand operational conditions. The research of this report focusses on the
components which the designated materials were for: solar collector glazing and absorbers, especially
the absorber coatings. Besides,only one group of applications was extracted for a further research viz.:
Solar Water Heating Systems. For other locations, applications and solar systems, the same procedure
as described here can be considered; some parameters and conditions can be derived from this review.
-------
1143
OPERATING CONDITIONS
Operating conditions not only justify the application of solar systems in particular local regions, they
also bind the selection of materials with their properties as well. As such, operating conditions and
material properties reflect the technical evaluation criteria for selection. It is shown before how solar
materials have to match in an extended field of operating conditions. Climates, application and system
design, they all yield numerous changing parameters resulting in a multitude of differing operating
characteristics. As performance prospects of materials are much related to applications, research on
materials was narrowed down to some climates, specific application fields and thermal evaluation
criteria as presented already before. The operating conditions were categorized as follows: Thermal
performance conditions, Degradation conditions, Failure conditions. Thermal performance conditions,
divided in environmental and operational conditions,directly affect the energy benefit of the applied
device (glazing, absorber). Solar radiation, ambient temperature on the one hand and the system in its
application on the other hand define the operating temperatures as well as the thermal performance.
Depending on local and consumer circumstances,active solar systems will sometimes act as a passive
system,e.g. in no-flow or no-heat demand conditions. Under these more or less stagnation conditions
not only the performance decreases, but the materials are also severely attacked with respect to
degradation. A continuously decreasing performance during lifetime will be effected. A main
requirement for solar materials is that they shall not adversely be affected by exposure to the previous
mentioned environmental factors to an extent that will significantly impair their function during their
design lives. The changes of material properties from initial values due to degradation can be expressed
in a percentage of the yearly performance during lifetime by the use of simulation programs. To
estimate this degradation or oppositely the durability of materials by means of a prediction and
accelerated testing the time occurrences of the particular conditions have to be formulated. The current
environmental data for UV-radiation, air pollution and air humidity, can be extracted from National
Statistical Data Handbooks for each particular region. Operating conditions, which cause tremendous
changes in the functioning of solar systems or which exceed the limits that components can withstand
(e.g. stresses, decolouration, deposits, corrosion or interference with other materials), are called
Failure Conditions. The particular failure effects with restriction to thermal performance and exceeded
limits were already discussed under degradation conditions. In another part of the IEA study,Hollands
and others (1991) stated a Performance Criterium for failure of absorber coatings. This criterium
Affecting:
TABLE 1 Operating Conditions
Overall range of
Boundary
outer
inner
absorber
operating values
values
panel
glazing
(glazed)
(indication)
(indication)
(if any)
THERMAL PERFORMANCE CONDITIONS:
- Solar radiation (W/m2)
0-900
1100
o
o
- Solar spectral distribution (nm)
320 - 2500
-50, +40
o
o
o
- Ambient temperature (°C)
-20 - +30
- Service operating and stagnation temperature
10-50
80-90
o
of material (°C)
10-60
150-160
0
10-100
130 - 390
0
- Total spectral distribution (nm)
200 - 5000
o
0
0
DEGRADATION CONDITIONS:
3.5-5
- Solar UV radiation (%)
9
o
UVB radiation (%)
0.03-0.1
0.5
0
- Air impurities (fig/m3) - S02
3-185
2500
0
0
0
-N02
2-160
1500
0
0
0
- Temperature exposure (hours/year) > 80 °C
100-1000
1300-2000
0
standardized SDHW system >160°C
-
0 - 600
0
- High temperature exposure with
+20 - +27
27
RH=95 %, ambient temperature (°C) =
o
o
0
- Temperature cycling
o
o
0
- Temperature shock
20-90
o
o
0
- Relative humidity ambient air (%)
o
o
0
- Time of wetness (hours/year)
100 - 4200
4200
o
of material
100 - 3300*
3300*
O
0
FAILURE CONDITIONS:
- Performance Criterium 5% efficiency loss
0
- Structural exposure, hail, diam(mm)
20-50
100
o
1
* The number 3300 is applicable for location De Bilt, N1
-------
1144
corresponds to a 5% loss in efficiency (annual). Based on literature and IEA Task X research, the
operating conditions affecting solar collector covers and absorbers were derived and presented in Table
1. These data can be used for the prediction of service life in Carlsson's (1991) work and consequently
for the estimation of the thermal performance during life time of solar collectors and systems. In this
way materials can be selected on the basis of an optimal cost/gain ratio.
MATERIALS
The search for more promising and improved materials and the increasing level of required performance
and quality strengthen the need of accurate knowledge of the applicable materials. This means that
questions arise such as : what are the properties; how it withstands the operational and climate
conditions; what is the stability of the material during lifetime with respect to its properties; how does
the material impact the environment, etc.? On the other hand,questions arise about the methodology and
the accuracy of measurements for characterizing and evaluating them. To organize the materials and its
specifications^ classification scheme is needed. In terms of physical entities/the specification of the
materials presented in a Database Solar Materials deals with the following : physical background, incl.
the composition and kind of material; optical and thermal properties; mechanical properties and
durability.
THERMAL PERFORMANCE
To improve the thermal performance of solar systems, as well as to estimate the effect of material
degradation on the thermal performance, the coincidence of some material properties was calculated with
simulation programs. For some base cases of Solar Domestic Hot Water Systems and several climates,
graphs were derived as Fig.l. Fig.la shows fraction solar Fs versus solar absorptivity as with the
plate emissivity X held constant at 0.1. Fig.lb shows fraction solar Fs versus plate emissivity with the
absorptivity held constant at 0.95. With some more investigation for a typically sized solar system,
these graphs were combined in single plots for different climate locations. These graphs present all
combinations of absorptivity decrease (-Aa) and emissivity increase (AX) which cause a 10% decrease
in the energy delivered by the solar system. See Fig.2. Graphs as such can be performed for other
climates and system variables as carried out by Hollands and ethers (1991). Fig.3. Investigation shows
a rather proportional effect of the system performance with respect to these combined property changes.
This results in a guideline to estimate the effect of property changes on system performance caused by
degradation or by improvements on materials. For Solar Fraction of 50 % or lower the effects of solar
system settings and locations are slight.
For a first estimation of the changes on thermal performanc of solar systems caused by changes of the
material properties,the results of location Toronto can be used, Fig. 4. This figure presents also the
following methodology , which can be used for other locations and system settings :
Determine the initial properties as and X of the absorber. Determine the values a and b (intercepts) from
previous -Aa/AX plots (AFs/Fs = 0.1) for specific location and system parameters (e.g. Fig.3). Mark
these values starting from the initial properties. All points on the straight line along these 2 points
correspond to the 10% solar fraction decrease. By interpolation or extrapolation(other supporting lines
can be drawn equivalent with a Solar Fraction decrease or increase.
It is evident that due to saturation (high solar fraction) and nonlinear effects,the accuracy of this method
is acceptable for small changes only.
CONCLUSIONS
The selection of materials for solar energy applications and new energy conservation techniques
complies with the obtainment of higher efficiencies. New materials and material assemblies are required
which possess a higher averaged energy performance during life time. Two main aspects for a material
as such are distinguished:
- thermal properties of materials in operating conditions, according to the maximum energy
benefits that it brings to the application
- degradation of materials during life time, which in turn do deteriorate the annual energy
performance of the assembly in this application.
-------
1145
0.2
0.4 0.6
Plate Emissivity
0.8
0.6
0.8
0.2
0.4
Plate Absorptivity
Fig la. Fraction solar versus absorptivity Fig lb. Fraction solar versus emissivity
- Aa
0.16
BEST FIT STRAIGHT
LINE
TORONTO
J-
EFFECT OF CITY
ON -Aa vs A«
LINE FOR 'BASE
-DENVER CASE CONDITIONS:
Ac * 4.8 m2, TgpS 50*C
-ZURICH, RAPPERSWIL
-HALIFAX
0.00
SEATTLE
SASKATOON
ALBUQUERQUE
COPENHAGEN
TORONTO
MADISON
STUTTGART
J_
00 0.10 0.20 0.30 0.40* 0.50 0.60 0.0 0.2 0.4 0.6 0.8 1.0 1.2
Fig. 2. Sample plot of changes absorptivity p-g 3 £ffect 0f locations on A a and A e
(A a) and emissivity (A £)
-------
1146
Material selection. A means of material selection for various solar applications throughout the world
is provided as follows : Define climate data and criteria; select application and system; select
simulation tool; establish operation conditions (Table 1); consult database of solar materials;
establish energy performance; concern degradation effects.
Methodology for material performance estimation. A comparison of performances between
absorber coatings was established and presented in Fig.4, with an absorber coating a = 0.95 and X
= 0.1 for a SDHW system, location Toronto as a reference example.
Degradation and performance. A 10% decrease in system performance is caused by an absorptance
decrease of 0.1 or an emittance increase of 0.4;both as a bound for most severe situations (locations
and system variables).
Tentative predicates for applying materials in solar technology with respect to energy and
environment are needed.
This information contributes to the greater efficiency of use of materials with the potential of
improving energy performance and environmental quality.
Main input data:
collector area = 4.8 m2
heat storage = 180 I
daily draw-off = 3501
temperature = 50 °C
1.04 Thermal
1-02 performance
i'JX factor for
Toronto
0.96
0.94
0.92
0.90
100
I 90
e-
0
1
I 80
70
%
III
IfifelS
: ¦" T ¦''i'x''
0 20 40 60 80
Thermal emittance, %
100
Fig. 4 Thermal performance factors of SHW-systems as a function of Solar Absorptance and
Thermal Emittance of solar collectors relative to Absorptance 95 % and Emittance 10 %.
REFERENCES
Brouwer, G. (1991). Solar materials research and development. Performance criteria for new
solar materials (Draft). IEA Solar Heating and Cooling. Task X. Van Heugten, Nijmegen, NL.
Brouwer, G. (1991). Solar materials research and development. Database Solar Materials (Draft
Working Document). IEA Solar Heating and Cooling. Task X. Van Heugten, Nijmegen, NL.
Carlsson, B. (1991). Solar materials research and development. Accelerated life testing of solar
energy materials (Draft). IEA Solar Heating and Cooling. Task X. Swedish National Testing
and Research Institute, Boras, Swe.
Hollands, K.G.T, A. Karagiozis, A.P. Brunger, and G. Brouwer (1991). On the way selective
surface degradation affects solar heating performance. Proceedings ISES Congress. Denver,
Colorado, U.S.
Trewartha, G.T. (1961). The Earth's problem climates. University Wisconsin Press, Madison,
U.S.
-------
1147
The Effects of the Pulse Current Electrolysis on the Electrocrystallization of a Black
Chrome Solar Selective Coating
Tai K. Lee, S. H. Cho, Y. H. Chea, P. Chungmoo Auh
New and Renewable energy research center;
Korea Institute of Energy and Resources
P.O. Box 5, Daedeok Science Town, Daejeon, Korea
ABSTRACT
Black chrome solar selective coatings were electrocrystallized on bright nickel
substrates utilizing the pulse current electroplating method with the newly prepared
bath solution composed of chromic acid, propionic acid and the second additive. The
optical properties, absorptance (a) and emittance (e), were measured for each films
obtained at different electrolysis conditions. It was observed that a = 0.973 and e =
0.17 for as-deposited film with the condition of 3 A/cm2 peak current density, 1/10 duty
cycle and 12 min. plating time. In addition, in order to investigate the thermal stability
of black chrome selective films thermal aging tests also were performed at 300 and
450 °C in air for 24 hours. After annealing, the optical properties were measured as a
= 0.97 and e = 0.16 for coating annealed at 300 °C, and a = 0.87 and e = 0.14 for film
annealed at 450 °C. Surface morphology were observed using SEM.
KEY WORDS
black chrome; solar selective coating; pulse current(PC) electrolysis; propionic acid
INTRODUCTION
The coating of black chrome solar selective films with distinctive initial optical pro-
perties, high solar absorptance(a; typically over 0.95) and low solar emittance(e;
below 0.25), has become fairly routine, and can be obtained by a number of different
techniques. It has been known that the electrocrystallization process demands the
simplest and cheapest equipment among other coating techniques. Therefore, many
researchers have reported plating procedures and surface analysis for stable
coatings obtained from the Chromonyx bath solution manufactured by Harshaw
Chemical Company or its modified solutions utilizing the direct current (DC) elec-
troplating method (Pettit, 1976, 1982, 1983; Ignatiev, 1979; Lind, 1980; Sweet, 1982;
Smith, 1981; Driver, 1982; Holloway, 1980; Zajac, 1980).
Pettit(1982, 1983) and Sweet(1982) reported that when the electrodeposited black
chrome coatings are thermally aged in air, the solar absorptance decreases from
initial values of 0.96-0.97 to values below 0.85-0.90. This decrease causes a
correspondingly large decrease in the operating efficiency of the solar collector.
As a result of the observed coating degradation, many efforts have been directed
-------
1148
toward developing a new bath solution, and searching for an alternative
electroplating technique to improve optical properties, especially solar absorptance
or thermal stability (Smith, 1981; Pettit, 1982; Driver, 1982).
The pulse current (PC) electroplating method has proven successful in the elec-
troplating industry for providing many advantages over the DC electroplating method.
The concept of the PC electrolysis is widely applicable to many different types of
electrodeposited metals, and general advantages over the DC-applied coatings
include the following : less requirement of additional agents, enhancement of
coating uniformity, higher current efficiency, and higher quality coatings, as far as
microstructure and surface consistency are concerned (Avila, 1970; Ible, 1980). In
spite of these advantages and characteristics of the PC electrolysis it has not yet
been applied for the electrocrystallization of solar selective coatings.
Since the modified bath solutions usually have been prepared from the formula of
Chromonyx solution, their major constituents are chromic acid, and acetic acid which
plays an important role as catalyst during electrodeposition. A fluosilicic acid,
fluoboric acid or hydroflouric acid also has been used as catalyst with chromic
acid and their formula have been reported in some articles(Smith, 1981; Driver,
1981,1982). However, the use of propionic acid other than acetic acid as a catalyst
has not recieved much attention for the electrodeposition of black chrome oxide.
In this study, new chromic acid-propionic acid bath solutions have been prepared to
deposit black chrome solar selective films onto nickel substrates employing the PC
electrolysis. Comparisons of optical measured values and surface structure also were
made between coatings obtained from the PC and DC electrolysis method with the
same bath solution. In addition, thermal aging tests also have been carried out to
investigate the change in optical properties and surface microstructure due to thermal
degradation of films. Surface microstructures for the as-prepared and thermally aged
films have been investigated by SEM microphotographs.
EXPERIMENTAL
All the black chrome-oxide coatings were electrodeposited on the bright nickel
substrates of 0.1 mm in thickness from the chromic-propionic acid plating baths. The
bath temperatures were regulated within ± 1 °C by combining active cooling bath
during deposition. In this experiment, the new plating bath solution is consisted of
250 g/l chromic acid (Cr03), 10 g/l propionic acid( CH2CH3COOH ) and 0.5 g/l the
other additive. These chemicals were dissolved into distilled water (> 300 kQ) that is
the balance of the solution. The monitored pH of bath solution was 0.1 -0.15.
Since there exists various plating variables in the pulse current electrolysis process
which affect on the quality of deposited coatings, careful control of electroplating
conditions as well as chemical composition of each black chrome bath have been
accomplished by employing the standard Hull Cell test. Through repeated this test,
the experimental plating conditions were selected as summarized in Table 1. The
electroplating conditions except Run 11 were for the pulse current electrolysis. Run
11 indicates the direct current electroplating condition with the pulse off time being
zero during electrodeposition.
Because the coating thermal stability was of utmost importance, thermal aging tests
were performed using an electric muffle furnace to investigate the effects of thermal
degradation on the optical properties of black chrome solar selective coatings.
Optical measuremens for coatings both before and after aging were obtained from a
-------
1149
TABLE 1 Experimental electroplating conditions for black chrome solar
Selective coatings on nickel substrates.
Run
peak current
^on '
^off '
duty
plating
bath temp.,
No.
density,A/cm2
msec
msec
cycle
time,min
°C
1
6.6
10
150
1/16
5
18±1
3
6.6
1
15
1/16
8
20±1
4
6.6
1
15
1/16
10.5
30±1
6
3.0
10
90
1/10
7.5
27±1
9
3.0
10
90
1/10
12
15±1
10
3.0
10
90
1/10
20
22±1
11
3.0
-
0
-
6
20±1
Varian Cray 17D integrating sphere Spectrophotometer (for solar absorptance), and a
Perkin-Elmer IR 882 Spectrophotometer (for solar emittance), respectively. A Phillips
SEM 505 scanning electron microscope was used to examine the microstructure of
coatings. SEM photomicrographs of DC-applied coating and its PC electrolysis
counterparts, of the deposition of the chrome oxide thin film, were obtained as part of
the microstructure characterization study.
RESULTS AND DISCUSSIONS
Optical properties
Measured solar absorptance and emittance values for each black chrome selective
coatings were reported corresponding to each electroplating conditions as tabulated in
Table 2. Comparison of PC electrolysis data and DC electrolysis data reveals that
higher a and lower e can be obtained from new bath solution by utilizing the PC
electrolysis, and that the most of PC-mode films outperform the DC-applied coating
obtained from Run 11 in this experimentation.
As indicated in Table 2, optical characteristics of films also can be specified with the
ratio of solar absorptance to emittance, a/e, called selectivity. It is obvious that the higher
value of selectivity is, the better coating performs. Taking into account selectivity it
can be said that coating produced from Run 9 which exhibits the good initial solar
absorptance as 0.973 and its 100 °C emittance as 0.17 is the best film in this
experimentation. Coatings obtained from Run 3 also might be acceptable because of its
high selectivity, 4.65. These imply that the excellent selective coatings could be
obtained from the pulse current electrolysis under plating conditions of Run 9 or Run
3. Coatings obtained from Run 9 were selected for a detailed investigation of thermal
aging behavior. Therefore, samples of black chrome film obtained from Run 9 were
thermally aged in air at two different temperatures, 300 and 450 °C for 24 hours.
Figures 1 and 2 illustrate the typical change of the measured spectral reflectance
values as a function of wavelength for the as-deposited, and for the aged coatings
at 300 and 450 °C, respectively. After thermal aging at 300 °C for 24 hours it was
observed that there was a marginal difference of solar absorptance as 0.97 comparing
-------
1150
TABLE 2 Average current density and optical properties of
electrocrvstallized black chrome selective coatings
Run
average current
a
e
a/e
No.
density, mA/cm2
1
400
0.95
0.33
2.88
3
400
0.976
0.21
4.65
4
400
0.975
0.36
2.71
6
300
0.962
0.30
3.21
9
300
0.973
0.17
5.72
10
300
0.955
0.36
2.65
11
500
0.90
0.31
2.90
with the value of as-deposited film. However, a remarkable decrease in solar ab-
sorptance as 0.87 has been observed after annealing at 450 °C.
Meanwhile, little change in reflectance spectrum for emittance values has been
observed after heating upto 300 and 450 °C as shown in Fig. 2. Their calculated solar
emittance values were exhibited as 0.16 and 0.14 for the aged films at 300 and 450
°C, respectively. From these results it is seen that as heating temperature was increased
both the solar absorptance and emittance values decreased after aging as indicated
elsewhere with DC-mode coatings(Pettit, 1982, 1983; Sweet, 1982, 1984). The cal-
culated absorptance for DC-applied coating was 0.9 and emittance was 0.3 as tab-
ulated in Table 2.
Regarding the effect of bath temperature on the optical response values of selective
black chrome coatings as presented in Tables 1 and 2, bath temperature should be
maintained in the range of temperature from 15 °C to 20 °C during electroplating. The
pulse current plating time in the bath also must be adjusted in order to obtain the
desired optical properties. As can be seen from Tables 1 and 2, the distinctive black
chrome oxide coatings can be obtained with plating duration of 8-12 minutes.
Surface analysis
SEM photomicrographs of as-deposited black chrome coatings obtained from Run 9
and Run 11 are shown in Figs. 3 and 4, respectively. Micrograph for PC-applied
coating surface as illustrated in Fig. 3 indicates that the black chrome coating consists
of sphere-like particles with a diameter in the range 0.08 urn-0.1 urn. It was also
observed that these small spherical particles have agglomerated to form clusters
which are typically 0.3 n.m - 0.5 |i.m across.
SEM analysis of the DC deposited coatings confirmed surface microstructural infor-
mation already presented by previous workers(lgnatiev, 1979; Driver, 1982; Holloway,
1980; Zajac, 1980). As shown in Fig. 4, DC-applied deposits exhibit a relatively larger
sphere-like particles with the typical size being about 0.2 |im in diameter and a con-
sistent distribution of nodular chromium oxide growths. In addition, very few clusters are
present.
Figures 5 and 6 show SEM micrographs of the annealed coatings at 300 and 450 °C in
air for 24 hours, respectively. Although slight change in microstructure after annealing
at 450 °C was observed from Fig. 6, it can be said that there is no significant change in
-------
1151
0.25
0.20"
aged at 450 C
0.15"
0.05-
aged at 300 C
as-deposit
0.00
—i—1—r
800 1000 1200 1400 1600
—i—¦—i—1—r~
200 400 600
I
1800
Wavelength , nm
Fig. 1. Spectral reflectance properties for solar absorptance of samples plated
from Run 9 for as-deposited and aged at 300 and 450 °C in air.
0.9-
0.8"
0.7-
0.6"
0.5
As-deposit
300 C
450 C
0.4-
0.3
0.2"
0.0
0 10000 20000 30000 40000
Wavelength , nm
Fig. 2. Spectral reflectance properties for solar emittance of samples plated
from Run 9 for as-deposited and aged at 300 and 450 °C in air.
-------
1152
•1 urn*
tmr
¦
Fig. 3. SEM photomicrograph of a PC- Fig. 4. SEM photomicrograph of a DC-
applied coating from Run 9 (X 14,800). applied coating from Run 11 (X 14,800).
h- *'
N
* 1
jib
T
nssieiatf
Fig. 5. SEM photomicrograph of a ther- Fig. 6. SEM photomicrograph of a ther-
mally aged coating at 300 °C initially mally aged coating at 450 °C initially
plated from Run 9 (X 20,000). plated from Run 9 (X 20,000).
-------
1153
coating surface microstructure and grain size after 24-hour exposure to either 300 or
450 °C in air. In view of this fact heating does not significantly change the coating
particle size or shape.
CONCLUSIONS
Through detailed experimentation it has been shown that the optical properties and
microstructure of plated black chrome solar selective coatings are dependent on the
chemistry of bath solution and on the plating process variables. This study also has
shown that the chromic-propionic acid bath solution can produce black chrome solar
absorbing films with good optical properties and thermal stability by using the pulse
current electrolysis method. From results of this investigation, it was concluded that
in order to obtain the best coatings which exhibit high solar absorptance and
selectivity, the chromic-propionic acid bath temperature should be maintained between
15-20 °C and the average current density from 300-400 mA/cm2 with duty cycle from
1/10-1/16 during electrodeposition.
The further detailed study should follow in order to ensure the reproducibility of
thermally stable coatings with good optical properties. Thus it appears that the search
for optimum composition of the bath constituents and pulse current electrodeposition
procedure is essential in order to maximize the solar absorptance values while keeping
the emittance values as low as possible.
ACKNOWLEDGEMENT
Support for research from which this study evolved was provided by the Ministry of
Science and Technology, government of Korea and is gratefully acknowledged.
REFERENCES
1. Pettit, R. B. and R. R. Sowell (1976). J. Vac. Sci. Technol. , 12 , 596 - 602.
2. Pettit, R. B. , R. R. Sowell and I. J. Hall (1982). Solar Energy Mater. , Z . 153 -170.
3. Pettit, R. B. (1983^. Solar Energy Mater. ,3,349-361.
4. Ignatiev, A., P. O'Neill and G. Zajac (1979). Solar Energy Mater. , 1, 69 - 79.
5. Lind, M. A. , R. B. Pettit and K. D. Masterson (1980). J. of Sol. Energy Eng. Trans.
ASME . 102 . 32 - 40.
6. Sweet, J. N. and R. B. Pettit (1982). Optical Modeling of Black Chrome Solar
Selective Coatings. Sandia Report SAND 82-0964.
7. Smith, G. B. and A. Ignatiev (1981). Solar Energy Mater. ,4,119-133.
8. Driver, P. M. and P. G. McCormick (1982). Solar Energy Mater. ,£,159- 173.
9. Sweet J. N. , R. B. Pettit and M. B. Chamberlain (1984). Solar Energy Mater. , 1Q ,
251 - 286.
10. Holloway, P. H. , K. Shanker, R. B. Pettit and R. R. Sowell (1980). Thin Solid Films
,Z2. 121 -128.
11. Zajac, G. , G. B. Smith and A. Ignatiev (1980). J. Appl. Phvs. , 51, 5544 - 5554.
12. Avila, A. J. and M.J. Brown (1970). Plating . 58 .1105 -1108.
13. Ible, N. (1980). Surface Technology. 10.81 -104.
14. Driver, P. M. (1981). Solar Energy Mater. . 4 . 179 - 202.
15. May, R. O. , O. T. Inal, and I. H. Gundlier (1983). Solar Energy Mater. , 2 , 253 - 266.
-------
1154
IMPROVED NICKEL BASED ALLOY COATING Ft®
SOLAR THERMAL APPLICATIONS
4r "ic^c
± L. Cindrella , C.E. Sooriamoorthi and S. John
School of Energy Sciences, Madurai Kamaraj University,
** Madurai - 625 021, S. INDIA.
Central Electrochemical Research Institute,
Karaikudi - 623 006, S. INDIA.
ABSTRACT
An electroplating bath with a new complexing agent (Ammonium acetate) was
previously reported by us for the development of Nickel-Cadmium alloy solar
selective coating. Improvements in the properties of this coating by the
incorporation of an inorganic additive, Boric acid, in the plating bath are
hereby reported and analysed. Effective concentration' of the additive and
the operating conditions of this improved system have been optimized by a
systematic and powerful method, the Hull cell study. Optimized system yields
coatings with solar absorptance (a) of 0.91 and thermal emittance (.e) of 0.11.
Cadmium has been incorporated into the system to improve the corrosion resis-
tance of the coating. Metallic contents of the deposit; studied by atomic
absorption spectroscopy shows a cadmium to nickel ratio of 3.846. Tape test
and thermal stability test prove good adhesion of the coating to the sub-
strate. Salt spray analysis infers higher optical stability of the coating
in corrosive environments. Topography of the coating analysed by scanning
electron microscopy reveals grain refinement of the deposit by the incorpora-
tion of boric acid into the plating bath. This coating recorded a stagnation
temperature of 128.7°C for an average insolation of 902.5 W/m2. Based on
the promising candidature of this coating for solar applications, enhancement
in efficiencies of solar flat plate collector and parabolic collector by the
use of this coating has been simulated.
KEY WORDS
Selective coating, Electrodeposition, Nickel-Cadmium alloy, Hull cell, Bath
composition, Optical properties, Topography, Simulation.
.INTRODUCTION
Tabor (1956) reported that selective black coatings, characterized by high
solar absorptance and low thermal emittance are very effective in minimizing
the radiation heat loss occurring from the solar thermal systems. Though
several selective coatings namely black chrome (McDonald, 1975) and black
nickel (Tabor and co-workers, 1961) have been developed successfully, their
application is limited due to the higher current density and cooling require-
ments of the former and poor corrosion resistance of the latter. Hence, we
attempted to develop a durable and optically stable coating, at low current
density, with high corrosion resistance which resulted in the evolution of
Nickel-Cadmium alloy coating. Selective property of nickel and additional
property of the improved corrosion resistance imparted by the presence of
cadmium in the coating promise the suitability of this coating for long term
applications. Also, we have adopted a powerful and systematic method, the
-------
1155
Hull cell study, to optimize the electroplating bath composition and operating
conditions of the electrolyte. Effect of the additive, Boric acid, on the
coating characteristics has been studied and the change in the topography
of the coating and its effect on the optical and thermal stability of the
deposit by the incorporation of the additive has been analysed.
EXPERIMENTAL PROCEDURE
Copper panels of 75 sq.ctn area were cleaned by conventional pretreatments
and nickel plated to a thickness of 10 ym from low concentration nickel bath.
Hull cell study was carried out from an electrolyte containing 35 g/1 of nickel
sulphate, 15 g/1 of cadmium sulphate, 40 g/1 of ammonium acetate and varying
concentration of boric acid to optimize the concentration of the additive
and operating conditions of the electroplating bath to produce quality black
coatings. The chemicals used for this studies were of laboratory grades and
distilled water was used for solution preparation.
Optical properties of th e coatings were evaluated using alphatometer and
emissometer manufactured by M/S Devices and Services Co., USA. Adhesion of
the coating to the base metal was assessed by tape test. The metallic compo-
sition of the coating was studied by atomic absorption spectroscopy.
High temperature stability of the coating to withstand short-term overheating
was assessed by thermal stability test. Corrosion resistance of the coating
was assessed by subjecting the coated panels to 5% sodium chloride neutral
salt spary for a period of 88 hours.
Topography of the coating was studied by scanning electron microscopy. Stag-
nation temperature of this coating was recorded in a stagnation temperature
measurement chamber on a clear day. Under identical conditions, the maximum
equilibrium temperature attained by an ordinary black coating was also recorded.
RESULTS AND DISCUSSION
Figure, la shows the codes used for recording the Hull cell patterns. The Hull
cell patterns obtained with changing concentration of boric acid is shown
in Fig. lb. Hull cell current of 3A was passed through the electrolyte.
Concentration of boric acid was varied between 0.25 and 2.0 g/1. Lower concen-
trations of boric acid (0.25 and 0.50 g/1) produced bright bluish black coa-
tings but in a higher current density range and the coatings were slightly
spotty in nature. 1 g/1 of boric acid produced the bright bluish black coating
in a wide current density range of 4.5 - 12 A/sq.dm. Higher concentration
of 2 g/1 of boric acid shifted the occurrence of the black coating to a higher
current density value. Hence the concentration of boric acid had been fixed
at 1 g/1.
With 1 g/1 of boric acid, a bluish black coating was obtained in a current
density range of 4.5 - 12 A/sq.dm. Streaky grey coating with black speckles
was observed in the current density range of 0.6 - 4.5 A/sq.dm. White coating
was obtained in the current density range below 0.6 A/sq.dm. Hence, the useful
current density range for this system was fixed as 4.5 - 12 A/sq.dm.
Hull cell patterns were obtained with different pH values of the electrolyte
(Fig. 1c). pH of the electrolyte was changed between 4.5 and 6.0 using acetic
acid and ammonia, respectively for lower and higher pH values. Lower pH value
of 4.5 shifted the occurrence of black coating to a high current density value
of 12 A/sq.dm. Interference was observed over a wide current density range.
At pH value of 5.0, the black coating was obtained in the current density
-------
1156
range of 6 - 12 A/sq.dm. Higher pH values (5.5 and 6.0) gave rise to black
coatings at high current density >12 A/sq.dm. At a pH value of 5.3, the bluish
black coating was obtained in a wide current density range of 4.5 - 12 A/sq.dm.
Hence, the pH of the bath was fixed to be in the range of 5.0 - 5.3.
The pH of the electrolyte near the cathode increases during the process of
electrodeposition. At higher pH values, deposition of metal hydroxides occur
and hence a grey coating was obtained.
3A. HULL CELL
12 9 7^5 6 45 3-6 3 24 l"8 1-2 0-6 0-15 current
1 J * 1 «... i i i 1.1 density
•n A/dm2
Hlack
white
Grey
©
©
Greyish
Streaky
grey
53
breyis
white
(a)
Unplated
3A HULLCELLl
12 9 75 6 4-5 3-fi 3 JMI-fi 1-2 060-I5 current
6-0
TemperGture
= 20 C
30 "C
VZ'Sr'SrOiI |kh I'
(b)
Fig.la. Codes for recording Hull Cell pattern. Fig. lb. Hull Cell patterns with different
c i ,, „ ^ , Concentration of Boric acid
hgJc. Hull Cell patterns at different pH values Fiq id Hull tpii nnt* ~ A-tt
k r ig.ia, Hull cell patterns at different temperatures.
Effect of temperature on the effective current density range of the system
was studied. Hull cell patterns obtained at different temperatures are shown
in Fig. Id. Lower temperature of 20°C of the electrolyte produced spotty
black coating in a current density range of 0.6 - 12 A/sq.dm. Hull cell pat-
tern obtained at room temperature (30 ± 2°C) showed the occurrence of black
coating in the current density range of 4.5 - 12 A/sq.dm. Higher temperature
of 40°C, produced non-uniform streaky greyish black coating. Coating obtained
at higher temperature was non-uniform due to the thermal agitation of the
molecules caused by the thermal energy. Though, black coating was obtained
in a wide current density range at 20°C and also at 30°C, based on the prac-
tical conveniences and the ambient nature of the temperature, 30°C, coatings
were obtained at room temperature (30 ± 2°C) only.
Plating time played an important role in governing the optical properties
of the coatings. Lower plating periods of 5 - 10 seconds produced coatings
with absorptance value varying between 0.91 - 0.96 and emittance between 0.11
and 0.19. A plating time of 15 seconds produced black coating with a value of
0.96 but an increased emittance (e) of 0.48. Higher plating time resulted
in coatings with increased thickness (Vitt, 1987) and this resulted in high
value of emittance. The plating time was hence optimized to be in the range
of 5 - 10 seconds.
-------
1157
Atomic absorption spectroscopic analysis showed a cadmium to nickel ratio
of 3.846. Presence of higher fraction of cadmium imparts increased corrosion
resistance to the coating. Tape test proved good adhesion of the coating
to the substrate.
25KV K1003 4t57 18.TO CECRI
(a) (b)
Fig. 2. Scanning Electron Micrograph of Nickel-Cadmium selective black coating
obtained from electrolyte without (a) and with (b) the additive,
Boric acid.
Scanning electron micrograph of the coating is shown in Fig. 2b. The photo-
micrograph of the coating obtained from electrolyte without boric acid is
shown in Fig. 2a. Presence of boric acid in the electrolyte highly favours
the procurement of grain refined deposit. Periodicity of the deposit materials
is revealed in the micrograph (Fig. 2b). The particle size varies between
0.59 - 3.53 ym as compared to the particle size range of 0.59 - 9.41 um when
boric acid was not used in the plating bath.
On thermal cycling, the absorptance of the coating remained almost constant
(0.91) whereas the emittance fluctuated (Fig. 3). Emittance of the coating
increased for the first eight hours of thermal treatment and fluctuated there-
after. Higher periods of thermal treatment resulted in decrease in the value
of emittance. This is due to oxidation and recrystallization of the deposit
in the furnace atmosphere leading to reduction in thickness of the coating
which is responsible for the decreased value of infra-red emittance of the
deposit.
/Wv.
O 20 40 60 80
Heat treatment period in hours
Fig. 3. Effect of heat treatment on emittance of
Nickel-Cadmium alloy selective black coaling
O'f -
0 20 40 60 So 00
Salt spray period in hours
Fig.4. Effect of salt spray on emittance of
Nickel-Cadmium alloy selective black coating.
-------
1158
On salt spray analysis, the absorptance of the coating remained almost constant
(0.91 - 0.90), whereas emittance changed between 0.11 - 0.28 (Fig. 4). Leach-
ing of the coating in the chloride environment was not experienced during
the test period. Slight spotty appearance of the coating was observed after
88 hours of salt spray treatment. Corrosion product of the underlying layer
was never witnessed during the test-period. This coating is highly stable
in salty environment due to the presence of higher fraction of cadmium which
inhibits the corrosion of the coating and the underlying substrate.
Stagnation temperature recorded for this coating was 128.7°C for an insolation
of 902.5 W/sq.m. Under identical conditions, ordinary black coating recorded
a stagnation temperature of 107.2°C only.
REACTIONS OF ELECTRODEPOSITION
Various chemical and electrochemical reactions characterizing the process
of electrodeposition of Nickel - Cadmium coating may be summarized as below:
AT ANODE (Pure Cadmium)
Cd ^ Cd2+ + 2e~
AT CATHODE (Bright nickel plated Copper metal)
(a) Reduction of acetate ion.
COOCH3" + 6H+ + 4e" —t C2H5+ + 2H20(aq)
(b) Formation of nickel and cadmium ions
NiS04 ^ Ni2+ + S042"
CdS04 -«-* Cd2+ + S042"
(c) Formation of nickel and cadmium oxides.
Ni2+ + H20(aq) —» NiO + 2H+
Cd2+ + H20(aq) —»• CdO + 2H+
(d) Evolution of hydrogen gas.
2H+ + 2e" H2t
(e) Deposition of Nickel and Cadmium metals from their ions.
Ni2+ + 2e" » Ni
Cd2+ + 2e~ ,=* Cd
(f) Function of ammonium ion.
NH4 + OH" ^ NH3 + H20
"^(aq) + «H3(aq) — W < NH3 > 6l2+4]2t(a,)
(g) Deposition of metallic nickel and cadmium from their ammonia complex ions.
tNi (NWf(aq) + 2e"-*Ni + 6NH3
[Cd (NH3)4]f^+ 2e~ —Cd + 4NH3
From the above reactions it is inferred that Nickel-Cadmium selective black
coating obtained from the present electrolyte is a heterogeneous deposit of
metallic Ni, Cd and their respective oxides.
-------
1159
THEORY OF ACTION OF ADDITIVE
Addition agents are substances (not necessarily ingredients) which are inten-
tionally added to the plating baths in small quantities to produce beneficial
changes in the character of the deposit. Edwards (1962) proposed the dirt
mechanism for the action of additives. The additive is viewed as a piece
of dirt interferring with some process. The addition agents are generally
substances that have a high surface activity, i.e., they tend to adhere to
the surface. Increase in cathodic polarisation is usually observed. The
increase in cathodic polarisation occurs as a result of either a sharp reduc-
tion in size of the active cathode surface and hence local increase in current
density or as an increase in activation energy for the penetration of cations
to the cathode surface across the adsorbed layer. Addition agents effective
in increasing the cathodic potential has the ability to decrease the crystal
size of the deposit or to alter the type and or degree of preferred orienta-
tion. If the colloid or other substance attaches itself to the surface of
the crystal nucleus and covers it, further growth of nucleus will be prevented.
When ions are discharged, they are thus compelled to start fresh nuclei and
the result is that, the deposit is fine grained. Sometimes the deposits have
in fact been found to contain a certain proportion of coarse grained crystals
as might be expected from the fact that they tend to coat the crystal nuclei.
It is easy to understand how an excess of additive can cause the deposit to
become brittle. Hence optimization of the quantity of the additive is very
essential.
CONCLUSION
From the above studies, the optimized bath composition and operating conditions
of the present system to produce grain refined selective black coatinq are
as follows.
Scanning Electron Micrograph reveals grain refinement of the Nickel-Cadmium
coating by the presenceof boric acid in the plating bath. This coating with
desired optical properties, high corrosion resistance and good thermal stabi-
lity is_a promising candidate for efficient thermal conversion of solar energy.
It is interesting to point out that this electrolyte is of low concentration
and hence economical also. The additional merits of this coating are that
it is well obtained at room temperature and the useful current density range
for this system is very wide. Also, high periodicity of the deposit with
smaller particle size not only imparts high optical selectivity but also pro-
vides mechanical stability to the coating thus improving its life period.
Enhancement in efficiencies of solar flat plate collector (Fig. 5a) and para-
bolic collector (Fig. 5b) by the incorporation of this coating has been simu-
lated and compared with that of systems with ordinary black coating. Work
on characterization of this coating is in progress.
Nickel sulphate
Cadmium sulphate
Ammonium acetate
Boric acid
PH
Temperature
Current density range
Plating time
5.0 - 5.3
30 ± 2°C
4.5 - 12 A/sq.dm
5-10 seconds
35 g/1
15 g/1
40 g/1
1 g/1
-------
1160
IOOi
• with selective black cocrting
o with ordinary black coating
IOOi
• with selective black coating
go _ o with ordinary black coating
O 0-02 0-04 006 0-08o-io
0 0-1 0-2 0-3 0-4 0-5
(Tfi+Jf0 _Toyit inywm-2
(5itn°-Ta)/|binVw.m-2
Fig.5a. Efficiency graph of Flat Plate collector
with and without Nickel-Cadmium
alloy selective black coating.
Fig.5b. Efficiency graph of Parabolic collector
with ana without Nickel-Cadmium
alloy selective black coating.
ACKNOWLEDGEMENT
UGC, India is thankfully acknowledged for the financial support to L.Cindrella.
Authors thank Dr.S.K. Rengarajan, Director, Central Electro Chemical Research
Institute, Karaikudi, for his kind permission to carry out this work at CECRI.
REFERENCES
Cindrella, L., C.E. Sooriamoorthy and S. John (1990). Nickel-Cadmium Selective
Black Coating for Solar Thermal Equipments, Proceedings of the World Renewable
Energy Congress, Reading, U.K., pp.1377-1382.
Edwards, J. (1962). Radiotracer study of addition agent behaviour, Trans.
Inst. Met. Finish. 39, p.33.
McDonald, G.E. (1975). Spectral reflectance properties of black chrome for
use as a solar selective coating, Solar Energy, 17,p.119.
Tabor, H. (1956). Selective radiation, Bulletin Research Council of Israel,
5A, p.119.
Tabor, H., J. Harries, H. Wenberger and B. Doron (1961). Further studies
on Selective Black Coatinq, U.N. Conf. on New Source of Energy, Paper E. Conf.
35/546, Rome.
Vitt, B. (1987). Black - Cobalt Coating for Solar Collectors, Philips Tech-
nical Rev.43, No.9, pp.244-252.
-------
1161
THE NEW METHOD TO DETERMINE TECHNIC CONDITIONS OF SPUTTERED
BASE LAYER OF SPECTRAL SELECTIVE SURFACE
Zhan Yu-ying, Song Gang, Wang De-rong
Liaoniiig Province Res. Inst, of Environnental Pro. Sci.,
Shen Yang, China
ABSTRACT
In this paper the regular variable relationship between spattering technic con-
ditions (argon pressure, background vacuus) and sputtered aluaimm layer ^infrared
eaittance in base layer of Magnetron sputtered selective absorbing surface are
discussed. According to these relationships, we propose that the aethod of us-
ing the principle of lowest eaittance to determine the optiaal technic condition
of sputtered aetal base layer.
This aethod is very siaple and high speed. The spectral reflectance can be ae-
asured without use of expensive spectrophotoaeter and has no use for integral
calculation.
KEYWORDS
aagnetron sputtering; spectrally selective surface; metal base layer; argon pre-
ssure; background vacuus; infrared eaittance; lowest eaittance.
INTRODUCTION
It is to be wished that spectrally selective absorbing surface has low theraal
einittance, therefore in the infrared wave zone an enhanced reflectance aetal ba-
se layer is needed, for all practical purposes Al and Cii as this aetal base layer
are usually selected, because they are cheap and broad aateo'al on the earth. The
aetal base layer can be got by cylindrical aagnetron sputtering technique which
also can be applied in the Baking of spectrally selective absorbing surface coat
of evacuated glass tubular solar energy collectors.
In these technical parameters, argon pressure and background vacuua have an im-
portant effect 011 the character of sputtered aetal base layer. The constructon
difference between coaters, the installing place of aeasuring vacuua apparatus
and different distance froa coated part to target cathode also have effect on the
technology parameters. So the new structure coater should be tested so the be-
tter technic paraaeters can be determined. S. Craig had used spectrophotoaeter
for aeasuring ref tec tance(<2ji) when studying effects of argon pressure on the str-
ucture of aagnetron sputtered thin copper filas. Shi Yue-Yan( 5fe H Ife > also applied
planar Magnetron sputtered resource to research effects of argon pressure on the
properties of sputtered thin Al filas by using spectrophotoaeter for aeasuring sp-
ec troref lee tance curve (wavelength <25jO. In a word, technic condition of sputtered
thin Al filas has been studied by using reflectance. However, nethod of spectropho-
toaeter is coaplex, it needs large and expensive apparatus and calculation of inte-
-------
1162
gration on Measured spectroreflectance curve.
Fundamental parameter of spectrally selective absorbing surfaces is thermal emit-
tance, which depends on infrared emittance of metal base layer in wavelength of
2.5-30(i, therefore, the experimentaI investigation on the dependence of sputtered
deposi tioiial technical parameter on the infrared omittance of sputtered metal base
layer can build a connection between sputtered technical parameters and thermology
parameters of spectrally selective absorbing surfaces.
ie have used coater on evacuated glass tubular solar energy collectors to study the
relationship between infrared omittance of metal aluminium base layer and deposited
technique. So the metal base Al films with lowest omittance can be easily obtained
The coater applied by experimental investigation is a doable-cylindrical targetaa-
gnetron sputtered system (diameter 900»m, length 1800m»>, which has horizontal con-
struction. In the cathode of magnetron there are some permanent magnets for prod-
ucing magneticfieId.Diameter of target cathode is 66mm. The Al content of alumin-
ium target is >93?*'. In the vacuum chamber 34 absorbing glass tubes substrate (di-
ameter 40mm, length 1520mm) are mounted on a shelf which can rotate and revolve.
The distance between cathode and substrate is changed from 150mm to 540mm., so does
the angle of incidence of the sputtered flux.
The substrate of sample is two pieces of glass, one is hung on the absorbing tube
which can rotate and revolute, the other is hung on the connecting link which can
revo lu te.
The infrared omittance is measured by D and S model AE emi ttance-meter, the measure
wavelength is from 3 (J to 30n.Detector temperature of apparatus is 80*C. Discharge
voltage is 460V., current is ?0A, deposition time is 10 mill.
Effect of Argon Pressure on Sputtered Aluminium Films Infrared Emittance
When the fixed background vacuum is 0.7X10"'pa, the infrared emittance of spu-
ttered thin aluminium films is measured under the different argon pressure, the
result is shown in Table 1.
EXPERIMENTATION
RESULT OF TEST AND DISCUSSION
TABLE 1 Effect of Argon Pressure on Sputtered Aluminium
Films Infrared Emittance
Argon Pressure
( X 10 'pa)
67?
Al Fi la emi t tance( e ,)
Sample revolution Sample revolution & rotation
TFTffF
0. 06
0. 06
0. 0?
0.0»
or
0.8
0.07
1.0
1.6
2.5
0. 07
0. 08
0. 13
According'to Table b the relationship carve of sputtered Al film emittance and
argon pressure is graphed (Figure 1).
-------
1163
1.0"
0.8-
0.6-
tn -
0.4-
0.2-
0 1——— L i i_ i
0.5 1.0 1.5 2.0 2.5 3.0
P( xio-'pa")
Fig. 1. The relationship of the sputtered thin aluainiua
fi 1ms infrared eaittance and argon pressure.
Table 1 and Fig. 1 show that under the constant background vacuum following the
deeretion of argon pressure infrared eaittance of sputtered aluainiua fila deer-
eases, however when argon pressure is l.OXlO'pa, infrared eaittance of deposi-
ted aluainiua filas is siniaun 0. 06
-------
1164
Pi scuss ion
The reason of effect of argon pressure and background vacuua on spattered thin AL
filas infrared eaittance can be discribed by the relationship of sputtered condi-
tion and thin fila surface aicrostrueture studied by Shi Yue-Yan and Craig. Shi
Yue-Yan considers that tinder the tow Ar pressure, the deposited AL fi las'surface
is relatively smooth, reflectance is also high. When argon pressure increases,
because of low aean free path of AL atoa sputtered, high collision probability of
AL with Ar particles, eossing of energy and bevel base plant deposition, generate
aany hole cylindrical crystalline grains Al fi las surface relatively is rough and
reflectance decreases.
As we know that rough aetal surface has relativity high infrared eaittance, on the
other hand, snooth and dense surface has low infrared eaittance . Therefore, the
reason of low sputtered thin fills infrared eaittance following the decreation of
argon pressure is explained.
Craig has been studied effect of adding low levels of oxygen and nitrogen in argon
gas, and fora there are soae conical hillocks on the fila surface which cause low
reflectance, The higher background vacuua shows tow levels of iapurity gases surv-
ived in vacuus chamber when gases are evacuated, therefore, the less gases iapurity
aixed in argon at sputtering, the saoother the sputtered deposited thin Al fila is,
thus infrared eaittance is low, conversely, infrared reflectance is high.
CONCLUSION
1. In the process of Baking selective absorbing surface Al base fill by aagnetron
sputtering technique, the best gas paraaeter can be got by the aethod of aeasuring
the lowest infrared eaittance.
2. In the process of deteraing the best gas paraaeter, the aethod of aeasuring in-
frared eaittance by eaittance-aeter is better than the aethod of aeasuring the re-
flectance by spectropotoseter. The foraer has soae advantages such as convenience,
fast speed, good reappearance tow cost and popularization.
REFERENCES
S. Crain, and G. L. Harding (1981). Effcets of argon pressuer. J. Vac. Sci..
1 9, 754-755.
ife f} 16 > ( 1987 ). Base layers of spectral selective absorbing surfaces,
8 , 341 -346, (china).
-------
1165
RADIATION LIMITATION OF STAGNATION TEMPERATURE IN HIGH TEMPERATURE
SELECTIVE ABSORBERS
A.G. Monger and D. R. Mills.
Department of Applied Physics
University of Sydney
ABSTRACT
A sharp-cutoff wavelength high temperature selective surface has been
modelled, with a view to determining the stagnation temperature as well as the likely
operating temperature. It has been found that, under conditions of low optical
concentration, a solar collector receiver coated in a sharp cut-off surface may be able
to operate with acceptable efficiency at temperatures approaching 400°C. Yet,
because of a very sudden increase in losses at elevated temperatures, radiation losses
may be enough to prevent stagnation (no load) temperatures above 500 °G
KEYWORDS
High Temperature, Selective Surface, Germanium, Stagnation, Temperature Limitation,
Edge Shift.
INTRODUCTION
This paper examines the maximum stagnation temperature and maximum
performance which is theoretically attainable from a selective surface taking account
of lower absorption edge placement and the effects of the temperature variation of its
temperature dependant optical properties. A number of workers in this field have
examined semiconductors as the absorbing layer, covering a Drude metal substrate
such as copper. They have compared these theoretical and actual surfaces by the use
of a standard figure of merit defined as the solar absorption coefficient divided be the
emissivity at infrared frequencies as/ eIR. Thogerson et. al. (1982), examined textured
germanium and achieved an cts/eIR of 13, White (1989) has examined combination films
composed of a-Ge:H and a-C:H, and produced films with very low emittances but with
relatively low absorption coefficients also (a=0.65). Trotter and Sievers(1980) have
examined the problem more generally and attempted to put limits on selective surface
performance, claiming that a. J eIR figures greater than 33 with as = 0.80 cannot reliably
be achieved. Mills (1985) has revised these limits on a J em to more than 77, and
suggested that the only appropriate limit is an ideal thermally broadened selective
surface.
Many workers have chosen silicon as the semiconductor and have achieved
good theoretical performance. In all cases the limitations have come about because of
the trade-off between absorption and emissivity. Since the surfaces are principally for
the purpose of collection of solar radiation, most emphasis has been placed on the
absorption at the expense of the emissivity. Due to the more moderate slope of the
absorption edge of silicon surfaces (Braunstein et. al. 1957), this has meant relatively
high emissivities which limit the power gain of the absorbing surface (Trotter and
Sievers, 1980) at high temperatures.
-------
1166
AIMS AND PROBLEMS
The work presented in this paper is based on the work of a number of the
authors mentioned above, but the semiconductor selective absorption layer is
polycrystalline or crystalline germanium, which has been seldom investigated as the
selective absorbing layer. Yet this material exhibits the qualities required for high
temperature operation with little alteration^ shown by Zhang et. al (1991).
The high temperature properties are moderate and reasonably well understood.
Further, the intrinsic edge is at 1.86 (im at room temperature, i.e. just under the
commonly chosen 2 nm point, which is appropriate for the 300-400°C operation range.
The combination of thermal broadening and edge shift with temperature does not
cause an excessive increase in losses, under ideal circumstances, until the absorbing
surface temperature is in excess of 500 °C. This means that applications up to 350 °C or
400 °C may be possible with this material.
One major problem with such a surface is stagnation. If the solar collector is
accidentally operated under no-load conditions, the selective surface will rise to very
high temperatures which could potentially cause irreparable damage to the surface
substrate. This must be guarded against by some means, usually preventing the no -
load condition in some operational way, or constructing the collector substrate of
expensive or exotic materials which can tolerate very high stagnation temperatures.
A possibly cheaper and more elegant solution would be to design the selective surface
to suddenly increase radiative losses as its temperature rises past a preset maximum.
As will be shown in the following, this can possibly be achieved using the
phenomenon of a thermally induced absorption edge shifting to longer wavelengths
and intersecting with the emission spectrum shifting to shorter wavelengths at high
temperatures, causing a sudden increase in losses.
THEORY
The selective surface emissivity and absorptivity vary as the surface heats up.
This will naturally effect the performance of the surface, and so it is important to
understand and be able to predict how these figures will vary. Three temperature
induced effects are important; the semiconductor edge position moves toward the
infrared region, the edge slope is decreased due to thermal broadening and the
emission spectrum of the surface, which is Planckian, increases dramatically and
shifts toward the visible wavelengths.
The effect of thermal broadening is to cause a blurring of the sharpness of the
cutoff of the spectral absorption curve. This effect is modelled by equation 1,
A(A.)=A0{(l+exp(-hc(lA - lAL)/kT))_1
(l+exp(-hc(l/X - lAu)/kT))-1} (1)
where XL is the lower edge wavelength, is the upper edge wavelength, T is the
absolute temperature, h is Planck's constant, c is the speed of light in a vacuum, k is
Boltzmann's constant and A0 is the maximum value of the absorption, which is
determined by film thickness. Simply dividing both sides of (1) by A0 normalises it and
evaluating the result produces a spectral map of the ideal absorptivity of a surface.
Varying the temperature and recalculating allows comparison of the shape of the
curves and thus illustrates the effect of increasing temperature on the edge
sharpness. This is illustrated in Fig. 1 which also includes the effect of edge shift
explained below.
-------
1167
300 K
400 K
500 K
600 K
700 K
800 K
2 3 4 5 6
Wavelength (microns)
Fig 1. Absorption Coefficient For An Ideal Selective Absorber
Showing The Effects of Thermal Broadening and Edge Shift
Only the lower edge is shown, as this is the region of interest. It can be seen that
the edge slope decreases with increasing temperature, and the corners of the curve
becomes increasingly rounded. In addition, a tail extends into the infra-red emission
region of the spectrum. These effects are due to the increasing rate of interband
transitions at higher temperatures.
As the temperature of the surface increases, more energy is supplied to the
electrons in the outer shells, and the band gap decreases; this causes the edge to shift
towards the infra-red. This effect has been measured up to 300 K by Braunstein et al.
(1957), who found that at low temperatures (<200 K) the edge movement obeyed a
quartic relationship, while at higher temperatures the relationship was linear. Since
the data only extend to 300 K, an extrapolation to 800 K has been assumed. Although
this result has never been tested, the work of Macfarlane and Roberts (1955) as well as
ma^y others suggest that this is a reasonable assumption. This rate was 0.1(im per 100°C
temperature increase.
To calculate the solar absorption, it is necessary to convolve the envelope
function (1) with the Planck emission spectrum (Kreith and Kreider, 1978) given by
E(X,T)=
Ci
[exp(C2AT)-l]^5n2
(2)
where Q = 3.7415 x 108 W-nm4m"z
C, = 14387.9 jim-K
n = refractive index
The convolution is integrated over all wavelengths, and the results
dividing by the spectral integral of E(X,T) alone, i.e..
f
Jo
A(X).E(X,T) dX
normalized by
(3)
E(X,,T) dX
The denominator in (3) reduces to the black body total emissivity, becoming
-------
1168
a =-
f A(r).E(),,T) dX
Jo
OT4
(4)
where T is 5762°K and a = 5.6697 x 10" is the Stefan-Boltzmann constant.
To calculate the emissivity of the germanium at infrared wavelengths and at the
normal, it is necessary to determine the level of overlap of (1) with the reemission
spectrum. The spectral emissive power of any material or grey body is determined by
equation (2) as before but with the refractive index (n) of the material accurately
known. For germanium this value is approximately 4 - 4.2 so for the purposes of this
calculation we have chosen 4.1.
vr=f A(X.).E(A,,T) dA,
Jo
(5)
The emissivities are calculated by convolving (2) with (1) and integrating over
wavelength for a range of temperatures then normalizing by the Plank function for
the corresponding values. The results can be seen in Table 1. It should be noted that
the emissivity of the Drade metal substrate must also be added to these values. AT 800K
the emissivity for copper is 0.02 making a total of 0.047 at 800K for the germanium
copper tandem.
Table 1 Calculated High Temneratiire Optic Properties of c-Oe
Temperature (K)
Absorption
Emittance
Total :
300
0.925
0.0
0.02
350
0.930
0.0
0.02
400
0.938
0.0
0.02
450
0.947
0.0
0.02
500
0.957
0.0003
0.0203
550
0.966
0.0012
0.0212
600
0.973
0.0036
0.0236
650
0.979
0.0079
0.0279
700
0.983
0.0139
0.0339
750
0.986
0.0206
0.0406
800
0.989
0.0270
0.0470
3500-q
s
3000-1
(A
£
25004
'
W3
2000-=
©
o
o
jliLl
U
<1>
1000-E
£
O
0-
500-j
Loss Rate n=1.8
Loss Rate n=4.1
27 127 227 327 427
Temperature (Deg C)
f
527
Fig 2. Power Loss Rate Due To Increasing Temperature For Two Ideal
Selective Absorbers With Differing Refractive Indices
-------
1X69
The losses in Fig. 2 are very low, below 400°C, and show that operation at that
temperature is a real possibility at low optical concentration. Moreover, the increase
in losses is very quick with increasing temperature. This is because of a 'scissor effect'
between the absorption edge and the emission curve illustrated by Figs. 3 and 4.
The results of this is that the collector would run efficiently at up to 400°C, but would
stagnate at just over 500°C because of the sudden increase in losses and corresponding
decrease in the nett collected power. Fig. 5 shows the radiation losses caused by the
variation in emissivity of the absorbing layer alone. It can be seen that the
contribution to the losses by reradiation alone is less than 10% at 650 K but over 60% at
800 K. The advantage of this is that the surface can be mounted on glass, which is
inexpensive, without any danger of stagnation causing damage to the collector.
300
S 200 -
a"
100 -
-0.4
- 0.2
0.0
12
2
4
6
8
10
0
Wavelength (Microns)
Fig 3. Absorption Edge Plotted With Infra Red Emission Spectrum For 700 K
300
6 200 -
a"
- 0.8
- 0.6
-0.4
100 -
-0.2
0.0
10
12
2
6
8
0
4
Wavelength (Microns)
Fig 4. Absorption Edge Plotted With Infra Red Emission Spectrum For 800 K
At higher concentrations, stagnation would occur at higher temperatures, so some
-------
1170
precautions to protect the collector would be necessary if it were made of glass;
alternatively, the selective surface may be redesigned with a higher loss rate
(different edge position) at a given temperature of operation; this would at least
partially negate the effect of the optical concentration on performance, but would
allow some reduction in area of expensive selective absorber surface.
1000
s
a
cr
CO
& 800
w
o 600
X)
400
o
u
o
U
200
300
400
500
600
700
800
Temperature (K)
Fig 5. Collected Power as a Function of Temperature For The Selective
Absorber Layer Only
CONCLUSIONS
It is clear that the performance of selective surfaces can be greatly enhanced
by a prudent choice of edge position and by ensuring that the shape of the edge is as
sharp as possible. Edge position can be modified if necessary using alloy materials
(Mills, 1985). However, it now also appears that the effect of temperature on the
position and shape of the edge and on the reemission profile can be used to limit the
stagnation temperature which the collector absorber is able to reach. The collector
concentration ratio may be used as an additional parameter to 'fine tune' stagnation
performance so as not to exceed the temperature limits of inexpensive substrates.
REFERENCES
Thogerson, Paul P., Cocks, Franklin H., Pollock, John T A. and Jones, Philip L. (1982)
Journal of Materials Science. Vol. 17, pp. 1377 - 1380
White, S. B., (1989) P.hD. Thesis, University of Sydney, Department of Applied Physics
Trotter, D M., and Sievers, A J., (1980) Applied Optics. Vol. 19, No. 5, p. 711
Mills, D.R., (1985) Applied Optics. Vol. 24, No. 20, p. 3374
Braunstein, Rubin., Moore, Arnold R., and Herman, Frank., (1957)
Physical Review. Vol. 109, No. 3, p. 695
Zhang, Q C., Kelly, J. and Mills, D R., J. Apn. Phvs. (Accepted for publication)
Macfarlane, G G., Roberts, V. (1955) Physical Review. Vol. 97, p. 1715
Kreith, F., Kreider, J F., (McGraw-Hill, New York, 1978) Principals of Solar
Engineering, p. 147
-------
1171
HIGH ABSORPTANCE SELECTIVE SURFACE FOR HIGH TEMPERATURE
SOLAR THERMAL COLLECTORS
Qi-Chu Zhang*, D. R. Mills* and J. C. Kelly**
~Department of Applied Physics
University of Sydney
Sydney, NSW 2006, Australia
**School of Physics
University of New South Wales
Kensington, NSW 2033 Australia
ABSTRACT
Selective surfaces with much reduced emissivity at elevated temperatures would allow
operation at 300-400 °C in a fixed collector. Work is under way to develop a selective
surface with dramatically reduced emissivity at elevated temperatures. Optical
reflectivity measurements show that the reflectivities of Ge are dramatically reduced
in the wavelength range 0.3 to 1.4 |im after high dose oxygen ion implantation. To
explain such greatly reduced reflectivity a model has been developed to calculate the
reflectivity for high dose oxygen implanted germanium. It has been confirmed from
infared and Rutherford backscattering spectrometry measurements that a
considerable decrease (to 0-10%) in the reflectivity in the main solar spectrum is
caused by the formation of a surface layer consisting of Ge and Ge02 mixture on the
bulk germanium. For a layered structure consisting of a Ge and Ge02 mixture on Ge on
CaF2 on a Cu substrate, a low reflectivity of 0-10% in the solar spectrum is achieved,
together with a high reflectivity, near 100%, in the wavelength range 1.7 to 20 nm
from our experimentally measured and our calculated reflectivities.
A second high temperature selective surface approach being investigated uses a new
thin film interference structure. This has already achieved an ale of 46, but for
commercial reasons will be described only in the delivered paper at the Congress.
KEYWORDS
Selective surface; Optical reflectivity; Ge; Ge03 ; Sputtering; Film; Solar Energy.
INTRODUCTION
An ideal selective surface for a solar thermal converter absorbs a maximum number of
incident solar photons while simultaneously suppressing the emittance in the
infrared. The suitable absorber material is semiconductor Ge. Because Ge lias a high
refractive index in the solar spectrum, a simple homogeneous surface layer of such
semiconductor is not efficient as a selective surface. The semiconductor-insulator
composite films have been used to reduce concentration-dependent refractive index
with same energy gap. Gittleman and co-workers (1977, 1979) have studied such
composite semiconductors produced by cosputtering CaF2 and MgO with both
germanium and silicon. They reported that the dependence of performance
parameters on Ge and Si concentration is very weak, and optimum performance was
found to occur about 40 vol.% Ge in Ge-CaF2 composite films. Another method to reduce
-------
1172
refractive index is that the semiconductor layer is graded smoothly to air, then the
effective bulk absorption will take place in the solar spectrum (Ritchie and Window,
1977). Zakirov, Khaibullin and Zaripov (1985) reported that the reflection coefficient
<3ro>jspedsignificantly in the solar spectrum region for germanium irradiated to 4xl016
ion/cm2 with 80 keV Sb ion. They suggested that this is due to the formation of an
ultra-dispersed implanted layer near the sample surface associated with the formation
of special surface topography. The topography of germanium surface under self-ion
bombardment shows numerous holes in the surface at dose above 1.2x1016 ions/cm2 at
45 keV (Wilson, 1982). Recently high dose ion implantation techniques have been used
to synthesize materials whose composition and structure would not be attained by
conventional techniques. It is possible to formation of special structure, top layer is
Ge02, sub-surface the mixture of Ge and Ge02, and substrate Ge, by high dose oxygen
implantation into Ge. The refractive index of Ge02 in the ultra-violet and visible
regions is about 1.6, so the reflectivity of this structure will be reduced in the solar
spectrum.
Recently, we reported that the reflectivity of germanium in the wavelength range of
0.2 to 1.4 p.m is greatly reduced after high dose oxygen ion implantation. The
reflectivity values at about 0.7 mm are near zero for a germanium sample implanted to
a dose of 1.5xl018 O/cm2 at 45 keV (Zhang, Kelly and Kenny, 1990a).
Mills and Botten recommended that such as a structure, semiconductor-insulator-metal,
could be a candidate of selective surface for high temperature solar energy thermal
converter (Mills and Botten, 1983; Mills, 1985). As the operating temperature of a
selective absorber rises, the reradiating spectrum increasingly overlaps the spectrum
of incoming solar radiation. In such conditions, the change from low surface
reflectance in the solar spectrum to high infrared reflectance must be as rapid as
possible in order to maximize the absorptance to emittance ratio. The absorption edge
of crystal Ge at room temperature occurs at about 1.6 nm (0.8 eV) with good steepness,
near the optimal position of absorptance curve of an ideal solar energy thermal
converter operating at high temperature.
A physical model has been developed to calculate the reflectivity of germanium
implanted with high dose oxygen (Zhang, Kelly and Mills, 1990b). Using this model we
calculated the reflectivity spectra for a series of special structure. The reflectivity
curve for some special structure is very close to that of an ideal photothermal
converter operating at high temperature of 300 to 500 °C.
EXPERIMENTAL
The target for oxygen ion implantation was the single crystal n-type germanium. The
implantations were performed using a 50 keV implanter. Germanium samples were
implanted to doses 1.0-1.5xl018 O/cm2 at 45 keV. The implanted germanium samples
were characterized by several experimental techniques. The Infrared (IR) absorption
measurement was used to monitor the germanium dioxide formed in germanium by
oxygen implantation. Scanning electron microscopy (SEM) was used for micro
topography of sample surface of implanted germanium by oxygen. Rutherford
backscattering spectrometry (RBS) was used to measure depth profile of atomic ratio of
oxygen to germanium. Absolute optical reflectivity measurements were made with a
Cary 2300 spectrophotometer in the wavelength range from 0.3 to 3 jim.
THEORETICAL CALCULATION OF REFLECTIVITY
Based on the experimental results, the Ge02 was formed in the germanium after oxygen
implantation, and the oxygen implant created no significant surface features and
revealed only smooth surfaces, we constructed the following model for calculating the
reflectivity spectrum.
-------
1173
The inhomogeneous high dose oxygen implanted layer is divided into a multi-layer
system with parallel plane boundaries, each sublayer being taken as homogeneous.
The volume fractions of Ge02 and Ge, and the thickness for every thin layer are
determined with the aid of the RBS measurements. The complex refractive index of
every thin layer is obtained from the relationship between the complex refractive
index and the dielectric function of the mixture of Ge and Ge02 calculated using the
Maxwell Garnett (MG), Bruggeman (BR), and Person and Liebsch (PL) theories for the
dielectric function of composite media. Using the expressions for the complex
reflection and transmission coefficients for an absorbing multi-layer system with
parallel plane boundaries the reflectivity dependence on wavelength can be
calculated.
Two models that have been widely used to analyse the dielectric function of composite
media were proposed by Maxwell Garnett early this century and by Bruggeman in
1935. In the MG theory, the contribution to the local electric field arising from
polarized particles in the vicinity of a given particle is neglected. Recently Persson
and Liebsch (Persson and Liebsch, 1982; Liebsch and Persson, 1983) proposed a lattice-
gas model to deal with the dielectric function of composite media, which includes
interactions between the particles of the composite and takes into account the disorder
in the spatial distribution of particles. This new theory predicts both a red shift and
broadening of the dipole resonance effects which have been observed in some metal-
insulator composites.
RESULTS AND DISCUSSION
The dielectric functions of the mixtures for 0.1, 0.2, 0.3, 0.4, and 0.5 Ge in a Ge02 host
medium, and 0.1, 0.2, 0.3, 0.4, and 0.5 Ge02 in a Ge host medium have been calculated in
the wavelength range 0.3 to 10.0 nm based on the MG, BR and PL theories.
The depth profiles of the volume occupied fraction of Ge and GeOz in the implanted
samples were deduced from the RBS spectra, assuming the oxygen atoms bond with
germanium, forming germanium dioxide. From the depth profiles of volume fractions
of Ge and Ge02 for germanium samples implanted to dose 1.5x1018 O/cm2, we divided the
implanted layers into 9 sublayers within which the volume occupied fraction is
assumed to be uniform. The first layer is pure Ge02 , and the 2nd - 9th layers are 0.1, 0.2,
... 0.8 volume fraction of Ge to Ge02 . The substrate Ge is taken to be infinitely thick. In
experimental measurements the back surface was roughened to scatter the reflected
light to simulate infinite thickness.
Figure 1 shows the experimentally measured reflectivities in the wavelength range 0.3
to 3.0 jam, and calculated spectrum in the PL approximation in the wavelength range
0.3 to 5.0 (im, for Ge implanted to a dose of 1.5x10' O/cm2. We found that the reflectivity
curve calculated in the PL approximation is closer to the measured spectrum than is
the MG and BR approximations. As seen from Fig.l that the reflectivity curve
calculated in the PL approximation is in good agreement with the measured spectrum
in both shape and magnitude in wavelength range 0.3 to 3.0 nm. It is shown clearly in
Fig.l that the reflectivities are greatly reduced for high dose oxygen implanted Ge.
Such layers are of considerable practical interest for absorbing photons in solar
energy thermal converters. Low emittance may be achieved with an opaque layer of
highly reflective copper, or other suitable metal, underlying the Ge layer. Figure 2,
for such an efficient solar absorber surface, shows the reflectivity curve calculated
for structure B, of nine thin layers having the composition of a Ge sample implanted
with 1.5xl018 O/cm2, the 10th thin layer being 0.5 nm of Ge on a Cu substrate. The solar
absorptance and the thermal emittance were calculated for this structure. The solar
radiation absorbed with AM2 was used in calculating the solar absorptance for these
selective surfaces, and black body thermal radiation was assumed in calculating the
thermal emittance. The solar absorptance for this structure is 0.8 and the absorptance
-------
1174
to emittance ratio is only about 12. The thermal emittance is about 0.07 for a surface
temperature of 300 to 500 °C
0.4
R(PL)
R(EX)
.1 1 10
Wavelength (nm)
Fig. 1. Reflectivities measured experimentally and
calculated for germanium implanted to a dose of
1.5xl018 O/cm2.
0.8 -
> 0.6 -
*->
O
0)
"S 0.4-
oc
0.2 -
0.0
10
Wavelength (|am)
Fig. 2. Reflectivity calculated for structure B, Ge and
Ge02 mixture on Ge with a Cu substrate.
Emittance can be reduced further by interpolating a dielectric layer between the Ge
and Cu. For example, Fig. 3 shows a reflectivity curve calculated for a structure C, of
nine thin layers with a composition corresponding to Ge implanted with 1.5xl018 O/cm2,
the 10th thin layer being 0.5 |im of Ge and an 11th of 0.4 |im of CaF2 with a Cu substrate.
The reflectivity of this structure is very close to that of the selective surface of an ideal
photothermal converter. The solar absorptance of structure C is nearly the same as
that of structure B. The thermal emittance, however, is reduced to about one third of
that of structure B, and is about 0.025, also corresponding to temperature of 300 to 500
°C The absorptance to emittance ratio for the structure C is as high as 35, and is almost
independent of temperature in the temperature between 300 and 500 °C, unlike
conventional selective surfaces in which emissivity rises steeply while increasing
temperature. It is possible to further reduce the reflectivity in the wavelength range
-------
1175
0.3-1.2 (im by changing the thicknesses of the surface thin layers to achieve a higher
solar absorptance.
P
> 0.6-
o> 0.4-
i
Wavelength (jim)
Fig. 3. Reflectivity calculated for structure C, Ge and
Ge02 mixture on Ge on CaF2 with a Cu substrate.
1.0
100 1000 10000
Wavelength (nm)
Fig. 4. Experimentally measured reflectivity spectrum
for a sample deposited by sputtering.
Figure 4 shows the experimentally measured reflectivity spectrum in the wavelength
range 0.3 to 3.0 Jim for a film deposited by sputtering technique. The layer structure is
similar to structure C, Ge and Ge02 mixture on Ge on CaF2 on Cu on a glass substrate.
Argon gas was used for sputtering Cu and Ge layers, argon and oxygen gas mixture for
layers of Ge02, and of Ge and Ge02 mixture. The absorptance of the film, calculated from
the reflectivity spetrum with AM2, is 0.88.
NEW THIN FILM SURFACE DEVELOPMENT
The physical model (Zhang, Kelly and Mills, 1990b) developed to explain experimental
results for the above semiconductor selective surface has been recently used to
calculate the performance of candidate thin film interference surfaces which might
-------
1176
also exhibit low emissivity. This has resulted in a new selective surface structure
which may achieve improved performance over the thick film approach.
An early sample of this surface has achieved a room temperature emissivity of at most
0.02 with a normal absorptance of 0.92, for an a/e of 46. This figure may be
conservative, as the emissivity of the surface at room temperature was identical to the
sputtered copper standard used, and our assumption of 0.02 for the standard is a higher
figure than many in the literature, where figures between 0.011 and 0.017 are
common. For an emissivity of 0.017, a/e would be 54, but this is not claimed as we have
as yet no absolute verification reflectance of the copper standard used.
Details of the thin surface structure cannot be revealed at this stage for commercial
reasons, but will be given in the delivered paper at the Congress.
While both thick and thin selective surface approaches should result in low loss high
temperature fixed collector systems, the thin film approach would, if successful, be
much easier and less expensive to deposit.
ACKNOWLEDGEMENTS
This project was supported by grants from the Australian Research Council and Dr.
Zhang is currently supported from a grant from the Energy Research and
Development Corporation. We wish also to acknowledge the financial assistance of His
Royal Highness Prince Nawaf Bin Abdul Aziz of the Kingdom of Saudi Arabia through
the Science Foundation for Physics within the University of Sydney. We are indebted
to the Australia Institute of Nuclear Science and Engineering for the provision of
experimental facilities.
REFERENCES
Gittleman, J. I., B. Abeles, P. Zanzucchi, and Y. Arie (1977). Optical Properties and
Selective Solar Absorption of Composite Material Films. Thin Solid Films. 45, 9.
Gittleman, J. I., E. K. Sichel, and Y. Arie (1979). Sol. Energy Mat.. 1, 93.
Liebsch, A., and B. N. J. Persson (1983). Optical Properties of Small Metallic Particles in
a Continuous Dielectric Medium. J. Phvs.. C16, 5375.
Mills, D. R., and L. C. Botten (1983). Lower Emissivity Limits Indicated for High
Temperature Selective Surfaces. Appl. Opt.. 22, 3182.
Mills, D. R. (1985). Limits of Solar Selective Surface Performance. Appl. Opt.. 24, 3374.
Persson, B. N. J., and A. Liebsch (1982). Optical Properties of Inhomogeneous Media.
Solid State Commun.. 44, 1637.
Ritchie, I. T., and B. Window (1977). Applications of Thin Graded-index Films to Solar
Absorbers. Appl. Opt.. 16, 1438.
Wilson, I. H. (1982). The Effects of Self-ion Bombardment (30-500 keV) on the Surface
Topography of Single-crystal Germanium. J. Appl. Phvs.. 53, 1698.
Zakirov, G. G., I. B. Khaibullin, and M. M. Zaripov (1985). Transformation of the
Structure and Optical Properties of Germanium Bombarded with Heavy Ions. Phvs.
Chem. Mech. Surf.. 2, 3110.
Zhang Qi-Chu, J. C. Kelly and M. J. Kenny (1990a). Germanium Implanted with High
Dose Oxygen and Its Optical Properties. Nucl. Instr. Meth.. B47, 257.
Zhang Qi-Chu, J. C. Kelly and D. R. Mills (1990b). Optical Studies of Germanium
Implanted with High Dose Oxygen. J. Appl. Phvs.. 68, 4788.
-------
1177
TIN OXIDE COATED SELECTIVE ABSORBER SURFACES WITH EXTREME THERMAL,
CHEMICAL AND MECHANICAL STABILITY
A. Roos*, E. Wackelgftrd* and G. Chinyama**
*Dept. of Technology, Uppsala University, Box 534, S-751 21 Uppsala, Sweden
**Physics Department, University of Zambia, P.O. Box 32379, Lusaka, Zambia
ABSTRACT
Tin oxide coated nickel pigmented anodized aluminium surfaces have been studied. Fluorine doped
tin oxide was pyrolytically deposited on heated substrates, and the resulting film is very hard and
chemically inert. The tin oxide film not only protects the anodic layer from chemical degradation
when exposed to both acidic and alkaline solutions, but also provides protection against mechanical
abrasion. The high temperature stability is excellent and the surface degrades very slowly at 450°C.
After as much as ten days at this high temperature, only small changes in the optical reflectance can
be noticed.
Contrary to most other selective surfaces the optical properties are nearly independent of the
thickness of the anodic and the tin oxide layers. Instead the selectivity depends mainly on the
doping of the tin oxide and the pigmentation of the anodic layer. Both deposition techniques are
suitable for large scale industrial deposition.
Solar selectivity is limited by the high refractive index (n = 1.7 - 2.0) of the tin oxide and cannot
quite match the best selective coatings on the market. However, a and e values of 0.92 and 0.17
can be achieved, and in view of the durability of this surface, this is of technical interest.
KEYWORDS
Anodized aluminium, Pyrolytic tin oxide, Solar selective surfaces, Durability, Accelerated aging
INTRODUCTION
Several selective solar absorber surfaces have been suggested over the last decades. Nearly all are
based on an oxide on an infrared reflecting metal. The oxide is either grown from the metal or
deposited in a chemical or a vacuum process (Garg, 1982). Well known examples are black nickel
(Mason, 1982), black chrome (Zajac, 1982), black cobalt (Hutchins, 1982), black copper (Roos,
1983) and anodized aluminium (Andersson, 1980). Low cost, high selectivity and high durability
are desired properties of these surfaces. Unfortunately it does not seem possible to optimize all
these properties simtitaneously. Increasing the selectivity usually means increasing the cost and/or
reducing the durability. In this paper a selective surface is presented which is not optimized as
regards the optical properties, but is instead incredibly resistant to accelerated aging. This surface is
based on nickel pigmented anodized aluminium which has been coated with infrared reflecting
pyrolytic tin oxide (Roos, 1991a).
-------
1178
SAMPLE PREPARATION
The aluminium substrates were anodized in a 2.5 M H3PO4 solution at room temperature giving a
porous and amorphous layer of aluminium oxide (Keller, 1953). Anodization can be carried out in
other acid solutions but the phosphoric acid gives the highest volume fraction of pores which is
advantageous in this case. This is because in the subsequent electrolytic process the pores are filled
with nickel in a buffered NiSC>4 solution. The anodization is a dc process and the pigmentation is a
ac process. The resulting film is a cermet with nickel particles embedded in a matrix of aluminium
oxide. This surface is an excellent selective absorber surface and with proper optimization of the
process parameters it is possible to obtain a and hemispherical e values of 0.95 and 0.13
respectively (Wackelg&rd, 1990).
In order to increase its durability this surface can be coated with tin oxide using the pyrolytic
spaying technique. The surface is heated to 400 - 450 °C in an oven and a solution of tin chloride in
alcohol is atomized and sprayed onto the heated surface, where the tin oxide is formed (Manifacier,
1979). The tin oxide is crystalline and, if doped with fluorine atoms, the infrared reflectance can be
as high as 90% for a 0.7 - 0.8 (xm thick film. The tin oxide is transparent in the visible and the light
is absorbed by the black anodic layer underneath. The absorption is limited by the high refractive
index of the tin oxide giving a front surface reflectance of more than 5 % in the visible wavelength
range.
OPTICAL ANALYSIS
The reflectance spectra in the visible wavelength region were recorded on a Beckman 5240
spectrophotometer equipped with an integrating sphere, and on a Perkin Elmer 983 in the infrared.
For some of the samples the true hemispherical emittance value was measured at 100 °C using a
calorimetric method. This value is around 30 % higher than the near normal value obtained from the
infrared reflectance measurements. This is the case for most kinds of selective surfaces, and should
be kept in mind when emittance values are quoted in the literature.
ACCELERATED AGING TESTS
Accelerated aging tests were used to establish the durability of the surfaces. The objective of these
tests was to compare the uncoated anodized surface with the tin oxide coated, and not to get a
realistic estimate of the life time of the surfaces (Roos, 1991b). The tests were therefore very tough
in order to degrade the surfaces within a short time.
The thermal stability was tested in an ordinary oven at a temperature as high as 450 °C. At this
temperature the surfaces degraded slightly over a period of one week. Most selective surfaces
degrade rapidly at a much lower temperature than this.
Humidity and condensation cause problems in all absorber designs, except those where the
absorbing surface is kept in vacuum. In particular this is the case when the humidity contains acids
or other substances from environmental pollution or from degassing within the absorber. To test
the resistance against chemical degradation the anodized surfaces were placed in solutions of
sulphuric acid, acetic acid and sodium hydroxide. These are very corrosive solutions and the test is
far more aggressive than any natural condensation the selective surface can be exposed to.
Testing the mechanical strength of a hard coating on a soft substrate is not easy. An investigation of
tin oxi4e on aluminium as compared with anodized aluminium is presented elsewhere (Hedenqvist,
1991). In this paper an abrasive wear test was used where the sample was held against a rotating
-------
1179
abrasive wheel of 10 cm diameter. The rotating speed was 50 - 100 rpm and the load varied
between 50 and 250 g. A small container with a slurry of aluminium oxide powder and water could
be placed under the rotating wheel to perform a more abrasive wet test.
RESULTS AND DISCUSSION
Reflectance spectra of two tin oxide coated anodized surfaces are shown in Fig. 1. The two
samples are identical apart from the thickness of the anodic layer. It is interesting to note that the
absorption of this surface is relatively insensitive to the thickness of the anodic layer. It can also be
shown that this is also true for the thickness of the tin oxide. More relevant parameters to the optical
properties are the fluorine doping of the tin oxide and the nickel pigmentation of the anodic layer.
100
Co- 80
uj
O 60
z
<
h
UJ 40
-j
u_
Ui
cc 20
0
Fig. 1 Reflectance spectra for tin oxide coated nickel pigmented anodized aluminium surfaces.
100
Annealing 450 °C
///
anodized Al
UJ
o
z
<
before test
2 days
h
O
UJ
_l
///
5 days
U.
Ui
cc
10 days
16
8
2
4
0.4
0.8 1
WAVELENGTH (|im)
Fig. 2 Reflectance spectra for anodized nickel pigmented aluminium before and after annealing
Reflectance spectra before and after annealing at 450 °C are shown for the anodic surface in Fig. 2
• SnQ,
anodic
layer
IK
u 0.7 urn
'/
ULl J
1/
i-
mivi
/
;/
d = 1.0 nm
a = 0.91 , e = 0.16
¦ — d = 0.4 |xm
a =0.88 ,e =0.16
_l I I 1 L-
0.4 0.8 1 2 4
WAVELENGTH (um)
16
-------
1180
and for the tin oxide coated surface in Fig. 3. It is clear that both surfaces are very stable at elevated
temperatures. It is especially interesting to note that in both cases the absorption and the emittance
values increase and, since the absorption is the more important quantity, this actually corresponds
to an improvement of the selectivity. The increased absorption in the near infrared wavelength
range is possibly due to a redistribution of nickel within the pores. This also causes the shift of the
reflectance step at around 4 |im for the anodic surface. The similar shift for the tin oxide coated
surface is due to a slight reduction of the oxygen vacancy doping and a corresponding shift of the
Drude plasma frequency.
100
^ 80
UJ
O 60
<
uj 40
-i
u.
in
K 20
0
0.4 0.8 1 2 4 8 16
WAVELENGTH (^m)
Fig. 3 Reflectance spectra for tin oxide coated Ni-pigmented anodized aluminium before and
after annealing.
Annealing 450 °C
SnCL / anodized Al
before test
2 days
5 days
todays
100
HAc - test
Anodized Al
80
before test
60
Sn09/ Anodized Al
40
before test
after test
20
0
j.
X
0.4
WAVELENGTH (nm)
Fig. 4 Reflectance spectra for uncoated and tin oxide coated nickel pigmented
anodized aluminium before and after acetic acid corrosion test.
-------
1181
100
NaOH - test
Anodized Al
before test
UJ
O
z
after test
SnOg/ Anodized Al
O
Ul
before test
—I
Li.
after test
Ul
EC
0.4
0.8 1
2
16
4
8
WAVELENGTH (|im)
Fig. 5 Reflectance spectra for uncoated and tin oxide coated nickel pigmented
anodized aluminium before and after sodium hydroxide corrosion test
The chemical stability is quite remarkable for the tin oxide coated surfaces. Tin oxide is chemically
very inert and deposited on a glass substrate it resists almost any acid or alkaline solution. When
deposited on the porous and softer anodic surface it is obviously vital that the tin oxide film is free
of cracks and pinholes, otherwise the corrosive solution would penetrate through the film and
attack the anodic layer. The results from the tests where the films were exposed to acetic acid and
sodium hydroxide are shown in Figs. 4 and 5 respectively. It is clear that the tin oxide almost
completely protects the surfaces from this type of chemical degradation. Alkaline condensation is
not very likely to appear in a solar absorber, but as both aluminium and anodized aluminium are
known to be very sensitive to NaOH, this test was performed to demonstrate the excellent
protective properties of the tin oxide coating. Protection against acetic acid is sometimes important
since many sealant materials contain small amounts of acetic acid. Inside the sealed box the
concentration can be high enough to cause degradation unless the absorber surface is resistant to the
acidic humidity. 100
jS 80
ui
^ 60
<
h
O
Ul 40
u.
Ul
c
20
0
0.25 0.5 1 2 4
WAVELENGTH (>im)
Fig. 6 Reflectance spectra for black nickel pigmented anodized aluminium before and after
abrasive test
Dry abrasive test
before test
5 min
Anodized Al
30 min
-------
1182
100
s1 80
ui
2 60
<
h
O
y 40
u.
HI
cc
20
0
0.25 0.5 1 2 4
WAVELENGTH (,um)
Fig. 7 Reflectance spectra for tin oxide coated nickel pigmented anodized aluminium before
and after abrasive test.
The results of the abrasive tests are shown in Figs. 6 and 7. In this case a dry test was performed
where the rotating wheel was kept dry and the sample exposed for 5 and 30 minutes. In Fig. 6 it
can be seen that the wear of the anodic layer was considerable during the test. In particular the shift
of the interference peaks after 30 minutes indicates that the thickness of the anodic layer has been
considerably reduced. The tin oxide coated surface in Fig. 7 has not changed much during the same
test. Similar results were obtained for a wet abrasive test using a slurry of aluminium oxide and
water.
SUMMARY
It has been shown that pyrolytic tin oxide deposited on nickel pigmented anodized aluminium
considerably increases the durability of the surface. The thermal stability of the surface is excellent
both for the tin oxide coated and for the uncoated anodized surface. The chemical and mechanical
stability are greatly improved by the tin oxide.
REFERENCES
Garg, H.P. (1982). Treatise on Solar Energy. John Wiley, London.
Andersson, A. Hunderi, O. and Granqvist, C-G. (1980). J. Appl. Phvs.. 51. 754.
Hedenqvist, P. and Roos, A. (1991). Surf. Coat. Technol.. Accepted publication.
Hutchins, M.G., Wright, P.J. and Grebenik, P.D. (1986). Proc. SPIE. 653. 188.
Keller, F. Hunter, M.S. and Robinson, D.L. (1953). J. Electrochem. Soc..l00. 411.
Manifacier, J-C., Szepessy, L., Bresse, J.F. Perotin, M. and Stuck, R. (1979). Mat. Res. Bull..
14, 109 and 163.
Mason, J.J. and Brendell, T.A. (1982). Proc. SPIE. 324. 139.
Roos, A. and Karlsson, B. (1983). Solar Energy Mat.. 7. 467.
Roos, A., Georgson, M. and Wackelg&rd, E. (1991a). Solar Energy Mat.. Accepted publication.
Roos, A. and Georgson, M. (1991b). Solar Energy Mat.. Accepted publication.
Wackelg2rd, E., Chibuye, T. and Karlsson, B.(1990). Proceedings North Sun 90. Reading, UK
Zajac, G. and Ignatiev, A. (1982). Appl. Phvs. Lett.. 41. 435.
_ — 1
1
i _
Dry abrasive test
SnO / Anodized Al
— — 5 min
2
-
//
A
//
-
//'
—VVV
jj
— 1
i
•
-------
2.2 Collectors II
1183
TEMPERATURE DISTRIBUTION IN
THE GLAZING ON SOLAR COLLECTORS
Svend Svendsen, Finn Kristiansen
Thermal Insulation Laboratory, Technical University of Denmark
Building 118, DK-2800 Lyngby, Denmark
ABSTRACT
An analytical model of the temperature distribution in the glass sheet and frame of a solar col-
lector has been made.
Two different collectors have been in stagnation at a solar irradiance of about 800 W/m2, and the
temperatures of the glass sheet and the frame have been measured.
The calculated and measured temperatures have been compared and found in relatively good
agreement.
KEYWORDS
Thermal stress in glass on flat plate collectors, modelling, tests.
INTRODUCTION
In Denmark non-tempered glass is normally used in collectors because of its lower price. The
experience has been good, but sometimes the glass will break because of the high temperature
difference between the center and the edge.
In solar collectors the cover will often have a quite large temperature difference between the
center area and the edge area. In stagnation the center of the glass sheet may rise to 60-70°C.
The frame will reach temperatures in the range from the air temperature to the glass temperature
depending on the level of insulation of the frame. The edge of the glass sheet is in thermal contact
with the frame and will, therefore, have a temperature somewhat above the frame temperature
depending on the details of the glass-frame construction.
These relatively large temperature differences in the glass cover will, due to thermal expansion,
cause stress in the edge.
If the edge is without any defect, the ordinary glass sheets will typically withstand temperature
differences of 40°C. For sheets with small defects in the edge the limit is lower.
When ordinary (non-tempered) glass is used, it is important that the collector box is designed in
such a way that the temperature difference in the glass sheet lies well below the critical tem-
perature limits.
-------
1184
DESCRIPTION OF THE INVESTIGATED SOLAR COLLECTORS
Two different collectors have been investigated. The collectors were about 2 m2 and they were
supplied with absorbers of the SunStrip type. The insulation thickness of the back was 50 mm,
and the insulation thickness of the edge was 15 mm.
Both collectors are supplied with a non-tempered glass sheet mounted in a frame of an extruded
aluminium profile.
The major difference between the two collectors was the design of the frame, see figure 1 and 2.
glazing bead
glazing bead
glass
absorber
insulation
glass
absorber
insulation
Fig. 1. Frame design of collector 1.
Fig. 2. Frame design of collector 2.
THE ANALYTICAL MODEL
In the analytical model an energy balance for the glass at the centre of the collector is made. This
is made traditionally as described in the book of Duffie and Beckman (1980). From the energy
balance the glass center temperature and the heat transfer coefficients in the air gab and at the
outside of the glass are calculated.
The temperature variations in the glass from the center to the edge are found by solving a set of
differential equations based on energy balances of small segments of the glass sheet, see figure
1-3, the glazing bead and the frame.
The edge segment of the glass sheet situated below the glazing bead is supposed to have a uni-
form temperature. An energy balance for this segment is set up taking the energy flows to the
segment from the glass sheet, the frame and the glazing bead into account.
The energy balance for a small segment of the glass sheet will be
». dT
9&9dx
L. dT
+ hi(Ta-T)Ax-h0(T-T0)Ax + S&x = 0
-------
1185
centerline
A.v
4* *
••
%
%
1
[ ~
L,/2
«S A.v
dx | x
Fig. 3. Energy balance for a small segment of the glass.
When dividing by Ax and letting A.v approach zero you will have:
d2T 1
dx2 kge:
•[(/ii + /i0)r-/iiTa-h07VS]
The two boundary conditions, necessary to solve this second-order differential equation, are
symmetry at the centre line and a known root temperature, Tg.
dT
dx
= 0 and T | x_0 = T £
x-L„/2
The root temperature Tg corresponds to the glass temperature at the edge.
If the following definition is made
2 hi + h-0
k g6g
the general solution will be
T = Cjsin h(_mx) + C2cosli(mx) -
h(T a + h0T 0 + S
h(T a + h0T 0 + S
where C, = C2 tan h(mL_/2); C2 = T
hi + h0
-------
1186
The temperature profile of the glass is shown on figure 4, Tg is set equal to 56 °C.
0.50 m
glass center
Fig. 4. The temperature profile of the glass sheet.
An equation for the energy conducted to the edge of the glass is then made:
cLT
-------
1187
The temperatures of the glass, glazing bead and the frame have been measured by thermocouples
and by infrared sensing techniques (AGA Termovision). The results of the measurements are
shown in table 1.
TABLE 1 Results of the measurements
Collector
Measuring points
Thermocouple
Infrared
Average
°C
°C
°C
Glass centre
79
81
80
1
Glazing bead top = glass edge
57
55
56
Frame
52
52
52
Temp, difference in glass
22
26
24
Glass centre
78
75
77
2
Glazing bead top
55
51
53
Glass edge
48
-
48
Frame
51
49
50
Temp, difference in glass
30
-
29
It was impossible to measure the glass edge temperature by means of the termovision equipment
because of the glazing bead.
The glass edge and the glazing bead top temperatures are probably identical for collector 1, be-
cause the glazing bead is very thin (0.7 mm) and very close to the glass
When measuring the temperatures for collector 2 the glass sheet broke, probably because of the
large temperature difference between the glass centre and the edge (29 °C).
COMPARISON OF CALCULATED AND MEASURED TEMPERATURES
The temperatures of the two collector designs have been found by calculation and measurements.
The results are shown in table 2.
TABLE 2 Measured and calculated temperatures
Collector 1
Collector 2
Measurements
Calculations
Measurements
Calculations
°C
°C
°C
°C
Glass centre
80
75
77
75
Glazing bead top
56
56
53
51
Glass edge
56
56
48
53
Frame
52
52
50
50
Temp, difference in glass
24
19
29
20
-------
1188
The calculated temperatures of the glass centre are 2-5 °C lower than the measured temperatures.
When making calculations it is difficult to find the heat transfer coefficient in the air gab and at
the outside of the glass. This is probably the reason of the lower calculated temperatures at the
glass centre.
The complicated geometry of collector 2,especially the geometry of the glazing bead, and the
connection to the glass has given some problems when making the analytical model. In the model
the length of the top of the glazing bead and the glass edge segment are the same and are in a
uniform thermal contact. This assumption is not very good for collector 2. That can be the ex-
planation why the measured and calculated temperatures of the glass edge show a difference of
5°C.
CONCLUSION
The analytical model has shown its usefulness even if some deficiencies must be admitted. The
model is less accurate for complicated frame profile designs. In the design of new or improved
collectors it will be possible to calculate the temperature distribution of the glass sheet. The
calculations ought to be supplemented with measurements on prototype collectors. By means of
these tools more reliable collectors based on non-tempered glass can be developed.
REFERENCES
Duffie and Beckman, 1980. Solar Engineering of Thermal Processes. John Wiley & Sons, New
York.
SYMBOLS
eg
: glass thickness
hi
: heat transfer coefficient in the air gab
h0
: heat transfer coefficient at the outside
kg
: glass conductivity
Lg
: width of the glass
S
: absorbed solar energy
T
: temperature
Ta
: absorber temperature
To
: temperature of the surroundings
Ax
: width of elemental region
-------
1189
INVESTIGATION OF FLAT-PLATE MONOLITHIC SILICA AEROGEL COLLECTORS
A Nordgaard* and W.A. Beckman**
*Norwegian Inst, of Technology, University of Trondheim, Norway
**Solar Energy Lab, University of Wisconsin-Madison USA
ABSTRACT
It has recently been experimentally shown (Jensen, 1989) that the flat-plate collector efficiency can
be significantly improved by filling the air gap between absorber and cover with monolithic silica
aerogel (MSA), and evacuating the system to 0.1 bar. The objective of this study is to model MSA
collectors and introduce the necessary quantities that enables MSA collectors to be treated as
ordinaiy flat-plate collectors.
KEYWORDS
Monolithic silica aerogel; flat-plate solar collector; transmittance; heat loss coefficient; theoretical
model.
INTRODUCTION
Monolithic silica aerogel (MSA) is highly transparent in the solar part of the spectrum (0.3-3.0 |J.m).
Furthermore, the material has a very low thermal conductivity (0.008 W/mK at 20°C evacuated to
0.1 bar), and is a substantially better insulator than still air (k=0.024 W/mK). The heat loss
reduction due to the low thermal conductivity is obtained without a dramatic reduction of the
transmittance-absorptance product, (ra). For medium and high temperatures the solar collector
efficiency can be significantly improved by using a 20 mm thick MSA cover.
When using MSA in collectors the expenses of selective absorber coatings and convection
suppressing devices can be eliminated and the cost effectiveness of the solar heating system
improved (Jensen, 1989).
CALCULATION OF THE EFFECTIVE TRANSMITTANCE ABSORPTANCE
PRODUCT OF MONOLITHIC SILICA AEROGEL
When solar radiation enters the MSA slab, a fraction of the incident energy is transmitted through
the material without being attenuated (direct-direct transmittance), and a part is removed by
scattering and further absorption. A portion of the scattered radiation is backscattered and a portion
is transmitted through the slab (direct-diffuse transmittance). It has been shown that the scattering
in MSA may be assumed to be isotropic, as long as the cover thickness is less than 50 mm
-------
1190
(Nordgaard, 1991). Furthermore, no correction for surface reflections will be required in the
analyses that follows because the index of refraction of MSA is close to unity, being between 1.01
and 1.05 depending on the density.
The monochromatic direct-direct transmittance, Tdir-dir. is defined by Bouguer's law. The spectral
direct-diffuse transmittance is given by the multiple scattering within the MSA. Analysis of multiple
scattering is usually mathematically very complex and requires a great deal of computational effort.
However, a new and fast method for isotropic multiple scattering within an absorbing and scattering
medium has been developed.
The F ("F-hat'"> Concept ^
Beckman (1971) defined a total exchange factor, F, between pairs of surfaces of an N surface
enclosure. This methodology has been extended to an absorbing and isotropically scattering
medium. Consider a 1-dimensional plane parallel system of optical thickness Kp divided into n
equal elements each of optical thickness Ak, with n+1 surfaces. The factor Fvi,sj is the total
exchange factor between volume element i and surface j. A general expression for Fvi,sj for a
system divided into elements is given by
Fvi,sj — Fvi.sj + fii Fvi,vk Fvk.sj
k=l
(1)
where FVjiSj is defined as the fraction of the energy isotropically leaving volume i that directly
impinges on surface j, that is, without being scattered or absorbed along the way. The factor,
Fvi,vj, is defined as the fraction of energy first scattered in element i that is attenuated in element j,
and Qj denotes the albedo for scattering, that is, the ratio of scattered radiation to scattered plus
absorbed radiation. An expression similar to Eqn 1 can be written for every combination of
elements and surfaces. The total set of equations for the F factors of n elements and n+1^ surfaces
results in a set of n(n+l) linear equations with n(n+l) unknowns. After evaluating the F factors,
the resulting reflectance and transmittance of the slab can be calculated by
n ^
P = (Fvk,sl £2k Iatt,vk)/Io (2)
k=l
n ^
t= ^dir-dir + (Fvk,s(n+1) ^k Iatt,vk)/l0 (3)
k=l
where Iq is the primary beam intensity, and Iatt,vk is the attenuated energy in element i, and is equal
to the decrease in the primary-beam intensity in element i. Eqns 2 and 3 can also be used to
determine the absorption, a, within the slab, by applying the following relationship:
a = 1-x-p (4)
The monochromatic angular transmittance,x^_(0) and reflectance, p^(0), values can be calculated in
the usual way, with k replaced by and £ by The total values are then obtained by
integration over the solar spectrum (0.3-3 [im) for each incidence angle.
-------
1191
Effective Transmittance-Absorptance Product
A new method, which allows the inclusion of scattering layers, based on the embedding technique
(Edwards 1977) was developed and used to calculate the transmittance-absorptance product. All of
the solar radiation that is absorbed by a cover system is not lost since this absorbed energy tends to
increase the cover temperatures and consequently reduce the losses from the absorber plate. In
order to maintain the simplicity of the Hottel-Whillier collector equation, an effective transmittance-
absorptance product (xa)e, was evaluated. The procedure described by Duffie and Beckman (1980)
was applied for this purpose. Figure 1 illustrates the difference between (xa)e and (tot) versus
incidence angle for an absorber plate temperature of 50°C.
1.0
0.8
0.6
0.4
« 0.2
0.0
Me
(rex)
\
. \
0 10 20 30 40 50 60 70 80 90
Incidence angle, 0
Fig. 1 (tot) and (xa)e versus incidence angle .
Incidence Angle Modifier
The incidence angle modifier approach used in conventional flat-plate collectors was found to be
valid for MSA collectors. For a 4 mm thick glass cover, a 20 mm thick MSA cover, and a black-
painted copper absorber the incident angle modifier coefficient, bg. was found to be -0.21 which is
about the same as a 2-glass cover conventional collector.
General Remarks
Effective incidence angles of isotropic diffuse solar radiation has also been evaluated. The
procedure is similar to the one described by Brandenmuehl and Beckman (1980). The effective
incidence angle for diffuse radiation from the sky is generally 5° lower for MSA than ordinary
glazings. This is a result of the scattering occurring in MSA at short wavelengths.
Svendsen (1989) has reported on a more transparent MSA (TMOS). The total transmittance at
normal incidence for this MSA type has been measured to 0.9 for a 20 mm thick sample, which is
4% higher than the type (TEOS) used in this study. However, spectral transmittance values for
TMOS have not yet been reported, and the TEOS type was therefore used.
-------
1192
CALCULATION OF THE INFRARED HEAT TRANSFER IN COLLECTORS BASED
ON MONOLITHIC SILICA AEROGEL
Heat transfer through MSA is due to conduction and infrared radiation. The conduction occurs both
in the skeleton and in the air in the pores. Convection in the very small pores can be ignored. MSA
absorbs and remits radiation, but shows no scattering in the infrared. MSA is partially transparent in
the infrared between 3 and 5|0.m. Due to large spectral variations in this region, the radiative
transport will not be a local phenomenon anymore, and direct radiative communication between the
boundaries may occur. Consequently, the radiative transport strongly varies with MSA thickness
and the emissivities of the boundaries. In this case, the coupling between the radiation field and the
heat flux caused by conduction has to be considered. To obtain an exact solution for the combined
conduction-radiation energy transfer in an absorbing-emitting (and scattering medium), the general
energy equation must be solved.
Calculation Procedure
The F-concept was modified to be applicable for the infrared heat transfer. The effects of boundary
reflections and emissions from MSA were included in Eqn 1 (Nordgaard 1991). The temperature
distribution within the collector cover was found by expressing the conduction term by finite-
differences, and solving the simultaneous nonlinear equations by Newton-Raphson. The inclusion
of property variations (MSA was treated as a non-gray medium) did add some complexity to the
functional form of the equations. The total exchange factors had to be solved on a monochromatic
basis, and the integrals within the radiation terms had to be evaluated in each iteration. Once the
temperature distribution is known, the heat transfer across the collector cover from the absorber to
the surroundings can be easily calculated by means of an energy balance on one of the boundary
surfaces.
Collector Overall Heat Loss Coefficient
The collector overall heat loss coefficient was evaluated using the procedure described by Duffie and
Beckman (1980). Results of the calculation of the overall loss coefficient for absorber plate
temperatures of 50,70 and 140°C are presented in Table 1.
TABLE 1. Calculated Overall Heat Loss Coefficient, Ul
Absorber plate temperature (C) 50 70 140
UL(W/m2°C) 1.26 1.37 1.73
COMPARISON WITH OTHER COLLECTOR DESIGNS
Instantaneous collector efficiencies are shown in Fig. 2 for three different collector designs: (1) one
cover with selective absorber plate; (2) evacuated flat-plate (HVL 20); (3) MSA collector. In all
cases, the incident radiation on the collector was 800 W/itA
It is clearly seen that considerable improvements of the efficiency can be obtained by using MSA
-------
1193
collectors for high temperature applications. The efficiency is better than the evacuated flat-plate
collectors for temperature differences larger than about 65°C. Even if the thermal performance of
one collector exceeds another collector over the actual temperature range for the application, cost of
the systems must be considered before a final judgment is possible.
CONCLUSION
The introduction of the transparent insulation material monolithic silica aerogel in flat-plate collectors
seems to be very promising. It is shown that the collector efficiency can be significantly improved.
Especially for medium and high temperature applications, the system performance can be increased.
The price of MSA is today so high that it has only been used for research purposes. However, it is
expected to become so low that it will be the same as -the price of a selective coating (which is
omitted in MSA collectors).
1.0
P- 0.8
I °"6
<*—
CD
O 0.4
o
05
o
O 0.2
0.0
0 20 40 60 80 100 120 140 160
(Ti-Ta)[oq
Fig. 2. Collector efficiency as a function of the difference between collector fluid inlet temperature
and ambient temperature.
—
(1
— - (2
) Collector with selective surface
) Evacuated flat-plate collector
) MSA collector
(3
'Nj
s
S
S
\
*N» ,
\
\
s
""S.
N» ^
¦v
\
N
N.
REFERENCES
Beckman, W.A. (1971). The Solution of Heat Transfer Problems on a Digital Computer, Solai
Energy. 13. 3.
Brandenmuehl, M J. and W.A. Beckman (1980). Transmission of Diffuse Radiation Through CPS
and Flat-Plate Collector Glazings. Solar Energy . 24.511.
Duffie, J. A. and W.A. Beckman (1980). Solar Engineering of Thermal Processes.
Edwards, D.K. (1977). Solar Absorption by Each Element in an Absorber-Coverglass Array.
Jensen, K.I. (1989). Transparent Cover with Evacuated Monolithic Silica Aerogel. Transparent
Insulation Technology for Solar Conversion. Proc. 3rd Int. Workshop, 58.
Nordgaard, A. (1991). PhD Thesis, University of Trondheim, Norway.
Svendsen, S. (1989). Solar Collector Based on Monolithic Silica Aerogel, ISES Congress, Kobe,
Japan.
-------
1194
DEVELOPING A GENERAL MODEL OF THE
WIND DEPENDENT HEAT LOSS OF A FLAT SOLAR RECEIVER
*G. Angermeier, **S. Harrison, *A. Richter and *Ms H. Soltau
"Lehrstuhl Prof.Dr.R.Sizmann, Ludwig-Maximilians-Universitat, 8000 Munchen 40, FRG
"Solar Calorimetry Laboratory, Queen's University, Kingston, Ontario, Canada K7L3N6
ABSTRACT
Combined measurements of wind field and heat transfer over a flat plate both outdoors and
indoors are presented, proving the enhancement of heat transfer rates due to free-stream turbu-
lence. In a first approach the forced convection heat transfer coefficient is modelled with wind
speed and angular velocity of the wind direction accurate within 1 W/m2K. Consequences for
the testing of solar receivers are discussed. Differences between indoor and outdoor measure-
ments stress the need for further investigations into the effect of frequency distribution of the
free-stream velocity fluctuations on heat transfer rates.
KEYWORDS
Uncovered collector; collector test; heat loss; forced convection; wind; turbulence.
INTRODUCTION
Every solar receiver is affected in its energy balance by wind induced forced convection at
its outer surface. However, so far no systematic uniform model of the forced convection heat
transfer coefficient exists, which fits the flow conditions and the dimensions that are relevant
for solar energy applications:
1. The flow field of the ambient air at the ground level of an urban environment is highly
turbulent and instationary. We observe turbulence intensities, i.e. changes in the wind
speed magnitude, between 35% and 50% \ and fluctuations around the mean flow
direction between 10°/s and 40"/s1.
The free-stream turbulence affects the heat transfer rate significantly. The experimental
data we present in Fig. 1 and Fig. 2 are enhanced between 0 % and 250 % above the case
of undisturbed parallel flow. The variation of the turbulence quantities determines the
large scattering of the data.
2. The spatial extension of flat solar radiation receivers comprises the large range between
0.1 m2 ( single solar panels ) and 1000 m2 ( building surfaces, large collector arrays ).
The state of the art [ (Kowalski, 1976), (McCormick, 1984), (Simonich, 1978) ] has been di-
scussed in detail in (Soltau, 1990). In addition we want to emphasize a recent publication of
Macijewski and Moffat (1990) who presented a very interesting summary of indoor measure-
ments over a broad range of turbulence levels and geometric structures.
'recorded over a flat area of 10 m2 with a sample rate of 1 Hz and an averaging period of 1 min
-------
1195
In the paper we report on a systematic experimental investigation of the forced convection
heat transfer rate of a flat area in turbulent flow. A heat transfer model is aimedfiu...subject
to the following specifications:
• simple, i.e. a minimum number of variables
• comprehensible, i.e. refering to heat transfer theory
• accurate to 1 W/m?K
• restricting to variables which are recorded by a pl&in measurement equipment, adapted to
longterm outdoor or indoor tests. The wind sensors are mounted to the heat exchanging
area in a distance of about 10 cm. Correlating the heat transfer coefficient with the
flow field in front of or significantly above the receiver involves a host of uncertainties
depending on the near-by surroundings and the oncoming flow direction.
The field of application of the model covers the identification of the relevant wind field quan-
tities in experimental investigations. The comparability of outdoor and indoor test results of
( wind sensitive ) solar receivers is aimed at. Technical guidelines are necessary for the gene-
ration of adequate turbulent flow conditions in environmental chambers. Comprehension of
the heat transfer mechanisms is essential for the optimum, construction of heat exchanging
surfaces. For simulation proposals will be made for the estimation of the wind field quantities.
THE PROJECT
(1) Outdoor measurements: At the Collector Test Facility of the University of Munich
we record the natural wind field above a 3.5 x 3.5 m2 uncovered collector array. In a distance
of 10 cm from the test area, direction and velocity of the horizontal flow are determined in
a spatial resolution of 90 cm and a time resolution of 1 Hz. We use cup anemometers and
wind vanes with a low inertia. Statistical wind field variables, describing the dynamics and
the structure of the flow field, are drawn from the data. They are correlated with the average
forced convection heat transfer coefficient of the plate which is determined from the energy
yield of the uncovered collectors. The experimental set-up is described in (Angermeier, 1987).
(2) Indoor tests: In October 1990, heat transfer measurements have been performed indoors,
at the Canadian National Solar Test Facility in Toronto. A 0.16 m2 heat flux plate which was
centred in a plywood plate of 2.6 m2 has been used. During the test the angle of attack
of the oncoming flow on the plate has been changed from —15° to 90°. The flow field in
the environmental chamber has been designed to resemble the natural wind field. A free-
stream turbulence level between 15 % and 55 % is generated by horizontal slats in front of
the ventilators and the development of the free jet in the surrounding air. Using a Ultrasonic
Anemometer (Kaijo Denki Co., Tokyo, Japan) we have been able to determine the air flow in
its vector components with a time resolution of 20 Hz.
(3) Iteration of the indoor tests: A second experiment in the Solar Test Facility of
Canada is planned with a 2.5 m2 heat flux plate approximating the dimensions of the outdoor
test. The experience of the first test series will enter the new experiment.
(4) Wind sensor test: A detailed analysis of the quality of the diverse wind sensors
has been ongoing for about a year, with a series of experiments indoors and outdoors. Cup
anemometers, wind vanes, an ultrasonic and thermal anemometers have been tested.
The main results axe summarized in (Richter, 1991). A detailed reader with guidelines for the
measurement of low-velocity high-turbulent wind fields is planned.
-------
1196
(5) Iteration of the outdoor measurements: A resumption of the outdoor experiments is
in preparation. Instead of the uncovered collectors heat flux plates of various sizes will be used.
They permit a certain local resolution of the heat transfer coefficient and a better accuracy
in the single data point. Additionally, the heat flux plate will be mounted on a rotating pillar
improving the control of the flow conditions on the test area. A relevant objective of this
Section is the analysis of the influence of the characteristic length scale of the plate on the
heat transfer rate, as expected from classical heat transfer theory. Moreover, the use of the
ultrasonic anemometer will allow a detailed analysis of the influence of the spectral distribution
of the wind field fluctuations.
(6) Simulation: Regarding simulation,work is under way on the estimation of the wind field
quantities from data provided by meteorological stations.
FIRST RESULTS
The first outdoor measurements were completed bf the end of 1988. Preliminary
results have been published in (Soltau, 1989) and at the International Heat Transfer Confe-
rence in Jerusalem (Soltau, 1990). The first indoor tests were performed in November
1990. The analysis of the heat transfer data is not yet finished. Outstanding is a detailed
analysis of the influence of the frequency spectra of the wind field fluctuations and the specific
characteristics of separation bubbles. New questions are put to the outdoor measurements.
However, .at the moment three dominant features prevail:
1. The enhancement of the forced convection heat transfer rate in a highly turbulent flow field
is large compared to undisturbed parallel flow conditions.
We observe an augmentation between 0 % and 250 % for our measurements, in- and outdoors.
Fig. 1 and Fig. 2 show the value of the forced convection heat transfer coefficient (hfc) in
dependence of the wind speed (uw). All quantities are averaged over the heat exchanging area
and a time period of about 30 min, presupposing stationary flow conditions. The value (uw)
gives the horizontal wind speed magnitude above the test area. For comparison the equivalent
heat transfer rate in an undisturbed flow field (h/c ~ Kharuw8) 's inserted in the Figures
(Kays, 1980). The data stress that considering an additional turbulence quantity is essential.
2. The average flow velocity and the angular velocity of the fluctuating flow direction sort the
heat transfer data.
Fig. 3 shows the heat transfer coefficient {hjc) in dependence of the product (ojuw) for the
open air measurement. The angular velocity (w) of the flow direction
^ N-1
" s (jv-i)At ij A1 = ls M
has been calculated from the data = 1,7V} recorded by the wind vanes with a 1 Hz
sampling rate. The magnitude of (to) pictorially reflects the proportions of the turbulent fluc-
tuations by summarizing the amplitude and the frequency of the flow direction changes. The
introduction of the angular velocity reduces the broad scattering of the data in Fig. 1 to the
error in measurement of about 0.7 W/m2K. A simple linear fit 2
specifies the dimensionless quantity divided by Si-Units
-------
1197
yields a standard deviation of 0.8 W/m2K. Equation 2 demonstrates an equal ability of the
average wind speed (uw) and the turbulence (w) to enhance the heat transfer rate. The angular
velocity varies between (10°/s < w < 40 °/s) for our measurements.
The same fit has been applied to the indoor measurements. Restricting data which show the
same magnitude and dynamics as the outdoor data 3 gives equivalent values for the parameters
4" = 6-4Jif + 0-148J¥"*»- <3»
and the standard deviation. The error in measurement is estimated ast 0.3 W/m2K.
However, ^ the entire measuring range shows a considerable variation in the relation
between (h/c) and (to uw), as visualized in Fig. 4. A linear fit reveals a scattering of 2.2 W/m2K.
3. The heat transfer rate is more sensitive to turbulent fluctuations in the high than in the low
frequency range.
The scattering of the data in Fig. 4 reduces to 1.1 W/m2K, if the sampling rate of the wind
direction measurement is changed from 1 Hz to 20 Hz. The time resolution of 20 Hz has been
enabled by the use of the ultrasonic anemometer. A large value of the highly resolved angular
velocity (50 °/.s < fi < 250 °/s) reflects that the turbulent fluctuations as they are experienced
by the local flow field are dominated by fast fluctuations with a frequency {v > 1 Hz).
The sensitivity of the heat transfer rate to the high frequency changes is as well demonstra-
ted by the good correlation between (hfc) and the turbulence intensity (TU(v>2Hz) Uw), as
visualized in Fig. 5. The turbulence intensity (Tu^^hz)) is defined as the relative standard
deviation (crUw) of the horizontal wind speed magnitude (uw)
Tu(u>2jjz) — , in restriction to the frequency domain (2 Hz 2Hz uw) yields a standard deviation of 0.5 W/m2K.
That value is significantly enlarged if low frequency fluctuations are included in the turbulence
intensity. For (0.1 Hz < v < 10 Hz) we calculate a standard deviation of 2.2 W/m2K. The
runaway data belong to backward ventilation causing a large amount of separation bubbles.
DISCUSSION OF THE RESULTS.
The present results support the feasibility of developing a plain model of the forced convec-
tion heat transfer coefficient in a turbulent flow field, as specified in the introductory part.
Nevertheless some relevant open questions have to be answered:
1. Though to be expected,the dominant sensitivity of the heat transfer to high frequency
fluctuations has to be proven for field conditions. Part 5 of the project includes measu-
rement with the ultrasonic anemometer.
2. Though a linear fit of the heat transfer coefficient is good, it plainly lacks reference to
heat transfer theory or the limiting case of vanishing free-stream turbulence. The general
validity of the parameter values is unlikely in the present form, as Fig. 4 indicates.
The dependence on the characteristic length scale of the heat exchanging area (lchar)
will be investigated. Part 3 and Part 5 include the variation of the plate size.
3to be precise 3.5 °/s < DiTu < 19°/s (Soltau, 1990)
-------
1198
20
15
1 O
5
0.5 1 1.5 2 2.5 3 0.5 1 1.5 2 2.5 3
Fig. 1.' Outdoor measurements Fig. 2. Indoor measurements
xx
15
xX
10
5
Xx'
uw[m/s}
x sr
x x *
• x xx
*XX x
B<
x» „
&X*
%
wum["/s * m/s]
20
IS
16
14
12
10
hMmm
u)uw[°/s * m/s]
2 O 40 60 SO
20 40 SO SO
Fig. 3. Outdoor measurements
Fig. 4. Indoor measurements
-------
1199
3. Aiming at a simple measurement procedure correlating the heat transfer coefficient with
the changes in the wind direction in 1 s time steps as it is measurable with a wind vane
with a low inertia is interesting. Yet, the context between the angular velocity (w) and
the spectral distribution of the wind field fluctuations is unclear. An alternative is the use
of a simple ultrasonic anemometer as provided e.g. by Thies (Thies GmbH, Gottingen,
Germany). Such a device is included in the investigations of Part 4.
18 -
*
X
x X
1© "
X
X
X
X *x
XX
X
14- "
X X
v & X
** x*xXj<
X x
X
12. "
>
10 -
s -
\
X x x x
X
xx *
X
X
Tuv>2H*Uw[m/s]
0.05 O.I 0.15 0.2
Fig. 5. Indoor measurements
REFERENCES
Angermeier, G., R. Pitz-Paal, and Ms H. Soltau (1987). The wind dependent heat transfer coefficient
of uncovered collectors. In Proc. of the ISES Solar World Congress. Hamburg (Germany).
Kays, W.M. and M.E. Crawford (1980). Convective Heat and Mass Transfer. McGraw Hill, New
York (U.S.A.).
Kowalski, G.J. and J.W. Mitchell (1976). Heat transfer from spheres in the naturally turbulent,
outdoor environment. J. of Heat Transfer, Trans, of ASME, 98:649-653.
Maciejewski, P. K. and R. J. Moffat (1990). A correlation for boundary layer heat transfer accounting
for free stream turbulence. In Proc. of the 9th IHTC Congress, pages 303-308. -Jerusalem (Israel).
McCormick, F.L. Test, and R.C. Lessmann (1984). The effect of free-stream turbulence on heat
transfer from a rectangular prism. J. of Heat Transfer, Trans, of ASME, 106:268-275.
Richter, A. (1991). Measuring and Characterizing Highly Turbulent Air Flows. Master's thesis,
Ludwig-Maximilians-University Munich, Lehrstuhl Sizmann, Amalienstr. 54, 8000 Miinchen 40
(Germany).
Simonich, J.C. and P. Bradshaw (1978). Effect of free-stream turbulence on heat transfer through a
turbulent boundary layer. J. of Heat Transfer, Trans, of ASME, 100:671-677.
Soltau, H. (1989). The Thermal Performance of Uncovered Collectors. Fortschrittsberichte der VDI-
Zeitschriften, Reihe 6: Energietechnik, VDI-Verlag, Diisseldorf (W-Germany).
Soltau, Ms H. and G. Angermeier (1990). Modelling the heat transfer of a flat plate in the highly
turbulent wind field of an urban environment. In Proc. of the 9th IHTC Congress, pages 151-156.
Jerusalem (Israel).
-------
1200
MEASURED TOP HEAT LOSS COEFFICIENTS FOR FLAT
PLATE COLLECTORS WITH INNER TEFLON COVERS
Yu Yiqint, K.G.T. Hollands*, A.P. Brunger*
* Solar Thermal Research Laboratory, Dept. of Mechanical Engineering,
University of Waterloo, Waterloo, Ontario, CANADA, N2L 3G1
tDept. of Second Mehanical Engineering, Hebei Institute of Technology, Tianjin, China
ABSTRACT
Recent work has shown that the equations often used for characterizing the heat transfer
across air layers significantly underpredict the heat transfer when the air layer is bounded
on at least one side by a thin plastic film. The reason is that the equations were derived
assuming both faces of the air layer are isothermal, whereas the film temperature can vary
spatially. The dearth of experimental data on this problem has prompted the present study,
which reports over 400 measurements of the thermal conductance of air layers of various
thickness having various temperatures at their faces, and have FEP films (either 25 /irn or 13
jjrn thick) spaced at various locations inside. Comparisons are made between the measured
conductances and those predicted using the Hollands and co-workers correlation equation;
they show mean bias differences of 10%. Comparisons are also made with using a modified
form of their correlation equation.
KEYWORDS
Flat plate collectors, heat conductance, Teflon films, plastic films, free convection.
INTRODUCTION
It is well known that one can decrease the thermal losses from a flat plate collector by adding
one or more inner glazings of thin transparent FEP (Teflon) film, provided one keeps the
various air layer thicknesses at appropriate values (roughly 10 mm or more). For example,
calculations indicate that two such inner glazings will reduce this top loss coefficient to as
low as 1.8 W/rn^K. Since these films are highly transparent to solar radiation (their solar
transmission for normal incidence is typically 0.96), their presence increases the collector
efficiency even at modest values of the collection temperature.
In order to justify the extra cost of the films, it is highly useful to quantify precisely the
top loss coefficient with them present. Methods have been developed (Hollands and Wright
1983, Edwards and Rhee 1984, Wijeysundera and Iqbal 1991) for calculating this coefficient,
using algorithms that account for the coupling of radiant and convective exchanges across
the layers, including the effect of infra-red transmission and absorption by the film as well
as non-gray effects. But recently the equation that these methods ordinarily incorporate to
characterize the free convective heat transfer across individual air layers has been shown to
be slightly inaccurate.
The convection equation of Hollands and co-workers (1976), assumes that each face of the
air layer is isothermal. Hollands and Wright (1983) demonstrated an important difference
between predictions based on using that equation and their own measured heat transfer
across two horizontal layers separated by an FEP film; in particular they observed a critical
Rayleigh for the air layer about 40% lower than would be expected. They attributed this
difference to the fact that the FEP film is not necessarily isothermal; convective cells are free
to establish slight variations in the film temperature.
Subsequently Catton and Lienhard (1984), Lienhard (1987), and Richards and Edwards (1989)
carried out stability analyses of two or more horizontal fluid layers separated by solid layers,
-------
1201
with the last workers incorporating the effects of radiation. For the very thin solid layers,
a condition approached closely by the commonly used 25 fim thick FEP film, the critical
Rayleigh number for one of a pair of identical layers is found to be 1296 for no radiation,
and the effect of radiation is expected to raise this slightly. On the other hand, the value for
isothermal boundaries, (and the value embedded in the Hollands and co-worker's equation),
is 1708. In addition, Lienhard and Catton (1986) treated the heat transfer across two fluid
layers separated by a solid, and found that the initial rise in Nusselt number Nu with Rayleigh
number Ra is more steep than with isothermal boundaries. In particular, in the equation
usually used to characterize the initial rise,
Nu = k{ 1 - Rac / Ra) (1)
the value of k (for very thin layers) should be 1.77 rather than 1.44, the value embedded in the
Hollands and co-workers' equation. Catton and Lienhard were able to show good agreement
with measured heat transfer of Hollands and Wright (1983) and Ulrich (1984).
This important analytical work described in the previous paragraph was restricted to horizon-
tal layers. For vertical layers, on the other hand, Wright and Sullivan (1988) have observed
quite reasonable agreement between measured and predicted heat transfer using algorithms
incorporating free convective correlation equations based on isothermal layers. There is, at
present, no information relating to tilted layers.
For this reason it was decided to undertake an extensive experimental study measuring total
heat transfer coefficients across tilted, horizontal, and vertical air layers containing FEP films,
covering a large range of bounding temperatures, film thickness, air layer thicknesses, and tilt
angles. This paper reports the results, and compares the data with predictions of the Hollands
and Wright method using various equations for the free convection coefficient.
METHOD
The apparatus of El Shirbiny, Raithby, and Hollands (1982) was used for the measurements.
It contains two parallel copper plates: a hot plate and a cold plate. Between these plates is
the air layer with films. The hot plate contains three imbedded plates, (guarded hot plates)
each heated electrically, for heat flow measurement. For the present experiments, the cold
plate had a 1.5 mm thick neoprene gasket (thermal conductivity = 0.134 W/m ICj laminated
to it, to simulate a glass sheet in emissivity and (to some extent) in thermal resistance. The
hot plate had a similar neoprene gasket laminated to it, and laminated on the outside of that,
a 50 (im thick copper sheet (selective surface foil). This copper sheet had an emissivity of
0.10, simulating the selective surface in the collector. The thermal resistance of these two
plate coverings were measured, and subtracted from subsequent measured overall thermal
resistance of the total layer. The heat fluxes from the three (vertically aligned) guarded hot
plates were averaged to give the heat flux across the layer. Generally the central plate flux
was within 2% of the average flux.
A total of 400 conductance measurements were made in all. Measurements having one film
separating two air layers were made for film thicknesses 25.4 jim and 12.7 /jm; for five tilt
angles, ranging form horizontal to vertical, for six combinations of air gap thicknesses, ranging
form 10 to 21 mm, for temperature differences ranging from 10 to 50°C, and for cold plate
temperatures ranging from 5 to 35°C. Other measurements have been done with two layers
of film and three air gaps, and they are also reported.
Based on the work by El Shirbiny (1980) it may be concluded that the accuracy (i.e. the
bias error) in the heat transfer measurement was quite small, less than 1%. But because of
various random effects, the precision error of the measurements is substantially greater. To
determine the precision, ten of the measurement (including setting the FEP film in place)
were repeated, several weeks after the initial measurement. The per cent difference in these
measurements was 6.7% if one includes one possible outlier having a difference of 19%, and
4% otherwise. From this we concluded that the precision of the measurements is about 5%.
RESULTS
All the results for the single film of 25.4 fim thickness bounding two air layers are given in
-------
1202
Table 1. Table 2 gives the results for a single film of 12.7 \im thickness separating two air
layers. Table 3 gives the results for two films separating three layers.
Figure 1 was prepared to assess the error in using the equation of Hollands and co-workers
(1976) in the algorithms of Hollands and Wright (1983) for predicting the total layer con-
ductance U. It plots the calculated conductance versus the measured conductance for each
measurement point. The prediction is better for the 25 fim film (mean bias difference 8.4%;
rms difference 10%) than for the 13 fj,m film (mean bias difference 13.6%; rms difference 14%).
Thus it seems that this procedure for predicting U will produce errors of the order of 10%.
Figure 2 shows similar plots using a similar prediction procedure except with the Hollands
and co-workers equation changed by replacing the coefficient 1708 by 1296 and the coefficient
1.44 by 1.77, making the equation
This equation is a phenomenological correction to the former equation, based upon the work
of Lienhard and Catton (1986). The comparison in Fig. 2 is for only those layers containing
just one film. Use of this equation (2) has reduced the mean bias difference in estimating U,
from 8.4% to 2% for the 25 fim film, and from 13.6% to 7.7% for the 13fim film. The rms
difference is 9.2% in both cases.
ACKNOWLEDGEMENTS
The authors are indebted to the government of the People's Republic of China for providing
one of us (Y.Y.) with an overseas research fellowship, to Energy Mines and Resources Canada
for providing financial support, and to Mr. Bert Habicher for technical help.
REFERENCES
Catton, I., and J.H. Lienhard, (1984), "Thermal Stability of Two Fluid Layers Separated by
a Solid Interlayer of Finite Thickness and Thermal Conductivity", ASME Journal of Heat
Transfer, 106, pp. 605-611.
Edwards, D.K., and S.J. Rhee, (1984), "Nongray Radiative and Convective Conductive Ther-
mal Coupling in Teflon-Glazed Selective-Black Flat Plate Solar Collectors", ASME Journal
of Solar Energy Engineering, 106, pp. 206-211.
El Shirbiny, S.M., (1980), "Heat Transfer by Natural Convection Across Vertical and Inclined
Air Layers", Ph.D. Thesis, Department of Mechanical Engineering, University of Waterloo,
Waterloo, Canada.
El Shirbiny, S.M., G.D. Raithby, and K.G.T. Hollands, (1982), "Heat Transfer by Natural
Convection Across Vertical and Inclined Air Layers", ASME Journal of Heat Transfer,
104, PP- 96-102.
Hollands, K.G.T., T.E. Unny, G.D. Raithby, and L. Konecek, (1976), "Free Convective Heat
Transfer Across Inclined Air Layers", ASME Journal of Heat Transfer, 98, pp. 189-193.
Hollands, K.G.T., and J.L. Wright, (1983), "Heat Loss Coefficient and Effective Transmittance-
Absorptance Product for Flat Plate Collectors With Diathermanous Covers", Solar En-
ergy, 30, No. 3, pp. 211-216.
Lienhard, J.H., (1987), "An Improved Approach to Conductive Boundary Conditions for
Rayleigh-Bernard Istability", ASME J. Heat Transfer, 109, pp. 378-387.
Lienhard, J.H., and I. Catton, (1986), "Heat Transfer Across a Two-Fluid-Layer Region',
ASME Journal of Heat Transfer, 108, pp. 198-205.
Richards, R.F., and D.K. Edwards, (1989), "Effect of Boundary Radiation on the Thermal
Stability in Horizontal Layers", Int. J. Heat Mass Transfer, 32, No. 1, pp. 81-86.
Ulrich, T.R., (1984), "Heat Transfer Across a Multi-Layered Air Enclosure", Master's Thesis,
in Engineering, University of California, Irvine, Irvine California.
Wijeysundera, N.E., and M. Iqbal, (1991),"Effect of Plastic Cover Thickness on Top Loss
Coefficient of Flat Plate Collectors", Solar Energy, 46, No. 2, pp. 83-87.
Wright, J.L., H.F. Sullivan, (June 1988), "Glazing System U-Value Measurement Using a
Guarded Heater Plate Apparatus", Window U-Value Measurements Symposium, ASHRAE
Summer Meeting, Ottawa, ASHRAE Transactions, Vol. 94, Pt.2.
1/3 "I +
-------
1203
Table 1. Measured Collector Top Heat Loss Conductance U for 25.4 fim Thick Film
Tilt
Angle
Air
Gap
AT = 9.6 ± 0.2°C
AT = 29.2 ± 0.4°C
AT = 48.6 ± 0.3°C
Size
T
J-mean
T
mean
Tmcan
Tmean
T
-1 mean
T
-1 mean
L
R
= 30 °C
= 40 °C
= 30 °C
= 40 °C
= 30 °C
= 40 °C
24 mm
5/7
1
2.41
2.32
2.74
2.50
2.60
2.69
2.82
2.69
2.82
3.06
2.93
3.17
0°
7/5
2.17
2.46
2.64
2.76
2.98
3.07
36 mm
5/7
1
2.13
2.14
2.18
2.23
2.65
2.64
2.69
2.73
2.90
2.88
3.04
2.98
7/5
2.13
2.21
2.56
2.68
2.89
2.92
5/7
2.39
2.74
2.52
2.75
2.77
2.90
24 mm
1
2.29
2.50
2.41
2.56
2.77
2.89
30°
7/5
2.18
2.46
2.46
2.59
2.72
2.80
36 mm
5/7
1
2.02
1.96
2.15
2.17
2.52
2.54
2.53
2.58
2.82
2.81
2.85
2.75
7lb
2.04
2.10
2.46
2.59
2.84
2.83
24 mm
5/7
1
2.40
2.28
2.74
2.50
2.47
2.33
2.71
2.56
2.64
2.52
2.83
2.70
45°
7/5
2.20
2.45
2.33
2.49
2.53
2.67
36 mm
5/7
1
2.00
1.90
2.12
2.14
2.33
2.36
2.44
2.57
2.61
2.63
2.68
2.53
7/5
1.94
2.00
2.28
2.42
2.63
2.65
24 mm
5/7
1
2.39
2.27
2.73
2.51
2.44
2.34
2.69
2.55
2.59
2.39
2.78
2.54
60°
V5
2.20
2.38
2.23
2.45
2.41
2.55
5/7
2.01
2.11
2.18
2.36
2.43
2.59
36 mm
1
1.88
2.13
2.19
2.28
2.45
2.33
7/5
1.83
1.94
2.14
2.26
2.45
2.45
5/7
2.37
2.59
2.44
2.50
2.48
2.56
24 mm
1
2.22
2.36
2.34
2.43
2.22
2.38
90°
7/5
2.20
2.28
2.25
2.34
2.30
2.07
36 mm
5/7
1
1.97
1.87
1.86
1.78
1.97
1.99
1.99
1.71
1.94
1.85
1.95
1.60
7/5
1.80
1.72
1.70
1.66
1.73
1.59
the actual
temperature difference ATa across the air can be obtained from ATa = AT/ (1+
0.023 U).
Tmean is the mean of the copper plate temperatures.
L is the thickness of the total air layer.
R = ratio of air layer thickness next to hot plate to that next to the cold plate.
(Heating is from below.)
The tilt angle is measured from the horizontal position.
-------
1204
Table 2. Measured Collector Top Heat Loss Conductance U for 12.7 fim Thick Film
Tilt
Angle
Air
Gap
AT =
9.8 ± 0.1°C
AT = 29.2 ± 0.2°C
AT = 48.3 ± 0.4°C
Size
Tmean
Tmean
Tmean
T
J-mean
T
J-mean
Tmcan
L
R
= 30°
o
II
o
0
O
= 30 °C
= 40 °C
= 30 °C
= 40 °C
5/7
2.44
2.67
2.64
2.82
2.79
2.94
24 mm
1
2.21
2.41
2.89
2.91
3.31
3.38
0°
7/5
2.21
2.36
2.79
2.97
3.02
3.17
36 mm
5/7
1
2.13
2.37
2.16
2.33
2.70
2.90
2.89
3.03
3.01
3.10
3.17
3.27
7/5
2.17
2.22
2.82
2.94
3.07
3.22
5/7
2.43
2.68
2.56
2.74
2.75
2.90
24 mm
1
2.18
2.39
2.53
2.58
3.05
3.04
CO
o
o
7/5
2.21
2.33
2.71
2.78
2.97
2.97
36 mm
5/7
1
2.00
2.12
2.11
2.08
2.69
2.76
2.74
2.85
2.95
3.01
3.10
3.19
7/5
2.05
2.20
2.70
2.75
2.96
3.11
5/7
2.42
2.67
2.50
2.68
2.64
2.80
24 mm
1
2.18
2.38
2.32
2.47
2.67
2.70
45°
7/5
2.20
2.33
2.52
2.57
2.70
2.75
5/7
1.88
2.04
2.51
2.53
2.73
2.87
36 mm
1
1.92
2.00
2.49
2.60
2.79
2.94
7/5
2.01
2.19
2.44
2.54
2.73
2.84
5/7
2.42
2.67
2.46
2.65
2.58
2.74
24 mm
1
2.18
2.36
2.24
2.39
2.42
2.50
60°
7/5
2.17
2.33
2.30
2.40
2.51
2.55
5/7
1.81
2.02
2.28
2.32
2.54
2.66
36 mm
1
1.84
2.00
2.31
2.39
2.57
2.69
7/5
1.94
2.13
2.26
2.35
2.52
2.63
5/7
2.42
2.62
2.44
2.64
2.50
2.62
24 mm
1
2.15
2.37
2.24
2.41
2.28
2.46
90°
7/5
2.17
2.29
2.19
2.28
2.15
2.20
5/7
1.80
1.97
2.04
2.21
2.12
2.27
36 mm
1
*1.91
2.00
2.03
2.16
2.20
2.29
7/5
1.94
2.13
2.06
2.20
2.15
2.28
*AT for this measurement was 9.3°C.
See bottom of Table 2 for meaning of symbols
-------
1205
3.5
3
ca 3
>
3
-a 2.5
<0
3
O 2
ca
o
1.5
1.5 2 2.5 3 1.5 2 2.5 3 3.5
Measured U Value Measured U Value
Fig„ 1: Comparison of prediction using Hollands and co-workers' (1976) equation with
measurements. The left plot is for 25.4 fj,rn thick film. The right plot is for 12.7 fim thick
film.
Table 3. Measured U Values Across 3 Air Layers, each 15 mm thick
Tilt
Thickness of film
Thickness of film next
Angle
to cold plate
next to hot plate
12.7 fim
25.6 jim
12.7 fim
2.54 (48.6)
2.46 (48.6)
0°
25.4 fim
2.58 (48.5)
*2.54 (48.6)
12.7 fim
2.43 (48.6)
2.38 (48.7)
30°
25.4 fj,m
2.49 (48.6)
*2.45 (48.6)
12.7 /im
2.12 (48.7)
2.06 (48.7)
45°
25.4 jim
2.21 (48.7)
*2.13 (48.7)
12.7 (im
1.97 (48.8)
1.92 (48.8)
60°
25.4 (im
2.06 (48.7)
*1.99 (48.8)
12.7 jj,m
1.85 (48.8)
1.82 (48.9)
90°
25.4 yum
1.92 (48.8)
* 1.86 (48.8)
* average of 2 separate measurements.
** The numbers in parentheses are the ATa
values associated with each U value.
Tmean for all tests was 30°C
Air Layer Tilt Angle
Horiz. o 30,45,60
Air Layer Tilt Angle
Horlz. b 30,45,60
2.5 3.
Measured U Value
D 2.5 :
Measured U Value
Fig. 2. As in Figure 1, but using eqn.(2) to calculate the convection heat transfer.
-------
1206
TWO DIMENSIONAL TRANSIENT ANALYSIS FOR THE SOLAR HEATING OF
A FLUID BY A PARTIALLY RADIATION ABSORBING MEDIUM
Y. B. Safdari, D. A. Witek, and A. Fakheri
Department of Mechanical Engineering
Bradley University
Peoria, II61625
ABSTRACT
The thermal trap effect and its advantages for the semi-transparent medium applied to the solar
heating of a fluid is studied. The fluid is heated by thermal radiation transmitted through a semi-
transparent medium. The radiation flux is reflected by an ideal paraboloidal concentrating
collector. The two dimensional transient temperature distribution for the semi-transparent medium
with a variable flux distribution is determined for the first time.
The temperature distribution inside the semi-transparent medium depends on the flux distribution.
The extent of radial temperature dependence follows the radial variation of the radiation flux at the
base of the medium. For the same hemispherical total radiation, the fluid temperature is
independent of flux distribution. Since the air side temperature of the semi-transparent medium is
lower than the fluid side, heating of a fluid with a semi-transparent medium reduces heat loss.
This work makes a significant contribution toward understanding and analyzing the thermal stress
related problems in the target materials due to high intensity flux distribution at the target.
KEYWORDS
Solar Heating of a Fluid; Absorbing Medium; variable flux distribution; Thermal Stress;
paraboloidal concentrating collector.
INTRODUCTION
It has been shown theoretically and experimentally by Safdari [1966] that the far side of a
transparent medium receiving solar radiation achieves a higher steady state temperature than the
side receiving radiation.
Same was shown to be the case if the medium is not perfecdy transparent, for which absorption acts
as a source of internal heat generation,as discussed by Cobble [1963]. Safdari [1981] analyzed the
transient one dimensional conduction for a semi-transparent medium, subject to a uniform flux at
its base. His analytical results for the semi-transparent medium are in better agreement with the
experimental results compared to those for the perfectly transparent medium [1966].
This work extends Safdari's [1966 & 1981] analyses to two dimensions. It also incorporates
radially variable flux distribution. Solutions are obtained for the two dimensional transient
temperature distribution in a semi-transparent medium located at varying distances from the focal
plane of an ideal paraboloidal concentrating collector that reflects the concentrated solar radiation on
the medium. The effect of a variable flux distribution is also compared to that of a constant flux for a
perfectly transparent medium.
-------
1207
ANALYSIS
Figure 1 shows the geometry under consideration. A semi-transparent circular disk (the medium)
receives radiation at its base exposed to ambient air. The radiation is reflected from an ideal
paraboloidal concentrating collector. The disk is insulated around its perimeter, while the top
surface is in contact with a fluid. The container surrounding the fluid is also insulated.
Partially
absorbing
medium
Mirror
Fig. 1 - Schematic of the system
The transient two dimensional conduction equation for the disk can be written as
cTT 1_3,9T\ q'" _ 1 8T
2 + r dr Tdr + k „ 3t (1)
OZ urn
Using Lambert's law, the radiation intensity is given as
I(z) = Iinf(|) e+z and q'" = (i. Iinf(£) e^ (2)
where the term f(r/R) is a nondimensional function accounting for the flux variation in the r
direction and R is the radius of the disk. The functional form of f(r/R) depends on geometrical
parameters of the system such as the dimensions of the paraboloidal mirror and the receiver
medium and their relative orientations.
Substituting Equation 2 into Equation 1 yields
dh 1 9 ,3T. i ax
+ — K- (r-5~) + : - = 5T (3)
2 r 3r 9r k n 3t
dz
The initial and boundary conditions are:
T(z,r,0) = T„, ^ (z,0,t) = 0, ^(z,R,t) = 0
-k (0,r,t) = hJT^-T(0,r,t)], -k ^ (L,r,t) = Iinf(£) e"^-hf[T(L,r,t) - Tf] (4)
The fluid is well stirred and therefore its temperature is only a function of time. The energy
transfer equation to the fluid is given below:
i>R
3Tf fK
Pfcpfv-g- = Qin. whereQin= I h^TfL.r.t) - Tf]2jtrdr
(5)
-------
1208
In the above equations Too is the temperature of surroundings, Tf is the fluid temperature, rf is the
fluid density, Cpf is the constant pressure specific heat for the fluid, V is the fluid volume. At the
medium-fluid interface, all the energy received at the boundary z = L is absorbed by the fluid. The
energy is then transferred to the fluid via a convective boundary.
The governing equations are nondimensionalized and solved numerically using an implicit finite
difference technique to obtain the temperature distributions for the medium and the fluid.
RESULTS
The results were first obtained for the perfectly transparent medium (|a. = 0.) with a constant flux
distribution as shown in Figure 2. These are in excellent agreement with Safdari's [1966] results
and serve as a verification. The results that follow were obtained using the same dimensions as
Safdari [1966].
380
370
360
350
Fluid Side-
(medium)
Air Side
(medium)
340
Fluid
330
320
310
300
290
280
0
2.0
4.0
6.0
8.0
10.0
0 .20 .40 .60 .80 1.0
TIME (hours)
Fig. 2 - Temperature rise with time of fluid, fluid side, and air side of transparent medium
Simon [1959] and De La Rue [1957] have developed a model for flux distribution of radiation
reflected from a paraboloidal mirror onto a receiver medium. The flux distributions used are
shown in Figure 3. Curve A represents a constant flux. To examine the effects of the flux
distribution, a moderate distribution (Fig. 3, curve B) and a highly intensified distribution (Fig. 3,
curve C) are used. Although the flux distribution varies, it is important to note that the total energy
received at the base of the medium is constant for all cases. Results are obtained for the radial lines
of the medium at three different axial locations: air side, center line, and fluid side.
Using the more intensified flux distribution C (Fig. 3), Figure 4 shows the air side, center line,
and fluid side temperature distribution plotted for times of 1, 5, and 10 hours. Notice the middle
(r*=0) to end (r*=1.0) temperature difference for the fluid side is larger than that of the air side or
center line. This is because the boundary at z=L acts as a source, where all of the energy is
-------
1209
absorbed. The middle to end temperature difference is much larger for a more intensified flux
distribution C.
f
20-
0
Fig. 3. - Flux distribution profiles
Next we consider the effect of flux distribution in the axial direction. The fluid, fluid side, and air
side temperatures are plotted for the nodal location r*=0. Figure 5 is a comparison of the fluid,
fluid side, and air side temperatures for an intensified flux distribution C and constant flux
distribution A at the center nodal point (r*=0). The temperature of the fluid side increases more
rapidly for an intensified flux distribution than for a constant flux distribution. This is also true
for the air side temperature. Notice however that the fluid temperature increases at the same rate
for both constant and varying flux distributions.
Fluid side
11 rm - lOhturs
limn ¦ahourg
11 rm ¦ i rnir
450
400
350
300
250
¦[Center linq-
11 rm - 10 h wit
firm -1 hnir
450
400
350
300
250
I Air sid
—
—
—
I1rm - lOhwr*
_
11 rm -3 hour*
11 rm ¦ 7 hair
1 1 1 1 1 1 1 1 1
.10
.20
.30
.40
.50
.GO
.70
.80
.90
1.0
RADIAL LOCATION ( r* )
Fig. 4 - Temperature vs. radial location
-------
1210
280
i—r
Flux distribution - A
Flux distribution - C
F1 uid Side
(medium)
J L
J L
Air Side
(medium)
I I L
1.0
7.0 8.0
9.0 10.0
2.0 3.0 4.0 5.0 6.0
TIME (.hours)
Fig. 5 - Temperature vs. time for fluid, fluid side, and air side of medium at r* = 0
This work establishes some important results. It has been shown that the radial temperature
distribution profile follows the flux distribution profile received at the air side of the medium. This
proves the temperature distribution within the medium to be two dimensional and is directly related
to the flux distribution. It has also been shown that, provided the total energy received at the base
of the receiver is constant, the fluid temperature is independent of flux distribution. The
temperature of the air side of the medium is less than the fluid side of the medium,resulting in less
heat loss. Therefore, the fluid reaches a higher steady state temperature with a semi-transparent
medium. More importantly, this work shows that thermal stress related problems caused by large
temperature gradients due to highly intensified flux distributions can be observed and analyzed.
CONCLUSIONS
It is noted that for constant flux distribution the temperature distribution is one dimensional (axial)
and transient. This agrees with results presented by Safdari [1966]. It is concluded that the
temperature distribution in a semi-transparent medium is two dimensional (axial and radial) and
transient with a varying flux distribution. It is also concluded that, if the total energy received at
the base of the semi-transparent medium is constant, the fluid temperature is independent of flux
distribution under the well-stirred fluid assumption.
NOMENCLATURE
Cpf = Heat Capacity for the fluid, Btu/lbm-deg R
Cpm = Heat Capacity for the semi-transparent Btu/lbm-deg R
hf = Fluid side film coefficient, Btu/hr ft2-deg R
hM = Air side film coefficient, Btu/hr ft2-deg R
Iin = Beam strength entering medium, Btu/hr ft2
k = Thermal conducitivty of the medium, Btu/hr ft2/ft-deg R
L = Medium thickness, ft
H = Fluid depth, ft
-------
1211
R
V
Radius of the medium disc
Volume of the fluid, ft?
Internal energy generation, Btu/ft3-hr
Energy transferred from the medium to the fluid, Btu/hr
Temperature of the medium, °R
Temperature of the fluid, °R
Temperature of the surroundings, °R
time, hours
Coordinate, ft
k/pm Cpm = thermal diffusivity, ft2/hr
q"
Qin
T
Tf
T c
x
a
Pf
Pm
Density of the fluid, lbm/ft3
Absorption coefficient, 1/ft
Density of the medium, lbm/ft3
REFERENCES
Cobble, M.H., "Radiation in Transparent Solids and the Thermal Trap Effect". (1963) Engineering
Experiment Station, New Mexico State University, University Park, New Mexico, Technical
Report No. 16.
De La Rue, R., et. al., "Flux Distribution Near the Focal Plane". (1957) Solar Energy, 1, No. 2-
3, 94.
Safdari, Y.B., "Radiation Heating Through Transparent and Opaque Walls". (1966) Solar Energy,
Vol.10, No. 1, pp. 52-58.
Safdari, Y. B., A.L. Kasu, and M. A. Cascia, (1981). Solar World Forum, Vol. 3, 2285-2292.
Pergamon Press.
Simon, A.W., "Calculation of the Concentration of Energy at Points Outside the Focal Spot of a
Parabolic Condenser".(1959) Solar Energy, Vol. 3, No. 4, p. 67-69.
-------
1212
THE BIFACIAL ABSORBER COLLECTOR:
A NEW HIGHLY EFFICIENT FLAT PLATE COLLECTOR
A. Goetzberger, J. Dengler, M. Rommel, V. Wittwer
Fraunhofer-Institut fiir Solare Energiesysteme
Oltmannsstrasse 22, 7800 Freiburg, Germany
ABSTRACT
A new flat plate collector has been developed which, especially during times of low radiation
conditions, reaches higher efficiencies than any other collector known so far. The absorber plate is
insulated with Transparent Insulation Material (TIM) on the front and on the rear side and both
sides are irradiated. A prototype was built. Indoor and outdoor measurements were carried out to
determine the optical efficiency T)q and the U-value of the collector.
KEYWORDS
Flat plate collector, Transparent Insulation Materials, bifacial absorber collector.
INTRODUCTION
A new type of collector, having higher efficiencies than any other flat plate collector under
conditions of low irradiation or high operating temperature^ has been realized. This collector
accomplishes effective light concentration by a factor of two without restriction of angular
acceptance of radiation. (The optical configuration of this design has also been proposed for solar
cells (Goetzberger, 1988)). The collector design is based on the properties of transparent
insulation materials with high thermal insulation qualities which have been developed in recent
years. For a review see (Wittwer, 1990). The concept which is shown in Figure 1 can be described
as follows:
- Capture of radiation from an area 2 F, where F is the area of the absorber sheet
- A thin absorber plate is illuminated on both sides
- The absorber sheet of area F is surrounded very tightly with transparent insulation.
According to these principles the absorbed radiative energy is proportional to 2 F, while the
thermal losses are proportional to F. Compared to the design of a conventional flat plate collector,
the losses of the opaquely insulated backside of the absorber are eliminated, and the area is used
as additional absorber area.
-------
1213
Fig. 1. Collector design
This design is advantageous when the thermal losses of the back side of the collector are no longer
negligible compared to those on the front side. For a conventional selective flat plate collector with
single glazing the ratio is Ufront/Uback = 3.5/0.5. For a collector with transparently insulated front
side (Rommel, 1987) this ratio is Ufront/Uback = 1/0.5 only.
The performance of a transparently insulated collector is therefore significantly improved when
the absorber is covered on both sides with transparent insulation material and is irradiated on
both sides. Thus the back side of the absorber is no longer just losing energy but also contributes to
the gains.
DESIGN OF PROTOTYPE
In order to test the validity of the concept, a prototype was designed and built. Two cylindrical,
trough-like mirrors direct the light to the rear of the absorber. The front surface is directly
illuminated as in a conventional flat plate collector. This arrangement accepts all incoming
radiation of area 2 F with some losses due to the reflectivity of the mirror. The absorber is
surrounded by transparent insulation of width a and thickness b. It is important to have a > > b in
order to reduce heat losses due to edge effects.
•1672 mm-
•836
Fig. 2: Cross-section with actual dimensions
-------
1214
Prior to the realization of the collector the dimensions and materials parameters were optimized
by computer simulation. The transparent insulation material was of the polycarbonate honeycomb
type as described in (Wittwer, 1989). Its thickness was 5 cm. The air gap between the absorber and
the honeycomb material was 2.4 cm. The actual dimensions are given in the cross section of Fig. 2.
The size of the absorber sheet is 0.826 m times 1.84 m, the absorber area is 3.08 m2. Figure 3
shows a photograph of the finished collector on the outdoor test facility.
Fig. 3: Photograph of the prototype collector.
THEORY
The efficiency was determined for different operating conditions. The angular dependence of the
transmission of the transparent insulation materials has to be taken into account. Figure 4 shows
both the total energy transmission and the optical transmission of a low iron glass pane attached to
the the honeycomb material used.
energy tronsmittance
o
o
c
a
s
solar transmittanco
w
c
g
0.0
0
15
30
45
60
75
90
incidence angle [degree]
Fig. 4. Angular dependence of solar transmittance and total energy tronsmittance of a low iron glass
pane attached to the 5 cm thick honeycomb material used
-------
1215
Computation of the angular dependence of the front side is straightforward. For the back side it
has to be considered that the beam directions are drastically altered by the circular reflector.
Results of a ray tracing simulation for incidence angles 0°, +40° and -40° are depicted in Fig. 5.
The inserts give the intensity distribution along the radius and the area integrated distribution of
incidence angles in arbitrary units. The reflectivity of the mirror has been assumed to be 0.9 in
these calculations. Strong inhomogeneities are expected at 0° and at negative angles. In the
experiment, however, the temperature was found to be constant along the absorber within the
accuracy of the measurement. This is probably due to two reasons:
a) small inhomogeneities of the mirror surface will dampen the intensity peaks
b) the transparent insulation material has the tendency to homogenize the light intensity by
multiple reflections.
This applies also to some extent to the direction parallel to the cylinder axis where strong inhomo-
geneities can be expected at high incidence angles. An optical efficiency T]q of 0.72 was determined
for direct normal irradiation and 0.59 for diffuse irradiation.
angle of incidence:
0 degree
Fig. 5. Results of ray tracing simulations for
incidence angles of 0° + 40° and -40° to
investigate possible inhomogeneities caused by the
circular reflector (p=0.9). The intensity
distribution along the radius and the area
integrated distribution of incidence angles are
given in arbitrary units.
angle of incidence:
-40 degree
-------
1216
EXPERIMENTAL RESULTS
Stagnation Parameters under diffuse irradiation fSolar Simulator^!
The stagnation temperature was measured for three irradiation intensities and three inclination
angles of the collector. From the temperature difference AT against the ambient temperature and
with an estimated total energy transmittance of 0.64 for the experimental conditions under the
solar simulator, the temperature dependent U-values of the collector were determined. The results
are listed in Table 1. These values are in good agreement with simulated results.
TABLE 1 Results of Indoor Measurements
tilt angle
irradiation, transmitted energy, Tcojj - Tamj3,
Wm"2 Wm"2 K
U-value,
Wm"2K4
134.1
85.8
70.6
1.21
horizontal
100.3
64.2
55.4
1.16
51.9
33.2
31.3
1.06
143.8
92.0
69.8
1.32
45°
105.2
67.0
53.0
1.26
54.9
35.2
31.5
1.12
138.4
88.6
67.1
1.32
vertical
100.5
64.3
50.5
1.27
53.3
34.1
29.3
1.16
Outdoor measurements
First the heat loss coefficient was determined by observing the cooling rate at night. The U-values,
normalized to the absorber area, are plotted in Fig. 6 . They exhibit a rather high degree of scatter
which is mainly due to different sky temperatures.
8 1-6+
a 1.5--
® 1.3--
-=<0.9--
0 20 40 60 80 100
Tcoll — Tomb, K
Fig. 6. V-values, determined from the cooling rate at night in outdoor measurements.
-------
1217
The average temperature dependent U-value was determined as
U = (0.949 + 0.0076 AT/K) W m"2 K"1
Then efficiency characteristics of the collector were constructed from experimental data. They are
plotted in Fig. 7. They were in good agreement with measured points under varying meteorological
conditions. For comparison, the characteristics of a selective flat plate collector and an evacuated
tubular collector are included in Fig. 7.
1 .0
0.9
-o- 0.8
a)
o 0.7
-------
-------
2.3 Collectors III
iceding page blank
-------
-------
1221
DYNAMIC COLLECTOR MODELS FOR 1 HR TIME STEP DERIVED
FROM MEASURED OUTDOOR DATA.
Bengt Perers, Hakan Walletun
Studsvik Energy
S-61182 Nykoping
Sweden
ABSTRACT
A measurement and evaluation method is described by which standard
collector performance parameters can be derived directly from
measured all-day outdoor 1 hour data. Multiple regression is
presently used to determine the model parameters. A one node
capacitance correction for dynamic effects and separate incidence
angle models for direct and diffuse radiation is essential for the
accuracy of the method. The model is set up for useful energy Qu
(and not efficiency) which forces the parameters to values that
are suitable for prediction of long term performance. The
collector model and parameters correspond closely to those used in
existing detailed simulation programmes like TRNSYS, WATSUN, or
MINSUN. The method can be used as an accurate bridge between short
term testing and long term prediction by simulation.
KEYWORDS
Solar Collector Testing; Collector modelling; Simulation; Multiple
Regression.
INTRODUCTION
The aim of this work is to find a practical and accurate enough
connection between measured outdoor all day collector performance
and collector efficiency parameters that can be used in standard
simulation programmes for long term performance prediction based
on 1 hour time steps.
In the Swedish climate stationary outdoor test methods are very
time consuming and expensive to use.
For testing of large area collectors and in situ testing of
collector arrays an improved outdoor test method is needed that
can use all day values from normal operating conditions.
The basis for this work is the experience with on-line simulation
directly in the measurement computer which has been gained since
1979 in all our testing and monitoring projects.
A standard collector model, with correction terms for incidence
angle effects and thermal capacitance effects, can describe the
hourly and daily performance accurately enough to show if the
collectors and the system are performing as expected, and also give
some idea of what is wrong when the agreement is not so
satisfactory.
Preceding page blank
-------
1222
The inverse method of on-line simulation described here has been
used at our laboratory for in situ testing of collectors and
collector arrays since 1985. Some of the first results are
described in (Walletun 1986), (Perers and Hoist 1987) and (Perers
1987) .
The collector model used here is based on the almost 50 year old
Hottel Whillier Bliss equation with improvements to take into
account thermal capacitance effects, incidence angle effects and
the temperature and wind dependence of the heat loss coefficient.
The basic model and the correction terms used are described in
(Duffie and Beckman 1980). A description and validation of
improved collector models can also be found in (Ambrosetti 1983)
and (Gemmel 1986) describing work within IEA&C Task VI and Task
III.
The 1 hour time step is chosen mainly because of the standard
resolution of available weather data for simulation programmes.
In practice we have found that the 1 hour time steps also imply
that thermal capacitance effects in the collector can be treated
with a one node model without going into detail in the collector
design. This corresponds closely to theoretical work by (Klein
1973) investigating different thermal capacitance models by
detailed simulation.
Improved outdoor test methods have been proposed and validated by
(Proctor 1984) and (Emery 1984) to derive standard stationary
collector parameters from outdoor testing.
This method utilizes a more complete characterisation of the
collector including incidence angle and thermal capacitance
effects. This leads to a wider range of acceptable climate
conditions for the test, and a shorter test period.
The thermal capacitance correction term is only based on
information about the mean fluid temperature variation within the
hour (a new on line data reduction in the measurement system is
required). This means that the hourly values do not have to be in
sequence. Databases with gaps or data from different test periods
can be used without any problem.
PROPOSED COLLECTOR ARRAY MODEL FOR GLAZED COLLECTORS
The model described below is a mixture of already existing and
validated correlation models for the complex instantaneous
thermal and optical behaviour of a solar collector. The extension
of the instantaneous models to hourly mean values is possible also
for dynamic conditions, as the integrated effect of thermal
capacitance within the hour is taken care of in the model.
Except for the wind dependence and thermal capacitance term, most
detailed simulation programmes already use the same performance
model. One problem is that some programmes such as TRNSYS and
WATSUN use models based on inlet temperature T. , whereas the
model in this case is based on the mean fluid -temperature T_. The
conversion between the two formats is described in (Duffie 1985).
-------
1223
The model is written as an energy balance as in the simulation
programmes and not as collector efficiency. This is important for
the parameter values, as the multiple regression programme should
minimize the error in useful energy and not in efficiency, as this
would lead to a too high influence of low radiation conditions on
the parameter values.
The present model used for glazed collectors is:
q = F1(TO) K . (0)1, + F'(ra) K ,1, - F'U.DT - F'U.(DT)2-
ni 1 'e rab1 ' b 1 'e rad d 0 lv '
- F'UwDT W - (mC)edTf/dt - UpDT (1)
qu = Collector array thermal output [W/m2]
F'(ra)e= Zero loss efficiency for direct radiation [-]
at normal incidence.
Kj-ab= Incidence angle modifier for direct radiation [-]
^rad = Inci^ence angle modifier for diffuse radiation [-]
1^ = Direct radiation onto the collector plane [W/m2]
Id = Diffuse radiation onto the collector plane [W/m2]
F'Un = Heat loss coefficient at T_-T =0 [W/(m2*K)]
U £ cl
F1= Temperature dependence in the heat loss coeff.[W/(m2*K2)]
F'Uw = Wind speed dependence in the heat loss coeff. [W/(m*K*s)]
U = Piping heat loss coefficient per m2 of coll. [W/m2*K]
P
(mC) = Effective thermal capacitance incl. piping
for the collector array. [J/(m2*K)]
DT = Temperature difference (Tf-Ta) between mean
collector fluid temp, and ambient temperature [C]
w = Wind speed near the collector [m/s]
dT^/dt = Mean time derivative for the average fluid
temperature Tf in the collector within the
time step. [K/s]
Tf = Mean fluid temperature in the collector
*0-5
T = Ambient air temperature near the collector [CJ
6 = Incidence angle for the direct radiation onto -
the collector plane [radians]
-------
1224
TEST REQUIREMENTS
Standard test equipment can be used, provided that all necessary
variables in the model above are measured. The only change
required is an on-line data reduction that calculates and stores
the average mean fluid temperature change per second within the
hour.
The only limitation for the use of the model is that the mean
fluid temperature can only be determined as long as there is a
continuous mass flow in the collector array during the hour.
At the solar collector test field in Studsvik the collectors are
operated with a continuous flow. This means that data for the
whole day can be used.
As this method uses the measured data much more intensively for
the whole day, some extra care is required when using the sensors
in practice. Placing, ventilation and radiation shielding of
ambient temperature sensors has become important in order to give
a representative ambient temperature for the collector. The
alignment of the pyranometers also becomes more important in this
case.
RESULTS
At the Studsvik Solar Test Field we are presently testing 11 m2
modules of LGB- (Long Ground Based Collectors) flat plate
collectors which can be site built in sizes up to 70 m long
(Wilson 1988). In this case the collectors have a selective
sunstrip absorber and a (teflon inner glazing).
Randomly chosen 65 hour periods of one hour meanValues for each
month April to September have been used to identify the model
parameters. Table 1 shows the results together with the expected
parameter values from the collector design and materials used.
TABLE 1. Model Parameters derived by Multiple Regression. The
input data come from 65 hour periods of one hour meanvalues each
month. There was no selection of the periods except the avoidance
of snow and frost on the collectors.
Month F1(ra), b„ F'(ra),K , F'UL (mC) St.Dev R2
v'b 0 1 'b rad v/e
[No.] [-] [-] [-] [W/(m2 K)] [J/(m2 K)] [W/m2 ] [-]
4 0.739 -0.137 0.652 -3.181 -8654 8.23 0.998
5 0.756 -0.125 0.712 -3.625 -12095 11.60 0.997
6 0.763 -0.196 0.738 -3.572 -9713 12.60 0.996
7 0.777 -0.146 0.742 -3.909 -9304 11.82 0.996
8 0.746 -0.123 0.660 -3.225 -9446 12.29 0.994
9 0.743 -0.130 0.671 -3.168 -8986 8.95 0.998
Expected 0.75 -0.15 0.68 -3.5 10000
All day values from sunrise to sunset have been used and in some
cases also the hours of darkness. This means that the parameter
-------
1225
values are valid not only for the operating time but will also
give a good estimation of the transient behaviour outside the
operating time.
The term F'UL in table 1 is the total heat loss factor. For these
randomly chosen data sets, the second order heat loss terms were
just on the limit to be statistically significant (T-ratio just
below 1) but were excluded in this presentation. The absolute
values were near what could be expected. Here the accuracy and
mounting of the climate sensors has a significant influence, and we
are now refining this part of our measurement installation in
Studsvik.
The seasonal variation is not completely
understood but the model
Krab(e) = 1_ bo(l/cos(0)-l) (2)
used here to describe the incidence angle dependence is not good
enough for angles exceeding 70°.
The summer data here contain many hours with incidence angles
greater than 70° which means that a seasonal variation in the
parameters can be expected.
An incidence angle model that is more accurate for high incidence
angles is described in (Ambrosetti 1983) :
KTab(0) = 1_ (tan(6/2))n (3)
the use of this model may reduce the seasonal effects.
CONCLUSIONS
Multiple regression can be used to identify standard collector
performance parameters from measured outdoor data.
Standard test eguipment can be used, but the mounting and placing
of the climate sensors become more important as all day values are
used and second order effects are taken into account.
By adding correction terms for incidence angle and thermal
capacitance effects to the Hottel Whillier Bliss collector model,
the all day performance on an hourly basis can be described
accurately for most climate conditions in Sweden during the period
April to September.
By calculating and storing the temperature change of the collector
mean fluid temperature each hour, the hourly thermal capacitance
effects can be described very well with a one node model for the
collector.
By running the pump in the collector loop continuously during the
test period, all day data can be used which increases the
variation range for the input variables. This will lead to better
accuracy for the individual parameters and a shorter test period.
The derived collector parameters for 1 hour time step can be used
directly in most detailed simulation programmes.
-------
1226
REFERENCES
Ambrosetti, P. (1983). Das neue bruttovaermeertragsmodell fuer
verglaste sonnenkollektoren, Teil 1 grundlagen. (The new Gross
Energy Contribution Model for Glazed Collectors, Part 1). EIR
Wurenlingen, Nov 1983. ISBN-3-85677-012-7.
Duffie, J. A., Beckman, W. A. (1980). Solar Engineering of Thermal
Processes. Chap.6 and 7. John Wiley and Sons Inc. New York 1980.
Gemmel, W. L. , Chandrashekar, M., Vanoli, K. H. (1986) Detailed
Modelling of Evacuated Collector Systems. IEA SH&C Task VI.
Klein, S. A. (1973). The Effects of Thermal Capacitance upon the
Performance of Flat Plate Solar Collectors. Msc. Thesis. Univ.
of Wisconsin.
Perers, B., Hoist, P. (1987). The Sodertorn Solar District
Heating Test Plant. Results 1982-1985. Studsvik AB, Sweden.
(STUDSVIK-87/1).
Perers, B. (1987). Performance Testing of Unglazed Collectors,
Wind and Long Wave Radiation Influence. Report for IEA Task III.
Studsvik Energy, Sweden.
Proctor, D. (1984). A Generalized Method for Testing all Classes
of Solar Collectors.Part I,II,III Solar Energy Vol.32. No.3.
Walletun, H., Perers, B. (1986) Vindens inflytande pa oglasade
solfangare respektive solfangare med konvektionshinder. (Wind
influence for unglazed collectors and collectors with convection
suppressing glazing.) Studsvik Energy, Sweden. (STUDSVIK/ED-
65/15).
Wilson, G. (1988). Construction, Performance and Cost Analysis
of The LGB-Collector. North Sun Conference 1988. Swedish Council
for Building Research D2:1989. ISBN 91-540-4973-3.
-------
1227
SOLAR COLLECTOR GEOMETRIES
AT HIGH LATITUDES
Lars Broman* and Gouri Datta**
*Solar Energy Research Center University College of Falun/Borlange
P. O. Box 10044, S-78110 Borlange, Sweden
**Physics Department, Deshbandhu College, Kalkaji
University of Delhi, New Delhi 110019, India
ABSTRACT
Insolation onto and energy output from solar collectors have been estimated for some typical
high latitude locations using the F-Chart computer program and weather data from the Swedish
Meteorological and Hydrological Institute. Optimum tilt of south-facing collectors at Swedish
latitudes is approx. latitude minus 20° for maximum yearly performance, and about latitude
minus 30° for maximum summer performance. If tracking is considered, vertical axis tracking
seems favorable with energy gain larger than normal 1-axis tracking with latitude tilted axis.
KEYWORDS
Solar energy; collector tilt; vertical axis tracking; F-Chart; high latitude.
INTRODUCTION
Solar collectors are either fixed or tracking. A stationary collector is usually facing the meridian
and tilted towards the equator. For temperate zones, say latitudes between 30° and 45°, a tilt
equal to the latitude is usually regarded as close to optimum. A larger tilt, latitude plus 15°, is
sometimes preferred, since this is a more favorable geometry during the winter.
At high latitudes, say over 55°, the situation is different. Days are so short during the winter that
not much energy can be collected with any collector tilt. Therefore, a solar installation is
frequently operated only part of the year. Furthermore, the sun rises far north of east and sets
far north of west during the long summer day, so much of the sun's daily radiation does not reach
a latitude tilt collector facing south. These facts together make a smaller tilt optimal. The
shadowing effect of collector rows, which can be appreciable when the sun is low above the
horizon, is also less when the collector tilt is smaller.
Tracking usually has to be employed when concentrators are used, but also the output from a
flat-plate collector is increased by tracking. This is true in the temperate zones, but the added
cost of support, bearings and a tracking motor is usually found trohigh. At high latitudes, however,
the situation may be different. The fact that the sun travels almost full circle during a summer
day makes the difference in output between a stationary and a tracking collector larger than at
lower latitudes. The added cost of tracking may in certain applications therefore be justified at
high latitudes.
-------
1228
The axis of a tracking collector is usually horizontal in the east-west direction or latitude inclined
in the north-south direction. In the second case, the daily rotation around the axis may or may
not be combined with seasonal adjustment of the angle between the collector and the axis. A
third tracking mode uses a tilted collector mounted on a vertical axis. If there is no large
difference between two axis and vertical axis tracking,the vertical axis mode can be chosen when
such a geometry is simpler and therefore less costly.
In the present paper, these geometries are qualitatively analyzed for three locations between
northern latitudes 55° and 68°.
METHOD
We have used the F-Chart computer program (Duffie, 1990) together with monthly global
insolation data from SMHI, the Swedish Meteorological and Hydrological Institute. This
program is easy to use when different collector geometries are to be compared with one another,
as in the present study. The calculated monthly and yearly insolation values are directly given in
the computer output, and the monthly and yearly utilized solar energy values are easily derived
from it.
For calculating the energy output from the collector, the collector parameters FR*UL = 4
W/m2,°C and FR*TAU*ALPHA = 0.8 were chosen as reasonably typical for the modern
Swedish solar collector. Furthermore, we used the default mass flow rate (0.015 kg/s,m2) and
the temperature of the water into the collector equal to 60°C. This gave a water output
temperature of just over 60°C, so the collector temperature was always close to 60°C, which is
typical for a collector in the middle of a large Swedish collector field. Finally, we made the hot
water requirement so large that the whole output from the solar collector was always used, and
the space heating requirement was set to zero (to simplify output data).
The F-Chart calculations were done for a number of different tilts of a south-facing collector as
well as for 2-axis and 1-axis tracking collectors at three of the different locations in Sweden for
which insolation data are available:
Kiruna, latitude 67.83° (northenmost Sweden)
Borlange, latitude 60.48° (middle Sweden)
Lund, latitude 55.72° (southernmost Sweden)
We used weather data from 1988 for all three locations. Weather data from Borlange 1989
produced only marginal changes as compared to Borlange 1988.
F-Chart has no vertical axis tracking mode, only normal 1-axis and 2-axis tracking modes.
Comparisons between these three tracking geometries were therefore done using clear weather
calculations, i. e. assuming a constantly clear sky. Absolute results from such calculations have
little relevance, but we felt that some conclusions regarding comparison between the modes
could be done.
According to Meinel and Meinel (1976), the intensity/(Z) in kW/m2 of direct sun is approxi-
mately given as a function of the sun's zenith distance Z by
I(Z) = 1.353 * exp(- 0.357 * cos Z - 0.678).
With the sun's declination equal to d, its hour angle equal to t (where t = 0 when sun is south),
and the latitude equal to I, Z is given by
Z = arccos(sin I *sin d + cos / * cos d *cos t).
-------
1229
For calculations of monthly clear weather insolation, monthly average solar declination values
are needed. These values were calculated from daily values given in astronomical ephemerides,
as well as the characteristic day of each month, i. e. the day with the (Swedish noon) declination
closest to the average.
The intensity of the direct sun on a collector surface is given by
/ = /(Z) * cos i,
where i is the angle of incidence. The values of i for the different geometries are for
2-axis tracking equal to 0, for 1-axis tracking equal to d and for vertical axis tracking equal to
(tc - Z), where tc is the collector tilt (i. e. the angle between the horizontal and the collector
planes).
Using the monthly declination values and the equations, the insolation during the characteristic
day was computed with a small program that sums 12-minute values over the day. Multiplication
with the number of days in the month finally gave the monthly clear weather insolation values.
CALCULATIONS AND RESULTS
The validity of the F-Chart results was first checked by calculating monthly insolation values at
a south-facing collector tilted 60° and 90° and with ground reflectance equal to zero. The results
•2-axis tracking
12 mo.
1000
6mo.
500
3 mo.
collector tilt (°)
2
Fig. 1. Variation of insolation onto an 1 m
stationary collector in Borlange with col-
lector tilt for three different time periods
in 1988 calculated with F-Chart. The
dashed lines are drawn at 95 % and 90 %
of the curve maxima. At the top of the
diagram are shown the calculated insola-
tion per m2 onto 2-axis and 1-axis tracking
collectors (left).
Fig. 2. Variation of the output from an 1 m
stationary collector in Borlange 1988 with
collector tilt (below).
(kWh)
500
12mo.
6mo.
3 mo.
0
0
40
collector tilt (°)
-------
1230
for Kiruna, Borlange and Lund were always close to those given by SMHI (calculated from
pyranometer and pyrfaeliometer data) and never more than 10 % off. Thus we felt confident that
F-Chart estimates indeed could be used in the present study.
Figure 1 shows the variation of insolation with tilt of the collector for Borlange 1988. Curves are
given for the whole year, for the six months April - September and for the three months June -
August. The peak for the whole year occurs at about 40°. Similar curves for Borlange 1989,
Kiruna 1988 and Lund 1988 gave the whole year peak at 40° for Borlange 1989 (the same as
Borlange 1988), at 45° for Kiruna and at 35 for Lund. However, if the collector is operated
during the summer half year only, the optimum tilts are quite different. The best tilt in this case
is for Borlange approx. 30° south, for Kiruna 40° south, and for Lund 25° south.
In case the collector is to be used for only the peak summer months, the tilts at which maximum
insolations are obtained, are for Borlange, Kiruna, and Lund 25°, 30°, and 20° south, respec-
tively. So it is seen that the thumb-rule of latitude tilt of collector for getting maximum solar
insolation the year round or latitude minus fifteen degrees for maximum summer half year
insolation does not hold for such high latitudes as in Sweden. In Figure 1, the horizontal dashed
lines refer to 95 % and 90 % of maximum insolation obtainable with best tilt for the whole year,
the summer half year, and the summer months, respectively. It is clearly shown that even quite
large deviations from the optimum tilt are possible if a 5 % or 10 % loss is acceptable.
Figure 2 shows the variation of the output from the collector described in the previous section
for Borlange 1988. Also, these curves are given for the whole year, for the six months April -
September, and for the three months June - August. The tilts for optimum performance of the
collectors are within 5 degrees of the corresponding values for maximum incident global
radiation. Noteworthy is the fact that there is no difference between whole-year performance
and six-month performance for low tilts and only a small difference for large tilts.
Using the method outlined in the previous section, some clear weather calculations regarding
vertical axis tracking were done. The results are summarized in Figures 3 and 4. In Figure 3, the
yearly direct insolation on a vertical axis tracking collector is given in % of the yearly direct
insolation on a 2-axis tracking collector as a function of collector tilt for latitude 60°. The relative
insolation on a 1-axis tracking collector is included for comparison. The maximum performance
of a vertical axis tracking collector is almost 97 % of the performance of the (more complicated)
2-axis tracking mode.
In Figure 4 a, the optimum tilt of a vertical axis tracking collector is given as function of latitude,
and in Figure 4 b the yearly insolation on vertical axis and 1-axis tracking collectors compared
with 2-axis tracking are given as functions of latitude. It is clear from this Figure that vertical
axis tracking is better than normal 1-axis tracking (with latitude tilt axis) for Swedish latitudes
and improves with increasing latitude.
From F-Chart calculations, 2-axis tracking in mid-Sweden (Borlange) is about 40 % better than
a stationary south-facing optimum-tilt collector; combining this with the clear weather results
indicates a 38 % improvement of vertical axis tracking over fixed, or 1600 kWh/m2 instead of
1150 kWh/m . The energy from the collector should increase by at least the same fraction.
Whether this increased yield is economical depends on the extra cost of a support that permits
tracking. Such calculations lie beyond the scope of the present paper. Finally,it must be pointed
out that shadowing effects will decrease the improvement in a collector field; Perers et. al. (1990)
have discussed this in a recent report.
-------
1231
rel. insolation
besttilt (°)
collector tilt (°) latitude(°)
Fig. 3. Vertical axis tracking collector with
different tilts compared with 2-axis and
1-axis (latitude tilt axis) tracking collec-
tors. Results from clear weather calcula-
tions (above).
Fig. 4 a. Best tilt of a vertical axis calcu-
lated for different latitudes (right top).
Fig. 4 b. Vertical axis tracking collector
with optimum tilt compared with 2-axis
and 1-axis tracking collectors for different
latitudes (right bottom).
rel. insolation
2-axis
vert, axis
1-axis
40 50
latitude (°)
For large-scale utilization of solar energy, a large number of collector panels have to be set up
on the ground. Typically, several long south-facing arrays oriented east-west are placed north of
each other. As seen from Figure 2, virtually all energy is collected during the summer half year,
i. e. when the sun's declination is positive. Then, shading between arrays does not occur when
the angle between the (horizontal) ground and the plane between the bottom of an array and
the top of the array south of it equals 90 -1 (or less) where I is the latitude.
Let collector area be^4, ground area G, and collector tilt tc. Then no summer half year shading
occurs when
G/A = cos t + sin t / tan (90 - /).
Figure 5 illustrates how G/A varies with tilt of collector for the three locations Kiruna, Borlange
and Lund. As seen from the Figure (and easily proved by derivation of the formula), most ground
area is required for latitude tilt collectors. The G/A ratios are, for whole year operation of
collectors at Kiruna (optimum tilt 45°) 2.5, for Borlange (optimum tilt 40°) 1.9, and for Lund
(optimum tilt 35°) 1.7. For summer half-year operation the optimum tilts and corresponding
G/A values for the three locations are respectively tc = 35°, 30°, 25° and GjA = 2.3,1.7,1.5.
In Figure 6, the relative cost (per installed power unit) of a large collector field in Borlange is
shown as a function of collector tilt for three different assumed ground costs (10 %, 20 %, and
40 % of collector cost). As the Figure shows, when taking ground cost into account, even lower
tilts are optimal. Calculations for Kiruna and Lund show similar results.
-------
1232
rel.
2.5
G/A
2.0
1.5
1.0
1 1
t 1 1 I I 1
0.6
_
/ -
r 1
i i i i i i
0.4
0
40 80
collector tilt(°)
Fig. 5. Required ground area G over col-
lector area .4 in a collector field in Kiruna
(K), Borlange (B), and Lund (L) in order
to avoid shadowing effects.
i i i
l i i i i
GC/CC
-
/ 40%
- ^
/ 20% -
/ 10%
i i i
i i t i i
0
AO 80
collector tilt (°)
Fig. 6. Relative total cost of a collector
field in Borlange with given collector area
for different ground costs (GC) to collec-
tor cost (CC) ratios.
CONCLUDING REMARKS
Our calculations using F-Chart and clear weather estimates have shown that at high latitudes,
much lower tilts of stationary collector arrays than latitude tilt give most energy per year. This
is principally due to two facts: Virtually nothing is gained during the winter half year at any tilt.
The sun travels almost full circle (in Kiruna full circle) during a summer day.
This astronomical fact also makes vertical axis tracking more favorable than at lower latitudes.
Whether the cost of tracking is worth the extra energy gained has not been computed in the
present study. In the case of PV panels, however, the cost of a typical 15 % efficiency panel is
about 1200 $/m2. To produce 1 MWh per year in Borlange with a south-facing panel requires
5.8 m2 at 7000 $, while a vertical axis tracking panel requires 4.2 m2 at 5000 $. It does not seem
unlikely that the extra cost of a simple rotating support over a fixed support is far less than the
difference in panel cost.
ACKNOWLEDGEMENTS
This work was done when one of us (Gouri Datta) was visiting scientist at Solar Energy Research
Center in Borlange, and she thanks Lars Broman for his hospitality during the stay. The study
was paid for by grants from the University College of Falun/Borlange.
REFERENCES
Duffie, J. (1990). Hot water systems - F-Chart analysis. Proc. Int. Conference North Sun '90.
Meinel, A. B. and M. P. Meinel (1976). Applied solar energy, Addison-Wesley.
Perers, B., B. Karlsson and H. Walletun (1990). Simulation and evaluation methods for solar
energy systems. Application for new collector designs at high latitudes. Report Studsvik/ED-
90/4.
-------
1233
OPTIMISATION OF FIXED REFLECTOR DESIGN TO MINIMISE BOTH BACKUP REQUIREMENTS AND
SOLAR COLLECTOR RECEIVER AREA.
D.R. Mills and A. G. Monger,
Department of Applied Physics, University of Sydney
and
G. L. Morrison,
School of Mechanical Engineering, University of New South Wales
ABSTRACT
A computer simulation of the relative performance of certain truncated
symmetrical and asymmetrical fixed reflector designs for solar energy
collection was performed. The major results were as follows :
1)Annual solar fractions in excess of 90% seems to be feasible with a fixed
load matching collector, in a climate where 7 0% of hot water requirements is
the norm from flat plate collectors. Consumer interaction could either improve
or lower this figure, depending upon circumstances.
2)Symmetrical CPC reflectors always gave the best annual output performance per
unit of mirror area, and allowed the lowest receiver area for situations of
constant annual load.
3)Asymmetrical fixed concentrators are most cost-effective for seasonally
asymmetrical load patterns.
4)Fixed parabolic systems were not competitive.
5) The concentration levels utilizable in fixed systems are much higher than
previously supposed, with approximately 3.1:1 in an asymmetrical reflector
being optimal for the domestic load pattern used.
6)With seaonal laod matching, the storage required to achieve solar fractions
above 90% appears to be minimal.
KEYWORDS
Non-imaging concentrators;.non-imaging concentrators; asymmetrical reflectors;
stationary solar concentrators; fixed reflectors; CPC.
INTRODUCTION
We present an abbreviated computer-calculated annual performance comparison of
fixed symmetrical and asymmetrical concentrating collectors for circular
cylindrical evacuated tube receivers, using an asymmetrical seasonal load.
Types investigated were:
Fixed Parabola
A parabola may be used as a fixed collector if the concentration factor is kept
very low. In the following, the cases were evaluated by varying the rim angle
and the mirror area.
Symmetrical Compound Parabolic Concentrator (CPC)
Figure 1 illustrates a symmetrical truncated CPC tilted for winter bias.
-------
1234
Construction is described elsewhere (Winston and Hinterberger, 1975). Truncated
versions (one is shown) are formed with an incomplete extension of the sides.
Extreme Asymmetrical Concentrator (EAC)
It was shown by Mills and Giutronich (1979(b)) that ideal extreme asymmetrical
reflectors may be formed for a cylindrical receiver using a pair of reflectors,
one of which is "quasi-parabolic" and one of which follows a "quasi-elliptic3l"
curvature. This configuration is also illustrated in Fig, 1, with the quasi-
elliptical section indicated. This ensures that all rays from a direction k'
parallel to the aperture to an acceptance angle limit k are accepted by the
collector; this behaviour is characteristic of an EAC, after Mills and
Guitronich, (1979(a)). The equation for the quasi-ellipse portion of the above
reflector section has been determined by Mills (1979), and will be given in a
much expanded version of this paper submitted to Solar Energy.
4?
Quasl-parabola
Involute
Quasi-ellipse
EAC
CPC
Fig. 1. Asymmetrical and symmetrical non-imaging reflectors.
MODELLING METHOD
Each collector configuration was modelled using a FORTRAN computer ray-trace
program which has been described in earlier work (Mills et al., 198 6) . The
raytraces were used as an input to a modified version of the TRNSYS simulation
program. The collectors were assumed to possess a perfect reflecting surface,
but losses were included. The Sydney weather data used were the Typical
Meteorological Year (TMY) data gathered at the University of New South Wales,
in Sydney. The daily and annual output energies were calculated, and each of the
collector designs was compared by modelling them at numerous fixed angles
throughout the year.
IMPACT OF LOAD: A UTILISABT.F. ENERGY COMPARISON CRITERION
For a load pattern 1 (m) , where L is measured in MJ per month and m is the
month, an appropriate criterion for comparison is the fraction of backup B,
measured as a fraction of annual load. For a collector system of aperture
-------
1235
area A, this may be found by calculating E(m), the monthly energy generated per
m2 by the collector or collector system, and then calculating the monthly
shortfall S(m) (if any), where
S (m) = E (m) *A - L (m) if E (m) *A < L (m)
and
S (m) = 0 if E (m) *A > L (m)
since the excess energy has no value. If the annual load, is given by
12
lANN = £ L (m) ,
m=l
then the annual backup fraction is
12
B = Z [ S (m) /Lann ] .
m=l
Collectors of identical backup requirement which use the smallest receiver area
for the load will be the most cost-effective. Collectors with seasonal output
variations which match seasonal load patterns will clearly have some advantage
under the above criteria, since they generate the energy seasonally when
required, minimising B.
RESULTS OF PERFORMANCE COMPARISONS FOR AN ASYMMETRICAL LOAD PATTERN
High Space Heat Assumption
SOOO -I
4000 -
Month
Fig. 2. NSW domestic thermal load with high space heating assumption.
In the State of New South Wales (NSW), the range of winter to summer domestic
thermal load is estimated to be between 3.25:1 and 4.75:1, depending upon space
heating usage assumptions. It would be desirable to bias annual output per tube
toward the heavy load season. Asymmetrical collectors can achieve this goal,
but suitably designed and inclined symmetrical reflectors can also simulate
this behaviour.
Table 1 shows a typical output from the sorting program for Sydney (34°S) which
indicates the best performer of each collector type in satisfying the load
profile in Fig. 2. Note that although annual collection of CPC's is higher
for a given receiver area, the useful energy obtained is lower per unit of
receiver area because much of the annual output is discarded in low-load
months. This means that, to achieve the backup criterion with a CPC, greater
receiver area (in this case, more evacuated tubes and metal tubing) are
-------
1236
required to satisfy the load, with consequently greater cost.
TABLE 1 ASYMMETRICAL LOAD PROFILE BACKUP FUEL FRACTION 5.0 -
Area in m2
Absorber
Reflector
Aperture
Ref/Abs
Ap/Abs
Tilt
As 0/225
5.91
23.64
15.23
4.00
2.58
58.0
As 30/235
5.16
25.81
15.93
5.. 00
3.09
58 . 0
As 30/225
5.31
22.56
14.76
4.25
2.78
58.0
As 30/180
6. 65
23.26
18.25
3.50
2.75
58.0
As 30/135
8.08
22.23
18.68
2.75
2.31
58 . 0
As 60/235
4.88
24.41
15.23
5.00
3.12
58.0
As 60/225
5.05
25.27
16.92
5.00
3.35
58.0
As 60/180
5.75
23. 00
17 .70
4 .00
3.08
58.0
As 90/235
4 .76
21.43
13.11
4 .50
2.75
61.0
As 90/225
4.87
24.33
16,09
5.00
3.31
58.0
As 90/180
5.64
22.58
16.50
4.00
2.92
58.0
As 120/235
4.92
24.61
15.10
5.00
3.07
58.0
As 120/225
4.89
24.46
15.69
5.00
3.21
58.0
As 120/180
5.41
27.05
19.19
5.00
3.55
58.0
CPC 12
7 . 68
19.20
14 .86
2.50
1. 93
49.0
CPC 15
7.38
22.13
15.67
3.00
2.12
46.0
CPC 18
6.80
23.79
15.2 6
3.50
2.24
43.0
CPC 28
5.28
26.40
11.40
5.00
2 .16
37 .0
CPC 30
6.36
19.08
11.41
3.00
1.79
40.0
CPC 35
6.65
19.94
11.18
3.00
1.68
43.0
CPC 40
7.26
21.78
11.40
3.00
1.57
46.0
Par 80
14.53
21.79
19.69
1.50
1.36
43.0
Par 90
13.33
20.00
17 .43
1.50
1.31
43 . 0
Par 100
11.43
17 .15
14.27
1.50
1.25
43.0
Par 110
11.09
19.40
15.21
1.75
1.37
43.0
The best choice for this highly asymmetrical (but typical!) load pattern were
fixed asymmetrical collectors using the quasi-ellipse profile.
1 40
120
2 100
-t—t
eL 80
o
E
60
40
20
0
* J.
mil 11
MP"'M III
irm
nun win
III!1 II M l
'I! II "'III1
]i 1 i'i
I'll Til
W\WWJ\
i
51
1 0 1
25 1
30 1
351
151 20 1
Day of Year
Fig. 2 Predicted collector output (no storage) of asymmetrical configuration As 60/235 at 100°C using TMY
insolation data for Sydney.
The optimum collector was approximately as shown in Fig. 1, but cusp locations
within 30° to either side would deliver nearly similar results. Optimal
geometrical concentration was about 3.1. Parabolic reflectors are not
competitive.
-------
1237
1 40
120
5 100
S. 80
Day of Year
Fig. 3. As for Fig. 2 but including 300 litres of storage and allowing the storage to range in
temperature between 60°C and 100°C. Dropouts shown would be carried by backup fuel, but at only 5% of total load,
these are infrequent. In this case, the solar fraction is 95% of load, and the underlying load pattern is clearly
distinguishable as the monthly 'step function' load. As can be seen from the Fig. 2, this is not for a desert climate,
but a climate with significant cloud cover.
Fig. 2 shows the predicted collector output of the optimal configuration from
the TMY data; the pattern requires significant backup because of poor weather
and lack of storage capacity. Fig. 3 shows the effect of including storage
with the collector array.
too
C
o
U_
W3
60 -
40 ¦
¦ No Storage
*• 300 Litres
20
0.5
1.5
1.0
2.0
Normalised Array Area
Fig. 4 Array size vs backup energy fraction supplied to load. Calculations for the AS60/235 asymmetrical
concentrator with asymmetrical load profile are shown. With storage in the system, steep rises in array cost do not
occur at solar fractions below 80-90%, indicating that very high renewable energy substitution levels are feasible.
Without storage, only about 60-70% solar fraction is optimal. The normalisation of array area is arbitrary and for
comparison only.
Fig. 4 shows that a high solar fraction can be expected from systems with some
storage. With 300 litres of storage, dumped solar energy is 16% of load for
a 90% solar fraction, and 35% of load for a 95% solar fraction. The results
suggest that an economic optimum closer
to 90% would probably be chosen, rather than the 95% used in Table 1. The
common practice of undersizing. solar systems means that necessary overcapacity
is relegated to the backup system, yet it may be cheaper to place some excess
capacity in the distributed solar equipment.
-------
1238
A similar simulation (not shown for lack of space) carried out for a constant
annual load profile showed the CPC 28 (28° half-acceptance angle) of 2.16
concentration to be optimal; a symmetrical collector would be expected for such
a profile. The 'knees' in the curves (analogous to Fig. 4) lie above the 90%
solar fraction (10% backup) level if storage is provided, and at about 75%
without storage.
It seems possible to minimise fixed collector cost for a given annual load
profile, while minimising the use of Greenhouse backup energy to below 10% of
annual load using some storage. Use of natural gas backup with a solar system
of this type would result in a C02 emissions 10 times less than a pure gas
thermal energy system, and more than 40 times less than a purely electric
thermal energy system based upon black coal.
It must also be emphasised that to provide an estimate of the maximum expected
depletion of backup energy forms such as hydroelectricity during exceptionally
poor weather periods, a full 10-20 year simulation would be required using real
hourly insolation data.
R. Winston and H„ Hinterberger, 'Principles of cylindrical concentrators for
solar energy. ' Solar Energy, 17., 253 (1975)
D.R. Mills and J.E. Giutronich, 'Asymmetrical non-imaging cylindrical solar
concentrators.' Solar Energy, 20., 1 (1978)
D.R. Mills and J.E. Giutronich (a), 'Symmetrical and asymmetrical ideal
cylindrical radiation transformers and concentrators.' J. Opt. Soc. Am., 69.
325-328 (1979)
D.R. Mills and J.E. Giutronich (b) , 'New ideal concentrators for distant
radiation sources.,' Solar Energy, 23. 85-89 (1979)
D.R. Mills, 'Non-tracking solar concentrators.' Doctoral thesis. University of
NSW (1979) .
D.R. Mills, I.M. Bassett and G.H. Derrick, 'Relative cost-effectiveness of CPC
reflector designs suitable for evacuated absorber tube solar collectors.'
Solar Energy, 36. 199-206 (1986)
ACKNOWLEDGEMENTS
The author wishes to acknowledge the financial assistance of His Royal Highness
Prince Nawaf Bin Abdul Aziz of the Kingdom of Saudi Arabia through the Science
Foundation for Physics within the University of Sydney.
PROJECT FUNDED BY THE ELECTRICITY COMMISSION OF NSW
-------
1239
THERMAL PERFORMANCE OF SOLAR CONCRETE COLLECTOR
J. K. Nayak, S. V. Bopshetty and S. P. Sukhatme
Department of Mechanical Engineering
Indian Institute of Technology, Bombay 400 076, India
ABSTRACT
This article presents thermal performance studies on solar concrete collectors
used for providing domestic hot water. Such collectors integrated with the
building structures are likely to be more cost effective than conventional
collectors having metal absorber plates. The collectors are made up of thin
concrete slabs with a number of PVC (polyvinyl) tubes embedded inside. The top
of the slab is painted black and glazed, while the back insulation is made out
of autoclave cellular concrete. A systematic experimental study of the
collectors has been conducted. A transient analysis of the thermal performance
of the collectors has also been carried out, and theoretical calculations have
been compared with the experimental data. It is seen that the theoretical
calculations are in good agreement with the experimental observations.
KEY WORDS
Solar Collector; Concrete Collector; Collector Performance;
Thermal performance.
INTRODUCTION
Flat-plate collectors are simple to construct, easy to operate and relatively
inexpensive. They are therefore widely used in low temperature applications
like water heating and air heating. At present, several different designs are
commercially available. But a large amount of metal, e.g., copper, aluminium,
G.I., steel used in these collectors. Studies on the energy inputs required
for the production of different materials indicate that metals need a large
amount of energy for their production (Payne, 1980). Secondly, individual
collector modules are connected to form a large array to meet the required
energy demand. Thus, the solar system forms a separate entity. In the long
run, it would seem desirable that solar collectors be made an integral part of
building elements like the roof and wall panels (Sonwalkar, 1984).
The present authors have undertaken experimental investigations of solar
collectors made out of concrete (Nayak, 1989). This paper discusses the
mathematical model developed to estimate the thermal performance of such
collectors. Calculations have been carried out to validate the model by
comparing its predictions with experimental data. The theoretical calculations
are found to be in good agreement with the experimental observations.
-------
1240
DESCRIPTION OF CONCRETE COLLECTORS
The concrete collectors are made from thin concrete slabs of thickness 3.5 cm
with a network of PVC tubes (20 mm O.D., .17 mm I.D.) embedded inside
(Fig. 1(a)). A layer of galvanised iron mesh on either side of the tubes
provides reinforcement to the concrete. The top of the slab is painted black
with an ordinary black-board paint, so that it acts as an absorber for the
collector. In order to reduce the top loss, an ordinary window glass 3 mm
thick,fixed in a teakwood frame,is placed over the concrete slab leaving an air
gap of 4 cm. Commercially available 3Siporex' slabs of cellular concrete are
used to support the concrete absorber plate. These slabs are light in weight
and have a low thermal conductivity. Thus they provide adequate insulation on
the back side of the collector. A sealant tape is used to hold the frame in
position and to seal the gap between the frame and the Siporex slab.
Figure 1(b) shows the tube network used in the collectors. Collectors with
tube spacing varying from 6 to 15 cm have been cast and tested for different
flow rates and inlet fluid temperatures (Nayak, 1989). $8.4cm
rWOODEN FRAME
(-RUBBER CASKET
-PUTTY
- GLASS COVER
rPVC TUBE
BLACK BOARD PAINT
•CONCRETE
r- WIRE MESH
O
.OSLO
SIPOREX SLAB E
V
tn
CM
(a) (b)
Fig. 1. (a) Cross-section of concrete collector (b) Schematic of tube network
ANALYSIS
Mathematical Model
Unlike the conventional flat-plate collector, the concrete collector has a
large thermal mass. Hence the temperature distribution in the collector has to
be estimated by solving the transient heat conduction equation with appropriate
initial and boundary conditions. From the arguments of symmetry, it is assumed
that the same temperature distribution exists between any two tubes. Hence the
region tinder consideration is reduced from the whole slab to a longitudinal
section of half-pitch width (Fig. 2(a)). The coordinate axes chosen for the
analysis are shoT-n in this figure.
From the experimental data, it is observed that the temperature gradient along
the length of the collector, the direction of the Y-axis, is an order of
magnitude less than that along the other two directions. Thus the governing
equation is
~blTl "bX + "& T/ ~bZl = yoC/k ( dT/Bt ) (1)
where p, C, k respectively are the density, specific heat and thermal
conductivity and t is the time co-ordinate.
-------
1241
Initial Condition
The initial condition is considered as the ambient temperature(Ta) at t=0,
i.e.,
T(X,Z,0) = Ta( t = 0) (2)
Boundary conditions
The top surface (ABCD) of the collector is blackened and glazed and is exposed
to solar radiation. The energy balance for the top surface can be written as
-k( "bT/bZ )l?=0 = S - Ut ( TIz=q - Ta ) (3)
where, S = 1^ (l"°veff , If is the solar radiation incident in the plane of the
collector, ("£-<*•)&££ is the product of effective transmissivity of cover and
absorptivity of absorber and Ut is the top loss coefficient. The empirical
expression for Ut has been taken from Duffie and Beckman (1980) for the purpose
of calculation.
The energy balance for the bottom surface (EFGH) is given by
-k( *T/bZ )lz=g = Uu ( TU=g - Ta ) (4)
where 6 is thickness of collector and Uu (=k^/l^) is the bottom loss
coefficient kb is thermal conductivity of the oase and lb is its thickness.
Because of symmetry, it is clear that the temperature distribution will pass
through a maximum midway between any two tubes. Hence, across the surface
BCGF, the boundary condition is
( DT/BX ) lx=4f/2 = 0 (5)
Because of symmetry, one can write for the boundaries AJKD and LMHE
( 3T/ax )lx=0 = 0 (6)
The rate of heat transfer from the tube wall, JNLMQK, can be written as
k ( 2>T/Sn )ltube wall = ha ( Tcb - Tf ) (7)
where Tcb and Tf are respectively local tube wall and fluid temperature and n
is the outward normal to the tube wall. The heat transfer coefficient, ha,
takes into account the conductive resistance of the PVC(thermal conductivity
kp) tube wall(inner and outer diameter as and DQ respectively) and the
convective resistance(hf) from the tube wall to flowing water, and is given by
1/ ha = l/hf + ( D± / 2^ ) In ( D / D± )
Water flowing through the tubes carries away heat from the tube walls. As a
result, its temperature increases and useful energy is available in the form of
hot water. The energy balance for the working fluid can be written as
1 dTf c
- ( mf Cf )-— = ha J ( Tcb -Tf ) dl (8)
N dy tube wall
where N is the number of tubes, m^ is mass flow rate and Cf is fluid specific
heat.The integration of the heat transfer rate is carried out over the tube
wall. The boundary condition for the fluid temperature is
( Tf )IY=Q = Tfi (9)
where T^ is the inlet fluid temperature.
Numerical Solution
Equations (1-9) have been solved by using an explicit finite difference
technique. The region is discretised with a uniform grid-spacing both in the X
and Z directions. In order to derive the finite difference approximation to
Eq. (1), along with its boundary conditions(Eq. 3-7), the control volume
approach has been adopted. The control volume around a grid point is chosen in
such a fashion that the grid becomes its centre, and its boundary extends half,
way between the neighbouring grid points. It may be mentioned that, as compared'
to an interior grid point, the control volume becomes half for any surface grid
point except corner ones, for which it reduces to one-fourth.
-------
1242
(a)
Fig. 2. (a) The region for the heat conduction model (b) Discretisation showing
stepped boundary.
The control volumes for grid points neighbouring the tube wall (JNL in
fig. 2 (a)) are irregular because of^the curved boundary. Hence the
corresponding difference equations need to be specially derived. For this
purpose, the following approximations are made:
(i)The curved boundary (i.e. tube wall) is replaced by a stepped boundary shown
in Fig. 2 (b). Consequently, the control volumes around the grid points
adjacent to and lying on the boundary become respectively full and half of the
regular control volume. However, there occurs a difference in the thermal mass
and the heat transfer area between the two cases.
(ii)The difference in the thermal mass of the two cases is suitably taken into
account by modifying the thermal mass of the grid points neighbouring the
boundary.
(iii)The heat transfer from the tube wall to the working fluid is accounted by
defining the following heat transfer coefficient (h_ ) for the new boundary :
'Vi = ha x Pa / P i (10)
where ha is the actual heat transfer coefficient, pa is the actual perimeter of
the tube wall and p^ is the perimeter of the new boundary.
With these considerations, the finite difference equations for grid points
lying on or adjacent to the new boundary could be written down. The solutions
of the difference equations yield temperatures at various grid points. With
these the wall temperature and hence the rate of useful heat gain(qu) can be
estimated.lt may be noted that the thermal mass of the concrete collector is
very large. Hence the concept of instantaneous efficiency does not carry any
significant meaning. On the other hand, the performance of the collector is
best judged on the basis of the daily efficiency ( ii ) which is defined as ,
ij = Qu /[Ap x Hj, ] (11)
where Q^, Ap and Bj, respectively are daily useful heat gain, absorber area and
insolation in the plane of the collector.
RESULTS AND DISCUSSION
In-order to validate the mathematical model, calculations have been carried out
for estimating daily efficiencies of solar c'oncrete collectors and the results
have been compared with several sets of experimental data.
The following values have been chosen for the parameters in the calculations,
k = 0.92 W/m-K k = 0.2 W/m-K
p = 2262 kg/m ku = 0.125 W/m-K
C = 840 J/Kg-K =0.85
Since the flow of water through the tubes is laminar, the heat transfer
coefficient from the tube wall to water is calculated using relevant
correlation (Kreith and Kreider, 1978). The transmissivity of the cover glass
-------
1243
has been calculated using standard expressions (e.g. Sukhatme, 1984), The
refractive index, extinction coefficient and the thickness of cover have
respectively been taken as 1.52, 20 m and 0.003 m.
Typical results of a whole day's calculation are shown in Fig. 3. This shows
the hourly variation of the useful energy gain of a concrete collector. The
corresponding experimental results, the solar flux incident on the collector
plane and the ambient temperature shown in the figure as well. It is seen that
there is very good agreement between the theoretical and experimental results.
Further, it is observed that the useful energy of the collector follows the
same pattern of variation as that of the solar radiation. However, unlike a
conventional collector, the useful energy from the concrete collector is
available only after some time has elapsed in the morning. This is because the
concrete collector has a large thermal mass and hence requires more time to
heat up to the operating conditions. For the same reason, the maximum of solar
flux, the phase lag being about 45 minutes. Furthermore, the collector
continues to deliver useful energy for sometime even after sunset.
Table 1 shows the theoretical^^) and experimental (ij^™) values of the daily
efficiencies of various collectors for a flow_rate of 1.2 1pm . The
corresponding values of average inlet temperature (Tf^), the average ambient
temperature (T ) and the total radiation received by the collector per unit
area in a complete day (Bj) are also presented in the table. Table 2 shows a
similar set of data for a flow rate of 0.6 1pm. It is seen that the model
developed predicts the daily efficiency reasonably well.
It is observed that the daily efficiency of a jconcrete collector varies
linearly when plotted against the parameter (Tfi~Ta)/Hx- Figure 4 shows one
such variation for a collector having a pitch of 6 cm and a flow rate of 1.2
lpm. The corresponding experimental data have also been shown in this figure.
It may be noted that this behaviour of linear variation of the daily
efficiency with (T^-T„)/B£ is similar to the dependence of instantaneous
efficiency on (Tfj-Ta)/Xj in the case of a conventional collector when its
characteristics are expressed on the basis of the Hottel - Whillier-Bliss
model. Thus, it is interesting to note that even when the colle'ctor has a
large thermal mass, its daily efficiency can be expressed by an equation of the
form
q (%) = A - B - Ta )/JLj ], where A and B are constants. (12)
V 'V ¦ -
TIME OF THE DAY , h
Fig. 3. Hourly variation of Hj,Ta & qu for
ihf=1.2 1pm, T^^=35°C and pitch= 6 cm
1000-
EXPERIMENTAL xxx
DATA
THEORETICAL
RESULTS
(Tfi-Ta)/HT, K-m2/MJ
Fig. 4. Variation of ij against
(Tfi-fa)/HT.
-------
1244
CONCLUSION
It can be concluded that the mathematical model developed predicts the thermal
performance of solar concrete collectors reasonably well. Secondly, even if the
collector has large thermal mass, its daily efficiency can be expressed by an
equation similar in form to the Hottel-Whillier-Bliss equation.
TABLE 1 Collectors daily Efficiencies TABLE 2 Collectors daily efficiencies
for a Flow Rate of 1.2 1pm
for a flow rate of 0.6 1pm
Pitch
%
T
-¦-a
ht
n
'exp
Pitch
Ta
ht
(m)
Co
(°C) (kJ/m-day) (%)
00
(m)
Co
(t) (kJ/m-day) (%)
(%)
0.06
30.5
31.5
17255
53.92
56.21
0.06
31.3
32.5
22343
51.06
50.28
31.0
37.0
24779
56.67
56.44
35.0
31.0
22686
46.06
43.71
35.0
32.0
21331
49.72
51.74
35.0
32.0
24577
47.65
44.64
35.0
33.5
24258
50.90
53.98
45.0
31.5
22478
37.76
38.67
40.0
28.5
19835
39.64
36.92
50.0
34.0
23917
35.74
32.58
45.0
32.5
22235
40.71
38.98
55.0
31.5
24630
30.54
24.06
50.0
33.5
22489
36.03
36.56
50.0
33.9
23785
39.31
36.91
0.10
28.0
31.3
23731
51.23
52.13
54.5
30.5
21862
31.02
28.64
45.0
31.0
26257
36.08
31.69
50.0
30.0
20215
29.08
24.57
0.10
27.0
31.0
24740
54.67
56.26
30.5
31.5
17255
51.64
49.74
0.15
31.3
32.5
22343
44.05
41.93
40.0
28.5
19835
37.21
33.38
45.0
31.5
22478
31.08
25.80
45.0
32.0
23773
38.33
37.84
64.5
33.5
22284
17.06
12.75
0.15
28.0
32.0
25471
47.54
45.30
31.0
37.0
24779
50.26
54.01
35.0
32.0
2.1331
42.33
38.85
45.0
32.5
22235
33.72
29.34
REFERENCES
Duffie, J.A. and W.A. Beckman (1980). Theory of flat-plate collectors. Solar
Engineering of Thermal processes. Wiley Interscience, New York, Chap. 6, pp.
197-249.
Kreith, F. and J.F.Kreider (1978). Fundamentals of fluid mechanics and heat
transfer. Principles of Solar Engineering. McGraw-Hill, New York. Chap.3,
pp.85-201.
Nayak, J.K., S.P. Sukhatme, R.G. Limaye and S.V. Bopshetty (1989) Performance
studies on solar concrete collectors. Solar Energy, 42, 45-56.
Payne, P.R. (1980). Which materials uses least energy. Chemtech, 10, 550-557.
Sonwalkar, N (1984). A study on the feasibility of utilizing building elements
and materials for making solar collectors. M.Tech. Thesis, Indian Institute of
Technology, Bombay.
Sukhatme, S.P. (1984). Liquid flat-plate collectors. Solar Energy: Principles
of Thermal Collection and Storage. Tata McGraw-Hill, New Delhi, Chap. 4, pp.
83-139.
-------
1245
Unglazed Transpired Solar Collectors:
An Analytical Model and Test Results
C. Kutscher, C. Christensen, and G. Barker
Solar Energy Research Institute
1617 Cole Boulevard
Golden, Colorado 80401
ABSTRACT
Unglazed transpired solar collectors offer a potentially low cost, high-efficiency option for
applications which involve the once-through heating of outside air such as preheating ventilation
air and crop drying. A linear model for convective and radiation heat losses is presented which
allows a simple means for performance prediction. Results of an iterative nonlinear model give
collector efficiency as a function of suction velocity, wind speed, ambient temperature, and
radiation. A small test collector has been built, and outdoor experimental results have been
compared with the model. The model has been used to investigate various perfoimance
sensitivities. Remaining research issues are identified.
KEYWORDS
Unglazed; transpired; collector; air-heating; low-cost; once-through; ventilation preheat; drying.
INTRODUCTION
Transpired porous absorbers have been studied for use in air-heating collectors to (1) increase the
convective heat transfer from the absorber to the air stream, (2) reduce infrared back losses if solar
radiation is absorbed within a thick absorber, and (3) minimize convective heat transfer from the
absorber to the glazing. In some cases, collectors were tested without cover glazings (apparently
for experimental convenience), but the intent seemed to be that in actual practice such solar
collectors would be used in recirculation systems and would include glazings. (See, for example,
Chau and Henderson, 1977.)
Once-through solar systems for direct heating of outdoor air are more efficient than typical
recirculation solar systems, because the collector inlet air temperature is equal to the ambient
temperature. Collectors with higher heat loss coefficients can be considered because the
temperature differences driving heat losses are reduced. Even unglazed backpass collectors are
possible, but efficiency may be significantly reduced by convective heat losses from the front
surface of such an absorber during windy conditions.
A relatively new type of solar collector for use in once-through air-heating systems is described
in this paper. Unlike typical matrix collectors, this collector does not employ a glazing and uses
a single thin perforated sheet instead of a thick matrix for the solar absorber. As shown in Fig. 1,
-------
1246
Win
Heated Air
Plenum
Ambient Air
Perforated
Absorber
Fig. 1. Unglazed
transpired solar
collector oriented
vertically for building
ventilation preheat.
Intake air is drawn
by the building
ventilation fan
through the
perforated absorber
plate and up the
plenum between the
absorber and the
south wall of the
building.
air adjacent to the front surface is drawn through the perforated absorber so that heat that would
otherwise be lost by convection is captured by the air flow into the collector. Although radiation
losses are somewhat higher than for a thick matrix glazed collector, this collector has higher
optical efficiency, and, we believe, is better suited for heating unfiltered outside air.
The unglazed transpired collector has the potential to be low-cost, lightweight and durable. If the
collector is site-built as part of the south wall of a new building and the fans and ducts are a
normal part of the ventilation system, then the only costs attributable to solar are the cost of
perforating the metal siding material and the cost of brackets and flashings (including installation).
A German patent (Wieneke 1981) describes an unglazed perforated roof absorber for heating
ventilation air. Schulz (1988) describes a fabric absorber used in Germany for crop drying. A
United States/Canadian company has recently begun selling unglazed perforated metal walls for
ventilation preheat (Hollick and Peter, 1990).
PREDICTIVE MODEL
A detailed analysis of heat loss theory for unglazed transpired solar collectors is given in Kutscher,
Christensen, and Barker (1991). Convective heat transfer from the front surface of the absorber
is characterized by an asymptotic boundary layer due to suction. Some convective heat will be
lost as the energy in the thermal boundary layer is swept by the wind past the end of the collector.
However, for a reasonably large collector, such heat loss has little impact on efficiency. Infrared
radiation losses from the absorber to the ground and the sky will also occur. Because the collector
is unglazed, there is no transmittance loss or glazing reflectance loss. The overall heat balance on
an unglazed transpired collector is
where a heat exchange effectiveness for air flowing through the absorber plate is defined as
As shown by Kutscher, Christensen, and Barker (1991), natural convection heat losses can be
Qu =
-------
1247
considered negligible. For forced convection with high suction ratios, we will assume an
asymptotic laminar boundary layer, and we can express the forced convection heat loss coefficient
as
U VDC
h = .82 " P p (3)
Vo
This assumes that the wind creates a two-dimensional boundary layer and that energy is lost off
a single downwind edge of the collector. (We will discuss an empirical modification to this
assumption later.)
Radiation loss occurs both to the sky and to the ground. The view factors depend on the tilt of
the absorber. We will assume that the wall behind the collector plenum is adiabatic and at a
temperature close to the absorber temperature so that radiation loss to this wall is negligible.
Assuming the absorber is gray and diffuse, the radiation heat loss coefficient is
h = e a ^I'coU ~ FcsTdiy ~ F"8Tg^ „ 4 e 0 x3 ^Tcoir (4)
The collector efficiency can be expressed as
n =
i+
hr/efa+hc
PCpV0
Note that thermal efficiency increases with increasing suction flow rate and that a heat exchange
effectiveness of less than one is manifested in increased radiation heat loss. Typical values of Tcoll
and Tsur can be used in Eq. 4 to calculate hj., and linear calculations of collector efficiency can be
made using Eq. 5 for typical operating conditions. The results presented in the remainder of this
paper, however, are based on iterative nonlinear solution of Eq. 1 using Eq. 3 and the exact form
of Eq. 4 for specific operating conditions.
The limiting (maximum) efficiency at high suction flow rate is equal to ac if the surrounding
radiant temperature is equal to the ambient air temperature. If these temperatures are not equal,
as is usually the case, the maximum theoretical efficiency is found from a simple energy balance
to be
¦n = a °ec CTamb" Tsur) (fi)
'max c j
EXPERIMENTAL RESULTS
A small test collector was constructed at SERI to acquire preliminary data to determine whether
an unglazed absorber can obtain reasonable efficiencies in the real wind. The first absorber which
we chose for testing is a black fabric because this most closely approaches the homogeneous
suction surface assumed in the model, and results for this absorber are reported here. (Preliminary
results for a perforated aluminum plate show similar performance.) The absorber area is 0.4 m x
0.4 m, and a small blower is used to drive the suction flow.
-------
1248
To measure flow rate we use one of several orifice plates (each plate corresponding to a specific
flow range) with AP measured by a capacitance-type pressure transducer. The AT across the
absorber is measured by an iron-constantan thermopile, and radiation is measured by a Kipp and
Zonen CM5/6 pyranometer. For the black fabric, we measured hemispherical reflectance using
a Perkin Elmer Lambda-9 UV-VIS-NIR spectrometer, and we determined the solar absorptivity to
be 0.86. A Gier Dunkle model DB100 infrared reflectometer was used to measure infrared
reflectance from which we deduced an emissivity of 0.88. An uncertainty analysis done in
accordance with ISO 5168 indicated an overall uncertainty of ±5% of the reading for T|.
Outdoor efficiency results are shown in Fig. 2 as test data points for a range of wind conditions
with wind speeds as high as 2.7 m/sec. Heat exchange effectiveness of the absorber was based
on direct thermocouple measurements of absorber temperature and outlet temperature and
computed from Eq. 2. The best agreement between the data and tne iterative model is obtained
if, in determining wind losses, we allow for three-dimensional flow. Because the wind striking
the collector spreads out over the surface and flows off more than one edge, we have used a
multiplier on the edge loss to adjust the model curve. As shown in Fig. 2, selecting a factor of
2.5 provides reasonable agreement over the range of suction values. We plan to better define this
effect using wind tunnel tests which will eliminate the spread in data caused by variable wind
velocities and direction.
0.8
O 0.6
UJ
o
it 0.4
III
0.2
-
-
• MEASURED DATA
1.0 x EDGE LOSS
1
2.5 x EDGE LOSS
i i i
0.02 0.04 0.06
SUCTION VELOCITY (m/s)
0.08
Fig. 2, Measured and
predicted thermal
performance of a
vertical unglazed
transpired solar
collector. Collector
size = .4 m x .4 m,
Tamb = 21°C,
t — t -1 «°r
* sky ~ 1 amb
T - T
x end * amb'
= 2.3 m/s, and
I,. = 792 W/m2.
MODEL RESULTS
The iterative model is currently being used to investigate various performance sensitivities.
Fig. 3 shows predicted thermal performance for a. vertical unglazed transpired solar collector as
a function of suction velocity for wind speeds of 0 and 5 m/s and absorber emissivities of 0.9 and
0.2. As suction velocity decreases, the effect of wind speed on collector efficiency increases,
especially for the low emissivity absorber. The benefits of the low-emissivity absorber generally
increase as the suction velocity decreases.
As the ambient temperature drops, so does the absorber surface temperature, and, because of the
nonlinearity of radiation heat loss, the collector efficiency increases. This effect is of course true
for any solar collector, but is especially so for this one. Chau and Henderson (1977) and others
have noted the magnitude of this effect for matrix absorbers.
Not quite so obvious - but evident from runs of the model - is that for decreasing levels of solar
radiation, efficiency can either increase slightly or decrease. If the average surrounding
temperature to which the collector radiates is near the ambient temperature, as insolation level
-------
1249
decreases, efficiency increases slightly due to the nonlinearity of the radiation heat loss term. For
very low surrounding temperatures, however, efficiency will decrease at lower solar radiation levels
because of the fixed radiation loss between ambient and the surrounding temperature. The
orientation of the collector and the emissivity of the surface strongly affect this dependence. Note
that whereas this fixed radiation loss term can be reduced by using a lower emissivity surface, it
is not affected by using multiple layers or a deep matrix.
O 0.6
t 0.4
Ul
70
60
50
o
a
40
E
¦
30
h-°
20
10
0
0.01
0.02 0.03 0.04
SUCTION VELOCITY (m/s)
0.05
0.06
0.01 0.02 0.03 0.04
SUCTION VELOCITY (m/s)
0.05
Fig. 3. Predicted
thermal performance
of a vertical ungiazed
transpired solar
collector as a
function of suction
velocity, absorber
emissivity, and wind
speed. Collector
size = 3 m x 3 m,
Tamb = 10°C,
^sky _ Tamb ' ^
Tgnd = Tamb> ?"d
I, = 700 W/m2.
(A) Efficiency versus
suction velocity.
(B) Suction air
temperature rise
versus suction
velocity.
For a typical ventilation preheat system with a suction velocity of 0.05 m/s, the collector
temperature rise is approximately 12°C at an insolation of 700 W/m2. Efficiencies are
approximately 78% and 84% for absorber emissivities of 0.9 and 0.2, respectively.
CONCLUSIONS
An analytical model accounting for radiation and convection losses has been developed to predict
thermal performance for ungiazed transpired solar collectors. A small test collector with a fabric
absorber has been tested at SERI, and initial outdoor results give efficiencies that are in reasonable
agreement with model predictions. The model indicates that high efficiencies can be obtained for
typical ventilation preheat flow rates and modest collector temperature rises. At lower flow rates
collector temperature rises are higher, but efficiencies are lower and wind effects are more
important. Selective surface absorbers would be useful in achieving higher collector temperature
rise.
-------
1250
FUTURE WORK
SERI is currently performing both experimental and analytical studies to better determine wind
heat losses and to optimize hole size and spacing. We are also investigating issues of suction flow
uniformity. In addition to these basic issues, we plan to utilize our analytical model to predict
overall performance of unglazed transpired ventilation preheat systems in a range of United States
climates. Finally, we will conduct field testing of full-scale absorbers at SERI.
ACKNOWLEDGMENTS
We thank Ed Hancock, Hassan Rafie, and Jay Burch for their assistance in setting up the
experiment and collecting data, Gary Jorgensen and Cheryl Kennedy for measuring the absorber
optical properties, and Ren Anderson, Rob Farrington and A1 Lewandowski for review comments.
We are also indebted to Robert Hassett of the Department of Energy Office of Building
Technology, Building Equipment Division, for providing the funding for this work.
NOMENCLATURE
A„ =
F =
xcg
p =
4 r. g
collector area (m)
specific heat at constant pressure
(J/kgK)
collector-to-ground view factor
collector-to-sky view factor
convective heat transfer coefficient
(W/m2K)
radiative heat transfer coefficient
T =
hj =
(W/m2K)
I,. = solar insolation incident on the
collector (W/m2)
Lc = collector length (m)
Qu = useful collected energy (W)
Tatrih = ambient temperature (°C)
Tcoii = collector temperature (°C)
avg
=
v0 =
ac =
ec =
eHX ~
T1 =
V
P =
a
collector output temperature (°C)
surroundings temperature,
(FcsTVvV1'4
average radiative temperature,
(Tcoll+Tsur)/*
free stream velocity (m/s)
suction velocity (m/s)
collector absorptance
absorber surface emissivity
absorber heat exchange
effectiveness
collector efficiency = QJlckca0
kinematic viscosity (m2/s)
density (kg/m3)
Stefan-Boltzman constant,
5.7 x 10-8 W/m2K4
REFERENCES
Chau, K. V., and S. M. Henderson (1977). Performance of a Matrix Solar Collector for Heating
Air, Transactions ASAE. 558-561.
Hollick, J. C. and W. Peter (1990). Method and Apparatus for Preheating Ventilation Air for a
Building, Patent No. 4,934,338, United States.
Kutscher, C. F., C. B. Christensen, and G. M. Barker (1991). Unglazed transpired solar collectors:
heat loss theory. Solar Engineering 1991. Proceedings 12th Annual ASME International Solar
Energy Conference. American Society of Mechanical Engineers.
Schulz, H. (1988), Das Solarzelt [The Solar Tent], report published by Landtechnik Weihenstephan
der TU, Munich, Federal Republic of Germany.
Wienecke, F. (1981). Solardach Absorber, Patent No. 29 29 219, Federal Republic of Germany.
-------
1251
THERMAL ANALYSIS OF A REFRIGERANT-FILLED SOLAR COLLECTOR
J. C. Baltazar and E. Torres
Instituto de Investigaciones Cientificas
Universidad de Guanajuato
L. de Retana #5, C.P. 36000, Guanajuato, Gto.
MEXICO
ABSTRACT
A mathematical model to predict the performance of a
refrigerant-filled solar collector operating as the evaporator of
a solar assisted heat pump system has been established. The
thermal behavior of the solar collector is investigated as a
function of the refrigerant quality and solar collection area. The
thermal analysis of the collector is developed by dividing it into
two parts; one of them including the external part of the
collector and its interaction with the environment, and tteother the
internal part including the phase change that occurs in the
refrigerant as a consequense of the operation of the Thermodynamic
cycle. The results of this study present alternatives to improve
the efficiency of the solar device, selecting the refrigerant in
according to the specific work conditions and applications.
KEYWORDS
Thermal analysis, Refrigerant-filled Solar collector, Plate
temperature distribution, Quality refrigerant, Two phases.
INTRODUCTION
Several studies on the behavior of flat solar collectors working
with refrigerants have been carried out. All the research done to
date has focused on a global analysis with respect to the energy
balance of the solar device. So far, the model most frequently
used is a simplification of the useful heat equation for flat
solar collector from which the behavior of the evaporator—
collector has been studied throughout the year in different
locations and with diverse obja±ives, as is shown in the following
references (Baltazar, 1989; Chaturvedi, 19S4; O'Dell, 19S4).
A common consideration among the studies mentioned earlier is
that the fluid temperature at the collector inlet corresponds to
the saturation temperature (assuming the evaporation process
occurs under conditions of saturation). In this case, the useful
heat from the collector is matched to the enthalpy change which
the refrigerant undergoes inside the evaporatoi—collector.
In order to simulate the thermal behavior of the refrigerant
inside the evaporatoi—collector, three different sections of the
-------
1252
total area were established, in this case the refrigerant goes
from a given quality to the state of saturation, and eventually,
leaves the collector at a certain superheating temperature.
The suggested model of the behavior of the evaporator—collector is
described and some results of the simulation runs are shown.
THERMAL ANALYSIS OF THE EVAPORATOR-COLLECTOR
The analysis of the collector is divided into an external and an
internal part. The former depends on the variations of the weather
conditions such as, solar radiation, ambient temperature and wind
velocity. The study of the internal part involves the changes the
fluid undergoes as a consequence of the operation of the
components of the solar heat pump.
The external analysis is based on the thermal balance for flat
plate collectors. This applied to an uncovered collector with
freon as working fluid, gives the expresion:
Qu = Ac [ Bsot -UL,(Tp-Ta)] (1)
where the global coefficient of energy losses (Ul) for this type
of collector is a function of the heat transfer coefficient due
to the wind (O'Dell, 1984).
On the other hand, the internal analysis is described on the basis
of convection heat transfer where the film heat transfer
coefficient in the region of evaporation is calculated as a two-
term sum, using the Chen correlation (Hsu, 1976). One of terms
is.caused by nucleate boiling vaporization and the other by
convection.
To simulate the behavior of the collector, this is divided in two
regions, one in which the evaporation takes place and the other
where superheating occurs. Furthermore, the evaporation region is
Superheating
Evaporation , .
SECTION
ATsc
0.95
XI
Fig. 1. Regions of the solar collector for the thermal analysis.
subdivided into two sections: Section I considers that the mass
vapor quality of the refrigerant range from 0.95 to 1 (almost
vapor phase) and section II, from inlet mass vapor quality to
0.95. This is shown schematically in Fig. 1.
For the first section with a two-phase flow, the useful energy
gained from collector can be calculated by the eq. (1) or from the
-------
1253
convection heat transfer, which is
Du = hd Ai (Tp - Ta) (2)
Considering the change of enthalpies, in the same section, the
useful heat is found as
Qu = m(0.05)\ (3)
In order to compute the plate temperature and the required
collection area, an iterative process is used until the value of
the Qu in the eqs. (1),(2) and (3) is the saiu.
On the other hand, when section II is evaluated, the mass vapor
quality is included between x and Ax, so it is necesary define a
small value of Ax.
The evaluation process is similar to that described above, with
the difference that the eq. (3) is written as
Qu = m Ax X (4)
In this form, an iterative method is required for each increment
in Ax, obtaining a increment of collection area and the
corresponding plate temperature. The methodology is continued to
cover the total collection area, such that, finally,the mass vapor
quality of refrigerant at the inlet the collector is obtained.
To determine the useful heat gain of the superheating region,
given a ATsc (superheating grade) the heat balance in the internal
part can be written as
Qu = m Cp (ATsc) (5)
or considering the plate temperature by
Qu = hg Ai (Tp - Tf) (6)
The heat transfer coefficient in the superheating region is
calculated by the Dittus-Boelter equation.
For the external side, the useful heat can be computed for a solar
collector equation, proposed by Willier (Duffie, 1974)
Qu = Asc F' ( Bsc* - UMTf - Ta) ) (7)
In a similar manner to the other region, the calculation of the
useful heat is achieved using an iterative process. All
thermod^namical and transport properties necessary in each section
are calculated by equations fitted to reported values.
RESULTS
The simulation of the evaporatoi—collector was performed for a
finned coiled pipe geometry with a solar collection area of 4,4
m®, which is considerated to be painted black matt, and no
insulation or glass cover was used. The refrigerants selected
were R-22, R-12 and R-ll.
-------
1254
The results obtained show the variation of the useful heat and
the plate temperature as a function of the inlet mass vapor
quality of the refrigerant and the solar collection area. Also
the collector efficiency was evaluated for each one of the
working collectors.
In Fig. 2, the plate temperature distribution along the collector
area is shown. From this figure, R-ll reaches the highest values
of plate temperature due to thermodynamic properties. All
temperature profiles were obtainedatthe same mass flow-rate and
solar flux.
320
310
300 .
£ 290 ,
0)
&
¦p 280 „
CD
%
R-ll
R-12
R-22
260
Fraction of area
Fig. 2. Distribution of plate temperature against the fraction of
collector area. Gs=700 W/m*, m=0.018 kg/s and Ta= 293 K.
The quality of refrigerants determinsd through the
collector—evaporator is plotted against the fraction area of
collection. In the Fig. 3 the different values of the inlet
quality for the three refrigerants are observed. These values are
related with the different sizes of required areas for the process
under the same operating conditions.
Figure 4 shows the accumulated useful heat achieved for each
fraction of the area of collector. It can be observed that R-22
removes more useful heat than the other fluids. Under the same
operating conditions R-ll absorbs the least amount of available
energy.
In Fig. 5 the efficiency of the collector for the three
refrigerants is plotted. The behavior observed on this figure
denotes that the efficiency curves have roughly slope values.
Therefore, it is posible to define the thermal behavior of the
refrigerant-filled collector with an unique characteristic curve.
CONCLUSIONS
The simulation of the device allows identifying of other alternatives
to improve its efficiency. This can be done with an appropiiate
-------
1255
1.0
1
o-
0.2 0.4 0.6
Fraction of area
0.8
Fig. 3. Mass vapor quality of the fluids against the fraction of
collector area. Gs=900 W/m®, m=0.0216 kg/s and Ta=293 K.
irt
rtf tj
V r
rC w
3 I
0)
s
4.50.
4.00.
3.50,
3.00.
2.50,
2.00,
1.50.
1.00.
0.50,
0.00
•1a Cm
R-12
R-ll
*4
0.2 0.4 0.6
Enaction of area
0.8
Fig. 4.
Accumulated useful heat versus the fraction of collector
area. Gs=900 W/m®, m=0.0216 kg/s and Ta=293 K.
selection of the refrigerant, according to the available solar
collection area, as a function of the weather conditions and the
intended application.
Once the thermal model of the evaporator—collector is integrated
into a Thermodynamic cycle of the heat pump, knowing the outlet
temperature of the refrigerant will be posible to achieve a better
prediction of the compressor performance and, thercrfore, also a
better prediction of the cycle performance.
-------
1256
1.200 ^
R-22
1.100 .
1.000 .
R-12
0.900 t
V
0.800 ^
0.700 „
R-ll
0.600
(Tf-Ta)/Gs
Fig. 5. Efficiency curves of the collector—evaporator for R-ll,
R-12 and R-22.
NOMENCLATURE
Ac Solar collector area [m*]
At Internal area [m23
Cp Heat capacity [J/kg K]
F' Efficiency factor
Gs Solar flux [W/m2]
hd Two-phase convection heat transfer coefficient [W/m* K]
hg Gas convection heat transfer coefficient [W/m2 K]
m Refrigerant mass flow-rate [kg/s]
Qu Useful heat [W]
Ta Ambient temperature [K]
Tp Plate temperature [K]
ATsc Superheating grade
Ul Heat losses overall coefficient [W/m2 K]
x Refrigerant quality
a Absorptance
\ Latent heat of vaporization
REFERENCES
Baltazar, C.J.C., R.E. Torres and J. Cervantes de G. (1789).
Theoretical performance of a solar flat collector, as part of a
solar heat pump. Proc. of the Natl. Solar Energy Soc. CtiexicoJ,
Morelia, (Mexico). 216-219. (In Spanish).
Chaturvedi, S.K. and J.Y. Shen. (1984). Thermal performance of a
direct expansion solar—assisted heat pump. Solar Energy, 33,
154-162.
Duffie, J.A. and W.A. Beckman. (1974). Solar energy thermal
processes. Ed. Wiley Interscience, New York. pp. 122-143.
Hsu, Y. (1976). Transport processes in boiling and two-phase
systems. Mc Graw-Hill. U. S. A. pp. 139-142.
~'Dell, M.P., J.W. Mitchell and W.A. Beckman. (19S4). Design
method and performance of heat pumps with refrigerant-fi1 led
solar collectors. J. of Solar Energy Engineering. 106. 159-164.
-------
2.4 Collectors IV
-------
-------
1259
EFFICIENCY OF A SOLAR COLLECTOR USED IN
A TYPICAL TRANSPARENT HONEYCOMB INSULATION
Tomohiro Mizuno, Hiroyuki Shimizu, Shunji Asai
YAZAKI Corp., LTD., Air conditioning R&.D Lab.,
1370 Koyasu-cho, Hamamatsu-shi,
Shizuoka-ken,435, Japan
ABSTRACT
The reduction of the radiation and the convection heat loss from the upper side of a
solar collector is very effective in improving the efficiency of the solar collector. The
thermal radiative heat loss from an absorber can be reduced by the application of
so-called selective layers, this method has been in use. And the thermal convective
heat loss from an absorber can be reduced by using honeycombs between the
absorber and the cover plate with transparent materials, and the Multi-covers with
glass or plastic films. There have been many reports about the characteristics of
these structural solar collectors. However, there are not many practical models due to
their cost and production problems.
We suggested using a poly-carbonate honeycomb, one-body, structural sheet in
place of a cover plate in a general single glazing solar collector. In this paper the
characteristics of the mentioned material and the efficiency of the solar collector
when this material is used have been investigated theoretically and experimentally.
The efficiency of the collectors used the mentioned sheet is higher than that using
the single tempered glass at the middle and high temperature level when (Tp-Ta)/I is
more than 0.07.
The mentioned sheet has many slats, which give directionality. When the direction of
the slats is pointing south-north, the long side of the slats are inclining. The effect
of reducing the natural convection heat loss and the efficiency of the solar collector
is higher than when the direction of the slats is pointing east-west.
KEYWORDS
Slat; Flat-plate solar collector; Solar collector efficiency; Natural convection heat
loss; Transmittance.
INTRODUCTION
An efficiency of the solar collector is determined by the balance of the absorption
value of a solar energy incident and heat loss values. A general flat-plate solar
collector is formed by a glazing glass, an absorber plate, an insulation, and a case.
The reduction of the radiation and the convection heat loss from the upper side of the
solar collector is very effective in improving the efficiency of the solar collector. The
thermal radiation heat loss from the absorber can be reduced by the application of
so-called selective layers) this method has been in use. And the thermal convection
heat loss from the absorber can be reduced by using honeycombs between the
absorber and the cover plate with transmittance materials, and the Multi-covers with
glass or plastic films. There have been many reports about the characteristics of
these structural solar collectors. However, there are not many practical models due to
'receding page blank
-------
1260
their cost and production problems.
Poly-carbonate honeycomb, one-body, structural sheet is a transparent material (
hereafter "slat-sheet" is used ), that has been in use for skylights, windows, roofs
of swimming pools, and so on. This sheet handles easily because it is very light and
rigid, and it has high-impact hardness. It has high insulative value, because it is a
double structure. The cost is lower than that of double glazing glasses.
We suggested using the slat-sheet in place of a cover plate in the general single
glazing solar collector. In this paper the characteristics of the slat-sheet and the
efficiency of the solar collector when this material is used have been investigated
theoretically and experimentally.
SLAT-SHEET
The slat-sheet is an extrusion
strangpressing production. It is a Multi-wall
rigid sheet like a Harmonica, has twin walls
and many partition walls", they are
perpendicular to the twin walls, and they are
parallel to each other. The structure of it is
shown in Fig. 1. The parallel partition walls (
hereafter "slat" is used ) serve structural
purposes. Moreover, the structure brings
about its high insulative value. The weight of
this sheet per square meter is less than 1/2 of
that of the single tempered glass. It s
elasticity and impact hardness is higher than
that of single tempered glass.
Therefore, it handles easily when we produce and install it. The honeycombs
researched until now have production problems, and they are expensive. However
the slat-sheet is suitable for mass production. And it s cost is low.
W=13.5mm
L=18.0mm
Fig. 1. Cross-sectional view of the
slat-sheet.
TRANSMITTANCE OF THE SLAT-SHEET
The solar radiation transiting the slat
is influenced by the reflection and the
absorption of the slat. The
transmittance is not reduced by the
reflection of the slats because the
radiation reflected by slats proceed to
the absorber. Part of the radiation is
absorbed when it transit through the
slats. When the angle of incidence
increases, the transmittance of the
slat-sheet is reduced while increasing
the number of slats and distance that
the radiation has to transit through.
The relation between the
transmittance and the angle of
incidence of the slat-sheet is shown in
Fig. 2. The angle of incidence is
represented by two parameters, fi,Q, (
see Fig. 3 ) due to the slat's
directionality. The curves in Fig. 2 are
the theoretical curve based on
characteristics of a Poly-carbonate.
W
O
H
E->
6 =0
H
ANGLE OF INCIDENCE & *
Fig. 2. Transmittance of the slat-sheet.
-------
1261
The transmittance is higher when the
radiation is parallel to the slats ( 0 =90* )
than when it is perpendicular to the slats
( 6 =0° ). For example, the difference
value is 1.896 at /?=60° . Generally, the
solar collector is installed facing south
on a tilted surface. When the slat-sheet
is used, there are two cases for
installation. In one case, the direction of
slats is pointing south-north ( hereafter
"slat-SN" is used ), in the other case, it
is pointing east-West ( hereafter
"slat-EW" is used ). When the efficiency
of the solar collector is evaluated, the
coefficient of dirt becomes 2~59f> ( Duffie,
1974 ). Therefore the difference of the
performance concerning the
transmittance between the two casescan
be neglected.
PERFORMANCE OF THE SLAT-SHEET COLLECTOR
Specifications of the applied solar collector for the theoretical calculation and
experimental of the efficiency of the slat-sheet collector is shown by TABLE 1.
TABLE 1 Specifications of the applied solar collector
Items
Specifications
Size
1002X2002X78
Collector area
1.91m2
Absorber type
Tube-in-sheet
Absorber surface
Black stainless selective surface
(absorptance a =0.93, emittance /S=0.11)
Insulation
Glass wool
Equations of heat transfer balance are constructed as follows.
I Ap (rgi rgzap) =Aphp-g (Tp-Tgi) +Aphb+e (Tp-Ta)
IAg (rgz) agi+Aphp-g (Tp-Tgi) =Aghg (Tg.-Tgz)
IAgaga+Aghg (Tg.-Tgz) =Aghw (Tgz-Ta) +Aghr,g-a (Tga-Tsky)
h p-g = h c,p-g i + h r,p-g,
hg=hc,g,-g2+hr,g>-g2+hs,g»-g2
hb+e=hb+he (Ae/Ap)
hb=kb/tb
he=ke/te
hs=ks (ts/W) /L
Symbols
I : Total solar energy incident [ W/m2 1
rx : Transmittance of x element
ax: Absorptance of x element
Ax : Area of x element [ ms ]
Tx : Mean temperature of x element [ °C ]
hc,x-y : Natural convection heat transfer coefficient between x-y element [ W/ms°C ]
hr,x-y : Radiation heat transfer coefficient between x-y element [ W/m=°C ]
Fig. 3. Two parameters of angle of
incidence.
-------
1262
hw : Forced convection heat transfer coefficient in regard to wind [ W/m2"C ]
h b : Heat transfer coefficient through back [ W/razoC ]
he : Heat transfer coefficient through edge [ W/mE"C ]
hs : Heat transfer coefficient in regard to conduction of slats [ W/m2"C ]
Kx : Conduction coefficient of x element [ W/m"C ]
tx : Thickness of x element [ m ]
Subscripts
a: Ambient air
sky:sky
p : absorber plate
g i : Bottom sheet of slat-sheet
g2 : Upper sheet of slat-sheet
b : Back of collector
e : Edge of collector
s : Slat
The experimental equation of Hollands and others ( 1976 ) for the natural convection
heat transfer from absorber to slat-sheet ( hc,p-g ) as well as the equation of Ozoe
and others ( 1982 ) for the natural convection heat transfer in the slat-sheet ( h
c>gi~gs ) are used, and otherpartsof calculation depend on the formula of reference (
ISES, 1978 ). The experiment has been carried out on the roof based on ASHRAE
STANDARD 93-77R method.
The theoretical efficiency curve and experimental plots of the solar collectors used
the slat-SN and the slat-EW are shown in Fig. 4, Fig. 5. After adjusting the
parameters of the total solar energy incident ( I ), the ambient temperature ( Ta ), the
velocity of wind ( V m/sec ), the angle of inclination ( S * ) with experiment conditions
the theoretical calculation were carried out. The theoretical values are in good
agreement with experimental values as shown in Fig. 4 and Fig. 5.
1 r 1 1 1—
Ta=10.0
1=815
V=3.0
S=60
Ta=10.0 .
1=815
V=3.0
S=60
,9
,8
.7
,6
5
,4
,3
«¦
.2
.1
0,
.05
,05
0
.10
(Tp-Ta) /1
(Tp-Ta)/1
Fig. 4. Efficiency curve and
experimental results of
the slat-SN collector.
Fig. 5. Efficiency curve and
experimental results of
the slat-EW collector.
-------
1263
The efficiency curve of the
solar collectors used the single
tempered glass, and the
slat-SN, and the slat-EW are
shown in Fig. 6. The efficiency
of the collectors used the
slat-SN and the slat-EW are
higher than that used the
single tempered glass at the
middle and high temperature
level when (Tp-Ta)/I is more
than 0.07. Heat loss values are
shown in Fig. 7. The heat
loss Value of the collectors used
the slat-SN and the slat-EW is
lower than that used in the
single tempered glass. The heat
loss value of slat-SN collector is
smaller than the single
tempered glass collector at the
rate of 37.1%, and the value of
the slat-EW is smaller than that
at the rate of 33.6 % , in
accordance with S=60 * ,
(Tp-Ta)=60deg. The efficiency
of the solar collector using the
slat-SN is high as well as the
heat loss value is small as
compared with the collector
using the slat-EW.
According to this result, when
the direction of the slats is the
south-north, inclining a long
side of the slats, the effect of
reducing the natural convection
heat transfer and the efficiency
of the collector are higher than
when the direction of the slats
is the east-west. This result is
close to the experimental result
of the literature by Symons
and others ( 1982, 1983 ).
CONCLUSIONS
The heat loss value at S=30° is
shown in Fig. 8. Comparing with
Fig. 7, the heat loss of the
slat-SN is smaller than that of
the slat-EW, but the difference
between them is reduced from
5.3% to 2.2% at (Tp-Ta)=60deg.
The effect of reducing the
natural convection of the slats
is not enough. The aspect ratio
( L/W ) of mentioned sheet is A
is 1.0 (see Fig. 1). The effect of
Ta=20.0
I =930
Tempered glass
V=3.0
Slat-SN
Slat-EW
.05 .10
(Tp-Ta) /1
Fig. 6. Efficiency curves of
the slat-sheet collectors and
the single tempered glass
collector.
a
\
j*
K
W
h
W
E-<
Eh
600
400
200
100
60
20
Ta=20.0
I =930
V=3.0
S=60
Tp-Ta (deg)
Fig. 7. Heat loss of the slat-sheet
collectors and the single
tempered glass collector when
angle of inclination is 60deg.
-------
1264
reducing the natural convection
become effective when A £ 1.0
due to references ( Symons,
1982, 1983 ). The transmittance
is reduced when the distances
600
among the slats are decreased.
And the convection sensitively
depend on h ( ISES, 1985 ). The
efficiency of the slat-sheet
collector will be improved due
to theoretical anedysis based on
these factors.
0,
0 20 40 60 80 100
Tp-Ta (deg)
Fig. 8. Heat loss of the slat-sheet
collectors and the single
tempered glass collector when
angle of inclination is 30deg.
REFERENCES
Jhon A. Duffie and William A. Beckman (1974). Solar Energy Thermal Processes, 156.
K. G. T. Hollands and others (1976). Trans. ASME of Heat Transfer, 98, 189.
Ozoe H. and others (1982). Proc. of 7th Int. Heat Transfer Conf. Munhen, Vol. 2, 257.
JSES (1978). Foundations and Applications of Solar Energy, 70.
Symons J. G., Peck M. K. (1982). Proc. of ISES Australia Newzealand section 21st Conf.,
103.
Symons J. G. (1983). Proc. ISES Perth Conf., 748.
R. h. D. Cane and others (1977). ASME of Heat Transfer, Vol. 99, 86.
JSES (1985). Handbook of Solar Energy, 180.
-------
1265
USB OP TRANSPARENT INSULATION AS COVER IN AN INTEGRATED
COLLECTOR CUM STORAGE SYSTEM
SURENDRA KOTHARI*, N.K.BANSAL** AND N.S.RATHORE*
* Renewable Energy Centre,Co I Iege of Technology and
Agricultural Engineering,Udaipur-313001, Rajasthan India.
** Centre for Energy Studies, Indian Institute of Technology,
De1h i, Ind ia.
ABSTRACT:
A new class of materials known as transparent insulation
materials with good transmitti v i ty to solar radiation and
very low thermal conductivity have been investigated for use
as cover in integrated col 1ector-cum-storage system. A
particular class of material capillary structure yield 38%
overall efficiency in comparision to a single glased system
which yields an efficiency of 21% only. Detailed time
dependent calculations have been made to show the
achisvsabla temperatures in such systems under winter and
summer conditions of Delhi,
Key words : transparent insulation, capillary
structure, pmma foam
INTRODUCTION :
In nearly all solar thermal conversion systems, the solar
radiation is absorbed by a black plate(absorber) which
converts input light energy into heat.
To achieve high temperatures and high efficiencies, the
plate has to be insulated against heat losses to the
environment. Glass has a high transmission for the solar
radiation,but as the thermal conductivity is slightly
more, result is, heat loss to the environment.
To reduce the heat losses from top transparent insulating
material can be used as cover.
Transparent Insulation Material:
Transparent insulation material has high solar transmittance
and low thermal conductivity. The collector is insulated on
the back side with conventional insulation. By providing
transparent insulation, front side losses can be minimized.
Advantages of transparent insulation material over glass
cover is more in case when thermal losses are significant as
compared to incoming radiation. These transparent insulation
materials are available with transparency of 0,6 and heat
loss coefficient U = 0.8 w/m k. Some of the transparent
insulation materials, aieag with their optical and thermal
properties, are given in tabie-1.
-------
1266
Table-1 Characteristic data of some of the Transparent
Insulation Materials.
Mater ials
Transmittance
Heat loss
X
coefficient
W/m°K
PMMA FOAM
58
3. 6
THIN FILM STRUCTURE
66
1.28
CAPILLARY STRUCTURE
60
0.8
AEROGEL (AIR)
53
1-. 25
AEROGEL (EVAC.)
53
0.8
These transparent insulation materials offer a new potential
for using solar energy in thermal applications such as:
1- Integrated co11eotor-cum-storage system for domestic hot
water system.
2- Large area solar ponds for hot water and process heat.
3- Passive solar energy utilization in buildings.
The analysis has been done for a system in which collection
of solar energy and its storage is done in one system.
Numerical calculations have been performed for different
climatic conditions of Delhi.
INTEGRATED COLLECTOR-CUM-STORAGE SYSTEM:
The integrated storage collector unit (Fig.l) is a flat box,
the front surface of which absorbs solar energy . The
collector is insulated on the front side with the
transparent insulation material and on the back and side
with conventional opaque thermal insulation. For the
collector cum storage system, it is necessary to consider
the time variation of the input parameters i.e. the solar
radiation S(t) and the ambient temperature Ta( C) and the
resulting, temperature of water in the collection tank.
.TRANSPARENT INSULATION
COVER
•COLLECTOR SURFACE
WATER TANK
-I NSULATION
A
w
FIG.1 INTF.GRATED STORAGE COLLECTOR UNIT.
-------
1267
The heat balance equations at the absorber and for the water
mass (Mw) in collection tank can be written as
(<£ ) S = hf (Tp - Tw ) + U ( Tp - Ta )
and
dTw
hf ( Tp - Tw) = Mw Cw + Ub ( Tw - Ta)
dt
eliminating Tp from above equations
dTw
+ aTw = b
dt
where
Ub
(hf + U) + U
hf
a
Mw Cw
{hf + U )
hf
and
Ta Ub
oCCS + U Ta + (hf + U)
hf
b =
Mw Cw
( hf + U )
hf
Though, strictly speaking, the constant b is a function of
time, over a small time interval it could be assumed to be
constant. The solution of equation can then be written as:
Tw (tn) = Tw tn-1.exp(-a(tn - tn-1))
+ b/a ( l-exp(a(tn - tn-1))
If the water is withdrawn from the system at the instant tn,
then we can write the following equation:
Cw dmw ( Tw - Ti ) = -Mw Cw dtw
This equation can be written in the form;
[TJ .= _ J_(r/¥wflfmhJ
'Toj Tu>~ 7i Jo
Integrating we get
Tj(€n)~Ti Tu>&k) -TL(*h)J (- )
The total useful energy gain by the system can be calculated
from the equation
and efficiency of the system
-------
1268
Useful energy gain
= x 100
Incident solar energy
Qu
x 100
(«
(Joules)
(%)
GLASS
51. 87
3.76X10S
21.56
PMMA FOAM
43. 68
3. 37X id5
18. 75
THIN FILM STRUCTURE
58. 08
5. 65X106
32. 40
CAPILLARY STRUCTURE
57. 65
5.8 X106
36.26
AEROGEL(AIR)
49. 55
4.58X icf
26. 26
AEROGEL(EVAC.)
52. 52
5. 12X10&
29. 36
With Aeroge1tEvac) the increase in water temperature is high
during off sunshine hours, where^as it is less during peak
sunshine hours. When Aerogel(Air! is used the water
temperature is more than the water temperature when glass is
used, during off sunshine hours,however, there is a slight
decrease in water temperature during peat; sunshine hours.
-------
1269
FIG2: TEMPERATURE OF WATER WITH DIFFERENT TRANSPARENT
INSULATION COVER.
a
2
LU
h-
10 12
HOURS
¦* AMB TEMP.
-o—PMMA FOAM
RAD/10
GLASS
X THIN FILM STRUCTURE
CAPILLARY STRUCTURE A AEROGEL (AIR )
-B-
AEROGEL ( EVAC. )
-------
120
105-
90-
75-
60-
45-
3 0-
15-
0
FIG. 3. Temperature of water with different transparent insulation
RAO —BAD/ 10
TG —GLASS
TA —AMB TEMf*
TG (PF)—PMMA FOAM
TIM(AA)—AEROGEL (AIR )
TIM(AE)—AEROGEL (EVAC.)
TIM(CS)—CAPILLARY STRUCTURE
TIM(TF)—THIN FILM STRUCTURE
DELHI (JUNE )
TIM tTF)
TIM (CS )
TIM (AE)
RAD
TG —TIMtA
TG ( PF )
HOUR
-------
1271
It is observed from the Fig.2-3 that PMMA foam, as a cover
material, is not beneficial. Temperature of water is even
less than that obtained with glass. From table-2 the useful
gain of energy is the highest when capillary structure is
used. It is 5.29x10 and the efficiency of the system
is as high as 38%. the useful energy gain is only 3.26 x 10
joules and efficiency is 21.5% when glass is used.
It is observed that the efficiency of the system with cover
material is not affected by particular environmental
cond i t i ons.
CONCLUS ION :
It can be concluded that the use of Transparent
insulation materials as a cover of integrated colleotor-cum-
storage system is highly beneficial. Eff ic iency of the
system is almost double when capillary structure is used.
Temperature of water is higher even during off-sunshine
hours. So cost of additional storage in traditional water
heating system can be avoided.
NOMENCLATURE
CP Specific heat of Transparent Insulating Material(KJ/Kg°C)
Cw Specific heat of water (KJ/Kg °C)
dT Differential temperature change in the temperature of
the water ( °C1
Eb Emissive power of black body source (KJ/h m'")
Ki Thermal conductivity of the transparent insulation
(KJ/hm°C)
Mw Capacity of the system(kg)
r coefficient of reflectivity of transparent insulation
T*^ TLnhi. )
j~ul HYysta.ntane.ons Cjdfesu LYX "the* systern
REFERENCES '
1. Duffie, J.A. and U.A.Beckman(1980). Solar Engineering
of Thermal Process. Vilify & Sons, New York.
2. Pfluger, A. (1988) and S.Spektraler Trahlungstransport
in Transpaenten Warmedammungen, Dessertation. Univ.
Freiburg, Se1bstver1ag.
3. Kothari S. (1989). Transparent Insulation in Solar Water
Heating Systems. M.Tech.Project Report.
-------
1272
FUTURE DEVELOPMENT IN THE FIELD OF TRANSPARENT INSULATION SYSTEMS
W. Platzer, V Wittwer
Fraunhofer-Institut fur Solare Energiesysteme
Oltmannsstr. 22, D-7800 Freiburg
ABSTRACT
The utilization of solar energy in the area of active and passive thermal systems needs cover
or glazing systems, which have both high total energy transmittance and good thermal
insulation qualities. The recent development of low-e coatings, of so-called transparent
insulation materials (TIM) and sealing technologies allows optimised transparent insulation
systems to be designed for different application fields from flat-plate collector technology to
day lighting systems.
When TIM are combined with selective coatings, low pressure or special gas fillings, heat
resistances from the absorber to the ambient of more than 1.5 m^K/W may be achieved,
where the total energy transmittance (equivalent to the effective transmittance-absorptance-
product) is still around 85 percent for normal incidence and around 70 percent for diffuse
irradiation.
INTRODUCTION
The energy efficiency of a thermal system can be characterized by the balance of solar heat
gains, i.e. the fraction of the solar input which is absorbed by the system, and the heat losses,
i.e. the fraction of the produced heat which cannot be stored and used. The heat gains are
proportional to the effective transmittance-absorptance (™)eff product (Duffie, 1980) or
total solar energy transmittance g (see ISO/DIS 9050 for glazings, Platzer, 1987 for a
generalised definition).
The heat losses are described by the heat resistances R or equivalently by the heat
conductances U (U-value) of the cover systems. For daylighting systems the visual
transmittance Tyjs is also an important quantity.
Applications for transparent cover systems with high thermal resistance and reduced g-value
have been discussed in many papers. The most important applications seem to be high-
temperature flat-plate collectors (Rommel, 1987), integrated-storage collectors (Schmidt,
1990), the transparent insulation of buildings (Goetzberger, 1984; TI 1,1986; TI2, 1988; TI3,
1989) and transparent daylighting walls and windows (Reilly, 1991).
-------
1273
AVAILABLE MATERIALS
In this section materials will be presented that are commercially available or in the process
of commercialisation. The most promising types are low-emittance coatings on glass or
plastic films, honeycomb-type materials and silica aerogel products. Other materials have
been investigated and are conceivable (see e.g. Wittwer, 1990 for a classification of
materials; Platzer, 1988 for other material types).
Low-e Glazings
Although low-e-coated double glazed units have been fabricated commercially for more than
a decade, the use for solar energy application was restricted. The usual units utilise a silver
based coating sandwiched between two antireflecting transparent films, usually oxides. The
reflectance of these films, is very high from the NIR to the FIR. This leads to emittance
values of 6-8 percent combined with very good visible transmittance (close to 80 percent).
However, a large proportion of the solar radiation will be reflected resulting in a
comparatively low solar transmittance of these glazings around 60 percent for normal
incidence. An additional disadvantage of these units was the sensitivity to environment.
Humidity usually destroys these "soft coatings" very soon. Therefore the sealing technology
had to be improved to increase the lifetime of these glazing units above 10 years.
Ag-based SnO,:F
1
0.8
0.6
0.4
0.2
'(•A
0
700
1100
1500
NANOMETERS
1900
2300
300
Figure 1: Comparison of two commercial low-e-coatings
Recently two major developments occurred. The first was the commercial production of a
pyrolytically fluorinated tin oxide "hard coating" (SnC^F) which is stable in a normal
environment. The solar transmittance may be determined by the doping level. Figure 1
shows the spectral transmittance of a commercial Ag-based and a commercial fluorinated tin
oxide coating. The emittance of this coating is considerably lower than that of previous hard
coatings (hemispherical emittance 15 percent).
-------
1274
Concerning window technology, another development was initiated by research of Lawrence
Berkeley Laboratory on triple-glazed krypton filled units (Arasteh, 1989). Table 1 shows
typical values of two commercial products measured at our institute.
TABLE 1: Measurements on low-e-coated units with different gas fillings
Material
T • D
V1?%]
T 1>
g1)
[%]
U-value
[W/m2K]
coating A,
coating B,
coating B,
coating A,
coating B,
double,
double,
triple,
triple,
triple,
Argon
Argon
Argon
Krypton
Krypton
76
67
55
64
60
52
47
30
36
32
62
56
44
47
44
1. 45
1.27
0.77
0.75
0.59
Normal incidence
One can readily recognize the differences in coatings for the two manufacturers which
influence both the transmittance and the U-value. Also, krypton filling of the glazings is
strongly recommended.
Honeycombs
The two main different versions of honeycomb-type plastic structures are a capillary
structure of hollow cylinders with diameters ranging from 1mm to 3mm, and a extruded
multiple channel structure with nearly square cross-section of the honeycomb cells. Both
materials may be fabricated in thicknesses of more than 15cm, which seems to be the upper
practical limit.
TABLE2. Honeycomb materials
Material
'AMI.
.1)
TAM1.
.2)
g2) u-value
[%] [W/m2K]
Capillaries
PMMA, 3mm diam., 10cm
PC, 3mm diam., 10cm
PC, 1.5mm diam., 10cm
94
94
80
70
69
58
78
71
0.91
0.80
0.69
Square honeycombs
PC, 4.5mm diam., 5cm
PC, 4.5mm diam., 10cm
97
97
83
73
87 1.43
82 0.93
Glass capillaries (+ 2 covers)
7mm diam., 10cm
normal incidence
2' diffuse incidence
73
55
63
0.98
-------
1275
The currently sold capillaries are produced from polycarbonate (PC) and
polymethylmethacrylate (PMMA), the square-celled materials from PC. Other plastics have
been tried, but up to now the results have been less satisfactory (Platzer, 1990). A very
promising approach is the production of glass capillaries, which overcomes the fire hazard
and the stagnation temperature problems. However, due to the higher thermal conductivity
of glass the wall conduction is no longer negligible. Theoretical considerations showed that
glass capillaries with 7mm diameter and 100/im wall thickness perform close to the optimum,
A prototype structure with these specifications has been produced and tested. There is still
some potential to improve the optical quality of the glass capillaries; in particular, dust free
manufacturing is important.
Silica Aerogel
Silica aerogel is a microporous open-celled material with pore sizes in the order of lOOnm.
The gas conductivity within the material is therefore restricted and the material looks almost
clear apart from some scattering (Fricke, 1986). Depending on the production technique the
material can be produced in monolithic blocks up to 60cmx60cmx2cm or in a cheaper
granular form (granules up to 10mm diameter) with lower quality. Problems of the aerogel
are the fragility and the hydrophilic tendency: water causes destruction of the gel. However,
research and development are currently concentrating on these problems. A future product
might be a evacuated aerogel-filled window unit, either clear or diffusing.
Table 3. Properties of prototype Aerogel glazings
(Tilling 2cm plus two 4mm glass panes')
Material
granular, 2cm
monolithic, 2cm
monolithic. 2cm
(evac. 10 1 Pa)3)
normal incidence
2| diffuse incidence
3) centre of glass value
r . 1 r
VIS TAM1.
T 2)
AMI.5.
U-value
. 9,
!%]
[%]
I%]
[W/m K]
46
52
43
0.95
75
78
69
0.72
75
78
69
0.46
IMPROVEMENT OF COMPLETE ABSORBER SYSTEMS
When absorber (or glazing) systems with TIM as a whole are optimised, different options
exist to improve the performance. Selective coatings may be used to reduce the thermal
radiation transport. However, due to the coupling of thermal radiation and heat conduction
a gap between the absorber and TIM is needed. An optimisation of the gap thickness is
needed because of convection which arises then. Other gases may be used in order to reduce
the coupling more efficiently. However, gases with lower conductivity usually also have lower
-------
1276
viscosities and are more prone to convection (Arasteh, 1989). Convection may be reduced
more efficiently by a low pressure evacuation to about 100 Pa. Then the Rayleigh number is
considerably lower and convection vanishingly small. Table 4 shows results from calculations
of an absorber system consisting of 10cm honeycomb material with a cover glass and a
selective absorber (e = 10%) utilising different gas fillings and variable gas pressures for air.
TABLE 4. Effect of different gas fillings on heat transport
CFIat-plate Collector, absorber temperature 80°C)
Gas Pressure U-value-*-)
[Pa] [W/m2]
air 105 0.83
air 10^ 0.64
air 103 0.63
carbon-dioxide 10® 0.79
argon 10® 0.78
krypton 10
xenon 10® 0.69
without side and back losses
Additional optimisation of honeycomb systems may be tried by using tilted honeycombs
(Platzer, 1988). Low-iron cover glasses have approximately 4-6% higher transmittance than
standard float glass.
FUTURE POTENTIAL AND OUTLOOK
For windows and glazing systems, research at LBL has promoted the industrial production of
high-performance glazings with the triple-pane krypton window (Arasteh, 1989). Also the
heat conduction through the frame has been minimised, which should be an increasingly
stringent requirement for future super-glazings utilising aerogel. According to Svendsen
(1990), a tight edge seal for aerogel windows has been produced. According to our own
research only low pressure and vibration of granular aerogel glazings can prevent the
undesirable settling of the granules. Low pressure convection reduction has been successfully
tried in commercial, flat-plate collectors. It has been experimentally investigated with
honeycomb collectors (Brunotte, 1990). Brunotte also showed that evacuated double-glazed
units with low-f-coatings and micro-supports may have U-values down to 0.45 W/m^K.
Robinson (1990) and Benson (1986) are working to develop the evacuated window
technology for full-scale units. The main problem certainly is gas leakage and heat losses
through the edge seal. However, these problems should be soluble in the near future. A new
possibility for producing large-area broadband antireflective coatings with so-called moth-
eye microstructures (Wilson, 1982) would then allow a large variety of windows and absorber
systems to be produced with both high thermal resistance and relatively high solar or visible
transmittance. U-values down to 0.3 W/m^K for windows with g-values larger than 0.5 seem
possible.
-------
1277
LITERATURE
Arasteh, D., Selkowitz, S. (1989). The design and testing of a highly insulating glazing system
for use with conventional window systems. J. Sol. En. Eng.. 111. 44-53
Benson, D.K. and Tracy, C.E. (1985). Evacuated window glazings for energy efficient
buildings. 29th SPIE Int. Techn. Svmp. on Optics and Electro-Optics. San Diego.
Brunotte, M. (1990). Warmetransportmechanismen bei transparenten Warmedammungen
im Vakuum. Diploma-Thesis, Univ. Freiburg.
Duffie, J.A., Beckman, W.A. (1980). Solar Engineering of Thermal Processes. Wiley&Sons,
New York
Fricke, J. (1985). Aerogels - a fascinating class of high-performance porous solids. In J.
Fricke (Ed.) Aerogels. Springer Proc. in Physics 6. Springer-Verlag.
Goetzberger, A., J. Schmid and V. Wittwer (1984). Transparent insulation system for passive
solar energy utilization in buildings. Int.J. Sol. En.. 2. 289-308
Wilson, S.J. and M.C. Hutley (1982). The optical properties of'moth eye' antireflection
surfaces. Optica Acta. Vol.29/7. 993-1009
Platzer, W.J. and Wittwer, V. (1987). Total energy transmission of transparent insulation
material. Proc. Workshop on Optical Property Measurement Techniques. Ispra. 27-29
Oct. 1987
Platzer, WJ. (1988). Warmetransportmechanismen und solare Transmission bei
transparenten Warmedammaterialien. Dissertation, Univ. Freiburg.
Platzer, WJ. and V. Wittwer (1990). Fortschritte bei transparenten Warmedammaterialien.
Proc. 7. Int. Sonnenfonim Frankfurt. Vol. 1, 527-532
Reilly, S., W. Platzer and V. Wittwer (1991). Transparently insulated windows: Daylighting
prospects and aperture control. Proc. Int. Workshop TI4. 28-30 May 1991 Birmingham.
Robinson, SJ. and Collins, R.E. (1990). Evacuated windows - theory and practice. Proc.
ISES Solar World Congress. Kobe. Japan
Rommel, M., V. Wittwer and A. Goetzberger (1987). Flat plate collector for process heat
with honeycomb cover - an alternative to vacuum tube collectors. Proc. ISES Solar World
Forum Hamburg.
Schmidt, Ch. (1990). Anwendung transparenter Warmedammung mit Wabenstrukturen in
integrierten Speicherkollektoren zur solaren Brauchwassererwarmung. Fortschrittberichte
VDI. Reihe 19: Warmetechnik/Kaltetechnik. Nr.43. VDI-Verlag
Svendsen, S. (1990). Solar collector based on monolithic silica aerogel. Proc. ISES Solar
World Congress. Kobe. Japan. 711-715
Til (1986). Proc. 1st Int. Workshop "Transparent insulation materials for passive solar
energy utilisation", Freiburg 27-28 Nov. 1986 (ed. L.F. Jesch), Franklin Comp. Consultants
Ltd., Birmingham
TI2 (1988). Proc. 2nd Int. Workshop "Transparent insulation in solar energy conversion for
buildings and other applications", Freiburg 24-25 March 1988 (ed. L.F. Jesch), Franklin
Comp. Consultants Ltd., Birmingham
TI3 (1989). Proc. 3rd Int. Workshop "Transparent insulation technology for solar energy
conversion", Freiburg 18-19 Sep. 1989 (ed. L.F. Jesch), Franklin Comp. Consultants Ltd.,
Birmingham
Wittwer, V. and W.J. Platzer (1990). Transparent insulation materials. Proc. SPIE Conf.
ECQ3 The Hague. 12-13 March 1990. SPIE Proc. Series 1272, 284-296.
-------
1278
NUMERICAL AND EXPERIMENTAL ANALYSIS OF CONVECTION SUPPRESION
FOR CHARGED GAS SOLAR COLLECTOR
Tatsushi Kobayashi, Hiroyuki Shimizu, Shunji Asai
YAZAKI Corp.,LTD.,Air conditioning R&D Lab.,
1370 Koyasu-cho, Hamamatsu-shi,
Shizuoka-ken, 435, Japan
ABSTRACT
Most of the heat loss of the flat-plate solar collector occurs through the blackened
absorber surface, principally by radiation and convection to the exterior
surroundings. Glass, film, honeycomb, etc. were inserted between the outer cover
plate and absorber surface, to minimize convection heat loss that has been frequently
reported. Whereby, a decrease of transmittance and an increase of weight is caused.
Thereupon, the authors took up the gas layers as the important point between the
absorber surface and the outer cover plate. Then, in regard to the use of different
types of gases, the convection suppression has been studied numerically and
experimentally.
In this paper, numerical analysis for the gases has been carried out to obtain the
natural convective heat transfer coefficient from the density, the specific heat at
constant pressure, the heat conductivity, the viscosity as well as the outer cover
plate surface temperature from the heat balance equation. Then, the collector
heat-loss conductance was determined from a heat balance calculation computer
program. The gases of Helium, Argon, Krypton, Xenon, Air, Nitrogen, Carbon dioxide,
which have a high stability in the atmosphere, were selected. Furtheiynore, the gases
of Helium, Air, Argon have been actually charged inside the solar collector, and the
experiment has been carried out according to the method of r ASHRAE STANDARD
93-77Rj . As a result, theoretical calculation and experimental values were conformed
as well as reduction of the convection heat loss were confirmed by using Argon,
Krypton, Xenon gases .
KEYWORDS
Convection suppresion; flat-plate solar collector; charged gas; natural convection
heat loss; tilt angle.
INTRODUCTION
Most of the heat loss of the flat-plate solar collector occurs through the blackened
absorber surface, principally by radiation and convection to the exterior
surroundings. Glass, film, honeycomb, etc. were inserted between the outer cover
plate and absorber surface, to minimize convection heat loss that has been frequently
reported. Whereby, a decrease of transmittance and an increase of weight is caused.
Thereupon, the authors took up the gas- layers as the important point between the
absorber surface and the outer cover plate. Then, in regard to the use of different
types of gases, the convection suppression has been studied numerically and
experimentally. Hereby,the effects and comparison of the results are reported.
-------
1279
CALCULATION METHOD
Fig.l shows the heat balance of flat-plate solar collector.
Outer cover plate —QGs^
Absorber surface y
Insulation
W/W/
4 Qre
Fig.l. heat balance of flat-plate solar collector.
Heat balance equation (1) of outer cover plate is shown as follows.
Qvl+QRl = Qv2+QR2+QG
J-r-dr
(1)
Absorber surface temperature (tp) and ambient temperature (ta) is fixed, then, heat
loss of every part can be determined from equations of (4)~(21) as follows.
Consequently, total heat collected per unit of collector area (Qc) can be calculated
from equation (2) as follows.
Qc= (J • r • ap) - (Qv2+QR2+Qre+Qed • Aed/Ac) (2)
Then, overall heat transfer coefficient (U) can be determined by equation (3) as
below.
Q° J
U=(r • ap )x (3)
J (tp-ta)
Consequently, Collector efficiency {v) can be determined by equation (4) as below.
r)= r-ap-U-At/J (4)
HEAT LOSS OF EACH PARTS OF SOLAR COLLECTOR
( 1 ) Radiation heat loss of outer cover plate QR1 [W/nf]
QR1 = FR1 • (tg-ta)
4.88 a g
FR1= X
tg-ta
t g + 273 /ts + 273
-(¦
(5)
(6)
100 / V 100
Where FR1: Radiation heat transfer coefficient of outer cover plate [W/nf • K]
The sky temperature is ts=ta-5
( 2 ) Radiation heat loss of absorber surface QR2 [W/nf]
QR2=FR2 • (tp-tg)
4.88 £0
FR2= X
tp-tg
/ tp+273 /tg+Z73\
^ 100 / ^ 100 1
t g+273
(7)
(8)
100 / x 100
Where FR2 : Radiation heat loss transfer coefficient of absorber surface
Available emissivity is £o=l/(l/£g+ 1/e p— 1 ) [W/m2 • K]
-------
1280
( 3 ) Forced convection heat loss of outer cover plate Qvl [W/nf]
Qvl= Fvl • (tg-ta) (9)
Where Fvl: Forced convection heat transfer- coefficient of outer cover plate
[W/nf • K]
The experimental equation of Jiirges was used.
Fvl = 4.83 +3.36V (V£5.0m/s) (10)
Fvl=6.15Va7S (5 X1708
Ra Cos^ J
(14)
Ra Cos# ^
'RaCos#-^ ~
1
. 5830
Where CX) °= ( | X| +X) /2
Ra=Gr • Pr* • * (15)
Gr=g -83'P (tp-tg) / (u/p) s (16)
Pr= Cp • n/X (17)
Where coefficient of expansion is /?= 1/ ( (tp+tg) /2+273)
However, the property such as p • n • Cp • A of different types of gases will be
changed in relation to the temperature. Accordingly, the different values of property
were analyzed by 2 order regression curve for calculation.
( 5 ) Conduction heat loss from absorber surface to back surface Qre [W/nf]
Qre=Kre- (tp-ta) (18)
Kre= 1/( 1/Fvl+dre/Are) (19)
Where Kre :Overall heat transfer coefficient from absorber surface to back
surface [W/nf • K]
( 6 ) Conduction heat loss from absorber surface to edge Qed [W/m!]
Qed=Ked • (tp-ta) (20)
Ked= 1/(1/Fvl+ded/Aed) (21)
Where Ked :Overall heat transfer coefficient from absorber surface to edge
[W/nf • K]
nomenclature :
a : Absorptivity QR1 : Radiation heat loss of
e : Emissivity outer cover plate [W/nf]
J : Reflectivity QR2 : Radiation heat loss of
T : Transmissivity absorber surface [W/nf]
t : Temperature [°C ] Qvl: Forced convection heat loss of
J : The radiation flux [W/nf] outer cover plate [W/ nf]
-------
1281
V : Wind speed [m/s ]
: Tilt angle of solar collector [ ° ]
Ac : Collector area [nf]
Aed : Edge area of solar collector [m']
d : Thickness of insulation material [m]
Nu : Nusselt number
Ra: Rayleigh number
Gr : Grashof number
Pr : Prandtl number
6 : Thickness of gas layer [m]
g : The acceleration of gravity [m/s ]
X : Heat conductivity [W/m • K]
Cp : Specific heat at constant pressure
[kJ/kg-K]
fi : Density [kg/m ]
U : Viscosity [>uPa
30 [ • ]
!>>
o
c
-------
1282
J
J - X • Ot 1
Qc Qv2 QP^R^ed
He
25.4
\
OJ
^r.
en
ro
18.6
Air
44 .
7
22.2"'
14 . 5
18.6
cn
22.0
ID
"
-------
1283
THE COMPARISON OP CALCULATION AND EXPERIMENT RESULTS
Fig.4 shows the calculation and experiment results for He,Air,Ar gases.
1.0
0.8
0.6
>,
o
a
«
o
4)
u
3
o
ig 0.4
0.2
1
1
1
experiment
1
calculation
-1
1
>
>
1
vsJ
vx
I o Air
I a He
1
1 s\
1
1 X X
1 to
r\Ni
i
1 c
~l
1
1
1 ^ \1
1
1
~1
1
1
1
1
1 1 s \
. ---VN
1 1 N
1 1
1 1
1 1
0.02 0.04 0.06 0.08
— At/J
Fig.4. The comparison of calculation and experiment results for
solar collector efficiency curves (for He,Air,Ar)
Fig.4 shows the experimental data and calculation data which are close for Air,Ar.
However (experimental data as compared with calculation data is not favourable for He
and there is considerable difference. Indeed, insulation was included the Air for
calculation while included the He for experiment that has 5.8 times larger heat
conductivity (X) ( for 40 °C) than Air that effected the result.
CONCLUSIONS
The results of simulation calculation and experiment of this paper for charged gas
solar collector gave evidence of variation of convection heat loss.
Particularly, the effect of Ar, Kr, Xe for convection suppression was cleared up.
REFERENCES
JSES(1978),Foundations and Applications of Solar Energy,Ohmusha,70-84.
Japan Society of Thermophysical Properties(1990),Thermophysical Properties Hand
Book,Yokendo ltd.,42-47,51-53,57-59.
Japan Society of Mechanical Engineers(1986),Heat Transfer 4th Edition,
,JSME,328-329.
VDI(1989),WARMEATLAS,Japan Management Association.
Yazaki corp.(1984),Blue Panel Solar System Design Manual.
K.G.T.Hollands and others(1976),Trans.ASME of Heat Transfer,Vol 98, 189.
-------
1284
THERMAL PERFORMANCE OF IMPROVED EVACUATED FLAT PLATE
COLLECTORS
N. Benz, W. Scholkopf and R. Sizmann
Sektion Physik, Ludwig-Maximilians-Universitat Munchen, Lehrstuhl Prof. Dr. Sizmann
Amalienstr. 54, D 8000 Munchen 40, FRG
ABSTRACT
In highly efficient collectors the use of selectively coated absorbers is common. For further
improvements the convective losses between the absorber and the cover have to be reduced.
By evacuating the collector housing the natural convection and the heat conduction losses
can be completely suppressed. By a moderate vacuum of 103 Pa only the natural convection
disappears. With a detailed simulation program the feasible improvements of partly evacuated
collectors were investigated. In a sensitivity analysis with stationary simulation the influence
of internal pressure of the collector housing, the geometric configuration of the collector casing,
the emittance of the absorber, and the operating conditions were evaluated. Results indicate
that an optical efficiency of 0.85 and at a residual pressure of 103 Pa a loss coefficient of
2.0 W/(m2K) are feasible.
KEYWORDS
Evacuated flat plate collector, dynamic collector modelling, simulation, senstivity analysis,
collector test
INTRODUCTION
Evacuated flat plate collectors (EFPC) are commercially produced in Germany by the com-
pany Thermosolar. The problem of supporting the outer pressure and keeping a lightweight
construction is solved by the insertion of 130 stainless steel spacers per m2 between the glass
cover and the 7 cm deep-drawn aluminium casing. To allow for the spacers, the absorber
which is suspended 2 cm from the top of the collector, is perforated, which causes about
2% of absorber area losses. A reduced pressure of about 103 Pa prevents internal convection
(C.B.Eaton and H.A.Blum, 1975; H.Buchberg, I.Catton and D.K.Edwards, 1976). The selec-
tively coated absorber reduces radiation losses.
A flat plate collector exploits direct and diffuse radiation as well. Advantages in comparison
with the evacuated tube collector (ETC) modules are: higher optical efficiency which leads to
a higher thermal efficiency at usual operating conditions for SDHW; higher area usage factor;
better handling and integration into roofs; lower costs.
The thermal efficiency is determined by
• thermal and optical properties of the employed materials, particularly of absorber and
cover,
• the obtainable and sustained vacuum,
-------
1285
Fig. 1. evacuated flat plate collector
• the geometric parameters of construction.
Here we report on a detailed simulation program which allows investigation of the effects
of loss mechanisms and optimizes the collector parameters. By correlating measured
variables to a dynamic simulation model the parameters of the EFPC can be
determined more easily in short term tests than in tests requiring steady state
operation.
THE COLLECTOR MODEL
The collector is represented by a multi-node model. For calculation the collector has been
divided into N absorber and fluid segments. The temperatures of the glass cover, the casing
and the absorber and fluid segments satisfy the time-dependent differential equations (for
j = 1...N):
Ca^ = Ga + K_r(T}tj-Ta,j) + (hl3g_a + h>_a)-(Tg-Taij) +
(Af_c + h"a_c) ¦ (Tc - TaJ) + haa ¦ (Taj-i - TaJ); (1)
Ta,0 — Ta< 1
Cjd^T = haa-r(Ta,j~Tf,j) + N.mcp.(Tf,j_1-TfJy, (2)
Tf,0 = Tftin
CS,in^j~ = ha^c-(Tc-T}iin) + mcp-(Tin-Tf,iny, (3)
dT
Cf,out—^ = haf_c-(Tc-Ti,0Ut) + rhcp-{T}iN-TJ,0Uty, (4)
CslT = G' + (h0*- + K-a)'(Fa-Tt) + hP_airF-(TairF-Ta) +
K-sky ¦ (Tsky - Tg) + h»_g ¦ (Tc - TJ; (5)
Ce~ = (htc + K_c).(Ta-Tc) + h"c_g-(Tg-Tc) +
(htairB + K-airli) ' (T*irB ~ Te) + h"f_e • (Tf
-------
1286
ambient(rear side)
ambient (front)
J(N)
TO)
T (out)
T(in)
T(N)
T(1)
T(2)
rhci
casing
fluid
casing
cover
absorber
conduciion-
: convection-resistance
Fig. 2: collector resistance-model
Ta is the mean absorber temperature
_ 1 N
Ta = jrET«j
i=l
The thermal capacities are assumed to be constant, whereas the heat transfer coefficients are,
in general, functions of the corresponding temperatures. The absorbed radiation is calculated
with an absorptance-transmittance product (ar) which depends on the radiation incidence
angle of each of the irradiance components (beam, diffuse from sky, diffuse from ground). The
collector model is applicable for dynamic collector tests and sensitivity analyses. The assump-
tions for calculating the heat transfer coefficients can be summarized as:
Radiation. The heat transfer coefficient of infrared radiation between gray surfaces is given
by hi_2 = E ¦ a ¦ (T'i + T?)(Ti + T22). The constant E includes the emittances of the components
which are assumed to be constant, the areas and view-factors, cr is the Stefan-Boltzmann con-
stant.
Convection. For the rate of convective heat transfer between two inclined plates, a model of
Hollands (K.G.T.Hollands, T.E.Unny, G.D.Raithby and L.Konicek, 1976) is used. Of obvious
-------
1287
importance is the temperature dependent critical pressure where convection disappears.
Conduction. The various heat transfer coefficients of conduction are assumed for one-di-
mensional geometry; they are dependent on temperature, pressure and materials properties.
Collector efficiency factor. The collector efficiency factor F' in the Hottel-Whillier-
Bliss equation (J.A.Duffie and W.A.Beckman, 1980) is calculated by a model of Lund
(K.O'Ferral Lund, 1989). F' determines the heat transfer coefficient between absorber and
fluid.
The set of differential equations is numerically solved by a fourth-order Runge -Kutta method
with adaptive stepsize control.
SENSITIVITY ANALYSIS
The simulation program (for steady state condition) is used to investigate the effects of loss
mechanisms and to determine the optimum parameters of the collector. This approach yields
detailed information about the influence of construction details and operating conditions on
collector performance. The improved collector parameters for all variations are:
• dimension: 2m x lm x 10cm,
• absorptance of absorber: 0.92, emittancefor 100°C absorber temperature: 0.1; serpentine
duct layout with liquid heat transfer medium,
• internal pressure: 2000 Pa.
Altogether the simulation program uses about 110 parameters. In figure 3 diagrams are given
for the variation of the parameters: internal pressure, height of casing (with centered absorber),
emittance of the absorber and thermal efficiency curves, where fluid inlet temperature and ir-
radiation are varied.
Internal pressure The overall heat transfer coefficient decreases for lower pressures. Due
to reducing convection it decreases drastically in the range between 105 to 7 • 103 Pa. In the
subsequent plateau of the curve,convection is completely suppressed but heat conduction still
remains independent on pressure. Only when the pressure becomes smaller than 1 Pa, con-
duction losses decrease with pressure. Below 10~2 Pa the heat losses of the absorber are only
determined by radiation and contact conduction. In order to prevent frequent reevacuation,
the collector is kept operating in the plateau range of the curve where the performance is in-
dependent on internal pressure. The back losses appear to be greater than front losses because
of heat transfer losses between the connection pipes and casing.
Depth of casing In the operating range of the collector, where heat conduction in the
residual gas is significant, the distances between absorber and cover, and absorber and casing
bottom respectively are of particular importance. Therefore the depth of the casing should
not be less than 10 cm.
Emittance of absorber The thermal efficiency is substantially influenced by the emittance
of the absorber. To achieve high performance, the emittance should not exceed 0.1.
Efficiency curves The thermal efficiency curves are calculated for Munich, March 22 at
noon time. The optical efficiency is dependent on the fraction of diffuse radiation. That is why
it decreases with decreasing irradiance. Below operating temperatures of 100°C the attainable
thermal efficiency is between 0.65 and 0.8 (x = 1000 W/m2). It is greater than the stan-
dard optical efficiency of evacuated tube collectors modules with a maximum of about 0.65
(S.P.Chow and G.L.Harding, 1985). Process heat production (100°C to 150°C) is also possible
with the evacuated flat plate collector with efficiencies greater than 0.5.
-------
1288
¦ UL (total) [W/(m2K)]
- Ul (front)
• - - Ul (back)
standard operating conditions:
location : Munich
date, time : March 22, 12—
tj>± = 800W/m2; vwi„d = 2m/s
Tamb = 15° C; T» = 60° C
0.01 1 100 1E4
internal pressure [Pa]
2.5
2
1.5
1
0.5
0
• UL (total) [W/(m2K)]
Ul (front)
Ul (back)
10 15 20 25
height of casing [cm]
thermal efficiency ( £ absorber= 0-1 )
[W/m2] = 1000
800
thermal efficiency = 1000 W/m )
0 50 100 150 200 250
T(fluid) - T(ambient) [K]
50 100 150 200 250
T(fluid) - T(ambient) [K]
Fig. 3. Sensitivity analysis of evacuated flat plate collector
EXPERIMENTAL VALIDATION
A commercial collector was tested under ^different operating conditions. Results of the meas-
urement show that the theoretical model reliably predicts the operation behaviour of the
flat plate collector. The overall heat transfer coefficient (in the SDHW operating range) is
-------
1289
3.5 ± 0.5 W/(m2K). The depth of the present commercial collector casing is too shallow and
the heat losses between the collecting pipes to the casing at inlet and outlet are excessive.
CONCLUSIONS
The sensitivity analysis shows that:
• a higher optical efficiency than 0.8 (refering to the aperture area) is attainable;
• in the operating range of a SDHW-system and in space heating systems a heat loss
coefficient of 2.0 W/m2K can be achieved;
• for SDHW purposes the yearly energy yield of an improved collector exceeds the perfor-
mance of an evacuated tubular collector modul;
• comparision of the simulation results with experimental data shows a high improvement
potential of the commercial collector.
Nomenclature
T : temperature [K]
C
: thermal capacity [J/K]
G : absorbed irradiance [W]
mc
p : thermal capacitance rate [W/K]
h : heat-transfer coeff. [W/K]
t
: time [sj
Ui : overall heat transfer coeff. [W/m2K]
Vth
: thermal efficiency [-]
± : irradiation in absorber area [W/m2]
£
: emittance [-]
subscripts:
a : absorber a : conduction
out : fluid outlet
g : glass cover /? : convection
airB : rear side ambient
/ : fluid p : radiation
airF : front ambient
c : casing in : fluid inlet
j : index
REFERENCES
C.B.Eaton and H.A.Blum (1975). The use of moderate vacuum environments as a means of
increasing the collection efficiencies and operating temperatures of flate-plate collectors.
Solar Energy, 17:151-158.
H.Buchberg, I.Catton and D.K.Edwards (1976). Natural convection in enclosed spaces — a
review of application to solar energy collection. Journal of Heat Transfer, 182.
J.A.Duffie and W.A.Beckman (1980). Solar Engineering of Thermal Processes. Wiley, New
York.
K.G.T.Hollands, T.E.Unny, G.D.Raithby and L.Konicek (1976). Free convective heat transfer
across inclined air layers. Journal of Heat Transfer, 189.
K.O'Ferral Lund (1989). General thermal analysis of serpentine-flow flat-plate solar collector
absorbers. Solar Energy, 42(2):133-142.
S.P.Chow and G.L.Harding (1985). Angular dependence of optical efficiency of evacuated
tubular collectors with antireflection coatings and stationary specular reflectors. Solar
Energy, 34(6):489.
-------
1290
DETAILED EXPERIMENTAL INVESTIGATIONS ON
EVACUATED AND OTHER HIGH PERFORMANCE COLLECTORS
Konrad R. Schreitmiiller, Martina Niemann, Gunther Rockendorf
Institut fur Solarenergieforschung (ISFH), Sokelantstr. 5, D-3000 Hannover, F.R.G.
ABSTRACT
The accurate knowledge of the optical and thermal characteristics is crucial both for the
development, improvement, and application of high performance collectors. The paper
describes some of the used equipment and some results both with flat plate and evacuated
tubular collectors.
KEY WORDS
Solar Simulator, Spectroscopy, IR Camera, High Performance Collectors, CPC, Thermal
Loss Values, Incidence Angle Modifiers
INTRODUCTION
Various solar energy applications ranging from domestic hot water and district heating
systems in adverse climates to process heat, absorption cooling, and water desalination
plants demand high performance collectors in the temperature range of 50...180 °C. That
implies an optimal combination of good optical performance, low thermal loss values, and
efficient absorber-fluid conductivity. The accurate knowledge of the respective charac-
teristics and other key values is especially essential for two purposes, i. e.
- for the technical and economical optimization in the stage of collector development, in
order to assess fast and accurately the effects of possible modifications and to detect
hidden flaws,
- for the detailed modelling of large or new installations or investigations of special
effects as e. g. matched flow concepts.
Collector characteristics are generally determined in outdoor tests. Detailed test methods
have been developed during the last years. However, due to the fluctuation of the
involved meteorological data (diffuse/beam radiation, incident angle, wind velocity,
ambient and sky temperature) the scattering of the respective data points is usually rather
high. This is especially valid with collectors operating with elevated temperatures. Thus
the corresponding results veil the details, and second order and/or combined effects are
seldomly accurately represented. Furthermore, as the collector is treated as black box,
local flaws are hardly detectable. Hence outdoor tests have to be complemented by
-------
1291
additional indoor investigations. This report gives an overview on the relevant ISFH
activities.
EXPERIMENTAL EQUIPMENT
Solar Simulator SUSI A
The solar simulator (Fig. 1. /I/) with an active area of 2.9 m2 (now 3.4 m2) was ac-
tuated November '89 and completed meanwhile appr. 1000 hours of operation (Feb. '91).It
consisted originally of 165 (now 218) halogen lamps with 150 W each and dichroic mirror.
The spectral distribution is improved by two panes of acrylic glass between lamps and test
object and is rather close to AM 1.5; as these panes are cooled by cold air below ambient
temperature, they serve as well as "cold sky". Due to the optimized arrangement of the
lamps the radiation density is very uniform, the standard deviation is appr. 3 %. The
periphery covers a thermostat/flow control unit (stability of inlet temperature ±.02 K, of
flow rate ±1%), a stabilized power supply (±.2%), an air conditioning system (±.3 K),
and a radiation meter checking before and after each test the distribition of the radiation
density. The collector measurements are performed with an accuracy of ±.035 K
(temperatures), < 1% (mass flow rate), and ± 1% (radiation density).
8 ventilators
(cooling of lamps)
v<=3 m/s
lamp field
(15x11=165
halogen lamps)
cold sky
(ventilated, trans-
parent channel)
curtains for
intensity Q
variation fTi
cold air v°»10m/s
irradiance
distribution
measurement
wind simulation
collector (test are:
max 1.3 x 2.2 m2)
Fig. 1. Solar Simulator SUSI A
Spectroscopical Equipment
The spectroscopical equipment allows the following investigations
direct-direct transmission (non-scattering materials) in the UV to FIR region (300 -
50.000 nm)
- direct-total/direct-diffuse transmission of extended specimen in the UV to MIR region
(400 - 15.000 nm, two integrating spheres, 50 cm0 each, ports 9 cm^, barium-sulfate
and gold coated, resp.); determination of the angular dependence within ±70°
-------
1292
diffuse-diffuse transmission in the UV to NIR region (400 - 2.500 nm)
direct-total/direct-diffuse reflexion in the UV to MIR region (400 - 15.000 nm)
Some of the respective investigations are reported in /2/.
The Infrared Camera
The IR camera with HgCdTe detector (wavelength range 8 - 13 pm) is applicable in the
temperature range of -40...2000 °C; the temperature resolution is .05 K (30 °C blackbody
temperature). Quantitative measurements are possible with known emissivity/reflectivity
of the investigated specimen, otherwise qualitative ones. The IR equipment will be
extended in 1991 by a cooled evacuated tank with ZnSe window (10 cm0, transmissivity 70
% in the .5-14 ^m wavelength region) in order to investigate the emissivity and uniformity
of extended specimen within vacuum.
Additional Experimental Equipment
The careful investigation of various key effects needs additional facilities providing highly
uniform and constant test conditions. These cover
a. combined indoor/outdoor collector test facility, serving as standard for the no-mea-
surements and to investigate the temperature dependence of the thermal loss factors,
an electrically heated test facility for the investigation of the energy transport
phenomena within heatpipes,
- an electrically heated test facility for the investigation of the temperature dependence
of the thermal loss factors of evacuated tubular collectors.
SOME RESULTS
The following collectors have been analysed
selective flat plate collectors with/without absorber parallel Teflon foils as convection
(and, partially, IR radiation) barrier,
selective flat plate collectors with honeycombs,
evacuated flat plate collectors (soft vacuum, p = 1...10 kPa),
evacuated tubular collectors (hard vacuum, p < 10 mPa), with/without CPC-reflectors.
A few results are discussed in the following.
Flat Plate Collectors
Fig. 2 shows the temperature dependence of the thermal loss values of two generic types
of flat plate collectors (FPC), measured with two different test facilities (/3/). Base
case (BC) is a selective FPC with U, = 4-6 W/m2K (AT = 20... 100 K), which is standard
for domestic hot water systems in Central Europe. An additional convection barrier
('teflon foil parallel to the absorber) leads to loss values of appr. 3.5 - 4.2 W/m2K
(BC/CB). A honeycomb convection barriers (cell diameter = 3 mm) reduces furthermore
the losses to appr. 2.7 - 3.7 W/m2K (4HC, thickness 4 cm) or 2.5 - 3.1 W/m2K (8HC,
thickness 8 cm), respectively. However, this difference is small enough to allow the lower
value for most applications, especially as the thicker honeycomb structure increases signifi-
cantly the optical losses with oblique irradiation (incidence angle modifiers). Very
-------
1293
effective and recommendable is an improved backside insulation (8HC/BSI, 8 cm
styrofoam), which leads to loss values in the range of 2.1 - 2.2 W/m2K. The last curve
(EFPC) corresponds to an evacuated FPC (soft vacuum, pj = 1 - 10 kPa). This collector
(different make) shows an only marginal increase of the U-value with temperature, which
is due to the suppression of convection and an apparently good selective layer; however, as
with this construction the loss factor starts from a rather high level caused by thermal
bridges (both at inlet and outlet the fluid tubes are directly coupled to the frame), there is
surely some space and some need for improvements.
W/m K
6 n
5 -
4 - F
3
2
1
0
20
40
60
Tin
80 100 120 140 160
T amb > K
0
BC —1— BC/CB —4HC
—8HC —8HC/BSI —<>— EFPC
Fig. 2. Thermal Loss Values (FPC)
Evacuated Tubular Collectors (ETC)
W/m K
5
4
3
2
1
0
300
200
250
150
100
50
0
Tabs Tamk , K
—^ Collector I Collector II
Fig. 3. Thermal Loss Values (ETC)
-------
1294
The thermal loss values of various evacuated collector tubes have been determined by
internal electrical heating to stagnation temperature. Fig. 3 shows these values for two dif-
ferent makes (refeifed to absorber area). Additional investigations (check of the internal
gas pressure by a high frequency/high voltage induced glow discharge and the statistical
analysis of the experimental data) show, that in both cases the losses are almost com-
pletely due to IR radiation. Both collectors are stagnation resistant at least up to
temperatures Tabs = 300 °C. Whereas collector I, an all glass ETC, shows an excellent
selective layer (e < 6 % for Tabs < 300 °C), that of the other one (metal absorber in
glass) has been seriously damaged probably in the process of manufacturing. The IR
investigation reveals the concentration of the IR losses in the absorber parts adjoining the
metal to glass connection.
A crucial point with many collector constructions is the low absorber to fluid conductivity.
Heatpipes tend to choke or even to fall partially dry especially with high radiation
intensities and/or low inclination The measurements revealed furthermore an often
insufficient filling of the heatpipes and an inadequately poor condenser to fluid coupling.
250
200
150
100
50
0
0 500 1000 1500 2000
distance from condenser, mm
— 500 W/m2 —750 W/m2
—1000 w/m2 -®- 1250 w/m2
Fig. 4. Axial Temperature Distribution on Absorber (ETC)
Fig. 4 shows the temperature distribution along the heatpipe (fluid temperature 40 °C).
Even with low intensities the absorber temperature exceeds that of the fluid by 50...60 K.
With high intensities, which rather often occur with this construction (equipped with a
backside reflector), the heatpipe falls dry and reaches stagnation. However, the energy
transfer by aqueous heat transfer medium, circulated in U-tubes soldered to the absorber,
may be inadequately low, too, if the collector tubes are connected in parallel, as then non-
uniform flow distribution may result in local overheating and consequently in temporary
blocking of the respective tubes by steam bubbles. This effect increases with low flow
rates and hence the respective collectors are unfit for microflow applications. Fig. 5 shows
the course of the relevant ^0-value with parallel and serial connection of the tubes.
absorber temperature, C
inclination 45
Fluid Temperature
-------
1295
%
80
serial connection
£> +
60
75 -
70 --
65 -
parallel connection
0
20
40
60
spec, mass flow rate (H20), kg/m2h
=o
Fig. 5. Mass Flow Dependence of ri0 - Value with Parallel
and Serially Connected Tubes (ETC)
The operational temperature of ETC may be significantly extended by CPC reflectors.
The first ISFH prototype (C = 6, aperture area 2.2 m2) shows a thermal loss value
U = 0.27 Wm"2 K"1 + 0.0035 Wm"2 K"2 * AT, which is surely adequate for applications up
to temperatures even above 200 °C. However, a new prototype with reduced con-
centration factor in order to allow only seasonal tracking and with increased aperture area
(4 m2) is now being investigated. The results of this collector type will be discussed at the
conference.
REFERENCES
/l/ Rockendorf, G.: Design and Application of a Collector Test Facility based on a Solar
Simulator. Proc. North Sun '90, Reading, UK, 17./21.09.1990
/2/ Jahn, K., Christoffers, D.: Valuation of Transparent Insulation Devices. Proc. 1991
ISES Solar World Congress, Denver, USA, 19-23 August 1991
/3/ Niemann, M.: Entwicklung von Kollektoren im Temperaturbereich bis 200 °C.
1. Nat. Symposium Thermische Solarenergie, Staffelstein/F.R.G., 13./14.06.1991
-------
I
I
I
I
I
I
-------
2.5 Solar Domestic Hot Water I
eceding page blank
-------
-------
1299
DESIGN AND PERFORMANCE OF SMALL SOLAR WATER HEATING SYSTEMS
Francis de Winter*, Adolfo Arata** and Maier Perlman***
* Altas Corporation, 1401 Laurent St., Santa Cruz, CA, USA
** Universidad Santa Maria, Casilla 110-V, Valparaiso, Chile
***Ontario Hydro, 800 Kipling Ave., Toronto, Ontario, Canada
ABSTRACT
It is shown there is an enormous day to day variation in the
consumption of domestic hot water (DHW) in the single family
home. This has a significant effect on both the design and the
performance of solar domestic hot water systems. Results are
included on the system size, behavior and performance of a
"two-tank" system, and of a prototype developed at Altas. The
designs are systems of type 4 and 5 respectively, using the
scheme of the Universidad Santa Maria in Chile. The "two tank"
system has serious disadvantages when DHW usage varies.
KEYWORDS
Solar DHW System Design, DHW Consumption Variability
INTRODUCTION
To date the solar energy field has been almost oblivious to the
consumption variability of domestic hot water (DHW) in the single
family home. Virtually every author has assumed (based on no
evidence) that single family homes provide DHW loads which do not
vary. Virtually no author has tried to see whether this is really
so, by simply mounting a flow meter on the water heater of his
own home (or the home of someone else) , and making one reading,
once every day at the same time. It is hardly possible to think
of a cheaper DHW test, yet hardly anyone has actually done it.
Some measurents on the domestic hot water consumption in actual
homes are shown in Fig. 1 to 3. The consumption from day to day
varies by as much as a factor of 10, and from week to week by as
much as a factor of 3. This variation is greater than that of
the insolation. When the family goes out or on vacation
consumption stops altogether. This has several implications:
a. In solar domestic hot water systems, once the water heated
with solar energy is used up, one uses backup energy. If
calculations are based on invariant hot water consumption, the
performance predictions can be wrong and misleading.
b. The design of the heater may also be wrong, with the
assumption of invariant hot water consumption. One gets the
wrong results for the size of the tanks to be used, and for the
necessary thermal coupling between the solar and backup tanks.
^receding page blank
-------
1300
c. There is no reason to believe that the hot water consumption
and the insolation or weather can be correlated well (or widely)
enough for the correlation to be of any use in equipment design
or evaluation. It is safest to assume such a correlation does
not exist, and that consumption variations are simply random.
d. DHW usage variability is unimportant in gas or electric water
heaters, since they have a more or less unlimited DHW capacity.
It has long been "¦w necessary to consider variable DHW usage.
Buckles, et. al.,(1979) stated: "The observed effect of the daily
load variation was larger than that of any of the design
variables (except collector area) studied." In another paper,
Buckles, et. al7(1980) stated: "Wide variability in the daily draw
pattern can significantly reduce the system thermal performance,
particularly if the daily draw frequently exceeds the storage
tank capacity." Nobody paid much attention to these conclusions,
including the listed co-authors of the Buckles,et.aL, papers.
The present paper deals with this problem, with the design
requirements imposed by hot water consumption variability, with
plausible design configurations, and with some results on
equipment performance and behavior which can be expected.
LIMITATIONS OF CURRENT DESIGNS FOR SOLAR WATER HEATING,
AND OTHER DESIGN CONFIGURATION POSSIBILITIES
For many years, the solar energy field has studied two solar
water heating designs to the virtual exclusion of all others.
These are the "one-tank" system, in which the backup heating is
added at the top of a single tank, and in which the bottom part
of the tank serves for solar storage; and the "two-tank" system,
in which water (or heat) is only transferred from the solar tank
to the backup tank when someone is using hot water.
Both of these devices have serious drawbacks, having to do with
stratification and with tank coupling. In the "one-tank" device,
thermal destratification can produce significant performance
penalties. In the "two-tank" device, when no hot water is being
used, all heat losses from the backup heater tank require backup
energy, even though the solar tank may be fully charged. At
EPRI, a two-tank device was used to supply hot water to an office
building, and the backup energy requirements were highest on
weekends, when nobody used any hot water (Purcell, 1984).
Sha, et. al., (1980) started the preparation of a computer code to
model the tank stratification behavior, but this code was never
properly validated. There is no really reliable way to design a
properly stratified water tank. At present the most reliable way
to ensure stratification is to divide a tank into successive,
segregated tanks, between which mixing or coupling can be
controlled or avoided altogether. If one wants to have optimum
system designs, it is useful to consider systems other than the
standard "one-tank" or "two-tank" systems.
Arata, et. al,, (1991) made a systematic study of various solar hot
water system configurations involving one or more tanks. Seven
distinct configurations were defined. The "one-tank" system is
of type 3, the "two-tank" of type 4, Altas has for long studied
systems of type 5, and Arata, et. al.» (1991) have built and studied
-------
1301
some very large heaters of types 6 and 7. Systems of types 5, 6,
and 7 have tank arrangements which enhance the stratification,
and coupling arrangements between the successive tanks. The
present study deals with small systems of types 4 and 5.
SYSTEM DESCRIPTION FOR ALTAS SYSTEM OF TYPE 5
Altas Corp. developed a new configuration of solar water heater
tanks, shown in Figure 4, on R&D contracts with the U.S. DOE and
the Gas Research Institute (see Morrison, Grunes, and de Winter
references). The solar tanks hold 270 liters, 30% of which is in
the drainback jacket. The backup tank holds 110 liters.
The heater corresponds to a system of type No. 5 according to the
Arata (1991) classification scheme. The solar tank and the
backup tank are coupled by a passive thermal diode, produced by a
set of two convection chimneys. This diode has a thermal
conductance for heat up-flow about 5000 times higher than for
heat down-flow. When the solar tank is colder than the backup
tank, the tanks are isolated. When the solar tank is hotter than
the backup tank, the tanks are coupled. The chimneys normally
operate in laminar flow, and the performance of the chimney-diode
system can be predicted accurately, by equating the buoyancy
driving forces and the friction losses (de Winter, 1980). The
"conductance" for heat up-flow is typically 100.0 W/C. The bottom
tank is never more than a few degrees C hotter than the top tank.
In small systems passive diodes are practical, for the relatively
small tanks can be mounted on top of each other. Some of the
Chilean systems involve tanks holding as much as 16 tons of
water. Passive diodes become impractical, but one can easily use
an active diode: a pump with a controller to do the same thing.
Since the drainback jacket has a significant thermal capacity,
one cannot calculate a heat exchanger factor using simple de
Winter (1975) equations. The effective heat exchanger penalty,
as confirmed by both experiment and analysis, is about 2%. The
system performance is quite insensitive to the fraction of the
water contained in the drainback jacket.
The backup heater is provided by a natural gas heated "two-phase
thermosyphon," (TPTS) somewhat similar to a heat pipe, but with a
separate condensate return (see Grunes and Morrison refs.).
Water is used as a working fluid, and firing efficiencies of
above 80% were obtained in the unit shown. The TPTS also behaves
as a thermal diode, with a conductance ratio also roughly equal
to 5000 to 1. The gas burner uses an intermittent igniter. The
unit shown has a thermal loss of about 1% of the stored heat
capacity per hour, which corresponds to a thermal time constant
of about 100 hours. Time constants of 160 hours are achievable.
SOME EXPERIMENTAL RESULTS
Two heaters of the type described above were built between 1980
and 1982. Both were tested extensively in the Altas laboratory,
one for a period of 10 months with a hot water consumption of 360
liters per day. The heaters worked very well. One unit has been
on field test in Hawaii since May 1984. Natural gas is not
available in Hawaii, and the synthetic gas used is about three
times more expensive than natural gas in the continental USA.
The unit uses two collectors of 4 ft x 8 ft.
-------
1302
N.B. 1 US Gallon = 3.785 L.
250
Month of July and
month of December
o t *00
Mar. Apr May Jun Jul Aug Sep Oct Nov Dec Jan Feb
Months
Fig. l. Measured Weekly Avg.
DHW Usage for Family with
"Reasonably stable water use
patterns" (Baker, 1982)
2 4 6 8 10 12 14 16 18 20 22 24 26 28 30
Day of month
Fig. 2. Measured Daily DHW
Consumption for Family of 5
July & December (Laue, 1982)
— i. Hot water consumption (litres)
Month of Feb. 1981.
Most of July 1981
On vacation
I I
Fig. 3, Measured Daily DHW
Consumption for Family of 3 in
Feb. & July "81 (Perlman, 1982)
Note on this Perlman reference:
It should be noted that Ontario
Hydro has obtained DHW usage
measurements on an hourly basis
during a period lasting 4 years
for a total of 58 families.
This was as part of an ASHRAE-
sponsored effort on DHW heating.
HOT OUT
BACK-UP
CONDENSER
STORAGE
TANK
THERMAL
DIODE
I mm-j
BURNER a
'EVAPORATOR
SOLAR
TANK
INSULATION
JACKET
WATER IN
Fig. 4- Altas/U.S.DOE/GRI
Solar Augmented Gas Fired
Water Heater for the single
Family Home.
Table !•Optimum Design Performance: Two-Tank (Type 4) Solar Water
Heater Meeting the DHW Usage Profiles of Fig. 3. Backup Heater
with Standby Losses of 4% per Hour, and an 83% Firing Efficiency.
Storage Cost $0.80 per litre $0.40 per litre
Month of February
(no vacations)
Month of July
(with vacations)
Opt. Solar Vol.
Cost per L DHW
Solar Fraction
Opt. Solar Vol.
Cost per L DHW
Solar Fraction
166 L
$0.0091
-0.376
166 L
$0.0192
-1.751
226 L
$0.0089
-0.346
226 L
$0.0187
-1.685
-------
1303
A full year of field test operation was documented (de Winter et
al, 1987). The heater delivered an average of 166 liters of
water per day at 65.6 C. It had been found that at 60 C there was
virtually no gas usage. The total gas required for the year was
439 kWh. A standard gas water heater with an efficiency of 54%
would have used about 6,300 kWh, so the associated replacement
solar fraction was about 93%. In a heater with a standing pilot,
typically using 220 watts of gas, the pilot would have used
1,924 kWh of gas per year. The heater used less than 23% as much
as such a pilot light. The field test is continuing.
Several factors are involved in the good performance of the
heater. Clearly one can not have a standing pilot. A daily
average usage of only 166 liters was less than expected, so that
the total tank volume of 380 liters was able to be very helpful
in handling DHW variations. The low heat losses from the tanks,
and the thermal diode coupling the tanks also clearly helped.
COMPUTER SIMULATION RESULTS
Extensive computer simulations were performed at Altas for the
DHW usage profiles shown in Figs. 2 and 3, and for a DHW usage
with the same monthly total but the same usage day to day. A
fairly simple Fortran computer code was written to do this, using
hourly calculations during a full month. Enough comments were
included to make the program self-explanatory, and it is
available as a listing or on a disc to all who want it.
Calculations were performed for systems of both type 4 and type
5, using backup heaters of 151.4 litres (40 US gallon) with a
firing (recovery) efficiency of 83%, and standby losses ranging
from 0.7% per hour (probably the best achievable) to 4% per hour
(common at present, see Anon, 1991). Collector costs were
assumed to be US $250 per sq. m, storage costs $0.80 per litre
(the effect of having storage at $0.40 per litre was explored),
and gas costs were taken to be $0.06 per kWh. This gas cost is
high. At current gas costs, solar DHW is however not
competitive, and it was the objective of the current study to
examine the dynamics of solar DHW systems which are competitive.
The fossil fuel "costs" offered by the "free market" are quite
artificial, and have never been otherwise. The solar storage tank
was assumed to have an L/D ratio of 3.0, and an insulation
wrapping of 0.36 W/sq. m/C. The tanks were assumed to be fully
mixed. The inter-tank diode for the Altas heater was assumed to
transfer 100.0 W/C up, and zero down. The collector area was
taken to be 1 sq. m for every 40 L/day in average DHW usage, with
the backup heater thermostat at 60 C, with a tempering valve (set
at 60®C) used at the backup heater exit, and the solar cut-off
temperature at 80"C. A fairly good Hottel-Whillier flat plate
collector was used [with a possible de Winter (1975) heat
exchanger factor], and the ambient temperature (as well as the
cold water supply temperature) was set at 15°C. The DHW
consumption profiles of Figs. 2 and 3 were used to describe a 30
day month, with the individual days following the hourly profile
of Beckman et al (1977). The month was made up of 3 ten day
periods, each having four good clear days and one poor day, then
three good clear days and two poor days. The cost of capital
plus service plus depreciation was taken to be 15% per year.
Each computer run was made with a range of solar storage tank
sizes so as to be able to choose a cost-optimum, and the monthly
calculations were repeated twice so as to have good starting
-------
1304
temperatures the second time through. The results included the
cost per litre of hot water, and the solar fraction-
Many solar fraction definitions are possible (Laue, 1982). The
one we used was: (useful heat delivered, minus gas heat used),
divided by (useful heat delivered).
Large backup heater standby losses impose large penalties on the
two-tank system especially when the water usage is low or when
people go on vacation. Table 1 shows performance numbers for a
type 4 heater with the profiles shown in Fig. 3, and with 4%
hourly standby losses in the backup heater. It should be noted a
solar fraction of -1.751 corresponds to an effective gas heating
efficiency of 1/2.751=0.363, or 36.3%. A good gas heater (83%
firing efficiency, 0.7% hourly standby losses) would give ar
effective gas heating efficiency for this case of 63.75% with nc
solar help whatever. Altas units get +0.648 in solar fractionl
Water heated to satisfy the variable profile shown as the solid
line on Fig. 2 was about 4% more costly than water heated tc
satisfy a constant daily usage of the same average usage, and the
solar fraction was only about 90% as high.
The system of type 5 (Altas) optimized at consistently smaller
solar tank sizes than the two-tank (type 4) system. This is
understandable because the backup tank in the type 5 system is
closely coupled to the solar tank, hence helping with solar
storage.
The type 5 system was generally between 1% and 2% more cost-
effective than the two-tank system, when used with good (low
loss) backup heaters. For the profiles shown in Fig. 2, this
held for variable DHW consumption, as well as for a constant
daily consumption at the monthly average (329 L/day for the solid
curve, 327 L/day for the broken line). Despite the fact that
these lines give a very closely equal average, water heated to
satisfy the broken line profile is between 1% and 2% cheaper than
that for the solid line.
CONCLUSIONS
It is clear that DHW consumption variability must be considered
in solar hot water system design and evaluation. Ultimately it
will be necessary to establish typical variability profiles,
which can be used with some confidence to describe what will
happen in the average household. Much more work is necessary
before it will be possible to establish such typical variability
profiles. One must study DHW consumption variability profile
statistics, and study the effect of these profiles on solar DHW
heaters. The variability measurements and the computer programs
will be made available to those interested in pursuing this.
This is an area which should not be ignored any longer, but which
is difficult to treat adequately within the limits of a six page
paper.
REFERENCES
Anon (1991). Updated Water Heater List (of recovery efficiency
and standby loss performance numbers for 142 gas-fired tank-type
-------
1305
DHW heaters offered in the USA for residential applications).
Blueprint, Publ. by the Building and Appliance Eff. Office of the
Calif. En. Comm., Sacramento, CA, USA, Jan./Feb. 1991, No. 33.
Arata, A.A. and F. de Winter (1991). Design and Performance of
Large Solar Water Heating Systems. A paper included in the
present Proceedings.
Baker, W.S. (1982) . Monitoring and Performance of Solar Domestic
Hot Water Systems in Oregon. Proc. of the Solar Hot Water Field
Test Technical Review Meeting, July 14-15, 1982, Atlanta,
Georgia, USA, US DOE Report DOE/CH/10122-7, Sept. 1982, Prepared
by ESG, inc. under DOE Contract DE-AC02-82-CH101222, pp 125-173.
Beckman, W.A. et al (1977). Solar Heating Design by the F-CHART
Method. John Wiley & Sons, NY, 1977, Fig. 4.1, Page 52.
Buckles, W.E. et al (1979). Analysis of Solar Water Heating
Systems. Proc. 1979 Conf. of ISES & ASES, Atlanta, GA, Vol. 2, pp
959-963, published by Pergamon Press, 1979.
Buckles, W.E. et al (1980). Analysis of Solar Domestic Hot Water
Heaters. Solar Energy, Vol. 25, 1980, pp 417-424,
de Winter, F. (1975). Heat Exchanger Penalties in Double Loop
Solar Water Heating Systems. Solar Energy, 1975, pp 335-337.
de Winter, F. (1980). Double Water Chimneys as Optimum Thermal
Diode Designs for the Interconnection of a Solar DHW Storage Tank
and a Gas-Fired Backup Tank. Proc. Annual Meeting of the Am.
Section of ISES, Phoenix, AZ, June, 1980, Vol. 3.1, pp 182-185.
de Winter, F. et al (1987) . Hawaiian Field Test Results on the
Altas Corporation Solar Augmented Two-Phase ThermoSyphon Gas-
Fired Domestic Water Heater. Proc. Annual Meeting of ASES,
Portland, OR, July 12-16, 1987, pp 280-284.
Grunes, H.E. et al (1982) . Development of an Advanced Solar
Augmented Water Heater. Final Report by Altas Corp. on Gas
Research Institute Contract 5014-343-0279, September 1982.
Grunes, H.E. and D.J. Morrison (1983). Two-Phase Thermosyphon
Heater. United States Patent 4,393,663, Issued July 19, 1983.
Lau, A. (1982) . Uses and Abuses of F-CHART. Same Proc. as Baker
(1982), pp 211-217.
Morrison, D.J. et al (1980). Development of a Gas Backup Heater
for Solar Domestic Hot Water Systems. Altas Corp. Final Report on
U.S. DOE Contract No. DE-AC02-78CS34696, June 1980.
Perlman, M. (1982). On the Performance and Reliability of
Residential Solar Water Heaters. Same Proc. as Baker (1982), pp
35-49.
Purcell, G. (1984). Personal Communication to F. de Winter, 1984.
Sha, W.T. et al (1980). COMMIX-SA-1; A Three-Dimensional Thermo
Hydrodynamic Computer Code for Solar Applications. Argonne
National Laboratory Report ANL-80-80, Nov. 1980.
-------
1306
DESIGN AND PERFORMANCE OF LARGE SOLAR WATER HEATING SYSTEMS
Adolfo Arata* and Francis de Winter**
*Universidad Santa Maria, Casilla 110-V, Valparaiso, Chile
** Altas Corporation, 1401 Laurent St., Santa Cruz, CA, USA
ABSTRACT
Solar energy has had problems not only because of cost, but also
because of limitations based on careless analysis and design.
This paper presents a comparative study of seven different system
configurations for solar water heating. This includes the system
configurations normally used, and handled in the normal analysis
and design methods. Also included are some other configurations,
with better thermal stratification using multiple tanks, coupled
thermally so heat can flow in only one direction. Several large
systems have been built in Chile. These have longterm test
results. Comparative measurements and calculations were made for
all systems. There are significant differences in performance.
It is important to choose the best design in any application.
KEYWORDS
Solar DHW System Configurations, Thermal Stratification Benefits,
Solar DHW System Test Results and Performance Predictions.
INTRODUCTION
Solar energy has often been too costly. Many installations have
however been poor (Spielvogel, 1980) . This is not only due to
problems in materials and maintenance, but also to the poor
designs often used. This paper presents a performance analysis of
seven system configurations using solar collectors for producing
domestic hot water (DHW). This involves different solar
conditions and different DHW daily draw schedules. These
conditions are kept the same every day. This is realistic for
hot water usage in many large applications, and for the weather
in many areas in Chile. The designs include the ones
traditionally used, and represented in the "F-CHART" and in the
"Phi-F-Chart" methods (Beckman et al, 1977; Liu and Jordan,
1978). They also include the "Altas/DOE/GRI" system proposed for
residential usage (de Winter, 1989), and several systems built
for large scale DHW applications in Chile (Arata et al, 1985).
DESCRIPTION OF THE CONFIGURATIONS STUDIED
The systems studied involve flat plate collectors for the
collection of solar energy, and water tanks for heat storage. All
systems use pumps and controllers, all tanks are fully mixed, and
backup heating is used when the solar heating is not enough.
-------
1307
The seven system configurations are characterized by the backup
heating arrangement, which can be in series, in parallel, or in
one of the tanks; and by the number and the interconnection of
the storage tanks. The configurations are shown in Fig. 1, and
comments on the designs follow below. In all the designs there
can be a heat exchanger (with the associated heat exchanger
factor) between the storage tanks and the collectors.
OEMANO
System No. 4
OEMANO
iSC
BC
supply
System No. 1
System No. 5
FA
iSC ¦
supply
System No. 2
BC
supply
^ fL OEMANO
System No. 6
—I JTH
II
.SC
Isc
BC
rBC
Supply
supply System No. 7
System No. 3
Fig. 1. Diagrams of the Seven Different Solar Water Heating
System Configurations considered in the Present Study.
SYSTEM NUMBER 1. This has a single storage tank with a pumped,
controlled loop to the collector. An instantaneous backup heater
is used in parallel, with solar heated water only used when it is
hotter than the control temperature of the backup heater.
SYSTEM NUMBER 2. This is similar, except that the backup heater
is in series so it is always fed with solar pre-heated water.
SYSTEM NUMBER 3. This has a single tank, with the backup heater
and the solar loop both heating the tank. This is the standard
"one-tank" system. Since it is assumed to be totally mixed,
solar heat can only be collected when the collector can operate
above the backup control temperature, or when the backup heater
alone can not keep up with the demand.
SYSTEM NUMBER 4. This has thermal storage in two tanks. One of
these is heated with the collectors, the other with the backup
-------
1308
heater, and they are coupled only when there is consumption.
This is the standard "two-tank" system.
SYSTEM NUMBER 5. This is similar to the two-tank system of type
4, except in that there is a controller and a transfer pump to
ensure that the backup tank is never colder than the solar tank.
The transfer pump serves as a "thermal diode" which readily
transfers heat from the solar tank to the backup tank, but offers
a very large resistance to heat flows in the opposite direction.
This is similar to the Altas/DOE/GRI DHW heater concept (de
Winter et al, 1991), which uses a passive thermal diode device
instead of a pump, with a conductance ratio of about 5000 to 1.
SYSTEM NUMBER 6. This has the thermal storage in three tanks.
It is similar to the two-tank system, except the solar tank is
divided in two for additional stratification, and these two parts
are coupled with a controlled transfer pump: a thermal diode.
Coupling between the solar tanks and the backup tank only occurs
when hot water is used. A system of this type has been in long
term operation in the Salvador Chilean copper mine of Codelco
(Arata et al, 1985).
SYSTEM NUMBER 7. This is similar to the system of type 6, except
there is also a thermal diode between the hot solar tank and the
backup tank. A system of this type has also been in long term
operation in the Salvador Chilean copper mine (Arata et al,
1985).
CONDITIONS USED FOR COMPARATIVE ANALYSIS OF SYSTEMS
Consumption Profiles Used
All systems used the same DHW consumption at 40 C, with four
different types of consumption profiles. Consumption is the same
every day, representative of many large applications. The highly
variable consumption in residences is treated by de Winter et al
(1991). The consumption profiles are specified below.
CONSUMPTION OF TYPE 1: DHW SUPPLIED FOR THREE WORKING SHIFTS.
30% in the hour after midnight, 30% in the hour after 09:00, 30%
in the hour after 16:00, and 10% in the hour after 17:00.
CONSUMPTION OF TYPE 2: DHW SUPPLIED DURING THE HOTTEST HOURS.
30% each in the hours after 15:00, 16:00, and 17:00, and 10% in
the hour after 18:00.
CONSUMPTION OF TYPE 3: DHW SUPPLIED DURING DAYLIGHT HOURS.
10% each in the ten hours following 08:00.
CONSUMPTION OF TYPE 4: DHW SUPPLIED AROUND THE CLOCK.
4.1667% each of the 24 hours of the day.
Solar and Meteorological Conditions
We used solar and meteorological conditions corresponding to a
representative day for three different regions. In each case we
need the following hourly information:
Radiation incident, and incident angle on the collector.
Ambient and input water temperature, and wind speed.
Fig. 2 shows the values used for regions of high, medium, and low
solar inputs. The water supply temperature was taken to be 18&,
15t, and lCfC for the high, medium, and low solar input regions
respectively. For all regions we used a wind speed of 4 m/s.
Collector Characteristics
All systems were analyzed with a flat plate collector having an
Fr and the normal tau.alpha product of 0.6885, and an Fr and the
U1 loss coefficient product of 3.75 w/sq.m/C.
-------
1309
»tl Illy/tn2!
90
T—I—I—I—I—I—r~l—I—1~1—I—|—r
HT11.337 Wh/mdi'o
TqCCJ Iffy/m2)
90
High Insolation
1.1 i i i i I 1 1 I I—>4 i.
Fig. 2. Ambient Temperature,
Incident Radiation, and Angle
of Incidence versus Solar
Hour for the Three Regions
Considered in this Study:
of High, Medium, and Low
Insolation Levels.
T0 CC)
Medium
T„rci
»,!_•) frtwi(rf)
to
Ht = 1552Wh/m di'o
- 1200
Low
Insolation
1C II 20
Storage Systems
All systems were considered to have a storage system equal in
capacity to the daily usage. In any particular system all tanks
are of equal size. All tanks and pipes are assumed to be
insulated well enough so that heat losses are negligible. In
large systems this is easy. The backup heater was assumed to
have a firing (recovery) efficiency of 80%.
METHODOLOGY
A computer program "SOL 10" was prepared at the Santa Maria
University to simulate the performance of the various system
types under various consumption and solar input conditions. The
program was validated with laboratory prototypes, and also with
industrial installations to confirm the effect of scale. The
difference between measured and predicted values was 2.6% on the
average, with a 5% maximum. Table 1 shows the differences
between predicted values and for those measured in the
laboratory.
Despite the advantages of this computer model, it is not a
practical design tool because of the need for much in data and in
calculation times. This led us to develop a design method
[(phi-bar)-(Ht-bar)] versus f (Arata et al, 1986), which yields
generalized plots allowing the rapid sizing of installations
involving any of the seven systems involved.
-------
1310
Table 1. Difference Between the Measured and Predicted Valnas
on the Daily Performance for the Seven Different Systems
System Average Difference Maximum Difference
1 2.2% 5%
2 0.7 2
3 2.0 5
4 0.6 2
5 2.6 4
6 1.0 2
7 1.0 1
To demonstrate the reliability of the SOL 10 (Arata et al, 1982)
model and of the [(phi-bar)-(Ht-bar)] versus f approaqh, Table 2
shows the measured fraction "fm" on a real installation, compared
to the SOL 10 results shown as "fs," and the [(phi-bar)-(Ht-bar)]
versus f results shown as "fp." It is an installation which
involves 320 sg. m of collection and about 50 tons of solar
storage, with a backup heater in a tank of 16 tons, involving a
configuration of type 7. It supplies 45 cubic meters of 40 C
water per day for the showers of several thousand copper miners
in three shifts at the Salvador mine, following a DHW demand
profile of type 1. It has been working very successfully for
many years, and was described earlier in SunWorld (Arata, 1985).
Table 2 Monthly Performance Of Water- Wearing Rystpma nf Typg
7. With the Backup Hpatgr 3n Final. Tank, and With Transfer
Pumps Between the Successive Tanks.
Month
fp
fs
fm
Jan
100%
100%
99%
Feb
85
88
85
Mar
88
90
84
Apr
79
85
78
May
60
62
57
Jun
59
60
56
Jul
65
70
66
Aug
70
72
73
Sep
78
76
78
Oct
80
78
83
Nov
89
89
89
Dec
95
95
94
Annual
Average
78
80
78
COMPARATIVE RESULTS THROUGH THE USE OF SOL 10
Considering the relationship between solar fraction and energy
savings, and between collector area and the cost of capital, the
performance of the different system types can be compared using a
ratio (solar fraction)/(collector area), normalized to a value of
1.00 for the system which performs best. The highest value
corresponds to the highest cost-effectiveness. Tables 3, 4, and
5 show the results obtained with SOL 10 for the various system
types, for the different consumption profiles and different solar
radiation regions.
-------
1311
Table 3 High Radiation Region Values of the Ratio
(Solar Fraction)/(Collector Area)
Consumption System Type
Type
1
2
3
4
5
6
7
1
.89
.89
.82
.93
.94
.98
1.00
2
.86
.87
.86
.91
.92
.95
.97
3
.88
.88
.88
.89
.91
.95
.95
4
.85
.85
.85
.92
.94
.96
.98
Table 4 Medium Radiation Region Values of the Ratio
(Solar Fraction)/(Collector Area)
Consumption System Type
Type
1
2
3
4
5
6
7
1
.40
.51
.38
.52
.52
.57
.57
2
.41
.51
.39
.52
.53
.56
.56
3
.41
.50
.41
.50
.50
.53
.53
4
.41-
.50
.39
.51
.51
.55
.55
Table 5 Low Radiation Region Values of the Ratio
(Solar Fraction)/(Collector Area)
Consumption
System
Type
Type
1
2
3
4
5
6
7
1
.05
.17
.04
.17
.17
.19
.19
2
.05
.17
.04
.17
.17
.18
.18
3
.05
.17
.04
.17
.17
.18
.18
4
.05
.17
.05
.17
.17
.18
.18
From these values one can conclude:
a. It is better to have a backup heater in series than parallel.
b. The daily draw schedule affects the solar fraction, but this
is evident only at the higher insolation levels.
c. Systems with more tanks have better performance. With the
improved stratification behavior, the collector is able to
receive colder water, enhancing collection efficiency.
d. The transfer pumps or thermal diodes are able to improve the
performance, especially when operating at high solar fractions.
GENERAL CONCLUSIONS
The effects of DHW system configuration were examined under the
following conditions:
a. Fully mixed tanks,
b. No heat losses from the tanks,
c. Same DHW consumption day-in-day-out,
d. Same solar and weather conditions day-in-day-out.
Under these conditions, it is shown that there are better system
designs than those generally used (systems 1-4). System 5 is
clearly better than system 4, system 6 is better still, and
system 7 is clearly the best. If tank mixing can not be
controlled, it is clearly better to use several successive,
segregated tanks in order to produce effective stratification.
-------
1312
The conditions used are quite reasonable for the large scale DHW
applications considered in a Chilean (largely sunny) climate. It
should however be noted that the assumption that conditions are
the same every day and that the tanks are adiabatic probably
tends to lead to an underestimation of the advantages of the
systems of type 5, 6, and 7, when used in smaller scale systems,
and in more variable climates (see de Winter et al, 1991).
The improvements are slight, and the question might be raised:
why bother with such design options? It is a matter of cost. In
the context of a system with 320 square meters of collector and
some 66 cubic meters of tanking, a few pumps costing perhaps
U.S.$100 each is a small price to pay for a 1% (or more)
improvement. In small residential systems, the passive thermal
diode device used in system 5 would probably cost less than $1.
The cost of having multiple tanks is also relatively small.
There is, for example, no reason to believe that two 25 cubic
meter tanks (installed at the plant site) are necessarily more
expensive than one of 50 cubic meters. In the Altas/DOE/GRI
system, the cost of going from a type 3 (one-tank) to a type 5
(two-tank) system was estimated to be only $30 in manufacturing
and material costs on the tanks, plus perhaps $1 for the thermal
diode.
REFERENCES
Arata, A.A. and S.J. icazategui (1982). Programas Computacionales
Simuladores SOL 10, SOL 5, Sol 25. Internal Reports, Universidad
Santa Maria, Solar Energy Lab, Valparaiso, Chile, 1982.
Arata, A.A., S.J. Icazategui, and E. Villalobos (1985). Result
and Performance Report of a Solar Heating Plant. SunWorld 1985.
Arata, A.A. and J.S. Icazategui (1986). A General Design Method
for Liquid Heating Solar Plants. (Plantas de Colectores Solares
Pianos), Manual de Proyectos, Universidad Santa Maria,
Valparaiso, Chile, 1986.
Beckman, W.A. et al (1977). Solar Heating Design by the F-chart
Method. John Wiley & Sons, NY, 1977.
de Winter, F., A.A. Arata, and J.S. Icazategui (1985). Thermal
Coupling Requirements and Possibilities of Backup Heater Tanks in
Solar Hot Water Systems. Proc. of the 1985 Ninth Biennial
(Montreal) Congress of ISES, Pergamon Press, Vol. 1, pp 601-605.
de Winter, F. (1989). Active Solar Water and Space Heating - Past
Accomplishments and Future Needs. Proc. of the Annual Conf. of
ASES, Denver, CO, June 19-22, 1989, pp 105-111.
de Winter, F., A. Arata, and M. Perlman (1991). Design and
Performance of Small Solar Water Heating Systems. A paper
included in the present Proceedings.
Liu, B. and R. Jordan (1978). Applications of Solar Energy for
Solar Heating and Cooling of Buildings. ASHRAE Report 170, NY,
1978.
Spielvogel, L.G. (1980) . The Solar Bottom Line. ASHRAE Journal,
November 1980.
-------
1313
DEVELOPING PERFORMANCE MODELS OF SOLAR ENERGY SYSTEMS
BY DAILY ENERGY INPUT/OUTPUT CURVES
William S. Duff
Eileen C. Boardman
Solar Energy Applications Laboratory
Colorado State University
Fort Collins, Colorado 80523
ABSTRACT
A method is developed for generating performance models of specific solar energy systems
from a few TRNSYS runs or from experimental data. The resulting models have simple
form and are extremely easy to use while retaining dynamic effects of system operation.
The method is demonstrated for a DHW and an IPH system and is shown to be very
accurate on a daily basis for a range of climates and operating conditions.
Performance data are first generated from TRNSYS runs or from experimental data.
Combinations of variables occurring in a conductance model of the collection system are
then chosen for inclusion in a statistical regression analysis. The model is tested for
statistical significance. Variables chosen for the weather input are those readily available:
daily radiation and mean twenty-four hour average temperature. An IPH example model
is included which has been developed in a two stage procedure with validation and
accuracy demonstrated. The resulting model is highly statistically significant and is very
accurate on a daily basis. Another example, a DHW model, is included to demonstrate
the flexibility of this procedure.
INTRODUCTION
The performance of systems, subsystems, and components may sometimes be represented
accurately by means of a small subset of parameters. The daily energy input/output curve
methodology developed in Task VI of the International Energy Agency Solar Heating and
Cooling Program is such an approach. See Duff[2].
The daily energy input/output curve takes the form
Q = a. + ax + a,x2 + — + a„x.
-------
1314
where the x, are values of daily insolation and a few other location and system parameters,
their squares, cross products, and so forth. The equation predicts daily energy delivered
by the collection system for these daily values. The coefficients a, in the equation are
obtained from a multi-variate linear regression analysis of points obtained from
experiments or simulations.
Since the required long term experimental data are not available, performance data are
generated by TRNSYS[4]. TRNSYS has been shown to accurately represent experimental
results. See Mitchell et al. [3]. Although the generation of such a TRNSYS model is
difficult and time consuming, the resulting elaborated daily energy input/output model is
extremely easy to use and is highly accurate.
The first example system included involves solar industrial process heating, where accurate
short-cut prediction methods are not available. The second example is a domestic hot
water system which demonstrates the flexibility of this approach. The daily accuracy of
the model developed by this method allows such a system to be evaluated with using
elaborate sensitivity analyses for any climate and location, using readily available weather
inputs.
MODEL DEVELOPMENT
A year long period and a variety of locations will be used to develop the model. Day
length then becomes an important variable since a January day and a June day with the
same amount of radiation on the collector have different daily energy outputs due to the
different radiation patterns and operating times.
Data generated by TRNSYS are used to develop a multiple regression model. These
points and candidate variables are entered into a stepwise regression analysis, where
possible models are developed and analyzed. Only those variables which have been
entered are considered, and only in the form which has been entered. Possible models
are developed. Since several models may be as good statistically, the engineer then uses
knowledge of these variables to select that model which will provide the closest match to
physical models. The goodness of any model can be measured by its ability to predict
accurately and its robustness. The models developed here are extremely accurate and
robust.
SELECTION OF VARIABLES
The performance model developed generalizes the concept of the daily energy input/output
line. The variables in the model are chosen from basic physical principles which evolve
from a conductance collection model by Guisan et al.[l] developed for evacuated tube
collectors. These variables and the significant set that resulted during the multiple
regression analysis process were nearly the same set or, in some cases, exactly the same
set. A conductance model is also used in the TRNSYS simulation. In the conductance
model T„ is the temperature of the absorber plate, T^, is the ambient temperature. The
collector output is Q. = U^T^ - T„), and the collector loss is UL = ((0.56 + .0065 * (T*. -
T„)) * (T*.-T_»), where U„ is the conductance between the absorber and the fluid.
Following Guisan et al., this leads to
-------
1315
Q. = A*I - B*aT*I - C*r - D*aT - E*aT (1)
Although equation (1) is for instantaneous radiation, these variables could be
integrated over operating conditions to obtain daily results. However, actual operating
time and conditions are not known in advance. Thus, a daily model is structured using
all-day radiation and by including day length, DL, in the terms relating to operating times.
The form of this all-day model is
0 = 3, + a,*H + a*H' + 3*aT*DL + a*AP*DL + a*AT*DL*H (2)
where H is the radiation on the collector for the entire day.
Day length, DL, in hours, is found from the equation
DL = (2/15) cos' (-tan 1 * tan d) (3)
where 1 is the latitude and d is the declination.
The day length multiplied at aT is proportional to daily losses and is the variable which
distinguishes lengthy summer operating periods from shorter winter operating periods. The
aT used here is the inlet temperature minus the 24-hour average ambient temperature,
where the inlet temperature is constant. The 24-hour temperature is used because it is
the information which is generally available from weather sources. Also, overnight
temperature has some influence on system performance since it determines the starting
temperature of the array.
THE EFFECT OF CLIMATE ON THE MODEL
The effect of climate on the model was investigated by developing the IPH model in two
stages. In the first stage the model was developed from Toronto data only. The resulting
model was tested for statistical significance and then validated for other climate types:
Albuquerque, Miami, and Seattle. In the second stage the model was developed from all
four locations, tested for statistical significance and then validated for two additional
locations: Boston and El Paso.
INDUSTRIAL PROCESS HEAT (IPH)
Solar industrial process heating involves unique system characteristics not found in space
heating and domestic hot water applications. Typically, the arrays are very large and pipe
runs may be extensive due to site limitations. The system may be open loop or closed
loop. The system may be direct, where the process fluid is circulated to the collectors, or
indirect, where a heat exchanger is used between the collector loop and the process loop.
Also, the load is often in phase with the collection.
The modeled system is an industrial process heat or district heating application without
storage, where return temperature from the load is constant. Return fluid temperatures
of 40°C, 80"C, and 110C are used in model development. The collector fluid is 50%
ethylene-glycol and 50% water. Collector tilt is 10° less than latitude. A heat exchanger
-------
1316
with 0.9 effectiveness is used between the collector and load loops. Extensive piping is
required for these large systems which can result in large losses.
A single climate performance model was developed from six months of Toronto, Canada
data chosen because they displayed a full representation of day length and seasonal effects.
The data for these months were chosen over a period of three years so that a greater
variety of weather is represented. This model, developed from one location was validated
against three diverse locations: Albuquerque, Seattle, and Miami.
The model to be investigated was
Q. = a + b*H + c*H! + d*DL*AT + e*DL*AP + f*DL*H* XT (5)
Stepwise regression analysis can be used to show the relative significance of the variables
entering this model and to eliminate any variables that are statistically redundant. The
variables chosen for consideration were H, IT, aT*DL, aT**DL, and DL*aT*H. The first
variable to enter the equation was H which provided an R' of 94.99%. This means that
94.99% of the variability in the energy collected is explained by the daily radiation on the
collector over the whole day. This high level of significance, (much less than 0.001),
explains the great accuracy of the a + bH model discussed earlier. The next variable of
importance, given that H had already entered, was DL*-aP which, together with H,
provided an R: of 99.47%. Additional variables come in the order of H!, DL*aT*H and
finally aT*DL, without any previous variables dropping out. All variables are statistically
significant. When all variables are in the model, the original order of entry does not
necessarily indicate the importance in the final model. In this case, the most important
variable is H, with DL*£>P being next most important, then H!, DL*/T*H, and DL*aT.
The specific equation for the energy collected is
Q. = 2.3290 + .6775*H - .0015389*HJ - .000387*aT*DL - (6)
.000012* aP*DL - .000015*H*DL*aT
Considering points predicted by the model and the points generated by TRNSYS for
May-June, the agreement is very good, corresponding to the high R2 value determined
from the regression analysis. If aT and DL are held constant for the month, a good linear
fit is obtained as well, although a better fit is achieved by treating aT and DL as variables.
Validation of the single climate model was performed at three locations: Albuquerque,
Seattle, and Miami: chosen for climate diversity. Weather input was provided by the
Typical Meteorological Year tapes.
THE MULTICLIMATE IPI I MODEL
Although the model developed from Toronto weather data performed very well for greatly
differing climates, it would be expected that a model based on a diversity of climates
would be even more robust. The multiclimate performance model was developed from
the three TMY locations on which the first stage model was validated plus the original
Toronto data.
-------
1317
The multiclimate performance model is based on the same variables as the first stage and
is of the same form. Again, stepwise regression analysis was performed. The first variable
to enter the equation was again H, which provided an R* of 96.49%. The next variable
of importance, given that H had already entered, was DL*aT which, together with H,
provided an R2 of 99.64%. Additional variables come in the order H2, DL*aP and then
DL*aT*H. The R2 for the complete model is 99.74%. In the final form of the model,
the most important variable was H, the second most important was DL*aP, then H2,
DL*aT*H and DL*aT.
The specific equation for the energy out of the collector is
Q.= -.571905 + .680446*H - ,001528*H! - .000009793*aP*DL (7)
- .00005582*DLVT*H - ,OOH38*DL*aT.
Results for the three TMY locations show a very close fit for all locations with this model.
The model is seen to be extremely accurate not only on a yearly basis, but also on a daily
basis.
Results of the single climate and multiclimate models were compared. As would be
expected, the addition of climate diversity produces an even better model.
The multiclimate model was validated against two other very diverse locations: El Paso,
Texas and Boston, Massachusetts. The accuracy shown for these locations indicates that
this model can be used with confidence for any location.
DOMESTIC HOT WATER
A domestic hot water system performance model is developed to demonstrate the
flexibility of this approach for generating performance models. The system consists of a
flat-plate collector, storage, and a heat exchanger between the collector and load loops.
The collector parameters are: area 6.5 m!, collector efficiency F'= 0.88, U„ = 0.75 KJ/hr-
m2-K, e = 0.90 and a = 0.95. The storage is 454 liters and the heat exchanger has 0.9
constant effectiveness. The load is 300 kg. per day modeled with a Rand profile. Ground
water temperatures of 10°C and 20°C were used. Weather data was provided by TMY
tapes for Albuquerque, Miami, Seattle, and Madison, Wisconsin.
Variables considered for inclusion in the domestic hot water model are again those which
are readily available, in combinations that successfully modeled year long performance in
the industrial process heating example. These are H, H2, DL*aT, DL* aP , and DL* aT*H.
In the domestic hot water system with the flat plate collector, the DL*aT! is not significant
and so is not included in the model. Here, aT is the difference between the ground water
temperature and the daily ambient temperature, since these are the available data. This
T is shown to be a very accurate predictor of system performance when combined with
the other variables. The most important variable is again the H, which alone provides an
R!=96.80%. The next most important variable is DL*aT which, together with H provides
R2 = 99.66%. For the complete model the R = 99.67%. The final model shows the most
important variable to be radiation, the second most important to be DL*aT, then H! and
J)L*aT*H The model is
-------
1318
Q = -.57947 + .621429*H - .001004*H' - ,00622*DL*aT - (8)
.00002457*DL*H*aT
Model validation was performed for Boston and El Paso. Ground water temperatures of
10°C and 20°C were used for both locations. In addition a 15*C ground water temperature
was run for Boston. Results show very close agreement with TRNSYS predictions.
CONCLUSIONS
A method has been developed for generating performance models of solar energy systems.
Although the generation of such models requires extensive experimental results or
TRNSYS runs, the resulting models are extremely easy to use and very accurate, even on
a daily basis. Model validation for industrial process heat systems and domestic hot water
systems demonstrates the robustness of this technique in greatly differing climates. The
specific models developed here are for specific systems, but with a range of operating
conditions. These models will be valid anywhere within the range of operating conditions
and for reasonable extrapolations. The contribution of this method is not the specific
models developed here but the technique which allows for the generation of such models
for any system for which adequate experimental data is available or for which TRNSYS
simulation is a valid approach.
REFERENCES
1. 0. Guisan, A. Mermoud, B. Lachal, 0. Rudaz, "Evacuated Collector Systems
Characterization," Report No. IEA-SHAC-TVI-3, University of Geneva (1985).
2. IEA Task VI Report, "Experimental Results from Eleven Evacuated
Collectors," W.S. Duff, editor, Solar Energy Applications Laboratory, G±ra±)
State University (1985).
3. J.W. Mitchell, W.A. Beckman, M.J. Pawelski, "Comparisons of Measured and
Simulated Performance for CSU Solar House I," Journal of Solar Energy
Engineering, 102 (1980).
4. TRNSYS, A Transient Simulation Program, University of Wisconsin, Report 38-11,
April (1981).
-------
1319
COMPARISON OF EXPERIMENTAL AND TRNSYS RATINGS OF GENERIC
DRAIN-BACK SOLAR WATER HEATERS
W. T. Carlson, J. H. Davidson, and W.S. Duff
Solar Energy Applications Laboratory
Colorado State University
Fort Collins, CO 80524
ABSTRACT
The use of a simulation model to predict SRCC energy ratings of solar domestic hot water systems
has been proposed as a cost effective alternative to laboratory ratings. A comparison is made
between laboratory and TRNSYS SRCC ratings for a drain-back system. A two-level, five-factor,
half-factorial experimental design is used to evaluate the ability of TRNSYS to predict the energy
ratings. The five component and operating conditions varied are collector area, collector flow rate,
recirculation flow rate, storage tank volume, and storage tank design. The factorial analysis
indicates TRNSYS accurately predicts changes in rating due to the changes in the five factors
considered and could be used to eliminate duplicate laboratory ratings required for component
substitutions. TRNSYS predicts the absolute values of net energy delivered and auxiliary energy
within an experimental error of two standard deviations and energy delivered from the solar storage
tank within a maximum error of 11 percent, well within three standard deviations.
KEYWORDS
Solar; hot water heating; drain-back; SRCC; TRNSYS.
INTRODUCTION
Certification of solar domestic hot water systems (DHW) by the Solar Rating and Certification
Corporation (SRCC) currently requires a four day laboratory test (SRCC, 1984) which must be
repeated for all component substitutions and operating modifications. This procedure is an economic
burden on a low volume industry and reduces the competitiveness of solar DHW versus
conventional DHW systems. The case for certification however is sound. The lack of standardized
ratings during the solar boom of the tax credit days had a detrimental impact on the industry. The
ability of the transient system simulation TRNSYS (Klein, et al. 1976) to accurately predict drain-
back system ratings is evaluated as a first step in implementing a more cost effective rating
procedure. The evaluation compares laboratory ratings (Carlson, 1990; Davidson et al., 1991) to
TRNSYS ratings over a range of component and operating parameters.
METHODOLOGY
The SRCC rating procedure as implemented in this study is included in Davidson et al. (1991).
SRCC energy based rating quantities shown in Fig. 1 include: useful collected energy, Qu; daily hot
water energy delivered by the solar storage tank, Qs; and net energy delivered from the solar storage
tank, Qnet, equal to Qs minus parasitic energy Qpar. The auxiliary energy, Qaux, is the energy input
into the auxiliary hot water heater.
-------
1320
del
loss.aux par
loss,s
aux
Auxiliary
Heater
Collector Array
Solar
System
and
Storage
Fig. 1. SRCC energy quantities
The comparative analysis is based on a two-level, half-factorial designed experiment in which
collector area and flow rate, recirculation flow rate, solar storage tank volume and tank design are
varied in sixteen trials. The high/low levels of each design or operating factor are based on current
industry standards.
The TRNSYS simulation data were provided by the University of Wisconsin. Detailed descriptions
of measured flow rates, piping lengths, heat exchanger effectiveness, overall tank conductances,
tank volumes and pump power characteristics were provided to TRNSYS engineers. A 3-node tank
model was used for the eight trials without the stratification manifold and a 10-node model was used
for the trials with stratification manifold.
RESULTS
Table 1 lists the system configuration used for each trial and summarizes the simulation results.
Scatter plots comparing the experimental and simulation results are shown in Figs. 2-5 for Qs, Qnet,
Qaux. Qu respectively. The experimental error bands on Qs, Qnet, and QaUX are two standard
errors based on the measurement transducer accuracies. The error bands include the measurement
error for the particular quantity plus the measurement error associated with Qu. The addition of
measurement error is necessary since any error in Qu manifests itself in the other quantities through
errors in the energy input into the system.
Figure 2 shows that the simulation results are within the error bands of Qs for the eight trials with
high collector area (5.56 m2). The simulation data in the low collector area trials (2.78 m2) falls
below the experimental data and outside the error bands in six of the eight trials. A bias also exists
in the high collector area trials in which the larger storage tank (310 liter) is used.
The comparison of ratings of Qnet shown in Fig. 3 indicates that even though TRNSYS sometimes
under predicts Qs, all the simulated determinations of Qnet are within the error bands of the
experimental values except in trial 9 where the simulated value is low. One trial outside of the error
bands is acceptable for a population of sixteen when using a 2o error.
Figure 4 indicates that except for trials 2 and 3, the simulated ratings of Qaux lie within the error
bands. In these configurations, parasitic losses are high relative to thermal output.
-------
1321
Table 1. Numerical Results of Two-Level Half-Factorial Experimental Design
Collector
Recircula
Storage
Flow
-tion
Collector
Tank
Storage
Rate
How
Area
Volume
Tank
Os
Qnet
Qaux
Qu
Trial
(1/s)
Rate (1/s)
(m2)
a)
Design
(kJ)
(kJ)
(kJ)
(kJ)
1
.057
.047
2.78
250
Basic
15900
10365
39140
18130
2
.114
.047
2.78
250
Manifold
17350
11910
37690
18980
3
.057
.095
2.78
250
Manifold
16470
11070
38570
18490
4
.114
.095
2.78
250
Basic
16850
11450
38180
18870
5
.057
.047
5.56
250
Manifold
25370
19970
29670
30620
6
.114
.047
5.56
250
Basic
25380
19980
29660
30970
7
.057
.095
5.56
250
Basic
24600
19200
30440
30370
8
.114
.095
5.56
250
Manifold
26830
21430
28210
32110
9
.057
.047
2.78
310
Manifold
16690
10919
38350
18680
10
.114
.047
2.78
310
Basic
16410
10872
38630
18650
11
.057
.095
2.78
310
Basic
15840
10308
39200
18290
12
.114
.095
2.78
310
Manifold
17250
11850
37790
19200
13
.057
.047
5.56
310
Basic
24150
18750
30890
30140
14
.114
.047
5.56
310
Manifold
26890
21490
28150
32280
15
.057
.095
5.56
310
Manifold
25360
19960
29680
31160
16
.114
.095
5.56
310
Basic
25690
20290
29350
31650
0 Experimental
~ Numerical
'i
ii !ii'
i i i i i i i i i i i
23456789 10111213141516
25 '
Trial
20-
0 Experimental
• Numerical
iiil
I I I I I I 1 I I I I I I I
123456789 10111213141516
Fig. 2. Comparison of Experimental and
TRNSYS determinations of Qs.
based on 2a experimental error.
Trial
Fig. 3. Comparison of Experimental and
TRNSYS determinations of Qnet
based on 2a experimental error.
-------
1322
45
40
W)
fc-
4>
G
w
35
30
25
0 Experimental
~ Numerical
[,
i i i i i i i i i i i
1 2 3 4 5 6 7 8 9 10111213141516
Trial
Fig. 4. Comparison of Experimental and
TRNSYS determinations of Qaux
based on 2a experimental error.
35 -
00
U
a
C
H
25
15
0 Experimental
• Numerical
¦ 11111111111111
123456789 10111213141516
Trial
Fig. 5. Comparison of Experimental and
TRNSYS determinations of Q„
based on 2ct experimental error.
As expected, since both experiment and numerical simulation use the same collector performance
model, Fig. 5 indicates good agreement in Qu . The only potential source of discrepancy between
the experimental and simulation data for Qu is due to differences in collector inlet temperature.
A factorial analysis of the simulation data indicates that as in the experimental trials, collector area,
collector flow rate, and tank design have the most significant effects on Qs, Qnet, Qaux> and Qu.
Collector area alone accounts for 97% of the variation in Qs, Qnet, and Qu and 99% of the variation
inQu.
A comparison of analysis of variance (ANOVA) models of the experimental and simulated ratings is
used to determine if any of the factors or factor interactions have a statistically significant effect on
the differences between the two sources of data. The ANOVA models calculated from the TRNSYS
data for responses Qs, Qnet, Qaux. and Qu are given by equations 1-4. The first term of each model
is the mean value of the sixteen trials. The remaining terms include a coefficient equal to one-half the
effect due to a factor multiplied by the factor level, L, equal to +1 if the factor level is high and -1 if
the factor level is low.
Qs = 21064 + 4470 Lcollector area + 517Lcollector flow rate " 462 Ljank design kJ (1)
Qnet = 15613 + 4521 Lcollector area + 546 Lcollector flow rate " 462 Ljank design kJ (2)
Qaux = 33975 - 4469 Lcollector area " 518 Lcollector flow rate + 462 Lxank design kj (3)
Qu = 24912 + 6251 Lcollector area + 427 lcollector flow rate kJ (4)
-------
1323
The experimental minus the simulation ANOVA models for Qs, Qnet, Qaux, and Qu are given by
equations 5-8.
AQs = 976 - 343 Lxank design " 278 Lcollector area - 72 Lcollector flow rate kJ (5)
A Qnet = 583 - 269 Lcollector area " 183 Lcollector flow rate - 8 L-ranj; design kJ (6)
AQaux = -531 + 415 Ljank design + 353 Lcollector area - 95 Lcollector flow rate kJ (7)
AQU = -615 - 190 Lcollector area " 31 Lcollector flow rate kJ (8)
The AQ ANOVA models predict the difference between the experimental data and the simulation data
for changes in the factor levels included in these models. The first term in each difference model is
the average bias between the experimental and TRNSYS ANOVA models. The ability of TRNSYS
to predict changes in a response is assessed by neglecting the bias terms and then summing the
absolute values of the remaining coefficients. This maximum relative difference is summarized in
Table 2 along with the experimental error associated with the baseline system of trial 16.
Table 2. Comparison of Error Predicted by AQ ANOVA Models to Experimental
Error
ANOVA
Model
Maximum Relative
Difference Predicted by
AQ ANOVA Model
(kJ)
Calculated
Measurement
Error
(kJ)
Statistically Determined
Random
Measurement Error
(kJ)
aQs
693
1648
966
AQnet
460
1649
880
AQaux
863
1638
898
aQu
221
1630
1028
The maximum relative differences predicted by the aQs, AQnet, AQaUX and aQu ANOVA models are
less than the calculated measurement error. Thus,TRNSYS can predict changes in the energy
responses Qs, Qnet. and Qaux for changes in collector area, collector flow rats, and tank design and
changes in the energy response Qu for changes in collector area and collector flow rate within the
accuracy of experimental error.
The biases in the differences limits TRNSYS' ability to predict absolute energy rating values for Qs
and Qnet. The bias may be due to several factors other than instrumentation. It is unknown how
much of the pump energy ends up in the fluid and how much is lost to ambient. The differences or
biases between experimental and simulation data for the responses Qs and Qnet are well within the
magnitude of the pumping energy. Another source of bias could be modeling errors of the heat loss
coefficients for the drain back module and the piping.
CONCLUSIONS
Comparison of ratings obtained with TRNSYS to experimental ratings validate the effectiveness of
TRNSYS in accurately predicting changes in rating due to changes in collector area, collector flow
rate, recirculation flow rate, storage tank volume and storage tank design within the limitations
imposed on these design factors in this study. The effects of more advanced concepts in domestic
water heating cannot be inferred from this study. Determining the ability of TRNSYS to predict the
absolute values and not just changes in Qs and Qnet is dependent on analysis of the sources of bias
between the experimental and simulation data. The ability of TRNSYS to predict system behavior is
-------
1324
also predicated on the accuracy of supplied system descriptions and the validity of the TRNSYS
input deck.
ACKNOWLEDGEMENTS
This work was supported by the United States Department of Energy. TRNSYS data were supplied
by Dr. William Beckman and his students at the University of Wisconsin, Madison.
REFERENCES
Solar Rating and Certification Corporation, Washington, D.C. (1984). Operating Guidelines for
Certifying Solar Water Heating: Systems. Document OG-200.
Klein, S.A., et al. (1976). TRNSYS. A Transient Simulation Program. Engineering Experiment
Station Report 38. Solar Energy Laboratory, University of Wisconsin, Madison.
Carlson, W.T. (1990) Comparison of Experimental and TRNSYS SRCC Ratings of a Generic Drain
Rack Solar Water System. Masters Thesis, Colorado State University.
Davidson, J.H., Carlson, W.T., and Duff, W.S. (1991). Impact of Component Selection on SRCC
Ratings of Drain-back Solar Water Heaters. In proceedings of 1991 Solar World Congress,
Denver, Colorado.
-------
1325
MODEL-TO-MODEL TESTING OF
SIX SOLAR ENERGY DESIGN PROGRAMS
William S. Duff
M. Chandrashekar
Solar Energy Applications Laboratory
Colorado State University
Fort Collins, Colorado USA
Department of Systems Design.
University of Waterloo
Waterloo, Ontario CANADA
ABSTRACT
This paper summarizes the results of a workshop to compare six of the most
popular solar energy design programs.
KEYWORDS
Simulation, solar energy, computer model evaluation, and tools.
INTRODUCTION
This paper is based on the results of a model-to-model comparison of six of the
best known active solar energy design programs. The comparisons were performed
in a June, 1989, workshop. The workshop participants were the modelers responsible
for ISFH (Germany), G! (Switzerland), TRNSYS and F-CHART (United States), MINSUN
(Sweden), and WATSUN (Canada). These countries also supplied the research funds.
PROGRAM DESCRIPTIONS
A brief description of the six programs spanning a wide- range of modelling
methodologies follows.
ISFH is an approach for the analysis, design and sizing of solar energy systems
based on daily energy input/output curves. The modelling approach may be
described as a mathematical compaction technique which links two stages of a
simulation process through a small number of variables and equations. Stage 1
Detailed Simulation: input/output relationships describing the desired
performance measures are derived using a detailed component based simulation
program and ranges of "typical" climate data and "typical" days of operation.
Stage 2 Daily Simulation: The I/O relationships thus obtained are applied to a
"real" system, climate, and load in order to predict "real" performance measures.
ISFH has been implemented for use on personal computers and requires only a few
minutes to calculate the annual performance of systems.
G8 is a personal computer based program for the analysis, design and sizing of
solar energy systems. The program requires daily sums of global and diffuse
radiation on the horizontal plane and daily average daywtime ambient
temperatures. Daily radiation absorbed is deduced for twelve representative days
of the year by accounting for collector geometries and optical characteristics.
Interpolation among these twelve days is used to determine the radiation absorbed
-------
1326
for the remaining days. The system specification is completed by specifying all
other component parameters and loads. The program calculates daily energy gained
by direct numerical integration of idealized radiation profiles, and predicts
monthly and annual performance.
TRNSYS is a well known modular transient system simulation program. It recognizes
a system description language in which the user specifies the components that
constitute the system and the manner in which they are connected. The TRNSYS
library includes many components. Annual simulations are performed on an hour-
by-hour basis using THY weather data. At each time step, the steps forming the
simulation are repeatedly called until the outputs from each component converge
within a user-specified tolerance.
F-CHART is an interactive solar energy system analysis and design program. The
program is based on correlation equations which model the results of a large
number of annual simulations using TRNSYS. The program requires minimal computer
resources.
MINSUN was originally developed for design studies on large seasonal storage
systems. The program uses hourly climate data in a detailed simulation model
developed from TRNSYS to calculate the daily collector array output and operating
time. Computer resource requirements are similar to TRNSYS.
WATSUN is an interactive hour-by-hour simulation program for analysis of a
variety of solar energy systems. The models were originally derived from TRNSYS.
It differs from TRNSYS in that: a) the system configurations are fixed and b)
at each time step, the energy calculations are performed without iterations.
Annual simulation on an IBM AT compatible machine takes a few minutes.
DESIGN OF EXPERIMENTS
Two applications -domestic hot water (DHW) and industrial process heat (IPH),
were used as the testing framework for comparing the six programs.
A base case was established for the IPH application. An evacuated tubular
collector, a parabolic trough collector oriented |3-W and N-S, four collector
areas, three climates, and two storage volumes were examined by exhaustively
enumerating all 72 combinations. Performance for each of the 72 combinations
was calculated by ISFH and F-CHART. Performance for the 24 combinations
involving the evacuated collector were calculated by G1. Four of the evacuated
collector combinations were run by WATSUN. The parabolic trough runs were not
made with G' and WATSUN because they had not implementad a tracking collector
capability. Though TRNSYS has a tracking collector and IPH capability, IPH
TRNSYS runs were not made because all the available workshop time was needed to
set up and run TRNSYS for the DHW system cases.
In contrast to the IPH cases, all computer programs were run for nearly all of
the DHW system cases. Therefore, we will focus on the DHW applications in this
paper. The reader is referred to the workshop report, Duff[1989], for the IPH
system results and additional details of the workshop.
The base case DHW system shown in Table 1 was formulated to eliminate as many
of the differences among models as was practical. For example, the collector
was horizontally mounted so as to use the horizontal radiation data directly and
therefore bypass the effects of different radiation processors.
A set of experiments was used to assess the impact on system performance of
changing values of system parameters one or two at a time. The various issues
-------
1327
TABLE 1 DHW Base Case Specifications
Parameter
Value
Parameter
Value
Collector Area
6 m»
Collector Tilt
Horizontal
F'Tt
.82
F'UL
3.8 W/m*-K
Collector Flow
.015 1/s-m*
Collector Cap.
0. KJ/m2-K
IAM
1
Collector Fluid
50/50 e. glycol
Pump Power
0
Controls
ideal
Pipe UA
0. KJ/m2-K
Pipe Cap.
0. KJ/m2-K
Pipe Dimensions
26.75 mm od
Pipe Length
lm/m' out x2
1.5 mm thick
6 m in x2
Heat Exchanger
none
Storage Volume
50 1/m2
Tank Model
Stratified
Aspect Ratio
3 to 1
Insulation Cond.
0.1 W/m-K
Insul. Thickness
0.1 i tanks
0.05 m pipes
Mains Temperature
10 °c
Draw
175 1 6-8am
Set Temperature
50 °C
175 1 5-10pm
Overheat Protection
100 °C
Tempering Valve
included
Auxiliary Tank Volume
175 1
Weather
Miami TMY
that the DHW experiments were designed to address were changing collector area,
differences in collector capacitance, different pipe capacitance, differences
in stratification models, collector flow rates, number of tanks, volume of
storage, climate differences, collector tilt, collector temperature dependence,
collector performance, draw timing, internal and external heat exchanger models,
incidence angle modifier effects, delivery set temperatures, constant inlet
temperature operation, and synthetic versus real weather data
The experiments chosen in this paper are given in Table 2. These were some of
the 28 DHW system experiments that produced more interesting or unexpected
results or addressed issues which were felt to be current or important. Other
experiments and results can be found in the workshop report.
RESULTS AMD DISCUSSION
Tank Stratification and Collector Flow Rate
Results from Experiments 1 and 3 can be used to compare the effect of fully
mixed and stratified tank models used in different programs. The programs
predicted a reduced performance for the fully mixed tank case.as expected, except
for G1 which only uses a stratified tank model. An average reduction of about
.2 GJ/year was observed for the other programs. Relative ranking of the programs
remained unchanged.
Results of Experiment 4 (low collector flow rate of 0.003 1/s-m1) were compared
to Experiment 3 (high flow rate of 0.015 1/s-m'). The results shown in Figures
1 and 2 indicate that, except for WATSUN and F-CHART, none of the other programs
predicted any significant change in performance. In the case of WATSUN, the
annual performance improved for the low flow case as one would expect. In the
case of F-CHART, the annual performance declined, which may be due to mixed tank
model used in the development of F-CHART method.
-------
1328
TABLE 2 DHW System Experiments
Exp. Description
1* Change the stratified tank model of experiment 3 to a fully mixed tank
model.
la1 Decrease the collector area of experiment 1 to 3 i!.
lb* Increase the collector area of experiment 1 to 9 m2 .
3 The base case as given in Table 1.
4 Run the same parameter values as experiment 3 with collector flow rate
changed to .003 l/s-ml.
6 Experiment 3 is run with a) Seattle and b) Albuquerque TMY weather data.
8 Run experiment 3 with a detailed collector model with F'Tt = 0.82 and F'Ut
= 3.50 + .012*Tflilj - -006*T M (/ where temperatures are in Celsius.
9 Run Experiment 3 with a)high and b)low performance collectors having F'T(
= 0.72 and 0.70 and F'UL = 1.70 and 7.00 W/m!-K, respectively.
15 Run experiment 1 with the following changes: Collector Capacitance of 10
KJ/m'-K as in experiment 2, Pipe UL as in experiment 2a, Pipe Capacitance
as in experiment 2a, Stratified Tank as in experiment 3, Collector flow
rate as in experiment 4, Climates: a) Seattle, b) Albuquerque, and c)
Miami, Collector Tilt equals latitude. Collector equations as in experiment
8, Heat Exchanger at 100 W/m!-K, IAM as in experiment 12
* G', has only a stratified tank model. Therefore, its results for experiments
1, la, and lb are for the stratified tank model.
SOLAR TO STORAGE AND AUXILIARY
EXPERIMENT 3
2.B -
a.s -
2.2
0.8 -
0.6
0.4 -
0.2
JAN FEB MAR APR MAY JUNE JULY AUG SEPT OCT NOV DEC HON AVE
«• FCHART X TRNSYS
~ G3 O ISFH a WATSUN V MINSUN
Fig. 1. Base case.
Collector Area and Climate
Results from Experiments 1, 1A, IB with 6, 3, and 9 ml collector area showed
consistent behaviour from all the six programs. In all cases, the monthly and
annual gains increased as the collector area was increased. The spread among the
different models was most pronounced for the 9 m1 case.
-------
1329
SOLAR TO STORAGE AND AUXILIARY
EXPERIMENT 4
JAN FEB MAR APR
JUNE JULY AUG SEPT OCT
DEC WON AVE
FCHART X TRNSYS
ISFH A WATSUN
Fig. 2. Base case with, low collector flow.
Results from Experiments 3, 6A, 6B, representing the base case simulated in
Miami, Seattle, and Albuquerque, respectively, show that all six models predicted
annual performances generally within 10% of each other. However, in Experiment
6B, as shown in Figure 3, monthly variations are quite significant among
different models. This is particularly noticeable for ISFH and G' which normally
predict results very close to each other.
SOLAR TO STORAGE AND AUXILIARY
EXPERIMENT 6B
MAY JUNE JULY AUG SEPT OCT
DEC MON AVE
FCHART
ISFH
X TPNSYS
WATSUN
Fig. 3. Base case at Albuquerque.
-------
1330
Collector Characteristics
Differences between the linear collector model used in Experiment 3 and the
nonlinear model used in Experiment 8 were not significant, mainly due to the DHW
system which does not develop very high operating temperatures.
Experiments 9A and 9B, using high and low quality collectors, respectively, show
that the spread in the annual (and monthly) performance prediction is quite large
for the low quality collector (20% points) compared to the high quality case
(within 5%). The spread is less between TRNSYS - WATSUN and ISFH - G', and may
be due to similarity of modelling methodologies.
Combined Effects
A typical DHW system was simulated in Experiment 15. It is evident that programs
based on similar methodology form groups, that is MINSUN and TRNSYS, and ISFH
and G1. WATSUN results were higher than that of any other program. This may be
due to differences in the way diffuse radiation components are calculated in
WATSUN. It is not possible to state which of the five programs predicted the
performance most accurately.
User-Friendliness and Improvements
The interactive programs (ISFH, G1, F-CHART, and WATSUN) were all very easy to
use and required very little learning time. TRNSYS and MINSUN required
considerable time for preparing inputs. Average execution time for the
interactive programs was at least an order of magnitude less than that required
by TRNSYS and MINSUN.
F-CHART, WATSUN, and TRNSYS are well documented. The documentation for TRNSYS
and WATSUN include details of the models used. ISFH and WATSUN have an on-line
help feature which is useful.
As a result of the workshop, many improvements were considered and, in some
cases, incorporated into the program either before, during, or after the
workshop. Since the actual developers of the programs were present, the changes
and improvements could be discussed and implemented quickly and efficiently. In
some cases, coding errors were detected and corrected when results from one
experiment were compared among different programs.
CONCLUSIONS
The accuracy and precision of ISFH, G*, and WATSUN were generally close to the
previously validated program TRNSYS. Compared to TRNSYS, ISFH, G', and WATSUN
were much more easily set up and ran rapidly. F-CHART, ISFH, and G1 required the
least time of all. TRNSYS and WATSUN results often grouped closely together.
The workshop itself proved to be very valuable to the developers of the programs
and resulted in numerous improvements to the programs.
REFERENCE
Duff, William S., Model Testing Workshop, June 25-30, 1989, Report of the
International Energy Agency Solar Heating and Cooling Program, Fort Collins,
Colorado USA, December 1989.
-------
1331
A PROCESS FOR MAKING SIMPLIFIED METHODS FOR
SYSTEMS WITH STORAGE
E. Mannik*, J. Atkinson* and G. Morrison**
*Dept. of Mech. Eng., Univ. of Syd., Australia
** School of Mech. and Ind. Eng., Univ. of N.S.W., Australia
ABSTRACT
This work describes a process for making simplified methods, to estimate the long-term performance of solar
energy systems, which largely overcomes many of the difficulties with existing techniques. The process is
outlined, and the results of using it to make a simplified method for the performance of a Solar-assisted Heat
Pump Hot Water System are described. A satisfactory simplified method (which computes the fraction of
'free energy', as a function of 5 system parameters and 4 climatic inputs, to an accuracy of better than 1%) is
produced without great difficulty or the use of excessive computer time.
KEYWORDS
Solar energy, solar energy systems, simplified methods, design tools, simulation
INTRODUCTION
Evaluating the economics of active solar energy systems requires the estimation of their long-term performance.
Detailed simulation packages such as TRNSYS (Klein and others, 1983) are one way to do this. They are very
flexible, allowing results to be obtained for a very wide range of systems, but their use requires significant
effort, expertise and computer time.
Simplified methods, which are formulae for monthly or annual performance, such as the well-known 0 f chart
(Klein and Beckman, 1979) are a much more convenient way of calculating long-term performance. They are
easy to use since little calculation or weather data are Tequired. Unfortunately, they are not available for all
system types, especially the more complex ones, be,cause individual simplified methods are both applicable to
very specific system configurations and difficult to make by existing techniques.
This work is concerned with making it easier to make simplified methods, so that they can be made available
for a wider range of system types without excessive effort. Before describing our approach we consider where
the difficulties lie in the existing techniques.
EXISTING TECHNIQUES FOR MAKING SIMPLIFIED METHODS
Simplified methods are rarely derived by purely analytic means because of the problems involved. Normally
system behaviour is too complex to be formulated in a way which is both simple enough for analytic solution
and which yields accurate results. These difficulties usually make it necessary to resort to the correlation of
long-term performance results generated by detailed simulation. Some authors (e.g. Huang, 1989) correlate
using large numbers of variables (system parameters and weather statistics), making it necessary to analyt-
ically derive the form of the performance expression, before the coefficients can be found numerically. More
commonly, the number of variables is reduced by combining them into 2 or 3 dimensionless groups. This makes
it much easier to find a suitable form of the performance expression, but identifying satisfactory dimensionless
groups may be difficult. Many more complex systems are not easily dealt with by a single correlation, making
it common for simplified methods to consist of a number of interconnected formulae. The method of Hobson
and Norton (1989), for instance, consists of 5 formulae involving 9 dimensionless groups. The utilizability
function, which estimates collector performance at fixed inlet temperature, is often a component of simplified
-------
1332
methods for more complex systems. How a simplified method can be divided into a number of 'pieces' is not
always obvious.
The difficulties likely with these techniques are :
(i) Significant analytic effort may be required to help choose dimensionless groups, terms appearing in the
performance expression, analytic simplifications etc.
(ii) Significant computer time will be required, whenever correlation techniques are used, to generate the large
number of long-term simulation results required. Some authors (e.g. Huang, 1989) have used compressed
data as input to the simulation to reduce the computer time required, often at some cost in accuracy.
(iii) The decisions, referred to in (i), involve elements of judgment and risk. Incorrect choices may result
in a simplified method which is only partially successful, or the need to alter or repeat sections of the
work. There are numerous simplified methods in the literature which are only partially successful, often
because they do not give accurate results across the full range of conditions.
A PROCESS FOR MAKING SIMPLIFIED METHODS
Systems without Storage
A process for making simplified methods more easily, which avoids or minimizes the problems described above,
is outlined in two sections. The first describes the process for systems without storage, and the second describes
how it is extended to deal with systems where storage is present. As is usual for making simplified methods,
a detailed simulation of the system is presumed available ; the task is to produce a formula which generates a
close approximation to the detailed simulation results, but uses orders of magnitude less calculation and data.
We begin by observing that long-term solar energy system performance, for systems with or without storage,
can invariably be expressed in terms of a small number of integrals over the time period for which performance
is required. For instance, for the Heat Pump Hot Water system considered below, the fraction of 'free energy'
is given by 1 — Ehp/El where Ehp = fom°* enp(T)dT and Ei = ei,(T)dT. Here enp(T) and e^{T)
are the rates, at time X, of electrical energy supply to the heat pump and energy supply to the load, and
performance is required for the period from time 0 to time Tmax. Producing formulae for long term performance
reduces to producing formulae for the relevant long-term integrals.
We also observe that if a system has no storage, then it also has no 'memory' of the past history of input values.
Any of the quantities of which we require a long-term integral, q(T), thus depends only on the current values
of weather (and possibly other) inputs, and on time-invariant system parameters, pn. Like any other
function, q(T) can, in principle, be expressed as a multi-dimensional polynomial in the variables on which it
is dependent, giving
i(T) = £ ajlh...kli3...p*>P*> ...h (Tf i2 cry*...
An expression for a required long-term integral can be obtained by integrating through the required time
period
g(T)dT = £ ... £m" h {T)h i2 (T?>... dT
The process generates formulae of this type for all the relevant long-term integrals. This makes it possible
to compute system performance given a chosen set of system parameters (pi,P2, ¦ ¦ ¦) and values of all the
input-dependent integrals ii (T)jl i? (T/2.. .dT which appear in the formulae. These input-dependent
integrals are the form of reduced data used by the simplified methods generated by this process. They can
either be reduced directly from detailed data, or estimated using statistical models of the inputs (not discussed
further due to lack of space).
An algorithm which generates multi-dimensional polynomial approximations for arbitrary, but reasonably well-
behaved functions is required to approximate each of the functions, q(T), of which an integral is required. As
well as producing approximations which are accurate and not too complex, the algorithm must meet these
additional criteria if an easier way to make simplified methods is to result :
• it must be able to deal with laxge numbers of variables. This will eliminate, or at least minimize, the need
to combine variables into dimensionless groups or to subdivide simplified methods into several 'pieces'.
-------
1333
• it must select the terms as well as determining the coefficients in the approximation. This is particularly
important when there are many variables as manual selection is difficult.
• it must make efficient use of computer time, requiring a minimum of time for the algorithm itself and
a minimum number of function evaluations. As the function evaluations are instantaneous values, not
long-term ones, a major reduction in computer time has already been achieved.
The following two-level algorithm, believed by the authors to be original, meets these criteria. The lower
level operates on a grid of function values, with a dimension corresponding to each variable. It generates
coefficients for multi-dimensional Chebysheff polynomial product terms (form Tni(* i)^r>2(®2) • ¦ • where the
Tni(xi) are Chebysheff polynomials and the set of X{ consists of all system parameters and inputs), one of
which corresponds to every function value in the grid. Though there may be many terms, coefficients are easily
generated for all of them, because of the efficiency of the method. It requires of the order of n In n operations
to find n coefficients, compared to the n3 required by least-squares regression. The next step is to remove
small terms, which are often numerous; terms are eliminated, beginning with the smallest, until a specified
'error increase' is reached. This achieves the automatic selection of terms. The final step in the 'lower part' of
the algorithm is to convert the approximation from Chebysheff form to the required power series form.
The lower level algorithm requires a very large grid of function evaluations if there are many variables, so we
incorporate it within an 'upper level' to improve the efficiency with which function evaluations are used. This
is based on noting that the overall function can be partitioned into subfunctions, there being a subfunction
dependent on every subset of the complete set of variables. The upper level of the algorithm assembles an
approximation to the overall function by adding together approximations to subfunctions, produced using
the lower level of the algorithm, until the required accuracy is achieved. Subfunctions are added beginning
with those dependent on the smallest number of variables and, provided that the function is reasonably well-
behaved, only low-order subfunctions will be required. The very large grid which would otherwise be required
is replaced by a number of small grids, resulting in far fewer function evaluations being required. In the
example discussed below, functions dependent on 11 variables were approximated using subfunctions involving
no more than 2 variables, and the number of function evaluations was reduced from ~ 10s to 831.
Using this algorithm is easy. Subfunctions, gridsizes and levels of 'additional error' must be selected as the
overall function approximation is built up, but the rules developed, used together with information generated
by the software, make these decisions almost automatic. There are few problems, in fact, in using the whole
process; the only other decisions to be made are whether any rearrangements of the system parameters or
performance output quantities might produce shorter formulae, or allow them to be obtained more easily.
Systems with Storage
The process for systems without storage assumes that system behaviour is dependent only on current input
values and not on their past history. More specifically, it assumes that the functions, q(T), of which integrals
are required, are dependent on input values at time T, but not on earlier values. This is not true of systems
with storage,and if the process is to be applied to them it must be modified so that the q(T) are dependent
on input values preceding time T, as well as those at T.
This could be arranged by explicitly making the q(T) dependent on earlier input values; for instance, in the
case of input im dependent on im(T — 6),..., im(T — k6) as well as on im(T). This is not practical, however,
because it makes the q(T) dependent on too many variables.
We choose instead to use a model for the behaviour of each input, preceding time X, as very few variables will be
required to describe each input. Each model describes an input as a function of one or more parameters, which
can be adjusted to make the input model approximate, given input behaviour. More explicitly, the parameters
are set so that the modelled input matches given values of certain input statistics. These input statistics
are used as variables in the approximation of the q(T). For instance, in the example discussed below we use
a simple 2-parameter model of radiation. The days preceding time T are all presumed identical, the daily
profile being obtained by taking a standard 'shape' and adjusting the total daily radiation and the daylength
(Fig 1). The adjustment forces the modelled input to take given values of the 'weighted average' statistics
/0T Jy (T — t)u(t)dt and JQT 7|.(T — t)u(t)dt. Here It is the collector plane radiation and u(t) is a weighting
function which falls with increasing t, reflecting the falling influence of input values on current behaviour as
we move further back in time. Whenever this radiation model is used, the functions q(T) are presumed to be
dependent on these two input statistics.
-------
1334
Original radiation 'shape', with possible profiles resulting from 'adjustment'
As solar energy systems tend to involve only a few different inputs (usually solar radiation, ambient temper-
ature, humidity and windspeed), it is feasible to develop a library of models so that the user will not usually
need to develop his own. This will need to contain several models of each input, to allow different tradeoffs
between accuracy and the number of additional input statistics, depending on the requirements of different
system models.
We now consider how the process is altered to adapt it to systems with storage. The changes are:
(i) Each of the quantities of which integrals are required, q(T), is now approximated by a multi-dimensional
polynomial in system parameters, and the input statistics used by the chosen input models. They are
no longer a function of instantaneous input values.
(ii) The calculation of function values now involves the use of the chosen input models. They are used to
generate the daily profiles corresponding to the given input statistic values, these daily profiles then
being used repeatedly as inputs to the detailed simulation until the behaviour 'converges' to a repetitive
cycle.
The adaption of the process to systems with storage would now be complete, if the choice of appropriate input
models for each system was obvious. As this is not necessarily the case, and as it is much too time-consuming
to experiment by creating formulae for different choices of input model, some other way of choosing input
models is needed. Our approach is to introduce a preliminary 'first stage' to the process, where input model
choices are made and confirmed, before creating the formulae in the 'second stage' (the existing process).
The first stage is not described in detail, due to lack of space, but involves testing sets of input models until a
satisfactory set is found. The testing is carried out by using the input models to generate sets of input profiles
(representative days), which sample the test data set; these are used as input to the simulation and estimates
of performance quantities obtained. The errors in these estimates are an indication of the errors which would
be introduced into the simplified method by this choice of input models. This testing technique can be used
as a method in its own right and is somewhat similar to other 'representative day' methods such as that of
Reddy, Gordon and de Silva, (1987).
APPLYING THE PROCESS
System Model
While parts of the process have been successfully applied to various system models, the only realistic model
to which the whole process has been applied is the TRNSYS simulation of a Heat Pump Hot Water System
described in Morrison (1990) . The system (Fig. 2) consists of a heat pump, an uncovered collector acting
as the evaporator, a storage tank into which energy is transferred through a wrap-around heat exchanger,
and a tempering valve to allow water to be delivered at a specified temperature. The total daily load varies
seasonally with mains water temperature. System performance is dependent on four inputs : solar radiation,
dry bulb temperature, wet bulb temperature and mains water temperature.
We require a formula for the annual fraction of 'free energy'. An 'error' (discrepancy from detailed simulation)
of about 1% is desireable, as this will not significantly increase overall errors beyond the 1 or 2 percent already
present in the detailed simulation. We also require a formula for the fraction of the annual load which is met,
to somewhat lower accuracy, so that system parameter choices which do not result in the system meeting most
of the load can be avoided. The model incorporates a detailed description of a particular heat pump, which
we will not alter, but we will vary other major system component sizes, including load, as well as the collector
-------
1335
TEMPERING
VALVE
-c2f1"7L
Draw OIF
STORAGE
TANK
HEAT
PUMP
Mains Water
Figure 2: Heat Pump Hot Water System
Table 1: First Stage Results, r.m.s. of errors for 8 testcases
Fraction of
Fraction of
Free Energy
Load Met
1 Representative Day
0.84%
1.7%
3 Representative Days
0.52%
1.7%
loop controller setpoint. The simplified method will then be a function of 5 system parameters : collector
area, storage size, heat exchanger capacity, load scaling and controller setpoint.
Applying the First Stage
The initial model choices were the 2-parameter model described above for radiation and a constant value for
the other three inputs (dry bulb, wet bulb and mains water temperatures). This constant is set to the local
weighted average of the appropriate input, f0 im(T — t)u>(t)dt. The inputs are thus described by a total of 5
input statistics.
We initially tried small numbers of representative days (1 and 3), generating them for typical Sydney data
and using them to produce results for 8 testcases. Results are listed in Table 1 and as r.m.s. errors are below
1%, the adequacy of the chosen models is confirmed. These results were easily obtained, apart from some
difficulties with the 'convergence' of the representative day calculations. The results quoted required 5 to 7
days of detailed simulation for each representative day result, allowing results to be obtained about 50 - 70
times faster (using one representative day) than with the full detailed simulation.
Applying the Second Stage
Before applying the approximation algorithm, we rearrange the performance quantities to simplify approxi-
mation. The fraction of 'free energy' is given by 1 — Ejjp/E£, so in theory we must approximate the integrals
Ehp and El- In practice, examining first stage results soon shows that a single integral, of the 'daily fraction
of free energy', (which can be obtained from each function evaluation) is a better option. This quantity varies
less than Ehp and El and is linear enough in the inputs for negligible errors to result from this simplification.
The fraction of the load met can be formulated as £/,/(total load), a formula for El only being required as
Table 2: Second Stage Results
No. of Terms
No. of Data Items
r.m.s. Error
Fraction of Free Energy
formula 1
457
12
0.77%
formula 2
170
4
1.57%
Fraction of Load Met
75
1
3.0%
Creating formulae required 831 function evaluations.
Computer time used was equivalent to: 20 annual detailed simulations
13 hours on 3 Mflop machine
-------
1336
(total load) is a known function of average mains water temperature. The actual quantity approximated was
El/{\o3.A scaling), to improve the distribution of errors. No rearrangement of the existing system parameters
was necessary. However, the standard method for dealing with the input statistics treats the annual average
dry bulb temperature as an additional 'parameter', so that other temperatures can be considered to fluctuate
about it and therefore take a much smaller range of possible values.
The approximation algorithm was applied without difficulty to the two quantities identified above, both of
which are functions of 11 variables (6 parameters, including the extra one, and 5 input statistics). Results are
given in Table 2, and show that the formulae produced are satisfactory and were obtained using reasonable
amounts of computer time. The two results quoted for the fraction of free energy use the same function
evaluations, but different choices of 'additional error'. Note the rapid increase in the amount of calculation
with increasing accuracy.
DISCUSSION AND CONCLUSIONS
We now consider, to the extent that the limited experience with the process makes this possible, whether it
is an easier way of making simplified methods. For this to be the case, the process must avoid the problems
with existing techniques, as well as meeting other criteria.
Most importantly, the methods produced must be satisfactory. This has been achieved with the example
described. A deviation from detailed simulation of 1% is quite adequate. The formulae are somewhat complex
for a simplified method, but as they can produce an annual result in one step, they will still require less
calculation and data than many existing methods using monthly formulae.
A further requirement is that the process must produce satisfactory simplified methods for a wide range of
system types. It is impossible to shW this conclusively without making a large number of simplified methods,
but we note that the process is very general and makes no assumptions about the system which would limit
its applicability. The factor limiting the applicability of the process is expected to be the complexity of
the simplified method required; if behaviour is dependent on too many system parameters and inputs, or is
dependent on them in too complex a way, then the computer time required to generate the formulae will
become excessive.
A further requirement is that the process require less effort for analysis and decision-making than the usual
techniques. We have attempted to ensure this by making the process as routine as possible. In the case
of the example, this appears to have been successful; the whole process has been routine, apart from the
minor rearrangements of the performance quantities in the second stage and some convergence problems. It
is believed that the finding of dimensionless groups etc. required for conventional techniques would have been
much more difficult.
A further requirement is that computer time be minimized. The process has been designed to ensure this.
Computer time, for everything other than function evaluations, first stage results (limited requirement) and
detailed simulations for checking accuracy is negligible. The approximation algorithm uses function evaluations
efficiently, and these, in any case, require much less computation than long-term results. In the case of the
example, overall computer time cannot be reduced much further, because the time required to create the
formulae (equivalent to about 22 annual simulations) is comparable to that required to produce sufficient
simulation results for checking.
The final requirement is that the process should rely as little as possible on the user making correct judgments.
As we have succeeded in making the process largely routine and removed much of the need to make decisions,
we believe that this has been achieved.
The process appears, on current evidence, to be an easier way to make simplified methods. Further experience
is needed to confirm this and develop some aspects of the process (such as the library of input models) further.
REFERENCES
Hobson, P.A. and Norton B. (1989). Solar Energy, 43, 85-95.
Huang, B.J. (1989). ASME J. of Solar Energy Engineering, 111, 124-131.
Klein, S.A. and Beckman, W.A. (1979). Solar Energy, 22, 269-282.
Klein, S.A. and others (1983). TRNSYS 12.1 User's Manual. Univ.of Wisconsin Solar Energy Laboratory.
Morrison, G.L. (1990). Extensions for the TRNAUS Simulation Program. Report No. 1990/FMT/l, School
of Mechanical and Industrial Engineering, U.N.S.W.
Reddy, T.A., Gordon, J.M. and de Silva I.P.D. (1987). Solar Energy, 39, 123-133.
-------
2.6 Solar Domestic Hot Water II
-------
-------
1339
COMPARISON BETWEEN EXPERIMENTAL AND SIMULATED DHW SHORT-TERM TEST
RESULTS
P. J. Schaefer, W. A. Beckman and S.A. Klein
Solar Energy Lab., University of Wisconsin-Madison
ABSTRACT
Short-term experimental testing of solar domestic hot water (SDHW) systems is a critical part of
the Solar Rating and Certification Corporation's (SRCC) current method for system certification.
Even short-term testing, however, is time consuming and expensive. Replacing some of the
experimental tests with computer simulations is one way to avoid these disadvantages.
Comparisons between short-term experimental and simulated test results for one well-instrumented
SDHW system are presented.
KEYWORDS
TRNSYS; solar domestic hot water, simulation; performance sensitivity; performance rating
EXPERIMENTS
The experimental tests were conducted in accordance to the ASHRAE-95/SRCC guidelines by
FL2
\ '
FU
TC10
TC4
Boiler
TC11
TCS
T x«>
TC15
TC12 TC13i i
TC2
Tank
Side
Pump
TC8
)(«./
Drain-Back Tank
Containing
Submerged Heat
Exchanger
TC7
Collector
Side
Pump
TC
Auxiliary
Storage
Tank
Solar
Storage
Tank
Fig.l. SDHW System
Colorado State University (CSU), Fort Collins, Colorado. A schematic of the SDHW system
under investigation is shown in Figure 1. Water is the working fluid throughout the system. An
receding page blank
-------
1340
electric boiler, rather than a solar collector, is used to transfer energy into the system. The
radiation and incidence angle profiles are as specified by SRCC. The mains water, tank ambient,
and collector ambient temperatures are all maintained at 22 C while the set temperature is 54 C. A
load draw occurs at 8:00 A.M., 12:00 noon, and 5:00 P.M. each day. A draw of 0.20 kg/s
continues until 16603 kJ are removed. The system is operated until the daily-integrated auxiliary
energy input between successive days agree to within 3%, or four days have elapsed. A total of 16
experimental tests were performed as indicated in Table 1. Additional information is availabel in
Schaefer (1991).
SIMULATION MODEL
The computer program TRNSYS 13.1 (Klein and others, 1990). was used to perform the
computer simulations. Standard models found in the TRNSYS library were used to replicate the
SDHW system, with two exceptions. The TRNSYS pump model was modified such that 85% of
the pump work acts to raise the fluid temperature. The CSU experimental pump power for both
the collector side and tank side loops are nearly constant, regardless of the flow rate. The second
non-standard TRNSYS subroutine is the load-flow on-off controller. The controller subroutine
turns the flow on at the time step closest to the specified time (8:00,12:00, or 17:00) and turns the
flow off at the time step in which the energy draw is closest to 16603 kJ, to match the experimental
conditions.
TABLE 1. Experimental Test Summary
Test
Coll
Tank
Area
Tank
Tank
#
Flow
Flow
Vol
Design
(kg/hr)
(kg/hr)
(m2)
(m3)
1
205.2
169.2
2.78
0.223
Basic
2
410.4
169.2
2.78
0.223
Manifold
3
205.2
342.0
2.78
0.223
Manifold
4
410.4
342.0
2.78
0.223
Basic
5
205.2
169.2
5.56
0.223
Manifold
6
410.4
169.2
5.56
0.223
Basic
7
205.2
342.0
5.56
0.223
Basic
8
410.4
342.0
5.56
0.223
Manifold
9
205.2
169.2
2.78
0.272
Manifold
10
410.4
169.2
2.78
0.272
Basic
11
205.2
342.0
2.78
0.272
Basic
12
410.4
342.0
2.78
0.272
Manifold
13
205.2
169.2
5.56
0.272
Basic
14
410.4
169.2
5.56
0.272
Manifold
15
205.2
342.0
5.56
0.272
Manifold
16
410.4
342.0
5.56
0.272
Basic
The CSU tests use an electric heater instead of a flat-plate collector to transfer energy into the
system. The Hottel-Whillier equation with FR(ta)n = 0.602 and FrUl = 5.56 W/m2-C and b0 =
0.42 is the governing equation controlling the amount of energy delivered into the system for both
the experimental and simulation cases. The collector area was either 2.78 or 5.56 m2. The turn on
and turn off dead band temperatures are 11.11 and 2.78 C respectively.
The experimental tests use either a manufacturer rated "80 gallon" or "65 gallon" solar storage
tank. The measured tank volumes are actually be 71.9 gallons (272 liters) and 58.9 gallons (223
liters). Both tanks are identical except in diameter. CSU determined the tank loss coefficients by
cool down tests to be 3.74 W/C and 3.41 W/C for the large and small tanks, respectively. All of
the experimental tests use a manufacturer rated "44 gallon" auxiliary storage tank. The actual
volume of the auxiliary tank was not measured and was assumed to have a tank volume 10% less
than the rated volume or 143 liter. The tank heat loss coefficient was determined to be 1.9 W/C
(1.13 W/m2-C) by measuring the auxiliary energy necessary to maintain the tank at constant
-------
1341
temperature. This tank is modeled as being fully mixed.
A major simulation question concerns what type of tank model to use. Kleinbach (1990)
investigated test data for the eight non-diffused tests and concluded that a three node tank model
performed the best in comparison to the other tank models investigated. Sufficient data to conduct
similar comparisons on the eight diffused tanks are not available. Simulations were performed
with 5, 10 and 15 node solar tank models. Increasing the number of nodes beyond ten results in
small simulation differences. A 10 node, tank model was used for tests with the diffuser.
The experimental tests use a manufacturer rated "8 gallon" drain-back tank. The tank's outer
height is measured to be 0.6 m. No other information about the tank is available. It is again
assumed the tank's actual volume is 10% less than the rated volume, and the inner height is 2
inches less than the outer height due to the tank's wall and insulation thickness. The tank volume
and height used for simulation purposes are thus 27 liter and 0.55 m, respectively. The tank's loss
coefficient was assumed to be 0.67 W/C.
The TRNSYS simulations only considered the pipe losses between the collector and drain-back
tank. Piping in the system other than that of the collector loop is relatively small. The pipe's inner
diameter is approximately 3/4" (19 mm). The pipe lengths leading to and from the collector are
11.77 and 13.41 m, respectively. The pipe insulation is rated at 4.7 hr-ft-°F/Btu. The simulations
used a slightly greater overall resistance of 5.0 hr-ft-°F/Btu to account for convection.
A submerged heat exchanger coil is located within the CSU drain-back tank and is modeled as a
constant effectiveness exchanger located outside of the tank. The heat exchanger effectiveness for
the various flow rate combinations was calculated from the experimental data to be 0.44 for test #7.
The effectiveness ranged from a low of 0-38 for test #11 to a high of 0.58 for test #6. The
variation in the effectiveness from test to test is not explained by differences in flow conditions.
For example, tests 7 and 11 both have the same collector and tank flowrates, but the experimental
heat exchanger effectiveness factors were 0.44 and 0.38, respectively. This degree of uncertainty
in the heat exchanger effectiveness is of minor concern, based on the sensitivity results presented
in a following section.
COMPARISON OF RESULTS
Even well-controlled experiments will not achieve steady-periodic conditions in the 4 day test
period. Hence, a bias exists when test results are compared to simulations that are necessarily
steady periodic. The change in internal energy of the solar tank over the last day, AU, was
estimated from the thermocouple tree in the solar tank and found to range from -300 to +700 kJ,
depending upon the test. Addition of AU to the measured delivered solar energy results in a more
accurate indication of the energy delivered to the storage tank from the solar collectors. The
following comparisons between the experimental and simulation results for the eight tests with the
basic solar tank take into account the energy storage by addition of AU to the delivered energy.
Sufficient data to determine AU for the eight experiments with the diffuser in the solar tank were
not available and hence, these data could not be corrected.
Five daily integrated energy flows are of interest:
Qu = Energy gain across the collector (boiler)
Qpar = Sum of the two pump works
Qs = Energy delivered from the solar tank to the auxiliary tank
Qanx = Auxiliary energy supplied to the auxiliary tank
Qua = Total energy, solar plus auxiliary, delivered from the system.
Figure 2 shows calculated and experimental energy quantities for the tests with the conventional
solar tank. Figure 2a compares the important energy flows and solar fraction for arbitrarily
selected test #11. Figures 2b, 2c and 2d compare experimental and simulated values of Qs, Q».,t
-------
1342
and solar fraction for the eight tests. Figure 3 presents results similar to those in Figure 2 except
that they are for the solar tank with the inlet diffuser.
The total delivered energy, Qjei, is specified by the testing standards to be 49809 kJ/day. Both the
experimental and simulated tests deliver the correct total delivered energy. The pump work, Qpar'
is typically small for a SDHW system. The CSU system under investigation is atypical in that the
pump work is substantial, being approximately 10% of the total delivered energy. However, Qpar
for the system is merely an indication of pump-on time since the power required by both pumps
during operation is nearly constant. The power input to the pumps used in the simulation was
determined from the experiments. In the simulations Qpar was calculated as the measured pump
power times the simulated pump on time.
CSU estimated two sigma errors for all energy quantities. The estimated Qu experimental error is
approximately ±6% of the measured value. The Qu simulation results all fall within this
experimental error tolerances except for test #8 . For test #8, the measured value of Qu was 30047
±1736 kJ while the simulated value was 32110 kJ, which is within a three sigma error bound.
Experimental Results
TRNSYS Results
Experimental Results
TRNSYS Results
Qu Qpar Qs Qaux Qdel
#1 #4 #6 #7 #10 #11 #13 #16
Experimental Results
TRNSYS Results
Experimental Results
TRNSYS Results
40
#1 #4 #6 #7 #10 #11 #13 #16
#1 #4 #6 #7 #10 #11 #13 #16
Fig 2. Experimental and simulated results for the basic solar tank.
The Qs and Qaux estimated experimental error tolerances are much smaller than those of Qu. The
estimated experimental measurement errors on Qs and Qaux are typically ±1% and ±0.5% of the
measured values, respectively. The Qs and Qaux simulation results do not fall within these very
narrow experimental error bounds. From the perspective of an energy balance on the auxiliary
tank it is clear that if Qs is, for example, too high then Oa,,T will be too low. The major reason that
the simulated delivered solar energy Qs (and consequently QaUx) is outside the reported error
-------
1343
bounds is that the energy input to the tank in the experiments did not exactly agree with the energy
input calculated in the simulation. The energy delivered to the solar tank is the sum of Qu and a
fixed fraction of Qpar minus the pipe losses. In the experiments, Qpar and the pipe losses are small
so that the uncertainty in the energy into the solar tank is approximately the same as the uncertainty
in Qu. CSU estimated the two sigma limits for Qu to be between ±1.0 and ±1.7 MJ. The
simulations produced values of Qu that were all within -0.1 to +2.0 MJ of the nominal
experimental values. In other words, whenever the energy supplied to the solar tank in the
experiments is 5% higher than expected, Qs will also be 5% too high in comparisbn with
simulation results, although the measured value of Qs will be experimentally known to ±1%. All
of the differences between the simulated and experimental values of Qs could have been caused by
uncertainty in the experimental value of Qu.
The solar fractions were calculated from sf = (Qs - Qpar)/Qdel- Since Qdel and Qpar are nearly the
same in the simulations and the experiments, only differences in Q$ affect the solar fraction.
Experimental Results
TRNSYS Results
Experimental Results
TRNSYS Results
Qu Qpar Qs Qaux Qdel
Experimental Results
TRNSYS Results
#2 #3 #5 #8 #9 #12 #14 #15
Exoenmental Results
Results
S 30
#3 #5 #8 #9 #12 #14 #15 #2 #3 #5 #8 #9 #12 #14 #15
Fig. 3. Experimental and simulated results for the solar tank with diffuser.
SIMULATION SENSITIVITY ANALYSIS
A sensitivity analysis were carried out to determine the importance of accurately knowing the
values of the system parameters used in the simulations. Simulations were performed with the 12
system parameters listed below being reduced one at a time by 10% of their nominal value.
1. Solar tank's heat loss coefficient per unit area (U^^)
2. Solar tank's volume (Vol^j^)
3. Heat exchanger effectiveness (Hx eff)
-------
1344
4. Collector gain coefficient at normal irradiance (Fr(ta)n)
5. Collector loss coefficient per unit area (F,Ul)
6. Collector area (Area)
7. Collector loop flow rate (Coll How)
8. Tank loop flow rate (Tank Row)
9. Percentage of pump work which acts to raise the fluid's temperature (Pump %)
10. Drain-back tank heat loss coefficient per unit area (U d-b)
11. Auxiliary tank heat loss coefficient per unit area (U aux)
12.Auxiliary tank volume (Vol aux)
Sensitivity results are reported in Figure 4 in terms of the fractional change of Qs with respect to a
fractional change to a variable for the conditions of test #6 and #11.
a
2
«
C
a
>
x
Sf
s
ts
¦c
!8
>
ro
Test #11
Test #6
s «
Fig. 4. Sensitivity analysis for tests 6 and 11.
Figure 4 shows the importance of accurately knowing collector area and the collector parameters
Fr(t
-------
1345
IMPACT OF COMPONENT SELECTION ON SRCC RATING OF DRAIN-BACK
SOLAR WATER HEATERS
J.H. Davidson, W.T. Carlson, and W.S. Duff
Solar Energy Applications Laboratory
Colorado State University
Fort Collins, Colorado 80524
ABSTRACT
Effects of changes in collector area, collector flow rate, recirculation flow rate, tank volume and tank
design on SRCC ratings of generic drain-back solar water heaters are experimentally measured.
Results indicate that only collector area, collector flow rate and tank stratification significantly affect
ratings.
KEYWORDS
Solar; hot water heating; drain-back.
INTRODUCTION
Effects of variations in collector area and flow rate, recirculation flow rate, solar storage tank volume
and design on the Solar Rating and Certification Corporation (SRCC) rating of a generic drain-back
solar water heater are investigated in a two-level, half-factorial experimental design. The levels of
each design or operating factor, shown in Table 1, are based on current industry standards.
Table 1 High/Low Levels for Each Factor
Factor Description
High Level
Low Level
Collector Flow Rate
.114 1/s
.057 1/s
Recirculation Flow Rate
.095 1/s
.047 1/s
Collector Area
5.56 nfi
2.78 m2
Storage Tank Volume
3101
2501
Storage Tank Design
Basic Drop Tube
Stratification Manifold
OVERVIEW OF SRCC RATING PROCEDURE
The SRCC rating procedure is a 4 day test in which environmental and load parameters are specified
(SRCC, 1984a,b). Three modifications to the standard testing procedure are made. Daily water
draw energy, Qdel. is increased to 49.8 MJ to conform with the Federal Trade Commission (FTC)
rating of electric and gas water heaters (FTC, 1989). Calculations to determine heater input
(ASHRAE, 1987) are modified to permit rating with a collector array different from that installed in
the test facility (Carlson, 1990). Since neither the bypass loop specified by SRCC nor the installed
solar controller can be used with a drain-back system, a dead-band controller is emulated with the
computer.
As shown in Fig. 1, SRCC energy based rating quantities include: useful collected energy, Qu; daily
hot water energy delivered by the solar storage tank, Qs; and net energy delivered from the solar
-------
1346
storage tank, Qnet, equal to Qs minus parasitic energy Qpar- Rating is completed in 4 days or when
the daily auxiliary energy input, Qaux. is within 3% of the previous day's value. At the end of the
test a continuous draw is made on the solar storage tank to determine the reserve capacity, Ores.
Collector Array
Auxiliary
Solar
Heater
0
s
System
and
—t ^
i Storage\
Q Q QDar Q
aux loss.aux M
-------
1347
Table 3 Drain Back Specifications
Component
Specification
Value
Collector
Total Area
(large)
5.56 m2
(small
2.78 m2
Frxa
0.602
FrUi
5.55 W/m2/K
K,>,>*^-12.7mm bolt
PVC 80 mm i.d.
Coupling
Two rows of holes
9.5 mm in diameter.
14hctes around"
circumference.
Circular Plate with open
cross sectional area of
1600 mmA2
150 mm
Fig. 3. Cross section of stratification
manifold
-------
1348
RESULTS
Ratings for the 16 trials are listed in Table 4. Experimental errors and random errors (determined
from statistical analyses) based on two standard errors are listed in Table 5. Since the error in Qu
affects other energy quantities, it should be added to the errors in Qdel. Qs. and Qaux to assess tetaL
experimental error.
Table 4 Summary of Experimental Ratings
Collec-
Recir-
tor
culation
Collec-
Tank
Tank
Trial
Flow
Flow
tor Area
Volume
Design
Qdel
Qs
Qnet
Ores
Qaux
Qu
Rate
Rate
(m2)
(0
(W)
(kJ)
(kJ)
(kJ)
(kJ)
(U)
(1/s)
(1/s)
1
.057
.047
2.78
250
Basic
49820
16647
11212
12309
39329
17554
2
.114
.047
2.78
250
Manifold
49815
19545
12981
14881
35480
18196
3
.057
.095
2.78
250
Manifold
49811
17975
11946
12772
37107
18254
4
.114
.095
2.78
250
Basic
49821
16996
11464
13157
38182
18401
5
.057
.047
5.56
250
Manifold
49811
26382
20331
18712
29312
29369
6
.114
.047
5.56
250
Basic
49815
25471
20008
18752
29529
30681
7
.057
.095
5.56
250
Basic
49829
24260
18890
19577
31449
29794
8
.114
.095
5.56
250
Manifold
49811
27273
21072
19299
28370
30047
9
.057
.047
2.78
310
Manifold
49811
18459
12394
14334
37265
18223
10
.114
.047
2.78
310
Basic
49822
17717
11993
14784
37753
18083
11
.057
.095
2.78
310
Basic
49816
16968
11312
14999
38675
18088
12
.114
.095
2.78
310
Manifold
49822
18482
12255
14808
36684
19084
13
.057
.047
5.56
310
Basic
49811
25113
19706
20402
30635
29796
14
.114
.047
5.56
310
Manifold
49815
27686
21474
21998
27642
31304
15
.057
.095
5.56
310
Manifold
49815
26956
20877
21686
28681
30125
16
.114
.095
5.56
310
Basic
49822
26712
21228
23378
29013
31746
Table 5 Experimental Error Based on Two Standard Errors
Collec-
Recir-
tor
culation
Collec-
Tank
Tank
Flow
Flow
tor Area
Volume
Design
Qdel
Qs
Qnet
Ores
Qaux
Qu
Trial
Rate
Rate
(m2)
0)
(kJ)
(kJ)
(kJ)
(kJ)
(kJ)
(kJ)
(1/s)
(1/s)
1
.057
.047
2.78
250
Basic
393
173
185
123
216
1103
2
.114
.047
2.78
250
Manifold
393
192
207
145
190
1103
3
.057
.095
2.78
250
Manifold
393
181
195
121
198
1113
4
.114
.095
2.78
250
Basic
393
174
186
129
207
1031
5
.057
.047
5.56
250
Manifold
' 393
239
249
165
157
1756
6
.114
.047
5.56
250
Basic
393
231
240
167
162
1657
7
.057
.095
5.56
250
Basic
393
223
232
177
171
1647
8
.114
.095
5.56
250
Manifold
393
245
255
168
153
1736
9
.057
.047
2.78
310
Manifold
391
184
198
179
199
1094
10
.114
.047
2.78
310
Basic
392
179
191
146
204
1051
11
.057
.095
2.78
310
Basic
392
173
186
152
207
1069
12
.114
.095
2.78
310
Manifold
392
184
198
142
193
1118
13
.057
.047
5.56
310
Basic
393
230
239
183
165
1658
14
.114
.047
5.56
310
Manifold
393
247
257
193
148
1744
15
.057
.095
5.56
310
Manifold
392
242
253
191
153
1754
16
.114
.095
5.56
310
Basic
393
241
249
208
158
1630
Random
966
880
1668
898
1028
Error
-------
1349
Standard normal quantile plots of the effects on each rating are shown in Figs. 4-8. Corresponding
analysis of variance (ANOVA) models which include statistically significant effects are given in
Eqns. 1-5. The first term of the model is the mean of the 16 trials. The coefficients for the factor
levels, L, are one half the effect due to the factor. The factor levels are ±1 representing the high and
low levels.
Qu = 24297 + 6061 ^-Collector area ^96 Lc0Hector flow rate ^ (1)
Qs = 22040 + 4192 I-Collector area ~ ^05 Ljank design + ^45 Lcollector flow rate ^ (2)
Qnet = 16196 + 4252 Lcollector area - 470 design+ 363Io0iiector flow rate kJ (3)
Qaux = 33444 - 4116 Lcollector area + 877 I-Tank design " 613 Lcollector flow rate (4)
Ores = 6470 + 3235 Lcollector area " 1058 I-Tank Volume kJ
Figures 4-8 show that changes in collector area significantly effects all energy ratings, and that
collector flow rate affects all ratings except reserve energy. Not surprisingly, collector area has the
largest impact. Figures 5-7 and Eqns. (2)-(4) show that tank stratification impacts Qs> Qnet and
Qaux- Since energy quality is not considered in Ores. Fig- 8 and Eqn. (5) indicate that tank volume,
not tank stratification, affects this rating.
CONCLUSIONS
Of the five design and operating factors examined, only the effects of changes in collector area and
flow rate and the use of a stratification manifold are statistically significant. Tank volume is
statistically significant only when determining reserve energy, Qres- The effects due to the other
factors and factor interactions are obscured by experimental error. As expected, collector area has
the largest effect on SRCC rating. Reducing the collector area from 5.56 to 2.78 reduces Qu by
38%, Qs, by 31 %, and Qres by 28%. Auxiliary thermal energy, Qaux, is increased by 28% .
Tank stratification is the second most significant factor on rating. Use of a stratification manifold
increases Qs by 6% and Qnet by over 4%. Qaux is decreased by 6%.
Halving collector flow rate from 0.114 to 0.57 1/s reduces Qu, Qs and Qres by approximately 3%.
Since flow rates have minimal effect on performance, pump selection is critical. This is evident
from the fact that parasitic energy represents 20% of solar output in the baseline system and even
higher percentages with less collector area.
REFERENCES
American Society of Heating, Refrigerating and Air-Conditioning Engineers, Inc.. (1987). Methods
of Testing to Determine the Thermal Performance of Solar Domestic Water Heating Systems.
Standard 95-1987. Atlanta, Georgia.
Carlson, W.T. (1990). Comparison of Experimental and TRNSYS SRCC Ratings of a Generic
Drain Back Solar Water System. Masters Thesis, Colorado State University.
Federal Trade Commission (1989). Uniform Test Method for Measuring the Energy Consumption of
Water Heaters. Code of Federal Regulations, 10 CFR Ch II, Part 430, Subpart B, Appendix
E.
Solar Rating and Certification Corporation, Washington, D.C. (1984a). Operating Guidelines for
Certifying Solar Water Heating Systems. Document OG-200. (November).
Solar Rating and Certification Corporation, Washington, D.C. (1984b). Test Methods and
Minimum Standards for Certifying Solar Water Heating Systems. Standard 200-82.
(November).
-------
1350
150001
10000*
Collector area
5000
£
Collector flow rate
-5000-
-2-10 1 ;
Standard Normal Quantile
Fig. 4. Standard normal quantile plot of
effects on Qu-
10000
8000
6000
4000-
2000 -
0
-2000
Collector area
Collector flow rate
Tank design
-2-10 1 2
Standard Normal Quantile
Fig. 5. Standard normal quantile plot of
effects on Q$.
£
10000
8000
6000-
4000 -
2000 -
o-
-2000
Collector area
Collector flow rate
B Tank design
~i—
-1 0 1
Standard Normal Quantile
2000"
o-
-2000-
-4000"
-6000 *
-8000"
-10000"
Tank design
D Collector flow rate
0 Collector area
—i »
-10 12
Standard Normal Quantile
Fig. 7. Standard normal quantile plot of
effects on Qaux-
8000"
6000"
4000*
as 2000 ¦
ui
-2000"
Collector area
Tank volume
-10 12
Standard Normal Quantile
Fig. 8. Standard normal quantile plot of
effects on Ores-
Fig. 6. Standard normal quantile plot of
effects on Qnet-
-------
1351
INTEGRATED COLLECTOR STORAGE DHW SYSTEM
NUMERICAL SIMULATION OF HEAT TRANSFER AND FLUID FLOW
L.A.M. Ramaekers*, C.J. van der Leun*
*ECOFYS Research and Consultancy,
Biltstraat 110, 3572 BJ Utrecht, The Netherlands
SUMMARY
This paper describes a numerical study of the behavior of an integrated collector storage DHW
system. The transient natural convection occurring in the hot water container is modeled
numerically. The simulation results are compared to experiments and serve as input for annual
simulations. A validated model will be a fast and cheap tool to test and design new variants.
KEYWORDS
Integrated Collector Storage; solar domestic hot water system; transient natural convection;
numerical simulation.
INTRODUCTION
One of the possibilities to arrive at more cost effective solar domestic hot water systems, is
to use an integrated collector storage (ICS). In this way costs can be reduced by a
simplification of the system. Experiments and simulations have been conducted by several
researchers in this area (e.g. Schmidt, 1990, Ecevit, 1990), but little attention has been paid to
the detailed physical processes occurring inside the hot water container. Studying these
processes can give a significant contribution to optimizing the energy yield of the ICS.
Simulations can give clues to explain experimental results and help in finding better designs.
In Fig. 1 the schematic cross section of an ICS is shown. This rectangular cavity is filled with
water, and the sides and back are thermally insulated. The front side, which also is the
absorber, is covered with transparent insulation. Hot water can be obtained by means of a heat
exchanger at the back or by directly extracting water from the container while refilling with
cold water.
A somewhat less realistic setup was used for our simulations, in order to be able to compare
our results to experiments conducted elsewhere (Visser, 1991). To simulate for forced heating
with a resistive heater, the heat losses at the front were set to zero. Heat losses at the back
and sides were assumed to be 0.375 W/m2-K. The ambient temperature was 13 °C. The
height (H) of the cavity was 1.54 m and its width (W) was 0.081 m, giving an aspect ratio
-------
1352
transparent insulation
hot water container
hot water
PUR insulation
heat exchanger
cold water
Fig. 1. Schematic cross section of an integrated collector storage DHW system.
(A) of 19.0. The collector has an inclination of 45 degrees.
MATHEMATICAL FORMULATION
The heat and fluid flow in the integrated collector is modeled using a two-dimensional
formulation1. The x and y direction are chosen parallel to the width and the height of the
cavity, respectively. The equations governing conservation of mass(l), momentum(2,3), and
energy(4) can be written as
* = o (i)
dx dy
du du du 1 dp Q/rr rrj \ d / a du\ d (u du\ (2
— +u—+v— = + g B(r-rn)+—_— + —_—K
dt dx dy p dx dx ^ p to I dy ^ p dy j)
^+«^+viH = -1^ + gvP(r-r0)+±fJiill) + ±(m) (3)
dt dx dy ' p dy y dx y p dx J dyyp dy J
1 An indication for the correctness of a 2D model is the fact that Visser (1991) finds no
three dimensional effects in his experiments.
-------
1353
dT dT dT d ( dT\ d ( dT\ (as
— +u—+v = a— + —a W
31 dx dy dx J dy ^ dy J
where the fluid is modeled as Boussinesq incompressible. This implies that the density is
regarded as a constant, except in the buoyancy term of the momentum equations, where the
density is assumed to decrease linearly as the temperature increases. The unsteady terms in the
equations are retained accounting for the unsteadiness of the heating of the container.
The temperature boundary conditions are determined by the incoming flux q" and the heat
transfer coefficients of the insulation. The velocity boundary conditions follow from the no-slip
condition and the impermeability of the walls (u=v=0). For initial conditions we took a volume
of water at rest of uniform temperature equal to the ambient temperature. At t=0 suddenly a
uniform flux is applied to the front of the container.
The above equations involve three dimensionless groups which characterize the problem,
namely, the geometric aspect ratio A = H/W, the Prandtl number Pr = v/a, and the Rayleigh
number based on the cavity height H and the incoming flux q", Ra = (g|3q"H4)/(avX). The
Prandtl number is a property of the fluid and equals 7.0 for water at 20 °C. The problem also
depends on the angle of inclination of the collector which determines the direction of the
acceleration of gravity, g, relative to the chosen coordinate system.
Turbulence was modeled using Reynolds averaging and an eddy viscosity model to find an
expression for the Reynolds stresses -uv and -~uT¦ We used a two-equation k-e model which
introduces two extra unknowns k and e, the kinetic energy of the turbulent fluctuations and the
turbulent dissipation. Two extra equations were used to solve for these variables.
NUMERICAL PROCEDURE
The equations Eire discretized using the finite-volume method. The equations are integrated on
a per volume basis, and the interrelations between the fluxes at the interfaces lead to a system
of nonlinear algebraic equations. The pressure field appears as part of the source term in the
momentum equations. But the pressure is only indirectly specified, via the continuity equation.
To solve for the correct pressure and velocity fields, the iterative SIMPLE pressure-correction
method as introduced by Patankar & Spalding (1972) is used.
For carrying out the calculations in this study, we used a code developed at Delft University
(Henkes, 1990). The nonlinear equations are solved using a line Gauss-Seidel method, while
the pressure-correction equation is solved directly. Relaxation is used to prevent divergence
from occurring. To test for convergence, we checked that the residuals in the discretized
equations and the changes in the solution between two iterations are below a small stop
criterion.
To calculate the transient behavior of the ICS, the equations had to be solved repeatedly for
successive small time steps. One time step was of the order of At = (H2/v)Ra~I/4. Once a
convergent solution had been reached for the first time step, only a few iterations were needed
to solve for the next time step.
-------
1354
NUMERICAL RESULTS
We calculated the heat distribution and flow pattern resulting from the transient heating of a
rectangular water container as described above. The heat flux at the front of the container was
603.2 W/m2. This results in Ra = 8.3-1013. We simulated heating for a period of 225 minutes.
These settings were the same as used by Visser (1991) in his experiments.
36
-0.05
-0.15
0.0
Fig. 2. Cross section of ICS showing solution for temperature field (upper, °C)
flmn /I(xiiTuT iftor frtr '1 rtf O O m 1 n Pn—ft 'i. 1 H
solution
of
ICS
for
showing
section
In Fig. 2 the resulting isotherms and streamlines are shown as they occur after 225 min. Along
the heated side of the ICS a boundary layer is formed. The flow takes the heated water up and
to the back of the container. A horizontal temperature stratification is observed. During transient
heating the energy content of the water in the container increased nearly linearly with time.
After 225 minutes 0.82% of the incoming heat flux had been lost through the insulation at the
back and sides.
Figure 3 shows the temperature profile in the middle of the cavity at x = 0.5-W for several
calculations. One calculation was conducted using a standard k-e turbulence model. Anjather
used a k-e model using low-Reynolds-number modifications (Lam & Bremhorst model, see
Henkes, 1990). A third calculation used no turbulence modeling at all. We saw that the
turbulence intensity died out quickly in the course of the first calculation and disappeared
altogether in the calculation using the low-Reynolds-number modifications. The difference
between the no-turbulence and low-Reynolds results is unobservably small. The calculated
temperature profile is compared with experimental results (Visser, 1991). The calculations show
a lower temperature gradient than is obtained by measurement.
-------
1355
70
standard k-e
60
-• low-Reynolds
modification
-• no tirbu-
lence model
' Visser. 1991
0}
b
40
(0
30
20
10
O.O
0.8
0.4
1.2
1.6
height
-------
1356
The findings of this study can be related to work conducted elsewhere to construct a simple
model for the ICS to calculate the annual efficiency (Visser, 1991). This model uses a coarse
segmentation of the ICS. Our results can be used to find a physically sound foundation for this
model.
When the simulation outcome has been shown to correspond well to the experimental results,
a cheap and fast tool to design and examine new ICS-variants has been obtained.
NOMENCLATURE
A aspect ratio = H/W
cp heat capacity (J/kg-K)
g acceleration due to gravity (m/s2)
H height cavity in y-direction (m)
Pr Prandtl number = v/a
p pressure (N/m2)
q" heat flux (W/m2)
Ra Rayleigh number =
(gPq"H4)/(av\)
T temperature (°C)
T0 reference temperature (°C)
t time (s)
u velocity in x-direction (m/s)
v velocity in y-direction (m/s)
W width cavity in x-direction (m)
a thermal diffusivity = V(pCp) (m2/s)
P thermal expansion of fluid (1/K)
X thermal conductivity (W/m-K)
(x dynamic viscosity (kg/m-s)
v kinematic viscosity (m2/s)
p fluid density (kg/m3)
ACKNOWLEDGEMENT
This investigation was financed by NOVEM (Netherlands Agency for Energy and the
Environment), Utrecht, the Netherlands.
REFERENCES
Ecevit, A., M.A. Chaikh Wais and A.M. Al-Shariah (1990). A comparative evaluation of the
performances of three built-in-storage-type solar water heaters. Solar Energy, 44(1), 23-36.
Henkes, R.A.W.M. (1990). Natural-Convection Boundary Layers. Ph.D. Thesis. Technical
University, Delft.
Patankar, S.V., and D.B. Spalding (1972). A calculation procedure for heat, mass and
momentum transfer in three-dimensional parabolic flows. Int. J. Heat Mass transfer, 15,
1787-1806.
Schmidt, Ch. and A. Goetzberger (1990). Single-tube integrated collector storage systems with
transparent insulation and involute reflector. Solar Energy, 45(2), 93-100.
Visser, H., A.C. de Geus (1991). Integrated collector storage: model development for
performance calculations and test evaluation. To be published.
-------
1357
INVESTIGATION ON SOLAR DOMESTIC HOTWATER SYSTEMS COMBINED WITH
AUXILIARY HEATERS
G.A.H. van Amerongen, H. Visser and A.C. de Geus
TNO - Building and Construction Research
P.O. Box 29
2600 AA Delft
The Netherlands
ABSTRACT
In the System Test Facility of TNO 7 combinations of SDHW's and auxiliary heaters were
tested. The aim of this test was to investigate the interaction between SDHW and auxiliary
heater (AH) and to estimate the yearly energy savings potential of each combination. The
manufacturers installed the systems in the test facility. Prior to the test the functionality of the
systems were checked and in some cases improved. It turned out that most combinations
performed reasonably well. Improvements can be realised by reducing the heat ss in the
auxiliary integrated SDHWS's , by reducing the hot water temperatures and by enhancing the
efficiency of the AH.
KEYWORDS
Solar domestic hot water systems, Auxiliary heaters, Integrated solar domestic hotwater
systems, Dynamic performance testing, control system, energy savings.
INTRODUCTION
During the stage of development most Dutch SDHW systems are tested inside using the solar
simulator of TNO. The applied test procedure results in the main parameters of the system.
This information gives the necessary feed-back to the manufacturers for their development
activities. However, to estimate the solar contribution of the SDHW or the yearly energy
savings, an indoor test is not sufficient. In order to determine these values the dynamic outdoor
behaviour of the total combination of SDHW and auxiliary heater has to be investigated.
Up til now experimental results about several possible combinations of SDHWS and auxilia-
ries were lacking. For that reason a side by side test of 7 SDHW/AH combinations was
performed. For the Dutch market typical combinations were chosen. This was realized by
inviting both manufactures and utilities, involved in renting SDHW systems to users, to
propose the combinations to be tested. This involvement of market parties increases the
applicability in practice of the results and promotes the transfer of knowledge about solar
systems.
-------
1358
SPACE HEATING
SPACE HEATING
TAP WATER
TAP WATER
GROUP 1A: PREHEATER - CENTRAL/TAP WATER
TAP WATER
TAP WATER
GROUP 1& PREHEATER - TAP WATER
GROUP 2& INTEGRATED - ELECTRICAL
Fig. l.Two groups of SDHW/AH combinations selected for the
tests.
Therefore the test aimed at a more broad set of results than merely the dynamic behaviour of
the combinations. The following goals were defined:
the evaluation of the installation of the combinations;
the detection of failures;
the evaluation of the control system;
the prediction of the energy saving.
INVESTIGATED SYSTEMS
Based on the information form manufactures and utilities, 7 combinations were selected. The
combinations can be divided into two groups which are shown in fig. 1. All tested combinati-
ons are given in table 1.
In the first type of system combinations the SDHW acts as a separate preheater. For all
systems in this group a gas auxiliary heater is applied. This heater is either a combined central
space and tap water heater or a dedicated tap water heater.
The second group is formed by SDHW systems with build-in auxiliary heating. This auxiliary
is either an electrical element or central space gas heater expanded with a boiler control.
-------
1359
Table I. Tested SDHW/AH combinations
Combination type; SDHW type: Auxiliary type :
1)
Preheater
1A
1)
2.6
m2, 120 1
l)
Central/tap water (gas)
2)
Preheater
1A
2)
2.5
m2, 100 1
l)
Central/tap water (gas)
3)
Preheater
IB
2)
2.5
Low-
m2, 100 1
flow concept
2)
Tap water (gas)
4)
Integrated
2A
1)
2.6
m2, 120 1
3)
Central/tap water (gas)
5)
Integrated
2A
2)
2.5
m2, 80 1
3)
Central/tap water (gas)
6)
Integrated
2B
1)
2.6
m2, 120 1
4)
Electrical,
Temp. 85 °C, night rate
7)
Integrated
2B
1)
2.6
Glas
m2' 120 1
S teflon cover
5)
Electrical
Temp. 65 C, night rate
Temp. 40 °C, day rate
EXPERIMENTAL SET-UP
The manufacturers of type 15 and type 2) SDHW systems were invited to install the 7 combinati-
ons in the System test Facility of TNO. All combinations were equipped with sensors for the
measurement of temperatures, flows and control signals.
From all systems hot water was automatically tapped 5 times a day by withdrawing each time
4.6 MJ. The total of 23 MJ/day is the equivalent to a withdrawal of 110 liters of water from 15
°C to 65 °C. The withdrawal of a fixed amount of heat instead of a fixed amount of water is in
our point of view a more realistic simulation of the use in practice. Moreover, in this way the
interaction between SDHW and AH will effect the system energy consumption and gives
therefor information about the best combinations. The experiment was carried on during 4
weeks in spring.
RESULTS
The technical realization
Especially the aspects concerned with heat were judged. In general these aspects in technical
realization of the systems were judged as good.
However, the technical realization of the top part of the storage tank in the integrated SDHW
concepts show a weak point. This part of the tank is always at a high temperature (due to the
auxiliary) and should loose as less heat as possible. Due to the way the piping is connected to
the tank (hot water pipes, connections to the external auxiliary gas heater and electrical
connections) the heat losses are high.
-------
1360
Failures
During 4 weeks in wintertime all systems were turned off. After starting up the systems again,
3 collector pumps from SDHW type did not function properly. This problem is probably
caused by the type of pump used, which is highly sensitive to calcareous deposits from the
water. Furthermore, this SDHW type is of a drain-down type in open circuit, which causes a lot
of calcareous deposits. Although only a simple action was needed to bring the pump in
operation again, users in practice possibly would not notice this failure immediately.
Other failures were not observed during the test period.
Control system
The hot water temperature of the SDHW systems were monitored. The temperature control was
in all cases judged as good. In the combinations 2) and 3) a valve is used, which bypasses the
auxiliary at higher temperatures. The control of these valves was judged as good. The
temperature level was high (60 to 85, °C), except for the dedicated tap water heater (type 2\ 55
°C) and the electrical heater type 5) (40 to 65 ° C). A low hot water temperature setting of 55
°C for all systems would increase the performance of the systems significantly.
The on/off control of the gas auxiliaries in the integrated SDHW types showed a rather high
number of heating cycles. Theoretically this number has to be lower or equal to the number of
tap water withdrawals. However, a maximum of 12 cycles and an average of 8 cycles at 5
withdrawals a day were noted. The high number of cycles will decrease the burner efficiency.
The collector pump control in all the systems was judged as weak. In the SDHW systems of
type the pump switched too frequent and in the SDHW systems of type 2) the pump switched
too less. This control type was examined in a separate investigation in more detail. It appeared
that accuracy of the hardware used is satisfactory. However, the measurement method used,
showed in both SDHW types weak points. In SDHW type the main problem was traced to
the location of the temperature sensor in the storage tank. In the SDHW type 2> the main
problem was traced to the construction of the temperature sensor in the collector.
Energy savings
The measurement data were evaluated with help of the STF-method which is developed by
TNO (de geus, 1987). The system performance is evaluated with dynamic efficiency curves,
build from the measured data, describing components and the total system. Based on these
curves the yearly energy consumption is estimated and the strong and weak points of the
systems can be examined. Furthermore, in some cases it is also possible to evaluate the
performance of other combinations with the tested SDHW and AH.
To determine the energy savings and the contribution to the heat demand of the SDHW, the
performance of the SDHW combinations are compared to a conventional DHW system
(without solar). This requires the definition of the yearly energy consumption of 3 reference
DHW systems:
Auxiliary (gas) for the preheater types SDHW;
Auxiliary (gas) for the integrated types of SDHW;
Auxiliary (electrical) for the integrated types of SDHW.
In table 2 and table 3 the results are given.
-------
1361
Table 2. Reference energy consumption and energy saving.
SDHW comb. Reference DHW energy Energy saving SDHW
no.: Natural gas kWh- el. %
1)
380
m3/ jr
43
2)
380
m3/jr
41
3)
380
m3/jr
31
4)
420
m /jr
39
5)
420
m3/jr
23
6)
2710
kWh/jr
39
7)
2630
kWh/jr
52
From these results the following conclusions were drawn.
The yearly energy savings were in general satisfactory.
The low-flow concept in the SDHW combination 3) was disappointing, probably due the
heat exchanger in the storage tank or the heat exchange in the collector.
The SDHW combination 7) performed very good, but due to the high energy consump
tion during the day rate (high!), the cost effectiveness is doubtful.
The SDHW concept S) performed bad. The storage setup needs changes.
The positive effect of an extra teflon cover in SDHW combination 7) was not noticed,
due to the rather low average working temperature of the collector.
The preheater types of SDHW performed better than the integrated types. This effect is
due to the heat losses of the storage tank.
Table 3. Reference DHWS heat demand and SDHWS heat contribution
SDHW comb. Reference DHW energy: Contribution SDHW:
GJ/jr
GJ/jr
1)
8,4
4,5
2)
8,4
3,8
3)
8,4
3,7
-------
1362
The positive effect of a low hot water temperature on the performance of the SDHW
was clearly seen. Combination 7) shows a high performance for this reason. The bad
performance of combination 7) is partly compensated by the positive effect of the low
hot water temperature.
The SDHW type 1) was in general performing better than the type 2). This is caused by
4% lower collector area and a lower instantaneous collector efficiency.
The auxiliary type ^ showed a low efficiency.
GENERAL CONCLUSIONS
This investigation showed clearly that the actual energy saving of a SDHW heavily depends on
the performance of the total combination of SDHW and auxiliary heater. Experiments under
dynamic conditions for both irrandiance and tap water withdrawal are necessary to investigate
the interaction between SDHW and the auxiliary heater and to check the control systems.
GENERAL RECOMMENDATIONS FOR IMPROVEMENTS
Improve the technical realization of the storage tank in the integrated SDHW types to
reduce the heat losses.
Install a diagnostic to show a malfunctioning of the collector pump.
Change the pump or the collector circuit in the SDHW type to avoid calcareous
deposit.
Choose a low hot water temperature.
Improve the control of the auxiliary gas heater in the integrated types to minimize the
number of burner cycles.
Change the collector pump control with respect to the measurement method to improve
the control accuracy.
REFERENCES
De Geus, A.C. (1987);
An evaluation method for solar energy systems monitored in the TPD System Test Facility.
ISES Congres 1987, Hamburg, West Germany.
-------
1363
LOW FLOW OR SINGLE PASS, THE HEART OF THE MATTER
C.W.J, van Koppen
Emeritus professor of Eindhoven University of Technology
Kosmoslaan 25, 5632 AT Eindhoven, Netherlands
ABSTRACT
For about ten years the so-called low flow control strategy has gained wide
acceptance. The physics of the strategy are not always well understood, however. It is
shown that it is essentially a singlepass strategy, which is only applicable in
combination with stratified storage. Some practical aspects are discussed.
KEY WORDS
Single pass control, low flow, stratified storage, solar water heating, solar space
heating.
INTRODUCTION, APPROACH
Up to about 1980 it was widely held that limiting the increase in temperature of water
flowing through a (series of) collector(s) to a low value of, say 10°C, yields/ in most
cases,, the best performance of residential solar systems, in particular water heaters.
The corresponding flow through the collector is typically about 50 kg/hr per m2 of
collector area (Duffie, J.A. and W.A. Beckman, 1980). However, H. Tabor (1969) noted
that in natural circulation DHW systems a much smaller flow and higher temperature
increase may yield about the same daily efficiency. He suggested that passing the
water in the storage tank about one time per day through the collector is best practice. It
is remarkable that no further conclusions were drawn at that time from Tabor's findings.
When designing the solar heating and DHW system for the Solar House at the
Eindhoven University of Technology it was calculated by T. van Wolde (1974) that the
highest heat gain would be obtained at a low flow rate of only 12 kgnr2hr1. Similar
results were found in several other systems (Van Koppen et al., 1979). In all these
cases much attention had been paid to the preservation of stratification in the storage
tank (floating inlet, low entry velocities, no mixing).
These contradictory views on the optimal flow rate have been reconciled in the early
eighties (Veltkamp (1980), Rademaker (1980), Van Koppen (1981)). The proceedings
of the ISES Congress in Perth (Szokolay, 1983) show that most leading research
institutes had by that time recognised the superiority of the "low-flow" control.
Sometimes, however, the physical and technical prerequisites for this superiority are
overlooked; even the name"low flow" is rather vague and may easily be misleading.
The purpose of this paper is to clarify these points.
Since the early eighties much progress has been made in the implementation of the
new control strategy in solar DHW systems. A comprehensive overview of the many
aspects that are important in practical realisation has been published by K.G.T. Holland
(1989), the most prominent researcher in this field. The present paper only discusses
-------
1364
some physical points that were not covered in Holland's paper. It finishes with a plea to
change and improve the name "low-flow" to "single-pass". It is shov/n that single-pass is
the essential characteristic of the control method.
In order to make the physical problems analytically tractable so as to avoid
untransparant numerical calculations, some simplifying assumptions are made. None
of these goes far from current practice.
- The collector efficiency is taken to be a first order function of the temperature
difference between fluid and ambient atmosphere.
- The collector loss coefficient is constant (which implies that the design of the
collector is always adapted to the flow rate).
- The storage is either perfectly stratified (no mixing) or well-mixed. These extreme
cases may represent actual situations quite well.
- The thermal losses of storage and lines and the pumping energy are negligible
compared to the amounts of heat gained and used.
- All thermophysical properties are assumed to be constant.
Further is is important to note beforehand that:
- Generality is achieved by reducing all quantities to the unit of collector area,
- Only systems and circumstances are considered in which the solar fraction is smaller
than one; such because it makes no sense to discuss optimal control when there is a
surplus of solar heat.
- In this paper only operational optimisation of a given system is considered, aiming at
minimisation of the auxiliary heating energy.
- The argumentation seemingly concerns only the solar water heater. However, it
applies equally well to air heating and to space heating provided that the return
temperature of any heating air or water is low and more or less stable. The reason is
that it is the thermal capacitance flow rates that really count.
A SIMPLIFIED SYSTEM UNDER STEADY CONDITIONS
Somewhat surprisingly a quick apprehension of the physical nature of the high heat
gain at relatively low collector flow rates may be obtained by analysing a simplified
system under steady conditions, encompassing (see fig. 1):
- a distribution system characterised by a constant return temperature (or water supply
temperature from the mains) T0, and a constant thermal capicitance flow rate per unit
of collector area, Hds
- weather conditions characterised by a constant ambient temperature Ta and a
constant, 24 hours per day, solar irradiance Gs, normal to the plane of the collector
- a collector characterised by a linear efficiency
n = no-Us{(Tf + Tc)/2-Ta}/Gs (1)
In accordance with the steady weather and heat demand conditions the thermal
capacitance flow rate through the collectors is put equal to a steady Hc,s- Further the
installation is assumed to form one single closed system.
We continue our discussion as if water were the working fluid in the system.
As regards the situation in the storage three different regimes may occur.ln the stratified
storage, with HC)S Hd,s (case 1) the water moves upward and the temperature in the
storage is T0. With Hc,s > Hd,s (case 2) the water moves downward and the temperature
in the storage is TCiS.'Entering the distribution circuit in case 1 is a mixture of HC|S at
temperature TC|S arid (Hd,s - Hc,s) at temperature T0. Therefore the temperature of the
water entering the distribution circuit and the collector respectively amount to
Te,i = To (1 ~ Hc,s / Hd,s) ^ Tc s Hc,s t ^d,s and Ty = T§t= To (2, 3)
If H0,s > Hd|S (case 2) it is easily derived that
-------
1365
Te>2 ~ Tst - Tc and T^2 - T0 Hd,s / Hc,s + (1 - Hd,s / Hc,s) Tc (4, 5)
Note that because of the stratification the hotter water is preferently extracted at the
upper end of the storage and the colder water is so at the bottom end.
For a well-mixed storage, case 3, it is found that:
Tf,3 = Te,3 = Tst= (To + Tc Hg s / Hd,s )/ (1 + Hc,s / Hd,s) (6)
In all cases the energy balance of the collector reads:
Cs = Gs,ut" UsfTf + Tc) /2 - T0} = HCiS (Tc - Tf) (7, 8)
where the utilisable solar radiation is defined as:
Gs.ut = 1o Gs - Us (To - Ta) (9)
HOT)
[COLD)
STRATIFIED STORAGE
WELL-MIXED STORAGE
SCHEMATIC OF A SOLAR HEAT SYSTEM
Hc / Hd
IINFLUENCE OF THROUGHPUT-RATIO COLLECTOR / USE
ON HEAT GAIN
Fig, 1.
Combining (7,8) with (2), (5) or (6) finally yields
CS,1 = Gs.ut I (1 + Us 12 Hc,s)
Cs,2 = Gs,ut / (1 + Us / Hd,s - Us / 2 Hc,s)
Cs,3 = Gs.ut I (1 Us / Hd,s "*¦ Us,s / 2 Hj^s)
Fig. 2
(10)
(11)
(12)
Fig. 2 illustrates the nature of (10), (11) and (12). The curves are based on fairly realistic
numerical values for the various quantities: Ta = 0°C, T0 = 20°C, Hd,s = 8.33 W°C-1nrr2
(a waterflow of 7 Kg h"1 m-2), n0 = 0,80, Us = 5/3 W m-2 °c-\ Gs = 750 / 3 W m"2 (the
denominator 3 in the last two values serves to compensate for the steady 24 hours a
day operation). The solar fraction of the system is about 50% if Td is 60°C. The curves
in fig. 2 have a general significance. For all realistic sets of parameters the highest heat
gain is invariably found for a stratified storage and at Hc,s/Hd,s = 1- That is for a single
pass system. Actauily the shape as such of these curves was discovered earlier than
the single pass kriterion (Van Koppen et al. 1979).
INCORPORATING THE DAILY CYCLE
It is easily shown that changing the steady system into a system with a three times more
intense operational period of only 8 hours per day does not change the daily
performance. If the daily average thermal capacitance flow rate through the collector
-------
1366
Hc ^ Ho (or Hd) the lower part of a stratified storage is always filled with water of
temperature T0. So Tf = T0, and (7) is valid but for the index s. During the operational
period (see also eq. (10))
Tc and the daily heat gain remain unchanged. During the operational period the upper
part of the storage vessel is gradually filled with hot water of temperature Tc. This water
is used during the "evening" and "night". Provided the storage vessel is sufficiently
large the draw off pattern as such is of no significance, only the daily average Hd plays
a role.
If He < Hd the daily feed of the collector consists of 24.3600 (Hc - Hd) plus 24.3600 Hd.
Hc is constant (see (13)) but the momentaneous value of Hd may vary. From (7) it
follows that Tc is lineary proportional to Tf, so the average value of Tf determines the
average of Tc and, consequently, the heat gain of the collector. But the average of Tf is
completely determined by the composition of the daily feed of the collector,
independently of any variability of Hd. By taking Hd = 3 Hd,s during the operational
period eq. (5) becomes valid, which implies again that C = 3 Cs (eq. (11)) during the
operational period and that the daily heat gain remains unchanged.
Note that this proof depends on the linear approximation of the collector efficiency.
The proof that the curve in fig. 2 for the well mixed storage also holds for the system
with an 8 hour operational period follows the same line of thought as the proof just
presented for the stratified storage with Hc < Hd. Because the proof is rather lengthy it is
not given here.
It remains to investigate some consequences of the actual variability of irradiance and
heat demand. More precisely we will investigate the associated optimal control for
these circumstances; incorporating in this way the daily cycle.
We start from the main result obtained so far, that the highest heat gain of a (simplified)
system is realised by making Hc = Hd. We take the average over one day because it is
the daily cycle that we are incorporating.
From (3) it follows that the optimal value Hc = Hd corresponds to the highest possible
collectorflow for which Tf has the lowest possible value T0. This observation is
essential. The total amount water available at T0 during one day, W?, is equal to the
volume of the storage vessel plus the water returning from the distribution system or
coming from the mains during the operational period of the collector. It can be
demonsrated by numerical analysis that making the daily throughput of the collector
equal to W0 results in the highest heat gain.
If the optimal daily throughput W0 is known the best momentaneous Hc(t) can be
determined from variational calculus. Rewriting (7) for Tf = T0 and refering to fig. 3 the
heat gain of the collector may be formulated:
In (14) the quantities C, Gut and Hc are functions of timet. Gut is known (perfect weather
forecast). Two conditions have to be satisfied (see fig. 3)
(G, U,Hc,C) = 3(Gs, Us, HClS, Cs)
(13)
C = Gut - U{(T0 + TC) /2 - To} = Hc 0"c - Tc) = Gut (^hj) (14- 15)
(16)
and
-------
1367
C0*~Ca"k&d"~
dt= max
(17)
From variational calculus it follows that the last condition is satisfied if
dkMMI
-!=—j?Hc + U = constant A
u He
constant A
(18)
which yields after some algebra
(18a, 20)
The dimension of A is temperature. On a bright day the value of A is about 1°C, in
cloudy weather it has a lower value.
It should be noted that several conclusions mentioned above have been published
independently in a somewhat different form by T.S. Scheffler (1987).
SOME PRACTICAL CONCLUSIONS, THE HEART OF THE MATTER
The systems considered in the foregoing are still of a rather theoretical nature. Actually
heat losses of the storage, mixing, non-linearity of the collector characteristics,
unpredictability of irradiance and heat demand during the day, pumping energy,
etcetera, all play a role in the optimal control strategy. Taking into account all these
effects requires extensive numerical calculations. Contributions of this kind have been
made a.o. by Veltkamp (1980, 1981). More fundamental insights have been presented
t1og
TIME (HRS)
EXAMPLE OF VARIABLE IRRADIANCE
-------
1368
by Rademaker (1980) and Van Koppen (1981). Again and again it is found that in a
well designed system the single-pass control strategy holds. To compensate for these
imperfections it usually suffices, to give an allowance of 15-40% to the single pass
throughput for the collector. The magnitude of the allowance is best determined from
computer simulations. Even with a constant collector flow and in the highly variable
Dutch climate the similarity between the steady system and real systems is apparanit in
all cases where stratified storage can be applied
The author feels that for reasons of clarity the new control strategy should be called
after its real nature: SINGLE PASS CONTROL.
NOMENCLATURE
A
a constant (see (18a) and (20)
-
C
collector heat gain per unit time
Wm-2
Q
irradiance
Wm-2
H
thermal capacitance flow rate
Wm-2°C-1
t
time
s
T
temperature
°C
U
collector loss coefficient
Wm-2°C-1
W
capacitance of collector feed water
J m-2°C-i
H(o)
(prime) collector efficiency, see (1)
-
Subscripts
a
ambient
s
simplified steady system
c
from collector
St
storage
d
distribution, use
ut
utilisable, see (8)
e
entry distribution circuit
w
water
f
collector feed
1,2,3
case 1,2,3, see text
0
return-, supply water
1,2,3,4 see fig. 3
REFERENCES
Duffie, J.A. and W.A. Beckmann, Solar Engineering of Thermal Processes, John Wiley
& Sons, 1980, ISEJN 0-471-05066-0, pp 419-21.
Hollands, K.Q.T. (1990) New Developments in Low-Flow Stratified-Tank Solar Water-
heating Systems, Proc. of the 1989 Congress of ISES in Kobe, pp. 463-71 (invited
paper), Pergamon Press, 1990, ISBN 0-08-037193-0.
Rademaker, O. (1980), On the Dynamics and Control of (Thermal Solar) Systems using
Stratified Storage, in C. den Ouden (ed), Thermal Storage of Solar Energy, Proc. of a
TNO Symposium, p. 61-72, Martinus Nijhoff, The Hague, ISBN 90-247-2492-9.
Scheffler, T.B. (1989), Optimal Flow in a Solar Water Heater, Proc. of the 1987
Congress of ISES, pp. 879-83, Pergamon Press, 1980, ISBN 0-08-034315-5.
Tabor, H (1969) See first reference p. 421
Van Koppen, C.W.J., P.J. Simon Thomas and W.B. Veltkamp (1979), The actual
benefits of thermally stratified storage in a small and medium size solar system. Proc. of
the 1979 Congress of ISES in Atlanta, pp 576-580, Pergamon Press, 1979, ISBN 0-08-
025074-2.
Van Koppen, C.W.J., (1981) Active Heating in Buildings, Proc. of the 1981 Congress of
ISES, pp. 8-19 (keynote paper), Pergamon Press, 1982, ISBN 0-08-026730-0.
Van Wolde, T., (1974). See first reference Van Koppen et al.
Veltkamp, W.B. (1980), Thermal Stratification in Heat Storage, in C. den Ouden (ed),
Thermal Storage of Solar Energy, Proc. of a TNO Symposium, p. 47-59, Martinus
Nijhoff, The Hague, ISBN 90-247-2492-9.
-------
1369
Low-Flow Solar Hot Water System
by
Torben Esbensen
Consulting Engineers FIDIC
Mollegade 54-56, DK-6400 S0nderborg
Denmark
IB.
ABSTRACT
2
The 44 m low-flow solar water heating system is the first larger
system in Europe using the attractive low-flow principle based on
lower flow, smaller piping and stratification in the storage tank.
The low-flow principle can increase the thermal performance with
10-20%.
KEYWORDS
First larger low-flow solar system,
with 10-20%.
44 m
increased performance
INTRODUCTION
A solar system for heating of domestic hot water has been con-
structed at S0fartsskolen situated in Senderborg in Jutland.
Sofartsskolen is a bording school with 96 seamen trainees.
The system has functioned since January 1988 and has been detailed
monitored since 1990.
-------
1370
The system is the first in Europe designed for larger applications
than single-family houses and making use of the attractive low-
flow principle based on lower flow, smaller piping and stratifi-
cation in the storage tank.
It is probably also among the first larger systems internationally
operating ^ad'ivg this principle.
As in most larger solar applications the system is supplying heat
for both the domestic hot water use and the considerable heat
losses in the circulating piping of the hot water.
S-.mall system experiments at the Thermal Insulation Laboratory
at the Technical University of Denmark have shown that the low-
flow principle can increase the thermal performance from 10 to
20%.
This is obtained with a system which is a little cheaper than
normal systems since piping have smaller dimensions.
The system is of significant importance for the future design of
larger solar systems in northern countries. If expectations are
met it will function as prototype for these system. Furthermore
the monitoring project will add knowledge about the low-flow
principle to be used for future design tools.
The monitoring project is financed by the Danish Agency of Energy.
The system has been designed by consulting engineers company
Esbensen.
The monitoring is performed by the Danish Solar Energy Testing
Laboratory, while the evaluation is performed by the Thermal
Insulation Laboratory, Technical University of Denmark.
-------
INVESTMENT
2
44 m solar collectors installed
Piping in the collector circuit
2000 litre storage tank including heat
exchanger
Piping in the storage circuit
Insulation of pipes and storage tank
Solar collector liquid
Electrical supply
Project fee
Danish VAT, 22%
3 0% Danish governmental grant
Total price for the system-owner
Output_for_a_normal_system
22,000 kWh (500 kWh/m2) x 0.075 US$
Simple pay-back period:
I!§£i!5§£®§_2y£Ey£_f or _a_ low-flow_ system
26,000 kWh (600 kWh/m2) x 0.075 US$
Simple pay-back period:
Total domestic hot water:
Heat loss in the circulating pipes:
1371
US$
10,000
US$
2,500
US$
5,000
us$
2,700
us$
3,000
us$
500
us$
700
us$
24,400
us$
4,000
us$
28,400
us$
6,200
us$
34,600
us$
10,400
us$
24,200
1,650 US$/year
15 years
1,950 US$/year
12 years
7 3,000 kWh/year
54,000 kWh/year
127,000 kWh/year
-------
-------
2.7 Solar Domestic Hot Water III
^receding page blank
-------
1374
LOW FLOW SOLAR HEATING SYSTEMS - THEORY AND PRACTICE
S. Furbo
Thermal Insulation Laboratory, Building 118
Technical University of Denmark
DK-2800 Lyngby, Denmark
ABSTRACT
Experimental investigations have shown that the thermal performance
of low flow solar heating systems is greater than the thermal
performance of traditional solar heating systems. Further, the low
flow principle makes it possible to reduce the cost of solar heating
systems. Therefore solar heating systems using low flow operation are
extremely attractive.
Both theoretical work and laboratory experiments concerning
optimization of small low flow solar heating systems for domestic hot
water supply have been carried out.
In 1989 low flow solar heating systems were introduced on the Danish
market. Measured thermal performances as well as experience
concerning the reliability of the first of these systems in practice
are presented.
KEYWORDS
Solar heating systems, domestic hot water supply, low flow operation,
laboratory experiments, thermal advantage, mathematical model,
optimum design and operation strategy, marketed systems, measured
thermal performance and experience from practice.
INTRODUCTION
Investigations (van Koppen, Thomas and Veltkamp, 1979? Furbo and
Mikkelsen, 1987; Hollands, 1988) have shown that the thermal
performance of low flow solar heating systems is greater than the
thermal performance of traditional solar heating systems.
As the low flow principle also makes it possible to reduce the cost
of solar heating systems the low flow systems are extremely
promising.
Therefore investigations were initiated at the Thermal Insulation
Laboratory in 1987 in order to develop optimum designed low flow
systems.
LABORATORY EXPERIMENTS
Since 1987 the thermal performance of three different small solar
heating systems for domestic hot water supply have been measured by
means of side-by-side experiments (Furbo, 1989,1990).
One of the tested systems is designed as most of the marketed solar
heating systems in Denmark today. The heat storage for this system
-------
1375
is a hot water tank with a built-in heat exchanger spiral at the
bottom of the tank. The solar collector fluid is pumped through the
heat exchanger spiral with a normal flow rate of about 1-1/min m2
solar collector.
Additional two identical systems were tested. Their heat storages are
mantle tanks where the solar collector fluid is slowly pumped through
the mantle from the top to the bottom. In this way it has been
possible to determine optimum flow rate and control strategy for a
solar heating system with a mantle heat storage.
The volume of the hot water tanks is 200 1 and electric heating
elements are placed in the top of the tanks. In this way the water
is also heated in periods without sunshine. The thermostat
temperature of the electric heating elements is set at 48°C.
Identical marketed solar collectors are used in the solar heating
systems, which are tested under uniform, realistic conditions. Test
periods with different solar collector types and areas were carried
out.
The investigations showed that the mantle storage system performs
best with volume flow rates in the solar collector loop between 0.1
and 0.2 1/min m2 solar collector.
The mantle storage system with this small volume flow rate performs
better than the spiral tank system. This can be seen in Fig. 1. which
shows the measured performance ratios between the net utilized solar
energy for the mantle storage system and the net utilized solar
energy for the heat exchanger spiral storage system as a function of
the solar fraction for the heat exchanger spiral storage system for
different test periods with durations of one week. The results from
periods with one particular collector type are not different from the
results from periods with the other collector type. Therefore the
solar collector type has no great influence on the thermal advantage
of the low flow system. This is at least true for the two tested
collector types.
The thermal advantage of the mantle storage system is strongly
influenced by the solar fraction. For decreasing solar fraction the
thermal advantage of the mantle storage system is increasing.
The main reason for the thermal advantage of the mantle heat storage
system with a low flow rate is the large advantageous thermal
stratification which is built up in the heat storage during operation
of the solar collectors. Only a short period with sunshine is
necessary for the water at the top of the tank to reach a temperature
level where the electric heating element will be turned off.
Continuous measurements of the thermal performance of the two systems
were carried out for a year from June 1989 to May 1990. Each system
had a 4.0 m2 Danish marketed solar collector with the efficiency:
0
.74 - 5.4 • Ta - 0.018 . Ta)'
-------
1376
Performance ratio=
Net utilized solar energy for the mantle storage system
Net utilized solar energy for the heat exchanger spriral storage sys.
1.6 r
1.4
1. 2
1.0
0.8 .
0.6 .
0.4 .
0.2 .
0 i i i i i
0 20 40 60 80 100
%
Solar fraction for the heat exchanger spiral storage system.
Fig. 1. Measured performance ratios for the mantle storage
system as a function of the solar fraction for the heat
exchanger spiral storage system.
The daily hot water consumption was 200 1 at 45°c. During the year
the mantle storage system performed 17% better than the heat
exchanger spiral storage system. The yearly net utilized solar energy
of the two systems were 319 respectively 272 kWh/year m2 solar
collector. The yearly solar fraction of the low flow system was 48%.
If the hot water consumption had been less than 200 1/day, the thermal
advantage of the low flow system would have been smaller. If the hot
water consumption had been larger than 200 1/day, the thermal
advantage of the low flow system would have been larger.
By the end of the year with identical operation the heat storage
tanks were inspected. At the bottom of both tanks were lime deposits.
The quantity of deposited lime was 2.5 times as large in the heat
exchanger spiral tank as that in the mantle tank. The main reason for
this difference is also the large thermal stratification in the
mantle tank. The higher temperatures of the tapped water, the smaller
amount of water is tapped from the storage, since hot water is tapped
in a realistic way from the storages. Therefore more water is tapped
from the heat exchanger spiral tank than from the mantle tank. The
increased amount of water tapped from the heat exchanger spiral tank
m ¦'« •
• ••
-------
1377
will result in an increased deposit of lime. Sooner or later the heat
transfer coefficient of the spiral and the system performance will
be noticeably reduced. The lime problems are therefore another reason
for using the attractive low flow principle.
SIMULATION MODEL
A detailed mathematical model simulating the thermal behavior of a
low flow solar heating system with a vertical hot water tank with a
mantle welded around a part of the surface of the tank as the heat
storage has been developed. The model has been validated by means of
indoor experiments with a mantle tank. Further, for the moment the
model is being validated by means of detailed measurements for the
solar heating system described in the previous section.
The theoretically determined influence of the flow rate in the solar
collector loop on the thermal performance of the system is in good
agreement with the experimentally determined influence. The optimum
flow rate is situated in the interval from 0.1 to 0.2 1/min m2 solar
collector.
MARKETED SYSTEMS
In order to show that the promising results from the laboratory can
be transferred to practice, a demonstration project was initiated in
1989. Three Danish producers of solar heating systems: Batec, Aidt
Mil jo ApS and Arcon Solvarme ApS participatedin the project. Each
producer built three small pilot test systems for different consumers.
That makes a total of nine systems which will be followed by means
of energy meters, water meters and hour meters until the end of 1991.
All nine systems have now been built in different locations in
Denmark. The systems have different solar collector areas, tilts and
orientations. All the systems have a mantle hot water tank as the
heat storage. The top of all the tanks can be heated by auxiliary
energy sources by means of an electric heating element and/or a heat
exchanger spiral or an extra mantle.
Much valuable experience is gained concerning the operation of the
systems. Most of the systems have been operating without any major
problems since the installation.
In two of the systems boiling occurred in the solar collector loop
during sunny periods without hot water consumption. We hope that
boiling can be avoided by using step 2 on the circulation pump
instead of step 1 in summer holidays.
In two of the systems the one-way valves in the solar collector loop
did not work. This caused thermosyphoning in the solar collector loop
at night resulting in a large extra heat loss from the heat storage.
After replacement of the one-way valves the problem have not occurred
again.
The heat loss from the top of some of the hot water tanks of the
systems are relatively large. This is extremely bad as large heat
losses from the top of the tanks strongly reduce the system
performances.
-------
1378
Above all, the large heat loss from the top of the tanks is caused by
a poor design of the pipe connections of the storage and of the
piping outside the storage.
Any pipe connection at the top of the tank causes an increased heat
loss. If the pipe is not turned downwards immediately outside the
tank, the water in the pipe will by natural convection always keep
a part of the pipe warm, and this will increase the heat loss. If the
pipe is a part of a loop through which a fluid circulates, there is
a risk that the fluid by natural convection will circulate in the
loop in periods when it was not projected. This might cause large
heat losses.
The loop must therefore be designed in such a way that the risk of
establishing a driving force which might start the thermosyphoning
in the loop is minimized. All parts of the loops should therefore be
well insulated and one-way valves should be installed in the loops.
Additionally, a poor design of the auxiliary energy source system of
the heat storage and a poor or missing control system for the
auxiliary energy source might result in too high temperatures in too
large parts of the heat storage. By determining the auxiliary energy
source system eind its control system it is therefore important that
only the required volume of water in the top of the tank is heated
by the auxiliary source and that the water only is heated to the
required temperature.
Five of the systems have now been in operation for a year from
February 1990 to January 1991. Some of the systems have in periods
or always had some of the above mentioned problems.
The total yearly horizontal radiations at the different locations in
the measuring year were close to 1018 kWh/m2 which is the total
yearly horizontal radiation in the Danish Test Reference Year.
The mean daily hot water consumptions during the year were for the
five systems 29, 30, 35, 41 and 44 1 per m2 solar collector. All the
systems are somewhat oversized, since a daily hot water consumption
of 50 1 per m2 solar collector is an adequate consumption for small
DHW systems in Denmark. The consumption; varies much throughout the
year, e.g. because of summer holidays. It is to be expected that the
modest consumption, the irregular consumption pattern and the above
mentioned problems will result in relatively small system
performances.
The measured yearly net utilized solar energy for the five systems
was 161, 227, 243, 336 and 371 kWh/m2 solar collector. The average
thermal performance of the systems was higher than previously
measured average performances for comparable small traditional solar
heating systems in Denmark (Mikkelsen, 1986).
Consequently, low flow systems can work with high thermal
performances in practice. Furthermore, some of the problems mentioned
above have been corrected during the year. It is therefore expected
that the performance of the systems will be improved in the next
year.
-------
1379
CONCLUSION
Laboratory experiments and model simulations have shown that low flow
solar heating systems for domestic hot water supply perform better
than traditional solar heating systems.
Measurements for some of the first installed small low flow solar
heating systems in practice showed that low flow systems can work
without operation problems with high thermal performances. The
investigations also showed that serious problems occurred if the
systems were not designed in the right way.
Low flow systems were introduced on the Danish market in 1989. In
1990 about 250 low flow systems were installed corresponding to about
20 % of the installed systems in Denmark.
The low flow systems marketed today are not cheaper than the
traditional marketed systems. The possibility to reduce the cost of
solar heating systems/ by making use of the low flow principle, is
therefore not utilized.
Based on the investigations it is strongly recommended to initiate
work with the aim to develope cheap, reliable and durable low flow
solar heating systems with high thermal performances.
ACKNOWLEDGEMENT
The author would like to thank the Danish Energy Agency which
financed the investigations.
REFERENCES
Furbo,S., and S.E.Mikkelsen (1987). Is low flow operation an
advantage for solar heating systems ? In W.H. Bloss and F. Pfisterer
(Ed.), Advances in Solar Energy Technology. Vol. 1., Pergamon Press,
Oxford, pp. 962-966.
Furbo,S. (1989). Solar water heating systems using low flow rates.
Experimental investigations. Thermal Insulation Laboratory, Lyngby.
Furbo,S. (1990). Small low flow DHW solar heating systems - status.
Thermal Insulation Laboratory, Lyngby.
Hollands, K.G.T. (1988). Recent developments in low-flow, stratified-
tank solar water heating systems. In L. Broman and M. Ronnelid (Ed.),
North Sun' 88 Proceedings. Swedish Council for Building Research,
Stockholm, pp. 101-110.
Mikkelsen, S.E. (1986). Output and experience with 31 solar domestic
hot water systems in Denmark. International Conference North Sun* 86.
Solar Energy at High Latitudes. DANHVAC. Charlottenlund, pp. 230-234.
van Koppen, C.W.J., J.P.S. Thomas, and W.B. Veltkamp (1979).
The actual benefits of thermally stratified storage in a small and
medium size solar system. In K.W. Boer and B.H. Glenn (Ed.), Sun II
Proceedings of the International Solar Energy Society Silver Jubilee
Congress.Vol. 1., Pergamon Press, New York. pp. 576-580
-------
1380
DYNAMIC METHOD FOR SOLAR SYSTEM TESTING -
MEASUREMENT RESULTS AND LONG TERM PERFORMANCE
PREDICTION OF DIFFERENT SDHW SYSTEMS
C. Arkar, S. Medved and P. Novak
Faculty for Mechanical Engineering, University of Ljubljana,
Ljubljana, Yugoslavia
ABSTRACT
Two years ago, a dynamic method joined existing methods of SDHW systems testing. It was
developed at the University of Munich and its major advantage is simple, short-term measure-
ments. We tested this method with the measurement of four simple systems. We additionally
checked the system models with short term predictions. The measurements showed that in
spite of the simple and short term measurements, the system models are such that the method
promises to be one of the more frequently used ones. On the basis of the obtained system
models, we have been able to perform long term performance predictions for Ljubljana.
KEYWORDS
Solar Domestic Hot Water System; thermisiphon system; dynamic testing; black-box meas-
urements; system model; short term prediction; long term performance prediction
INTRODUCTION
Many methods of SDHW system testing have been developed and are in use in the world.
These methods are distinguished both by testing procedures and in the form of the final results.
The dynamic method of solar system testing developed at L-M University in Munich (Spirkl,
1989,1990b) achieves some considerable simplifications in comparison to other methods, in
solar system measurement procedures, in system modelling and in predicting solar system
performance. This encouraged us to evaluate the quality and usefulness of this method on the
basis of concrete measurements.
DYNAMIC TEST METHOD
Measurements by the dynamic method (Spirkl, 1990a) are made using the black-box method,
which means that no measurements are necessary in the system, nor previous separate
measurements of any system component. Only the basic inlet variables and the output system
gain need be measured, everything as a function of time. Such easy measurements enable a
model and algorithm for system model parameter identification to be developed especially for
this method. Measurements are short term, three weeks being generally sufficient. It is also
possible to measure already installed systems without any interruption to the operation of the
system. The software developed enables automatic checking and condensing of data files, the
working out of a SDHW model of the measured system and long term performance prediction
-------
1381
on the basis of defined load data. The software additionally enables short term prediction
whereby the modelled and measured mean load power in the analyzed sequence is compared.
SYSTEMS MEASUREMENTS
We tested the described method (Arkar, 1991) with measurements of four simple SDHW
systems:
- termosiphon system with external jacked heat exchanger (system A) ;
- termosiphon system with coil heat exchanger (system B);
- termosiphon system without heat exchanger (system C) ;
- integral solar collector with TIM insulation (system ICS).
Short description of the systems :
Systems A, B and C had a solar collector of the same manufacture with aperture area of 2 m2
(A and B) and 1.44 m2 (C), tilted to an angle of 30°. Systems A and B are distinguished by their
heat storage, insulated with 5 cm insulation with X = 0.04 W/m2K. Heat storage of system C
was practically uninsulated. System ICS had black painted heat storage of volume 50 liters,
insulated with 1 cm Transparent Insulation Material (TIM). The cover was of acryllic glass with
pyramidic shape and aperture area of 1 m2. The measurements of the described systems were
made at the outdoor test loop of the Laboratory for Heating, Sanitary and Solar Technology
of the Faculty for Mechanical Engineering in Ljubljana.
On the basis of measurements over two to three weeks, we obtained the following parameter
values:
Table 1: Parameter values and fit error c for four measured systems
A
B
C
ICS
*
Ac
1.045
0.8708
0.5650
0.3372
m
uc
8.658
1.9907
0.8255
0
W/m2K
Cs
0.416
0.3188
0.4261
0.1655
MJ/K
Us
2.324
3.7878
6.9238
1.9336
W/K
Dl
0.027
0.5168
0.0569
0.5845
-
Sc
0.269
0.0654
0.0021
2.383
-
c
7.556
8.969
24.589
2.383
W
Where:
*
Ac* effective collector area, defined as Ac = AcFr(ra)
uc* effective collector loss coefficient, defined as uc* = uc/(ra)
Cs total store heat capacity
Us total store heat loss coefficient
Dl mixing constant, describing mixing effects during cold water inlet in the heat storage
Sc stratification parameter, describing stratification of temperature in heat storage
c is not a model parameter. It is a value introduced for model error estimation. On the ba-
sis of this estimate,we decided on when to end the measurements. The results clearly show
that the largest error is with system C, which was to be expected since measurements were
stopped after 11 days because we had problems with the system (frequently having to take
off the rubber connecting pipe between the solar collector and heat storage).
-------
1382
The described systems were tested at different seasons: summer, autumn and part/f winter. It
is not possible with such short measurements to embrace the entire spectrum of meteorological
conditions (summer - winter). We therefore verified the suitability of the systems models under
different weather and load conditions by the use of short term prediction. Figures 1,2,3 and 4
show the results of these analyses: the values of measured and predicted mean load power Pl
of analyzed measurement sequence and variability (higher > or lower <) from values
embraced by the model of the measured system.
These variables are:
- average daily load
- average daily ambient temperature
- average daily solar radiation
Vl (1/day)
Ta (°C)
Gc (W/m2)
system gain (W)
232 232
system gain (W)
I«
CD MEASURED
~ MEASURED
Fig. 1. Measured and predicted solar gain for Fig. 2. Measured and predicted solar gain for
system B in observed sequences
system A in observed sequences
system gain (W)
system gain (W)
'VQc>Ta
• WTa
CD MEASURED
CZD MEASURED
Fig. 3: Measured and predicted solar gain for
system C in observed sequence
Fig. 4: Measured and predicted solar gain for
system ICS in observed sequences
-------
1383
Results show good agreement between measured and predicted gain. The only exception is
system C, where the fit error (c) clearly indicates inadequate performance of the system.
LONG TERM PERFORMANCE PREDICTION
On the basis of the system models, we carried out long term performance prediction for
measured systems in Ljubljana (YU). Ljubljana is a smallish industrial city with a continental
climate. For long term performance prediction, meteorological data is required - test reference
year and load profile.
The following load profile was defined:
- daily load draw off volume 1 volume of heat storage
- draw off profile:
from 6 to 7 33 %
from 12 to 13 33 %
from 18 to 19 33 %
- required temperature of hot water 45 °C
- temperature of cold water inlet 15 °C
As a result of the long term performance prediction, we get:
- average required heat power Pd = Cl(Td-TCw)
- average system gain Pnet = Cl(Tl-Tcw)
- fractional system gain f=Pnet/PD
- solar efficiency As = Pnet/'Gc
Figure 5 shows the values for fractional system gain (f) and solar efficiency divided by the
collector aperture area (As/Ac).
i I As/Ac
A B C ICS
SOLAR SYSTEM
Fig. 5. Long term performance prediction of the measured systems
-------
1384
CONCLUSION
The practical application of the dynamic method by which we tested four simple SDHW
systems confirmed all the advantages of this method: simple and short term measurements,
automatic calculations, long term performance prediction and a sufficiently precise system
model, despite the small number of measured variables, naturally assuming that there is
sufficient variability in meteorological conditions during measurements. The short term
measurements are particularly useful in verifying the suitability of system models.
NOMENCLATURE
Ac
(n2?
Collector aperture area
Ac*
(«o
Effective collector area
As
(n^)
Solar efficiency
c
(W)
Fit error
CL
(W/K)
Load capacitance rate
Cs
(MJ/K)
Thermal capacity of the heat storage
dl
0
Mixing constant
f
(-)
Fractional system gain
Fr
(W/m2)
Heat removal factor of the collector loop
Gc
Solar irradiation in the collector plane
PD
(W)
Demanded heat power
Pnet
(W)
Net system power
Sc
(W)
Stratification parameter
Ta
(°C)
Ambient temperature
Tew
(°C)
Cold water temperature
td
(°C)
Temperature demanded by the user
Tl
(°C)
Temperature of the water delivered to the user
uc
(W/m K) Effective collector loss coefficient
Us
(W/K)
Total store heat loss coefficient
VL
m
Daily load draw off volume
ta
(-)
Effective transmission-absorption product
REFERENCES
Arkar, C. (1990). Kakovost solarnih sistemov. Diploma work. University of Ljubljana,
Ljubljana
Arkar, C. (1991). Dinamicna metoda za testiranje sistemov za pripravo tople potrosne vode.
Sunceva energija, Accepted publication
Spirkl, W. (1989). Dynamische Vermesung von Solaranlagen zur Warmwasserbereitung.
Disertation, L.-M. University, Munich
Spirkl, W. (1990a). Dynamic SDHW system testing. Program manual, L.-M. University,
Munich
Spirkl, W. (1990b). Dynamic solar domestic hot water testing. Transactions of the ASME.
J. S. En. Enginee., 112,98-101
-------
1385
OPTIMIZATION OF THE PRIMARY CIRCUIT OF SDHW SYSTEMS
A.C. de Geus, H. Visser, G.A.H. van Amerongen
TNO-Building and Construction Research
P.O. Box 29
2600 AA Delft
The Netherlands
ABSTRACT
Primary circuits of SDHW can be optimized with respect to dimensions, material contents,
control and flowrate. This paper gives results for optimizing the pump and control for
typical Dutch SDHW systems. The single pass principle plays an important role in this
optimization. The collector flowrate for a number of controlstrategies was investigated. A
low flowrate makes the pump on/off switching criterium less critical. A mantle heat
exchanger is well suited for a low flowrate and in combination with pump control based
on radiation intensity this gives good thermal performance for the SDHW. Besides
improvement in the thermal performance,low flow has other advantages, among others,less
critical control, low pump energy, etc. However with respect to the installation technique,
some disadvantages appear, such as performance,critical to flowrate ( i.e. obstructions),
difficult to maintain the drainback principle,etc. The commercial introduction in The
Netherlands depends much upon a suitable pump (low energy consumption, high head,
low flowrate). In a later stage the system can be optimized for low flow (absorber, piping,
heatexchanger) with respect to installation technique and the manufacturing techniques.
KEYWORDS
Single pass, low flowrate, pump energy, control, radiation switch, mantle heatexchanger,
solar domestic hotwater system.
INTRODUCTION
Nowadays Dutch research and development in the field of active solar applications is
mainly focused on the Solar Domestic Hotwater Systems (SDHW). The goal is to
introduce commercial attractive systems at a big scale after 1994, resulting in 300,000
installed systems in the year 2010. In order to achieve this goal,R&D activities are carried
out in three fields. In the first place 'new' concepts are under investigation with respect to
possible future cost reduction (Visser 1991). Secondly productdevelopment has to result in
-------
1386
designs optimized for mass production. In the third place investigations are going on to
improve the overall performance of SDHW systems. This paper will deal with this last
item.
STATE OF THE ART
Since van Koppen (1979) introduced the low flow concept, or preferable known as the
single pass concept, a lot of work has been carried out to fully understand and introduce
this concept. Discussion is still going on for the benefits of the single pass concept for
solar domestic hotwater systems. So far the commercial introduction of single pass
systems in the Netherlands has not started.
This is mainly due to the fact that the water distribution companies are very restricted to
the use of water/glycol mixtures. Otherwise the Canadian (lifeline) or Danish low flow
concepts were already introduced. In this case a double wall heat exchanger is mandatory,
which destroys the possible performance for single pass systems. All Dutch systems are
drainback systems (see fig. 1.)
Fig. 1. Typical Dutch drainback SDHW system. The mantle heat exchanger is also the
drainback tank.
One of the possibly advantages of the single pass principle is the increase in thermal
performance. Van Koppen (1979) showed an improvement of about 3% per (W/Km2)
collector heatloss compared to a fully mixed storage. As a rule of the thumb the optimal
collector flow is:
control
4,
1.2 * tankvolume
(ltr/hi .m2)
(1)
collectoraiea * 8
-------
1387
This relation is based on a clear day ( 8 hours high solar irradiation).
Note that this formula can only be applied if the storage tank volume has about the size
of the daily draw-off volume. For typical Dutch SDHW this optimal flow is about 30 1/hr.
Besides the improvement in solar performance,single pass has other advantages and
disadvantages. In this paper these other aspects will be mentioned. The accent will be on
low flow because this is easier to realize than single pass and less critical.
Marketed systems in the Netherlands are mostly equipped with temperature difference
controllers and low power central heating pumps (50 W). A recently finished project (Van
Amerongen, 1991) showed that controllers of this type have low accuracies partly
caused by the measuring technique. Combined to the rather high primary flowrate due to
the applied pumps, most systems show poor control actions. For instance this can result in
frequent and unwanted on/off switching of the pump.
On the other hand the flowrate is also sub-optimal with respect to the thermal
performance. Moreover the electrical energy consumption by the pump is not neglectable
compared to the thermal output of a SDHW. A development going on in The Netherlands
is based on electronically reducing the pump speed and thus the power consumption. This
technique is, however,limited with respect to the pump speed.
The above mentioned reasons are the basis for a coherent R&D programme in order to
realize within a short time commercially available single pass systems. Moreover within
the TASK 14 of the IEA Solar Heating and Cooling Programme, the development of
single pass SDHW system is a important activity for 1991/1992.
R&D PROGRAMME
The R&D programme comprises the following activities:
* Selecting, testing and improving of promising low flow pumps with respect to
flow characteristics, temperature resistant, efficiency, etc.
* Improving the pump controller.
* Adapting both the collector absorber and primary heatexchanger to a low primary
flow. Calculations and measurements support the development of optimized
components.
A first part of the programme is carried out. For 1991 the programme is concentrated on
realising a single pass system based on the nowadays available absorber and
heatexchangers.
Secondly in 1991 heatexchangers and absorbers will be investigated in order to make a
plan of requirements for applying low flow.
The planning for 1992 is to implement an optimized absorber and heatexchanger in the
system. The basis for this optimizations is,among other,the work of Hollands (1989).
Moreover this activity will also be carried out within the framework of IEA Task 14,
Advanced solar systems.
RESULTS
During 1990 the pump and control strategy were investigated. This resulted in conclusions
and new research activities for 1991.
-------
1388
Control strategy
For optimizing the control in combination with the flowrate a study was carried out. In
this study two control systems were taken into account. One is the temperature difference
control in which two temperature sensors are used; one in the storage tank,the other on the
collector. Subject for the study was the on/off temperature difference criteria and the
location of the sensor in the store. The flowrate was the third variable.
The other control system is a switching the pump on/off only based on the radiation
intensity (a simple radiation switch). If the intensity is above a certain level the pump will
be switched on, otherwise the pump will be switched off.
In fig. 2. the thermal performance of a SDHW is given for three situations; the delta T
control with sensor located on the bottom of the tank and in the middle of the tank, a
radiation sensor switching at 150 W/m2.
Form this figure it is clear that the best control is the delta T with the sensor at the bottom
of the store. The flow optimum is for all three situation about the same, but for the
radiation control this optimum is less stressed.
| Solar Qaln as function of flowrate and aansortype
4300
4200
'¦5
11
J=
« 4100
i
4000
3800
0 20 40 60 80 100
CollMtor (LOW (L/HR) *"
a T-aenaor A T-aenaor 4 Radiation
middle bottom sensor
Fig. 2. The thermal performance as function of collector flow and controlsystem.
Apparently a radiation control does not destroy stored energy in situations when the pump
is running and the store is at high temperature. This is caused by the mantle
heatexchanger. The return flow searches for the right temperature level in the
heatexchanger. Calculations showed that running the pump continuously only reduces the
thermal performance slightly.
Based on the results the influence of the accuracy of the radiation sensor is investigated.
In fig. 3. the number of pump running hours is presented as function of the intensity level
of the switch.
One can see that an inaccuracy of ± 50 W/m2 results in 2000 ± 200 pump running hours.
Application of a low power pump ( 10W) makes this inaccuracy unimportant.
i
_L
-------
1389
Cumulative distribution aolar radiation
eouth 45 dograa
£pread numbar ol pump hours
dua to Inaccuracy radiation aantor
100 300 900 TOO 800 1100
Solar radiation (ix) (W/mS)
Fig. 3. Accuracy of the radiation switch.
Pump
Based on the Dutch SDHW a programme of requirements for low flow pumps is made.
This looks as follows:
* low flowrate , high head (15 meter water column)
* low power consumption (less then 10 W)
* drain back via the pump must be possible
* low price (less then US $90)
* durable: temperature resistant up to 85 °C, low noise etc.
So far 5 pumps are selected. Four of these pumps are working on 12 V DC, which makes
adjusting the speed easy.
During the spring and summer 1991 the selected pumps will be tested.
ON-GOING RESEARCH
At this moment in the Netherlands research is and will be carried out on the heat
exchangers. Although the mantle heatexchanger has a very good performance the
durability can cause problems. The main problem with respect to the durability is the
corrosion on the welding partly caused by the drain back principle (open systems), by the
long joints and the high lime content of the Dutch water.
For this reason also the helix type heatexchanger is further investigated. At first by means
of computer simulations, later also by practical experiments.
For the optimization of the absorber the results of Hollands (1989) will be used and
"translated" to the Dutch situation. First results will be ready during summer 1991.
The low flow concept makes also the use of other piping possible. The necessary drain
back gives restrictions to the tube diameter and therefore the installation design is an
important aspect. Good drainback must be possible to avoid freezing problems.
-------
1390
DISCUSSION
Single pass systems have a number of advantages such as:
* higher thermal performance
* enabling thermal stratification
* less critical control
* possibly low pump energy
The discussion is always focusing on the higher thermal performance, which is mainly
caused by the stimulation of the thermal stratification caused by the low flow.
Not all SDHW concepts enabling stratification and therefore no improvement of the
thermal performance can be expected by applying single pass to these systems.
This is confirmed by a number of studies especially by Furbo (1989).
The single pass concept also has some disadvantages:
* the fluid dynamics
* high temperature raise over the collector
* drain back principle
* accuracy of the installation technique
If there is for some reason a small extra resistance,the flow rate will become lower,
resulting in lower thermal performance.
Moreover the influence of variation in the daily draw-off hotwater volume on the perfor-
mance with low flow systems is not clear.
The durability is an important aspect.
Single pass systems can become marketable when a good low flow pump and control is
found. Hereafter the installation technique with respect to piping, absorber and
heatexchanger must show that possible problems are solved.
ACKNOWLEDGEMENT
The research activities are made possible by financial means from the Dutch
Ministry of Economic Affairs in the framework of the National Solar Research
Programme, which is managed by the NOVEM, the Netherlands Management
Office for Energy and the Environment.
REFERENCES
Van Koppen, C.W.J., J.P. Simon Thomas, W.B. Veldkamp (1979). The actual benefits of
thermally stratified storage in a small and medium sized solar system. Proceedings ISES
Silver Jubilee Congress Atlanta.
Hollands.K.G.T., A.P. Brunger and P.G. Charalambous (1989). Optimization of the flat
plate absorber for low flow. Proceedings Solar World Congress Kobe.
Furbo, S. ( 1989). Solar Water Heating systems using low flow rates. Thermal Insulation
Laboratory. Technical University of Denmark. Report no. 89-9. August 198':.
Van der Linden, J., A.J. Dalhuijsen and E.M. Keizer-Boogh (1989). Optimalisatie
zonneboilers met gelaagde opslag voor huishoudelijk gebruik. TNO-Institute of Applied
Physics Report no. 714015. March 1989 (in Dutch).
-------
1391
USING A PLUG FLOW MODEL FOR SHORT TERM TESTING OF
SOLAR DOMESTIC HOT WATER SYSTEMS
W. Spirkl and J. Muschaweck
Sektion Physik, Ludwig-Maximilians-Universitat Munchen,
Amalienstr. 54, D 8000 Munchen 40, FRG
ABSTRACT
For modelling Solar Domestic Hot Water systems (SDHW), a plug flow store model is pre-
sented. It is used in conjunction with the Dynamic System Testing algorithm (Spirkl, 1990a)
to predict the long term performance of SDHW systems from short term test data.
The basic property of a plug flow model is its capability of modelling drawoffs without any
mixing inside the store. As an extension, the model developed in this paper covers the range
between pure plug flow and full mixing. The degree of draw off mixing is characterized by a
continuous parameter instead of a (discrete) number of thermal nodes which actually is defined
neither for a plug flow model nor for a real system.
Less than ten model parameters characterize a system under test. It was found that the
parameters can be identified in short time (two to four weeks outdoors) for a large class of
systems, e.g. thermosyphon systems or Integrated Collector Store (ICS) systems, systems with
load side heat exchanger, auxiliary heater, and heat pipe collectors.
A summary of experimental results for different systems is given.
KEYWORDS
Dynamic System Testing, Performance Prediction, Plug Flow, Short Term Test, Solar Domes-
tic Hot Water Systems
INTRODUCTION
At the University of Munich a method was developed to predict the long term performance
of Solar Domestic Hot Water (SDHW) systems from a short term test. This method, called
Dynamic System Testing (DST) method (Spirkl, 1990a; Spirkl, 1990b), was adopted by the
International Energy Agency (IEA), and an IEA working group concerned solely with the DST
method, the Dynamic System Testing Group (DSTG), was established in 1989. This paper
describes the progress made within the DSTG, with special emphasis on the plug flow model
used in the DST method.
In (Spirkl, 1990a), an algorithm is presented which allows parameter identification from mea-
sured data using non-linear, dynamic models. Although the correlation model with two state
variables (store temperatures) described there was found to be sufficiently accurate in many
cases, a more detailed model had to be developed to include systems with integrated auxiliary
heater and/or good store stratification.
Therefore we decided to retain the DST method, but to replace the correlation model by a plug
-------
1392
Load outlet
Load inlet
Fig. 1: Drawoffin the Plug Flow Model
to vary only in one dimension (i.e. in vertical direction), and that any flow through the store
will just shift this temperature distribution - upwards for load flow, downwards for collector
loop flow (see Fig. 1).
In principle, a simulation program like TRNSYS (Klein, 1988), which also provides a plug
flow model, could be used. But from a parameter identification point of view, a SDHW system
modelled by TRNSYS has too many free parameters. Its set of parameters is degenerate and
it is therefore impossible to identify the parameters from short term test data.
In comparison to the TRNSYS model, we made simplifying assumptions; e.g. extra store losses
at the bottom and the top are summarized in a global loss coefficient.
On the other hand, the plug flow concept was extended to take into account mixing caused
by load flow. The drawofF mixing effect is modelled using a diffusion equation.
THE PLUG FLOW MODEL FOR SDHW SYSTEMS
The Store
The main assumptions on the store for the model presented here are listed below. For now,
we restrict ourselves to systems with a solar loop heat exchanger immersed at the bottom of
the store and no load side heat exchanger:
1. The state of the system at time t is characterized by the one-dimensional distribution
T(t,h) of the store temperature, with the normalized height h in the range [0,1], i.e.
horizontal temperature gradients are neglected.
This assumption implies neglecting of different efFects;such as convection caused by store
losses, as well as horizontal temperature gradients caused by the geometry of the solar
loop heat exchanger.
2. The cold water is injected at height h = 0, the load is withdrawn at height h—l. The
remaining water in the store is shifted by an according value (plug flow).
3. The collector power is brought in at height h = 0 and is transported to water above
by natural convection, see (5). This is equivalent to a heat exchanger with no vertical
extension immersed at the bottom of the store.
4. The auxiliary power is brought in at height h = 1 — Jaux and is transported to water
-------
1393
Normalized Outlet Temperature
1.0 — — — ~ -
0.6
0.0
0
1 1 1 i
'—1—'—1—'—1—1—I-
0.0
0.5 1.0 1.5
Normalized Load Volume
2.0
Fig. 2: Step response of the store to a change in the cold water temperature for different values
of Dl (0, 0.01, 0.1, 1).
5. Local convection makes sure that dT/dh > 0 holds.
6. Mixing or heat conduction do not occur for zero load capacitance rate (Cs — 0), except
for fulfilling the relation mentioned in (5).
7. Cold water mixing is modelled by a diffusion term:
dT
dt
±(d9^
dk I dh
(1)
It is assumed that the (time dependent) diffusion coefficient D is correlated with the
load:
D(t) = DlCs/Cs. (2)
Figure 2 shows the resulting step response for different values of the drawoff mixing
constant Dl- Note that, due to Eq. (2), the step response does not depend on Cs- For
Dl = 0, the drawoff accords to Fig. 1 (pure plug flow). For Dl > 0, each drawoff is
associated with mixing (extended plug flow).
8. The loss coefficient is equally distributed over the height (dUs/dh = 0).
9. The heat capacity is equally distributed over the height (dCs/dh = 0). The capacity does
not depend on the temperature (latent heat storage presumably cannot be modelled this
way).
The Collector Loop
The collector loop is modelled according to the Hottel-Whillier-Bliss equation (Duffie, 1983)
neglecting thermal capacitance and fluid flow. The collector loop power Pc for solar irradiance
Ic in collector plane, collector ambient temperature Tca and inlet temperature T is modelled
-------
1394
Pc = A*c[Ic-u*c(T-Tca)]+, where H+ = { q otheZl ' (3)
For a collector with net area Ac, absorptance-transmittance (ar), heat removal factor Fr and
specific loss coefficient uc, the model parameters A*c and u*c are given by:
A*c = Ac Fr (ar), u*c = (4)
The Partial Differential Equation
In the sequel, the plug flow model is formulated in a partial differential equation for the special
case of a heat exchanger immersed at the bottom of the store. The function T(t,h) is modelled
as the limit of the solutions of the following parabolic partial differential equations for e —> 0:
dT{t, h)
dt
= 8e(h)A*c[Ic-u*c(T-Tca)]+
+ 8t{h - Jaux)Pavx
- US(T-Tsa)
+
Cs
d
-f
dh
d
+
dh
H'S)
_adT\ dT\
e dh) dh) '
(5)
The meaning of the terms of the right side is: collector gain, auxiliary power, store losses, plug
flow, diffusion and convection, respectively. Here, Tca and Tsa denote ambient temperatures
of the collector and the store, respectively; Paux denotes the auxiliary power. The model
parameters Us and C's represent the total loss coefficient and the thermal capacity of the
store.
For e —~ 0, the function Sc converges to a Dirac distribution at zero:
( e~x/e
St(x) - < ~T— for x > 0 (6)
[ 0 otherwise
Convection is modelled by a diffusion process with a diffusion coefficient depending on the
temperature gradient (a and b are arbitrary positive constants). This makes sure that in the
limes e -t 0 the requirement dT/dh > 0 is fulfilled exactly. Furthermore, with e —* 0 the heat
exchanger, the cold water inlet and the auxiliary heater are modelled with zero extension.
We now go beyond the restrictions made at the beginning of this section:
Load side heat exchanger
A load side heat exchanger is characterized by its thermal resistance Rl- The capacitance rate
Cs is reduced by the factor w(RlCs), '-p(x) — (1 — e~x)/x, see Fig. 3.
-------
1395
C's Cs
Fig. 3: Modelling of a load side heat exchanger. Within the model, the heat exchanger is
replaced by a mixing valve outside the store. The capacitance rate through the store is C's ~
Cs o
Effective collector area.
u*c
[ Wm-2K-1]
> o
Effective collector loss coefficient.
uv
[Jm-3K->]
> o
Wind speed dependence of Uq.
Us
[WK-1]
> o
Total store heat loss coefficient.
Cs
[MJ/K]
> o
Total store heat capacity.
Iaux
[-]
e]o,i]
Fraction of the store volume used for auxiliary heating.
dl
[-]
> o
Mixing constant, describing mixing effects during cold
water inlet (Dl = 0 for no mixing).
Sc
["]
> o
Stratification parameter, Sc = 0 is equivalent to a heat
exchanger immersed at the bottom.
Rl
[K/kW]
> o
Thermal resistance of the load side heat exchanger (if
any). A value of Rl = 0 is equivalent to no load side
heat exchanger.
External collector loop heat exchanger
An external heat exchanger is modelled using an additional parameter Sc by a direct collector
loop with fixed capacitance rate Cc-
Cc = A*cu*c/(1 - e~Sc), Sc > 0
(?)
The height of the inlet is chosen such as to match the collector outlet and the according store
temperature. An immersed heat exchanger (Sc = 0) is equivalent to Cc = oo.
Table 1 shows the complete list of model parameters to be determined in a test.
-------
1396
EXPERIMENTAL EXPERIENCE
The model was applied experimentally to about 50 different systems in national (Spirkl, 1990c;
VELS, 1988-1990) and international (de Geus, 1989) projects. The class of systems included
e.g. thermosyphon systems, Integrated Collector Store (ICS) systems, systems with load side
heat exchanger, auxiliary heater, and heat pipe collectors. The main results are:
• It is possible to describe systems whose design differs from the model assumptions. E.g.
an ICS system is described by lumping together collector losses and store losses.
• The model is accurate to about 0.5 K in the thermal output.
• The prediction error is about 1 K, corresponding to an error of about 3% referred to
35 K load temperature difference.
CONCLUSIONS
For the considered class of systems, with collector area less than 10 m2 and store volume
less than 10001, the model presented here is found to be adequate for the use in parameter
identification from short term tests.
ACKNOWLEDGEMENTS
Thanks are due to Prof. R. Sizmann for helpful suggestions, and to the participants of the
DSTG group for testing the DST method program package (Spirkl, 1990b).
This work was supported by the Bundesministerium fur Forschung und Technologie, grant
numbers 03E8101B, 032 8768 A and 03E8101C.
REFERENCES
de Geus, A.C. and E.M. Keizer-Boogh (1989). The dynamic systems testing group. Working
paper nr. 23 of the IEA (International Energy Agency) Working Group DSTG.
DufBe, J.A. and W.A. Beckman (1983). Solar Engineering of Thermal Processes. John Wiley
& Sons, New York.
Klein, S.A., W.A. Beckman, and P.I. Cooper (1988). TRNSYS: A Transient System Simulation
Program, Version 12.2. Solar Energy Laboratory, Madison Wisconsin.
Spirkl, W. (1990a). Dynamic SDHW testing. J. of Solar Energy Eng., Transactions of the
ASME, 112:98-101.
Spirkl, W. (1990b). Dynamic SDHW Testing Program, Manual (Version 1.16). University of
Munich (LMU), FRG. Available from DIN, Postfach 1107, D-1000 Berlin 30, FRG.
Spirkl, W. (1990c). Dynamische Vermessung von Solaranlagen zur Warmwasserbereitung.
Fortschrittsberichte der VDI-Zeitschriften, Reihe 6: Energietechnik, Nr. 241, VDI-Verlag,
Diisseldorf (W-Germany). ISBN 3-18-14 4106-6.
VELS (1988-1990). Verbundforschung zur Ermittlung der Leistungsfahigkeit von Solaranlagen
(VELS). BMFT-Projekt Nr. 0328768A.
-------
1397
PERFORMANCE OF ONE-DIMENSIONAL MODELS FOR STRATIFIED THERMAL
STORAGE TANKS
E. M. Kleinbach, W. A. Beckman and S. A. Klein
Solar Energy Lab, University of Wisconsin-Madison
Madison, WI
ABSTRACT
A study of the TRNSYS (Klein and others, 1990) tank models has been carried out using
experimental data. The results are discussed and recommendations are given as to which tank
model should be used under which conditions.
KEYWORDS
TRNSYS, Stratification, Domestic hot water, Storage tank models
INTRODUCTION
A number of models have been developed to account for thermal stratification in hot liquid storage
tanks e.g., (Duffie and Beckman, 1980; Sharp and Loehrke, 1979; Phillips and Dave, 1982;
Klein, 1976; Kuhn, et al, 1980; Phillips and Pate, 1977; Zurigat, et al, 1989; Pate, 1977). The
levels of sophistication with which these models were developed is quite different. The simpler
and therefore computationally less expensive models are suitable for simulating annual
performance. The more detailed models are computationally demanding but can shed considerable
light on the phenomena inside the tank, such as temperature and velocity distributions and thus lead
to better tank designs.
In the multi-node approach (Klein and others, 1990 and 1976), the tank is modeled as N fully
mixed volume segments (nodes) resulting in N first order ordinary differential equations. The
degree of stratification is determined by the choice of N; higher values of N result in more
stratification. A maximum number of 15 nodes can be chosen in the TRNSYS implementation.
For the special case of N=1 the tank is modeled as a fully mixed tank and no stratification effects
are possible. Unequally sized nodes can be specified. The model provides the option of
specifying fixed or variable inlet positions. The mains water enters at the bottom of the tank. At
the end of the time step, any temperature inversions are eliminated by mixing of appropriate nodes.
For variable inlet positions, the flows enter the nodes that are closest in density (and therefore
temperature) and no temperature inversions are created.
The plug flow model (Klein and others, 1990; Kuhn and others, 1980) simulates the behavior of a
temperature-stratified storage tank using a variable number of variable size segments. The number
of segments and their volumes actually empioyed cannot be controlled but vary depending
primarily on the tank volume, the net (heat source plus load) flow and the simulation time step.
The maximum number of segments in the TRNSYS implementation is 50. This upper bound is
-------
1398
maintained by merging of small segments, if necessary. The segments of liquid are assumed to
move through the tank in plug flow. The net shift of the initial profile is equal to the difference
between the total heat source volume and load volume. The segments and/or fraction of segments
whose positions fall outside the bounds of the tank are returned to the heat source and load. The
model provides the option of fixed or variable inlet positions. The incoming fluid mixes with
existing segments if its temperature is within 0.5 C, otherwise a new segment is created. The load
flow enters at the bottom of the tank.
In the late afternoon or during cloudy periods when the availability of solar energy has decreased
and the top portion of the tank is still hot as a result of higher energy input earlier in the day, the
temperature of the incoming fluid may be cooler than the upper portion of the tank. As a result, a
downward directed buoyancy force will drive the incoming fluid down into the tank and because of
its turbulent motion and viscosity, the hot fluid in the tank will be entrained in the falling plume.
Thus, the incoming stream is heated and it will fall down to the position in the tank where its
density and therefore temperature matches that of the tank. This phenomena is known as "plume
entrainment" and it will decrease the degree of stratification in the tank. The mathematical model of
plume entrainment, developed by Phillips and Pate (1977) and extended by Lightstone (1987), is
built into both the plug-flow and multi-node models in TRNSYS Version 13.1. In the following
sections, the multi-node and plug-flow models with and without plume entrainment will be
compared to experimental results.
EXPERIMENTAL DATA
The data for the low flow system were taken at University of Kingston, Kingston, Ontario
(Cataford and Harrison, 1990). The experimental apparatus consisted of a tank and a simulated
collector. The solar collector array was experimentally simulated using a conventional thermal heat
source as described in ASHRAE standard 95-1981. The collector area was 2.9 and the values
of Fr(toc) and FrUl were 0.743 and 4.54 W/m^ C, respectively, at a flow rate of 72 kg/h. The
collector loop heater was adjusted every minute to deliver energy equal to that of the specified
collector operating under a daily irradiance of 12 MJ/day and the measured heat source return
temperature. The daily irradiance profile on the collector surface was simulated as a sinusoidal
profile between hours 7 and 17 of the day. Effects of different flow rates were accounted for by
adjusting the collector heat removal factor, Fr. The collector loop pump was turned on at hour 7
and turned off when the rate of the useful energy gain for the collector dropped below zero in the
late afternoon. A storage tank with volume equal to 180 liters and length to diameter ratio of 1.84
was used. The UA-value was determined to be 4.57 W/C from a tank cool-down test. Water was
drawn at a flow rate of 6 liters/min from the system. Tests for various collector flow rates and load
profiles were performed. Successive test days were repeated until the system was determined to be
functioning in "steady-periodic" state. This state was identified when the daily delivered energy
was found to vary less than 3 % between test days. The data acquisition and control system
supervised load draws, calculated thermal performance (delivered energy) and collected
temperature and volume flow rate data every 5 minutes. Experimental tank temperatures show
very definite stratification.
The data for the high flow system were taken at Colorado State University, Fort Collins, Colorado
(Carlson, 1990) using a drain-back solar domestic how water system with a solar storage tank and
an auxiliary tank. The experimental test procedure consisted of a test in which the solar radiation
and water load profile, ambient temperature (22 ± 2 C), mains water temperature (22 ± 1 C) and
hot water set point temperature (> 48.9 C) were specified. The tests were completed at the end of
four days or when the daily value of the added auxiliary energy was within 3 % of the previous
day's value, whichever came first. The solar storage tank was preheated to about 40 °C at the
beginning of the tests to achieve faster convergence. The tests start at hour 17 of the day. Solar
radiation input was simulated with an electric boiler downstream of the collector array, located in a
-------
1399
constant temperature dark room. The boiler input was controlled according to an hourly profile
specified by the Solar Rating and Certification Corporation (SRCC, 1984) and calculations
outlined in ASHRAE standard 95-1987. The energy input occurred between hour 8 and 17 erf the
day. The daily total radiation was 17.03 MJ/m^ . The collector parameters, Fr(t
-------
1400
about as-good as a 3 node tank model.
The results for QI, QD and P for the high flow system are generally different than the low flow
systems. The fully mixed tank model leads to an underprediction in the experimentally measured
energy quantities but less underprediction than for the low flow case. The plug flow models
overpredict the energy quantities. Two to four nodes works well for the multi-node models with
fixed or variable inlets. Four nodes gives the best result for for the multi-node model with plume
entrainment. An increase in the number of nodes results in an overprediction of the energy
quantities for the multi-node models.
Since the performance of the multi-node models depends on the number of nodes chosen, a
relationship between the number of nodes and the conditions under which the tank operates is
needed. The quantities that were varied significantly during the experiments or simulations are the
values of the heat source flow rate, the load draw profiles (only for the low flow system), the
collector area and the collector flow rate (the latter two quantities only for the high flow system).
The variation of the collector area and the collector flow rate have little influence on the results
obtained for a multi-node model with a particular number of nodes. Therefore, the number of
nodes to be used was related to the mean number of tank turnovers, T, the sum of the daily load
and collector flow divided by the tank mass.
PF - PLUG PLOW, MN - MULH NODE, PE - FLUME ENTRAINMENT
H - FIXED INLETS, VI - VARIABLE INLETS, N - NUMBER OF NODES (QI /QD)
a
t«<
o
Q
o
H CP
Fig. 1 Results for QI and QD for the Low Flow Test #2
PF - PLUG FLOW, MN - MULU NODE, PE - PLUME ENTRAINMENT
FI-FIXED INLETS, VI-VARIABLE INLBTS.N-NUMBER OF NODES
0.05 I ¦ I l I I I ¦ I I I I I I I l I I 1 ¦ I I 11 I I 11 I I 111111 11 11 11 L I 11 I I L I
Fig. 2 Results for P for the Low Flow Test #2
-------
1401
Equations 2 and 3 relate the recommended number of nodes to the tank turnovers
variable inlets respectively.
Nflxed = 45.8-rL218
vr _„1t-0.966
^variable - 23.IT
The recommended number of nodes is the smallest possible number of nodes for which the relative
errors in the energy quantities QI and QD do not exceed 5 % and the performance number P is
within 10 % of the best value of P, Pbest- The criteria were established with two considerations.
First, the number of nodes which give the best results for QI, QD and P was often found to be
fifteen for low flow rates indicating that the performance numbers might be improved if the number
of nodes could have been further increased. From this point of view, fifteen nodes is an
"artificially" introduced maximum number of nodes. By introducing the above 5 and 10 % criteria,
it was possible in eleven out of seventeen cases to avoid the "artificial" limit of fifteen nodes. The
errors in predicting the energy quantities when using the number of nodes as proposed by the fitted
curves showed not more than 13 % deviation from the experimental results
The computational efficiency depends strongly on the number of nodes chosen. A large number of
nodes requires more CPU time. The multi-node model with fixed inlets results in higher values for
the appropriate number of nodes than the multi-node model with variable inlets.
The performance of the plug flow models and the multi-node models depends on the chosen
simulation time step. To investigate this dependence, simulations with a simple system consisting
of a collector (2.9 m2) and a tank (180 liters) were performed. Hourly radiation and load profiles
were specified and time steps ranging from 1 to 60 minutes were used. The values of the delivered
energy were taken as the criterion for the time step dependence. For the collector flow rate of 20
kg/h all the models, except the plug flow model with plume entrainment, exhibit a change in
delivered energy with respect to the value for the time step of 60 minutes of less than 1.5 %. The
value of delivered energy for the plug flow model with plume entrainment changes by 8 %. For
the collector flow rate of 180 kg/h all the models, except the plug flow model with variable inlet,
show a change of less than 3.3 %. For the plug flow model with variable inlet the relative change
in delivered energy is 9 %. In both cases, a plug flow model shows the greatest difference. This
difference is related to the change in the daily average number of segments employed as a function
of the simulation time step. The two models which exhibit the greatest change in the average
number of segments also have the largest time step dependence. This means that for different time
steps and therefore different average number of segments the plug flow model is an essentially
"different" model with a different number of nodes.
The models were also compared with respect to their computational efficiency. A TRNSYS deck
with three forcing functions, one integrator, one printer and the tank model under investigation was
used. Two hundred similar days for various heat source flow rates and time steps were simulated.
The heat source flow rate was operating eight hours per day. The temperature of the fluid from the
heat source was specified as hourly step profile with first rising and then falling temperatures in
order to force the variable inlet option and plume entrainment to be employed. An hourly load
profile was specified with four equal draws equally distributed during the time of the heat source
flow. Comparisons are obtained by subtracting the CPU time for a deck without tank model from
the CPU time for a deck including thfe tank model.
The plug flow models are faster for the higher heat source flow rate (200 kg/h) than for the lower
heat source flow rate (25 kg/h) since the higher heat source flow rate results in a smaller number of
nodes employed. The CPU times for the multi-node models depend strongly on the number of
nodes specified. The plug flow models are faster than the multi-node models. For the lower heat
source flow rate the plug flow model with variable inlets uses significantly more CPU time than the
for fixed and
(2)
(3)
-------
1402
other plug flow models due to a large number of segments involved in the algorithms for finding
the heat source flow inlet position. Including plume entrainment significantly increases the CPU-
time for the large simulation time step and the high heat source flow rate for the multi-node models.
The small increase in CPU time for the plug flow model including plume entrainment for a high
heat source flow rate and a large simulation time step is due to the fact that the number of segments
employed is very low (mostly one).
CONCLUSIONS
Several one-dimensional models for stratified thermal storage tanks have been investigated and
compared to experimental data for a wide range of conditions. The assumption of a uniform tank
temperature (instan taneous mixing) leads to a considerable underprediction of the energy input into
the tank and the delivered energy under all conditions considered. The developed relationships
between the number of nodes to be used and the mean number of tank turnovers are useful as a
guideline for choosing the most appropriate number of nodes under given operating conditions.
Use of the multi-node model with variable inlets is recommended. The plug flow models are
computationally more efficient than the multi-node models. However, the plug flow models with
fixed and variable inlets tend to overpredict the energy quantities. The plug flow model including
plume entrainment is recommended as an alternative for the multi-node with variable inlets for a
mean number of tank turnovers less than five. The reader is reminded that the results obtained
with the plug flow models depend on the simulation time step chosen, which introduces some
uncertainty in the results.
ACKNOWLEDGEMENTS
The authors wish to thank Queen's University, Kingston, Ontario and Colorado State University,
Fort Collins, Colorado for providing the experimental data which made this study possible.
REFERENCES
Carlson, W.T. (1990). MS Thesis, Colorado State Univ., Fort Collins, CO.
Cataford, R.J.. and S.J. Harrison (1990). Factors Affecting Storage Tank Stratification and the
Thermal Performance of SDHW Systems. SECI, Hallifax, Nova Scotia, Canada.
Duffie, J.A. and W.A. Beckman (1980). Solar Engineering of Thermal Processes. Wiley, New
York.
Hill, B.J. (1972). J. of Fluid Mechanics. 51. 773-779.
Klein, S.A. (1976). PhD Thesis, Chemical. Engineering, U. of Wisconsin-Madison.
Klein, S.A. and others (1990). TRNSYS 13.1 Users Manual. Report 38-13, Solar Energy Lab,
U. of Wisconsin-Madison.
Kuhn, J.K., G.F. VanFuchs and A.P. Zob (1980). Developing and Upgrading of Solar System
Thermal Energy Storage Simulation Models. Draft Report for DOE, Boeing Computer Services.
Lightstone, M. (1987). Mathematical Model of Plume Entrainment. Solar Thermal Research Lab,
U. of Waterloo, Ontario, Canada.
Pate, R.A. (1977). PhD Thesis, Utah State University.
Phillips, W.F. and R.N. Dave (1982). Solar Energy. 29. 111-120.
Phillips, W.F. and R.A. Pate (1977). Mass and Energy Transfer in a Hot Liquid Energy Storage
System. Proc ASES, Orlando, FL.
Schlichting, H. (1968). Boundary Laver Theory. 6th edition, McGraw-Hill, New York.
Sharp. M.K. and R.I. Loehike (1979). J. of Energy. 3 (2). 106-113.
Solar Rating and Certification Corp. (1984), Standard 200-82, Washington, D.C.
Zurigat, Y.H., K.J. Maloney and AJ. Ghajar (1989). Trans of ASME. 111. 204-210.
-------
1403
THERMAL BEHAVIOR OF A HEAT EXCHANGER COIL
IN A STRATIFIED STORAGE
J. van Berkel, W.B. Veltkamp and A.B. Schaap
LEVEL energy technology, Gascognehof 31,5627 KJ Eindhoven, The Netherlands
ABSTRACT
A top-bottom distributed helically wound coil is an attractive option for a stratified solar energy stor-
age. Literature survey provided relevant data covering natural and forced convection heat transfer in
and outside the coil. Numerical models, based on the experimental heat transfer data, yield the storage
and the solar system performance. The heat exchanger geometry is improved with respect to the ther-
mal performance.
KEYWORDS
Stratified storage, helical: heat exchanger coil, natural convection, DHW-system, heat transfer,
"single pass", 'low flow".
INTRODUCTION
In most solar Domestic Hot Water (DHW) systems, a heat exchanger separates the collector circuit
hydraulically from the mains water circuit to facilitate overheating- and freeze protection by collector
drain back or use of glycol, also protecting the collector circuit from corrosion and scaling by mains water.
The figure below shows two configurations using a helical heat exchanger coil.
storage vessel ^ storage vessel
collector Pre-heated DHW collector /y-T^^pre-heated DHW
mains water
mains water
pump
pump
Fig. 1. Solar DHW system configurations
In the configuration shown left, the power to be transferred by the coil equals the relatively low collec-
tor power output whereas in the configuration shown right the power to be transferred equals the higher
DHW-load. If, however, a DHW-coil is applied the storage vessel does not need to be mains water resis-
tant.
Main objective of the research is to gain insight in the thermal behavior and to provide design guide-
lines for both options. As natural convective heat transfer phenomena are temperature dependent,de-
tailed modelling is required. Ongoing research concentrates on thermally stratified storages.
Conservation of high temperature differences in a stratified storage showed to be beneficial for reducing
the heat exchanger size. This advantage adds up to the more generafyknown advantage of a stratified
storage (reduction of entropy production and therefore conservation of exergy).
-------
1404
EXPERIMENTAL HEAT TRANSFER DATA
Inside Heat Transfer
Like in a straight tube the flow in a helically coiled tube shows a laminar, a transition and a turbulent
flow region. Apparent discrepancy with the straight tube flow is the higher transition point from lami-
nar to turbulent flow caused by the stabilizing effect of centrifugal forces, Gnielinski (1986):
Hir:
Recrit = 2300
in which :
Recrit critical Reynolds number
d tube diameter
D coil diameter
(1)
H
[m]
[m]
The transition to a fully turbulent flow occurs at Re = 2.2 104.
In the laminar region the flow may be influenced by buoyancy- and centrifugal effects. Figure 2 shows
the influence of both effects on the flow field.
1 V2D
buoyancy effect» centrifugal effects buoyancy effect« centrifugal effects
Fig. 2. Buoyancy and centrifugal effects on the flow in a curved tube
Laminar Heat transfer inside a helically coiled tube is studied by Futagami and Aoyama (1988). The
Nusselt number Nu^ for flows in which buoyancy effects dominate :
fNubj45 = j + [0 19(Re Ra prp]4'5
\Nuo;
in which :
Nub Nusselt number for buoyancy effects
Nu0 Nusselt number for Poisseulle flow in a straight tube
Re Reynolds number
Ra Rayleigh number
Pr Prandtl number
(2)
The Nusselt number Nuc for flows in which centrifugal effects dominate :
(Nuc j6 = j + [0195 (dh pj.054)0-5]6
\Nuo;
in which :
Nuc Nusselt number for centrifugal convection
The dimensionless number Dn depicts the Dean number:
Dn = Re
(3)
[-]
(4)
-------
1405
For flows with buoyancy and centrifugal effects, the Nusselt number is an appropriate combination of (2)
and (3). Due to the chaotic flow patern, turbulent flow is appreciably less influenced by buoyancy- and
centrifugal effects. Gnielinski (1986) provides an equation similar to the Pethukhof-Popov equation, in
which the Nusselt number depends mainly on the Reynolds and Prandtl number.
Outside Heat Transfer
Experimental research concerning natural convective heat transfer in a stratified environment is con-
ducted by Eichhorn (1974). The temperature distribution is described in terms of a stratification index S :
S=aji
AT
in which :
S stratification index
a axial temperature gradient
At temperature difference
(5)
[-]
[Km"1]
[K]
The stratification index indicates the thermal level to which the plume rises. Figure 2 gives an impres-
sion of natural convection flow from a heated globe for several stratification indexes.
5 = 0,3
5 = 0,8
S>=1,5
S-2,2
S = c
Fig. 3. Natural convection heat transfer flows from a heated globe in
thermally stratified environment, Eichhorn (1974)
Rough experimental data showed a Nusselt number depending on the Rayleigh number and the stratifi-
cation index. The experimental research, however, did not yield comprehensive heat transfer data.
Natural convection heat transfer from a vertical array smooth tubes is studied by Marsters (1972).
Experiments showed that, for moderate Grashof numbers (1.2 10^) and a tube pitch larger than twice the
tube diameter the heat transfer from an upper tube is hardly influenced by a lower tube.
Henderson (1982) conducted thorough research on heat transfer data outside smooth or finned tubes in a
solar energy storage. For outside heat transfer of smooth tubes the data correlated :
Nu = 0.53 Ra0-25 (6)
For finned tubes:
Nus = 0.3365 (Ras^j0'285 (7)
in which
Nus fin distance based Nusselt number [-]
s fin spacing [m]
d fin diameter [m]
MODELLING OF HEAT STORAGE AND INTEGRATED HEAT EXCHANGER COIL
Based on the experimental heat transport data, a numerical segment model is developed. In the model,
the storage is thought to be divided into several temperature layers, each of them fully mixed.
-------
1406
The helically coiled heat exchanger is thought to be distributed over the entire height of the storage.
The heat exchanger inlet temperature in a layer equals the heat exchanger outlet temperature of the up-
stream layer. For simplicity, a perfect horizontal heat transfer is assumed, vertical heat transfer and
heat loss to the environment are not taken into account. As temperature and heat transfer coefficient de-
pend on each other, the computations have an iterative nature.
Step Charge Test Simulation
To gain insight, storage step charge- and discharge tests are simulated. As an example, the simulated
step charge test results are presented for the configuration with the heat exchanger in the collector cir-
cuit.
poor heat exchanger high heat exchanger
performance performance
outlet flow
Fig. 4. Step charge test
The dimensions of the storage and the heat exchanger selected are based on a Dutch solar storage design
comprising a cylindrical 0 450 x 730 mm, 100 liter storage vessel, and a 6 m 0 25 x 23 mm smooth stain-
less steel tube heat exchanger coil. The coil diameter amounts to 340 mm, the pitch equals 5-times the
tube diameter. An initial uniform storage temperature of 10 °C is assumed. At t =to the storage is
charged with a 80 °C supply flow. The 20 lh"1 "low" flow rate is selected according to the "single pass"
collector flow strategy which requires a daily collector throughput matching the daily DHW
consumption. The heal: transfer data are presented for a time interval of twice the characteristic charge-
time interval t* = 18000 s, the time in which the cumulative heat exchanger throughput equals the
storage volume. If the heat exchanger would be absent (or perfect), and the storage perfectly stratified,
the storage would be charged completely in the time interval t*. Computations showed that the flow is
laminar up to a flow rate of 300 lh"1. Figure 5 presents the heat transfer coefficient a in- and out-side
the heat exchanger coil. For graphical readability, the results are presented for a 5-layer model.
? 1200-
800
top layer
tf
top layer
1000
bottom layer
800
•g 600-
!
¦° 500-
600
400
400-
300
200
-I—I
36000
0
12000
24000
charge time [s]
12000
24000
Charge time [s]
36000
Fig. 5a,b. In- and outside heat transfer coefficients
The step-wise changes shown in the graph correspond with changes in flow regime caused by changes in
the temperature dependent fluid properties. The average heat transfer coefficients in- and outside the
tube are equivalent, indicating a well selected tube geometry.
-------
1407
As can be seen from fig. 5 both the in- and the outside heat transfer are governed by natural convection.
The heat transfer coefficients depending on the local temperature differences move through the storage
like a wave. Figure 6 shows the storage in- and outlet temperature.
Sgese t = 36000 s
inlet temperature
t = 18000 s
Outlet temperature
1 I 1 1 i 1 ' I 1 1 1 ' 1 I
12000 24000 36000
Top layer Layer number [-] Bottom layer charge time [s]
Fig. 6a,b. Storage-, in- and outlet temperatures
The gradually varying storage temperature over the height of the storage and the S-shaped variation
in time of the outlet temperature indicate that the storage is not charged perfectly stratified. With an
absent or perfect heat exchanger, the storage temperature would slide step-wise over the storage height,
whereas the outlet temperature would show a step-wise variation from T = Tiow to T = Thigh at t = t\
The thermal performance of the storage with integrated heat exchanger coil is indicated by the charge
fraction defined as the ratio of energy charged after a time interval t* compared to the maximum energy
content of the storage. Graphically the charge fraction corresponds to the area beneath the storage tem-
perature line at t = t* = 18000 s (Fig. 6a) in comparison with the total area beneath the 80 °C tempera-
ture line. The charge fraction for the 20 lh'l flow rate amounts to about 80 %. Perfectly stratified and
fully mixed storages without a heat exchanger yield charge fractions of 100- and 63 %, respectively. A
charge fraction of 100 % is hardly achievable with a heat exchanger. The more the storage is charged,
the smaller the temperature differences driving natural convection. This mechanism limits the charge
fraction, especially at high energy flow rates (the heat transfer coefficient does not vaiy proportionally
to the mass and energy flow rate through the heat exchanger tube). As expected, the simulated step test
at the "standard" flow rate of 200 Ih"1 (t* = 1800 s) shows a worse thermal performance, the charge frac-
tion being about 50 %.
Improvement of the ''low flow" Heat Exchanger Geometry
The charge fraction of the low flow heat exchanger is computed for various coil geometries (tube- and
coil diameter, heat exchange area). One of the dimensions is varied, while holding the other two con-
stant. Data are based on the 6 m 0 25 x 23 mm smooth tube collector circuit heat exchanger at the flow
rate of 20 lh"1. Computations showed an increasing charge fraction with smaller tube diameters.
Apparently the decreasing natural convection for smaller tubes is lower than the increasing forced con-
vection. Moreover, outside natural convection heat transfer roughly depends on d"®-^\ The charge fraction
showed a 6 % increase over an inner diameter range from 23 to 10 mm. Obviously the tube diameter is
limited by construction- (the smaller, the longer the tube) and pressure drop considerations. Provisional
calculations showed a 0.04 bar pressure drop (0,022 W power loss) for a 20 m long 0 8x6 mm tube at
20 lh'l. The charge fraction seems hardly sensitive to the the coil diameter, indicating that at 20 lh"^
centrifugal effects are weak. For the 0 25 x 23 mm tube, the charge fraction showed a 0.7 % increase over
a diameter range from 340 to 50 mm. With respect to an adequate horizontal distribution of heat the coil
diameter might be selected so that the storage vessel plan area is equally divided between the inside
and outside of the coil. Variation of the heat exchange area (by varying the tube length) showed an apt
heat exchange area of about 0.45 m^. Below this value, the charge fraction drops sharply, above this
value, the charge fraction increases weakly. Cost considerations will determine the optimum coil size.
-------
1408
Solar Energy System Simulation Results
For solar energy system performance computations several numerical models are added to the stor-
age/heat exchanger model. For simplicity the numerical model of the 2.6 m2 flat plate collector is
based on elementary collector theory (Duffie & Beckman,1980). Detailed modelling, required for 'low"
flow operation did not fit the research framework. Additional weather input data are derived from the
Dutch reference weather file. The daily 100 liter DHW usage is concentrated in the morning and
evening. An energy conservative mixing routine is added to the storage model to redistribute layer
temperatures in case a negative storage temperature gradient might occur.
Computations are executed for "standard" and "low flow" systems with a heat exchanger in the collector
circuit, the DHW circuit and, for reference purposes, for systems without any heat exchanger. Tentative
model computations indicate that the system with an improved heat exchanger coil in the 20 lh~^
collector circuit yields a solar fraction of 56.4 %, which is merely 2,4 % lower than the highest solar
fraction computed for the "low flow" system without any heat exchanger, indicating a well designed
heat exchanger geometry. Apparently a heat exchanger yields a higher thermal performance when
placed in the collector circuit. The charge time is longer, and the heat transfer rate is lower. Moreover,
deteriorating heat transfer is counteracted by an increased heat exchanger supply temperature, the
collector power output remaining nearly constant. The DHW coil encounters a serious penalty, as the
limited discharge fraction of the DHW coil limits the yearly solar fraction considerably. Moreover, the
heat exchanger supply temperature remains constant during the discharge mode. It is expected that the
thermal performance of the DHW coil can be improved easily by increasing the storage volume, avoid-
ing low temperature differences at the end of a discharge mode. Further research on this point is re-
quired.
CONCLUSIONS
1) Heat transfer coefficients of a storage vessel integrated heat exchanger vary considerably and can-
not be treated as constant
2) For a "low flow" solar DHW-system with a heat exchanger in the collector circuit,the design guide-
lines comprise:
w distribution of the heat exchanger over the entire height of the storage.
«• a tube diameter as small as pressure loss and construction considerations permits.
a coil diameter according to an equal storage vessel plan area in- and outside the coil,
a tube length which results in a heat exchange area of about 0.45 m2 per 1001 storage volume.
3) A solar system designed according to these guidejines yields a yearly solar fraction which is merely
a few percent lower than a system without any heat exchanger.
4) Further research is required on model validation and DHW-coil optimization
REFERENCES
Berkel, J. van (1991). Ontwerprichtlijnen van thermisch gelaagde zonneboilers met een spiraalvormige warmtewis-
selaar, (Design guide lines for thermally stratified solar DHW systems with a helically coiled heat exchanger),
level energy technology, Eindhoven.
Futagami, K., and Y. Aoyama (1988). Laminar heat transfer in a helically coiled tube, Int. journal of Heat and Mass
Transfer. Vol. 31, No. 2, pp 387-396.
Gnielinski, V. (1986). Heat Transfer and Pressure drop in Helically Coiled Tubes, Heat Transfer 1986. Proceedings of
the 8th International Heat Transfer Conference. San Francisco, 1986, pp. 2847-2854.
Henderson, J.B. and A.C. Caolo (1983). Optimization of'Radial Finned Tube Heat Exchangers for Use in Solar
Thermal Storage Systems, DOE/R1/25247-T1
Duffie, J. A., and W.A. Beckman (1980). Solar Engineering of Thermal Processes, Wiley & Sons, New York.
Jansen, L. (1976). Warmteoverdracht en Axiale Dispersie bij Laminaire Stroming in Schroefvormig Gekromde
Buizen, (Heat transfer an axial dispersion in Laminar flow inside helically coiled tubes). Dissertation Delft
University of Technology, The Netherlands.
Eichhorn, R., J.H. Lienhard and C.C. Chen (1974). Natural Convection from Isothermal Spheres and Cylinders
Immersed in a Stratified Fluid, 5th International Heat Transfer Conference. Tokyo, Proceedings NCI.3, Vol.
Ill, pp. 10-14.
Marsters, G.F. (1972). Arrays of heated horizontal cylinders in natural convection, Int. lournal of Heat and Mass
Transfer. Vol. 15, pp 921-933.
-------
2.8 Passive Domestic Hot Water
-------
-------
1411
ROOFTOP GREENHOUSE/SOLAR WATER HEATER:
AN URBAN RETROFIT FOR ENERGY SELF-RELIANCE
Gary R. Swindler
School of Architecture, College of Architecture and Environmental Design
Arizona State University, Tempe, Arizona 85287
ABSTRACT
The American urban population is described as contributing little to their own food production and
contributing most to the consumption of fossil fuel resources. A system is analyzed as to its
potential for reducing the non-renewable energy consumption, and rural food dependence, of the
urban population in America. The proposed energy system is a thermosyphon, fin-type solar water
heater and greenhouse to be investigated in a northern U.S. climate. The parameters of the system
simulation are described. The system is to be simulated using DOE-2. ID. Functional values are
defined for DOE-2 as they relate to fin-type performance and water heater - greenhouse heat transfer
dynamics. Results from the proposed simulation are described to supplant some of the non-
renewable energy consumption of multi-family apartment dwellers.
KEYWORDS
Greenhouses; computer simulation; water heaters; collector performance; energy conservation;
renewable energy resources; passive systems; agriculture; retrofits.
INTRODUCTION
Urban Americans consume an amount of non-renewable energy and food resources disproportionate
to their respective contribution. This is determined by acknowledging that urban Americans
comprise 4.6% of the world's population while consuming greater than 19% of the world's non-
renewable energy (U.S. Department of Commerce, 1989). Additionally, most of the nation's food
production is grown in the rural sector, which, along with the energy used in storage and
transportation to the consumer, accounts for 14% of the total energy consumption, mostly in fossil
fuels (Naar, 1990). This condition exists in an age when the burning of non-renewable resources
and the decline in the productive capacity of agricultural land are contributing steadily to the
deterioration of the global ecology (Meadows and colleagues, 1974; Brown and colleagues, 1989).
The provision of food and energy for urban apartment dwellers will be investigated through
computer simulation (DOE-2. ID) of the system schematically represented in Fig. 1 and Fig. 2. This
research will investigate the performance of the integrated greenhouse/water heater system that is
designed based on the climate and solar geometry of Chicago, Illinois. The water heater will be
simulated through polynomial expressions based on fin-and-tube performance equations (Lunde,
1980; Duffie and Beckman, 1974). The system elements to be investigated are best described by the
Greenhouse Solar Collector (Honarbakhsh, 1984) and the BigFin™ thermosyphon solar water heater
(Bliss, 1983).
Receding page blank
-------
water storage
^greenhouse vent
kZ / opaque roof
see schematic
v F'g- 2
glazing collector
Greenhouse
Thermal
Mass "
angle of
latitude +10'
^-Planters^^
Existing roof mass
Fig. 1. Schematic cross-section of rooftop greenhouse/solar collector system.
'Winter Sun Angle
1 and Radiation
'Winter*
Conditions
Louvered /,
Hot Water
Collector Fins
Greenhouse
Glazing
transmitted ( \ ^
radiation \
> absorbed
Greenhouse radiation
Interior ^
Fig. 2. Schematic of movable collector fins interior to the greenhouse glazing.
-------
1413
MODEL DESIGN
In regards to Fig. 1, the greenhouse retrofit to be analyzed is modeled to be of moderate to heavy
thermal mass and tight, double-glazed construction. The structure encompasses 140 m2 of roof area
and has a north wall with a mass of 392 kg/m2. There is assumed to be 620 kg/m2 interior
greenhouse mass (soil and planters). The glazing is of double-paned construction with insulated
aluminum/sheath framing, and the roof is angled at 50° to the south. The greenhouse/water heater is
assumed to be constructed on an existing roof of 115 kg/m2 mass serving a 6-story, 24-person
apartment building.
The fin-type water heater is modeled as an interior shading device adjustable to 50°, 30° and 10°
based on the time of year, and controlled by a shading schedule. As seen in Fig. 2, this louver is
modeled to both track the sun and provide increased shading as the summer sun approaches. Based
on recommended sizing requirements for a fin-type system (Zomeworks), there is 65 m2 of collector
area and 5.0 m3 of insulated water storage. The domestic hot water load is 1590 liters per day with a
maximum demand of 265 liters/hour.
POLYNOMIAL DESCRIPTION OF WATER HEATER PERFORMANCE
In order to measure the performance of the fin-type collector, DOE-2.1D's accessory, functions, will
be used. This component opens a 'window' for the programmer to access and utilize the keywords
used by DOE-2's algorithms. For this research, global weather and solar data, as well as other
building performance keywords, will be integrated into polynomial equations for hourly calculations
of water heater performance and greenhouse temperature change.
The expressions to be used in this investigation, while based on Hottel-Whillier's equation for the
performance of flat-plate collectors, are gathered from several sources (Harris, Miller and Thomas,
1985; Lunde, 1980; Duffie and Beckman, 1974). The simulation equations are based on generalized
empirical data of tube-and-fin collectors with flat black surfaces working under double-glazed
coverings. Additionally, equations for radiative and convective heat transfer between th'e collector
and the interior greenhouse environment will be integrated into the collector performance function.
The efficiency equation to be used in this research, as provided by Lunde (1980), is:
Efficiency = Nf * (Kr«
-------
1414
fin efficiency (Nf) of this collector is .9920.1
The radiation incident on the collector, IG, will be calculated from equations as provided by Duffie
and Beckman (1974). The equations are as follows:
I0 = Idn * cos(0) + la + Ir (4)
where, Idn = direct normal radiation
la = diffuse radiation
Ir = reflected radiation\
cos(0) == cosine of the angle of incidence of beam radiation.
While the radiation values are global variables in DOE-2's solar data, cos(0) will be further calculated
by:
cos(0) = sin(3)*sin(l)*cos(s) - sin(3)*cos(l)*sin(s) + cos(3)*cos(l)*cos(s)*cos(w) (5)
where, 0 = angle of incidence of beam radiation
3 = declination of the sun = 23.45 * sin(360(284 + day of year)/(365)) (6)
s = slope of collector
1 = latitude of location (42°)
w = hour angle = 15 * solar time (morning = positive, afternoon = negative). (7)
The variables of Fr , PrUL and K will be based on empirical and assumed values as recommended
by Lunde (1980). For a double-glazed, flat plate aluminum absorber with flat black paint, Fr =
.747 and FjUl = 5.14 W/°C-m2. For the incident angle modifier, assuming a modifier constant of
-.15 (typical for black paint, double-glazed), the derivation is:
K = 1 - .15 * (1 + ((Iq/Idn) * ((l/cos(0)) -2)). (8)
Finally, the ambient temperature (Ta) will be supplied by DOE-2 simulation data, and the inlet fluid
temperature (Ti) will be assumed to be an average temperature of 32°C. This is based on the local
source water temperature (Chicago) of 11.67°C and a demand temperature of 60°C. As explained by
Lunde (1980), a time-averaged collector temperature can provide adequate performance estimates.
The resulting efficiency equation for the fin-type collector is therefore:
E = .9920 * (K * (.747) - ((5.14 * (32 - Ta))/I0)). (9)
TOTAL SYSTEM PERFORMANCE AND OUTPUT
In order to evaluate the water heater/greenhouse interaction, equations for radiative and convective
heat transfer will be used. First, the hourly efficiency (E) of the collector will be multiplied by the
hourly incident radiation (Iq) and the collector area (65 m2) to arrive at the total, static energy content
of the system, Q. From this will be subtracted the radiative and convective losses (gains) as obtained
through the following equations (Lunde, 1980):
Qr = 5.673 * (.981) * (65 m2) ((Tp/100)1'4 - (Tg/100)1/4) and, (10)
Qc = 10.22 W/°C-m2 * (Tp - Tg), (11)
where Tp and Tg are the underside plate temperature and greenhouse temperature, respectively. The
plate temperature will be derived by considering the total, static thermal mass energy of the aluminum/
JUL can be assumed to be an. average, a reasonable assumption for two-glass collectors. For black paint, double-glazed and
average collector temperatures of 40°C to 65°C, the total collector loss, Ul > would be 3.8 W/°C-m2 (Lunde, 1980).
-------
1415
copper/water system and solving for the temperature of the system. This is described by the
equation:
Q = mw*cw*t + mc*cc*t + ma*ca*t (12)
where, m* = mass of water, copper or aluminum (kg)
c* = specific heat of water, copper or aluminum (kj/kg-°C)
t = temperature of system (°C)
Q = static energy content of system.
The resulting energy produced, Qr = Q - Qr - Qc, will be taken as the hot water energy produced for
that hour, and totalled throughout the run period. The values of Qr + Qc will be added to the thermal
heat content of the greenhouse air with the resulting temperature change incorporated into the DOE-2
space temperature algorithm.
The performance of the greenhouse growth, or food production, will be based on experimental
results and recommendations described in several sources (Smith, Jurgrau and Farrer, 1981;
Deminet, 1976). Tomatoes will be used as the theoretical food. The two key variables of the DOE-
2 simulation will be temperature and light levels. For moderate to optimum tomato production,
temperatures between 13°C and 38°C, and light levels throughout most of the daylight hours of 2500
lux or greater will be required (Smith, Savage and Mills, 1984). For times when these parameters
are not maintained throughout the simulation period appropriate reductions will be applied to the
optimum greenhouse performance. In terms of other parameters for covered agricultural growth,
this research assumes that proper management of water, pests and natural fertilizers would prevail.
CONCLUSION
This research will take the year's performance of the system simulation and equate two, fossil-fuel
saving outputs: hot water and tomatoes. The hot water production will be described in terms of
kWhs of electricity saved. The tomato production will be described in terms of kilograms of
tomatoes produced. As food (e.g. tomatoes) embodies measurable fossil fuel energy inputs used in
production, transportation, storage and distribution (Pimentel, 1975)2, the mass production of
tomatoes in the urban environment will be equated directly to megajoules of energy saved. The total
system production will be equated to lowered fossil fuel consumption by the American urban
population.
This research is an on-going thesis project of the author and is intended to be completed by the
Summer of 1991.
REFERENCES
Bliss, Steve (1983). Tapping the greenhouse for hot water. Solar Age, 8:12, 51-52.
Brown, Lester R., Alan Durning, Christopher Flavin, Lori Heise, Jodi Jacobson, Sandra Postel,
Michael Renner, and Cynthia Pollock-Shea (1989). State of the World 1989: a WorldWatch
Institute Report on Progress Toward a Sustainable Society. W.W. Norton, New York.
Deminet, Czeslaw (1976). Glass solar collectors for greenhouses and integrated greenhouse-
residential systems. Proceedings Solar Energy Food and Fuel Workshop, Tucson, Arizona.
Duffie, John A. and William A. Beckman (1974). Solar Energy Thermal Processes. John Wiley &
Sons, New York.
Green, Maurice B. (1978). Eating Oil: Energy Use in Food Production. Westview Press, Boulder,
Colorado.
^Based on preliminary estimates, each kilogram of tomatoes grown in the open field in the U.S., harvested, transported
to the urban market, stored and delivered to the average American household requires 1230 kilocalories of fossil fuel
input (Pimentel and Pimentel, 1979). This does not include the estimated 27% of the tomato crop wasted throughout
the supply system (Green, 1978).
-------
1416
Harris, Norman C., Cydney E. Miller, and Irving E. Thomas (1985). Solar Energy Systems
Design. John Wiley and Sons, New York.
Honarbakhsh, Ahmad (1984). Dynamic modelling of solar thermal devices. Ph.D. diss., Chalmers
University of Technology, Gotenburg, Sweden.
Lunde, Peter J. (1980). Solar Thermal Engineering: Space Heating and Hot Water Systems. John
Wiley & Sons, New York.
Meadows, Donella H., Dennis L. Meadows, J0rgen Randers, and William W. Behrens, HI (1974).
The Limits to Growth: A Report for the Club of Rome's Project on the Predicament of Mankind.
Universe Books, New York.
Naar, Jon (1990). Design for a Livable Planet: How You Can Help Clean Up the Environment.
Harper & Row, New York.
Pimentel, David and Marcia Pimentel (1979). Food, Energy and Society. John Wiley & Sons,
New York.
Pimentel, David (1975). World food, energy, man and environment. In William J. Jewell (Ed.),
Energy, Agriculture and Waste Management: Proceedings of the 1975 Cornell Agricultural Waste
Management Conference. Ann Arbor Science Publishers, Inc., Ann Arbor, Michigan.
Smith, C. C., M. Jurgrau, and R. G. Fairer (1981) . Solar heating conservation in a residential
attached greenhouse. Sunworld, 5:5, 152-158.
Smith, I. E., M.J. Savage, and P. Mills (1984). Shading effects on greenhouse tomatoes and
cucumbers. Proceedings 3rd International Symposium on Energy in Protected Cultivation, 2,
Columbus, Ohio.
United State Department of Congress (1989). Statistical Abstract of the United States. Government
Printing Office, Washington, D.C.
Zomeworks Corporation. Passive solar energy products' specifications guide. P.O. Box 25805,
Albuquerque, NM 87125.
-------
1417
THERMALLY STRATIFIED HOT WATER STORAGE
Ee-Tong Pak, ISES member
Dept. of Mechanical Engineering
Sung Kyun Kwan University
300 Chun chun-Dong, Suwon 440-746
Republic of Korea
ABSTRACT
This paper deals with experimental research to increase thermal storage
efficiency of hot water stored in an actual storage tank for solar
application.The effect of increased energy input rate due to stratification has
been discussed and illustrated through experimental data,which was taken by
changing dynamic and geometric parameters.Ranges of the parameters were defined
for flow rate,the ratio of diameter to height of the tank and inlet-exit water
temperature difference.
During the heat storage,when the flow was lower,the temperature difference was
larger and the ratio of diameter to height of the tank was higher,the momentum
exchange decreased.As for this experiment,when the flow rate was 8 liter/min,the
temperature difference was 30°C and the ratio of diameter to height of the tank
was 3,the momentum exchange was minimized resulting in a good thermocline and a
stable stratification.In the case of using inlet ports,if the modified
Richardson number was less than 0.004,full mixing ocurred- and so unstable
stratification occurred', meaning* that this could not be recommended as storage
through thermal stratification.Using a distributor was better than using inlet
ports to form a sharp thermocline and to enhance the stratification.lt was
possible to get storage efficiency of 95% by using the distributor,which was
higher than a storage efficiency of 85% obtained by using inlet ports in the
same, operation condition.
Furthermore, if the distributor was manufactured so that the mainpipe decreases
in diameter toward the dead end to maintain constant static pressure.it might be
predicted that further stable stratification and higher storage efficiency are
obtainable ('£ra. ,f tUsSs. 95%).
KEY WORDS
Storage efficiency; thermal stratification; geometric paraments;
momentum exchange; distributor.
INTRODUCTION
Energy storage is employed in a solar thermal energy system to shift excess
energy produced during times of high solar availability to times of low solar
availability. Storage can also be used to provide energy during events such as
cloudy days.Thermal energy may be stored as sensible heat or latent heattl-3:>
-------
1418
Sensible heat storage of thermal energy is perhaps, conceptually, the simplest
form of storing thermal energy'4'. In many analyses, the temperature of the
storage has been assumed to be uniform after storage, but in real situations
liquid temperature in the tank will not be uniform after storage, especially in
the vertical dimension. The warmer, lighter liquid is stored on top of the
colder, heavier liquid resulting in thermal stratification. In practice, perfect
stratification is not possible since the water entering the tank will cause a
certain amount of agitation and mixing.Having obtained good thermal
stratification by eliminating mixing, it is equally important to maintain the
temperature 1ayersc 3 5.
Many thermal storage devices consist simply of a tank of water which is assumed
to be mixed. However, there are definite advantages to operating the storage in
a stratified mode, with the hot water separated from the cold water'6•7>.The
separation of the fluid may be accomplished by the use of simply a single tank
operated to allow stratification to occur due to buoyant forces.The problem of
thermal stratification in solar energy storage systems has been considered by
several investigators. For example, Davis and Barerac8> observed from
experiments that improvement in the performance of solar water heating systems
due to strtification is of the order of 10 percent. Sharp and Loehrke19}
conducted detailed investigations of the system performance when stratified
water storage is employed. Recently, Pak and Cho conducted an experiment
involving flow analysis of buoyant jets into a storage tank through variable
nozzles'10' and Pak, Hwang and Choi also conducted an experimental study of the
thermal storage efficiency through variable porous manifolds in a test storage
tank5115.
The purpose of this investigation was to experimentally determine what conditions
produce optimum stratification during charging in thermal storage using water as
a storage media. The investigation included a large number of parameters, such
as inlet condition, mass flow rate, tank height to diameter ratio, and inlet-
exit water temperature difference.
Finally, it was intended to increase thermal storage efficiency of hot water
stored in an actual 'size storage tank by stratification enhancement.
Fig. 1. Schematic diagram of experimental apparatus
1. Automatic Voltage Regulator
2. Hybrid Recorder
3. Hot Water Supply Device
4. Temp. Controller and Heater
5. High-Temp. Bath
6. Pump
7. Pump Speed Controller
8. Temp. Controller
9. Flowmeter
10. Bypass Valve
11. M&ia Valve
12. Manometer
13. Testing Tank
14. Distributor
15 Acril Bar
16. Thermocouple Probe
17. Exit Port
-------
1419
EXPERIMENTAL APPARATUS
The experimental apparatus consisted of a charging loop, an experimental tank,
interchangeable inlets and distributor, and an outlet kept stationary at the
bottom of the tank. A data gathering system was involved in the apparatus. A
schematic diagram of the experimental apparatus is shown in Fig. 1.
Experimental Tank
The experimental tank is cylinderical,1680 mm tall and 516 mm in diameter( H/D =
3 in this case ). The tank, containing 350 liters of water, is made of
transparent fiberglass in order to enable picture taking and is insulated with
100 mm of batt insulator. It was possible to interchange inlet and distributor
inlet position on 3, 2 and 1 of tank height to diameter ratio( H/D ).
Inlet Port and Distributor
Tank inlet(including distributor) and outlet positions are also shown (Fig.l.]
The inlets were constructed of 20 mm straight fiberglass tube with 2 mm
thickness and screwed into the inlet port. The distributor perforated with the
same diameter holes was also constructed of the same material and diameter as
the inlet tubes and the total area of perforation to cross section area of the
distributor was 2 (a = Ab/Ai, = 2).
Charging Loop
The charging loop consisted of a pump, a hot water supply device, a heater and
an exit port. The hot water supply device was used to insure a constant supply
temperature of hot water, and the exit port to insure a constant flow.The maximum
flow rate in the charging loop was 12 liter/min which corresponds to a "
turnover " time( total tank fluid mass divided by charging loop flow rate ) of
29 minutes for the experimental system.
140
120
I
u
£
0
M
w
s
TKMPUKATIIK!-: < *C >
100
i
u
u
X
Fig. 2. Temperature profile in test tankUnlet port-without Distri
• butor) (AT=20«C, h/D.3, Q=8 LPM, unit of timed) . minute)
Fig. 3 ~ icmporature profile in test tmikOnkn port-without Distri-
butor) ("r*20eC, 1I/|J«3, Q~ 12 LPM, unit of Utncft) i minute)
-------
1420
Data Gathering System
The data gathering system consisted of a temperature sensor of a thermocouple
probe,a flow sensor of a flowmeter , a manometer and a Hybride Recorder for
temperature recording. The inlet temperature, outlet temperature and vertical
temperature profile were measured using T-type thermocouples.The vertical
temperature profile was taken with 26 probes for the case of H/D = 3, 25 probes
for H/D = 2 and 21 probes for H/D =1 in a 10 mm diameter polycarbonate tube
located along the centerline of the tank. It was assumed that the vertical
temperature profile was one dimensional. The accuracy of the temperature
measurement was ±0.05*C. The flow measurements were made using a Rotameter with
an accuracy of 0.5 % of measurement.The pressure measurements at the inlets were
made using a manometer with Carbon-tetrochloridel CCI4 ).
RESULTS AND DISCUSSION
Stratification through Temperature and Profiles
The temperature profiles of each experiment were plotted in the form shown in
Fig. 2, Each line represents a different point in time. The lines are plotted
from -t •= 0 until the end of the experiment. One of the most striking features of
these graphs is the nearly parallel temperature profiles in case of using the
distributor( see Fig. 5 ).From the graphs.it can be observed that there is a
region of nearly constant temperature gradient which moves down the tank as the
tank is charged. This is the boundary or thermocline region.
When using only inlet ports( without distributor ) for the heat storage,
momentum exchange has increased in case of the larger flow rate, the lower
inlet-exit water temperature difference and the smaller ratio of diameter to
height of the tank as indicated in the figures. Especially, when the flow rate
was 8 liter/min,the inlet-exit water temperature difference was 30 °C and the
ratio of diameter to height of the tank was 3, the momentum exchange was
minimized( see Fig. 3 and 4 ).
In the same operation condition, using the distributor was better than using the
inlet ports only to form the stratification. Also, in the casa of 8 liter/min.
flow rate, 30"C inlet-exit water temperature difference and 3 as the ratio of
diameter to height of the tank, the stratification enhancement took place.It was
also observed that the momentum exchange was minimized when the stratification
enhancement took piace(see Fig.5).
Degree of Stratification and Richardson Number
Degree of stratification of the thermal storage is characterized by the
magnitude of the temperature gradient in the boundary region. The magnitude of
the gradient was observed to be a function of the Richardson Number.It was
observed that the depth H*, at which this boundary region first occurred in the
tank varied from experiment to experiment and correlated well with the modified
Richardson Number defined as :
An example of the variation of H* with the Richardson Number can be seen by
comparing Figures. The relationship between H*/H and the modified Richardson
Number is shown in Fig. 6.
It is important to note that the sharp cut- off in the graph is not at the point
where the tank becomes ful ly mixed but only at the point where the tank no
longer stratifies at the inlet. The point of complete tnixing( i.e. H*/H = 1 )
corresponds to a Richardson Number of around 0.004.
-------
1421
~\ I r
15 20 25 30 35 40 45 50
tkmpkkatuui-: < ~c- >
Fjg,4. Temperature profile in lest tankllnlct port-witliout Distri-
butor) (&T=3Q8C, H/D=3, Q=8 LPM, unit of time(t) : minute)
~ : ll/KS Q: H/D=2 X : MM
0.C02 O.CO't 0.M 0.008 O.OI
ElcWdaw Hub
Fig. (j. HVH as function of Ricliordson number
1r
lbeofitiol cum
Experiieiitii
Fig. 5. Tcmgcrnluro profile in tc«l InuKlwitl) Distributor)
(ATs3u«C, H/U*3, 9=8 I-I'M, unit of timed) » mttiule)
*S>
'M
\
Distributor /"" |
¦
iniot Port
ji i i ii i ¦ i ¦ -
15 20 25 20 35 40 45 50
Tesperature IX)
fig. 7 • Temperature variation with height in the tutik ut 9 tnin.
from charging for H/D=3, Q=8 LPM, aT=30®C
0,5 1 15 2 25 3
du^ice ti*(t')
Fis- 3 . Effect of ctarging tim« on stocga unit
-------
1422
Thermocline Energy Storage
The ultimate reduction in storage tank volume is achieved when the storage tank
volume equals the storage fluid volume. An attempt to achieve this represent a
thermocline system in which both the hot and cold storage fluids occupy the same
tankcl25.Accordingly, thermocline energy storage systems have received much
attention because of their potential for low cost resulting from minimized
tankage volume as mentioned above.
As can be seen in the above figures, when the flow rate was 8 1iter/min.,the
inlet-exit water temperature difference was 30°C and the rate of diameter to
height of the tank was 3, a sharp thermocline could be formed and the mixing of
the hot and cold water layers was small. In comparing using the inlet ports(
without distributor ) with using a distributor, it could be seen that using the
distributor resulted in a more sharply stable thermocline and a higher
temperature of water in the tank than using inlet ports( see Fig. 7 ).
Effect of Stratification on Charging the Storage
Stratification of the thermal storage results in an increased rate of energy
input. The equation for energy content as a function of time of a perfectly
mixed tank with constant temperature input and constant mass flowrate is :
E = Mm Cp AT {1 - e -CMi/MnOt} (1)
The equation for the energy content of a perfectly stratified storage, with
constant temperature input and constant flowrate is :
E = Mi CP AT t (2)
Equation 1 and 2 can be non dimensionalized since storage efficiency is defined
as the ratio of theoretical total energy to be stored to actual energy stored in
the tank, so that for the perfectly mixed tank :
Tjm =
Ms CP AT { 1
Ms CP
-CMi/Ms >t}
1 - e
(3)
and for the perfectly stratified tank :
Mi CP AT t
= mTcT^t <4)
= t*
where t* = t / Ms / Mi.
Fig. 8 illustrates the charging rate of a perfectly mixed and perfectly
stratified storage and the results of an actual stratified storage comparing the
theoretical curve with experimental data.
As can be seen from Fig.8, it is possible to increase storage efficiency to 95%
by using the distributor in case of Q = 8 1 iter/min. , AT = 30 °C and H/D = 3,
which is much greater that a storage efficiency of 63 % obtained by using the
inlet ports in case of Q = 12 liter / min., AT =30 °C and H/D = 3.
It is also possible to get storage efficiency of 84% by using the inlet ports in
the same operational conditions of geometric and dynamic parameters(Q, AT, H/D )
as were applied 95% storage efficiency was obtained. Therefore, when the
distributor was used, storage efficiency has increased by 11% to 95%.
-------
1423
CONCLUSION
During the heat storage, when the flow was lower,the temperature difference was
larger and the ratio of diameter to height of the tank was higher,the momentum
exchange decreased.As for this experiment,when the flow rate was 8 liter/min.the
temperature difference was 30°C and the ratio of diameter to height of the tank
was 3,the momentum exchange was minimized resulting in a good thermocline and a
stable stratification
In the case of using inlet ports,if the modified Richardson number was less than
0.004,full mixing occured and so unstable stratification occured,which means that
this could not be recommended for storage through thermal stratification.
Using a distributor was better than using inlet ports to form a sharp
thermocline and to enhance the stratificationlt was possible to get storage
efficiency of 95% by using the distributor,which was higher than a storage
efficiency of 85% obtained by using inlet ports in the same operation
condition.Furthermore,if the distributor was manufactured so that the mainpipe
decreases in diameter toward the dead end to maintain constant static
pressure.it might be predicted that further stable stratification and higher
storage efficiency are obtainable (ie.,more than 95%).The differential equation
was solved for the hydraulic diameter as a function of distance from the dead
end to obtain equal flow :at each perforation.
REFERENCES
1.Telkes.H.(1974). "Solar energy storage" ASHRAE.JL, pp.38-44.
2.Telkes,M.(1973). "Energy storage media," Proc. of the Solar Heating and
Cooling
for BuildingWorkshop. Washington.D.C. NSF/RANN 73-004, pp.57-59,(NTIS ACCESSION
No.PB-223 536)
3.Close,D.J., "Rock Pile thermal storage for Confort air conditioning", Mech.
and Chem. Engng Trans.Australia, MCI, 11, 1965.4. 4. Ldf,G.0,G.,E>-Wakl i,M,M, and
Chiou.
4. J. P. (1964)."Design and performance of domestic heating system employing solar
air-The Colorado House,"Proo.UN Conf. on New Source of Energy 5,185.
5.Duffie,J.A.and Beckman.W.A.,Engineering of Thermal Processes"(1980). John
Wiley & Sons.
6.Cuplinska,E.L.(1976). ASHRAE JOURNAL, pp 29-30.
7.Brumleve.T.D.."Sensible Heat Storage in Liquids" (1974). Plowshare and
Transducer Technology Division 8184, Sandia Laaboratories, Report
SLL-73-0263, july.
8.Davis,E.S. and Barera.R.(1975)."Stratification in solor water heater storage
tank,"Proc. Workshop on Solar Energy Storage Sobsystem for the Heating & Cooling
of Building,Charlottesuille, Virginia,pp 38-42.
9.Sharp,M.K. and Loehrke.R.I. (1978)."Stratified versus well mixed sensible heat
storage in a solar space heating application," Paper No. 78-HT-49,Presented at
the AIAA-ASME Thermodynamics and Heat Transfer Conference,Palo alto..California.
10.Pak,E.T. and Cho.W.."Flow analysis of Buoyan jets into Storage Tank through
Variable Nozzles.(1989)." J,of Solar Energy Society of Korea,Vol.9, No.2,pp
42-50.
11.Pak.E.T.,Hwang,S.I. and Choi,Y.I.(1989)."Experimental Study on the Thermal
Storage Efficiency Through Variable Porous Manifolds in a Test Storage Tank,"J.
of Korea Solar Energy Society, Vol.9,No.2 pp 37-43.
12.Stine,W.B. and Harrigan,R.W.(1985)." Solar Energy Fundamental and
Design,"p.268 John Weleyand Sons.
-------
1424
DESIGN OPTIMIZATION OF A TWO-PHASE SOLAR WATER HEATER
OPERATING IN FORT COLLINS WITH R-123
H.A. Walker and J.H. Davidson
Solar Energy Applications Laboratory
Colorado State University
Fort Collins, Colorado 80523
ABSTRACT
Design of a two-phase, self-pumping solar water heater using R-123 is optimized
in terms of life cycle cost for Fort Collins, Colorado. Results show the most
economical solar system meets around 90 % of the load under cloudless conditions.
KEYWORDS
Economic Optimization; Passive Solar; R-123; Self-pumping; Solar Water Heating.
NOMENCLATURE
A
=
area, m2
U
loss coefficient, W/m2C
a
=
present worth factor for
V
=
volume, H
general inflation rate
z
=
height of liquid lift, m
b
=
present worth factor for
Greek letters
fuel inflation rate
a
=
absorptance
c
=
specific heat (J/kgK)
X
=
mean lifetime, years
d
=
discount rate
1
=
efficiency
e
=
energy escalation rate
T
=
transmittance
F
=
heat removal factor
Subscripts
K
=
incident angle modifier
a
=
ambient
k
=
thermal conductivity (W/m2K)
aux
=
auxiliary heater
M
=
mass (kg)
c
=
collector
IC
=
initial cost, US$
cond
=
condenser
LCC
=
life cycle cost (eqn. 3), US$
i
=
year index
LCCM
=
multiplier used to predict
in
=
indoor
costs of component failure
j
=
component index
LCS
=
life cycles savings, US$
1
=
thermal loss
n
=
life span, years
L
-
daily water heating load
OM
=
operation and maintenance cost
main
=
water mains properties
excluding electricity, US$
min
=
minimum
P
=
price of electric energy, US$/MJ
pre
=
preheat system
Q
=
thermal energy, MJ
Rb
=
heat removal factor for
R
=
replacement cost, US$
salvage value, US$
boiling collector
S
=
s
=
storage water properties
SF
=
solar fraction (eqn. 1)
set
=
water heater set temperature
T
=
temperature, °C
sol
=
solar
INTRODUCTION
The two-phase, self-pumping solar heating system uses vapor pressure to transport
vapor downward from the collector to the condenser as well as to force the upward
return of condensate. Reliability is high and maintenance costs low since no
pump, controller, nor external power is needed. Use of the environmentally safe
refrigerant R 123 as the transfer fluid eliminates freezing and scaling while
enhancing heat transfer with two-phase operation. Experimental investigations
(DeBeni and Friesen, 1987; Hedstrom and Neeper, 1986; Davidson and colleagues,
-------
1425
1989) have proved the technical viability of the self-pumping concept, but the
design of prototype systems has been heuristic. The validated simulation of
Walker and Davidson (1990) was used in an optimization study (Davidson and
Walker, 1991) based on the Solar Rating and Certification Corporation (SRCC,
1984) standard rating procedure for solar hot water systems. In this study, a
design optimization is presented which minimizes life cycle cost for operation
of a self-pumping system in Fort Collins, Colorado. Thermal and economic
performances are studied in terms of four design parameters: collector area, in-
tank condenser area, water storage volume and refrigerant volume.
Up per Accumulator
refrigerant liquid
jr-.v-M refrigerant vapor
— copper tubing dia. 0.64 cm
— copper tu&fng dla 1.27 cm
^ tempering valve
Preheat Tank Conventional
Heater
Fig. 1. Self-pumping solar water preheat system with electric backup heat.
The dual accumulator solar preheat water heater shown in Fig. 1 operates in
cycles made up of run, pressurizing, and pump phases. In the run, or heat-
collection phase, liquid refrigerant is gravity fed to the collector from the
upper accumulator tank. Vapor travels downward to the in-tank condenser. When
the upper accumulator is empty, a float valve in the upper accumulator forces the
vapor to flow directly to the lower accumulator through the pressurizing line.
Pressures and temperatures in the collector and lower accumulator increase until
the pressure differential between the upper and lower accumulators is sufficient
to return the condensate through the return line to the upper tank.
METHODOLOGY
Fixed system components are specified in Table 1. The reservoir is sized
conservatively at 10 I to avoid dry-out of the collector. Pipes are selected to
minimize annular volume and thermal capacitance while keeping pressure drops
negligibly small.
TABLE 1. Fixed System Specifications
Component
Single Collector:
A0 (m2)
(F^ra)
(FrjjU) (W/C)
£
(Mc) (J/C)
V (J)
Lifting height (z)
(UA). (W/C)
Accumulator/reservoir wall
Piping Type L copper
Insulation: k=0.026 W/mC
Specification
1.7
0.79
4.74
l-0.09[(l/cos0) -1]
2376
1.1
6.2 m, two-stories
0.9 + 0.0087/£
3 mm/1.6mm steel
1.27 cm and 0.64 cm
3.8 cm, 1.9 cm and 2.5 cm
-------
1426
Daily operation is modelled using a First Law numerical simulation (Walker and
Davidson, 1990) modified to include operation with R-123 and a double wall in-
tank condenser using a heat transfer correlation for tubes in an infinite
quiescent fluid (Churchill and Chu, 1975). Computer time is conserved by
choosing four cardinal days to represent seasonal effects and assigning each
cardinal day to 91.25 days of the year. The four days chosen to represent all
the seasonal possibilities are: Spring Equinox, a medium length, cold day; Summer
Solstice, a long, hot day; Autumn Equinox, a medium length, hot day and Winter
Solstice, a short, cold day. Environmental temperatures, insolation and load
information are listed in Table 2. Two types of load profiles model the hot
water demands; an energy draw fixes the amount of energy drawn from the system
relative to water mains temperature (SRCC, 1984) and a mass draw fixes the mass
of water drawn. The energy of the mass draw equals the energy draw if the mains
water temperature is 20 C. Water is drawn from the preheat tank at a rate of
0.2 kg/s three times a day, at 0800, 1200 and 1700 hours.
TABLE 2. Operating Specifications
Specification
Cardinal
Dav
Spring
Summer
Fall
Winter
80
173
266
356
5.0
8.0
12.0
4.0
2.8
18.5
15.7
-1.3
22.0
22.0
22.0
22.0
49.0
49.0
49.0
49.0
24.6
24.9
24.6
18.3
42.3
42.3
42.3
42.3
68.5
64.8
58.4
71.1
Julian Day
W (°C)
t„ rc)
Tm <°C)
Tset (°C) ,
Qsol (MJ/m day)
Ql (MJ/day) *
2l (MJ/davl **
* energy draw (3 draws per day of 14.1 MJ each at T.et)
** mass draw (3 draws per day of 124 kg each at Tset)
Solar fraction (SF), defined as the fraction of the daily water heating load met
by the solar system, is,
SF
(1)
where Q_re is the daily thermal output of solar alone and QL is the daily load.
To simulate periodic steady operation, SF is evaluated on the second day of
constant heat output (within 3%) or on the fourth day of operation, whichever
occurs first. Thermal efficiency, ij, is given by the ratio,
_ Qpre
IAr
(2)
The present worth of the lifetime costs of owning and operating the combined
solar and electric water heater (LCC) is determined by (Ruegg and Fav, 1979),
LCC - IC - (Snan) ~ I I a^OMy) + £ R.LCCM :
i-l 3
| (1-SF )QLPbi
i-l
(3)
where i refers to the year and j is a component index. Initial purchase and
installation costs are retail costs quoted in early 1990. At the end of the 20-
year analysis the system has no salvage value. Routine operation and maintenance
costs (e.g. cleaning the collector) are assumed to be handled by the owner at no
cost. The replacement multipliers, LCCMj, are based on the probability
distribution for a given component failure (Short, 1986) listed with other
economic parameters in Table 3. Replacement costs equal initial capital costs
plus installation. Present worth factor for costs escalating at the general
inflation rate is a1 and that for costs escalating at the fuel inflation rate is
b1. Each year in the 20-year lifetime is assumed to be the same as that
represented by the four days simulated. Life cycle savings, LCS, is the
difference between the life cycle cost of meeting the load with auxiliary energy
alone and that of meeting the load with the solar system and auxiliary back-up.
-------
1427
The economic optimization procedure is a global search method in which LCC is
minimized for each variable with the other three design variables held constant.
This procedure is repeated until the change in LCC between successive iterations
is less than US$ 10.
TABT.K 3 . LCC Parameters
Life span, n — 20; Energy escalation rate, e — 2% greater than general inflation
Discount rate, d = 3% (general inflation rate of 5% and interest rate of 8%)
Price of electricity, P - $0.0763/kWh1 (GAMA, 1988)
Component
IC
-------
1428
Life cycle cost and life cycle savings in the mass load case are higher than
those in the energy load case because water mains temperatures below 20 °G cause
the energy required to provide the mass load at the set temperature to exceed the
SRCC specified energy load.
Figure 2 (a) shows solar fraction as a function of collector area with the other
design variables held at their optimal values for the energy load case and Figure
2 (b) shows the same quantities for the mass load case. Solar fraction increases
with increasing Ac but the rate of increase decreases with increasing Ac
(3ZSF/3A0Z is everywhere negative). With the other design variables at their
optimal values, the most economical collector area is the point above which
initial and replacement costs of additional collector area (ICc+Rc - 134 $/m2)
are not returned by fuel savings. This is expressed by the relation,
ICc+R (4)
i-l laux aA0
For the energy draw case the summation on the right side is equal to $7,586. The
most economical collector area for this situation occurs when 3SF/3AC - 0.016/m2.
Referring to Fig. 2 (a), this value of 3SF/3AC occurs at Ac - 6.3 m2. For the
mass draw case, the summation on the right hand side of eqn. 4 is equal to
$11,714 and the most economical collector area is 8.9 m2. Higher collector costs
would reduce the optimal solar fraction and as a result the optimal values of the
other design variables would also decrease.
Optimal values of condenser area and water storage volume shown in Table 4 are
determined similarly. For the energy load case with A equal to 6.3 hi2,
performance is rather insensitive to values of Acond above 0.5 m2 and Vs above 170
Z. Increasing A00Jld from 0.4 m2 to 0.5 m2 increases SF from 0.82 to 0.88 but
doubling AconlJ to 1.0 m2 increases SF to only 0.94. Maximum SF of 0.918 is
predicted at Vs of 230 i. Above this volume SF decreases due to increased losses
from the preheat tank. Increasing the volume of refrigerant volume circulated
with each cycle increases cycle duration. This decreases losses incurred due to
pumping but exacerbates losses that occur at the end of the day when insolation
is not sufficient to maintain the elevated temperature required to pump. Thus,
as shown in more detail by Davidson and Walker (1991) increasing circulating
refrigerant volume increases cost without improving performance.
1.0
0.9
0.7
0.5
CO
Spring
Summer
Fall
0.4
0.3
0.2
Winter
Year
0.1
0.0
3 4 5 6 7 8 910111213
1.0
0.7
Spring
Summer
Fall
0.4
0.3
0.2
Winter
Year
0.0
9 10111213
7
5
Ac
(a) (b)
Fig. 2. SF as a function of Ac for a) energy load and b) mass load
-------
1429
CONCLUSIONS
Under the cloudless conditions assumed here, life cycle cost is minimized by
meeting about 90% of the load with solar energy. System performance is dependent
upon all of the design variables considered, but is most dependent upon collector
area. For a fixed load, collector area limits the maximum solar fraction. The
optimal sizes of the condenser and preheat tank increase with increasing
collector area.
Savings over using an electric water heater alone are also realized with
collector areas smaller than the optimum. A suggestion for future work would be
to determine the breakeven system which has a life cycle cost equal to that of
meeting the load with an electric water heater alone.
ACKNOWLEDGEMENT
The support of the U.S. Department of Energy through grant No. DE-FG03-86SF16036,
Freon Products Division of E,I. DuPont de Nemours & Co. (Inc.) and Colorado State
University Computer Center are gratefully acknowledged.
REFERENCES
Churchill, S.W., and Chu, H.H.S., 1975, "Correlating Equations for Laminar and
Turbulent Free Convection From a Vertical Plate," Inter. J. of Heat
and Mass Transfer. Vol. 18, pp. 1323-1329.
City of Fort Collins, Light and Power/Architectural Energy Corporation, 1989,
"REM/RATE Home Energy Rating System Manual," Fort Collins, CO.
Davidson, J. H., Walker, H. A., and Lof, G. 0. G., 1989, "Experimental Study of
a Self-Pumping Boiling Collector Solar Hot Water System," ASME J.
Solar Energy Engineering Vol 3, No. 3, pp. 211-218.
Davidson, J. H. and Walker, H. A., 1991, "Design Optimization of a Two-Phase
Solar Water Heater Using R-123," submitted to ASME J. Solar Energy
Engineering and presented at ASME/JSME/JSES Solar Energy Conf. , Reno,
NV, March 17-22, 1991.
DeBeni, G., and Friesen, R. , 1987, "Experimental Results of a Solar Hot Water
System with Spontaneous, Downward, Non-Freezing Heat Transport
System," Energy Conserv. Mgmt., Vol. 27, No. 3, pp. 293-300.
Gas Appliance Manufacturers Association (GAMA), 1988, "Consumer's Directory of
Certified Efficiency Ratings," GAMA, Arlington, VA.
Hedstrom, J., and Neeper, D., 1986, "Passive Space Heating with a Self-Pumpiing
Vapor System," Proc. ASES Eleventh National Passive Solar conf..
Boulder CO, p.493-498.
Ruegg, R.T., and Fav, G.T., 1979, "The Microeconomics of Solar Energy," The
National Bureau of Standards, Washington, D.C.
Short, W. D., 1986, "A Method for Including Operation and Maintenance Costs in
the Economic Analysis of Active Solar Energy Systems," Solar Energy
Research Institute, Report SERI/TR-253-2626.
Solar Rating and Certification Corporation (SRCC), 1984, "Operating Guidelines
for Certifying Solar Water Heating Systems," Document OG-200,
Washington, D.C.
Walker, H.A., and Davidson, J.H., August 1990, "Analysis and Simulation of a Two-
Phase Solar Water Heater," ASME J. Solar Energy Engineering. Vol.
112, No. 3, pp.153-160.
-------
1430
Testing the copper cricket™ Solar water heater
by an Electric Utility
Eldon Haines*, Douglas Boleyn**, and Charles Dallas*
*Sage Advance Corporation, Eugene, Oregon, USA
**Portland General Electric Company, Portland, Oregon, USA
ABSTRACT
Copper Cricket passive solar water heaters were monitored for performance in ten
households in Portland, Oregon, by Portland General Electric Company. Performance
closely matches that predicted by the Oregon Department of Energy.
KEYWORDS
Solar water heating; Copper Cricket™; electric utility; monitoring; performance.
INTRODUCTION
Portland General Electric Company (PGE) has recognized the need for definitive
research into the on-site performance and customer acceptance of solar water heaters in
its Portland, Oregon, area. This research is important for two reasons: First, the State
of Oregon provides tax credits for solar water heating, where the amount of the credit is
based on the energy saved by the solar water heater. Secondly, energy conservation
and renewable energy resources are being acquired by utilities of the Pacific Northwest
as a result of Least Cost Planning and the Pacific Northwest Power Plan. These
programs require the establishment of energy savings potential and cost effectiveness of
renewable resources. PGE began a monitoring program involving the Copper Cricket
as one potentially valuable renewable resource.
PGE limited its current monitoring to the Copper Cricket for several reasons: (1) PGE
has already monitored a substantial number of solar water heaters installed during its
Water Heater Incentive Program from 1980 to 1985; (2) Over the past five years the
Copper Cricket passive solar water heater, developed and manufactured by Sage
Advance Corporation, has entered the marketplace; (3) The Copper Cricket has not
been monitored in a household setting; and (4) The Copper Cricket appears to have
characteristics which would assure high reliability and long service life, both of which
are important factors to a utility acquiring renewable resources. Thus PGE began the
monitoring project reported here in order to supplement earlier programs involving
other solar water heaters.
-------
1431
DESCRIPTION OF THE COPPER CRICKET SYSTEM
The technology of the Copper Cricket solar water heater has been described previously
(Adams, 1985; Haines, 1985; Haines, Block, and Northcutt, 1987; McPhee, 1989).
The system uses "geyser-pumping", a downward-pumping passive technology , to
circulate freeze protected working fluid from the collector to a heat exchanger in a closed
solar loop. The collector, whose aperture is 3.2 m2 (34.2 fit2), is singly glazed and uses
a black chrome-on-copper absorber with fluid channels soldered along its long
dimension. The solar loop is filled to a defined level, evacuated, and sealed during
installation. The fluid boils in the absorber channels and the vapor bubbles elevate the
hot fluid into the upper manifold. The head provided by this elevation overcomes the
balance between the fluid columns in the pipes descending to and ascending from the
heat exchanger. Hot fluid is thus pressed down to the heat exchanger, where its heat is
delivered to the potable water, and cold fluid is lifted back up to the collector. Potable
water circulates through the heat exchanger and into the storage tank by
thermosiphoning. No electricity, pumps, valves, or ancillaiy equipment are needed.
The working fluid has been freeze tested to -40°C.
THE MONITORING PROJECT
The goals of the monitoring project were three-fold. The first was to determine actual
energy savings by the Copper Cricket in real households. The second goal was to
survey customer responses to the system and its installation. The third was to identify
any installation or operational concerns which could lead to limited life or customer
dissatisfaction. A fourth goal was added during the analysis phase, which was to
validate the energy savings prediction method employed by the Oregon Department of
Energy (Robison, 1988).
The Copper Cricket system was offered to a limited number of employees of PGE in
June, 1988. Eight employees and two customers purchased systems. The participants
paid for the systems, received financing through PGE, and were eligible to receive state
income tax credits. Households ranged in size from 2 to 5 occupants. Each household
agreed to participate by recording and forwarding meter readings over a one-year test
period.
Before the Copper Cricket systems were installed, each household had been equipped
with a flow meter placed on the cold water inlet, and an electric meter on the electric
service to the water heating tank. Each participant recorded meter readings on a daily
basis for at least two weeks before the installation of the system to provide baseline
data. These data were recorded during August and September, 1989. Solar radiation
was measured daily at one site in Portland by means of two Eppley Black and White
pyranometers, one oriented horizontally, the other facing south at 60° elevation.
The Copper Cricket systems were installed during September and October, 1989. All
ten participants recorded data daily during the first month after installation. Again in
April, 1990, seven participants provided daily readings during a two week period. Six
-------
1432
participants provided sufficient data throughout the year to permit periodic analysis.
Most data were recorded at monthly intervals. However, the participants differed in
interval length and consistency; while one recorded data regularly at weekly intervals,
another recorded data episodically at intervals as great as six weeks.
Direct Comparison Method.
The original plan was to monitor the hot water volume (V) and electric energy (E) used
in water heating before and after the installation of the Copper Cricket. We expected
that we could calculate solar savings from a year's accumulated data by comparing the
ratio (E/V) before and after the installation by means of
where So is solar energy savings in kWh by this comparison.
The value of (E/V)before was calculated from August data, when cold water
temperatures were at their highest and conductive losses from tanks and pipes lowest.
Thus (E/V)before was misleadingly low and resulted in inaccurately small solar savings
calculations. To further illuminate this problem, we obtained data on hot water energy
savings for over 200 monitored homes in the Pacific Northwest. Hot water volume and
energy data revealed that the ratio (E/V) varied widely throughout the year as cold water
temperatures and ambient air temperatures changed. We applied the variation pattern
derived from this data as a correction to (E/V)before for each household in the
monitoring program, and found that the calculated solar energy savings, Sc, were in
good qualitative agreement with predictions by the ODOE method (Robison, 1988).
The ODOE method is based on empirical correlations between monitored energy
savings and measured insolation, cold water temperature, and ambient temperature. It
uses the standard SRCC certification results for whole systems as a basis for the
prediction (SRCC, 1988).
Energy Balance Method.
As an alternative to the comparison calculation shown above, the empirical energy
savings were also calculated from simple energy balance. An energy balance calculation
assumes that energy input, consisting of solar energy (Sb) and electrical energy (E), is
balanced by energy output, consisting of energy required to heat water (W) and heat
loss (L) from the tank. The solar savings may be written as:
THE ANALYSIS METHODS
Sc — [(E/V)before - (E/V)after] Vafter
(1)
Sb= W + L - E
(2)
The term W is calculated from
W = V k (Thot-Tcoid)
(3)
where k is the product of density and heat capacity and Thot and Tcold are the hot and
cold water temperatures. Loss, L, is calculated as
-------
1433
L —24 UAioss (ThofTamb) (4)
where UAi0Ss is given a value of 0.0030 kW/°C (5.7 Btu/hr °F). Tamb is the average
ambient temperature, taken from Portland, Oregon, weather data.
This energy balance has the weakness that the resulting savings, Sb, is dependent
on hot and cold water temperatures, and to a lesser degree on the temperature around the
storage tank. None of the three temperatures were measured in the households during
the tests. The cold water temperature, Tcold, was modeled from anecdotal information
as an annual sine function with its maximum value in late August. Air temperatures
around the tank vary depending on degree of protection given the tank. Lacking this
information, we used outdoor ambient temperature, Tamb- The hot water temperatures,
Thot, were derived by minimizing the root-mean-square difference between the savings
from the energy balance method, Sb, and the energy savings, Sodoe, predicted by the
ODOE method. The resulting correlations for the six sites are shown in Fig. 1, where
empirical values, Sb, for each monitoring interval are plotted against the predicted
values, Sodoe-
Hk
A
Aft
*
un 4
*
0
I
o a
cA
r-
o °
<*~/"
r
1 A
&
A
a a
A
t
&
i
0 2 4 6 8 10
Sodoe
~
Site 1
o
Site 2
¦
Site 3
~
Site 5
4
Site 7
A
Site 9
—
Unit
Fig. 1. Correlation plot of empirical energy savings, Sb, and
predicted savings, S0doe, both in kWh/day.
We discovered that the collector at Site 9 was not functioning during the winter half of
the monitoring year; thus creating the poor correlation for those periods.
Because the two values, Sb and Sodoe. were matched in a least-squares sense, it is not
surprising that the annual average values of Sodoe and Sb, displayed in Table 1, are
close to one another. That result is required by the method. However, we may reverse
the argument with meaningful outcome: If the optimized hot water temperatures, Thot,
have reasonable values, and if the sensitivity of the savings, Sb, to Thot is not too great,
then the close comparisons between S0doe and Sb are also reasonable.
-------
1434
TABLE 1. Comparisons between predicted and empirical savings.
Predicted Empirical Hot water Sensitivity
Sodoe
Sb
Thot
to Thot
[kWh/day]
[kWh/day]
ra
[kWh/°C]
Site.l
4.85
4.91
49.4
0.485
Site.2
4.57
4.59
52.2
0.171
Site.3
4.91
4.78
52.2
0.243
Site.5
4.75
4.71
54.4
0.169
Site.7
4.86
4.87
46.1
0.194
Site.9
4.54
2.74
—
Table 1 shows, iin addition to predicted and empirical annual savings, the optimized hot
water temperatures and the sensitivities of the savings results to the hot water
temperatures. Hot water temperatures ranged from 46.1 to 54.4°C (115 to 130°F).
Sensitivities for four sites are about 0.2 kWh/°C. Sensitivity for Site 1 is higher
because the predicted and empirical savings group themselves more closely around the
unit correlation line. Temperature optimization and sensitivity are not shown for Site 9
because the results have no significance. For Site 1, a variation in Thot of about 2°C
(3.6°F) would change the calculated empirical savings by about one kWh/day.
However, for Sites 2, 3, 5, and 7, variations as great as 5°C (9°F) from the optimized
values of Thot would change the calculated empirical savings by only about one
kWh/day.
Annual solar savings ranged from 4.59 to 4.91 kWh/day and averaged 4.77 kWh/day.
The average insolation at the Portland site was 3.42 kWh/m2, whereas the insolation
used by ODOE for savings calculations was 3.87 kWh/m2. When we normalize the
average savings to the higher ODOE insolation, we get 5.40 kWh/day, which is 4%
greater than the savings determined by ODOE for tax credit purposes, 5.21 kWh/day.
From this we conclude that the ODOE method may be used with confidence to predict
solar savings achieved by the Copper Cricket.
SUMMARY AND CONCLUSIONS
Copper Cricket solar water heaters have been demonstrated to provide good energy
savings, close to that predicted by Oregon Department of Energy's method. Energy
savings ranged from under 2 kWh/day in the winter to over 8 kWh/day in the summer,
and averaged 4.77 kWh/day throughout the year, 4% more than the savings predicted.
From this we conclude that the ODOE method may be used with confidence to predict
solar savings achieved by the Copper Cricket, and that the method is useful for
predicting energy savings on a large scale by electric utilities.
The analysis would have been more complete if hot water, cold water, and tank
environment temperatures had been collected along with water volume and energy data.
-------
1435
REFERENCES
Adams, J.(1985) "The Revolutionary Geyser Pump Collector". Solar Age. October.
Haines, E. (1985) "The Geyser Pump Solar Collector," Svmp. Proc. RETSIE '85.
p.203.
Haines, E., R. Block, and D. Northcutt (1987) "The Geyser Pump Solar Collector:
Comparison of efficiency with an electric pump solar collector," 12th Passive
Solar Conf. Proc.. p.417.
McPhee, M.(1989) "Solar Heat with Perks", Popular Science. July.
Robison, D.(1988) "Performance monitoring to validate site-specific estimates of
passive solar water heaters" in Performance Predictions for Passive Solar Water
Heating Systems. Bonneville Power Administration, Portland OR.
SRCC (1988) Directory of SRCC Certified Solar Collector and Water Heating Systems
Ratings. Solar Rating and Certification Corporation, Arlington VA.
-------
1436
THE CONTRIBUTION OF ENERGY-CONSCIOUS BUILDING REGULATIONS TO
CREATIVE DESIGN PROPOSALS
David Galor, Architect
13 Fichman St. Tel Aviv Israel 69027
ABSRACT
Energy-conscious building regulations should encourage the
appropriate use of geometry and materials in building design.
Characteristics such as envelope area, elements' ratio, volume,
insulation level, glazing proportion and orientation, are the basic
passive measures controlled by the architect. The paper suggests a
geometric-thermal performance index as part of common building
regulations which enables the prediction of space heating
requirements in buildings and the impact of passive measures on
them.
KEYWORDS
Building regulations; energy-conscious design; building geometry
and materials; thermal performance index; energy targets; early-
stage prediction.
INTRODUCTION
Energy-conscious building regulations are aimed at improving
visual, acoustic and thermal comfort in buildings, and reducing
their energy bills. Normally, architects are not familiar with the
scientific basis of these regulations . This is one of the reasons
why energy-conscious building regulations are sometimes simplified
into a compact form, which does not necessarilly guarantee energy
savings.
AN INTEGRATED DESIGN SYSTEM
The three groups of factors which together combine a comprehensive
environmental system are: the building internal comfort
requirements, the external environmental conditions and the
building structure (BSI 1988). The first two are part of the
architectural brief, and can be considered as "fixed". The building
structure, namely its geometry and materials, is the architect's
relevant domain and can be more "flexibly" determined. Building
regulations assume correlation between the characteristics of the
building's structure and its thermal performances. The factors
which affect space heating requirements in buildings are numerous
and interrelated and should be taken, therefore, into account
simultaneously.
-------
1437
BUILDING REGULATIONS APPROCHES
The thermal performance of buildings is determined and/or evaluated
by building regulations which are based on different thermal
concepts at various levels of complexity. Building regulations
approaches can be classified also according to the level of freedom
they leave the architect during the decision-making process: The
prescriptive (or elemental) approach ("do this!"), trade-offs ("Do
this, but you can do that instead") and energy targets ("We don't
mind what you do, as long as you get there").
The Prescriptive Approach
This approach, most frequently, tries to control heat losses
through the building fabric by demanding maximum U-values for each
of the building elements. This can affect heat losses in different
ways, part of them illustrated in Fig. 1:
(a)
(b)
(c)
(d)
Ih
11.55
8.07
3ow
q0
4.47
(a)
(b)
(c)
(d)
Floor Area (m2)
86.4
43.2
30.0
60.0
Walls Area (m2)
49.0
76.®
66.0
86 .6
Roof Area (m2)
86.4
43.2
30.0
60.0
Window area (m2)'
17.0
17.0
12.0
24.0
Heat Loss Coe. (W/0C)
130
112
84. 3
148
U-values: roof - 0.25, wail - 0.45, floor - 0.45, window - 2.8
Fig. 1. H.L.C of semi-detached house in various sizes
Although all cases comply with 90' UK building regulations (DOE
1990), heat loss coefficient per m2 of floor area of a typical
86.4m2 semi-detached house can range between 1.50 W/oC/m2 floor
area (a) to 1.30 W/oC/m2 floor area (b). Similarily, this approach
taxes small buildings which have relatively large envelope: (c)
with 2.3 m2 envelope area/m2 of floor area (1.40 W/oC/m2 floor)
compared to (d) with 1.9 m2 envelope area/m2 floor area (1.23
W/oC/m2 floor).
Lower roof U-value, compared to other elements of the envelope,
encourages the design of buildings with larger roof area ratio,
although in terms of actual envelope area and construction costs,
they will be less favourable (Fig. 2):
-------
1438
Area (m2)
U-value
Area (m2)
U-value
Roof
100.0
0.25
500.0
0.25
Walls
1,200.0
0.45
537.0
0.45
Floor
lOtKO
0. 45
SOO.O
0.45
Heat Loss Coe.
610 W/oC
590 W/oC (-3*)
Envelope area
1,400.0
1,537.0 (+10*)
Fig. 2. The impact of elements' ratio on H.L.C.
A method to overcome this problem, has been suggested by some
countries (Germany, Netherlands etc.): a trade-off procedure
between the building geometry and its thermal properties.
Trade-offs
This procedure is usually based on the calculated fabric heat loss
coefficient of a design proposal, according to the elemental U-
values required by the same building regulations. It allows, for
example, building with smaller envelope area to have higher
elemental U-values, or lower window U-values but higher
transparent/opaque ratio and vice versa. The German building
regulations (Ehm 1975) relate a geometric index, F/V
(envelope/volume ratio), to the average U-value of the envelope
(Km). Dutch regulations (NNI 1981) limit this correlation to a
certain required thermal target, "It" - Insulation Index. The form
factor, F/V, represents the compactness of the building envelope,
but as illustrated in Fig. 3, can inaccurate:
5.0
(a)
(b)
5.0
1.(0
Floor Area A (m2)
86.4
86.4
Envelope F (m2)
181.0
216.0 (+20%)
Volume V (m3)
216.0
260.0 (+20%)
F/V Ratio
0.84
0.04
F/A Ratio
2.10
2.50 (+20%)
Fig. 3. F/V Ratio of shapes
-------
1439
Energy Targets
To assess correctly the thermal performance of buildings at an
early stage of design, the following factors should be taken into
account: fabric and ventilation heat losses, internal and solar
gains. These parameters can be incorporated into a single notion,
the "Energy Target", expressed, for example, in kWh/m2 heated floor
area per season. The French G- and B-Coefficients (CSTB 1988), the
Swiss k-zul (SIA 1988) and the CIBS B-Number (CIBS 1981) are some
of the examples to this approach. It leave the architect the
freedom to determine the appropriate passive measure but requires a
standardized calculation procedure. The Swiss building regulations,
for example, takes into consideration only 4 factors (basic, form,
climate and internal temperature factor) combined into a single
equation to establish "k-zul". The effect of different orientations
of glazing as a source of heat gains is also incorporated within
the expression.
A GEOMETRIC-THERMAL PERFORMANCE INDEX
A quick assessment of space heating requirements in dwellings is
possible using a geometric-thermal performance index, based on a
limited number of factors, applicable for various dwelling types in
different climates and energy savings levels. It can be used for
early-stage space heating assessments, comparison of design
proposals, evaluating the impact of different passive measures, and
suggesting design improvements. The index includes three groups of
parameters which describe a comprehernsive environmental system:
internal comfort requirements (internal temperature and fresh air
supply), external climatic conditions (ambient temperature and
solar radiation) and building structure characteristics (floor
area, shape, insulation and fenestration). It is based on a "m2
floor area" unit, as this can be related both to other geometric
factors (like envelope, volume, fenestration) as well as energy
targets (kWh/m2 floor area/season). The index includes the
following factors:
- Form Factor (plan shape, floor height)
- Insulation factor (opaque and transparent elements
characteristics)
- Ventilation Factor (fresh air supply, natural ventilation)
- Degree-Days Factor (internal-ambient temperatures, regional
factor)
- Internal Gains Factor
- Solar Gains Factor (orientation and tilt, glazing proportion,
regional factor).
Each of the factors was given an average value and a possible range
of minimum-maximum values. The example in Fig. 4 relates to the UK
semi-detached house and the values in the tables were
proportionally weighted according to their contribution to heat
losses or gains (the average semi-detached house is characterized
by these results: total heat losses 8,700 kWh/season (=1.00), total
heat gains 5,980 kWh/season (=0.69), and auxiliary heating 3,550
kWh/season).
To assess space heating requirement of a design proposal, the
architect has to take the following steps:
-------
1440
1. Give each factor a value (using the tables).
2. Calculate Total Heat Losses (L).
3. Calculate Total Heat Gains (G).
4. Calculate Gains/Losses ratio (G/L).
5. Calculate Auxiliary heating/Losses ratio (ALR).
6. Calculate auxiliary heating per m2 floor area (A).
7. Modify Auxiliary Heating (Amod) per m2 floor area according to:
- Required Floor Area Factor (HI)
- Heating Schedule Factor (H2)
- Construction Type Factor (H3)
e jj
u o
D flJ
fr. b
ss
> u.
u» u
O IB
O fe
Factor (A):
Factor (B):
Factor (D):
(D«D1*D2)
5 o
U-U
¦S a
c *
Factor (E):
Factor (F):
(F»F1*F2*F3)
v o
" s
x «
Total Heat Losses Factor (L):
(A*B+C)*D
Total Heat Gains Factor (G):
fi+F
Gains to Losses Ratio (G/L):^
ALR Factor:
Auxiliary Heating Factor
(A) = 100*L*ALR =
. KWh/m2 floor/season
A(mod) = _
A*H1*H2*H3
, kWh/m2 floor/season
Fig. 4. A Geometric-Thermal Performance Index
-------
1441
If the architect wishes to comply with current UK Building
Regulations, the above factors should be given the value of 1.00
(=average value). It is also possible to trade-off between factors
to achieve the same result (e.g. higher glazing proportion together
with lower insulation level). However, better energy-conscious
design should follow the logic of lower heat losses (e.i. below the
total value of 1.00) and higher heat gains (e.i. above the total
value of 0.69) to achieve lower auxiliary heating compared to the
average case.
REFERENCES
BSI 8211, Energy Efficiency in Housing. Part 1 (1988). BSI, London.
CIBS, CIBS Building Energy Code. Part 2 (1981). CIBS, London.
CSTB, Regies Th-B (1988). CSTB, Paris.
DOE, Conservation of Fuel and Power. LI (1990). Her Majastey's
Stationary Office, London.
Ehm, H., Enerqieeinsparender Warmschutz im Hochbau (1975).
Bundesblatt No.12.
NNI, Nederlandse Norm NEN 1068. Thermische Isolatie van Gebouwen
(1981). NNI, Delft.
SI A, 180/1 - Nachweis des Mittleren k-Wertes der Gebaudehulle
(1988), SIA, Zurich.
Yannas, S., Passive Solar Energy Efficient House Design:
Principles. Objectives. Guidelines. ETSIT, AA Graduate School,
London.
-------
1442
INTEGRATED COLLECTOR STORAGE: MODEL DEVELOPMENT FOR
PERFORMANCE CALCULATIONS AND TEST EVALUATION
H. Visser and A.C. de Geus
TNO Building and Construction Research
P.O. Box 29, 2600 AA Delft, The Netherlands
ABSTRACT
A coherent package of tools for both design and characterization of Integrated Collector
Storage systems is presented. Special attention is paid to the development of a simplified
ICS calculation model. Tests for supporting the model development are described. The
calculation model is to be used for evaluation of test results in order to estimate the long
term performance of tested or designed ICS systems.
KEYWORDS
Integrated Collector Storage; calculation model; experiments; indoor test method; black box
test method; test evaluation; parameter identification; yearly performance.
INTRODUCTION
Economic Improvement of Solar Domestic Hot Water Systems
The ratio between investment and both financial and environmental yield is too large for
most present-day solar domestic hot water (SDHW) systems. From turnover increase of
SDHW systems in future only a small cost reduction can be expected. Hence, cost
reduction has to be obtained by development of a new generation of systems which contain
less components, demand a smaller use of materials, are suitable for batch production and
can be installed easily.
Simplification and improvement of the heat transfer mechanism from the collecting surface
to the heat store may give a further cost reduction. Present-day SDHW systems are often
provided with a pump controlled by temperature difference sensors, taking care of the
circulation of the heat transfer fluid between solar collector and heat store. Both pump and
control are relatively vulnerable components of the installation. Moreover, in many cases the
energy consumption of the pump is not negligible and reduces the energy savings.
Research on New SDHW Concepts in the Netherlands
In the Netherlands two main lines in the development of new concepts can be observed. On
the one hand there is the development of compact SDHW systems in which collector and
heat store are separate components, on the other Integrated Collector Storage (ICS) systems
are designed. The designer is supported by theoretical and experimental research financed
by the Dutch Ministry of Economic Affairs. Within the framework of the National Solar
Energy Research Programme several projects on this subject are carried out, i.e.:
-------
1443
detailed measurements on the heat transfer in the heat store of a compact SDHW
system as well as inside an ICS system;
detailed flow and heat transfer calculations in ICS systems;
development of simplified models for performance calculations of compact
SDHW and ICS systems;
development of a test method specificly for ICS systems.
Figure 1 indicates the connection between the various research projects for the ICS concept.
In the following the relationship is further explained.
MODEL VERIFICATION
verification
detailed flow
calculations
experimental ICS
parameter
Identification
parameter
identification
Simplified
ICS
calculation
model
parameter
identification
performance
calculation
specific
ICS
system
test
short and long
term performance
prediction
MODEL CALCULATIONS
Fig . 1. Dutch projects on ICS model development and calculations.
COMPLICATED SIMPLICITY OF THE ICS SYSTEM
The operation of an ICS system is based on a direct heat transfer mechanism. In Fig. 2 the
ICS building construction is outlined. The system is characterized by heating through direct
solar radiation on the outer surface of the heat store. Hence, the store acts as solar
collector.
hot water out
store
Insulation
cover
transparant
Insulation
£o!d water in
Fig. 2. Outline of an ICS system.
-------
1444
Such a SDHW system looks like simplicity itself. However, appearances are deceptive.
Three major points of attention can be indicated:
the high degree of insulation to prevent too large heat losses;
the prevention of too high and too low store temperatures;
the construction of the ICS system.
As the complete system is placed outdoors the insulation has to be good. For the non-
collecting walls this can be achieved by using conventional insulation material. The solar
radiation absorbing surfaces need to be insulated in a different way. The heat loss
coefficient of a conventional solar collector is 4 W/m2K; for an ICS system this should be
less than 1.5 W/m2K. A transparant insulation material (TIM) might be used to achieve this.
These materials incorporate a good thermal insulation and a fair solar transmission.
The dimensions of ICS systems will be more critical than for conventional systems as it is
more difficult to avoid freezing and overheating of the system. This might occur under
extreme operating conditions such as no tap water draw-off during a cold period without
solar radiation or during a period with lots of sunshine. Requirements with respect to the
system parameters such as the heat loss, the transmission of the TIM, the collector area and
the store volume might be strict.
Best performance of an ICS system can be expected if the tap water is heated directly. No
tap water heat exchanger is needed then. Consequently, the system will be at mains
pressure which is hard to realize for a flat box shaped ICS. That is why the heart of many
. ICS systems consists of one or more cylinders.
As TIM's are rather new materials there is relatively little experience with respect to their
way of processing and durability. Much attention is paid to these materials within
international research programmes, such as those of the International Energy Agency.
Hence, more knowledge will be available soon.
ICS MODEL DEVELOPMENT
A calculation model is being developed in order to evaluate test results and investigate the
performance of tested ICS systems. In Fig. 1 this model is indicated as the simplified ICS
calculation model. The model development is supported by measurements on an
experimental ICS system.
Experimental setup
The most interesting in the model development is the distribution of the solar energy
over the store. Modelling of the solar radiation on the absorbing surface is very much the
same as for a solar collector with incident angle dependency and so is the heat loss.
Therefore, the experimental ICS system only consists of a well insulated water tank: a
rectangular water container of 1.60 (height) x 0.80 x 0.08 meter having an inclination of
45°. The solar radiation is simulated by electric heating of the upper surface.
Test sequences were carried out at different heating rates: 200 W/m2, 300 W/m2 and 600
W/m2. Temperatures were measured on the axis of the ICS system along with the height.
Figure 3 shows the temperature development for the test using 600 W/m2: first the water
volume is heated from a uniform temperature (a), then the ICS is partially discharged (b)
and finally the heating is continued (c). It is clear that microflows cause thermal
stratification irrespective the initial temperature profile. Six hours after the start of the
heating a top temperature of 60°C is obtained. Three-dimensional effects were found
negligible.
The internal temperature measurements were used to derive a function which describes the
distribution of the solar heat over the store. Secondly, as indicated in Fig. 1 these
measurements were used for evaluation and verification of the model for detailed flow and
heat transfer calculations (Ramaekers, 1991).
-------
1445
70
60 ~
ZOO m&n
SO -
160 min
40 ~
0 mint
1.6
.6
0
0.4
0.6
0.6
1.4
1.
0.2
1
Location in ICS system fmj
70
60
SO
10
0
0.4
0.6
0.6
1
1.6
1.3
1.2
Location in ICS systsm fmj
70
60
SS
SO
40
10
.6
0
0.2
0.4
0.6
0.6
1
1.2
1.4
1.6
1.
Location in ICS system fmj
3. Temperature development in the experimental ICS system during heating with 600
W/m2 from a uniform temperature (a), partial discharge (b) and continued heating
(O.
-------
1446
Identification of the distribution function
The simplified ICS model is a fixed segment model: for each segment the heat balance
containing all gain and loss terms is set up. The distribution of the solar energy over the
store segments is taken into account by defining two flows: one along the absorbing
surface, f|aye„ the other from this surface into the segment, fmijt. These flows should not be
mixed up with the real microflows which strongly depend on the operating conditions and
the location in the store. The flows f,ay(!r and 4ilt should only be used to describe the
distribution function: these are constant throughout a specific test and equal for all
segments. Top and bottom segments as well as those with large temperature gradients are
exceptional cases.
Parameter identification techniques (Van Dijk, 1991) have been used to identify f,ajr„ and
4,,. For the complete test sequence shown in Fig. 3 fUy„ = 2.8 g/s and fmiI = 0.56 g/s was
found. Figure 4 shows the resemblance between calculated and measured temperatures in
the top, middle and bottom part of the ICS system. The largest deviations come to 2 K.
60 -
40 -
5
36 -
30 -
26 -
10
300
400
100
0
Hma fm&rU
Fig, 4. Measured and calculated ICS temperatures during heating with 600 W/m2.
Further investigations revealed that the calculated temperatures are rather insensitive for
changes in the values of flay„ and fmiJ1; long term performance predictions for ICS systems
even more. This means that there is no need for a very accurate determination of the
distribution flow rates, and secondly that the solar gain characterictics of an ICS system
might be determined by a test using only one level of solar radiation.
USE OF THE ICS MODEL
Evaluation of ICS System Tests
Primarily the simplified ICS model will be used for evaluation of test results; see Fig. 1.
An indoor test method for ICS systems is being developed and will be tried out on a
commercially available system. The test method consists of these successive parts:
a charge temperature step test at the mains inlet or outlet of the system, e.g.,
from 20°C to 50°C, according to the CEC Test Procedures for Short Term
Thermal Stores (Visser, 1991). From this test the heat capacity and heat loss are
determined.
-------
1447
a direct discharge temperature step test to mains temperature, e.g. from 50°C,
using a flow rate as prescribed by the manufacturer; also according to the CEC
Test Procudures. This test yields the draw-off mixing.
heating with solar radiation and simultaneously tap water draw-off periods. The
draw-off profile is being investigated. This test reveals the solar gain parameters
such as the optical efficiency and the components of the distribution function.
All tests are of the black box type: no internal measurements are carried out. Only the flow
rate, the water temperature at the inlet and outlet of the system as well as the solar
radiation are measured. Parameter identification techniques in conjunction with the
simplified ICS model are used again for determination of the system characteristics.
Prediction of the Yearly Performance
After characterization of the tested ICS system the simplified ICS calculation model is
ready to determine the system efficiency and yearly performance predictions; see Fig. 1.
Also simulated test results, e.g.,as calculated by the detailed flow and heat transfer model
indicated in Fig. 1 (Ramaekers, 1991), might be evaluated in order to derive parameters for
the simplified model. In this" way long term preformance predictions can be coupled to
detailed ICS calculations. Hence, the effect of changes in the system dimensions on the
yearly performance can be investigated easily.
It is expected that after the completion of the projects indicated in Fig. 1 a coherent
package of tools will be available for both design and characterization of integrated
collector storage systems. The role of the simplified calculation model is to compile test
results of different types of ICS systems to a yearly performance prediction. The model is
less suitable for investigation of freezing and overheating. For this more detailed models
should be used which might be a problem for irregularly shaped ICS systems.
REFERENCES
Van Dijk, H.A.L. (1991). The PASSYS method for testing passive solar components.
Proceedings
ISES 1991 Solar World Congress.
Ramaekers, L.A.M. and C.J. van der Leun (1991). Integrated Collector Storage DHW
system numerical simulation of heat transfer and fluid flow. Proceedings ISES 1991 Solar
World Congress.
Visser, H. and H.A.L. van Dijk (1991). Test procedures for short term thermal stores.
Commision of the European Communities. Kluwer Academic Publishers, Dordrecht, the
Netherlands.
-------
I
-------
2.9 Solar Water Heaters
receding page blank
-------
-------
1451
IN SITU SOLAR STEAM GENERATION WITH CPC MODULES1
2
V.Balasubraraanian*, S.Jayaraman** ,
K.Perumal** and G.Sankarasubramanian***
*Radio Astronomy Centre, Tata Institute of
Fundamental Research, Post Box No.8,
Udhagamandalam (Ooty) - 643 001, India
**Department of Physics, S.R.M.Vidyalaya
Arts College, Coimbatore - 641 020, India
***G.M.R.T.Project Office, TIFR, Post Box No.3
Poona University Campus, Ganeshkhind,
Pune -411 007, India
ABSTRACT
Experiments were done to generate steam with the CPC modules
made of aluminised polyester foil and acrylic sheet reflectors
having glazed, nonevacuated heat receivers by boiling water
right inside the absorber tubes themselves. Out of the several
modifications carried out in the experimental set up in the
course of testing the CPC modules, the important one was to
reduce the water holding capacity of the absorber tubes without
reducing the absorber surface area. Instantaneous steam
generation efficiencies were determined. The results showed
that these CPC modules could be operated as stand alone solar
low pressure steam generators with steam generation efficiency
of about 40% or higher.
KEY WORDS
CPC trough; solar steam generation; modified absorber; absorber
wetting.
INTRODUCTION
In our earlier paper {Jayaraman'90) the details of two different
engineering designs (CPC I and CPC II) of double glazed, medium
concentration (2.5x) CPC modules with non-evacuated, selectively
coated receivers were presented. One of the glazings is a plane
transparent acrylic sheet of 2 mm thickness at the aperture
and the other is a borosilicate glass tubing closely surrounding
the absorber. CPC I uses a sheet and foil type reflector.
24 SWG aluminium sheet pasted with solar reflecting aluminised
1. Project funded by University Grants Commission
2. ISES Member
Preceding page blank
-------
1452
Polyester foil is supported by aluminium ribs of suitable shape
to form the required CPC trough. CPC II is of thick acrylic
mirror type. Curved half mirrors for CPC troughs are made
by thermo forming 3 mm thick aluminised acrylic sheets. The
half mirrors are supported by suitable T-shaped pillars. Both
CPC I and II are designed for 19 mm OD tubular absorber and
incorporate reflector truncation at top and bottom. The
truncation at bottom is to permit accommodation of the glass
tube envelope around the receiver. The experimentally deter-
mined F'qo value for both the CPC modules is nearly 60%.
A potential capability of compound parabolic concentrators
is to attain operating temperatures higher than the boiling
point of water. The time constant of CPC modules is
considerably less than that of flat plate collectors. Hence
rapid warm up and operation in the steam generation mode is
possible with CPC modules. The above said modules CPC I and
CPC IIj each consisting of five adjacent trough, are used to
try out generation of steam by boiling water right inside the
absorber tubes themselves, and the details are presented here.
PRELIMINARY EXPERIMENTS AND RESULTS
Preliminary experiments were carried out with 19 mm OD copper
tube absorber assembly. One end of the absorber tubes are
blocked; and the other ends are connected to a common header.
The bottom end of the header is closed leaving the upper end
open to the atmosphere. Initially the absorber assembly was
filled up with known quantity (nearly 2 litres) of cold
distilled water. A graduated measuring jar was positioned
below the output end of the header in order to collect any
possible over flow of water. A thermo electric pyrheliometer
driven by a suitable sun-tracking system was used for monitoring
the beam component of solar irradiance. A pyranometer in the
same plane as the aperture of the CPC module was also used
for measuring the total irradiance. Measurements of temperatures
at the inlet end, outlet end and ambient, and irradiances were
recorded at intervals of five minutes.
It was observed that the water temperature rose to about 98°C
in about 60 to 70 minutes. As the temperature reached about
75°C to 80°Cj some quantity of hot water was suddenly ejected
from the output end,and the same got collected in the measuring
jar. As the output temperature reached 98°Cja somewhat steady
output of steam started coming out accompanied by a spray of
hot water droplets. The output was allowed to escape to the
atmosphere. Though we did not have any arrangements for
measuring the; flow rate of steam during these preliminary
experiments, it was clearly noticeable from visual observations
that the rate of output of steam was continually decreasing
with time.
The average thermal efficiency rj' for the duration of the
experiment was computed by deducing the quantity of steam
produced, from the knowledge of the initial quantity of cold
water and the output of hot water. In computing this efficiency
diffuse and direct irradiances accepted by the CPC were
-------
1453
estimated along the lines suggested by Rabl (1980). There
were large day-to-day fluctuations in the value of average
thermal efficiency (24% to 44%). The continuous reduction
in output of steam observed could be due to progressively
insufficient wetting of the inner surface of the absorber tubes
by water during steam generation. To ensure wetting of the
inner surface of the absorber tubes throughout the experiment,
to measure the flow of output steam, to reduce the escape of
hot water in the form of spray and to reduce the initial warm
up time, the following modifications in the experimental set
up were incorporated.
To reduce the water holding capacity of the absorber tubes
and to ensure better wetting of the walls of the absorber tubes,
13 mm OD Aluminium tubes were inserted concentrically inside
the absorber tube, with both ends of the inner tube sealed
by brazing metal plugs at the ends. With this arrangement
the water holding space inside the absorber is only the
cylindrical annulus between the inner surface of the absorber
tube and the outer surface of the blocking tube. The support
arrangement for the inner blocking tube is shown in Fig. 1 .
Similar blocking arrangement was carried out for the header
tube also. By this modification the water holding capacity
was reduced to about 30% of the unmodified absorber assembly.
MODIFICATIONS TO THE EXPERIMENTAL SET UP
HEADER
BRAZED TO
INNER TUBE
BRAZING
BRAZED WITH
PLUG /
METAL PLUG
ABSORBER TUBE
WASHER WITH HOLES
BLOCKING TUBE
WASHER WITH HOLES
BRAZED TO INNER TUBE
FIGURE 1 MODIFIED ABSORBER
-------
1454
To measure the rate of steam flow, a Liebig's condenser made
of borosilicate glass was attached at the output end of the
CPC module. The outer jacket of the condenser was cooled with
a continuous flow of cold water. The condensate was allowed
to drop into a graduated measuring jar so that the cumulative
collection of condensate could be recorded as a function of
time from which flow rates could be computed.
To have a steady output of steam, constant wetting of the
absorber tubes should be ensured. For this,feed water reservoir
was included which would feed cold water at the bottom of the
header by gravity flow as water inside the absorber tubes
evaporates to steam. The reservoir is of adjustable height
by means of a length of flexible rubber tubing.
To prevent the entry of hot water into the condehser, a water
trap arrangement was used. It was made of cylindrical stainless
steel vessel of about 4" diameter and 3" depth, closed on all
sides except for two bits of short length piping attached to
diametrically opposite sides of the cylindrical surface. The
trap was included in between the absorber output end and the
condenser. The water trap was well insulated with wrapping
of glasswool on its outside surfaces. The complete experimental
setup with all the above modifications is shown in Fig. 2.
Experiments in in situ solar steam generation were carried
out with the modified setup which had a nonselectively coated
aluminium tube absorber.
COLO WATER
OUTLETv
CONDENSER
COLD WATER
INLET
STEAM
Insulation
constant .
HEAD TANK
WATER
TRAP
MAINS
TEMPERATURE
SENSOR ~~
(emfffA ¦
CPC
INSULATION
U-TUBE
FROM
MEASURING
JAR
TEMPERATURE
SENSOR
FIGURE 2 EXPERIMENTAL SET-UP FOR IN SITU
STEAM GENERATION
-------
1455
RESULTS AND DISCUSSIONS
The above modifications resulted in considerable decrease of
warm up time and also a steady output of steam free from hot
water droplets. In addition, significant increase in thermal
efficiency was also obtained. The modified modules gave
instantaneous efficiencies of 39% for CPC I and 29% for CPC
II. The possible reason for the improvement could be as follows:
The annular spacing between the inner and outer tubing of the
absorber assembly is only 2.5 mm. At such small spacing,
bridging of the two surfaces by drops or sheets of water might
be easily possible owing to effects of surface tension of water.
The action of the outer surface of the inner tube might be
to support layers of water all over the inner surface of the
absorber tube. In the absence of the inner tubes, wetting
of the inside of the absbrber tubes could be highly non-uniform.
Steady instantaneous steam generation efficiency rj. observed
for both the modules is higher by nearly 5%, when compared
to the values obtained without the feed water reservoir. We
can infer that the provision of feed water reservoir ensures
good wetting of the inside surfaces of the absorber tubes
thcughout the duration of the process of in situ solar steam
generation by CPC modules.
One can visualize extensive applications of such solar steam
generators in developing countries with tropical climates.
Two such areas of application are steam cooking of foodstuff
and sterilization of hospital tools such as syringes and
hypodermic needles. CPC-based solar steam generators operated
in the in situ mode described here makes them truly stand-
alone systems, without any need for auxiliary power sources
for pumps etc. From the results reported here one may note
that further improvement in thermal efficiencies are possible
by using selectively coated absorbers.
ACKNOWLEDGEMENT
The grant provided by University Grants Commission, New Delhi,
to carry out this research work is gratefully acknowledged.
REFERENCES
Jayaraman,S., V.Balasubramanian and K.Perumal (1990).
Development of compound parabolic concentrators in India,
Energy and the environment into the 1990s. Proceedings of
the 1st World Renewable Energy Congress Reading, UK.
Rabl, A., J.O1Collagher and R.Winston (1980). Design and test
of non-evacuated solar collectors with compound parabolic
concentrators. Solar Energy, 25, 335-351.
-------
1456
DEVELOPMENT AND ENHANCEMENT OF A
SOLAR BOOSTED DOMESTIC HOT WATER SYSTEM
J.I. Stewart, Senior Lecturer in Electrical Engineering
Ballarat University College, Ballarat, Victoria, Australia
ABSTRACT
In conjunction with various Government and Private Industry organisations,
staff at the Ballarat University College have been working over the last decade
on the development of an electrically powered solar boosted domestic hot water
system. Designed on the vapor compression principle, one of the primary aims
encompassed in the original concept was to develop a viable package capable o£
mass production at a realistic cost, whilst demonstrating as many of the
desirable properties of th^ Heat Pumping principle as possible.
KEYWORDS
Heat pumping systems; domestic water heating; variable speed control; efficient
energy usage.
INTRODUCTION
This paper describes work which has taken place in the last eighteen months
aimed at enhancing the operation of a production model solar boosted heat pump
based domestic hot water system, building on experience gained in the several
thousand units installed in Australia and overseas.
BRIEF DEVICE HSITORY
During the middle of the last decade, an Australian company, (Siddons-Ramset
Pty Ltd), released for production and sales a range of solar boosted heat
pumping hot water systems aimed at providing a new means by which domestic hot
water could be generated using electrical energy as the primary fuel source,
whilst at the same time offsetting the inherent inefficiencies of cost which
are normally associated with conventional resistance element electric storage
hot water units. (Carter, 1988) .
To cater for normal demand variations of differing consumer groups, three
capacity sizes are available, all sharing the functional layout shown in Fig. 1.
Based on the vapor compression principle, utilising a sealed motor/compressor
unit, the conventional refrigeration components of expansion valve, evaporator
and condenser are combined to enable low level heat gathered by the externally
mounted collector (evaporator) plates to be transferred to the storage tank at
temperatures in the range of 50 to 70 degrees Celsius.
-------
1457
Expansion Valve
Compressor
Solar Evaporator plates
Receiver
Water Outlet-dfl
Water Tank
Condenser Coil
Insulation
Water Inlet — d b
Fig. 1. System Functional Layout
A properly designed heat pumping system based on these principles can achieve
an operating efficiency, or co-efficient of performance (C.O.P.) figure of
anywhere between 200% and 500%, making the use of such equipment attractive
both in terms of primary energy consumed and any resultant energy related costs.
Equipment Based Operational Limitations
Keeping in mind the domestic market application, and remembering that the
original design took place in the early 1980's, current production has been
based around a single phase alternating current sealed compressor unit, with
voltages and frequencies appropriate to the various regions in which marketing
has taken place. Because of this unavoidable necessity to obtain power from
conventional fixed frequency alternating current (a.c.) supplies, the thermo-
dynamic design therefore has to be based on the assumption of a constant speed
motor/compressor combination.
This requirement was accepted at the outset as being a compromise in terms of
overall system efficiency, since it is known that performance optimisation can
be achieved on a dynamic basis by varying the displacement of the compressor
using some form of motor speed control, instantaneous speed being determined by
one or more system variables incorporated into a feedback control algorithm.
It was with the removal of this limitation as a goal that a project was under-
taken aimed at carrying out an indepth series of tests on efficiency relation-
ships as a function of such things as compressor speed, evaporator temperature,
ambient operating conditions, time of day and draw off patterns. Also to be
investigated was the possibility of optimising the efficiency of the system as
viewed by the consumer, that is a study of end costs for energy, necessitating
examination of such things as tariff structures and user expectations.
Figures 2 and 3, are photographs of the storage/compressor unit used as a
reference for tests carried out, and its associated externally.mounted evaporator
plates.
-------
1458
Fig. 2. Compressor/Storage Unit
Fig. 3. Evaporator/Collector Plates
-------
1459
MACHINE TESTING PROGRAM
As the proposed testing program was centred around obtaining sufficient inform-
ation about the operational characteristics of the motor drive component of the
sealed compressor unit, it was necessary to carry out numerous individual tests
relating to such parameters as input and output powers, machine speed, electrical
power factor, and of course overall efficiencies. Early attempts to modify an
actual sealed compressor unit to enable these tests proved unsatisfactory, due
mainly to the very special operating conditions for which these machines were
designed. A solution to this problem proved to be the acquisition of a conven-
tional stand alone general purpose motor with characteristics matched as closely
as possible to the heat pump drive.
Once this had been achieved, the above tests were carried out for many combin-
ations of applied voltage and frequency, with computerised data collection and
handling methods being used to automate the calculation of derived character-
istics .
At this stage in the research program it became apparent, that as originally
postulated, a form of variable frequency (in conjunction with voltage) control
was likely to yield the most satisfactory results for this application. This
conclusion raised the possibility of also considering, at the very least for
comparative purposes, the operation of a three phase machine of similar output
rating, as the technologies involved for three phase speed control are very
similar to those for single phase units.
As part of this consideration of three phase options, a computer model was
created which enabled the prediction of machine operation over a useful load
range, with subsequent theory/practice comparisons proving that such a model
was a useful tool for design purposes.
The Variable Speed Option
A decision to proceed with a conventional variable frequency control strategy
required as a next step the design and construction of a suitable custom made
motor speed controller (inverter drive), as an extensive market survey up to
mid 1990 failed to unearth a readily available commercial alternative. Reasons
for this include the historical lack of defined variable speed single phase
drive applications, as well as the relevantly recent emergence of cheap means
by which such controllers can be built.
Resources were committed at this point to designing, building and testing a
simple inverter drive, and to subsequently assessing its performance in the
crucial area of overall efficiency. Quite obviously, gains made by improving
fundamental heat pump efficiencies could not be offset significantly by inverter
introduced inefficiencies without cancelling the main aim of the program.
With a prototype drive unit developed sufficiently to allow bench testing of
performance, it was quickly apparent that when tested as a stand alone unit,
the inverter was capable of high efficiency operation, but that the generated
voltage waveform was likely to cause inefficiencies in the operation of the
motor due to extra generation of heat within the machine windings. Figure 4,
demonstrates the comparison of test machine efficiencies when operated from
both sinusoidal mains and inverter derived sources.
As predicted, machine performance with the prototype inverter drive suffered
significantly, and an analysis of current and voltage waveforms provided
explanations of that change.
-------
1460
SINGLE PHASE, .75 HP MOTOR EFFICIENCES
VERSUS SHAFT SPEED FOR TWO DRIVES.
EFFICIENCY (%)
80
SINUSOIDAL SUPPLY.
60
INVERTER SUPPLY.
40
20
SUPPLY FREQUENCY = 50 Hz.
1500
1400
1300
1200
SHAFT SPEED (RPM)
—— Sinusoidal Supply. —Inverter Supply.
Fig. 4. Comparative Drive Efficiencies
REFINED SOLUTIONS
Experience with machine testing, coupled with extensive experience with the
laboratory test system (Fig. 2.), showed that it was possible to define an
initial compressor control algorithm which took into account known significant
heat pumping variables, which was also capable of being implemented in an
improved version of the prototype inverter drive.
Evidence has also recently come to hand that independent equipment manufacturers
may be starting to show some interest in the single phase speed control market,
and tests are concurrently being performed on the first commercial example of
such a drive, with the aim of comparing all significant performance parameters
with the custom built unit.
USEFUL OUTCOMES
For the reasons outlined above, dynamic control of the type of heat pumping
system described will yield worthwhile improvements in the already impressive
performance figures that such systems are capable of. One extra benefit, which
in itself is a commercially useful outcome, will be that by the nature of the
inverter drives being utilised, the shift between countries which is necessary
for international marketing, but which also involves changes of supply voltage
magnitude and frequency, can be incorporated into the original package design.
The necessity to provide specially engineered models to suit different local
conditions should be eliminated.
-------
1461
REFERENCES
Carter, A. R. , (1988) . Energy Conservation in Practice - The Solarplu's Water
Heater.
Stewart, J. I., (1990). Modern Control Methods For Use With Solar Boosted
Heat Pump Systems.
Charters, W., and de Forest, L., and Taylor, L., (1984). Demonstration of
Solar Boosted Heat Pump Systems.
Drives For Heat Pumps. Proc. 2nd. Essen Heat Pump Conference (1978).
-------
1462
ICC SOLAR WATER TECMC-ECONOMIC ANALYSIS
WU QING
Guangzhou Institute of Energy Conversion
Chinese Academy of sciences
81 Xianlie Middle Road, Gyangzhou 510070, China
ABSTRACT
This paper illustrates the workings of the ICC solar water boiler, shows a system diagram
of the Pilot Experimental Installation, demonstrates its technical and economic feasibility
by comparative analysis of the system's operational conditions and construction investment
with a conventional energy water boiler, as well as provides the scientific basis for the
further popularization of its application.
I. The ICC Solar Water Boiler
This water boiler is a 3-stage open-cycle mid-temperature heating system which consists of
a flat-plate collector (FPC), a honeycomb collector (HC) and an Involute Circle Concentrator
(ICC), or a two-stage system with an FPC and ICC. It can gradually heat water to a boil and
Flat plate Constant temp. Honeycomb Vacuum tube Solenoid Water Cool water
collector water discharge collector collector valve guage tank (4m3)
valve
I Solar hot
- water tank
„ (4m3)
-N-i
(Upstair solar collector)
i—
Inverse
power
supply
Voltage
stabili-
zer
Display
control
device
Solar boiled
water tank
(0.5m3)
Downstairs
boied water
supply room
38o V
Fig. 1 System Diagram of the Solar Water Boiler
-------
1463
can supply boiled & hot water from two direct-current heat-supplying limes. To ensure a
boiled & hot water supply on rainy & cloudy days, the boiled-water tank is equipped with an
auxiliary automatic electric heater and the hot-water tank with an auxiliary steam heater
or a diesel direct-heater according to the user's choice. Fig.1 shows the system diagram
of the Pilot Experimental Installation.
II. The Analysis of the Technical & Economic Feasibility
1. Technical feasibility
The Pilot Experimental System of the multi-stage high-efficient solar water boiler was set
up and began operation in 1985. The first pipe-flush and repainting of the supporting
framework was finished in Oct., 1990. After that, it continued to work well. This system
was awarded a patent right by the Patent Office of China in 1988. Three dissertations on
the solar water boiler— "The Research of ICC Solar Collector" , "ICC Solar Water Boiler"
& "The Multi-stage High-efficient Solar Water Boiler" —were publicly published in this
order at two international conferences and at the 1989 annual conference of the Chinese
Solar Energy Society. Having been certified, the system was installed and used in railway-
stations, hotels,middle-schools as well as some places in Bangkok and Hong Kong. This has
proved that ICC solar water boiler is a feasible technology and has a bright and great
prospect on the market.
2. Economic feasibility
A. An analysis of the savings the heat supply system of the Pilot Experimental system
The heat supply performance of ICC solar water varied during the year. The analysis was
made on the basis on the data of solar resources collected in the area of Guangzhou, as
2
shown in Fig. 2 and in Table 1. 52m Pilot Experimental System.
Table 1.
Type of water
supply
Water output
(Kg/day)
Temp.* diff.
CC)
Quantity of heat
(Kcal/day)
savings annually
(KWH) (Yuan**)
Hot water(50'C)
708
27.3
193 28.4
8203. 5 1804.8
Boiling waterdOOT!
560
77.3
43288.0
18372. 7 4042.0
Total
62616.4
26576. 2 5846.8
* the annual average temperature of running water is 22.7 degrees.
** according to, 1 kwh costs o.22yuan
5000
(kg)
4000 ¦
2000
1500
3000
1000
2000
1000 -Back-up energy
500
6
10
2
(month)
0
4
8
Fig.2 Heat supply performance of ICC solar water boiler system (collection area 52m2)
-------
1464
B. The initial investement in the Pilot Experimental System and the operational expenses
The initial investment in components of the 52ffl2 Polit Experimental System are listed in
Table 2 . Unit area investment is
Cs = 234.6 yuan^m2
The operational expenses of Pilot Experimental System is quite low because the system is
automatic and no needs supervision. The annual operation & maintenance pen unit area,
Ms := 8.98 yuanym2
C. The initial investment, operational expenses and the expenses of the energy consumption
of a conventional energy water boiler, _ ,
STRUCTURAL
INVESTMENT
TATE
COMPONENTS
COST(yaun)
(M)
Collector
7675.0
62.9
Supporter
639.9
5.2
B.W. tank*
1223.2
10.0
H.W. tank
1055.9
8.7
Piping
607.5
5.7
Auto-control
1000.0
8.2
Total
12201.5
100
*B.W. = Boil ling water H.I. = Hot water
1) An electric water boiler,
The one selected, 112KW Xset Price, 17IT4.3 yuan
Neccessary accessories such as switches,'71. 7 yuan
Total, 1856.0 yuan
The initial investment being equivalent to the solar unit area,
Ce = 35.7 yuan/m2
The operation and maintenance expenses (supposing the electric water boiler is fully
Automatic), Me = 0
The power-consumption expenses, the daily demand of the electric boiler being equal to
the capacity of the 52m2 Pilot Experimental system,
62616. 4 Kcal
= 81flKg/day
77.3 Kcal/Kg
The power consumption of the electric water boiler, 0.12265 kwh/kg
The cost of electricity at that time, 0.22 yaun/kwh
When equivalent to the solar unit area, the annual power expenses are,
810X365 X 0. 11265 X0. 22
Ee= =139.66 yaun/a • year
52
2) A coal-fired boiler
According to the mini-boiler installed originally by the user, converted into,
The initial investment equivalent to the solar equipment's capacity
Cc = 403.85 yaunxm2
The operational maintenance expenses,
Mc = 35.93 yaun/m • year
-------
1465
Fuel expenses,
p
Ec = 21.82 yaun/m • year
D. Research of the Economic-math Model and the Calculations.
Having studied the economic-math model of the conventional energy and the solar energy
systems, and having made a synthetical analysis of the references 1—5,we've a discriminant
of the economic feasibility of the solar water boiler,
< n-1 n-1 \ f n+1 N
Y = C(l+r)n+ £ M(l+r>k+ £ E(l+r)k • (1+j )ntk - Cs(l+r)n+ £ Ms(l+r)k
^ k=0 k-0 ' K k=0
feasible, cheaper
= A-B^O
unfeasible, move expensive
In the discriminant,
A—The expenses when the service life of the conventional energy system is n years; when
^ the annual interest rate of the fuel cost is j.
B—The expenses when the service life of the solar energy system is as same as n year;and
the interest rate of the capital is r; and the solar energy supply is free, so item E
is equal to zero.
Supposing the annual rise of the power costs j = 3%, the rise in the price of coal j,
separately, = SH & 10%, and when compared with the solar energy system in both cases
considering and ignoring the initial investment of the conventional energy system , the
calculation are shown in the Table 3 and Fig. 3.
(year/a2)
7000
6000
5000
4000
3000
2000
1080
0
-1000
« 1 1 1 J
~ Compared with,
1.electric water boiler
" Y„ C ^ 0, M=0, 1=3%
2.coal-fired boiler
Y2,c, M*0, J=5M
Y3,c,M=0, J=5M
Y4, c, M*0, J=10.!¥
Y5,c,M=0, J=10%
1 1
Yi^
1 1
ft.
• '
_3-—-~
y3
1 1 . 1 1 1 1 t 1
10
12
14
16
NCyear)
Fig.3 Comparison of economic benefit of ICC solar water boiler
III. Conclusion
1. Compared with the electric water boiler
Yl (as in Fig.3) shows that the investment in the solar energy system may be paid back in
-------
1466
Comparison of Economic Benefits of ICC Solor Hater Boiler
Table 3
| Type
| Solor
| Electric
:i
Coal-fired lioiler
1 C( Yuan/in2 )
I 234,(4
I 35,67
I 403.85
1
0
I 403.85
0
1 H,(Yuail/in'-y)
1 8,58
I 0
I 35,93
I
0
I 35,93
0
1 li (Y u'a n / in1-v )
1 1
1 139.66
I 21.82
I 21,82
1 21,82
21,82 I
I ,T (% )
1 0
1 3
1
5
1
5
10
1
1
1
1 B
1
1 1 Yi-
1 A i 1
1 1 Ai-B
1
1 a2
1
1 Ya-
1
1 Az-B
1
1 Aj
1
"1 Yd =
1
1 A3-n
1
1 A4
1
IY„-
1
1 A,, (i
Aj
Ys" 1
As-B 1
i i
I m
I 182,34 -1-73,86
1 494,68
I 232.47
I 22,91
1-239,29
I 435,77
I 233,56
24
-238,2 I
1 2
i 291,95
1 344,95 | 53
1 593.84
1 301,89
i 48,7!
1-243,17
1 597,36
I 305,41 ¦
52,31
-239.64 I
1 3
I 324.03
1 524.89 | 209.83
I 792,06
I 378,01
I 77,911
1-246,15
I 709,65
| 385.6
85.49
-238,56 |
1 4
1 358,?
1 723,64 | 364,94
1 820,12
1 461,42
| 110,6
1-248,1
I 833,73
1 475,03
124.21
-234,49 |
1 5'
I 3%,B9
1 942,86 | 546.77
I 948.85
1 552,76
I 147,21
1-248,88
I 970,83
I 574,74
169,13
-226,9 |
1 6
! 436,44
1 1184,29 I 747,86
I 1039.1?
1 652.73
I 183,1
1-248,33
1 1122,31
I 685,87
221,24
-215,2 I
1 1
1 479,98
1 1449,85 | 9(9,87
I 1242.07
I 762.08
[ 233.71
1-246,28
I 1289,64 | 809,66
281.28
-198,7 |
1 8
1 526,98
I 1741,6 | 1214,62
I 1408,6
I 881,63
I 284.45
1-242,52
I 1474,49 | 947.51
350.34
-176,64 |
1 9
1 577.69
1 2861,76 | 1484.87
I 1589,95
I 1012.25 | 340,83
1-236,86
1 1678,65 | 1160,95
423.53
-148,16 |
1 IB
I 632.43
1 2412,74 | 1780,32 1 1787,34
I 1154,92
I 403,3?
1-229,06
1 1904,12
I 1271,69
520.151
-112,28 |
1 11
I 691,49
1 2797,15 | 21195,(6
I 2092.15
I 1310,65
I 472,63
1-218,86
I 2153.11
I 1461.62
623.6
-6?,9 |
1 12
I 755,24
I 3217,81 I 2462,57 J 2235,84
I 1480.59
I 549,25
1-205,99
I 2428,05
I 1672,81
741,47
-13.7? I
1 13
1 324,04
1 3677,76 | 2853,72
I 2489.99
I 1665.95 | 633.9
1-190,14
I 2731,61 | 1907,57
875,52
51,48 |
1 14
I 898.28
I 4180,28 | 3282
I 2766,33
1 1868,05
I 727,31
1-170,97
I 3066,75 | 2168,4?
1027,72
129.44 |
1 15
1 978,4
1 4728,35 | 3750,54
I 3066,71
1 2038.31
1 830,27
1-148,13
I 3436,71
I 2458,31
1200,27
221.86 |
1 16
1 1064,8?
1 5327,59 | 4262,72
I 3393,16
1 2323.28
I 943,66
1-121,22
I 3845,89
I 2780.22
1395,59
330.72 |
I 1?
1 1158,13
1 5980.38 | 4322.18 | 3747.84
1 2589,65
1 10(8,41
1-89,78
1 4295,84
1 3137,65
1616,41
458,22 |
1 18
1 1258.9
1 6691.78 | 5432,88
I 4133.11
I 2874.21
I 1205,54 1-53,36
I 4793,32
I 3534,42
1865,75
606.84 |
1 1?
1 1367.59
1 7466.(7 | (039.08
I 4551.52
1 3183,93 | 135(,15 1-11,43
I 5342,33
I 3974,74
2146,96
779,38 |
I 20
1 1484.88
1 8310.27 | (825,39 | 5005,82
I 3520.95
I 1521,46 | 36.58
1 5948,16
I 4463,29
2463.8
378,92 |
A. D=--C (1 +r) An + EM(I +r) Ak
A. A = C(Hr)An + EM(l+r)Ak + EE(l+r)Ak(|+J)A
-------
1467
about two years. If the service life of a solar water boiler is 15 years, the benefit
would, be near ¥ 4000yuan/m2 of the collector. If 20 years, the benefit Could be near ¥ 7000
yuan/m2.
2. Compared with the coal-fired boiler
When the initial investment of the equipment and the maintenance costs are taken into
account (ie. C,M^O), Y2, Y4, as in Fig. 3, show A>B, so the benefit will be gained in the
beginning; when they aren't taken into account (ie. c,M =0), Y5 shows that the investment
would take 12 years to be paid back in towns where the coal price goes up relatively fast,-
and it would take 18 years in the places where the coal price is low and stable, as Y3
shows in Fig.3. In a rational comparison the investment costs of the equipment and the
maintenance costs should be taken into account.
In the places where solar energy is abundant, the conventional energy supply is short, or
the power costs are high, the use of the solar water boiler to replace the conventional
energy water boilers has a great social-economic benefit.
REFERENCES
1. Peking Electric Tube Factory, "Techn-economic Feasibility Report of the Vacuum Tube
Solar water Heater" Nov. 1978.
2. J.A.Duffie, W.A.Beckman, "Solar Energy Thermal Processes" . 1974.
3. S.Y.Szokolary, "Solar Energy and Building" znd edition, 1977.
4. Guangzhou Institute of Energy Conversion , Chinese Academy of Sciences, "Technical
Evaluation Data Collection of a Hybria Solar Drier Pilot Experimental Installation"
P2.8 ~ 2.11 & P4.l~4.12.
5. Lou Wei-Qiu, "Solar Thermal Energy Conversion Engineering" part 2 Pi 1 ^—15. Shanghai
Institute of Machinery, China.
-------
1468
TRANSPARENTLY INSULATED SOLAR POND
M. Rommel, V. Wittwer
Fraunhofer Institut fur Solare Energiesysteme
Oltmannsstr. 22, 7800 Freiburg, FRG
ABSTRACT
A shallow, transparently insulated solar pond was investigated, which was designed for
applications in the temperature range around 80°C. Measurements with and without heat
extraction were carried out to determine the collector characteristics data t)Q and U. The
investigations show that, using a transparently insulated solar pond, a continuous 24-hour energy
supply with approximately constant operating conditions can be provided.
KEYWORDS
Solar Pond, Shallow Solar Pond, Transparent Insulation Materials, Integrated Storage
Collector, Desalination, Membrane Distillation
INTRODUCTION
In general, storage elements with transparently insulated walls form a new concept (Rommel,
1987a) which became possible due to improvements in Transparent Insulation Materials (TIM)
(Wittwer, 1989). In principle, thermal energy can be stored without losses, if the losses can be
compensated by solar gains during the day. If the gains exceed the losses, the storage unit warms
up and the thermal energy can be utilized. A storage unit with transparently insulated walls thus
forms a complete solar system consisting of storage and collector components.
Following this concept, an integrated collector storage for domestic hot water, which is
applicable to Central European climatic conditions, has already been developed successfully at this
institute (Goetzberger, 1987; Schmidt, 1988, 1990). In Israel, transparently insulated solar ponds
are used for the centralized supply of domestic hot water in villages.
Since 1989, a transparently insulated shallow solar pond has been investigated at the institute.
Our experiments aim especially at applications in the temperature range near 80 °C. An important
example for a possible application is continuous supply, i.e. a supply for 24 hours a day, of thermal
energy for seawater desalination by membrane distillation (Fane, 1987).
SOLAR POND DESIGN
Fig. 1 shows the design of the investigated shallow solar pond. The transparent cover consists
of a low iron glass pane and a 10 cm thick honeycomb structure made of polycarbonate. As the
experiments aim at the temperature range around 80 °C, a selective absorber and an air gap
between the honeycomb and the absorber are essential. The opaque insulation is 16 cm of
polyurethane. The absorber area is 2.69 m x 1.92 m and, with a water depth of approx. 25 cm, the
total water volume is 1300 litres.
-------
1469
Fig. 1. Design of the investigated transparently insulated solar pond
global
radiation
o
outlet
middle
bottom
inlet
T3
O
in
13.7.90
12.7.90
11.7.90
10.7.90
10.7.90
11.7.90
12.7.90
13.7.90
T °C
aw>
Ta.°C
VTa-°C
42.2
50.9
57.2
57.1
15.6
16.9
20.4
22.4
31.6
34.0
36.8
34.7
H, Wh/m day
3344
7254
7864
7509
"^outlet' ^
—
57.9
65.7
61.6
out kWh/day
0
4.0
9.8
13.83
Fig. 2: Measurements July, 10-13, 90.
m denotes the mass flow rate during heat extraction, measured in 10 l/h, TL-.-jj,. and Thnftnm are
water temperatures in the storage tank.
The table summarizes the 24 h mean values of the storage water temperature Tw the ambient
temperature Ta, the temperature difference (T -Ta), the outlet temperature TQu^ef the sum of the
daily global radiation H and the total extracted heat out.
-------
1470
The water storage tank of the investigated pond, built for experiments at the test area of the
institute, is made from iron sheets. Along the absorber edges, which due to the absorber design are
not wetted by the stored water, heat conducting fins have been soldered onto the absorber. It
has to be mentioned, however, that we encountered condensation problems, presumably caused by
a leaking seal between the bottom part of the storage tank and the absorber forming the top.
MEASUREMENTS
From July 1989 to May 1990 the transparently insulated solar pond was investigated exclusively
under 'stagnation conditions'. This means that neither heat nor water was from the storage tank.
Consequently, the mean water temperature follows the weather conditions and typically varyied
between 10°C during the winter months and 90°C during the summer months.
From May to September 1990 measurements with heat extraction were carried out. The water
of the storage tank was pumped through a water to air heat exchanger in a closed cycle. Thus only
heat was extracted from the solar pond, without changing its thermal mass.
As an example, fig. 2 shows some of the more important measurements taken. (In total, 12
temperatures were measured.) During these 4 consecutive days in July 1990, the solar pond was
operated in three different modes: On the 10th without extraction of heat (under 'stagnation
conditions'), on the 11th and 12th with extraction of heat intermittendly during daytime, and
finally on the 13th with continuous heat extraction over 24 hours. This last operation mode is
particularly important for applications, as will be explained later.
EVALUATION OF MEASUREMENTS
Fig. 3 shows the overall heat losses of the solar pond that have been determined from the
cooling of the stored water during the night. For comparison, the solid line in Fig. 3 also shows the
U-value which was expected from theoretical considerations. The difference is believed to be
caused mainly by the condensation problems already mentioned above.
Further, the efficiency curve of the solar pond has been determined in the following manner: If
the energy balance equation of the storage tank is integrated over 24 hours and then is divided by
the total radiation on the absorber area, the 24-hour efficiency is obtained:
output (Tw - Ta)
= 7? - U , where:
A I 24h I
t2
output = (mc)st (T24h - Toh) + J m c (Tout - T±n) dt
tl
Here (mc)s{ denotes the thermal mass of the storage medium and storage tank, T24h an<^ ^0h
the mean water storage temperatures at 24:00 and 0:00 h, T^y, Ta, I the 24 h mean values of the
storage water temperature, the ambient temperature and the global radiation, t^ and the start
and the stop time of the heat extraction, TQUt and T^ the outlet and inlet temperature during heat
extraction, m the mass flow rate and c the specific heat of the stored water.
Under stagnation conditions the second term of the output equals zero and only the
temperature increase of the stored water during 24 hours is regarded as the collector-'output'.
Fig. 4 shows the efficiency curve determined from all measurements under stagnation
conditions (July 89 to May 90, 234 measuring days). The collector data T)a = .55 and
U = 1.72 W/(m^K) were deduced from the linear regression line as annual mean values.
-------
1471
3.00 ——
U 2.75
CjJ
'e 2.50
Fig. 3. Measured U-values (dots), determined from night cooling of the solar pond, compared to
expected U-values from theoretical considerations (solid line).
(Tw-Ta)/I, K m2 W"1 (Tw-Ta)/I, Km2W"1
Fig. 4. Efficiency curve, determined
from measurements
Fig. 5. Efficiency curve, determined
from simulated solar pond water temperatures.
-------
1472
However, these values depend strongly on the time of the year. As the solar pond was operated
under stagnation conditions and thus had higher temperatures during summer, the U-value is also
higher during summer. At the same time, a higher value of r)Q is also measured during the
summer. This is caused by better transmission conditions, as the sun then stands higher in the
zenith. In contrast to instantaneous collector efficiency measurements, here the effect of the
Incidence Angle Modifier is included in the value of r)Q, as the efficiency is determined from
measurements integrated over 24 hours. (The angle dependent transmission of the cover of the
solar pond is given in (Rommel, 1987b).) In Table 1 all results are summarized.
TABLEi. Data of different Transparently Insulated Solar Ponds:
year summer winter
solar pond (measured): r?Q= 0.55 0.68 0.49
U, (W/m2K)= 1.72 2.2 1.0
solar pond (simulated): T)Q= 0.57 0.68 0.42
U, (W/m2K)= 1.54 2.01 1.10
AREL solar pond (sim.): JJQ= 0.50 0.58 0.38
U, (W/m2K)= 1.97 2.42 1.40
Furthermore, Table 1 contains a comparison to solar pond data determined from simulation
calculations. Detailed hour-to-hour simulations were carried out to calculate the temperature of
the water in the solar pond under stagnation conditions. From these calculated temperatures, the
efficiency curves have been determined in exactly the same manner as was used for the evaluation
of the measured data and has been described above. Firstly, a solar pond with the same desibn as
the investigated pond was simulated. Additionally, a solar pond with a different cover, consisting of
a low iron glass pane, 10 cm honeycomb structure, a second glass pane and an air gap above the
open, not separately covered storage water was simulated. This cover construction was used in
Israel by the AREL company for solar ponds supplying domestic hot water to villages. Fig. 5 shows
the efficiency curve for the experimental solar pond, determined from simulation calculations.
HEAT EXTRACTION
The features of a transparently insulated solar pond as a solar system for a continuous 24 hour
heat supply (for example in a membrane distillation system) shall be explained using the
measurements of 13th July 1990, see Fig. 2. On this day (24 h mean value of the irradiation
I = 316.12 W/m2), the solar pond delivered 13.83 kWh in total, which corresponds to
114.13 W/m2. The mean outlet temperature was 61.6°C (min 54°C, max 70°C). The mean ambient
temperature was 22.4°C (min 14°C, max 29.2°C), thus the mean temperature difference was
39.2°C. Therefore, the daily efficiency of the solar pond as a complete system was 36.1%.
It is interesting to compare this to the daily collector efficiency of an improved flat plate
collector with honeycomb insulation (Rommel, 1987b), which has the same cover construction as
the investigated solar pond. Under the weather conditions of that day a - naturally higher - daily
collector efficiency of 60% can be reached. However, the useful energy is then not delivered
smoothly and spread out over 24 hours, but fluctuates with the irradiation conditions.
-------
1473
3*
T3
u
.53
£
€0
"O
10000
8000
6000
4000
2000
AT, C:
— 60 -40
continuous 24 hours output, W m'
fig. 6. Comparison of the performance of different solar ponds at LT =50K (dotted as measured, solid
as simulated, dashed AREL as simulated). The solid lines also compare the performance at different
temperature levels (kT-70, 50 and 30 K) for the solar pond as simulated
CONCLUSION
The advantages of the transparently insulated solar pond are its simple construction and thus
its cost effectiveness. Further, a continuous 24 h energy supply can be provided with approximately
constant operating conditions. This may have a very positive effect on the efficiency of the system
which is supplied with thermal energy by the solar pond, for example the seawater desalination by
membrane distillation.
Summarizing, Fig. 6 shows which continuous output power can be expected from different solar
ponds. To estimate the performance for sun rich areas, the values of the 'summer' results from
Table 1 have been used. For example, to get a continuous heat supply of 60 W/m^ (i. e. 1440
Wh/m day) at 50°C above the ambient temperature it is necessary to have average daily radiation
conditions of 7500 Wh/(m^ day) for the AREL solar pond, but only 5700 Wh/(m^ day) for the
pond investigated by us (both based on data from simulation results). Fig. 6 also shows that for
high temperature applications a cover design like the one investigated by us is necessary, whereas
for lower temperatures simpler designs are also acceptable.
REFERENCES
Fane, A.G., R.W. Schofield and C.J.D. Fell (1987). Desalination 64. 231-243
Goetzberger, A. and M. Rommel (1987). Solar Energy 39.211-219
Rommel, M. and A. Goetzberger (1987a). Proc. ISES Solar World Congress. Hamburg. Pergamon
Press, pp. 1553-1557
Rommel, M. and V. Wittwer (1987b). Proc. ISES Solar World Congress. Hamburg. Pergamon
Press, pp. 641-645
Schmidt, Ch., A. Goetzberger and J. Schmid (1988). Solar Energy 41.487-494
Schmidt, Ch. and A. Goetzberger (1990). Solar Energy 45. 93-100
Wittwer, V. (1990). Proc. 1st World Renewable Energy Congress. Reading. Pergamon
Press, pp. 1344-1351
-------
1474
COMPUTER PROGRAMS FOR DESIGNING
AND SIMULATING OF A SWIMMING POOL SOLAR HEATER
D. Gudiflo, M.P. Salas and J.J. Hermosillo
Unidad Academica de Tecnologia Intermedia
Division de Ingenieria, ITESO.
Fuego 1031, Jardines del Bosque,
Guadalajara, Jal., 44520, Mexico.
ABSTRACT
A structured software to simulate the heating of a swimming pool by
means of solar energy is presented. The user can perform a numerical
estimation of energy balance and temperature increments. Solar
collectors and pool covers can be considered in calculations for any
place and date. In addition, hydraulic calculations to design the
solar collector can be performed as well as its related costs.
Experimental results have been found to be in reasonable accordance
with predictions.
KEY WORDS
Software; swimming pools; solar heater desing; hydraulic design;
swimming pool covers.
INTRODUCTION
Solar energy is attractive from the ecological viewpoint, and as an
alternative to fossil fuels. However, when fossil energy is
relatively cheap, as in recent years, there are only a few
applications where solar energy is economical. In the thermal field,
these applications are generally related to the use of large amounts
of heat at low temperature.
One of these applications is the use of solar energy to warm the
water in a swimming pool. When utilized in an efficient form, the
sun can supply the energy required to keep a swimming pool at
comfortable temperature (22 to 30 °C depending on use of the pool and
weather).
Several papers about solar swimming pool heating have been published
in recent years (Govaer, 1981). Most of them are numerical estimates
of the thermal response of the swimming pool to the heat incomes and
losses. Usually, the energy input due to solar collectors is
considered in addition to the direct incidence of solar radiation
into the pool (De Winter, 1975; Fernandez, 1985; Szeicz, 1983).
Performance of pool covers to avoid or reduce thermal (convective
and radiative) and evaporative losses have also been studied
-------
1475
(Czarneki, 1963; Francey, 1980, 1981a, 1981b; Szeicz, 1983). It has
been shown that the simplest and most economical way to heat a
swimming pool is by means of a floating cover. Unfortunately, in
non-tropical latitudes or in high places, the use of a cover is not
enough to maintain the pool at comfortable temperatures the whole
year.
The software we describe in this paper is based mainly in De
Winter's work (1975). He compiled the equations required to perform
not only the thermal balance, but the hydraulic design of a swimming
pool solar heater, along with some useful tips for its construction.
The energy balance is performed considering the following heat
fluxes: a) Heat income by means of the solar collector (bare or
glazed), b) Heat income directly over the pool, c) Heat losses due
to radiation, convection and evaporation. These losses can be
reduced by means of the pool cover. In the software, the use of
various covers can be considered.
The program is able to simulate some meteorological variables used
in heat balances, like solar radiation, air temperature, humidity,
etc., by means of mathematical models published elsewhere (Duffie,
1980; Fernandez, 1990). Alternatively, these data can be obtained
experimentally and fed manually or read from a secondary file. The
software allows the user to change any primary variable, in a
dynamical form, so that the results affected by that change are
immediately corrected. User interacts with program by means of menus
and submenus. This fact makes a friendly program. Results from
design and simulation can be displayed in tabular or graphical form.
ENERGY BALANCE AND ESTIMATION OF POOL TEMPERATURE
Consider a swimming pool exposed directly to sunshine and heated, in
addition, by solar collectors. Let QcoU be the rate of heat income
into the pool due to solar collectors (W) , Qp the rate of heat income
directly over the pool (W) , and Ql the rate of heat losses (W) . Then,
the net rate of heat, Qn (W) over the pool is:
Qn = QcoU + QP " Qi (1)
Qn produces a rate of temperature change in the pool, accordingly
with the equation:
Qn = mCp dT/dt (2)
where mCp is the heat capacity of the pool (J K1), considered to be
the heat capacity of the water contained in the pool. T is
temperature (K) and t represents time (s) . In simple models, Qn could
be an easily integrable function of time. However, accounting for
efficiency factors in solar collectors and heat losses from the
pool, creates difficulties obtaining an analytical solution of equation
(2).
Equation (2) may be expressed in a finite difference form. Qn is
estimated at initial time, t0, where T is known. Then, equation (2)
may be solved for small time increments. The temperature increment
is added to the initial temperature to obtain the new value of T
corresponding to the next time step. In the new conditions Qn is
-------
1476
evaluated again. This is a cyclic process that is performed until
the desired time is reached.
In essence that is what software does A user can simulate the
behavior of temperature of a few hours, a single day (with or
without night) or a full week, to study the accumulated effect of
making changes in heater design, using a pool cover, etc. If weather
data (temperature, humidity and wind velocity) are simulated or read
from a secondary file, sunrise is taken as starting point for
calculations.
Following De Winter (1975) , the components of Qn in equation (1) are
given by:
where F3 is the overall efficiency of the spollector; Qideal is the
ideal f^.ux of energy in the collector (W m ) ; Ac is t|ie collector
area (m ) ; Q; ijp the flux of energy over the pool (W m ) ; Ap is the
pool area (m ) ; e is the emissivity of water; R is the radiation
heat lops to the sky of a black body at ambient temperature (about
79 W m" ) ; hr is the radiation heat transfer coefficient (about 5.7
W m" K~ at ambient temperature); hca is the convective heat transfer
coefficient, a function of wind velocity (about 8.5 W m" K~ for a
wind velocity of 10 km/h) ; TM and Ta are temperatures in the pool
water and ambient, respectively; Pw and Pa are vapor pressures of
water at pool temperature and ambient temperature, respectively.
Equations (3-5) require a lot of estimations or experimental
measurements to be ^solved. The next paragraphs describe the form
fSr performing these estimations.
Meteorological variables are calculated using appropriate models.
For beam diffuse irradiance in clear atmosphere, the Hottel's
model is used (Duffie, 1980) . For ambient temperature, two sine
curves are used to interpolate the variation from minimum to maximum
temperature and.viceversa (Fernandez, 1990). For humidity, the vapor
pressure of water is supposed to be constant during the day (constant
absolute humidity).
Qp takes different values if a cover is or is not used, and depends
on the type of cover. If a transparent cover is used, Francey
(1981b), estimates that 60% of the incident radiation is absorbed by
water. If cover is opaque, sunlight does not reach the water, and Qp
is very small.
In equation (5), the first term represents radiative heat losses to
the sky; the second one represents radiative heat losses to ambient;
the third one represents convective heat losses and the last one
represents evaporative heat losses. This term represents the main
percentage of the total losses. Similarly to Qp, Qt takes different
values depending on the use of a cover and the kind of cover used.
If a transparent cover is used, evaporative losses are practically
reduced to zero; convective losses are greatly diminished. In this
conditions, the two last terms in equation (5) are removed. If the
Qcoll ~ F3 Qideal) Ac
Qp = 0.8 Qi Ap
Qt = Ap[eR + hr(Tw-Ta) + hca(Tw-Ta) + 200hca(Pw-Pa)
(3)
(4)
(5)
-------
1477
cover were opaque, the radiative losses to the sky would be also
greatly diminished. In this case, the first term is also removed.
The cover temperature is considered to be equal to the water
temperature and the equation (5) conserves only the second term.
Some ambiental values should be supplied, like maximum and minimun
ambient temperature, relative humidity at sunrise (when usually is
highest) and initial water temperature. Also, the user should define
the periods of day when the solar heater will be under operation and
the cover will be on the pool. The program determines heat gains and
losses along the day, at specified time increments, as well as pool
temperature.
DESIGN OF SOLAR ENERGY COLLECTOR
The design of a swimming pool solar collector includes two main
parts: thermal design and hydraulic design. The purpose of a solar
collector is to convert sunlight into heat and transfer it to the
pool. To be economical, this should be done at the minimum cost. Due
to the low operation temperature of this system , the solar
collectors may be bare or glazed. In the software, both options are
considered along with its costs.
Thermal Design
Equation (3) describes the heat collected by solar collector. The
overall efficiency factor F3 depends on collector design, on rate of
water flow, and on wind velocity. Estimation of F3 requires some
previous calculations. When a certain sketch of the solar collector
has been made (tube diameters, tube spacing, etc.,) it is possible
to find other efficiency factors. First, the fin efficiency should
be estimated. This calculation includes geometrical and physical
variables, as the fin lenght and thickness, thermal conductivity of
the material and heat losses. The fin efficiency is then used to
calculate the section efficiency that includes other physical
parameters, as tubing diameter, solder conductance and convective
heat trasnfer from tubing to flowing water. Section efficiency is
then used to calculate the overall efficiency, F3, including water
flowrate and other variables.
This software calculates the fin, section and overall efficiency
after De Winter (1975). There are many variables that influence the
final value of F3. This part of the software allows the user to
change any geometrical or physical property of the collector and
find immediately the corresponding values of the efficiencies. The
optimal tubing spacing is also found.
Hydraulic Design
The objective of the hydraulic design is to set the diameter and
number of tubes and headers to obtain the desired flow rate and a
reasonable pressure loss. In the collector tubes, the pressure loss
should be approximately 0.2 bar. In the headers, the heat loss
should be 1/3 to 1/4 of the pressure loss in the tubes. The
procedure to calculate the pressure loss is described in the next
paragraph.
-------
1478
a) Calculate the volume of the pool, b) Define the time of
recirculation and find the required flow, c) Define the length and
width of the solar collector, d) Define the separation between
collector tubing and calculate the number of tubes required, e)
Define tubing diameter, f) Calculate the flow rate in each collector
tube, g) Calculate the pressure loss in each tube, h) Obtain the
pressure loss in the whole collector.
As in the thermal section, the user may change any dimension and the
corresponding results will be changed immeditely.
ESTIMATION OF COLLECTOR COST
When the thermal and hydraulic design are made, the dimensions of
the solar collector and the various materials are defined. Using a
special file containing unit prices, the program calculates the
total cost of the collector, including glazing and boxes, if
required in the design. If unitary prices are different from
information contained in the special file, they can be modified and
saved in such file, if desired. As in previous sections, any
modification in design parameters produces an immediate change in
cost estimation.
EXPERIMENTAL RESULTS
Experiments have been performed v^ith two different swimming pools.
The first one has an area of 2^8 m and mean depth of 1.9 m. The area
of the second one is 117 m and its mean depth is 1.8 m. Air
temperature, water temperature, global radiation, and relative
humidity were measured along the day. Wind was not measured, but
observations indicate that there was no wind. Therefore, for
simulation purposes, hca was considered 4.25 W m K , the minimum
recommended.
Both pools were heated only by direct sunshine over the pool. The
smaller one was studied with and without a cover (blue polyethylene
0.17 mm thick). Table 1 shows a comparison of simulated and
experimental results. Global irradiance, Q,-, initial water
temperature, Twj, maximum water temperature, Tumax, and temperature of
water at next day sunrise, Twnd, are shown. For measured data,
percentage of shadowing on the pool is also shown.
TABLE 1. Comparison of Simulated and Experimental Data.
SIMULATED
RESULTS
EXPERIMENTAL
DATA
Date
T •
Wl
°C
KJ/m dia
T
wmax
•c
Twnd
•c
Qi 2
KJ/m dia
T
wmax
•c
Tur>d
•c
Shadow
%
02/01/91
02/11/91
02/12/91*
17.9
19.4
19.9
20225
21364
21483
19.0
20.8
21.2
17.7
19.7
20.7
18563
19282
21202
19.7
21.2
22.0
17.8
19.9
20.5
32
23
23
* Results with a pool cover.
-------
1479
DISCUSSION AND CONCLUSIONS
The preliminary results in Table 1 show that measured next day
temperatures, Twnd, are similar to calculated results. However,
measured temperatures during the day are slightly higher than
predicted ones, as may be noted from the columns of T^^. It was
expected to obtain higher results from the simulation, taking into
account that simulated radiation was higher than measured radiation
and that simulations did not consider shadowing, as did the real
thing. Posterior experiments under a variety of meteorological
conditions and geometrical configurations of pool and solar heater
will produce information to adjust the model to obtain better
results.
REFERENCIAS
Czarneki, J. T. (1963). A method of heating swimming pools by solar
energy. Solar Energy. 7(11. 3-7.
De Winter, F. (1975). How to Desina and Build a Solar Swimming Pool
Heater. Copper Development Association Inc. New York.
Duffie, J. A. and W. A. Beckman (1980) . Aviable solar radiation.
In Willey-Intercience (Ed.), Solar Engineering of Thermal
Processes. J. Willey & Sons, New York. pp. 62-81.
Fernandez, J. L. (1985). Modelo para disenar calentadores solares
para albercas. Memorias de la IX Reunion Nacional de Energla
Solar. Merida, Yuc. pp. 129-132.
Fernandez, J. L. , and N. Chargoy (1990). Destiladores Solares de
Agua. Curso de Actualizacion en Eneraia Solar. UNAM, Mexico,
D.F.
Francey, J. L. A., and P. Golding (1981a). The wheathering of solar
pool covers. Solar Energy. 26. 237-242.
Francey, J. L. A., and P. Golding (1981b). The optical
characteristics of swimming pool covers used for direct solar
heating. Solar Energy. 26, 259-263.
Francey, J. L. A., P. Golding, and R. Clarke (1980). Low-cost solar
heating of community pools using pool covers. Solar Energy.
25. 407-416.
Govaer, D. and Y. Zarmi (1981). Analytical evaluation of direct
solar heating of swimming pools. Solar Energy. 27(61. 529-533.
Szeicz, G., and R. C. McMonagle (1983). The heat balance of urban
swimming pools. Solar Energy. 30(31. 247-259.
-------
1480
FLOW PATTERN AND HEAT EXTRACTION FROM A HORIZONTALLY
DISPOSED SINGLE ENDED EVACUATED ALL GLASS SOLAR
ABSORBER TUBE
Lu Weide, Wang Liting and Ho Zhichen
Beijing Solar Energy Research Institute, Beijing, China
ABSTRACT
In this paper, the flow pattern inside and heat extraction mechanism from the
single ended evacuated all glass solar absorber tubes which are horizontally
connected to a vertical header pipe has been studied with aid of visual tech-
nique. Pictutfesclearly show the processes of natural convection taki'ngplace in
such special partial enclosure. Quantitative experimental study on natural
convection from the absorber tube to,header pipe was conducted and the results
are reported as well.
KEYWORDS
Flow pattern; visual technique; heat transfer; single ended evacuated all glass
solar absorber tube; natural convection.
INTRODUCTION
Single ended all glass concentric tubular evacuated collectors were success-
fully developed in some laboratories in past ten years and are commercially
available nowadays. Their main advantages are considered as to be able to
achieve efficient operation at temperature range 70 - 80°C above ambient, sim-
ple structure and at low cost. Some simplified fluid in glass manifolding sys-
tems have been developed. A typical design has collector tubes with open ends
up vertically connected to a single header pipe without any feed tubes or par-
titions inside ( so called N - S type ). In this system, heat is extracted to
the header pipe by buoyancy effects and fluid flow between the header and sto-
rage tank could be established either by forced or thermosyphoning circulation.
This type of manifolding design, however,sti11 leaves the risk of fluid free-
zing as the collector system is not diai.nable. Another option, which has the
absorber tubes horizontally disposed with their open ends connected to a ver-
tical header pipe symmetrically or at its either side ( so called E - W type ),
was developed in Department of Electronics Engineering, Tsinghua University,
China. This system has the advantage of easy drainage, thus it can avoid freeze-
damage as well as fully use the hot water contained in the collector tubes.
The performance of N - S and E - W system was tested side by side. It is shown
that the E - W system got a whole day useful gain 6 % more than that of the N
- S system , even though the daily efficiency of N - S system is 10 % higher
than that of E - W system, since the apparent amount of hot water contained in
the absorber tubes can not be neglected ( Wu Jiaqing, 1985 ) ¦.
In China, the E — W systems have been extensively demonstrated for year
-------
1481
round hot water ^upply as well as for so.la,r wine aging( and got encouraging re-
sults, yet still leave some problems in heat extraction , particularly in forced
circulation or once through system operation.
Up to now, knowledge about the heat extraction from such horizontally dis-
posed collector tube system is notyet obtained.Window and Harding (1983) repor-
ted their research on buoyancy effects and heat extraction mechanism of N - S
system based on detailed measurement of temperature on tube banks as well as
temperature of inlet and outlet flows. This paper deals with the mechanism stu-
dy of flow pattern inside and heat extraction from an E - W collector tube with
the aid of visual technique. For collector performance evaluation, the natural
convective heat transfer coefficient in such system is also experimentally
studied.
VISUAL TECHNIQUE FOR FLOW PATTERN STUDY
A dye tracing technique was used to show the development of natural convective
flow pattern inside the horizontally disposed single ended glass tube. The hea-
ting element is a semi-conductive thin film deposited on the outer surface of a
glass tube prepared by sputtering technology hence possessiftgthe transluctant
characteristics. Thus, both dye tracing and transluctant electric heating film
on glass surface create a simple but effective visual technique, and the pic-
tures of flow pattern can be easily taken by an ordinary camera. In this ex-
periment, the electric heating film is divided into two parts which are connec-
ted with two electrodes respectively. The ratio of power dissipation to the top
and the bottom p^rts of the tubeis taken as P /P = 2 under the incoming flux
of I = 1.0 kW m . The liquid dye was prepared to have its density and thermal
expansion coefficient value approximate to that of the water as well as its
characteristics of stability. Dye liquid enters the glass tube via an injector.
RESULTS OF VISUAL EXPERIMENT
Pictures taken at the moment of t = 0, 1, 2, 6 and 9 min during the heating pro-
cess are shown in Fig. 1. From Fig.1(a) to 1(c), we can clearly see a three di-
mensionalaminar flow patternoccurringin the tube. The fluid near the tube wall is
heated and flows upward along perimeter of the tube and the cold fluid in the
header pipe is able to flow in from the bottom of the tubo partly moving upward
along the tube wall,and rest of them moves forward. The upper part of the fluid
flow near the open end firstly turns dark; after a period of 9 minutes, the whole
tube turns dark except its middle horizontal section, see Fig.1(e), as if there
existeda stagnation partition separating the tube flow into two parts — upper
and 'lower. This flow pattern could be simply interpreted as a !>Ch type collec-
tor-tube manifolding flow pattern, see Fie. 2.
Fig.3 shows the flow patterns at the entrance of the tube with different flow
rate in the header. It is apparent that the cold fluid flow in the header pine
can smoothly enter the absorber tube when the flow rate in the header is' quite
low, see Fig.3(a). As soon as the header flow rate increases, it will become
a sort of buffer to prevent the tube'flow moving out, see Fig. 3(b) to (d). .This
visual experiment clearly explains the cause of the heat extraction problem
occurring sometime in such horizontally disposed tube system when forced circu-
lation or once through type operation is taken. So thermosyphon circulation is
preferably suggested for this system in order to avoid the choke phenomenon at
the entrance of the tube.
-------
(d)
(e)
Fig. 1. FLow pattern inside the horizontal tube
(a): t = 0 min.; (b): t = 1 min.; (c): t = 3
(d): t = 6 min,.; (e); t = 9 min.;
I = 1,000 W/m ; 1 = 600 mm; d = 35 mm.
fr
4=
iff
.out
Am
Fig. 2. "CM Type flow pattern model
-------
1483
Fig. 3 The effect of flow rate in the header to the
flow pattern at the entrance of the tube
NATURAL CONVECTIVE HEAT TRANSFER PROCESSES IN THE COLLECTOR
TUBE - HEADER PIPE SYSTEM
In order to get a uniform heat flux along the longitudinal direction of the
tube, the electric heating film is divided into six segments in same length,
the heating power of each segment can be adjusted to reach the same value by
a variable resistance. Six pairs of copper constant thermoeouples spaced
along its length at the middle of each segment with four thermocouples spaced
around the circumference of the tube. Two mixing cups are located at the inlet
and outlet of the manifolding to measure the temperature rise of the header pipe
flow due to the natural convective heat transfer processes ocurfngi-n the horizon-
tal tube. During the experiment, the power input varies from 100 to 4300 W/cm
and the Re number of the header flow varies from 23 to 1600. Four sets of hori-
zontal tube with different length to diameter ratio (1/d = 1500/40, 1200/40,
900/40 and 600/40) were under experiment.
The definition of the local convective heat transfer coefficient from tube wall
to the inlet fluid from the header is given as :
h = q / ( T . - T. ) ( 1 )
t w w,i in
-------
1484
T . is the average local temperature on tube wall and T. is the water tempe-
rature in the header; q is the power input. THe al*g. fteat transfer coeffi-
cient of whole tube is cfefined as:
h = 1/A h • dA ( 2 )
av t t t
The governing heat transfer equation form is suggested as:
;p
Nu = C ( Gr"Pr )m ( 1/d )n Re" ( 3 )
from which, Nu = h 1/k; Gr = g l^q , ,2 Re = wD/w ( in the header pipe);
av ° nw jy k; '
D — diameter of the header pipe; y— kinematic viscosity of the fluid;
° diameter of the tube. ^ length of the tube; A • tube surface;
The results of the heat transfer experiment are concluded as follows:
Nu = c\ Gr " Pr )m ( 1/d )" ( 4 )
for Re > 410, n = 0.47, C* = 0.0476Re~0,395+0.00849Re~°"383+0,7160Re~2'36;
120 n = 2.6x10 ^Re +0.341; C = 300 Re 400Re
/ O O / o /
40 < Re <120 ' n = 16.5x10 Re + 0.183, c" = 2.15x10 Re- * + 0.39x10 Re" ' +
+2.5xl06Re_4'°.
From the heat transfer governing equations mentioned above, it is found that
for a given heat flux input, there exist a Re (optimum value), at which the
h value reaches maximum, increasing with the^increase of the input heat flux.
It is also indicated that the h decreases with the increase of tube length.
av °
CONCLUSION
The flow pattern inside a single-ended all-glass evacuated tubular absorber is
studied and some operational criterion is suggested under this fundamental re-
search. The governing equations of convective heat transfer of such system are
given by experiment. It is expected that the results would be helpful for the
Llieiifial system design.
REFERENCES
Harding, G. L., Yin, Z. Q. and Mackey, D. W. (1985). Solar Energy 35> 71 - 80.
Harding, G. L. and Yin Zhiqiang (1985). Solar Energy ,34. 13 _ 18.'
Ostrach, S. (1972). Advances of Heat Transfer, 8 , 161 - 227.
Tien C. L. and Chen Y. L. (1985). Int. J. Heat and Mass Transfer, 28, 603 - 612.
Window, B. and Harding, G. L. (1983). Solar Energy,31, 153 - 158.
Window, B. (1983). Solar Energy, 31, 159 - 166.
Yin Zhiqiang, Harding, G. L. Graig S. etal. (1985). Solar Energy 35, 81 - 92.
-------
2.10 Active Heating I: Seasonal Storage
-------
-------
1487
HEAT TRANSFER IN BOREHOLES FOR
SEASONAL SOLAR STORAGE IN THE GROUND
Svend Svendsen, Peter Berg, Karsten Duer
Thermal Insulation Laboratory, Technical University of Denmark
Building 118, DK-2800 Lyngby, Denmark
ABSTRACT
In order to investigate different proposals of borehole design, an indoor test facility has been built
and used at the Laboratory.
The facility consists of a 3 m3 vertical steel tank in which a 1.80 m long section of a borehole
can be placed in a sand filling. By a liquid heating/cooling loop and a measuring system, the heat
transfer coefficient from the liquid to the filling material in the borehole and to the surrounding
material can be measured. Two examples of borehole constructions have been built and tested.
A detailed numerical model for a single borehole and the surrounding soil has been developed.
Several analyses comparing calculated results and data measured in the indoor test facility show
that a very good thermal model of a borehole may be obtained. The model will be used as a very
effective tool, investigating the thermal behaviour of different borehole designs.
KEYWORDS
Seasonal solar storage in ground, heat transfer in boreholes, indoor test facility, computer
simulations.
INTRODUCTION
A central solar heating plant consisting of a solar collector field, a seasonal heat storage and a
distribution system (district heating system) can deliver almost all the heat needed for space
heating and domestic hot water in a building complex. (TEA Task 7, Status Report').
The cheapest way of storing heat seasonally on a large scale is by using the ground. Transfer of
heat to and from the ground can be done by placing plastic tubes in boreholes in the ground. The
storage losses can be limited to about 20%, even if only the top of the storage is insulated. No
insulation is needed on the surfaces against the surrounding soil if the storage is big enough due
to the small surface to volume ratio.
The boreholes and the top insulation are therefore the only cost of the storage. The design/con-
struction of the boreholes is important for the performance and price. The borehole must have a
sufficiently high heat transfer coefficient to the surrounding soil and be reasonably low priced.
receding page blank
-------
1488
DESCRIPTION OF BOREHOLES
The typical soil in Denmark consists of sand, clay or a mixture (moraine). Therefore, it is easy,
by an auger drilling method, to make boreholes with a diameter of 30-40 cm and a depth of up
to 20 m. A number of U-shaped plastic tubes are placed in the hole, which is then filled with
concrete or sand. By circulating warm water from the solar collectors through the plastic tubes
heat can be stored in the ground.
All the boreholes investigated in this first phase had a diameter of 30 cm and, furthermore, 3
U-shaped polyethylene (PEX) tubes placed in a circle with a diameter of 19 cm, see figure 1. The
filling materials investigated were concrete, dry sand and wet sand. The borehole with concrete
was cast in a circular form. The borehole with sand was constructed by use of a 30 cm diameter
tube made of 5 mm specially strong paper material. The test specimens were both 1.8 m long and
the dimensions of the plastic tube were 12 mm/8 mm. Instead of using the separate U-shaped
tubes, a single tube was used to simulate the U-shaped tubes. In this way, only one inlet and one
outlet of the tube were sticking out at the top of the borehole specimen. In order to secure that
the plastic tubes were correctly placed, as shown on the cross section on figure 1, the tubes were
fixed to a number of steel wire rings fixed to the form before it was filled with concrete or sand.
polystyrene|JM|foam j
plywood
7WT/T7T
3x100mm
/polystyrene foam
Fig. 1. Drawing of borehole configuration and experimental set-up.
-------
1489
EXPERIMENTAL SET-UP AND MEASUREMENTS
The stationary heat transfer coefficient from the fluid to the surface of the borehole was measured
for all three types of filling material. The measurements were performed at the Laboratory by
circulating warm water through the plastic tube at a constant temperature.
By insulating the bottom and the top of the boreholes a two-dimensional heat transfer from the
warm water to the surface of the borehole was obtained. From the surface the heat was lost by
natural convection to the surrounding air.
After an operation time of about 8 hours stationary conditions were reached.
The heat transfer coefficients were found by measuring the heating power of the water flow and
the average temperature difference between the water and the surface of the borehole, ie. the
concrete surface and the inner side of the paper tube containing the sand. The heat supplied to the
borehole was measured by means of a mass flow meter and a thermopile of type T thermocou-
ples.
The results of the measurements are shown in table 1.
TABLE 1 Heat transfer coefficient of boreholes based on the temperatures of the
water and on the surface of the concrete/sand filling. Length 1.8 m.
Type
Water temp.
Flow
Surface
Heat transfer
inlet outlet
temp.
coefficient
°C
kg/s
°C
W/m-K
Concrete
74.6 72.4
0.056
48.9
11.4
Dry sand
73.7 72.0
0.032
41.3
3.9
Wet sand
49.7 47.7
0.030
38.2
12.9
The results show that the heat transfer coefficient is strongly dependent on what type of filling
material is used in the borehole.
The dynamic heat transfer between the fluid and the ground was measured in an experimental
set-up shown on figure 1. The borehole with the concrete filling material was placed in the centre
of the steel tank, which was filled with dry sand. Insulation at the bottom and the top secures that
practically only radial heat transfer will occur. The bottom and top losses were calculated to be
less than 4% of the radial heat transfer. A heating and cooling sequence was run with constant
mean temperatures of the water of about 60 °C and 26° C, see figure 2. Several temperature
measurements were made in the borehole and in the sand as shown on figure 1 for making a
detailed analysis of the heat transfer and temperatures possible. The heating and cooling power
was measured as in the previous experiments. The result is shown on figure 3.
-------
Mean fluid temperature, °C
65 -|
60 -
55
JP
50 -
45 -
40 -
35 -
30 -
25
10
hours
Fig. 2. Heating and cooling sequence used .in the test of the concrete filled bore-
hole in dry sand.
Power kW/m
i
0.9
0.8
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0
-0.1
-0.2
-0.3
-0.4
-0.5
-0.6
-n
+ measured •
o detailed model
-
o simplified model
!
1 1 1 1 1 1
time
i i i i
Fig. 3.
0 2 4 6 8 10 hours
Heating and cooling power measured and calculated for the concrete filled
borehole in dry sand.
-------
1491
MODELLING OF THE HEAT TRANSFER
A detailed numerical model of the dynamic heat transfer in the borehole was developed using a
control volume technique. A simplified model of the heat transfer in the borehole was set up on
results from the detailed model.The simplified model is used in a computer program for calcu-
lating the heat transfer in a full scale ground storage. Both the detailed model and the simplified
model have been compared with the measurements. The results are shown on figure 3.
The two models correspond quite well with the measurements. The most important disagreement
occurs when the temperature changes suddenly. This is not going to take place under realistic
operating conditions. The simplified model will, therefore, be quite good for application in the
computer program.
It should be noted that possible effects of flows of moisture and water in the ground are neglected
in the calculations. Under ideal conditions negative effects of flows of moisture and water can be
avoided and have, therefore, not been considered.
CALCULATION OF THE PERFORMANCE OF SOLAR HEATING PLANTS
By means of the computer program (Berg, Olesen 1990) the performance of a solar heating plant
with seasonal storage in the ground has been calculated. The data used for the calculations are
listed in table 2. The results are shown in table 3.
TABLE 2 Data for solar heating plant calculations.
Solar sollector area
36,000 m2
Collector efficiency (fluid temp. = air temp.)
75%
Collector heat loss coefficient
3 W/m2K
Volume of short term storage, buffer, (water tank)
2,000 m3
Volume of ground storage (dry sand)
295,000 m3
Depth of ground storage
20 m
Number of boreholes
5776
Borehole distance (square system)
1.6 m
Thermal conductivity of borehole filling material
0.4 W/mK
Thermal conductivity of ground (dry sand)
0.5 W/mK
Heat capacity of ground
1.4 MJ/m3K
Top insulation , U-value
0.07 W/m2K
Heating load of district heating system
50,000 GJ/year
Temperatures of district heating (out/return)
60°C/30°C
-------
1492
TABLE 3 Results of solar heating plant calculations for different designs of
boreholes.
Borehole
Number of
Tube dia-
Max. temp, in
Solar fraction
diameter, m
U-tubes
meter, mm
buffer, °C
%
0.30
3
8/12
95.1
60.0
0.30
1
8/12
100.4
57.8
0.30
1
16/22
99.2
58.3
0.10
1
8/12
102.1
56.9
0.10
1
16/22
101.3
57.3
0.10
2
8/12
99.7
58.1
In general the borehole design has only a very small influence on the performance. This is
important because the prices of the different designs differ as much as about a factor of two.
The calculations can only be seen as an example of what kind of performance a solar heating
plant with a seasonal ground storage can attain. The plant is not optimized, but the results show,
that a relatively large solar fraction can be obtained.
CONCLUSION
The test facility, the models and the programs for characterizing and evaluating the heat transfer
in boreholes are useful tools for optimizing the design of boreholes to be used in seasonal solar
storages. By means of these tools the influence on the performance of the geometry and filling
material of the borehole as well as the material, size and number of tubes can be investigated in
a cheap and quick way. This activity has, of course, to be supplemented by full scale experiments
with a number of boreholes in order to study the technique for making the boreholes, or other
methods for inserting the heat exchanger in the ground. Full scale experiments are also necessary
for studying the problems related to flows of moisture and water in the ground.
ACKNOWLEDGMENT
This work was part of an R&D project funded by the Danish Energy Agency.
REFERENCES
IEA Solar R&D, Central solar heating plants with seasonal storage - status report. June 1990.
Berg, Peter, and Olesen, Ole. 'Szeson-Sol, Manual til Saesonsol Version 1,0'. February 1990. In
Danish.
-------
1493
PREDICTION OF THE LONG TERM PERFORMANCE OF
SEASONAL SPHERICAL UNDERGROUND SOLAR THERMAL ENERGY STORES
Mazhar Unsal and Recep Yumrutag*
'Department of Mechanical Engineering, University of Gaziantep
27310 Gaziantep, Turkey
ABSTRACT
An analytical solution for the transient temperature field outside an underground spherical
thermal energy store with arbitrarily time-varying store temperature and arbitrary initial
temperature distribution in earth surrounding the spherical thermal energy store is presented.
The solution obtained has been, utilized to investigate the long term performance of spherical
thermal energy storage systems with year-round charging of solar thermal energy via solar
collectors and extraction of stored thermal energy by a heat pump for space heating purposes
during winter months. Annual variation of store temperature and yearly energy balances of a
system with a spherical seasonal thermal energy storage have been estimated and are presented
in graphical form.
KEYWORDS
Solar collector, seasonal storage, active solar heating, heat pump, spherical store.
INTRODUCTION
Seasonal storage of solar thermal energy has been the subject of many investigations in the past.
Brunstrom et.al.(1987) and Hahne et.al.(1989) reported results from two experimental
projects. Bankston(1988) peresented a comprehensive review of the many experimental and
theoretical studies on the subject. Numerical simulation of the long term performance of space
heating systems with seasonal storage, for solar thermal energy is a very time consuming task
due to three dimensional and transient nature of the problem and because of the large span of the
time domain.
The present study is an analysis based on a simplified formulation of the problem considering
spherical symmetry for the transient heat transfer problem outside a spherical thermal energy
store and assuming a spatially lumped bulk temperature inside the thermal energy store. The
transient heat conduction problem outside the thermal energy store is formulated as a one
dimensional transient problem which can then be analyzed by a combined analytical-numerical
solution methodology.
Geometry of the problem considered is depicted in Fig. 1. Solar collectors mounted on roofs of
houses are operated over the whole year charging energy to a single spherical thermal energy
store. Heat is supplied to the houses only during the winter months by the heat pump which
extracts heat from the thermal energy store. The heat pump operates only when the temperature
of the store is not sufficient to keep the house at the design inside air temperature. Modelling of
the heat pump system with seasonal storage utilized in the present study is the same as that
reported by Unsal(1990) and will not be repeated. Analysis of the transient heat conduction
problem in earth surrounding the thermal energy store is extended in the present study to
account for an arbitrary initial temperature distribution in earth outside the spherical thermal
energy store, and this is presented in the next section. Modelling of the heat pump
systerrfreported by Onsal (1990) was applied in conjunction with the analysis given in the next
sectionlo predict the long term performance of a system with a residence complex located in
-------
1494
Solar Collectors.
Spherical
Water Store
iv++i+ivwv,t+*
Fig. 1 Heat pump system with a spherical thermal store in deep earth
Gaziantep. An interactive computer program was prepared in Fortran77 for simulating the
operation of the system, and the results obtained from application of this code to a specific
residence complex with 500 houses are presented in this paper. Listing of the interactive
computer code and details of the numerical procedures are available in Yumruta§(1991).
ANALYSIS
A spherical store located in deep ground is considered, and the far field temperature around the
store is assumed constant and equal to the deep ground temperature I*,. The water in the
spherical thermal energy store is assumed fully mixed and at a spatially lumped time varying
temperature Tw(t). The surrounding earth outside the store is assumed to be at initial
temperature distribution F(r). The problem formulation for the transient temperature in earth
outside the store is given by :
d2J | 2 dT_ 1 dT
0r2 r 3r a <5t
T(a,t) = Tw(t) (2)
T(oo,t) = (3)
T(r,0) = F(r) (4)
An energy balance relating energy charged to the store, energy accumulation rate in the store
and energy loss into the surrounding earth gives:
dT r)T
Q=pwVwcw—p-kA —(a,t) (5)
at dr
The problem consisting of equations (1)-(5) is cast into dimensionless form by introducing
dimensionless variables and by transforming the dependent variable according to
x = r/a, x = at/a2, $ = (T- Too)/Too - <(>w = OV Too)/Too
Y = a(one year)/a2 , y(x,t) = x(x,T) (6)
The resulting dimensionless problem formulation is given by:
-------
1495
(7)
\|/(1 ,i) = <|>w(t)
\|/(°°,T) = 0
V(x,0) = f(x)
(8)
(9)
(10)
(11)
The problem posed by equations (7)-(10) will be decomposed into two problems by letting \\i =
¥1 + ¥2 ancl using superposition. This leads to two problems. The problem for will be
required to satisfy \|f 1 (1 ,x) = and Yi(x-°) = 0- The problem for \(f2 will be required to
satisfy v2(1'x) = 0 and V2(x-°) = fM- The problem for y-j can then be solved by an
application of a similarity transformation and Duhamel's superposition principle. The problem
for \jr2 can be solved by an application of the Fourier Integral transform technique. Solutions
obtained for and \|/2 are as follows.
Substitution of \|/ into (11) yields an integro-differential equation for the temperature of the
thermal energy store. This equation was solved numerically using finite differences. Derivation
of the finite difference equation and description of the solution procedure are omitted.
Modelling of the heat pump system reported by Unsal(1990) and the solution for the transient
temperature outside the spherical thermal energy store given by equations(12) and (13) were
utilized to estimate the long term performance of a single spherical thermal energy store
coupled to a heat pump supplying heat to a residential complex consisting of 500 houses. The
methodology is based on hour by hour calculations of all system variables. Winter heating
requirements of the complex were estimated on an hourly basis. Thermal performance of the
heat pump was calculated using a model algebraic equation designed to give 25% of the COP
calculated from an idealized thermal cycle. If, at any instant, the temperature of water in the
thermal energy store is sufficiently high to meet the heating load of the complex, the water
from the thermal energy store is directly circulated through the load side heat exchangers
bypassing the heat pump,which then is not functional, residing in a standby status. The design
heat load of each house located in Gaziantep was assumed equal to 10 kilowatts yielding a total
design heat load of 5000kW for the complex. The winter design temperature difference used was
29 Celcius degrees. Heat may be supplied to the houses by panel heat exchangers or fan coil
units. UA-value of one house, defined as the ratio of design house heat load to design inside to
outside air temperature difference, was assumed equal to the total UA-value of all heating
elements located in the house. A total collector area of 40 square meters facing due south and
inclined at an angle equal to the latitude was considered for each house in the complex. Inside
design air temperature was assumed equal to 20C uniformly for the whole winter. Year long
hourly solar radiation data for a horizontal surface and hourly outside air temperatures were
¦f f g(z)sin(A,z) sin(A.z) e ^ 1 dzdA,
Jo Jo
A SAMPLE SIMULATION
-------
1496
1 00t TEMPERATURE
GRANITE, RADIUS = 40m
¦*" 1.YEAR
5. YEAR
10. YEAR
JULY 1 'OCT 1 JAN 1 APR 1 JULY 1
Fig. 2 Variation of Water temperature in the store for Granite R=40m
1 .Or
0.5'
0.0-1
GRANITE, 1. YEAR
,^^,1 *
an
M R = 10m
B3 R = 20m
0 R = 40m
+
SOLAR ELECTRIC STORED LOSSES TO LOAD
Fig. 3 Annual Energy Balances for Granite during the First Year of Operation
stored in a computer data file. Hourly solar radiation data available only for a horizontal surface
in Gaziantep was converted to hourly radiation on a tilted surface facing due south. These- were
utilized in hourly numerical calculations. Problem was studied considering four types of earth;
clay, sand, granite and coarse gravelled. Thermal property data for thermal conductivity and
thermal diffusivity used for the different types of earth are as reported by Onsal(1990). The
simulation was started on July 1. The temperature of earth outside the store was assumed to be
at a spatially uniform (radially invariant) initial temperature equal to the deep ground
temperature. The initial temperature of the store was also assumed equal to the deep ground
temperature. Numerical solution of the integro-differential equation over a time span of several
years, using hourly time increments, was found to require extremely large computational CPU
times. The computational strategy was then changed in search for a computationally more
efficient scheme. The alternate strategy considered was that of treating the transient heat
conduction problem in earth for a month at a time. The temperature distribution in earth
surrounding the store was determined at the end of a month, the time was reset to zero,and the
numerical procedure restarted using the thermal field residing at the end of the previous month
inside earth outside the store.This procedure proved useful for only a few years following the
-------
1497
0
GRANITE, STORE RADIUS = 40m
0.5
0.0
" | ¦ " ' ¦" t ¦ ^ , ¦ my.*- i- ¦ [
SOLAR ELECTRIC STORED LOSSES TO LOAD
M 1. YEAR
ES 5. YEAR
0 10. YEAR
Fig. 4 Annual Energy Balances for Granite
1.0t
0.5-
o.o-
ENERGY BALANCES DURING THE FIRST YEAR
Store Radius = 20m
SOLAR
[¦—*-"¦ [
B
Sand
Coarse
~
Clay
Granite
ELECTRIC
STORED
LOSSES
TO LOAD
Fig. 5 Annual Energy Balances for the First Year of Operation
start of simulation, and numerical truncation errors were found reduce accuracy with increasing
years. Procedures to decrease truncation errors were found to increase the CPU time.
DISCUSSION
Results obtained from numerical simulations as reported by Yumruta§(1991) for a complex of
500 houses located in Gaziantep and heated by a heat pump system which extracts heat from a
single spherical thermal energy store are reported in this section. Variation of the thermal
energy store temperature with time following start of system operation is given in Fiq. 2 for a
spherical store located inside a granite rock bed. It is seen from this figure that after several
years of operation, store temperature attains a sufficiently large value. A heat pump is not
necessary for operation of the system when the store temperature are sufficiently high, and the
heat pump electric energy requirements become very small.
Annual energy balances of the thermal energy store for three different store sizes is given in
Fig. 3. Solar heat gain from the collectors and electric energy input to the heat pump make up
for the annual energy requirements of the system, while annual thermal energy accumulation in
the store, annual conduction loss into earth and annual thermal energy input to the heat pump
are the energy expenditures of the system. It is seen from Fig. 3, that, increasing the store
radius decreases the electric energy requirements of the heat pump up to an optimal value of the
store radius. Electric energy requirements increase with an increase of store radius beyond
this optimal value.
Annual energy balances of the system during the first, fifth and tenth years of operation are
depicted in Fig. 4. Fig. 4 indicates an increase in heat losses into the surrounding earth with
-------
1498
years. The annual mean store temperature increases with years resulting in a decrease in the
annual electrical energy requirements of the heat pump system.
Annual energy balances for the system are depicted in Fig. 5 for four different earth types. The
annual average store temperature was found to be largest for sand and smallest for granite.
Granite rock bed yields the worst system performance,while sand yields the best.
Results indicate that store temperatures and annual heat pump COPs are high favoring technical
feasibility of seasonal solar thermal energy storage in deep earth. The simple model described in
this paper, which is based on a thermal energy store with spherical symmetry, can be further
utilized for system optimization purposes at minimal computational costs. Effects of system
parameters such as collector slope, collector type, annual load distribution, solar radiation
data, outside air temperature, wind velocity and load side heat exchanger size on system thermal
performance is currently under investigation.
NOMENCLATURE
A
Surface area of sphere
a
Radius of spherical store
c
Specific heat of soil
c w
Specific heat of water
OCR
Coefficient of performance
g(z)
= f(x)
k
Thermal conductivity of earth
p
pwcw/(3pc)
q
Q/^rcakToo)
Q
Net energy charge rate to the store
r
Radial coordinate
t
Time
T
Temperature of soil surrounding the thermal store
Tw
Temperature of water in the thermal store
Too
Farfield deep ground temperature
Vw
Volume of store
a
Thermal diffusivity of earth
P
Density of soil
pw
Density of water in the store
z
= x-1
REFERENCES
Bankston, C.A. (1988l.The Status and Potential of Central Solar Heating Plants with Seasonal
Storage: An International Report, in Advances in Solar Energy, vol 4, Ed. K.W. BOer, Plenum
Press
Brunstrfim, C. and C-G. Hillstrflm (1987). The Lvckebo project, solar district heating with
seasonal storage in a rock cavern , Swedish Council for Building Research, Stockholm,
Sweden.
Hahne.E., N. Fisch and R. Giebe (1989). Experience with a Man-made Aquifer in Short-Term
and Long-Term Cycles, in Energy Storage Systems: Fundamentals and Design, Kluwer
Academic Publishers, Dordrecth, The Netherlends
Onsal, M. (1990a). Long Term Performance of Heat Pumps with Spherical Underground
Thermal Stores. Heat and Mass Transfer in Building Material and Structure, Hemisphere
Publishing Corporation, Washington DC, USA, 389-398
Yumrutaf.R. (1991). Computer Simulation of Solar Aided Heat Pump Systems with Underground
Spherical Thermal Energy Stores , M.S. Thesis, University of Gaziantep
-------
1499
LONGTERM STORAGE OF LOW TEMPERATURE SOLAR HEAT
IN THE GROUND
M. Beck, M. Reuss, H. Schulz, J. Spannig, B. Wagner
Institute of Agricultural Engineering, Technical
University of Munich, Voettinger Str. 36,
8050 Freising, F.R. Germany
ABSTRACT
One of the major problems concerning solar thermal energy uti-
lization in European climate is the considerable time delay bet-
ween periods of solar energy supply and heating energy demand.
Both from a technical and economical point of view, the seasonal
storage of low and moderate temperature heat in soil seems to be
a good solution to this problem.
For this purpose our institute developed the concept of a ground
heat storage with vertical heat exchangers. Based on theoretical
computer simulations and laboratory and field experiments,a pilot
plant was designed and built. It consists of a 101 m2 unglazed
solar collector, a 2850 m3 ground storage and a diesel engine
heat pump of 58 kW heating power.
This installation supplies thermal energy to the low temperature
heating system of a two family house and an adjoining office
building. The coefficient of performance related to primary
energy is about 1.7.
KEYWORDS
Longterm storage of energy, ground storage, heat and moisture
transfer, vertical heat exchangers, diesel-engine heat pump,
heat recovery.
INTRODUCTION
Increasing environmental problems force us to develop and improve
new techniques for energy saving, for using renewable energies,
waste heat and other ecologically beneficial energy sources.
One of the major problems in using these techniques in the Euro-
pean climate, especially solar heating systems, is the longterm
storing of low and moderate temperature heat.
Among others, the ground storage with vertical heat exchangers is
promising from a technical and economical point of view.
-------
1500
Our project investigates the performance of these vertical heat
exchangers under charging and discharging conditions, especially
the basic problems of heat and moisture transfer in the surroun-
ding area of the heat exchangers.
A computer model describing these effects was developed and vali-
dated by several laboratory and field experiments.
These findings and previous research work in heatpumps and solar
collectors convinced us to provide a two family house and an
office building near Donauwoerth/Bavaria with a solar assisted
heatpump equipped with a ground storage.
A computer based data acquisition system recording the process
data monitors the longterm performance of the whole plant to
optimize the system.
THE SIMULATION MODEL
The simulations were performed with a numerical finite differen-
ce computer programme taking into account coupled heat and mois-
ture transfer in the surrounding area of a single heat exchanger.
It is located within the ground storage and is fully surrounded by
other heat exchangers.
Due to the geometry of the problem, a radial 2-D grid was chosen
for the discretization of the area.
The programme is based on the Crank-Nicholson method with a two-
step time scheme utilizing a predictor corrector method to calcu-
late time dependent heat and moisture transport characteristics.
The resulting sparse matrix of linear algebraic equations is
solved by successive overrelaxation, a modified Gauss-Seidel
iteration method with accelerated convergence.
Comprehensive climatological data were taken into account to
calculate heat and moisture exchange to the environment. These
are solar radiation, ambient temperature, wind speed, air humi-
dity, and precipitation. Two of these five variables were deemed
applicable and measured at the site of the ground storage in in-
tervals of 15 minutes. The other values were taken from the
nearest weather observation facility at Neuburg a.d.D./Bavaria.
The following assumptions were made for simplification:
- the soil moisture consists of pure water
- neglect of osmotic effects
- the soil matrix is rigid
- there is no biogenic or chemical heat production
- the different layers in the domain are homogeneous
- emission and adsorption factors of incoming and outgoing ra-
diation are constant
- evapo*transpiration due to plants is disregarded
- the neglect of further biological influences (wormholes, roots)
- the neglect of the irregular pattern of water infiltration
during ponding
-------
1501
DESCRIPTION OF THE PILOT PLANT
As shown in Fig.1., the system consists of the following circuits:
- collector, storage and heat pump circuit
- refrigerant circuit
- engine cooling liquid circuit
- heating water circuit with 6 m3 shortterm storage
- domestic hot water circuit
DHW
collector
unglazed
collector 101 m2
domestic
hot water
compressor
diesel engine
condenser
heating system
heatrecovery
from exhaust gas
evaporator
exhaust
r t
I <
L
u
ground coup, storage
I collector storage and heat pump circuit
II refrigerant circuit
III engine cooling liquid circuit
IV heating water circuit
V domestic hot water circuit
Fig. 1. Scheme of the pilot plant
The used unglazed collector has proved its efficiency in similar
applications of low temperature systems, especially in swimming
pool heating plants. The collector is made of black, corrugated
polypropylene tubes with an outer diameter of 25 mm, which are
mounted on the south facing roof of the office building covering
an area of 101 m2.
From May to September, when there is usually no heating demand,
the heat pump is switched offj and the energy gained by the col-
lector is transferred to the ground storage.
During the heating period the heat pump is connected either to
the collector or the ground storage, depending on the temperature
level of the device. If there is no heating demand,,and the tem-
perature from the collector is sufficient, the ground storage is
charged.
-------
1502
The monovalent heat pump, using R 22 as refrigerant, is directly
driven by a 25 kW diesel engine designed for industrial applica-
tion. Due to this fact it has a long life expectancy of almost
30 000 h and long maintenance intervals of 3 000 h. The energy of
the cooling liquid and of the exhaust gas, including the latent
heat of the water vapour, is extracted by several special heat
exchangers. Thereby 83 % of the fuel's lower calorific value can
be utilized.
By using either the ground storage or the collector,the solar
heat is transferred to the evaporator of the heat pump leading to
an evaporation temperature of - 10 °C to 15 °C. The condensation
temperature is in accordance withthe required input temperature of
the central heating system in the range of 45 °C - 50 °c.
The ground storage consists of 103 vertical heat exchangers ar-
ranged in six circuits, built by a simple and economical proce-
dure developed at our institute. U-shaped bent corrugated poly-
propylene tubes (as is found in the collector) are put in bore-
holes with 150 mm diameter and an average depth of 10.3 m.
To obtain good thermal contact,the bores were refilled with a
mixture Of bentonite, cement, sand and water. This composition
was optimized in preliminary experiments under several aspects.
It has proved to be resistant against freezing and heating, can
be pumped for refilling, and maintains reasonable thermal con-
ductivity of 0.785 W/mK at 0 °C up to 1.1 W/mK at 70 °C.
In order to design the ground storage) the thermal and physical
properties of the soil were analysed. The water content is 21 %
(per weight), the bulk density 1650 kg/m3, the porosity 37,8 %
(per volume). The heat capacity was found to be 1660 J/kg K, the
thermal conductivity 1.1 W/mK.
Based on the results of the computer simulations, the ground sto-
rage was designed with the heat exchangers having an average dis-
tance of 1.6 m and a total volume of 2850 m3. Due to the local
conditions,the geometry is almost cubic.
THERMAL PERFORMANCE MEASUREMENTS
The construction of the plant was finished in the summer of 89.
The system has been working since the winter of 89/90,interrup-
ted by a few heat pump failures.The data acquisition system was
installed in fall'89. Problems with the computer are presently
fixed, the system is now capable of restarting in case of a
voltage drop.
Fig. 2 shows the attained temperatures in a vertical cut through
the ground storage and the circumambient at the end of the charg-
ing period '90. The hatched area represents the edge of the
ground storage, assumed to be the influence sphere of the outside
heat exchangers.
In spring, when the charging mode began, the average temperature
of the ground storage was found to be 10 °C. The temperature of
the center almost reached the designed objective of 25 °C.
-------
1503
o
23.5 22.5
21.2
2
12.4
23.3/21.7
20.6
20.4 20.2
•15.2
12.0
4
21. i
6
20. i
5.0
-C
a a
10.6
10
10.9
12
14
« measuring point
edge of the ground storage
results in C
o
5
10
15
20
distance to the center in m
Fig. 2. Temperatures in the ground storage 28.09.1990
Isotherms marked in this figure show the distribution of the
energy from the center to the edge. On the top of the storage a
decline in temperature caused by losses to the environment can be
seen. The figure also shows the low rise in temperature outside
the storage.
In fig. 3 the energy flow during the charging period 1990 is
shown.
Irradiated Energy
Pumpenergy
1.3 GJ
342.2 GJ
Unuseable
Share /
125.0 GJ
Losses of the
\ Unglazed
/ Collector
15.9 GJ
Losses of
Pipes
96.5 GJ
Energy transferred to
the Ground Storage
Fig. 3. Energy flow in the charging period 1990
-------
1504
The input energy, the solar radiation and the electrical energy
for pumping is divided into three almost equal parts, which
represent:
- the energy transferred to the storage
- the losses of the unglazed collector
- the unusable part
The last part originates from the relatively high temperature of
the ground storage when starting the charging operation. By opti-
mizing the current through the several circuits of the ground
storage,a decrease of this portion to the stored energy and the
collector losses should be possible.
CONCLUSIONS
Within this project different basic problems of thermal energy
storage, in the ground are investigated.
Detailed computer modelling was performed and validated by labo-
ratory and field experiments. Based on this modelja solar assist-
ed heat pump with a ground storage was designed and built. This
system supplies monovalentlya two family house and an office
building with a heat power demand of 58 kW. The longterm perfor-
mance is recorded by a data acquisition system to verify the cal-
culations and to optimize the plant.
The results of the charging period '90 were as expected, the ana-
lysis of the heating period 90/91 had not yet been finished, as
it is still under work.
ACKNOWLEDGEMENT
The authors would like to express their graditude to the German
Minister of Research and Technology who sponsored this research
project.
REFERENCES
M. Reuss, T. Schmalschlaeger, B. Wagner (1989): Laboratory and
Field Studies of Heat and Mass Transfer. Phenomena in a Ground-
Coupled Storage with Vertical Heat Exchangers.
STES Newsletter, Vol X, No. 4, 2 - 4
M. Reuss, H. Schulz, B. Wagner (1990): Solar Assisted Heatpump
with Duct Storage in Donauwoerth, Proceedings of the Workshop on
Seasonal Thermal Energy Storage in Duct Systems.
Zeitschrift" fur angewandte Geowissenschaften, Vol. 9, 79-91
B. Wagner, R. Herold (1989): Modellversuch und numerische 3-D
Simulation zur thermischen Beeinflussung der Feuchteverteilung in
der ungesaettigten Zone durch einen vertikalen Waermetauscher.
Zeitschrift fuer angewandte Geowissenschaften, Vol. 8, 121 - 143
-------
1505
CENTRAL SOLAR HEATING PLANTS WITH SEASONAL STORAGE-
RESULTS OF A SITE SPECIFIC FEASIBILITY STUDY IN GERMANY
R. Kubler, N. Fisch, F. Miiller, E. Hahne
Institut fur Thermodynamik und Warmetechnik
Universitat Stuttgart
Pfaffenwaldring 6, FRG-7000 Stuttgart 80
ABSTRACT
Site specific studies for solar heating plants are carried out in a project
"Solar District Heating". Measurements have shown that high efficiency flat
plate collectors are available today with an optical efficiency of 80 % and a
heat loss coefficient less than 3 W/m2K. The design of a solar heating system
for 195 serial houses in Hamburg is described. If the houses are insulated
according to existing German regulations, the annual heat demand can be covered
to 53 % by solar energy with available roof area. Improved insulation with a
smaller heat demand (78 % of reference) increases the solar fraction to 64 %. In
the climate of southern Germany the solar fraction increases to 77 % and the
heat gain per unit collector area amounts to 398 kWh/m2«a. An economic analysis
shows, that solar heat costs between 185.- DM/MWh and 251.- DM/MWH can be
expected, which is about double the price of heat from a conventional district
heating system.
KEYWORDS
Solar Heating, District Heating, Solar Collectors, Seasonal Storage, Solar
Systems.
INTRODUCTION
The work carried out under the Task VII "Central Solar Heating Plants with
Seasonal Storage (CSHPSS)" of the IEA Solar Heating and Cooling Programme has
shown (Dalenback 1990) that it is advantageous to heat housing areas with a so-
lar fraction of 60 - 80 %. Experience with technologies applied in Sweden has
demonstrated that - with the potential performance improvement and cost re-
duction for seasonal heat storages and with high efficiency flat plate collec-
tors - that solar heating can become competitive with conventional heating tech-
nology within a few years. Work of the authors (Hahne 1989) in Task VII has
shown that similar systems could deliver heat for less than 200 DM/MWh to resi-
dential homes in Germany.
Results of predesign studies for 8 different sites were given by Kubler (1990).
Here design, cost and thermal performance are presented in detail for one of the
sites, and results are presented for three different climate regions in Germany.
-------
1506
SYSTEM DESIGN
Central Solar Heating Plants require a district heating system to distribute the
heat to the houses. The study has shown that due to space limitations in new
housing areas only roof mounted collectors can be applied. The heat from the
collectors is delivered to the central heat storage via a heat exchanger, a
water/glycol mixture is used as heat transfer fluid. The collector pipes are
placed in the same ditch as the heat distribution system. A schematic of the
system is shown in Fig. 1.
The heat storage is an earth pit of cylindrical shape, partly buried in the
ground with a concrete top. The storage is covered with soil and the area on top
of the storage can be used for play grounds, gardens or similar purposes. The
projected insulation thickness on top of the heat storage is 0.4 m and 0.2 m on
the side walls. Insulation of the bottom is only required, if the ground water
level is close to the bottom of the storage. A pre-design by a construction
company has shown, that the cost for this type of storage would be about 160.-
DM/m3 for a volume of 10,000 m3 (including heat exchanger, pump, piping and
charge/discharge facilities).
Heat is discharged from the storage into the heating system via another heat
exchanger. If the storage temperature is not high enough to meet the
requirements of the heating system, the auxiliary boiler will supply additional
heat. The domestic hot water (DHW) is heated centrally and distributed by
separate pipes. This allows for a better use of the heat storage due to the low
fresh water temperature, and leads to up to 5 % higher solar fraction. Moreover
it allows for an operation of the heating network independent of the needs for
DHW-heating (minimum supply temperature 55 °C) in the individual houses.
a
~
OQ.
Collector-
Circuit
Domestic
hot water
distribution
Auxiliary
boiler
Heat Storage
Fig. 1. Scheme of the System
-------
1507
SOLAR COLLECTORS
Experience in Sweden has shown (Dalenback 1990) that high efficiency large
module flat plate collectors today represent the most economic choice for solar
district or group heating systems. To base the cost calculations on realistic
performance data, six different collectors were investigated on the outdoor test
facility of the Institute during summer 1990. Two of the collectors were large
ground mounted modules (No. 1 and 2 in Table 1.), two were roof integrated
systems (No. 3 and 4). In addition a small array of 4 vacuum-flat-plate
collectors (pressure about 20 mbar) and an array of six modules were included in
the test. Recently another roof-integrated collector and a small array of
evacuated tube collectors have been installed.
TABLE 1. Summary of Collector Data
No.
Type convection
suppression
Absorber
area
m2
F'.(to) F'.(UA)
W/m2K
1
gm
teflon sheet
Sunstrip
12.5
0.79 4.34
2
gm
teflon sheet
black chrome
10.6
0.77 2.92
3
n
teflon sheet
Sunstrip
5.3
0.72 3.01
4
ri
none
Sunstrip
3.3
0.76 5.41
5
sm
near vacuum
Maxorb
4*1.8
0.80 3.74
6
sm
teflon sheet
Al, selective
6*1
.8
0.75 4.25
gm = ground mounted, ri = roof integrated, sm = single modules
The tests were run and evaluated following a method proposed by Perers (1990),
which is based on a dynamic collector model. The method requires a very simple
test setup (no heating nor control, only cooling is required), results are
available quickly, as tesing is not limited to steady state conditions and small
incidence angles. In addition to the key collector parameters F'-(ra) and
F'.(UA) listed in Table 1, the thermal capacity (m«c ) and the incidence angle
modifier constant b according to ASHRAE 93-77 are evaluated. Test results have
shown (see Table if that F'»(ra) = 0.8 and F'-(UA) below 3 W/m2K was achieved
already for prototype collectors. While F'«(ra) is quite close to the
theorect;cally possible value for a well designed flat plate collector with
convection suppressing sheet, the loss parameter can be reduced to about
2.5 W/m2K or even less in an optimized design.
These high efficiency flat plate collectors are a precondition for central solar
heating plants, where the storage is heated up to 95 °C. The prototypes 1-3
and 6 were all based on collectors available on the market today with an
additional teflon sheet inside to reduce convection losses. Calculations have
confirmed the benefit of the teflon sheet which reduces the loss coefficient for
a flat plate collector typically from about 5 W/m2K to about 3.0 W/m2K (see the
difference between Collector No 3 and No 4 in Table 1). Precondition for a
proper function of the teflon sheet is that it is fixed and stretched
sufficiently and does not touch the absorber nor the glass cover when the
collector is heated up. This was the case in collector No. 1 where at normal
operation temperature the sheet was partly lying on the absorber leading to
increased heat losses. Other interesting designs for high efficiency flat plate
collectors are considered presently such as the application of transparent
insulation material instead of teflon sheets which will reduce the collector
loss coefficient to about 2 W/m2K or even less. The reduced loss coefficient,
however, is accompanied by reduced optical efficiency. The major design problem
for these collectors is caused by destruction of the transparent insulation
material under stagnation conditions.
-------
1508
DESIGN AND THERMAL PERFORMANCE OF THE SYSTEM AT HAMBURG
Housing area and system description
The planned housing area contains 195 serial houses in 47 blocks of 3 to 7
houses. On the 169 houses with south facing roofs 6760 m2 collectors of the roof
integrated type No. 3 shall be mounted. The thermal performance of the system
was calculated using the program MINSUN (Mazzarella 1989), which was especially
designed for the simulation of this type of systems by the IEA task VII
participants. The climatic data, heat demand of the houses and the thermal
performance data for the system are listed in table 2. The first calculation was
done assuming that the house will be insulated according to actual German
building regulations. An investigation showed, that with available technology
(better windows, increased insulation thickness) the heating energy demand of
the houses can be reduced by about 30%, resulting in a total heat demand
(including hot water) of only 78 % compared to present day standard. The
equivalent heat cost (i.e., the annuity divided by the yearly savings) for these
additional investment is less than the cost of solar heat from the central
system (see section cost estimations).
Results for Hamburg Climate
Together with the available roof area the system with the reference heating load
only allows for a solar fraction of 53 %, while the system with the reduced
heating load achieves a solar fraction of 64 %. Figure 2 shows the heat balance
of the system with reduced heating load. The lower heat demand leads to a
reduction of the solar heat gain per unit collector area by 11 kWh/m «a. Design
of the heating system for a return temperature of 30 °C instead of 40 °C
increases the solar fraction from 53 % to 57 % and the heat gain per unit
collector area to 296 kWh/m2«a. The solar contribution per unit collector area
and thus system efficiency is quite low for this system due to a very high
fraction of diffuse radiation in June, July and August (62 %). This leads in
combination with the low solar radiation to high solar heat costs (see Table 4).
500
450
400
3
+*
350
c
o
2
300
£
5
250
S
200
«
«
•C
150
100
50
0
¦1 solar aneroy
1 1 auxiliary
-
>1.
-
1
-
1
-
'mi
3.00
2.50
2.00
1.50
1.00
0.50
0.00
¦C
§
a
«
x:
4 5 6 7 8 9 10 11 12 year
Month
Fig. 2: Solar energy contribution and total heating load, Hamburg
System, reduced heating load
-------
1509
TABLE 2. Climate and Thermal Performance of the System at Hamburg
Item
Unit Reference 78 %
Heating Load Heating Load
1
Number of degree days
K»d
3837
2
Solar radiation horizontal
kWh/m2-a
978
3
Total heat demand houses
MWh/a
3120
2418
4
Distribution losses
MWh/a
417
419
5.
Heat from Collectors
MWh/a 2156
2101
6.
Collector circuit and storage loss MWh/a
286
300
7
Solar contribution to load
MWh/a
1870
1801
8
Auxiliary to load
MWh/a
1667
1036
9
Solar fraction
% 53
64
10
Solar contribution per unit
area kWh/m2«a 277
266
Heat storage: volume 11 000 m3, depth 10 m, diameter 37.4 m
Results for Different Climate
8 sites were considered in the feasibility study (Kubler 1990) covering most of
the climatic regions in Germany. As the housing areas are quite different, a
comparison with respect to climate influences is difficult. Most information was
available for the housing area at Hamburg and therefore the detailed cost
calculations were carried out for this system. To demonstrate the influence of
climate on performance and solar cost, simulation runs for the Hamburg system
were carried out with the weather data from two of the other sites, namely
Stuttgart and Leutkirch. The climate at Leutkirch is well represented by the
test reference year Stotten, a sunny and cold place in southern Germany, for
Stuttgart measured data from 1989 were used.
TABLE 3. Thermal Performance of the System for other Climate
Item Unit Stotten Test Stuttgart
Reference Year Measured 1989
1 Number of degree days K-d 4553 3468
2 Solar radiation horizontal kWh/m2-a 1218 1129
3 Total heat demand houses MWh/a 3669 3043
4 Distribution losses MWh/a 435 427
5. Heat from Collectors MWh/a 2957 3060
6. Collector circuit and storage loss MWh/a 362 372
7 Solar contribution to load MWh/a 2595 2688
8 Auxiliary to load MWh/a 1510 782
9 Solar fraction % 63 77
10 Solar heat per unit collector area kWh/m2^a 384 398
Heat storage: volume m3 13500 13000
Both locations require a larger storage volume than Hamburg and show a
considerable increase in both solar fraction and especially the solar gain per
unit collector area. It is interesting to see, that at Stuttgart, despite lower
radiation and a higher solar fraction the solar gain is sligtly larger due to a
peculiarity of the weather. March 1989 was sunny and mild, allowing the solar
system to produce about 100 MWh surplus heat to be stored, which could be
utilized in the following cold April, a very favourable situation for a solar
-------
1510
system with seasonal storage. It should be noted, however, that the economic
calculation must be based on long term average weather data.
COST ESTIMATIONS
Investment costs for heat distribution system, auxiliary boiler, collectors and
heat storage were calculated in detail for the total system. For the heat cost
calculation the annuity method was used with a constant real interest rate of
6 % and 20 years lifetime for all system components. Annual operation and
maintenance costs were assumed to be 1.5 % of the investment for the solar
system and 2 % of the investment for the conventional system. The auxiliary
boiler was assumed to operate with natural gas (40.- DM/MWh) at an annual
efficiency of 94 %. The results of the cost calculations are summarized in Table
4. Solar costs are annual costs of the solar system divided by the solar heat
delivered to load, total heat costs are total system costs (including solar
system, distribution system, auxiliary energy) divided by the heat demand of the
houses. The results show, that under favourable climate conditions solar costs
of less than 200.- DM/MWh can be achieved. It is interesting to note that for
high solar fraction total heat costs may well be higher than solar costs (the
costs for the distribution system and the auxiliary are added), while at lower
solar fraction total costs are lower than solar costs due to the low price for
the natural gas. Total heat costs with splar system are only about twice as high
as if the system were entirely gas heated. The central heating system requires
about 60 % of the investment per house (7431.- DM) compared to individual gas
heating systems in each house (about 12 500.- DM).
TABLE 4. Solar heat cost and total heat cost
System
Solar cost
Total cost
DM/MWh
DM/MWh
Climate Hamburg - No solar system
108.-
Reference System
251.-
228.-
Climate Stotten - Reference building
195. -
201.-
Climate Stuttgart 1989
185. -
230.-
REFERENCES
Dalenback, J-0. (1990). Central Solar Heating Plants with Seasonal Storage -
Status Report. Document no D14:1990, Swedish Council for Building Research,
Stockholm. (Final report, Task VII, IEA SH&CP)
Hahne, E. (1989), Ml. Hornberger and N. Fisch. Solar Assisted District Heating
Plants with Long-Term Heat Storage, Proceedings of the ISES Solar Uorld
Congress, Kobe, Japan, (Seitenzahlen einfugen!)
Kubler, R.(1990), N. Fisch and E. Hahne. Central Solar Heating Plants with
Seasonal Storage - A Site Specific Feasibility Study for 8 Locations in the
Federal Republic of Germany. Proceedings of the North Sun Conference, Reading,
UK, (Pergammon Press, in print)
Mazzarella, L. (1989). Central Solar Heating Plants with Seasonal Storage - the
MINSUN simulation Progamme, application and user's guide, Politecnico di
Milano,(report, Task VII, IEA SH&CP)
Perers, B. (1990), B. Karlsson and Hakan Walletun, Simulation and Evaluation
Methods for Solar Energy Systems. Application for New Collector Designs at
High Latitudes, Studsvik Report ED-90/4, Studsvik Energy AB, Nykoping, Sweden.
Acknowledgement: This project has been funded by the Federal Minstry for
Research and Technology (BmFT) under the contract number 032 8867 A, the authors
gratefully acknowledge tljis support.
-------
1511
ANNUAL CYCLE SOLAR HEAT STORAGE
W. William Wisniewski, P.E.
ACCEPT/STEPS
334 North Shore Rd.
Cuba, New York 14727
ABSTRACT
In this study a central heating plant with seasonal solar storage (CHPSS) is being designed and
evaluated. The system utilizes an insulated fresh water solar pond with a black bladder collector
on the surface to collect and store solar heat energy during the summer. A commercially available
heat pump is used to deliver the heat to a school building during the winter.
KEYWORDS
Solar; seasonal storage; heat pump; unglazed solar collector; central heating plant; pond.
INTRODUCTION
Whitesville Central School in Whitesville, N.Y. was selected for characterization for the study.
The heating system design is capable of supplying heat for the 30,000 sq. ft. school building. The
system collects heat during the sunny summer season in an insulated, lined storage pond for use
later during the heating season. During the cold snowy winter months the collection of solar
energy is not attempted, and the temperature of the pond is allowed to drop, reaching a minimum
temperature in the spring, at which time solar heat collection is resumed.
The pond requirement is the basis of the annual cycle heat storage, and its specific size is
determined by the energy requirements of the school. The pond is sized such that the water in the
pond will reach a maximum temperature of 99 degrees F in October and a minimum temperature
of 35 degrees F in March. The ability of the collector to absorb solar heat at these relatively low
collector water temperatures reduces the transmission losses and increases the collector
efficiency.
The system is storage dependent. Storage dependency means that the overall system design is
more dependent upon the size and cost of the storage pond than any other single component.
Pond size is determined by the energy required by the building. If the depth of the pond is limited
to 20' for ease of construction, and if the entire surface of the pond is used for collector area, then
the collector area available is more than sufficient.
The objectives of the study:
1. Provide a pond design that can collect and store an ample amount of solar heat energy.
2. Provide a heat pump with efficiencies high enough to make the system operating costs
competitive with the existing heating system operating costs.
3. Provide a total system cost that will allow for a 10-year payback from operating cost savings.
-------
1512
BUILDING ENERGY USE
TABLE 1 Natural Gas Use at Whitesville Central School
Year
HDD
CCF
BTU Equiv.
CCF/HDD Tot.Cost
Cost/CCF
83-84
6,951
30368
3,036,800,000
4.37
$17,749
$0.58
84-85
6,397
24920
2,492,000,000
3.90
$14,339
$0.58
85-86
6,449
26755
2,675,500,000
4.15
$14,678
$0.55
86-87
6,301
29495
2,949,500,000
4.68
$13,755
$0.47
87-88
6,364
29225
2,922,500,000
4.59
$13,526
$0.46
88-89
6,734
29590
2,959,000,000
4.39
$15,248
$0.52
89-90
6,604
27915
2,791,500,000
4.23
$15,115
$0.54
AVE.
6,543
28324
2,832,400,000
4.33
$14,916
$0.53
Using 30,000 CCF and a boiler efficiency of 70%, only 21,000 CCF of gas is actually turned into
useful heat for use in the building. Using a heat equivalent for natural gas of 100,000 BTU/CCF,
the heat load of the building is 2.1 Billion BTU's per year.
HEAT STORAGE
Using a heat load of 2 Billion BTU's, that same amount of heat energy must be stored in the
pond. If the temperature of the pond is designed to drop 50 degrees F then AT = 50 degrees F.
Q = quantity of heat m = mass AT = temperature change c = specific heat
Q = 2X109BTU c = 1 BTU/lb./°F AT = 50°F
m = 2X10 BTU = 4 x 1q7 lb of water
1 BTU/lb./F X 50 F
4X10 lb. ^ ^ ^
= 641,000 ft of water
62.43 Ib./ft"
Pond construction parameters that were selected for ease of construction were a pond depth of 20
ft. and a square shape with 1:2 sloping sides.
Volume
in cu. ft.
2000000
1800000
1600000
1400000
1200000
1000000
800000
600000
400000
200000
0
100
¦
\
s
i'
1
+=riJ
150
300
200 250
Side Length in Feet
Fig. 1. Volume vs. side length for square pond with 1:2 sloping sides.
350
-------
1513
Fig. 1 shows a side length of about 220 ft. for a volume of 641,000 cu. ft. The actual side length
selected was 200 ft., to accommodate the placement of sixteen, 50 ft. square, insulating collector
modules on the surface. The new pond volume is 522,667 cu. ft., and AT is slightly greater than
the 50 degrees F originally assumed.
522,667 ft3 X 62.43 lb./ft3 = 3.26 X 10? lb.
AT
~Q~=
mc
2X10 BTU
3.26X10 lb.X 1 BTU/lb./ F
¦ 61° F
Fig. 2. Pond surface with 16 collector modules.
This simple method of sizing the pond ignores heat loss but is adequate, since the pond is so
massive that the minimum pond temperature will merely be a little lower. Accounting for heat
loss,the minimum pond temperature will be around 35 degrees F.
The pond must be insulated to store the heat until it can be extracted by the heat pump. The pond
has 10 inches of R5 per inch foam (R50) floating on the surface, 5 inches (R25) around the
perimeter to a depth of 8 feet, and 2 inches (R10) on the bottom. The foam on the bottom and
sides is held in place by the pond liner.
Fresh Water Solar Pond
R50
pond vater
R25
RIO
Fig. 3. Detail of pond insulation.
SUMMER SOLAR HEAT COLLECTION
The solar collector consists of an unglazed black bladder collector. During operation the average
depth of water in the collector is 0.1 ft. The approximate effective area of the collector is 90% of
the surface area of the pond or 36,000 sq. ft. When the collector is operating there is a total of
3,600 cu. ft. or 224,424 lb. of water in the collector.
A surge pump is used to fill a combination storage tank/header with pond water to start the
system. The circulating pump then supplies pond water as required to the header tank to maintain
the water level. A float activated valve controls the circulating pump to maintain the water level
in the tank.
-------
1514
Header
Tank
float
control
velve
v&ter
B lack B ladder Collector
Foam Insulation
pond water ¦&?/?$#&?'
Fig. 4. Collector operation.
The modules on the surface of the pond serve to both collect the solar energy and to insulate the
top of the pond. The 50 ft. by 50 ft. modules are meant to be assembled on site on a rigid
construction deck used also as a lifting platform, and lifted into place on the surface by a crane.
The modular units consist of two laminated layers of 5-inch thick foam glued together with
construction reinforcing fabric between the layers and on the top. The solar collector is made
from two layers of black rubberized roof membrane with 5-foot wide channels 50 feet long on the
top of the foam.
Black Bladder Cells
Insulating Foam
Sandwich
pond vater '
Fig. 5 Collector/insulation module cross section.
HEAT DELIVERY SYSTEM
A heat pump uses the warm pond water as a heat source, delivering hot water to radiators in the
school for space heating and to heat exchangers for heating water for showers and
kitchen/maintenance use. A heat pump with an average COP of 4.5 for an incoming source water
temperature range from 40 degrees F to 100 degrees F (the range of the pond water temperature)
was selected for this application. A commercially available unit was selected to avoid reliance on
a custom design.
CONTROL SYSTEM
A small calibrated microprocessor controlled test section of collector with its own separate
circulation pump will begin operating 2 hours after sunrise every day from mid-March to mid-
October. If the temperature rise in the test collector is sufficient and sustained for 15 minutes the
entire system is activated. The system will cease collection and drain down when a temperature
rise is no longer realized.
-------
1515
refrigerant
changed to
vapor(gas)
gaa compressed
to high pressure
70 degrees F aye,
Solar heated 'X
norid vater
Xy Compressor
^ I LillP
'Evaporator
<1
Condenser «j
gas to liquid \i» ¦
phase change \' ,
s:;
liquid to gaa
phase change
pond heat
absorbed
liquid
receiver
pressure
reducing valve
120 degreeg F
Hot vater for heating
heat delivered
to school building
liquid refrigerant
under high pressure
Fig. 6 Schematic of water to water heat pump.
TECHNICAL ANALYSIS
Heat pump performance is determined by the temperature of the pond. Pond performance is
determined by the amount of heat energy collected,less the amount of heat energy lost,and less
the amount of heat energy used by the school. Below is a graph of pond temperatures taking all of
these energy flows into account.
T POND (AVE)
Fig. 7 pond performance.
-------
1516
ECONOMIC ANALYSIS
Initial estimates for the cost of installing the system range from $165,000 to $200,000. Using:
$165,000 installed cost, $5.40 per MBTU for natural gas, $0.0965 per KWH for electricity, and
an average COP of 4.5; the payback time for the system is over 65 years, hardly worth
considering under the given circumstances. Even if the cost of gas increases considerably, the
payback times are not reasonable.
Alternative arrangements are possible which could make the system more economical to operate.
The most obvious is to run the heat pump at night using off-peak electric rates to heat up a tank
of water capable of meeting the heating needs the following day. Using this design, the payback
time could be as low as 15 years depending upon the off-peak rate.
Perhaps the basic notion should be that the system is not at this time reasonable for the school to
build as a replacement for natural gas heat. It may be appropriate as a replacement for other types
of heating systems where gas is not available. For example if the system were competing with
electrical xesistance heat the payback time would be under 4 years.
Although constructing the system is not feasible for the school, it is conceivable that a separate
corporation acting as a heating "utility" could lease the land from the school, build and install the
system and sell heat to the school at a rate lower than what the school currently has to pay for
heat. Such a corporation might be able to supply heat competitively because of the depreciation
expense they could claim. Further analysis of this concept is currendy being done as part of the
study.
CONCLUSIONS AND RECOMMENDATIONS
The first two objectives of the study have been accomplished. A pond capable of collecting and
storing an ample amount of solar heat energy for heating the school has been designed, and a heat
pump with efficiencies high enough to make the system operating costs competitive with the
existing heating system operating costs has been selected.
The third objective, to provide a total system cost that would allow for a 10 year payback from
operating cost savings, has not been accomplished. However, it appears that seasonal solar
storage heating systems can, under certain circumstances, be economically attractive. Economic
feasibility depends upon the installed and operating costs of the competing system and the
financial status and operational structure of the organization constructing and operating the
system.
The completed study will contain specific recommendations for CHPSS systems including
applications guidelines and engineering design guidelines.
ACKNOWLEDGEMENT
The study is being conducted by members of ACCEPT (Allegany County Conservation and
Energy Project Taskforce), a division of the non-profit environmental organization STEPS
(Southern Tier Environmental Protection Society) for NYSERDA (New York State Energy
Research and Development Authority).
Copies of the completed study will be available after May 1,1991 from
W. William Wisniewski, P.E., ACCEPT/STEPS, 334 North Shore Rd„ Cuba, N.Y. 14727.
Ref: Heat Pump Assisted Solar Heating Systems (With and Without Annual Cycle Heat Storage).
-------
1517
THE SARO PROJECT - SOLAR HEATING PLANT WITH SEASONAL STORAGE
J-O. Dalenback
Building Services Engineering,
Chalmers University of Technology, Sweden
ABSTRACT
A new solar heating plant, with a novel storage design, has been built in a new residential
building area south of Goteborg, Sweden. The plant will provide space heating and DHW
requirements in 48 flats, using roof-integrated solar collectors and a new type of insulated
and water-filled steel tank.
The main scope of this RD&D project is to evaluate system performance and economics
with advanced roof-integrated collectors and seasonal storage in new small residential
building areas. The paper describes the overall project and its main features.
KEY WORDS
Solar heating; roof-integrated collectors; seasonal storage; rolled tank; residential heating.
INTRODUCTION
The evaluations of the solar heating plants Ingelstad and Lambohov (1980-83), showed that
both cheaper and more efficient solar collectors and heat stores were required to make
solar heating with seasonal storage competitive with conventional smaller heating plants
(100-500 residential units). They also revealed substantial scope for cost reductions in the
system design, particularly in piping and control (Dalenback, 1988a).
A pre-study for a new smaller solar heating plant was initiated in 1984, with the objective of
designing a cheap insulated heat store together with a cheap and simple system, as new
collector designs were developed in other projects. The Saro project is the most recent
result of this study (Graslund, 1988).
The solar collectors are financed using the same kind of governmental building loans as for
the houses, while the storage is financed by experimental building loans. The Dept of
Building Services Engineering, Chalmers University of Technology, is responsible for the
evaluation, in co-operation with the Monitoring Centre at the University, on behalf of the
Swedish Council for Building Research.
-------
1518
DESCRIPTION OF THE SURROUNDINGS
Saro is located on the Swedish west-coast about 20 km south of Goteborg. The average
annual outdoor temperature is 7 °C, the heating design temperature is -16 °C and the annual
global solar radiation on a horizontal surface is close to 3 600 MJ/m2. The building area is
located in a small valley between rocky hills and consists of a group of nine small
multifamily houses with two floors.
SYSTEM DESIGN
The houses have ordinary roof slopes (27°) and the collector roofs are facing SSE to SE.
The collectors cover about 60 % of the south facing roof area.
The solar collectors, as well as the space heating and domestic hot water (DHW) systems in
the houses, are connected to a small central heating plant via insulated iron pipes in the
ground. The storage is close by the heating plant and supplementary heating is at present
managed by a conventional oil boiler.
Solar Collectors
The solar collector used is an advanced version of the roof-integrated collector type
previously used in several Swedish projects (Sunstrip absorber, acrylic cover).
In Saro this collector type is complemented with a convection barrier (TPFE-foil), thus
improving the thermal performance using a large temperature difference in the water heat
store. The total collector area is 775 m2 (aperture area 725 m2), and the collectors were
built and put into operation during the late summer 1989.
Absorber
Aluminium foil—
Mineral wool"
Corrugated metal'
Timber —
framework
Fig. 1. Roof-integrated collectors. The collector in Saro is
complemented with a convection barrier (TPFE foil)
between the acrylic cover (plexiglass) and the absorber.
-------
1519
Heat Store
The final choice for the storage design in Sard fell on a cylindrical conventional tank
construction, in combination with loose mineral wool insulation (blown into place), placed
in a rock pit. A spray polyurethan insulation in combination with plastic sheets (HDPE,
TPFE) or reinforced spray concrete as water-proofing liners, were two other alternatives
that were studied in advance.
The tank was rolled together on site out of 2 mm galvanized steel plates with a special
equipment. As this tank can have a self-supporting top, the outer roof construction was
changed from an original design with concrete slabs to aluminium sheets on iron beams,
which is much cheaper. The height of the tank is 7 m and the water volume is 640 m3.
The heat store was originally supposed to be a rectangular insulated and water-filled rock
pit, water volume of about 1 400 m3, with a self-supporting top similar to an earlier R&D
project in Vaxjo (Dalenback, 1988b).
The pit excavation was done already in 1988 in connection with the planning of the building
area and the final construction was supposed to be completed in the beginning of 1990.
The change of storage design to a cylindrical tank (inside a rectangular pit) resulted in a
reduced storage volume. The main advantages with this construction is that all parts are
established technologies and that the usable temperature range can be large, about 35-95 °C
compared to 35-80 °C for the original design.
This kind of tank is normally used for crops in farms, different liquids in industries and as
water storage in developing countries. One disadvantage is that it is difficult to apply
insulation on the outside because of the thin plate. Ongoing work aims to investigate the
possibility to use some kind of reinforced plastic skirt to keep the insulation in place on
above ground applications.
Roof
Mineral
wool
Roof support
Drainage
Fig. 2. Saro - Schematic drawing showing the storage design.
-------
1520
Space heating and DHW
The load is space heating and DHW for a small residential building area with 48 flats in
well insulated multifamily houses. The houses are calculated to have an annual heat
requirement of the order of 1 300 GJ, including DHW and distribution losses, which
corresponds to about 360 MJ/n# heated floor area. The space heating system is a low
temperature hydronic radiator system with a 55 °C supply temperature required at an
outdoor design temperature of -16 °C. The characteristic supply temperature requirement
is 50 °C for DHW, and the characteristic return temperature is 35 °C.
Heating plant
The design of piping and control in the heating plant is similar to the system design in other
late R&D projects (Dalenback, 1988a). The main difference is that there is a direct
connection between the collector circuit and the load circuit in Saro. This will not improve
the performance but it reduces the amount of piping in connection to the storage. The
interaction between these to circuits will be evaluated.
Load
Collectors
Water
storage
Suppl.
Boiler
Buffer
Rad DHW
Heating Plant
Fig. 3. Saro - System schematic diagram.
System operation
The heating plant in Saro has been in operation with solar collectors and the short-term
storage tank (boiler buffer storage) since autumn 1989 and the seasonal storage was put
into operation in May 1990.
The plant is controlled using a few conventional and simple thermo-stat control units and,
so far, there has been no draw-backs concerning system operation.
-------
1521
THERMAL PERFORMANCE
System Rating
Previous evaluations of solar heating systems with seasonal storage have resulted in some
general guide-lines concerning the rating of these systems (Dalenback, 1988a and 1990). A
desirable solar fraction of around 75 % requires about 0.7 mz of collectors per GJ annual
load and 2-3 m3 of water storage volume per m2 of collector area depending on storage and
load characteristics.
The Saro Project
Thermal performance calculations for the Saro plant were carried out at an early stage
(Graslund, 1986). Simulations were made for an average climatic year and an estimated
annual heat load of about 1 300 GJ. 800 mz of roof-integrated collectors (with convection
barrier) and a 1 400 m3 water heat store were expected to give a desirable solar fraction of
the order of 65 %.
Saro - Total load
(GJ)
(oC)
200
100
160 --
120 --
80 --
40 --
0
-- 80
-- 60
-- 20
-- 40
0
JFMAMJJAS0ND
~ Suppl. heat ¦ Solar heat O Storage temperature
Fig. 4. Saro - Preliminary calculations showing total load, utilized
solar heat and average storage temperature with actual
R&D storage volume.
-------
1522
The actual built plant consists of 725 m2 of collectors, which gives about 0.6 m2 of collector
area per GJ of annual load. The volume of the water heat store is, however, only 640 m3
due to the R&D characteristics of the project, and the storage volume is thus less than 1 m3
per m2 of collector area. This will of course result in a lower solar fraction.
The real performance with actual sizing will be evaluated with the simulation model
SIMSYS (Dalenback, 1988a), as well as TRNSYS, using the SST storage type (Eftring and
Hellstrom, 1989). Preliminary calculations (SIMSYS) of the thermal performance gives a
solar fraction close to 40 % (Fig. 4.).
DISCUSSION
The main scope of the S&rd project is to gain experience from the new seasonal storage
design. The final design is based on conventional technologies to a large extent. An
estimated investment cost, based on construction experience in Saro, is around 900 SEK/ms
for a storage volume of the order of 2 000 m3.
This cost is less than the investment cost for a conventional insulated steel tank, and there
are great possibilities to reduce this cost by a careful commission of a new larger storage.
This storage design is, furthermore, also possible to use in an accepted pit design in ground,
as well as, in an ordinary tank design above ground.
A future investment cost around 500 SEK/m3 (storage volume of 5 000 - 10 000 m3) is
judged to be realistic and creates a large potential in solar heating and other applications.
The investment cost for the solar pilot-plant was, in this case, 4.6 MSEK and the total cost
for the whole building area, including pilot-plant, was 46.5 MSEK (February 1990). It could
thus be possible to reduce the amount of purchased heat with 70 % by an investment that
only amounts to 5-10 % of the total investment cost for a similar building area in the future.
REFERENCES
Dalenback, J-O. (1988a). Large-scale Swedish Solar Heating Technology - System Design and
Rating. Document no D6:1988, Swedish Council for Building Research, Stockholm.
(National report, Task VII, IEA SH&CP).
Dalenback, J-O. (1988b). The Swedish Pilot Plant Kronhjorten - Description and some
Experiences from Design and Construction. Report no 1988:2, Dept. of Building Services
Engineering, Chalmers University of Technology, Goteborg.
Dalenback, J-O. (1990). Central Solar Heating Plants with Seasonal Storages - Status Report.
Document no 1)14:1990, Swedish Council for Building Research, Stockholm. (Final
report, Task VII, IEA SH&CP).
Eftring, B. and G. Hellstrom (1989). Stratified Storage Temperature Model - Manual for
Computer Code, February 1989, Dept. of Mathematical Physics, University of Lund,
Lund.
Graslund, J (1988). Solar Heating with Seasonal Storage in a Residential Building Area in
Sard. Report no R16:1988, Swedish Council for Building Research, Stockholm. (In
Swedish)
-------
2.11 Active Heating II: Heating System Performance
-------
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
-------
1525
ENTROPY ANALYSIS ON A SOLAR SPACE HEATING SYSTEM
Yulin Shao
Photovoltaic Program, University of Lowell
Lowell, MA 01854, U.S.A.
ABSTRACT
A solar space heating system at a steady state is analyzed in the present paper. The
study shows that the overall entropy increase is dominated by the contributions from
the tank and the heat exchangers. Although the entropy production from the solar
collector decreases as the temperature of the storage tank (or the temperature drop
across the heat exchanger) increases, the overall entropy production for the whole
system increases because the contributions from the tank and heat exchanger increase
very rapidly. For a given system, excessively large collectors are detrimental because
much of the exergy absorbed by the collectors is wasted in the storage tank and heat
exchangers. The study suggests ways to improve the design of a solar heating system.
For example, the tank may be designed so that only part of the water is kept at high
temperature. Also, the system may utilize two smaller exchangers to lower entropy
production.
KEYWORDS
Entropy analysis; solar; heating system; thermal energy; exergy.
INTRODUCTION
Solar space heating systems (SSH5) has shown a good potential for reducing energy
and cost consumption of building; in heating seasons. A couple of methods have been
developed to design a solar heating system and analyze its thermal performance
at design stage (Collares-Pereira, 1979; Klein, 1976, Holtz, 1988). The energy balance
approach is widely used in system design. However, thermal energy at different
temperatures has distinct value in both ability of doing work and economics. The
concept of exergy is useful to illustrate the difference. Furthermore, an understanding
of the sources of lost exergy in a thermal system may help to improve the system
design and economic performance. The lost exergy of a system is proportional to its
total entropy production. For SSHS, the entropy production is equal to the net sum of
the entropy changes from each of the system components (solar collectors, storage
tank, heat exchangers, solar heating space, and the surroundings). The different
contributions from each component are functions of temperatures, coolant flow rates,
and thermal resistances. On the basis of a simple steady-state analysis, it is possible to
suggest that the total lost exergy in a solar heating system which uses water as a
receding page blank
-------
1526
coolant can be lowered effectively by incorporating a double cylindrical storage tank
and a compound heat exchanger into the system.
ENTROPY PRODUCTION IN THE SYSTEM
A SSHS usually consists of solar collectors, a storage tank, heat exchangers, fans and
pumps. Sometimes, it includes an auxiliary heating system. For the analysis
presented here, the entropy production arising from every component is calculated
consistently for a given system load (i.e., for a constant heat delivered to a room at
constant temperature) and external temperature, assuming different temperature
drops across the heat exchanger.
A schematic of a simple liquid-state SSHS is illustrated in Fig. 1. The thermal energy
absorbed by the collectors is transfened to the coolant in the collectors, and stored in a
storage tank when the hot coolant returns to the tank. A load pump delivers the hot
coolant to the heat exchanger in the space to heat the air or other medium, and then
the coolant returns to the tank.
Space
Collector
Storage tank
Aux>
heater
Collector pump
Load pump
To
Fig. 1 Solar space heating system
For any process, the total change in the entropy of a system, dS can be written as
dS = dSe + dSi, (1)
where
dSe = the contribution to the change in entropy which arises from the exchange of
entropy with the surroundings,
dSj = the contribution from internal entropy production.
dSe can be positive, negative or zero depending on the process. dS; is always positive,
(or zero for reversible process).
In a SSHS, entropy productions arising from each component and process are: SI,
arising from solar energy absorbed by the collectors; S2 , from heat exchange between
-------
1527
surrounding and collectors; S3, entropy production of the coolant in collectors; S4, the
result of the lost heat of the storage tank into surrounding; S5, entropy change of the
storage tank caused by the coolant from the heat exchangers; S6, produced by the heat
exchange between the space and heat exchangers; and S7, from the lost heat of space
into surrounding.
For heat exchange between two constant temperature heat sources, the entropy
production is (Sama, 1983)
(2)
as = q2^
TcTh,
where
Th temperature of hot source,
Q heat transfesed.
temperature of cold source,
Equation (3) can be used to compute the entropy production in the processes of heat
absorbption or exhaustion
S=Q/T, (3)
All of the entropy contribution from each component and process are listed below
Ac(x (X)ave
51 =
52 =
Tel ,
AcFrUc(Td-T0)2
T0Tci ,
g _ EfllCpifTin-Tout)
3" TinTout
S4 =
ASUS(TS-T0)
T0To
g _ lh2Cp2(Tj1(in"Th/out)
Th,inTh,out
>
g _ CthAh(Th_ Tr)2 __ m2Cp2(Th/jn"Th/out)(Th" Tr)
6" = ThTr
S7 =
UrAr(Tr-T0)2
T0Tr
(4)
(5)
(6)
(7)
(8)
(9)
(10)
where
Ac
area of collectors,
Fr
fraction,
Is
solar insolation,
Uc
heat loss coefficient of collectors,
Tci
collector temperature,
To
surrounding temperature,
m
coolant flow rate,
cp
heat capacity of coolant,
Tin
inlet temperature of collector,
Tout
outlet temperature of collector,
Ts
average temperature of tank,
As
surface area of tank,
«h
heat transfer coefficient,
Tr
room temperature,
Th,in
inlet temperature of heat exchanger, Thout
outlet temperature of heat exchanger,
Ah
area of heat exchanger,
us
heat loss coefficient of tank,
ur
heat loss coefficient of house,
Ar
surface area of house,
a)ave average transmittance-absorptance product.
-------
1528
If the initial temperature of the storage tank, load and useful gain of the collectors are
known, the temperature change of the storage tank can be calculated (Hsieh, 1986).
NUMERICAL ANALYSIS
For the analysis, the entropy production arising from every component is calculated
consistently for a given system load (i.e., for a constant heat delivered to a room at
constant temperature) and external temperature. At steady state, the storage tank
temperature is proportional to the absorbed solar energy as well as to the temperature
drop across the heat exchanger. Therefore, the temperature drop will fix the storage
tank temperature. But the temperature drop across the heat exchanger varies at
different steady states.
The important parameters used in the analysis are listed in table 1.
Table 1
T0 = -20 C, - 5 C, 10 C.
Tr = 21 C.
Uc = 3.48W /m'K.
ASUS = 12 W/K.
miCpl = 125.7 W/K.
m2CP2 =125.7 W/K.
thermal capacity of storage tank,
6688 kj/ C.
Fr = 0.848.
Qu = Qumax sin {180(day time/total
day time)}.
Qload = (UA)h(TrT0).
Qumax = 76 MJ/hr.
(x oc)aye = 0.87.
The total entropy production and the contributions from the tank collector and heat
exchanger are plotted in Figure 2, 3, 4 and 5 as functions of the temperature difference
between storage tank and surrounding (AT), and surrounding temperature.
S (W/K)
SCW/K)
total
solar
tank
hex
2000
2000
hex
solar
tank
total
Cr°-cr
1000
1000
1—I
80
A T
20
40
60
80
A T
20
40
60
Fig. 2. (To = 10 C, Sun=l .0) Fig. 3. (T0 = -5 C, Sun = 0.8)
DISCUSSION AND SUGGESTIONS
The lost work or lost exergy of a system depends on its total entropy production. In
-------
1529
SCW/K)
S (W/K)
2000 -\
total
3000 -[
¦*— solar
total
2000 -
1000
0
0
0 20 40 60 80
A T
Fig. 4. (T0 =-20 C, Sun = 0.6)
0 20 40 60 80
A T
Fig. 5 (T0 = -5 C, Sun = 0.6)
the analysis, it is found that the overall entropy increase is dominated by the
contributions from the tank and the heat exchanger. Although the entropy
production from the solar collector decreases as the temperature of the storage tank
(or the temperature drop across the heat exchanger) increases, the overall entropy
production for the whole system increases because the contributions from the tank
and heat exchanger increase very rapidly. On the other hand, a high tank temperature
allows the collectors to absorb more exergy from solar energy. Therefore, the usual
siptem approach is to improve the insulation of the tank to increase the tank
temperature while keeping a low entropy production. An alternative approach is to
change the structure of the tank, letting part of the coolant to be kept at a high
temperature and the rest at a relatively low temperature, so that the average
temperature is relatively low.
The study suggests ways to improve the design of a solar heating system. For
example, the tank may be designed to consist of two coaxial cylinder tanks, so that
only part of the coolant is kept at high temperature. Because part of the coolant would
be at low temperature, a better (and more expensive) insulator may not be needed. At
noon, the "hot" coolant with a high temperature returns to the inner storage tank; at
other times, the "warm" coolant returns to the outer storage tank. Also, the system
which has a big infiltration and a very low surrounding temperature may utilize two
smaller exchangers to lower entropy production. The hot coolant is only used to heat
the air at high temperature and the warm coolant used to heat the cold air. When the
tank temperature is required very high, for example, the daily service water is heated
by the coolant from the tank, these considerations will be specially suitable. Using the
modified design, the temperature of the tank can be relatively low, resulting in a low
total entropy production of the solar system. A SSHS design and performance
analysis based on these considerations will be discussed in a later paper by the author.
For a given system, there are limitations on how small the temperature drop can be,
however. If the temperature drop is too low, the storage tank and the heat exchanger
have to be too large. In addition, there is less exergy in the heat from the solar
collector when the tank temperature is low, i.e., the quality of the heat is low. The
study also shows that excessively large collectors are detrimental because much of the
exergy absorbed by the collectors is wasted in the storage tank and heat exchanger.
-------
1530
ACKNOWLEDGEMENT
The author is indebted to Dr. Jose G. Martin of University of Lowell for his
encouragement and useful suggestions.
REFFERNCES
Collares-Pereira, M. and A. Rabl (1979). Simple procedure for predicting long term
average performance of nonconcentrating and of concentrating solar collectors.
Solar enerev. Vol. 23, 235-253.
Holtz, M.J. (1988). A residential energy design guide line development methodology.
Proceedings of 13th NFSC. 26-32.
Hsieh, J.S. (1986). Solar energy engineering,. Prentice-Hall, Inc.
Klein, S.A.,W.A. Beckman and J.A. Duffie (1976). A design procedure for solar
heating system. Solar Energy. Vol. 18,113-127.
Sama, D.A.(1983). Cost saving through lost-work analysis at heat exchangers.
Energy process. Vol. 3,21-2-219.
-------
1531
COMPARISON OF LARGE SIZED SOLAR AIR
AND LIQUID SYSTEMS FOR INDUSTRIAL SPACE HEATING
C. C. Smith
Solar Energy Applications Laboratory
Colorado State University
USA
KEYWORDS
Solar Heating; Solar Air Collectors; Solar Liquid Collectors.
ABSTRACT
Two large solar heating systems were constructed in 1978 and 1979
to heat bus service buildings in the Denver area. One system
utilizes air collectors with rock bed storage and the other liquid
(antifreeze solution) collectors with water tank storage. The Air
system is 4300 square meter collector area for a 32,500 square
meter floor area building. The liquid system is 3700 square meter
collector area for a 23,800 square meter building. Both col-
lectors have single glass covers with optically selective absorber
surfaces. Solar energy is stored in 2200 cubic meters of one-two
centimeter sized rock in the air system, and in 300 cubic meters
of water in the liquid system. Seasonal (annual) performance was
obtained by measurement on approximately half of the days with
heating requirement, together with f-chart modeling and Typical
Meteorological Year (TMY) data for Denver. These results gave an
annual system efficiency of 38 percent for the air system and 26
percent for the liquid system. Solar fraction of the total heating
load was 22 percent for the air system and 15 percent for the
liquid system. Maintenance and operational difficulties were also
noted in the study.
INTRODUCTION
The two buildings in this study serve the same function, as bus
maintenance and storage facilities, and are in the same climatic
location. The solar systems have operated without interruption,
except for routine maintenance, for more than ten years. Seasonal
energy monitoring was performed for the first time on these
buildings in 1987 (air system) and in 1989 (liquid system) by the
Solar Energy Applications Laboratory at Colorado State University.
Due to bus fumes in the buildings, large quantities of fresh air
must be supplied; accounting for as much as eighty percent of the
heating requirements. The fresh air (ventilation) heating load
corresponds to the frequency of cold engine starting and operation
of busses within the building. The highest traffic periods are in
the early morning and early afternoon.
-------
1532
AIR SYSTEM DESCRIPTION
The solar system was initially well designed in two respects: the
collectors heat only fresh outdoor air rather than recirculated
building air, and solar energy is used in combination with exhaust
air heat recovery. The outdoor air heating feature is suited to
the building demand for as many as fifteen fresh air changes per
hour. The solar system is capable of supplying approximately one
fourth of the total building ventilation requirement.
Figure 1 is a diagram of the air systems. Together on one building
there are five individual systems which are identical except for
size, and one which includes a domestic hot water preheating coil.
Control dampers and fans in each of the five systems determine the
modes of operation in response to the availability of solar energy
and the building heating demand. These modes are: heating directly
from the solar collectors, storing solar energy from the
collectors, and heating the building from storage.
The systems can simultaneously heat the space and store energy in
varying proportions, and exhaust excess solar energy through the
storage units; a means for summer ventilation of the collectors.
When the heat recovery plus solar energy are insufficient to meet
the heating requirements, fuel heated steam coils are activated.
The solar and heat recovery systems continue to operate as air
preheaters when the steam coils are on.
Outdoor air enters the collectors at point A. It is drawn through
the collectors and into the building by means of a collector fan.
The air at point B is directed in varying proportions to building
heating (point C), or through the rock storage bed and exhausted
outdoors (point F). At point C the solar heated air mixes with
outdoor air that is preheated by the heat recovery system. The
mixed air is heated by steam coils, if needed, and finally is sup-
plied to the building through overhead duct distribution (point E) .
Solar
Collectors
Fresh Air Intakes
Heat Recovery
Coil
Steam Heat
Coil
Solar Fan
Storing Heat
nviuny 11 uui
Storage
(D V
N:
Heated Air
Supply To
BuiMng
Fig. 1. Diagram of the solar air system.
-------
1533
The proportion of air at point B directed to storage depends upon
the building heating demand as well as the requirement for adequate
total ventilation. While operating in the heating from solar
storage mode, outdoor air enters the rock bed storage unit at point
F and proceeds past point B to point C. There it mixes with air
off the heat recovery coils as in the direct solar heating mode.
The solar collector array consists of 2420 pre-manufactured panels.
The panels are mounted in rows supported at a 55 degree tilt, due
south, by steel supports. Collector construction includes a single
pane of low-iron tempered glass, an air channel below the glass
cover, an internal manifold arrangement, and insulation on the
backside of the steel casing. The absorber surface, which is one
side of the air duct, is coated with a selective black chrome
surface. Minimum absorptance is 0.95, and maximum emittance is
0.15. Transmittance of the glass cover is 0.89.
There are 2200 cubic meters of rock bed storage, or .51 cubic meter
per square meter of collector area. The rocks are heated by air
flow in one direction (downward) , and later release heat to air
flowing in the reverse direction. An important factor of this rock
bed is the open loop design, wherein the bottom of the bed is open
to the outdoor air. The outdoor air intake to storage was
specified, because the solar system heats only outdoor air, not
recirculated building air. A disadvantage in this design is seen
when air is discharged from the bed at a temperature higher than
the ambient air temperature. Under such a condition the air should
be returned to the collector inlet for best performance. In
practice, however, this has not represented a significant
performance penalty.
A factor in storage performance was the non-uniform size and
irregular shape of the rock. As a result, the packing density of
the bed was high, and the air flow restricted. In operation this
restriction resulted in air flow from the collectors to rock
storage of about one half that of air flow from the collectors to
the building directly. However, this difference was not
exclusively due to the restriction of the rock bed. There are two
fans in series when the air is directed to the building, and only
one, the collector fan, when directed through storage. While
higher air flow through collectors increases solar performance by
lowering collector temperatures, there are times when heating the
rock bed from the collectors benefits from the reduced flow.
Particularly in this design, where the rock box discharges air to
the outside, moderate air flow is desirable. High air flow
following a period of high energy storage could actually discharge
rather than charge the storage. A low flow rate causes the
temperatures in storage to increase, resulting in more stored
energy available for heating the building.
LIQUID SOLAR SYSTEM DESCRIPTION
The liquid collectors supply energy via a double loop piping
design. Thus an antifreeze loop is separated from a storage water
loop by a tube-in-shell heat exchanger. Energy received at the
heat exchanger is delivered directly to space heating, or to a 300
cubic meter water tank, as stored energy for subsequent heating.
Figure 2 is a flow diagram of this system.
-------
1534
SOLAR
PANELS
T—11/
TO BUILDING
HEATING
P—33
HE—1
HE—3
P—32
BOILERS
STORAGE
Fig.. 2. Diagram of the liquid solar system.
A control system, which is integrated with the overall building
heating and ventilating system, manages the collection, storage,
and distribution of solar energy. When the collector liquid
reaches a temperature of 8 degrees above the storage tank,
collection pump P-33 and storage water pump P-32 start
automatically to bring energy from the collectors into the building
through heat exchanger HE-1. As seen in Figure 2, the storage water
leaving HE-1 passes directly through HE-2, the solar energy to
space heating heat exchanger, and then returns to the storage tank.
The liquid system is not capable of supplying both solar and fuel
energy simultaniously as the air system does. The selection of
solar vs. fuel energy supply depends upon the solar storage
temperature in relation to the outdoor temperature. The building
will not accept solar energy for heating if the storage water
temperature is less than the function 56 - .73 X Outdoor Temp.
When the storage temperature is greater than this function, the
space heating distribution water is directed through HE-2, which,
in combination with the storage pump P-32, delivers solar energy
to the building. Any combination of collection, storage, and
supply of solar energy to the building can thus take place
depending upon the status of the associated pumps and valves.
Whenever the solar storage tank temperature is insufficient for
space heating, the requirements are met by natural gas fuel
boilers. The hot water distribution is then directed through HE-
3 and bypasses HE-2.
The liquid collector panels are oriented due south and at a tilt
angle of 55 degrees from horizontal. Collector construction
includes a single cover sheet of 4 mm tempered water white glass.
The solar absorbing surface is black chrome selective on steel. All
piping, internal and External to the collectors, is steel.
Insulation consists of mineral wool and polyurethane. The entire
assembly is housed in a galvanized steel box.
MONITORING RESULTS
The air system was monitored over four months, February through May
of 1988. The liquid system also was tested over a four month
-------
1535
period, January through April of 1989. The results of these tests
were used as input to an f-Chart computer model, together with
typical meteorological data for the Denver region (TMY) to generate
annual performance levels for the two systems.
Collector efficiency is presented as a function of collector
temperature minus ambient temperature/ divided by solar radiation
for both the air and the liquid systems (Figure 3). The pre-
installation test performance is shown also in these figures.
Data using the inlet temperature would provide no characteristic
curve for the air collectors, as the inlet temperature is always
equal to ambient. Thus there is a mathematical conversion from
the outlet temperature efficiency for the air system in Figure 3.
.60
AIR
-.50
OPERATION
6.5 l/sec-m'
.30'
LU .20
OPERATION
5.5 l/sec—m'1
.10
.00
0.00 0.02 0.04 0j06 0j08 0.10 0.12
LIQUID
.70
DSAT TEST
.60
<4-
M-
LLI
20
OPERATION
.10
.00
0.00 0.02 0.04 0.06 0.08 0.10 0.12
-------
1536
COLLECTOR LOSS
2 ELECTRICITY
SOLAR
DIRECT
SUPPLY
COLLECTED
STORED
2 LOST
COLLECTOR LOSS
SOLAR
2 ELECTRICITY
DIRECT
COLLECTED
SUPPLY
7 STORED
1 LOST. ^
Fig. 5. Energy distribution in air (left) and liquid
(right) systems.
CONCLUSIONS
The thermal performance of the air system was higher, due primarily
to taking only outdoor air into the collectors. The liquid system
could, in principle, also operate at temperatures near to the
ambient with a redesign to reduce flow rates and maintain highly
stratified temperatures in storage.
The air system exhibited some reduced performance due to a lower
than specified air flow rate through the collectors. One of the
air systems operated at 62% of the specified flow rate. The rocks
in storage were highly compacted, preventing proper air flow
through storage. Also, air filters filled quickly from bus fumes.
The liquid system showed two significant performance-reducing
factors. The entire collector piping array outside the building
was not insulated. The low operating efficiency shown in Figure
3 reflects the large pipe heat losses. In addition, the
temperature setting for solar heating was adjusted higher than
originally specified. This resulted in higher temperature, less
efficient, liquid collector operation than otherwise.
Operation and maintenance of the two systems was also considered
in the study. The air system had more problems with air pollutants
at the site, and also had to have more adjustments to the
equipment. The liquid system had some frozen pipes, pipe
corrosion, and cover glass breakage.
ACKNOWLEDGEMENTS
The author expresses his appreciation to the US Department of
Energy, the Urban Mass Transit Authority for their financial
support, and the Denver Regional Transportation District, and the
Rockwell Corporation for their assistance.
-------
1537
PERFORMANCE OF SOLAR PREHEATED VENTILATION AIR SYSTEMS
Stephen C. Carpenter, P.Eng.
John P. Kokko
Enermodal Engineering Limited
368 Phillip Street
Waterloo, Ontario N2L 5J1
ABSTRACT
Three active solar systems for pre-heating ventilation air in industrial buildings are examined. Monitored
and predicted performance are compared. Several advantages other than the solar gains are identified
and quantified where possible. Average solar collection efficiencies of as high as 45% were observed.
Total energy savings of over 45 GJ/m2 are possible.
KEY WORDS
Active solar; energy conservation; solar collectors.
INTRODUCTION
One of the most cost-effective applications for active solar heating is the preheating of building
ventilation air. In a solar preheat systems, air-based collectors can be mounted vertically on the south
wall of the building. Outdoor air is pulled through the collectors and blown into the building at ceiling
level to meet ventilation requirements. The collectors operate at their maximum efficiency because the
collector inlet air temperature is at ambient temperature. In the summer, bypass dampers open to avoid
solar heating of ventilation air. In industrial buildings, further energy savings are achievable through
both the reduction and recapture of wall heat loss and the destratification of building air.
Over the past three years, the performance of three different solar preheated ventilation air systems in
Ontario has been monitored. In the first system, ventilation air is drawn between a corrugated absorber
and a fibreglass glazing up the wall and into the building. The second system consists of an unglazed
wall painted dark brown. Air intake grilles mounted on the underside of a canopy capture solar heated
air as it rises up the wall. With the third system, a perforated plate is used instead of a glazing. Air
is drawn through the perforations, up the passage between the perforated absorber and the south-
facing wall and into the building. The three systems are described according to their top cover:
glazed, unglazed and perforated-plate. All three systems were manufactured and installed by Conserval
Engineering Ltd. of Toronto under the trade name "Solarwall". This paper describes the operation and
the monitored and predicted performance of the three systems.
GLAZED COLLECTOR SYSTEM
The glazed collector system is illustrated in Figure 1. This system was installed at the Ford Motor
Company Plant in Oakville, Ontario in 1986. The system combines solar energy collection with the
energy conservation features of added insulation and building air destratification into one package. The
system is made up of standard building wall components including 38mm (1 1/2") of fibreglass
insulation, corrugated steel wall cladding for the solar absorber, and translucent fibreglass-reinforced
plastic glazing. Make-up air is drawn through openings at the bottom of the solar panels, heated and
-------
1538
Summer
by-pass -
dampers
Outside
air intake
Fig. 1 Glazed Collector System
then distributed through a perforated flexible duct. The fans operate 24 hours a day, 7 days a week.
The project consists of 16 independent systems, each with its own fan, distribution duct and control
system. The total collector area is 1877 square metres.
Collector outlet air is normally colder than the air in the building. The control system operates a
modulating damper mixing collector and indoor air to avoid thermal discomfort for building occupants.
Destratification occurs as a layer of cool air is introduced across the top of the building and mixes with
indoor air as it falls. The system is operated such that collector flow is reduced and recirculation flow
increased to maintain minimum 10°C delivery temperature.
In late 1987 a data acquisition system was installed by INRS-Energie of Varennes, Quebec on one of
the systems with a collector area of 237 square metres. Data was collected and sent via modem to their
offices for the one-year period of December 19S57 to December 1988. The data analysss and computer
simulations using the SIMAIR program [Enermodal, 1989] were carried out by the authors.
Table 1: Solar Energy Collected - Glazed System
Collector Area: 237 m2 Solar Collected plus
Solar Collected Recovered Building Heat Loss
Monitored Simulated Monitored Simulated
Month (GJ/day) (GJ/day) (GJ/day) (GJ/day)
Dec./87
0.56
0.50
0.64
0.68
Jan.
1.09
0.80
1.24
1.05
Feb.
1.33
1.25
1.56
1.47
Mar.1
(1.36)
(1.49)
(1.55)
(1.67)
Apr.
1.40
1.73
1.53
1.85
May
0.70
0.88
0.85
0.98
Sept.
0.83
0.86
1.20
0.94
Oct.
1.30
1.17
1.55
1.29
Nov.
0.99
0.81
1.16
0.95
Dec./88
1.84
1.66
2.07
1.88
Year2
311.0
306.9
365.6
350.1
GJ/m2/year
1.31
1.29
1.54
1.48
1 Based on linear interpolation between February and April.
2 Year GJ = (GJ/day) * (total days/months), summed over months shown. December values were
averaged.
-------
1539
The results of the monitoring and simulation are given in Table 1. Because of difficulties with the
monitoring system and data transmission, most of the months did not contain a complete set of data
Results have been presented on a per day basis, on the assumption that the performance on the
missing days was equal to the monitored days for that month. The system was not monitored for June,
July and August because it was felt that the building would require no heating, and the solar
contribution would therefore be negative. No monitored data were available for March; results are
interpolated from February and April.
Maximum air flow through the collector was 7500 l/s (15,900 CFM). This was lower than expected,
partly because of extra resistance from the flow measuring station, and partly because of a higher than
expected pressure drop through the rest of the system. The full implication of a low air flow was not
analysed. It is expected that collector efficiency will improve when the higher air flow is used (i.e.,
collector outlet temperature is above 10°C). The results indicate that the glazed system delivered a total
of 1.54 GJ/m2 over the one year period. The monitored performance compares favourable with the
simulated performance of 1.48 GJ/m2. The delivered energy is a combination of solar energy collected
during daylight hours, heat lost through the building wall that is captured by the system, and solar
energy that is stored in the wall during the day and captured by the system at night. Monitoring
showed that the energy collection was 1.31 GJ/m2 per year during daylight hours and 0.23 GJ/m2 during
the night.
No conclusive results could be drawn as to the energy benefit of the destratification aspects of the
system. While the system was operating, there was very little temperature stratification. The air
temperature at the ceiling was generally only 2 Celsius degrees warmer than the air temperature at the
floor. It was difficult, however, to determine what the long-term average temperature stratification was
before the wall was installed. The energy benefit of insulating the wall was calculated to be 1.0 GJ/m2.
Excluding any energy benefit due to destratification, the annual energy savings from this system is 4692
GJ (1,304,515 kWhr), assuming similar performance for the entire wall.
UNGLAZED COLLECTOR SYSTEM
To reduce the cost of solar preheated ventilation air systems, Conserval developed an unglazed version
of their Solarwall. In this system, the south wall of the building is painted a dark colour and air intakes
are installed across the top of wall. Incident solar radiation warms the air near the wall, which rises up
the wall by natural convection to be pulled into the building through the intakes. As with the glazed
system, the generally cooler solar air is intended to destratify building air. Further energy savings are
realised because solar radiation absorbed by the wall reduces building heat loss. With this system, the
outdoor air that enters the building comes from closer to ground level than with conventional roof-
mounted make-up air units. If the air on the south side of the building is warmer than at the roof (due
to warming from the ground), there is a further energy benefit.
The performance of an unglazed system was examined by monitoring the Solarwall system installed at
the McDonnell-Douglas Canada plant in Toronto. The plant is a large airplane hangar approximately
250 metres by 100 metres by 18 metres high. Hangar doors make up most of the south-east wall of
the plant. Most of the doors have been closed permanently. The 2677 square metres of hangar doors
were painted dark brown, and five fan/perforated-duct air handling systems (similar to those at the Ford
plant) were installed. Intake plenums were installed in a canopy overhanging the wall and connected
to the air handling system. To keep construction costs down the intake plenums constitute only 60%
of the canopy, and are distributed uniformly along the full length of the hangar wall.
A PC-based data acquisition system was installed and performance monitored over the 1989-90 winter.
Wind speed was measured on the roof, to obtain a representation of free air stream velocity, and in
the middle of the Solarwall. Monitored results showed that on average the wind speed on the Solarwall
was only 46% of that measured on the roof. The ambient air on the south side of the building was 2
C° warmer during the night and up to 3.5 C° warmer during the day than the air on the roof. The air
flow rates were significantly lower than expected. No definitive reason was obtained for this though it
is believed that higher than expected pressure drops occurred in the air handling system.
-------
1540
Initial readings showed surprisingly high air temperatures at the intakes, even during the night. A site
inspection revealed that there was considerable air leakage at the interface between the hangar doors
and the wall. It was determined that between 6 and 25 % of the air entering the building was actually
recirculated building air. This situation is probably unique to this building because of the use of hangar
doors. As such, the performance of this system may not be typical of unglazed Solarwalls.
Nevertheless, an attempt was made to quantify system performance.
Figure 2 shows the computer simulated and monitored performance of the McDonnell-Douglas system.
The simulations were performed with two different collector areas: the full collector area (2677 m2) and
that portion of the wall serviced by the air handling system (1606 m2). Two monitored values are shown
for January and February, representing the range of uncertainty caused by the air recirculation. The
SIMAIR computer predictions for these two months are within the range of monitored results.
Extrapolating to a full year, the annual solar energy delivered is estimated to be 0.4 GJ/m2. This can
be compared to the "1.3 GJ/m2 of solar energy collected for the glazed collector system. Excluding the
summer solar radiation, the solar heat collection efficiency is predicted to be between 13.5 and 18.2%.
Solar Energy Collected
(Simulated and Monitored)
250-
200-
150-
Futl Collector Area
i
3
100-
Reduced Collector Area
50-
Monitored Low
JUL AUG SEP OCT NOV DEC
FEB MAR APR MAY
JAN
Month
For the two months of monitoring there were additional energy savings due to reduced wall heat loss
of 233 kWhrs per day and 610 kWhrs per day due to ambient temperature gains. It was not possible
to extrapolate these savings to annual values; however, each of these savings is of the same order as
the solar air heating savings. Monitoring of indoor temperature showed that the Solarwall system
reduced stratification by at least 3 C°. SIMAIR calculations put this benefit at 1.8 GJ/m2. Again, the
reader is cautioned that these results may not be representative of a typical unglazed Solarwall.
Another system at a Hayes-Dana plant is to be monitored in 1991-92 to obtain more accurate
performance information on unglazed systems.
PERFORATED-PLATE COLLECTOR SYSTEM
An alternative to the previous two systems is the perforated plate collector. It removes the need for
unsightly and fire-prone fiberglass glazings. This system has the enclosed air channel of the glazed
system but is 35% less expensive to install because of lower overall material costs. The system is
similar to the glazed system, except that the fibreglass glazing is replaced with an aluminum sheet
containing small holes, the opening at the bottom is closed off, and a second skin of inner material is
not required. In fact, at the Ford plant in Oakville, the 1877 square metres of glazing were replaced
with perforated-plates with both 1% and 2% void fraction. The perforated plate was installed in 1990
and monitoring began in late 1990. Initial results showed the system was performing below
expectations. Site measurements revealed that the maximum air flow was slightly lower (at 4100 l/s,
or 15,100 CFM) than with the glazed system. This would indicate a slightly higher pressure drop
associated with the perforated-wall design.
-------
1541
Field Test vs NSTF Test Data
Ford of Canada, Oakville
NSTF No Wind
0.7-
NSTF HI Wine
(3.5 m/s)
0.5-
I
I
0.3-
0.2-
+ Field Test Data
Fig. 3 1 1
Figure 3 shows early test results for the efficiency of the perforated-plate design as a function of air
flow rate. The three curves are a quadratic fit to collector test results produced at the National Solar
Test Facility (NSTF) using an indoor solar simulator. The curves are for no wind, low wind speed (1.5
m/s) and high wind speed (3.5 m/s). The "+" signs indicate hourly data points obtained from the
monitoring at the Ford plant. The wind conditions for these test points were between 0 and 3.5 m/s,
as measured in front of the collector.
Infra-red thermography (see Figure 4) of the system revealed two areas which might be responsible
for degradation in performance. One is bypass damper air leakage: wall temperatures generally rise
as a function of height, yet, the bypass damper itself was much colder than the surrounding wall, as
shown in Figure 4a When the bypass damper was covered with a tarpaulin (Figure 4b), the lower half
of the wall dropped in temperature, indicating a greater air flow distribution. It was concluded that a
significant amount of air is leaking through the bypass damper. The other problem is that, because
of low air flow, only a portion of the wall is being utilized. As shown in Figure 4, the hottest spots
occur around the upper corners of the wall (i.e., little if any heat is removed from this area of the wall),
(a) (b)
Fig.. 4 Thermograph of Solarwall with Bypass Covered (a) and Uncovered (b)
-------
1542
Pressure measurements showed suction dropped quickly as distance from the fan increases. It was
felt that the air at the upper corners is stagnant or in fact may be leaking out due to buoyancy effects.
NSTF testing showed that hot spots in the upper corners can be eliminated with higher collector air flow
rates. The air flow has been increased about 20% by replacing the air flow measuring device with one
of lower resistance and increasing the fan speed. Continued monitoring will quantify the effects of this
change for its potential to improve the performance of the system.
The curve-fit of the NSTF system efficiency was used in the SIMAIR program to predict the potential
annual solar energy contribution. Assuming no solar contribution for the summer months, the solar
energy delivered is projected to be 1.4 GJ/m2, for an overall solar collection efficiency of 29%. This
value can be directly compared to the 1.3 GJ/m2 for the glazed system. Based on the results of
collector testing, the perforated-plate system has a potential performance equal to that of the glazed
collector design. The perforated-plate system would also have the energy benefits of reduced wall heat
loss and building air destratification.
CONCLUSIONS
Systems for solar preheating of ventilation air can save building energy consumption in five ways: solar
heating of ventilation air, reduced wall heat loss during the day, recovery of heat lost out walls at night,
destratification of building air and use of the warm ground-level air as opposed to rooftop air.
Extrapolations of the field-monitored results were done. The glazed system showed a potential for 1.3
GJ/m2 energy delivered by solar heating or 45% efficiency. The unglazed system value was 0.4 GJ/m2,
or 14% efficiency, although, this is in question due to irregularities at the test site, and another site is
being monitored. Extrapolation of the lab test results showed a potential solar energy delivery of 1.4
GJ/m2, or 49% efficiency, for the perforated-plate system.
Energy savings due to other features of the system vary with site conditions. For the buildings studied,
recapture of nighttime wall heat loss saved 0.2 GJ/m2 and building air destratification saved 1.8 GJ/m2.
Accurate monitored results for the savings due to reduced wall heat loss and use of warm ground level
air could not be obtained. Simple calculations suggest these savings to be approximately 0.5 to 1.0
GJ/m2 each.
Combining all the modes of energy savings shows that the total energy savings for solar preheated
ventilation air systems could be over 4.5 GJ/m2.
ACKNOWLEDGEMENTS
The authors would like to acknowledge the financial support of the Federal Panel on Energy R&D
(PERD), Energy, Mines and Resources Canada, the technical and administrative support of Doug
McClenahan of Energy, Mines and Resources Canada and the assistance of Ford and McDonnell-
Douglas in the monitoring.
-------
1543
EXPERIMENTAL STUDIES ON LARGE SIZED CONVENTIONAL SOLAR AIR
HEATER FOR AGRICULTURAL USE
Bruno Daniotti* , Vinod Kumar Sharma** and Simonetta Fumagalli***
*Politecnico di Milano, Department of Building and Territorial System
Engineering, 20100 Milan, Italy
** Visiting Scientis at the Commission for European Communities,
Ispra Establishment, 21020 Ispra (VA), Italy
***ENEA Laboratory, CCR Ispra Establishment, 21020 Ispra (VA), Italy
ABSTRACT
General objective of the investigations described in this paper is to discuss the
performance of a conventional solar air heater with an intention of utilizing solar
energy for crop drying. The experimental results have been analysed and the
concluded parameters have been included. Collector efficiency and solar
energy contribution for hay drying is determined by testing the air heater
outdoor and monitoring the output energy and the incident solar radiation.
INTRODUCTION
As a device of solar energy utilization, solar air heaters can find application in
many processes requiring low and moderate temperatures. Majority of air
heaters are used either for space heating or for solar drying applications. As an
economically v.iable solar energy device, the solar air heater is welcomed in the
agricultural industry. Simplicity of design, need for little maintainance and
possibility of using considerable cheap materials makes the solar air
heater suitable for large scale adoptionJBut despite all these factors, the use of
solar air heaters as a primary heat source in drying and other applications, is
still very low. Several factors are responsible for this situation. These include
non availability of technical know-how, poor performance of some prototypes
and non existence of demonstration programmes of the solar drying units.
-------
1544
Efforts should be made to fabricate such devices with the help of locally available
technical skill, materials and labour. Efforts should also be made for the field
trials of cheap and efficient solar devices.
Keeping in mind the fact mentioned above, a large sized conventional solar
air heater was designed and installed at Tavazzano in Italy. The purpose of this
installation was to provide preheated air for hay drying purposes. From the
experimental results it has been observed that it is possible to achieve a net
gas-oil saving of 4237 kg, for the whole season.lt is worth mentioning that the
present saving corresponds to 39.3% of the total energy needs for drying
approximately 4080 q of hay per year, from an initial moisture content of 85%
to a safe storage moisture content of 15%.
AIR HEATER DESCRIPTION
Because of the favorable environmental conditions and of the particularly
exposed nature of the ground, the project included the construction of a building
with a solar system on its south side. The heating plant is an air system with the
advantages of being quick to alterate and it making integration with a solar
system possible.
The air heater investigated (shown in Fig.l) was projected by the Solar Energy
Laboratory of the Institute of Applied Physics of Milan. It consists of a black
painted metal plate based on the southern slope of the hay loft. Over the
absorbing surface a layer of transparent fibre-glass gives the greenhouse effect.
The air within the collector is moved by a fan of suitable power situated at the end
of ducts to connect up with the air channel. The solar system is constructed with
a single panel having a surface area of around 336 m2 and is installed at an
inclination angle' of 15°.
EXPERIMENTATION
For the evaluation of the performance of the system under investigation,
outdoor calorimetric tests were performed in the summer months from May to
September, following the crop drying periods. The diurnal variation of inlet air
temperature, outlet air temperature, the convey ducts'temperature, the drying
air temperature and the instantaneous total solar radiation incident in the plane
of the collector has been plotted in Fig. 2.
Radiation intensity on the horizontal plane was measured using a Kipp and Zonen
solarimater; all temperatures were measured using thermoresistances; the mass
-------
1545
flow rate of air was determined from the air flow velocity; an axial ventilator
was used to supply air flow to the assembly.
All data was measured by a Solartron 7066 datastore voltmeter, recorded at a
regular interval of half an hour and processed on the Digital computer of the
ENEA Laboratory in the CCR Establishment of Ispra(VA).
PERFORMANCE EQUATION
The thermal performance of a solar collector is determined by passing the heat
transfer fluid through it at a steady rate under clear sunny conditions. The basic
measurements used to qualify the collector is the difference between the inlet
and outlet air temperatures. The integral efficiency is calculated using the
formula:
J Qdt
T| = (1)
J I La dt
Q = m (To -Ti) Ca (2)
where :
Q = useful power transfer by the air (W)
m = mass flow rate ( kg/s)
Ca = specific heat of air at constant pressure (kj/kg °C)
I = Incident solar radiation (W/m2)
La= Area of the absorber (m2)
To = outlet air temperature (°C)
Ti = intlet air temperature (°C)
ENERGY BALANCE RESULTS
The traditional field crop drying in the Northern regions of Italy has been
abandoned recently because of the losses in the nutritional I and quantitative
efficiency in the obtained product,It has been replaced by the artificial
drying technique, which consists of a first field drying period to get the hay
humidity from 80% to 50%, and a second period, in particular hay lofts, to
obtain the optimal storage humidity of 15% J10 do that the hay is placed on grilled
-------
1546
planes, to obtain warm air fluxes through it.
The enerav consumption of such systems are due to electrical blowers and
gas-oil burners to heat the air: it has been evaluated that the mean gas-oil
consumption to dry lq of hay is of 2.8 kg.
As the crop harvesting periods are during summer, when solar energy is most
available, the use of solar air heaters would achieve good savings. Besides,
the agricultural nature of the surroundings, the absense of shadings, optimizes
the exposition of the solar heater. The integral efficiencies, calculated as in
formulas 1 and 2, give an average value of 50.8%; good result as for the low costs
of the solar air heater.However the global seasonal efficiency is strongly
influenced by the plant managing: to optimize performance we have to try to
change some behaviour about cutting timing, drying timing, excluding night
auxiliary inputs, and eventually'using" the solar energy for other uses , such as,
heating houses. In the energy balance of the drying plant we have to consider,
besides the energy given by the solar heater, the nightly auxiliary
conventional gas-oil burner and the losses due to ducts.
From the experimentation it has been observed that during different periods of the
year, the use methods of solar1 and conventional energy showed substantial
differences. In June at night the conventional burner is on and then the
energy saving due to solar heaters is , 41 %. In July we have almost 91% of
solar energy to dry the hay, because of little daily use of conventional heaters.
However in September, we have much less available solar energy, giving only 28
% on the total.
In an other experience with solar drying in Pasturo, no conventional heater
was used. This demonstred that the whole drying is possible with solar
heaters (Daniotti, 1983). As showed in Table 1 and Fig. 3 for energy balance data,
we can conclude that it is possible to obtain good crop drying with solar air
heaters/ with no mold problems, no reduced nutritional losses and considerable
energy saving. On the basis of the performance measurement, the crop
seasonal calendar, and on solar radiation data, evaluated the
seasonal energy balance of the plant (see Table 2). There has been noted a gas-oil
saving of 4237 kg for the whole season, corresponding to 39.2% of the needs for
drying the 4080 q of hay year production.
REFERENCES
Bhargava, A. K, Garg, H. P. and Sharma, V. K. (1982).
I, 523
-------
1547
Bhargava, A. K, Garg, H. P. and Sharma, V. K. (1985). Energy. 10 . 589
Garg, H. P., Bandyopadhyay, B. and Sharma, V. K. (1981). Energy Convers.
Mprmt- . 21 .275
Daniotti, B. (1984). Physical aspects in the application of renewable energies in
agriculture. Physics graduate thesis. Milan University.
TABLE 1 Energy balance for measurement campaigns
MEASUREMENTS CAMPAIGNS
MEASUREMENTS PERIODS
(total campaign hours)
A
15.6-18.6
49
B
19.6-22.7
30
c
13.9-23.9
79.$
AVERAGE
Panel energy output
kWh
2494
3142
3346.5
2994.2
Incident solar radiation
kWh
4785.5
6251.8
6670.2
5902.5
Panel global efficiency
%
52.1
50.3
50.2
S0.8
Burner energy
kWh
2852.4
235.5
6672.7
3253.5
Energy distribution losses
km
506.8
714.8
769.2
663.6
Total drying energy
kWh
4839.6
2662.7
9250.0
5584.1
Panel energy/total energy
%
41
91
28
53.4
TABLE 2 Energy seasonal balance
MONTHLY PERIODS
MAY
JUNE
JULY
AUGUST
SEPTEMBER
TOTAL
AVERAGE
(total campaign hours)
14
9
8
9
10
49
16.3
Panel energy output
kWh
13573
8613
9007
8437.0
6971.0
46601.0
9320.2
Incident solar radiation
kWh
25904
16438
17188
16100.0
13303.0
88933.0
17786.6
Panel global efficiency
%
52
52
52
52
52
52.4
Burner energy
kWh
14817.0
9403.0
9832.0
9210.0
13958.0
57220.0
11444.0
Energy distribution losses
kWh
2755.0
1749.0
1828.0
1713.0
1596.0
9641.0
1928.2
Total drying energy
kWh
25635.0
16267.0
17011.0
19637.7
17638.6
96189.2
19237.8
Panel energy/total energy
%
42
42
42
42
28
-
39.3
Hay crop
<7
1200
720
720
720
720
4080
816
Gas-oil savings
kg
1240
787
823
771
616
4237
847.4
GLASS FIBER
WOODEN FILLETS
ABSORBING METAL PLATE
FIBER-CEMENT SLAB
TILE ROOF
Fig.1 Internal structure of the built solar air heater.
-------
1548
Temperature (°C)
65 t—
60 -
SS -
S« - A~
Solar radiation (W/rrf2)
Fig.2.a. Plant temperatures and solar radiation for june campaign.
Temoerature (°C)
SS -
60 -
ss -
sa -
0 S 10 IS 20 25 30 3S 40 4S SB SS 60 65 70 7S 80 p
Tim© (h) Solar radiation (W/m )
Fig.2.b. Plant temperatures and solar radiation for sept, cairpaign.
Leaenda
1 Convey ducts temperature
2 External temperature
3 Drying air temperature
4 Panel output temperature
5 Incident solar radiation
Fig.2 Plant significant temperatures and solar radiation.
kWh
14000 t
12574
12000
10000
H Incident solar radiation
[1 Panel energy output
0 Burner energy
B3 Energy distribution losses
B Total drying energy
June July September
Fig. 3. Energy balance for solar drying measurement campaigns.
-------
1549
Technical Inspections of a Large Number of
Commercial Solar Water Heating Systems on Florida Schools
M. Yarosh, J. Huggins, T. Tiedemann
Florida Solar Energy Center
300 SR 401, Cape Canaveral, FL 32920
ABSTRACT
A large number of commercial size solar water heating systems installed on
Florida schools were inspected by the Florida Solar Energy Center (FSEC). Each
system's design, installation, operation, maintenance and performance was
examined. Many of the systems were repaired and restored to good working order
by FSEC. Numerous drawings and specifications for new solar systems were also
reviewed. The research program enabled FSEC to identify both technical and
organizational problems and shortcomings which significantly impact the
successful application of solar water heating systems in Florida.
BACKGROUND INFORMATION
Florida has 67 counties, each a separate school district with independent
authority to include solar systems on schools if they so desire. In the latter
1970s a few Florida schools began to install solar water heating (SWH) systems
to provide hot water to meet specific school needs. In 1982 new legislation
called for all new schools in Florida which would use over 1,000 gallons per day
of hot water, to investigate the use of a SWH system. If consideration of a
solar system revealed that an installation would be the most cost effective
method, then it should be installed. Spurred by this legislation the
installation of solar systems on schools accelerated.
The Florida Solar Energy Center (FSEC), in 1983, began a small voluntary program
of inspecting existing SWH systems on schools. In 1983 and 1984 the Center,
under sponsorship by the Governor's Energy Office (GEO), prepared a design
guideline on school SWH systems and conducted a series of workshops for
architects and engineers and others involved in designing large solar systems.
In 1985 under a contract with the State Department of Education (DoE) , FSEC was
requested to review drawings and specifications for new SWH systems and also to
initiate the inspection and testing of all new systems. Unfortunately, the
review process was not in the direct chain of requirements for new school drawing
reviews and FSEC had no authority to require compliance with any of its
recommendations. There were also no accepted standards for large system design
to which FSEC could refer.
In 1987 FSEC, under a GEO contract, began a survey to identify the location and
condition of all SWH systems installed on Florida schools. All 67 county school
districts were contacted with survey forms and requested to identify all existing
SWH systems and reveal any firm plans for new systems. Ultimately, about 74
systems (either existing, under construction, or firmly planned) were identified
in 21 counties. As of this writing we are aware of approximately 80 systems.
-------
1550
Individuals in each county identified as most knowledgeable and familiar with
their solar systems were asked their views on the condition of their solar
systems. Was the system operating in a satisfactory manner or not? Responses
were received on 48 systems and over 65% of the responses indicated that the
solar systems were operating satisfactorily. Subsequent on-site inspections of
these systems by a technically qualified team from FSEC revealed that only 15%
of the systems within the group were, in fact, working properly. For a variety
of reasons, 50% of the systems inspected were not working at all. We concluded
that many of those presumably knowledgeable concerning solar systems installed
in the schools are not cognizant of the true condition of their SWH systems.
SYSTEM DESIGN REVIEWS
FSEC was requested by the State DoE to review drawings and specifications for
solar systems submitted by system designers. These drawings and specifications
were usually prepared by the architect/engineer responsible for the design of a
new school. Occasionally they were provided by the mechanical contractor or
contracting plumber for the new school or school addition.
The overwhelming majority of the designs we reviewed contained wholly inadequate
information to assure the installation of an acceptable SWH system. Often no,
and usually inadequate, specifications were provided and the drawings provided
depended extensively on having a very knowledgeable installer. Drawings usually
identified the collector to be used, the number of collectors, the tank size, and
the type of controller, but little else. Piping drawings were generally
schematic in nature and the design was not carried beyond the conceptual stage
to a set of working drawings. Less than 10% of the drawings and specifications
reviewed had sufficient information to produce an acceptable working system.
Only one system we reviewed identified the magnitude of the load to be met.
Such inadequate drawings and specifications require the services of a very
knowledgeable installer. However, very few of the designs or installations on
Florida schools were carried out by experienced designers or installers or by
those associated with the solar industry within Florida.
SITE VISITS TO INSTALLED SYSTEMS
Site inspections have now been conducted at almost sixty installations. Usually
the inspections were conducted by a team of three engineers experienced in large
system design and operation. Inspections normally required several hours
depending on the specific installation. The team inspected the general layout
of the system, the piping, system control, system instrumentation, freeze
protection, and design and installation acceptability. If the system designer
was known, he was invited to attend the system inspection.
In most cases the inspection of system operations, installation, and design is
a subjective process. There existja large number of gray areas where independent
judgement on the adequacy of design and installation practices were called for.
The system design and quality of installation was judged on a rating scale of
poor to excellent. We looked at location, type and selection of component,
access to equipment, operability, system function, maintenance, and other
factors. For some of the new systems inspected, we had reviewed drawings and
specifications, but for most systems, the inspection team had seen neither
drawings nor specifications. In such cases it was difficult to know whether a
problem had its origin in poor design or in errors of installation.
In many cases the inspection team identified simple problems, such as incorrectly
located sensors, and remedied the problem by relocating tfhe sensors during the
-------
1551
inspection. About 25% of the systems which were not operating or operating
poorly during the inspection had problems that could be corrected by the
inspection team to significantly improve or restore operation. About 40% of the
systems inspected were in operation after the completion of the first inspection.
For inspection purposes the SWH systems were divided into the following
categories:
o the collector array and piping (including all external piping).
o the control subsystem (includes all sensors and controllers).
o internal piping and storage (tanks and backup systems).
o valves and instrumentation,
o freeze protection,
o operation and maintenance.
Our primary findings as they relate to each of these categories are given below.
Collector Array and External Piping
Most collectors we saw were in good condition and had been certified by the
Florida Solar Energy Center. Almost all of the systems had collectors mounted
properly and correctly oriented to the sun with appropriate tilt angles.
Mounting hardware, in general, was in good condition and sturdy.
Many arrays we inspected did not have the flow uniformly balanced among the
collectors. We found many balancing valves, but also found that they were seldom
used to balance flow through the array. The simpler system of utilizing reverse
return piping eliminates the need for manual flow balancing and is far more
likely to result in balanced flow.
We found piping insulation deficiencies in 47% of the systems. The majority of
these deficiencies were the result of ultraviolet degradation of inappropriately
protected insulation. The balance of the insulation problems we found were in
the equipment room and at locations where valves, instruments, or other devices
were attached to the system.
In 20% of all systems, either the collectors or the systems would not adequately
drain. Usually this was because of improper piping configuration or improper
canting of the collectors.
Control Subsystem
All but one of the systems inspected operated with a differential temperature
controller receiving signals from sensors located to determine when collector
temperature is above the coldest water temperature in the solar storage tank.
When heat is available the controller turns on the system pump(s).
Of all subsystem problems, those associated with the control subsystem proved
both the most destructive in terms of system performance and generally the
easiest to correct. They certainly were the most commonly encountered. The
control subsystem fails to function properly if:
1) any of the sensors are improperly located.
2) the sensors, sensor wiring or wiring connections are faulty.
3) multiple sensors are improperly wired together to give an erroneous signal
to the controller.
4) the controller function itself is faulty.
We found that:
o seventy percent of the systems had a significant problem with the control
-------
1552
system. The problems usually had their origin in design or installation,
o seventy percent of the controllers tested were satisfactory. The balance
were either older and somewhat abused controllers or were in systems which
obviously had received little maintenance attention,
o either the collector sensor or the tank sensor was incorrectly located on
2/3 of the systems visited,
o thirty percent of all control systems had improperly wired sensors.
Storage and Internal Piping
This subsystem includes all piping within the equipment room as well as the solar
and backup storage tanks.
We saw four different techniques for connecting the solar system with the backup
system. It is essential that water heated by the backup system cannot enter the
solar system. This is best achieved by a separate storage tank for each of the
two systems. We found that 25% of the systems we inspected did not have separate
tanks. This is an extremely serious error.
In general, we recommend sizing the solar storage tank to provide approximately
one to two gallons of storage capacity per square foot of collector area. Almost
70% of the solar storage tanks we inspected were within this range. Only 10% of
the solar tanks inspected were seriously missized (most commonly oversized) . An
oversized tank can result in reduced water temperature which requires additional
backup heating. Most solar storage tanks had adequate insulation.
A common and serious error which we encountered was in the distance between the
solar storage tank and the load. Commonly, solar collectors were installed on
the roofs of schools, and generally the storage tanks were located in the equip-
ment rooms on the main school floor. In these cases the distance from collector
array to storage tank was short, often under 100 feet. In some schools, the
designer located the collector array and solar storage tank on the ground
adjacent to the school building. In some of these cases the collector array and
solar tank were several hundred feet from the equipment room to which the solar
heated water must return. In such cases, the intermediate piping between the
solar storage tank and the equipment room is a source of significant heat loss.
Valves and Instrumentation
The correct location of all valves (including pressure and/or temperature-relief
valves, check valves, bypass valves, automatic air vents, air bleed valves, and
vacuum breakers) is essential to proper and safe system operation, and in some
cases, to permit operation at all.
The most common problem we found with pressure-temperature relief valves was the
lack of piping on the discharge of the valve. This could result in serious
injury to personnel near the collector array.
Check valves were missing or improperly installed in fifty percent of the systems
we visited. This created problems ranging from night thermosyphoning to complete
blockage of flow to the collectors.
Some systems had manual air bleed valves. These were often located on a roof
where they were seldom checked. We favor the use of automatic air vents.
Solar systems should be installed with a bypass valve to allow water to bypass
the solar portion of the system during solar system repairs. Without this valve
(or if the valve is left open) cold city water bypasses the solar storage tank
-------
1553
and goes directly to the backup system. Thirty percent of the systems we
inspected had no bypass valve installed in the bypass line. We also found six
systems where the valve had been installed, but left open.
Instrumentation on a solar system informs operators when correct system operation
is occurring and enables operation and maintenance personnel to identify and
diagnose problems when they occur. Thirty percent of the systems inspected had
no instrumentation at all. Only 10% of the systems had acceptable minimum
instrumentation for system operation. Minimum instrumentation includes: a flow
meter for the collector loop, temperature measurements to and from the array, and
an indicator light to show when the circulating pump is operating. Of all of the
instrumentation we saw on systems, only one half of it was functional.
Freeze Protection
Although Florida has a very mild climate, freezing conditions do occur within the
state. This can and has caused serious damage to inadequately protected solar
systems. Twenty percent of all systems inspected showed some freeze damage.
Because of Florida's mild climate and the infrequency of freezing, certain freeze
protection methods employed in Florida are not common elsewhere.
The four common methods of freeze protection employed on Florida schools are
recirculation, draindown, drainback and antifreeze systems.
The method most commonly used in Florida is recirculation freeze protection.
This method utilizes a sensor which detects when the collector temperature
approaches freezing. The controller then operates the solar loop pump to
circulate warm water from the solar storage tank through the collectors to
prevent their freezing. Although heat is lost during this period, if freezing
conditions occur only rarely, the system simplicity makes its use attractive.
Recirculation freeze protection systems failed for two principal reasons:
improper location of the freeze sensor on the collector array and failure to
activate the freeze protection switches in the controller when it was installed.
Since recirculation systems depend on pump circulation for success, they do not
work in the event of a power outage during freezing conditions. To handle this
unlikely, but possible, event special valves called "freeze" or "dribble" valves
are installed in the system. These valves open at temperatures of 35°F - 45°F
and allow city water to flow through the collectors and out the freeze valves.
City water is generally above 60°F. The sensing element in freeze valves is
sensitive to high temperatures, and we believe a common cause of freeze valve
failures may be exposure to excessively high temperatures.
The number of systems we inspected with freeze protection methods other than
recirculation was too small to draw conclusions about their effectiveness in
freeze prevention.
Operation and Maintenance
For solar systems to be successful they must be operated properly and well
maintained. A number of O&M problems are influenced by features which are
considered and included in good design practice. For example, in our inspections
the following items in particular were notable by their frequent absence,
o an established operator training program.
o an adequate set of operating and maintenance instructions.
o an experienced and solar trained maintenance department,
o easy access to all the system, including the collector array.
-------
1554
o an appropriate set of instrumentation.
With the absence of an operation and maintenance infrastructure for solar
systems, it is not surprising that only one system that we inspected had an
adequate operation and maintenance manual. Ninety percent of the systems had no
information provided at all. We found little or no evidence of training of
system operators. Often no individual operator had been designated at the school
for solar system operation. We found that in almost all systems visited,
additional specific training would have been useful for maintenance personnel.
STATUS
In the 1970s schools and school districts in Florida moved aggressively into
solar water heating applications. The schools were cognizant of the need to save
energy and money, and in Florida schools shared the common indigenous interest
in solar energy.
Unfortunately, few of the architects and engineers who design schools, and few
of the contractors who build schools had adequate knowledge, background,
experience, or understanding of SWH systems. As a result, many systems that were
poorly designed and poorly installed found their way onto Florida schools. These
problems were compounded by a lack of operator training and operator training
programs, a lack of well trained solar maintenance personnel, and an almost
complete lack of information on operation and maintenance.
FSEC entered into this process largely in an advisory capacity and lacked any
enforcement authority. Often school construction schedules and occupancy
certifications were much more influential in dictating decisions than the need
to rework a poor SWH design or installation and hold up a school's certification.
Progress was made, however, in influencing some system designs and many systems
have been returned to proper operation through an ongoing program of system
repair. At present: .
1. between 35 and 40 SWH systems are now operating satisfactorily on Florida
schools.
2. about ten additional systems are currently targeted for repair.
3. about 20 systems (mostly in Dade County) have not yet been inspected.
4. perhaps a half dozen systems are not working or are working poorly, but
would require substantial investments to correct major design errors. The
cost of such corrections and the availability of funds will determine
whether any of these systems will be repaired.
5. a few older systems (perhaps 3 or 4) are not worth restoring to operation.
6. a very few SWH systems are still under construction.
LESSONS LEARNED AND RECOMMENDATIONS
The following reflects only the more important lessons and recommendations from
this study. A much more complete set of findings and recommendations are
available in the "Survey Report"1.
o System drawings and specifications must provide complete information on
system design and installation. The system hot water load must be known
and the solar system installation must not be included as an afterthought.
A&Es must follow SWH system installation.
lnA Survey of Solar Water Heating Systems on Florida Schools, Findings,
Conclusions, Recommendations", FSEC-CR-314-90, M. Yarosh, J. Huggins, T.
Tiedemann, 1990.
-------
1555
o Until a cadre of A&Es experienced in SWH design are available to Florida
school districts, FSEC should be requested to review all drawings and
specifications for school systems and it must have compliance authority,
o Until a cadre of installers experienced in installation of commercial size
solar systems are available, FSEC should continue to perform final
inspections and tests on school systems and have compliance authority,
o SWH system designers must provide operation and maintenance manuals,
o The State DoE and the school districts must develop a program for training
designated operators and designated maintenance personnel.
From the review of numerous drawings, designs and specifications and the
inspection of many systems, we have developed a few important recommendations and
observations.
o Adequate instrumentation is absolutely essential to the successful
operation of an SWH system,
o Experience suggests that designs with "reverse return" piping on collector
arrays usually produce good flow distribution in the array. Most systems
which required valve settings for balanced flow had not undergone flow
balancing operations.
o The placement of valves such as check valves, air bleed valves, relief
valves, vacuum breakers, and bypass valves should be specified by the
design drawings and specifications and installed by a knowledgeable
installer using good installation practice,
o Access to the collector array and to all valves which require manual
attention must be easy. Otherwise systems get ignored. The system
designer must consider this,
o We recommend completely separate storage tanks for solar and backup
systems. This will avoid preheating of solar water by fossil fuels,
o The collector array should be close to the solar storage tank which in
turn should be close to the system load. This will avoid the unintended
intermediate heat exchange occurring in long piping runs.
o The control system is the source of more problems affecting system opera-
tion than any other subsystem. Care and knowledge is required in proper
controller selection and installation/setup, in proper sensor placement,
and in the correct wiring of the control system. The installer of the
control system must understand both the control system and how the solar
system is supposed to work. The sensors should be checked annually using
procedures given in an O&M manual,
o All solar system collectors and piping must be drainable.
o Exterior piping insulation on collector arrays should be covered or
jacketed to prevent UV deterioration. Paints which we saw did not last
more than a year or so and seemed to be unsatisfactory,
o We believe recirculation freeze protection is overused in Florida. We
recommend it only for south Florida. When used, great care must be taken
to place freeze valves in locations that allow flow to move through all
collectors before exiting at the freeze valve,
o Two-year warranties should be required of the system designer/installer on
commercial SWH systems.
If commercial sized SWH systems are to become commonly used, the solar industry
must first demonstrate that these systems, when correctly designed and installed,
are highly reliable. System designs and installations need careful attention
both during the design and the installation. Finally, a well developed program
for training system operators and maintenance personnel is essential to continued
success of solar water heating systems.
-------
1556
PERFORMANCE OF SOLAR WATER HEATERS IN FIVE
HOSPITALS IN MALAWI
B.W. ZINGANO, H.M. KAN J ERE AND D, K.ADUYA,
MINISTRY OF WORKS - MALAWI.
ABSTRACT
The paper reviews insitu performance of sandwich solar water heaters
installed to supply hot water to patients in five hospitals in Malawi. The
design philosophy of the solar heaters is presented and the particular
choice of renewable energy is justified by demographic and energy demand
figures for a developing country. The particular absorber metal choice of
each hospital installation is shown to have been decided upon from an
analysis of onsite water chemistry to avoid pit corrosion and scale
formation.
In the analysis that follows, Laboratory Instantaneous Efficiency curves are
shown followed by insitu performance "Average Day Efficiency" test curves.
The Efficiency tests are shown for both half and full days to match typical
peak hours for hot water use in the hospitals. The results obtained are
used to redesign future installations in other hospitals and health centres
which are still on the drawing boards.
INTRODUCTION
Although thoughts to utilize solar energy to heat water for domestic
purposes started in 1953 in the Ministry, serious practical attempts were
made in 1978. Hitherto all government hospitals had no hot water available
to bathrooms and showers. Generally the lowest air temperatures in Malawi
rarely drop down 5 although zero temperature sometimes occur in high
altitude areas. However, even during the cold seasons,hot water was not
available to patients for baths. This was a deliberate policy because the
traditional sources of energy were too expensive to be borne on recurrent
budgets. Table 1 shows the energy consumption and sources in Malawi for
seven years.
In 1978 the Ministry of Works decided to try to design and install solar
water heaters on two district hospitals Mangocni and Nsanje, which were
being funded by the European Economic Commission under an aid grant to
Malawi. The first set of solar water, heaters were made out of galvanised
steel and due to corrosion problems'1'; these were removed in 1983 and
replaced by an improved and more efficient set of solar water heaters. This
set was made locally and the Ministry had a chance to test this model in a
in laboratory. Installed systems were thermosypnonic and the collectors
were sandwich type out of copper painted with normal black matt acrylic-
paint and covered with 3mm crown glass of 1.16% ferrous content.
-------
1557
In subsequent years the same model was installed in three other hospitals.
These solar water heaters are being tested in situ every two years to
monitor their performance.
This paper reports on the experiences with the original galvanised grid type
and the replacement model of solar of solar collectors.
IN SITU TESTING PROCEDURE
A simple method for in situ testing was adopted in water to get quick
results. The hot water storage tanks were installed in the roof space for
both types of installations. In the former model the collectors were
installed below the caves. Whereas in the latter model the collectors were
on the roof. The transmission pipes were not lagged. The former model had
the disadvantage of long transmission pipes whereas the latter was compact.
Instantaneous Efficiency tests using the standard method were done on both
models and the results are shown on table 2. The in situ tests were aimed
at finding the average efficiency for the day at no draw off conditions.
The equation modelled was the following:-
Q„ =F (ax)-
V
FUlj VTa)+ UA jT,fTa ]+ U?V (VTa
__
V
ACI
(1)
Where Qu is the useful heat gained
Ac is aperture of collector
F is the drain factor.
a is the absorptivity of the collector surface.
t is the transmissivity of glass.
I is the irradiance.
UL is the average thermal transmittance of collectors.
T# is the average fluid temperature in the collector.
T is the ambient temperature.
Ut and U
are the transmittance values of collector, storage
tank, and pipes respectively.
Tic, T|t, and T are the mean temperatures of fluid, in collector,
storage tank, and pipes.
At and A are the surface areas of the storage tank and the
transmission pipes respectively.
Prior to each test all the water in the storage and collector tanks was
drained and the mains inlet and draw off to the storage tank were closed.
This was done after sunset in the evening prior to
the test. The initial water temperature was taken on that evening and at
dawn the following morning.
At dawn a Kipp Zonen pyranometer and integrator were set up. The collector
glass was thoroughly cleaned and the sensor mounted at the same angle to
horizontal as the collector. From 0600 hours; the ambient temperature,
-------
1558
cloud cover, and the insolation were taken at intervals of 30 minutes.
After sunset the hot water was drawn off in quantities of two litres at a
time and the average temperature noted. This was done until both the
storage tank and the collector were empty. Typical results of a full day
and a half day results are shown on tables 3.
ON SITE OBSERVATIONS AND PRACTICAL PROBLEMS
When the first solar water heaters were installed in the hospitals, a few
showers were not connected to the hot water supply from the solar water
heaters. Although the patients were not told which showers had hot water,
after a few months the showers without hot water were hardly used. In the
course of time the patients realised that hotter water was available in the
showers from about noon onwards. In a separate test ' ' the maximum hot water
in the storage tank was observed to be reached at two hours after the solar
noon.
When the first set of collectors had corroded after a period between 24 -36
months,each collector was removed and thoroughly examined. The source of
water at the Mangochi hospital was from a fresh water lake while that at
Nsanje was. from a borehole. The water from the latter source was hard and,
on analysis, showed high content of sodium sulphates and carbohydrates and
carbonates. The water sample from the former source had lower levels of
salts but high chlorine content from the water treatment works. These
collectors had been sealed using arch welding and thereby removing the zinc
layer exposing the steel to contact with water. The corroded samples from
Nsanje hospital where the rate of corrosion was fastest, scale formation
occured to both the collector channels and pipes. In the collector samples
from the Mangochi hospital where the rate of corrosion was the slowest,
there was little scale .formation but quite a wide spread pit corrosion.
These differences could have been due to the water chemistry.
The replacement solar water collectors were designed to withstand the water
quality. Copper plates were used at the Mangochi hospital while stainless
steel plates were used at the Nsanje hospital. The plates were not welded
but fixed together using rubber gaskets and screws on the edges, and rubber
block separators between the plates were used to create the water channels.
Since this set was installed in!985.todate there has not been any problems,
and the efficiencies shown on table 3. relate to this model.
It was observed that the hospital administration usually sent a cleaner once
in a while to clean the first set of solar water heaters. However, the same
hospital administration was unable to send a cleaner to clean the
replacement collectors. This change of attitude occurred because the first
set was mounted within reach of human height (at eaves height) while the
second set was mounted nearer to the roof ridge. The dirt on the collector
meant that the insolation on to the absorber surface was reduced.
In table 3 it will be observed that the average effiency is higher when the
draw off is done halfway through the day than when it is done after sunset.
This can be explained by increasing heat losses from the collector surfaces,
transmission pipes and storage tank according to equation (1). From the
same table it is very clear that the same collector when cleaned,performed
better than when it was dirty by very considerable margin.
-------
1559
CONCLUSION
In this programme, a lot of practical information has been gained. The next
set of installations will combine the features of both sets. Mounting will
be at human height for easy cleaning to maximise the efficiency,and water
chemistry will always be matched to the absorber pi ate.This approach will be
used in the current programme to install rural health centres with solar
water heaters. The source of most of these rural health centres are
boreholes and the water chemistry, in every borehole, will be analysed.
Solar water heaters are a feasible and a cheap form of heating bath water in
hospitals in Malawi. Although it is difficult to make a financial
comparison, it is obvious that patients'prefer hot baths and hot showers to
cold ones. However, the patients washing times will have to be in the
afternoons if hot water is preferred and, in the initial observations,
patients learned this discipline quickly.
So far the public acceptance of solar water heaters in Malawi has been
overwhelming. The supplier of the second model of solar water heaters can
now not meet the orders. Other institutions and individual persons have now
got this model installed on buildings and houses since 1984.
REFERENCES
Zingano.; B.W. "Types of Plate Collectors and Their
Performance as related to the Corrosion Problem" Renewable
Energy Development in Africa Volume II 1985. Commonwealth
Science Council.
Zingano.; B.W. "Efficiency of Solar Collector System Using
an Indirect Tank". Paper Presented at the Annual Conference
for the Association for the Advancement of Science of malawi
1981.
-------
1560
Table 2.
Laboratory Tests on Instanteneous Efficiency of Models 1 and 2
Collector Model Maximum
Efficiency
MJL
at (T« - Ta) - 0 w»-2 des "i
54
Stagnation
Temperature at
zero - efficiency
aperture
construction uetan
'Jb) °C
I Ac J m'
86
1.53
Sandwich construction using
0.5 mm copper or steel leaving
a 5 mm water channel and an 8
mm gap between plates and
giasr. (not »uch difference
between copper and steel}.
Sandwich galvanized corrugated
sheets using 0.5 m thickness.
Sides welded and 15 mm
airgap.
TftBLE 1 ENERGY SOURCES AND CONSUPTI0N IN MALAWI OTHER THAN SOLAR ENERGY
(per capita worKed out from iv/n -iv%8, ail figures in "Billion Hega jomes)
YEAR
1978
Mn MJ
1979
Mn MJ
1980
Mn MJ
1981
Mn MJ
1982
Mn MJ
1983
Mn MJ
i984
Mn MJ
1985
Mn MJ
1986
Mn MJ
1987
Mn MJ
1988
Mn MJ
END USE
EFFICIENCY
FACTOR
SOURCE
H-ELECTRICITY
882
iOiO
1083
1051
789
1205
1242
1224
1307
1428
1441
0.85 *
DIESEL
803
850
891
754
666
751
717
811
778
744
763
0.25
PETROL
380
394
366
372
343
333
325
319
316
327
343
0.20
PARAFFIN
123
103
80
64
59
55
42
58
76
81
82
0.30
COAL
899
549
1059
775
991
710 "
671
572
676
751
768
0.65
MGODFUEL
-
-
10764
H005
ili45
11526
12145
12240
i2574
13098
13813
0.11 *
ETHANOL
-
-
-
-
51
55
60
68
80
78
76
0.30
TOTALS
-
-
-
14054
14635
15202
15929
15807
16507
17286
Ave. 15540 Mn MJ/'pa.
*An average value has been taken for two or more uses,
(all figures are after converting by end use efficiency factor)
-------
1561
Table 3
LOCATION
DATE
TEST TYPE
—
Ac
I
Til
vn
C'i
Ave. tff
RtflARKS
8LANTYRE
2.09.88
UHOjd
2.82
19.42
20.09
180
168
<0.125
57.19
Almost clear day
6LANTYRE
23.09.88
UHOjc
2.82
19.44
23.3
180
162
<0.i25
78.61
Almost clear day
HANGOCHI
01.10.88
ISi(FD)d
1.6?
21.98
Si.i
180
154
<0.125
69.91
Clear day
HANGOCHI
02.i0.88
IS*(FD)c
1.69
20.86
31.5
i80
160
<0.i25
77.86
Clear day
HSAHJt
22.il.88
IS4tF0)ti
1.63
24.04
29.77
i80
70
<0.125
24.3%
Part ciear day
fiSANJk
24.li.BB
ISJ(FD;c
i.63
18.34
29.60
180
122
<0.125
54.4
Cloudy day not overcast
HCHINJI
17.01.91
ISi-(FD)
1.98
15.02
25.3
100
iOO
5.3
55.0
Partial clouds and rain
HCHINJI
iB.0i.9i
IS#(FO)
1.98
15, .02
20.7
100
100
6
44.0
Partial clouds and rain
KARuNGA
06.02.91
1S?(FD)
i. 52
20.51
27.8
250
192
6
37.3
Partial clouds and rain
KARONGA
07.02.9i
ISl(HO)
i .52
8.20
28.4
218
218
7
67.6
Partial clouds and rain
IS(FD) : Field Full Day Test vn ; normal manufacturers volume
F(HD) ^ Field Half Day Test V ^ volume collected.
L(FD) = laboratory Test Cm - mean cloud cover in octas.
d = test done when collector was dirty,
c - test done when collector was cleaned the following day.
-------
-------
2.12 Active Heating III
eceding page blank
-------
-------
1565
EXPLAINING SELECTIVE SURFACE DEGRADATION
EFFECTS ON SOLAR HEATING PERFORMANCE
K.G.T. Hollands A. Karagiozis A.P. Brunger
Solar Thermal Research Laboratory, Dept. of Mechanical Engineering,
University of Waterloo, Waterloo, Ontario, CANADA, N2L 3G1
G. Brouwer
Van Heugten Consulting Engineers, Solar Energy Department,
Nymegen, The Netherlands
ABSTRACT
Selective surfaces often degrade in the field. Their solar absorptivity as and thermal emittance
e change with time in service by some amount, say Aas and Ae, from their starting values.
In order to project their service life, it is important to quantify the effect this degradation
has on the annual fraction solar Fs. A given relative change in Fs can be caused by different
combinations of Aas, and Ae. Recently Hollands et al. (1990) used computer simulation of
solar domestic hot water systems to graph these combinations, in a plot of Aas versus Ae, for
relative changes in Fs, of 10% and 5%. The slope and intercepts of this plot, which was found
to be linear, were studied for their dependence on a wide range of solar system parameters,
such as geographical location, collector area, and set point temperature. They found that
the slope, and—for starting values of Fs less than about 0.5—the intercepts, are relatively
insensitive to the system parameters. The present paper shows that this result is consistent
with a simple model. The paper also explains why, for Fs > 0.5, the intercepts rise sharply
with Fs, in a way that is strongly (and to some extent, only) dependent on the geographical
latitude of location.
KEYWORDS
Selective surface; Absorptivity; Emittance; Solar collectors; DHW system; Durability.
INTRODUCTION
Selective surfaces often degrade in the field; their solar absorptivity and thermal emittance
e change with time in service, say by amounts Aas and Ae from their starting values as0 and
e0. In order to predict their service life, it is necessary to quantify the effect this degradation
has on the system's annual fraction solar, Fs. Thus, if we (arbitrarily) define the "service
life" as the time required for the system performance to decrease by some arbitrary amount,
say 10% or 5%, then finding the service life reduces to a two step process: first determining
how the corresponding changes affect Fs and in particular how much change in as and e is
required to produce the arbitrarily-chosen permissible relative degradation in fraction solar
A Fs/Fs0.
A specified decrease in Fs can be caused by different combinations of Aas and Ae, and it is
of interest to map these combinations, in say a plot of Aas versus Ae. In such a plot, each
point represents a combination of Aas and Ae that will produce the specified decrease in Fs.
Recently Hollands et al. (1990) used computer simulations to generate such plots. The solar
system they chose was a solar domestic hot water heating system of fairly conventional design,
using a flat plate solar collector, also of conventional design. They examined the effect on
the plot of many system parameters: geographical location, set point temperature, collector
area, collector flow rate, water draw-off pattern, and other parameters. They found that the
plots were essentially linear, and that their slope was relatively constant independent of the
Preceding page blank
-------
1566
parameter setting. On the other hand, the vertical intercept (with Aas on the vertical axis)
varied considerably.
The purpose of this paper is to interpret these plots from a physical point of view. This will
permit a deeper insight into the way in which as and e effect system performance, and it will
also permit the results to be generalized, to some extent. With this interpretation, future
studies assuming different solar systems and different collectors can more quickly arrive at
the key effects.
The simulations of Hollands et al. were performed using the computer simulation code called
"WATSUN" to simulate a standard solar water heating system. Like other similar codes,
WATSUN marches through the simulated year in hourly time-steps of simulated time. At
each step it reads from storage the mean weather data (solar irradiance on a horizontal
surface and ambient temperature), and using modeling algorithms, calculates the mean solar
irradiance on the collector for the hour, calculates the collector output for the hour, revises
the status of the storage tank, and finally calculates the solar energy delivered to the load.
Summed over the year, the latter quantity represents the annual solar energy delivered to the
load. Dividing this value by the yearly energy demand gives the fraction solar Fs.
Collector characteristics FR(ra)e and FrUl, required for input into the WATSUN code, were
determined as a function of as and e using a separate simulation code assuming a repre-
sentative value for each of the many collector properties. The assumed collector was flat
plate, single (low iron glass) cover, with a fin-and-tube absorber plate. The program used
the algorithms described by Hollands and Wright (1983) to determine Utop and (ra)e, and it
assumed back and side loss of 1 W/m2K, which was added to Utop to obtain U^. The plate
efficiency factor F' was then calculated as a function of t, using standard equations. These
values were then substituted into the standard equation for heat removal factor Fr, yielding,
finally, FrUl and F/j(ra)e as functions of as and e. The system contained a stratified storage
tank, and a heat exchanger of effectiveness ex — 0.8 was used between the tank water and the
collector fluid. The collector flow rate (per unit of collector area) was assumed to be either
0.0022 kg/sm2 (low flow) or 0.025 kg/sm2 (high flow), in terms of water equivalent heat
capacity. Fig. 1 shows a typical graph obtained for Fs versus as0 and e, and Fig. 2 shows
some typical plots of Aas versus Ae, obtained from plots like Fig. 1, using aso = 0.95 and
to = 0.1, and based on a relative decrease in Fs of AFSJFs0 = 0.1. ( The subscript 0 on any
quantity indicates its starting value.) Thus a point in the curve represents a combination of
Aas and Ae, that will give a 10% decrease in system performance. Both curves were found
to be very close to linear, so that, to generalize, it was assumed that
where intercept a and slope /? are parameters which will depend on the geographical location
and other system parameters.
Plots like Fig. 2, 3 and 4 were prepared to show the effect of various circumstances: on
how the plot depends on geographical location, on collector area Ac, on thermostat set point
temperature Tsp, on the (arbitrarily-chosen) ratio AFa/Fs0, and the collector flow rate. The
resulting plots all showed a nearly constant value for the slope j3 of about 0.27, but the
intercept varied — from about 0.05 to about 0.35.
INTERPRETATION
A physical interpretation of the analysis of the plots will help in understanding the various
trends shown in the Figures. A simplified analysis (Tabor et al. (1961), Edwards et al. (1961),
Hollands (1963), Edwards (1977), Collier (1979)) has shown that constant collector output is
achieved if variations de and das in e and as respectively are related such that
REVIEW OF PREVIOUS FINDINGS
Aas = a — /?Ae
(1)
(2)
-------
1567
where f3 is given (roughly) by
/3 = a(T?-T?)/rG (3)
in which T„ is the mean plate temperature, Tc is the cover temperature, r is the cover transmit-
tance, ana G is the solar irradiance, usually interpreted as its daily mean value. Integrating
equation (2) gives
Aas + /3Ae = a (4)
where a is a constant. Since equations (1) and (4) are equivalent, the slopes of Figures 2-4 can
be estimated analytically if one can get representative values for Tp, Tc, and G. Each of these
is, however, a highly variable quantity and not easy to evaluate exactly. Some simplification in
evaluating /3 is afforded by the following analysis. We first use a refinement on (3) suggested
by Hollands (1963):
P = AaT^Tp~Ja)B (5)
where Ta ta (Tp + Ta)/2 and B is a multiplier depending on the details of the collector (see
Hollands (1963) for details) whose value is roughly 0.72 for the design under consideration.
Second we solve the standard equality
F'R[(Ta)eG - UL{Txi - T0)] = (¦ra)eG - UL(TP - Ta) (6)
for Tp — Ta and then substitute the result into equation (9), approximating (ra)e by ra.
giving finally
ra-^A), , Fk(T«-Ta)
UL s+ tG
/3 = AaTgB
(?)
In these equations F'R is the modified heat removal factor (modified for the presence of the
heat exchanger, see Duffie and Beckman (1980 a)), and Txi is the inlet temperature to the
heat exchanger which the system model takes to be equal to the temperature Ttb at the
bottom of the preheat tank. For a low-flow stratified tank system with preheat tank capacity
greater than the nightly draw (which it normally is), the value of Tij is equal to the the mains
temperature Tm which,on a yearly average, is equal to the ambient temperature. Thus Txi — Ta
in equation (7) is roughly zero on average for "low flow systems", and therefore the first term
in the bracketed part of the right hand side should dominate the second for these systems.
We now substitute the appropriate values for the low flow system simulated by Hollands et al.
(1990), which are as follows: B = 0.72, T0 — 300K,Fr ~ 0.71, Ul = 4.7 W/m2K (its average
value over the range e = 0.1 to e = 0.6), a3 — a30 = 0.95 and calculate F'R as follows:
F'r = Fr
j _i_ AcFrUl ( 1
(?7zCp)c
G-0
(8)
with the relevant heat exchanger effectiveness, ex = 0.85, we obtain F'R = 0.94Fr = 0.667.
Now substituting in equation (7) with Tx{ — Ta~ 0, we obtain fi = 0.29. Assuming Txi — Ta =
+4C and G = 300 W/m2 gives (3 = 0.335, and assuming Tx{ ~Ta = —4C and the same value
of G gives /? = 0.251. The values of (3 obtained from the low flow simulations, are in good
agreement with the above theory, having an average of 0.27 with a range from 0.23 to 0.31.
To further test the analysis, simulations were performed in which the key term, 1 - F'r, in
equation (7) was altered, by altering the heat exchanger effectiveness ex. Simulations were
made with er — 0.4 and 0.6, and the resulting values of (3 were plotted against the prediction
of equation (7) with Txj — Ta made equal to zero. The results are shown in Fig. 5. Reasonably
good agreement is observed.
The above discussion was limited to low flow systems, but examination of equation (5) would
suggest that, since on average Tp should be about the same for high flow systems as for low
flow systems, the slope should be independent of which flow rate; is used, and the simulations
performed by Hollands et al. (not detailed here) has borne this out. It appears, then, that to
obtain the slope for a high flow system, one can convert the collector's Fr to what it would
be if the flow were changed to low flow (that is, sufficiently low that the assumption Txi = Tm
-------
1568
is valid) using standard equations (Duffie and Beckman (1980 b)), then determine F'r from
equation (8), and then use equation (7). Since none of the other system parameters, such as
collector area etc., varied to obtain Figures 2-4 enter into the evaluation of /?, it follows that
the slopes in these Figures should all be about the same, and equal to about 0.27, which they
are: the slopes in these figures average 0.26, with a range of from 0.18 to 0.31, despite drastic
changes in system parameter settings.
This simple model has therefore effectively explained the slopes observed in the figures, but
the vertical intercept a also needs explaining. If all the heat absorbed on the absorber plate is
transferred to the load, then, other things being equal, F3 will be proportional to as. It follows
that a 10% relative drop in Fs would be produced by a 10% relative drop of as, or an absolute
drop in as (for as0 = 0.95) of 0.1 x 0.95 or 0.095. Since "other things being equal" implies
At = 0, the vertical intercept a should by these arguments be 0.095 for Fs/Fs0 = 0.9 and
0.0475 for Fs/Fs0 = 0.95, or, in general, a = as0AFs/Fs where AFs = Fs0 — Fs. Inspection of
Figures 2-4 shows that indeed these values of 0.095 to 0.0475 form a lower bound of the value
of a, produced by the simulations. The requisite condition of the above discussion — that
"all heat absorbed at the plate is transferred to the load"—should be approached most closely
at low fraction solar since the fluid is being heated to a low temperature above ambient, so
thermal losses should be small, and also no heat will be "dumped". As the fraction solar
approaches 100%, however, the proportional effect of Fs on as, can no longer hold: in an
extreme example, if F, = 0.99, a 10% relative increase in as can clearly not produce a 10%
relative increase in Fs, since Fs cannot exceed unity. (What happens in practice is that tank
maximum temperature Tmax is already exceeded or equalled very often in systems with very
high fraction solar, so the extra energy made available by having a higher value of as is
not utilized, but is effectively "dumped".) If Fs is less than proportional to as, it follows
mathematically that the change, a, in as required to produce the specified change in Fs
must exceed asoAFs/Fs0 (the proportionate value determined above for the low fraction solar
case). It follows that a should be an increasing function of Fs0. In Figure 6, we plot the
ratio a/(as0AFs/Fsa) versus Fso, the plotted points covering all the simulations included in
the study of Hollands et al. As expected, at low fraction solar, the ratio approaches unity
(its value when Fs is proportional to as), and increases with F,o. However, the ratio is also
seen to be different for different cities. One can explain the dependence as follows. If one
month of the year had no solar radiation, the maximum value for Fs would (for the presently
considered storage capacities) be, 11/12 rather than unity. Thus the effect of dumping (in
the other months) would be felt at lower values of Fso¦ For extreme northern locations, the
available solar radiation in the months near the winter solstice is very small, and thus, their
effective upper bound for Fs is also less than unity, so that for a given value of Fso, they would
have higher values for the ratio a/(as0AFs/F3). This is indeed what is observed in Figure 6,
in which the latitude of the individual cities has been indicated by a different symbol.
ACKNOWLEDGEMENTS
This work was carried out as part of the activities of Task X Materials of the International
Energy Agency (IEA) and supported by the Department of Energy, Mines and Resources
Canada. We also wish to thank our colleagues in the International Energy Agency Solar
Heating and Cooling Programme Task X, for their encouragement and interest.
REFERENCES
Collier, R.K., (1979), "A Simplified Technique for Comparing the Effectiveness of Solar Ab-
sorber Coatings", Solar Energy, Vol. 23, No. 5, pp. 455-458.
Duffie, J.A. and W.A. Beckman, (1980a)(1980b), Solar Engineering of Thermal Processes,
pp. 353, and pp. 264-268 respectively, John Wiley and Sons, New York.
Edwards, D.K., (1977), Solar Collector Design, pp. 31-32, The Franklin Institute Press,
Philadelphia, Pa.
Edwards, D.K., J. T. Gier, K.E. Nelson, and R.D. Roddick, (1961), "Spectral and Directional
-------
1569
Thermal Radiation Characteristics of Selective Surfaces for Solar Collectors", U.N. Conf.
on New Sources of Energy, Rome, Aug. 1961, E. Conf. 35/S/43, United Nations.
Hollands, K.G.T., (1963), "Directional Selectivity: Emittance and Absorptance Properties of
Vee Corrugated Specular Surfaces", Solar Energy, Vol. 7, No. 3, pp. 108-116.
Hollands, K.G.T., A. Karagiozis, A.P. Brunger, and G. Brouwer, (1990), "Effect of Selective
Surface Degradation on the Performance of Solar Water Heating Systems", Proceedings,
16th Annual Conference of the Solar Energy Society of Canada Inc., (SESCI), June 18-20,
1990, Halifax, NS, SESCI, Ottawa, 1990, pp. 229-234.
Hollands, K.G.T., and J.L. Wright, (1983), "Heat Loss Coefficients and Effective ra Products
for Flat Plate Collectors with Diathermanous Covers", Solar Energy, Vol. 30, No. 3, pp.
211-216.
Tabor, H., J. Harris, H. Weinberger, and B. Doran, (1961), "Further Studies on Selective
Surface Black Coatings", U.N. Conf. on New Sources of Energy, Rome, Aug. 1961, E.
Conf. 35/S/46, United Nations.
NOMENCLATURE
Ac
collector area
a
vertical intercept on a plot of
Aas versus Ae (see eqn (1))
B
multiplier derived by Hollands
(1963)
Fs
yearly fraction solar
A Fs
F — F
¦L so s
F'
collector efficiency factor
Fr
collector heat removal factor
F'r
modified collector heat removal
factor (accounting for effect of
heat exchanger)
G
solar irradiance on collector
(mCp)c
collector loop heat capacity
rate
Tc
collector cover temperature
Tm
mains water temperature
T
-L max
maximum temperature allowed
at top of solar preheat tank
T
1v
absorber plate temperature
T
-*¦ sp
DHW set point temperature
Ttb, Ttt temperature at top and bottom
of preheat tank, respectively
Txi inlet temperature to tank-side
of heat exchanger
Ul collector heat loss coefficient
Greek Letters
as solar absorptivity
ftsO '
j) slope of plot (see eqn (1))
e thermal emittance surface
Ac 6 — CO
<7 Stefan-Boltzmann constant
r cover transmittance
(ra)e effective transmittance absorp-
tance product
Subscripts
o value of a quantity when solar
system first goes into service
-------
1570
LOCATION
TORONTO
Q52 u.
0.950 "
0.925
0300"
0.875.
0.850
Q50
0.48
0.48
0.46
0.46
L. 0.44
0.950
0.925
0.900
0-875
0.850
0.42
LOCATION
RAPPERSWIL
0.40
0.2
03
0.4
€
0.5
0.6
0.7
Figure 1: Plots showing how Fs depends on e and
as in two locations (from Hollands et al., (1990)).
0.20
•DENVER
ZURICH ( RAPPERSWIL)
(S>
O
<3
-COPENHAGEN
0.05
O.OOL.
0.0
0.2
0.4
0.8
0.6
1.0
Ae
Figure 2: Dependence of Act, versus Ae plot
on geographical location (from Hollands et al.,
(1990)).
0.15
ALBUQUERQUE
Tsp« 90*C
STUTTGART
Tsp« 50 *C
7.2 (0.64)
4.8 <0.53)
3.4 (0.42)
0.10
005
£ 000
COPENHAGEN
TORONTO
Tgp ¦ 50"C
0.15
7.2 (0.70)
4.6 (0 58)
3.4 (0.45)
0.10
(0.50)
0.05
-(0.40)
(0.20)
O.oo I 1—
0.0 0.2
0.4 0.6 0.0 0.2
0.4
0.6
0.8
Figure 3: Dependence of Aas versus Ae plot on
collector area Ac (from Hollands et al., (1990)).
0.25
— FOR TORONTO WITH
Ac ¦ 4.8 m4
— - FOR COPENHAGEN WITH
Ac ¦ 3.36 m*
0.20
0.15
tn
a
<3
O.IO
0.05-
0.001-
0.0
0.2
0.4 0.6
As
0.8
Figure 4: Dependence of Aas versus Ae plot on
the set point temperature Tsp (from Hollands et
al., (1990)).
0.50
LOCATION : TORONTO
0.40
030
•S 020
01
•LINE OF PERFECT AGREEMENT
0.10
0.0
0-0
0.10
Q20
0.30
0.40
0.50
/^analytical
Figure 5: Checking of theory predicting the
slope, /3, of the Aas versus Ae lines.
LOCATION:
¦ TORONTO
A STUTTGART
• AL8U0UER0UE
O COPENHAGEN
O RAPPERSWIL
* OENVER
3.50
3.00
35 #
1.50
,4e°i
LOO
QO
0.20
0.40
0.60
0.80
10.
Figure 6: Plot showing the dependence of the ra-
tionalized value (a/a)/(AFs/Fso) of vertical in-
tercept a on F$o and the geographical location.
Plot contains all data in Figs. 7-10, of paper by
Hollands et al., (1990).
-------
1571
POTENTIAL OP SOLAR DISTRICT HEATING IN FINLAND
Seppo S. Peltola
Helsinki University of Technology
Department of Technical Physics
SF-02150 Espoo, Finland
ABSTRACT
Central Solar Heating Plants with Seasonal Storage (CSHPSS) systems are appropriate for solar space
heating applications when high solar fractions are desired. For lower solar fractions, Central Solar
Heating Plant with Diurnal Storage (CSHPDS) might be considered. In this paper we study the pos-
sibilities to apply the CSHPDS concept to existing Finnish district heating systems by means of
numerical simulation. Two cases are dealt with: small district heating systems with heat production
only and larger ones with combined heat and power production. This work has been carried out with
financial support from national research programme on new energy sources, NEMO.
KEYWORDS
Solar heating; short term storage; district heating; simulation.
INTRODUCTION
During last years, considerable efforts have been laid on studies o f Central Solar Heating Plants with
Seasonal Storage (CSHPSS) (IEA 1990). These large solar heating systems show today cost level
approaching those of conventional heat production systems. The key issues in promoting the use of
CSHPSS technology are (IEA 1990) bringing the cost of collector manufacturing to lower level by
means of industrial manufacturing and to develop cost-effective storages also in smaller sizes.
Large scale solar heating systems could also be built using short term heat storage in stead of the
seasonal one. In this case, the annual solar fraction decreases considerably and the system would
operate mainly in the summer time. On the other hand, when combining these CSHPDS systems with
existing district heating (DH) plants and networks, smaller investment costs could be achieved due to
smaller system size and, in some cases, due to the possibility to take benefit from existing system
components. These cheaper installations could smooth the way for high solar fraction CSHPSS systems
by creating a market for collector industry so that the benefits of serial production could be gained.
There are already today some experimental CSHPDS projects. For example, in Nykvarn near Stock-
holm, Sweden, a CSHPDS installation has been built in 1985 (Isakson 1989). This plant featured
originally 4,000 m2 of high performance flat plate collectors and a 1,500 m3 steel tank storage producing
heat at cost level of approximately 41 ore/kWh (7 tf/kWh) and a decision has been made to increase the
collector area by an additional 2,000 m2. Another Swedish experiment is in Falkenberg (Swedish
Council for Building Research, 1990) where 5,500 m2 of flat plate collectors feed a DH network through
a 1,100 m3 short term storage tank. This system should cover some 10% of the annual heating demand.
District heating is quite common in Finland; approximately 40% of buildings are connected to a DH
network (District Heating Statistics, 1990). In 1989, 34.3 TWh of fuels were used in DH plants to
produce 22,1 TWh heat and 6,8 TWh electricity. The main fuels were coal (47%), peat (18%), natural
gas (18%) and heavy fuel oil (12%). The smallest networks employ often only oil burners without
-------
1572
combined heat and power production. In some networks having combined heat and power production,
too small heat load disables the use of large back pressure power plants in the summer. The utility
must then buy electricity and use heavy fuel oil fired boilers to meet the heating load. Both of these DH
network types provide a potential application for summer time solar district heating as they use the
most expensive fuel during the time when solar energy is most abundant.
During last years, small short term storages, typically one or more steel tanks with volume of 10,000
m3 , have been built for combined heat and power production systems and used for peak shaving and
for optimizing combined heat and power production. They provide also an extremely interesting chance
for applying CSHPDS technology as the storage usage during summer months is minimal. Accordingly,
a CSHPDS system could be constructed with no extra cost due to short term storage by simply building
a collector field and using the existing steel tank(s) for short term storage.
METHODS APPLIED
In this study the TRNSYS model (Klein and others 1983) has been applied for thermal analyses. The
programme is very flexible, including a variety of different component models needed for this kind of
analyses. Furthermore, the TRNSYS model includes a special programmable component that can be
used for modelling,for example,specific control devices. The lately published pre-processor PRESIM
(Nordlander and others, 1990) was used to produce the TRNSYS input.
The basic CSHPDS system includes collector field and a short term heat storage connected to a DH
network through a heat exchanger. The DH system itself comprises the consumers, network, necessary
pumps and heat producing elements (boilers and/or back pressure power plants). A schematic of the
system studied is shown in Pig. 1.
COLLECTOR FIELD DIURNAL STORAGE CONVENTIONAL BOILER
LOAD
Fig. 1. Schematic of a CSHPDS system as modelled.
For simplicity, we have modelled the DH system by using a TRNSYS component type 15, Algebraic
Operator B to calculate the momentary heat load, desired forward temperature in the network and the
necessary mass flow rate. The return temperature has been kept constant. The collector field has been
modelled using TRNSYS component type 1 with linear efficiency mode. The heat exchanger between
storage and collector field has been accounted for by modifying the collector characteristics with heat
exchanger factor (see for example Williams 1983). The collector circuit has a variable flowrate to reach
either the momentaiy DH delivery temperature or the storage top layer temperature, whichever is
higher. The collector loop was coupled to the storage with fixed inlet and outlet.
Short term storage was modelled using TRNSYS component type 4 with 10 nodes. A fixed inlet/outlet
pair was used for DH network connection without a heat exchanger. Direct coupling between storage
and load is normally used in short term DH storages,as it is much cheaper than using heat exchanger.
-------
1573
Also, the storage can then easily be used as a hot water reservoir for DH network in case of pipe
failure. Additional TRNSYS components used comprise Solar Radiation Processor (type 16) and various
I/O devices. The weather data have been measured in 1979 in Helsinki, 60 °N, and corresponds on
annual basis to long time average values (Tammelin & Erkio, 1987). There is, however, a major
deviation in July data with lower ambient temperature and radiation levels than average values
resulting in larger collector area and storage volume than if average weather data were used.
RESULTS
Small CSHPDS Systems for District Heating Networks without Co-generation
The reference system in this case comprises a DH network located in southern Finland with annual
heat load of 8.2 GWh. District heat is delivered in the actual system both to multi-family houses and
commercial buildings. For simplicity, we have described the load in the calculations as a pure multi-
family house load of 500 flats. The return temperature has been set to 45 °C and the feed temperature
varies according to ambient temperature with minimum of 65 °C. The collector costs have been taken
from International Energy Agency (IEA) Solar Heating and Cooling Program (SH&CP) Task 7:250$/m2
for array sizes around 1,000 m2 (IEA 1990). The storage tank and DH network connection costs have
been derived from a national district heat storage technology survey (Sipila 1989) ending up to 360
mk/m3 for storage and 500,000 mk for the connection between storage and delivery network.
The thermal performance results from TRNSYS analyses were combined with the cost estimates in
three ways. First, the costjwere taken as such to indicate the commercial cost of solar heat. The other
two cases studied reflect the effects of governmental subsidies; reduction of VAT (20%) which will
become effective later in 1991, and a total of 40% investment subsidy corresponding to the maximum
possible subsidy for a heating plant according to current Finnish practice. The resulting heat costs have
been plotted versus solar fraction in Fig. 2. In all cases 25 yr lifetime and 5% interest rate was used.
The results indicate a relatively flat optimum around 10% solar fraction. The optimal collector area is
2,000 - 2,500 mz and the storage should be kept as small as possible. For comparison, the fuel cost of
a heavy fuel oil fired boiler plant was in 1989 at a level of 70 mk/MWh (Energy review 1990) but was
during 1981 - 1985 about 100 mk/MWh. Accordingly, the current low oil prices do not allow for
economic solar heat production with this kind of system but if the oil prices return to the level they
were at some five years ago, the solar alternative would become almost competitive.
Another factor affecting the applicability of a CSHPDS for summer time district heating is the time
during which the system can meet the heating load alone. The daily solar fractions with different
500
A 400
s
\
g 300
C0
V
f 200
" 100
$ I
t...
Y, V
* ^
"f
No subsidies
VAT reduction
40% subsidies
/
10
Solar fraction, 7.
15
-i 125
100
A
ie
S
75 ^
50
a
-------
1574
100
Collector area
— 4000 sq.m.
— 3500 sq.m.
— 3000 sq.m.
— 2500 sq.m.
— 2000 sq.m.
— 1500 sq.m.
— 1000 sq.m.
80
a
.2 60
o
a
u
«tH
fe 40
ffl
o
m
20
300
50
250
0
100
150
200
days
Fig. 3. Duration curve of daily solar fractions with different collector areas.
collector areas and storage volumes are given in Fig. 3. as a duration curve, i.e. the solar fractions have
been ordered in decreasing sequence. The short term storage size has been varied systematically for
each collector array size and the corresponding range of days at each solar fraction is given.
With 2,000 m2 of collectors 100% solar coverage can be reached during at least one day. For longer self-
sufficient operation, the CSHPDS should however, have at minimum 3,000 m2 of collector or preferably
even more. Full summertime solar coverage can be achieved with 4,000 m2 collector field. These sizes
are above the economic optimum, but the associated heat costs are only a few percent higher. If the
benefits of self-sufficient operation are taken into account, these larger system are of interest.
The role of short term storage is marginal with small collector arrays, as there is not much excess
energy above the heat load that could be stored. Consequently, a quite narrow band of daily solar
fractions with different storage volumes can be seen in Fig. 3 with smaller array sizes. For areas above
3,000 m2 a broader band of days in the solar fraction range between 60 - 90% can be seen,and with a
4,000 m2 array size the widening extends also to 100% coverage.
The storage volume can be determined from the original time series of daily solar fractions. If the
CSHPDS system should meet the load alone during summer, the number of consecutive days with 100%
solar coverage becomes important. These are shown in Fig. 4. for array sizes between 2,500 and 4,000
m2 as a function of storage volume. The rise in curves for array sizes above 3,000 m2 results from the
long rainy period in the middle of July. If the storage can compensate for that period, the length of
totally self-sufficient period is increased sharply. As the investment costs originate mostly from the
collector array, the cost of heat is not heavily affected by adding up some storage capacity. Therefore,
the most propable CSHPDS system would have a collector area of 3,500 - 4,000 m2 and a 2,000 m3
storage with heat cost range from about 240 (no subsidies) to 145 (40% subsidies) mk/MWh.
CSHPDS Systems for District Heating Networks with Co-generation
More promising alternative, although having smaller number of applicable DH systems, is a network
with combined heat and power production and an existing short term heat storage. In these systems the
heat storage is operated mostly outside the main solar collector operating period,and the additional cost
of constructing a summer time CSHPDS system would come from the collector field.
The basic system is the same as for our previous case. The main differences are fixed storage size of
10,000 m3 (typically used in DH networks) and a larger heat load. Also the collector cost has been
lowered to 200 $/m2 due to larger array areas in accordance with the IEA studies (IEA 1990). DH
-------
1575
4000 sq.m.
3500 sq.m.
3000 sq.m.
2500 sq.m.
1000 1500 2000 2500 3000
Volume, cub.m.
Fig. 4. Length of full 100 % solar coverage period.
system data has been adopted from a small town near Helsinki with annual heat load of 230 GWh. The
number of flats is 9,600, and the rest of the load is due to commercial and industrial customers. For
sake of simplicity, the load is again calculated as a normal house heating load only.
The solar fractions and heat costs as a function of applied collector area are shown in Fig. 5. Avery flat
optimum was obtained at solar fraction of about 7%. Because the costs come from the collector field
only, the resulting heat cost is even lower than in our previous example. It is worth to notice that the
cost level achieved with the maximum heating plant investment subsidy obtainable in Finland is almost
the same as the fuel costs of a heavy fiiel oil boiler. Therefore, these systems should be economically
attractive in very near future.
Although the cost estimates are encouraging, we must note that the solar yield has been calculated
separately from the rest of the system i.e. without any interaction with other heat producing com-
ponents. In real life, the storage operating strategy would be affected by the economics of, for example,
the marginal costs of own heat and power production with respect to the momentary costs of bought
electricity. Specially during spring and autumn, the storage may well have different state than that
10 r
150
125 S=
• Solar fraction
a Cost, no subs.
Cost, ZOZ subs.
O Cost, 407. subs.
10000
20000 30000 40000 50000
Collector area, sq.m.
Fig. 5. Solar fraction and cost of heat of a CSHPDS with 10,000 m3 storage.
-------
1576
predicted in our simulations. For more accurate analysis, a full system model would be needed includ-
ing models for national power production and associated time-dependent marginal costs of electricity,
models for back pressure power plant etc. but this is beyond the scope of this study.
CONCLUSIONS
The concept of CSHPDS seems to be most attractive specially if an existing short term storage can be
used during summer. In this case, the cost of solar heat is almost competitive when compared to oil
fired boilers. Smaller DH networks with heat production only and without an existing storage facility
show economics comparable to solar costs of CSHPSS systems. In IEA SH&CP Task 7, the costs of a
200 flat CSHPSS with flat plate collectors and a water pit storage, the costs were found to be 240 - 320
mk/MWh (60 - 80 $/MWh) whereas this study indicates cost level of 214 - 254 mk/MWh.
The potential of CSHPDS systems can be estimated from district heating statistics. In 1989, 863 GWh
of heat was produced in DH networks using heavy fuel oil only. A 10% annual solar fraction results in
nearly 90 GWh annual solar energy production potential. As the simulated specific collector yield was
approximately 360-370 kWh/m2 the 90 GWh annual solar energy production could offer market for
about 240,000 m2 of high performance flat plate collectors.
Additional potential can be found from those DH networks that have combined electricity production
but its magnitude cannot be estimated directly. However, the total volume of existing steel tank
storages is today about 90,000 m3 and an additional 57,000 m3 of existing but unused steel tanks could
be taken in use. Even a partial usage of this storage capacity would increase the potential collector
market by an additional 100,000 m2. Altogether, the CSHPDS concept should be considered as a good
alternative for solar heat production.
REFERENCES
International Energy Agency (1990), Solar Heating and Cooling Program Task 7 Final Report (Ed. J-0
Dalenback). Swedish Council for Building Research, Document D14:1990, Stockholm.
Isakson, P. (1989), Nykvarn - a Solar District Heating Plant witn Short Term Storage. Proc. North Sun
'88, August 28-31,1988 Borlange, Sweden. Swedish Council for Building Research, Document D2:1989,
Stockholm, Sweden.
Swedish Council for Building Research (1990), Solar Energy for Buildings, Report G7:1990, Stockhom,
Sweden (in Swedish).
District Heating Statistics 1989 (1990), Finnish District Heating Association, Helsinki (in Finnish).
Klein, S.A. and others (1983), TRNSYS, a Transient System Simulation Program, Univ. Wisconsin,
Madison, USA.
Nordlander, S. and others (1990), PRESIM User's Manual. Solar Energy Research Center, Univ. College
Falun/Borlange, Sweden.
Williams, J.R. (1983), Design and Installation of Solar Heating and Hot Water Systems. Ann Arbor
Science, Michigan, USA.
Tammelin, B., Erkio, E. (1987), Weather Data for Energy Analyses, Finnsih Meteorological Institute,
Report 1987:7 (in Finnish).
Sipila, K. (1989), Insulated Steel Tanks and District Heating Network, Proc. Nat. Seminar on Heat
Storages, December 13, 1989. Technical Research Center of Finland (in Finnish).
Energy Review 4/89 (1990), Ministry of Trade and Industry, Energy Department, Helsinki, Finland (in
Finnish).
-------
1577
PRELIMINARY DESIGN DEVELOPMENT OF A
CENTRAL SOLAR HEATING PLANT WITH SEASONAL STORAGE AT THE
UNIVERSITY OF MASSACHUSETTS, AMHERST
D. S. Breger, J. E. Sunderland and H. Elhasnaoui
Department of Mechanical Engineering, University of Massachusetts
Amherst, Massachusetts 01003 USA
ABSTRACT
The preliminary design of a Central Solar Heating Plant with Seasonal Storage (CSHPSS) located
at the University of Massachusetts in Amherst is discussed. The project would represent the first
implementation of this solar technology in the United States and results from the International
Energy Agency collaboration on CSHPSS since 1979. The preliminary design calls for a large
11,000 m2 flat plate collector array, 75,000 m3 storage volume in clay with heat transfer through
600 boreholes. Design optimization is based on computer simulations using MINSUN and
TRNSYS. The design is expected to provide 89% of the 3500 MWh heating and hot water load.
A project cost of $3.2 million is estimated, which provides an annualized cost of $73.6/MWh per
unit solar energy delivered. The project will proceed into an engineering phase in Spring 1991.
KEYWORDS
IEA (International Energy Agency), CSHPSS (central solar heating plants with seasonal storage),
solar heating, seasonal storage, duct storage, simulation, economic optimization.
SUMMARY AND STATUS OF CSHPSS TECHNOLOGY
Building and hot water heating are attractive solar energy applications due to the low temperature
requirements. Past solar approaches to space heating in cold winter climates have not fared well
technically or economically since the winter resource in northern climates is not adequate. The
seasonal storage approach allows for efficient solar collection throughout the year to charge a
thermal storage facility so that a sufficient heat source is prepared to meet nearly 100% of the
winter load. Seasonal thermal energy storage can be efficiently accomplished in various geological
formations, such as boreholes in clay deposits or bedrock, aquifers, and excavated (water filled)
rock caverns or earth pits. When developed on a large scale, CSHPSS are very efficient and can
provide cost competitive and reliable energy.
CSHPSS technology has been the subject of an International Energy Agency (IEA), Solar Heating
and Cooling Program Task VII since 1979 (Dalenback, 1990; Bankston, 1986). Over 30 projects
are now in operation in the IEA countries and these projects demonstrate the technical and
economic viability of this solar approach (Dalenback, 1990). CSHPSS is most highly developed
in Sweden where 13 projects are in operation and several new large projects are planned.
-------
1578
SITE AND PROJECT DESCRIPTION
Overview
The University of Massachusetts (UMass) campus at Amherst was selected for the CSHPSS project
after a careful evaluation of other State and private facilities. The site geology is well-known and
attractive for seasonal storage, land area for a large collector array is available, and suitable
building heating loads are located nearby.
The site of the proposed CSHPSS project will occupy roughly seven acres of land and is isolated
from pedestrian traffic and has limited value for alternative use. The site is underlain by over 30
meters (100 feet) of soft saturated clay which is a predominant geological feature of the region.
The clay depth and characteristics in the vicinity of the site are well-known from U.S. Geological
Survey studies and boring records from the University Physical Plant and Town of Amherst.
Two large buildings located about 550 meters (1800 feet) across athletic fields from the project site
will serve as the primary space heating and hot water loads for the CSHPSS project. The Mullins
Memorial Arena, now under construction, will be a new 10,000 seat multipurpose facility and its
HVAC design has been modified to accommodate the low temperature solar hot water heating
source. The Boyden Gymnasium is an existing building with attractive retrofit opportunity,
including existing utility conduits to facilitate installation of the piping to the loads. The total load
for the CSHPSS project is approximately 3500 MWh, with a nominal delivery temperature of 55
°C (130 F) and a return temperature of 40 °C (104 ' F).
An illustration of the CSHPSS project is shown in Fig. 1. Solar thermal energy is stored directly
in the clay by circulating hot water through U-tubes which are drilled or driven vertically through
the clay. Heat is extracted from the clay using the same heat exchange pipes and distributed
through a hot water district heating network to the loads.
COLLECTOR
ARRAY
Wastewater Treatment BgjjlJ
[BOREHOLE
STORAGE .
. MULUNS/
'ARENA 1
Commonwealth Avenue
BEDROCK
Fig. 1. Illustration of proposed CSHPSS project at UMass.
-------
1579
Collector Array
The site for the collector array is flat, open land with little value for other purposes and reasonably
isolated from campus activity. The proposed site has been flagged for wetlands which cover a
portion of the area. The marginal wetlands are of concern for ground water recharge which will
not be adversely influenced by the collector array.
The solar collectors will be high efficiency flat plate collectors. The collectors are based on the
IEA Task parameters used to model currently available Swedish technology (Dalenback, 1990).
The collector efficiency parameters are FRxa = 0.75 and FRUL = 2.90 W/m2K, and an array factor
of 0.88 to account for net efficiency losses associated with large collector arrays compared with
single modules.
Seasonal Storage — Duct Storage in Clay
The seasonal storage will be a duct (or borehole) system in the deep clay deposit. Large duct
storage systems have been demonstrated in several European projects, though only Groningen
(Wijsman and den Ouden, 1983) in The Netherlands combines the clay storage and high
temperature (no heat pump) which characterizes the UMass project. The boreholes will consist of
polymer U-tubes circulating the heat transfer fluid; the tubes will be drilled or pushed into the soft
saturated clay. The U-tubes will be connected on site and the piping network directed to a site
mechanical equipment station. The top of the store will be finished with a layer of sand,
insulation, a liner (possibly clay), and reclaimed top soil.
Knowledge of the geotechnical and thermal properties of the clay is based on data from laboratory
analysis of core samples down 24 meters (80 feet) taken by the UMass Department of Civil
Engineering in August 1989. This data has been incorporated with hydrogeological relations and
an analytical model (Ingersoll, 1988) to calculate the essential properties for analyzing and
simulating the thermal energy storage.
Below the top 3 meters (10 feet) of stiff clay, the
clay properties are rather homogeneous. The
average values based on the laboratory analysis of
the core samples and the calculated thermal
properties are shown in Table 1. Compared with
thermal properties documented for other European
(non-rock) duct storage projects (Chuard and
Hadorn, 1983), the UMass geology appears to
provide very favorable conditions for seasonal
thermal energy storage. The varved characteristic
of the clay is not considered in the Ingersoll
model used to derive the thermal properties and
may improve heat transfer through local
convective fluid flows and by maintaining a
wetted surface around the U-pipes during heat
injection to counteract the induced thermal
diffusion.
TABLE 1. Properties of Clay Storage
PROPERTY
VALUE
Water Content
0.757
Porosity
0.597
Density
1805 kg/m3
Permeability
0.373x10"6 cm/s
Specific Heat
2220 J/kg°K
Heat Capacity
4.0 MJ/m3oK
Therm. Cond.
1.9 W/m°K
Horiz. Water Flow
< 1 meter/yr
-------
1580
DESIGN SIMULATION AND OPTIMIZATION
Analytical Method
The MINSUN program (Mazzarella, 1989) is a preliminary design tool for CSHPSS developed by
the IEA Task and has been used to simulate and optimize the UMass design. The DST (Hellstrom,
1982) subroutine is used with MINSUN for the duct storage subsystem. For the optimization
process, an iterative MINSUN simulation is run over a range of design parameters defining
collector area, storage volume, and number of boreholes.
The TRNSYS model (Klein, 1988) has been adopted for further evaluation of the CSHPSS design
optimization and alternatives in system configurations and operation. Other programs from Lund
University (Eskilson, 1986) will be used to study the borehole design and heat transfer in detail.
The detailed (standalone) DST version will allow evaluation of radial temperature stratification in
the storage.
Cost Optimization
The performance results of the system simulations are combined with cost data to optimize system
design. Economic optimization is based on each system's unit solar cost, calculated as the
annualized total capital cost divided by the solar energy supplied by the CSHPSS.
The solar cost is plotted versus solar fraction in Fig. 2 and the expansion path is indicated as the
envelope of system configurations which are of economic importance. Systems above the
expansion path produce the same solar energy output as the design on the envelope but at higher
cost, and are therefore easily dismissed as non-optimal.
90
Storage Volume, m3 50000, 75000, 100000, 125000
(volume changes along connected data points)
No. of Boreholes 4O0, 600, 800, 1000
(boreholes change between connected data sets)
84-
82-
80-
o
DC
78-
8000
7000
Collector Area, m2
12000
11000
9000
EXPANSION PATH
10000
72"
SELECTED DESIGN
70-
0.60
0.70
0.75
SOLAR FRACTION
O.90
1.00
0.80 0.8!
FRACTION
Fig. 2. Expansion path diagram for design optimization.
-------
1581
The expansion path shows that the CSHPSS project
can deliver a solar fraction as high as 0.90 at a cost of
$74/MWh. The economic optimum design occurs
along the expansion path at the point where the
marginal cost of increasing the solar fraction (revealed
by the slope of the expansion path) is equal to the
displaced conventional cost. This optimum is
commonly reached at the point on the expansion path
just before the slope increases dramatically at the high
solar fraction. The selected system design is indicated
in the figure and its design specifications and
predicted performance and cost are shown in Table 2.
To support the MINSUN selection and prepare for
more detailed design studies, TRNSYS 12.2 has been
used to simulate the CSHPSS based on the MINSUN
design configuration. A comparison of the MINSUN
and TRNSYS performance simulation results for the
selected design and several other close parameter
values indicate that the two models closely agree.
TABLE 2. Design Specifications
DESIGN PARAMETERS
Annual Load
3500 MWh
Collector Area
11,000 m*
Storage Volume
75.000 m3
No. Boreholes
600
CSHPSS PERFORMANCE
Solar Fraction
.89
Collector Pert.
374 kWh.m2
Collector Eff.
28.7 %
Storage Temp.
52.3-73.3°C
Storage Eff.
81.4%
CSHPSS ECONOMICS
Capital Cost
$3,222,000
Ann. Solar Cost
$73.6'MWh
PERFORMANCE AND ECONOMIC ANALYSIS OF SELECTED DESIGN
Annual Performance Results
The annual performance of the selected design is shown in Fig. 3. Results are presented for both
the MINSUN and TRNSYS simulations and a favorable comparison is noted. The graph shows
the monthly energy flows of the CSHPSS and the associated temperature level in the clay storage.
The solar energy collected in the summer is stored as reflected by the increasing storage
temperature. The winter load is then met primarily by the store and the clay temperature decreases.
The solar collectors capture 374 kWh/m2 annually providing a 28.7% efficiency relative to the total
annual incident insolation. Storage losses represent 18.6% of the energy stored.
System Economic Results
Based on financial conditions of a 5% real discount rate and 20 year economic lifetime, the
annualized cost of the supplied solar energy is $73.6/MWh. The annual cost of the CSHPSS
project is greater than the displaced heating costs based on projections issued by the Massachusetts
Energy Office; but under assumed cost reductions in collector technology and savings in
environmental costs, future CSHPSS applications are attractive.
FURTHER PROJECT DEVELOPMENT
Phase HI Project Engineering is scheduled to begin in Spring 1991 with support from the U.S.
Department of Energy. The primary objective of the Phase IH effort is to advance the development
of the CSHPSS project at UMass through an engineering design resulting in pre-construction
specifications of all significant components and construction tasks and detailed cost and economic
analysis on which to base a construction phase budget and system operation business plan. The
duration of the Phase IH work plan is 18 months.
-------
1582
700-
MINSUN
TONSYS
SOLAR ENERGY
CCUECTH)
BUILDING
LOAD
600-
u. 300-
Si 200
ENERGY LOST
FROM STORE
100-
APR MAY Jltd JUt AUG SEP OCT NOV DEC JAN F€B UW
MINSUN
TRNSYS
70-
O
UJ
56-
50i
MAR
APR MAY JUN JUL AUG
OCT NOV DEC JAN
(a) Monthly energy flows. (b) Storage temperature.
Fig. 3. Annual performance simulation — MINSUN and TRNSYS.
REFERENCES
Bankston, C.A. (1986). Central solar heating plants with seasonal storage: evaluation of concepts.
IEA Solar Heating and Cooling Program, Task VII, Report # T.7.2.B., U.S. Government
Printing Office, Washington, D.C.
Chuard, P. and J.-C. Hadorn (1983). Central solar heating plants with seasonal storage — heat
storage systems: concepts, engineering data and compilation of projects. IEA Solar Heating
and Cooling Program, Task VII, Sorane SA, Switzerland.
Dalenback, J.-O. (1990). Central solar heating plants with seasonal storage — status report. IEA
Solar Heating and Cooling Program, Task VII, Swedish Council for Building Research,
Stockholm, D14:1990.
Eskilson, P. (1986). Superposition borehole model: manual for computer code. Department of
Mathematical Physics, Lund University, Sweden.
Hellstrom, G. (1982). Model of duct storage system: manual for computer code. Department of
Mathematical Physics, Lund University, Sweden.
Ingersoll, J.G. (1988). Analytical determination of soil thermal conductivity. Trans. Am. Soc.
Mech. Engrs., 110. 306-312.
Klein, S.A. (1988). TRNSYS - a transient system simulation program. Engineering Experiment
Station, Report 38-12, University of Wisconsin, Madison.
Mazzarella, L. (1989). Central solar heating plants with seasonal storage: the MINSUN simulation
program — application and user's guide. IEA Solar Heating and Cooling Program, Task
VH, C.N.F.-P.F.E., Rome, Italy.
Wijsman, A.J. and C. den Ouden (1983). Groningen: a group of 96 solar houses with seasonal heat
storage in the soil. National Report of The Netherlands for IEA Solar Heating and Cooling
Program, Task VH, Sub-task 1(e), TNO-TH, Report No. 103-220.
-------
1583
STUDY OF A 20,000 m3 SEASONAL HEAT STORAGE FED
BY SOLAR COLLECTORS
0. Guisan, B. Lachal, A. Mermoud. D. Pahud
University of Geneva, Applied Physics, Geneva, Switzerland
ABSTRACT
This study deals with an underground seasonal heat storage built under a
large industrial building and operating between 5 and 35 °C. We present its
characteristics, results obtained after 20 months of detailed monitoring, as
well as a modeling approach of such a storage.
KEYWORDS
Seasonal storage; heat storage; solar storage; underground storage; diffusive
storage.
INTRODUCTION
A large industrial building (75'000 m3) built near Geneva (Marcinhes, Meyrin)
was carefully designed for efficient energy use by Matthey and Roulet (1986,
1987). It involves 950 m2 of flat plate collectors on the roof (for heating
and domestic hot water), an underground seasonal heat storage (20*000 m3)
with 258 vertical heat exchangers (or wells) 14.5 m deep and 2.3 m apart
below the building, solar gains through passive solar walls (two double
glazing), a heat pump, auxiliary furnaces (oil, gas and wood), floor heating
as well as conventional heating devices. The fuel consumption should be below
150 MJ/m2y, where the reference area is the heated floor area.
The storage is fed by the solar collectors, mostly in summer time and
occasionally during winter when excess heat is available. Heat is extracted
from the storage by the heat pump in winter time when heating is required.
Solar gains may also be driven directly to the heat pump. The storage, and
consequently, the solar collectors are operating at low temperature
(typically 5-35°C for the storage).
We are monitoring the entire building in detail. The solar collectors are
working properly and as expected. They were specially built as part of shed
structures and they constitute the waterproof cover of the flat roof of the
building. They are single glazed collectors with selective absorbers. Our
measurements show that they are characterized by an optical efficiency of
0.74 and by thermal losses (including piping of the array) of 5.1 W/m2K.
The gas powered heat pump (~200 kW) failed most of the first winter operation
period (1989-1990), because of a control problem. The solar collectors,
therefore, had to be covered in summer 1990, in order not to exceed the
temperature limits (~50 °C) for the plastic tubes of the storage. The heat
pump is now working properly with an overall coefficient of performance
-------
1584
(useful heat/gas energy content) of around 1.7 which is quite satisfactory.
We now focus on the underground seasonal storage for which measurements
started May 1989. This study is still going on, but we can already present
some interesting and preliminary results.
THE UNDERGROUND STORAGE
The underground storage is located under the building and separated from the
building by a thermal insulation layer. Its characteristics are described in
the following table.
TABLE 1 Characteristics of the Underground Storage
Volume
Horizontal area
Upper insulation
Ground
20*000 m3
1'400 m2 (44 ix 32 •)
5 cm of foam glass
dry moraine
Number of boreholes
Borehole diameter
Borehole depth
Distance between boreholes
Total length of boreholes
258
114 mm
14.5 m
2.3 m (square network)
3740 m
Heat exchanger
Tube diameter
Tube material
Thermal contact between ^
tubes and ground J
Total length of tubes
Tube connections
Thermal stratification
Heat transfer fluid
4 tubes, i.e. 2 U-tubes/borehole
32 mm
polyethylene
fine sand
+ water injection at the head of
the boreholes if or when necessary
15500 m
by series of 13 wells
2 series in parallel for both
U tubes of each well (in case of
leak, one series can be shut off)
along storage radius, heat
injection: fluid goes from center
to periphery, heat extraction:
fluid goes from periphery to center
water
Maximum injection power
Maximum extraction power
Temperature range
Annual stored energy
200 W per m of borehole
25 W per m of borehole
~ 5-35 °C
~ 400 MWh
MEASUREMENTS
We restrict ourselves, in this paper, to the measurements concerning only the
underground storage.
Temperature measurements are achieved by the use of platinum resistances
(Pt-100) with an accuracy of the order of 0.1 °C.
-------
1585
We measure continuously, with the data acquisition system devoted to the
entire building, the total heat flow going in and out of the storage (i.e. 1
flowmeter and 2 temperature sensors). We also measure, with the same system,
the heat flow for the first, the second and all wells of a series of 13 wells
(i.e. 1 flowmeter and 4 temperature sensors). We also recently connected 3
additional channels for temperature measurements at the interface between the
storage and the building, in order to investigate thermal gradients and
related thermal losses at the upper surface of the storage (these losses are
recovered as heating by the building).
We measure once a week, 5 vertical temperature profiles in 5 special wells
located between and equidistant from adjacent storage boreholes. The first
well is in the middle of the storage, the second one at half a radius of the
storage and is considered as representative of the whole storage, and the 3
last ones are aligned along a same radius at the periphery of the storage in
order to investigate temperature gradients and heat losses from the storage.
Each measurement well is equipped with 10 temperature sensors at different
depths from the upper surface to a few meters below the storage (0.5, 1, 2,
6, 10, 13, 14, 15, 16.5 and 19 m). The temperature sensors are mounted in a
plastic tube (diameter: 30 mm), filled up with glycerine to insure a good
thermal contact and to prevent thermal convection. These weekly measurements
are performed by hand with a special device.
RESULTS
We show on the 3D-plot of Fig. 1 how the vertical temperature profile of the
storage (well 2) has evolved with time. Because of the heat pump failures,
the storage temperature did not return after the first winter of operation to
its expected minimum value (5-10°C). Various gradients appear clearly on such
a figure.
On Fig. 2 we present vertical temperature profiles for the 5 monitored wells
at the same time (or date), during a loading period (a) and during an
unloading period (b). Temperature gradients at the periphery of the storage
appear clearly and may vary with time.
We show on Fig. 3 how the average temperature of the storage has evolved
versus time (a) and versus the heat injected into the storage (b). We
consider, as a preliminary and arbitrary estimate, the average temperature of
well 2 for the depths 6 and 10 m as representative of the storage. The
injected heat includes negative contributions related to extracted heat. On
Fig. 3b the storage losses during one cycle can be evaluated by the fact that
the storage temperature after one cycle returns to the same value; for
instance ~300 MWh at 25 °C, which corresponds (see Fig. 3a) to a 13 month
period (particular case of an annual cycle, not yet stabilized). We also see
that, for the winter of 1990-1991, 150 MWh were recovered from the storage in
a 3 month period. The thermal capacity of the storage is expected to be
around 15 MWh/K (20*000 m3 x 2.7 MJ/m3K). When the operation of the storage
started, losses were negligible and the initial slope of the curve on Fig. 3b
corresponds to the inverse of the thermal capacity of the storage. Then we
find 14 MWh/K in good agreement with the expected value.
Let us further mention that the observed storage losses are higher than what
should be in normal conditions for two reasons: the average temperature of
the ground around the storage is not yet stabilized at its final value (which
will take a few years) and, due to the heat pump failures, the storage
temperature went higher than expected. All these aspects are still to be
carefully investigated during the coming years.
-------
1586
o
30
a) -
20
Date
me
.91
7.1.89
Fig. 1. Temperature of well 2 versus depth and time.
3 2 1
5 4 3 1 2
.well number
\
\
n
)
a
'1
X
y
r
A
Maj
t 28
990
-10
10 20
Temperature
30
-20
1
K
_b_
)
t
/
f/
/
Dec
. 11
1989
40 10
20
30
°C Temperature
Fig. 2. Temperature profiles of the 5 monitored wells at given dates.
-------
1587
40
o
o
30
30
20
£ 20
s~
>- 1 n
• ••• measurements
— model
.91
0 100 200 300 400 500 600 700
Net heat into storage MWh
Date
Fig. 3. Storage temperature versus time and versus injected heat.
MODEL
The purpose of modeling the underground storage is to describe reasonably
well, and consequently to be able to predict and to optimize, temperature
behaviours (versus space and time) and energy flows.
We first developed the one cell model. We consider one vertical heat
exchanger and its surrounding ground as a cylindrical cell. We solve
analytically the heat diffusion equation along one dimension, the radius, by
means of Bessel functions. For instance, we define initial conditions, we
give the heat flow injected inside the cylinder, we assume no external losses
from the cylinder and we then compute temperature distributions. This model
was developed and described in detail by Pahud (1989). It works rather well
for cells in the middle part of the storage.
We now consider the ground surrounding the storage as a second larger cell
(also cylindrical) and we consider the external temperature of the single
cells as the internal temperature of this larger cell. Then we can compute,
using the same routine as for single cells, the heat flow entering this large
cell, that is, the lateral heat losses of the storage. We call this procedure
the double cell model. We account for the losses below the storage by
increasing correspondingly the lateral area of the storage. We may now
correct the heat flow injected in the single cells for the external losses
and iterate the process. We still evaluate the losses at the upper surface of
the storage as usual conductive losses. Finally, by knowing the heat injected
in and extracted from the storage, we can evaluate the temperature evolution
of the storage and its losses.
The validation of such a model is illustrated on Fig. 3a. It has to be
considered as a first attempt. Many points have still to be investigated and
clarified in greater detail. But as a first try, it is very encouraging and
satisfactory.
CONCLUSION
Our measurements show that the solar collectors, the gas powered heat pump
and the underground storage behave as expected.
-------
1588
A simple modeling of the storage, based on the analytical solution of the
heat diffusion equation, in case of cylindrical geometry, gives satisfactory
results.
ACKNOWLEDGMENTS
This work is part of a research activity contracted by the Swiss Federal
Energy Office. We are grateful to the private owner of the industrial
building under study, Mr. Rey, who has offered the necessary facilities for
such a project.
REFERENCES
Guisan, 0., B. Lachal, B. Matthey, A. Mermoud and D. Pahud (1990). A
20'000 m3 solar seasonal heat storage under an industrial building at
Meyrin-Geneva, measurements and calculations. Zeitzschrift filr angewandte
Gewissenschaften, Munchen, 9, 35-55.
Guisan, 0., B. Lachal, B. Matthey, A. Mermoud and D. Pahud (1990). Stockage
saisonnier de 20'000 m3 sous un immeuble industrial a Meyrin-Geneve.
Proceedings OFEN-CH, Okt. 1990, Berne.
Matthey, B. , and C. Roulet (1987). A passive solar industrial building
combined with a 20'000 m3 seasonal storage at Meyrin (Geneva). Proceedings
ICBM 1987, IV, 197-203.
Matthey, B. (1988). Accumulateur saisonnier de chaleur solaire de 20'000 m3
par sondes verticales sous un batiment industriel a Meyrin-Geneve.
Proceedings Jigastock 1988, 2, 561-565.
Pahud, D. (1989). Stockage souterrain de chaleur: Calcul de diffusion,
mesures et comparaisons, Diploma work, Physics, University of Geneva.
-------
1589
A SIMULATION STUDY OF A DUAL SOURCE SOLAR
ASSISTED HEAT PUMP SYSTEM FOR A FAMILY
HOUSE HEATING IN POLAND
D. Chwieduk
Inst, of Fundamental Technological Research
Polish Academy of Sciences, Warsaw, Poland
ABSTRACT
This paper is dealing with an analysis of a dual source solar
assisted heat pump system for family house heating in Poland.
Considered system, which is a hybrid of a typical solar liquid
system and a ground system with vertical heat exchangers, coupled
with a heat pump, is described. Modes of system operation and
simulation model are presented. The calculations are made for
typical Polish heating season COctober-Apri13 for different number
of ground heat exchanger tubes and collector area. Results show
when the bivalent operation mode is possible in Polish climatic
conditions. Performance factors of studied system are prestented.
KEYWORDS
Solar energy; energy from soil; heat pump; solar assisted heat
pump system; dual source system; family house heating.
INTRODUCTION
Poland is located between 49° and 55° N latitudes in a moderate
climate zone. An average annual insolation on horizontal plane is
about 1016 kWh/m with a high share of diffuse radiation and with
1600 hours of solar operation. Nearly 80 percent of the annual
insolation occurs during spring and summer. This obviously limits
the possible utilization of solar energy during heating season and
determines some solutions of solar systems constructions. Solar
energy applications for heating purposes in Poland is the object
of several theoretical and experimental studies. Due to heavy
energy demand during heating season and poor insolation
conditions, studies on solar energy application for heating
purposes in Poland are concentrated on small objects, only e.g.
single family house. To be efficient, solar heating systems ought
to cooperate with another subsystem, e.g. a heat pump. Some types
of solar-heat pump systems have been proposed depended on the
basis of the source of energy to the heat pump evaporator. In an
earlier simulation study CChwieduk, 19903 a typical series solar
assisted heat pump system has been studied in order to answer the
question if such system is useful for P.olish climatic conditions
and building technology. The results are not promoted.
The dual source system is a hybrid of the series and parrallel
-------
1590
system in which the heat pump can draw energy from either the
solar storage or other source. The objective of this paper is to
analyze the dual source solar system, in which ground is used as
alternative source of energy. The main aim of this study is to
optimize the sizing of the major subsystems, particulary the
collector area and the number of ground heat exchanger tubes and
to determine the performance of such system in Polish conditions.
It has been mentioned, that any solar assisted heat pump system
for space heating has not yet been put in practise in Poland.
SYSTEM DESCRIPTION
Figure 1 shows a schematic of a dual source heat pump system
considered in this study. The main components of the system are
solar collectors, an energy storage tank, vertical ground heat
exchangers, a heat pump, space heating system, auxiliary heaters
a residence.
HEAT
PUMP
STORAGE
TANK
HOUSE
AUXILIARY
GROUND HEAT
EXCHANGERS
Fig. 1 Schematic of a dual source heat pump system.
The solar system is the standard liquid heating system. Solar
energy is collected by flat plate solar collectors and stored in a
water storage tank. Energy from the storage can be used directly
or via a heat pump. The earth heat pump installation consists of
earth heat exchangers and a heat pump. Subsoil heat exchangers
operate by indirect mode. Brine circulates through the heat
exchanger and the evaporator. The vertical concentric tubes are
coupled together and connected to the heat pump. Heat pump is a
liquid-to-liquid, vapour compression type. In a dual source system
the heat pump has two evaporators. This allows the heat pump to
use either the collected solar energy from the storage tank or
energy extracted from the ground by the vertical heat exchanger,
depending on which results in a higher COP.
Space heating system is low temperature system suitable for heat
pump and solar applications and consists of liquid-to-air heat
exchangers. Electrical resistance heaters are the auxiliary for
space heating. The energy required to meet the space heating load
in dual source system can be supplied in four modes of system
operation: direct solar heating, solar heating via a heat pump,
ground heating via a heat pump and auxiliary heating.
A residence investigated is a typical Polish two story, single
family house Ctotal floor area of 180 m , volume of 500 m 5. The
construction is standard type of foamed concrete with styrofoam
insulation and with overall loss coefficient of 0.6 W/m K. This
house has 15 percent of its outside wall area as windows, which
are double glazed. Ventilation is done passively.
-------
1591
SIMULATION MODEL
A computer simulation is used to study the performance of the dual
source heating system for a single family house for one typical
heating season Cfrom 1 October to the end of April} near Warsaw.
A measure of the thermal performance of the combined
sol ar-ground-heat pump systems is the fraction CF3 of the total
load CQ 3, that is met by non purchased C"free"3 energy CQ -solar,
Q -ground}, defined as: s
9
F = CQ +Q 3/Q C13
s g L
Several numbers of ground heat exchanger tubes and collector areas
are studied. The size of energy storage tank is kept in direct
proportion to the collector size.
Solar System, Heat Pump and Climate Model.
The simulation calculations of solar system are performed with an
hour computational time-step using a modified model of a
"first-cut" method CEstes, Kahan, 19783. Some assumptions and
sinpli fication are made to reduce the complexity of the
calculations. The solar system modeled is a conventional liquid
solar system. Solar collectors are flat plate type, shielded by
glass covers and insulated on the back side. The collector model
is a zero capacity model originated by the Hottel and Whi 11 i er
C19553 with a constant heat removal factor CFr3, heat- exchanger
efficiency factor CF'3, transmittance - absorbtance productCroO
and heat loss coefficient CUl,3. According to the suggestion given
by de Winter CI 9763 the basic equation of collector is:
Q =A F' CH CtoO - U CT - T)] C23
u c R T L s a
where T is an ambient temperature and H is a solar radiation
incident on the tilted surface of collector and calculated using
the method of Liu and Jordan C19603. The water storage tank is
modeled as being fully mixed and has constant heat loss
coefficient. The governing differential equation of storage is:
dT
M c — = Q - Q - Q C 33
• P dt u L
where M is the mass of water in storage, c is the specific heat
of water? The energy balance C33, accounts p for energy gain from
collectors, energy removed , by the load and energy lost from
storage to the surroundings CQ 3. Q , Q , Q modeled are functions
of a storage temperature and time. uThe energy removed by the load
depends on the mode of the system operation. Storage tank is sized
in accordance with recommendation for liquid solar system. Other
equipments needed for the solar system Cheat exchangers, pumps,
fans, controls3 are not discussed in this paper, because they are
not explicity considered in the calculations.
To simulate the heat pump, a constant temperature difference is
assumed between evaporation and average temperature of low
temperature source Caverage temperature of water in solar tank or
average temperature of brine from ground heat exchangers) and
between condensation and average temperature of water in heating
system. Carnot efficiency is also supposed to be constant.
The climate is described by the ambient temperature and available
insolation Cdirect and diffuse} on horizontal surface. These
-------
1592
weather data are represented by periodical waves with constant
values over each time-step Cone hourD. Daily ambient temperature
modeled is an asymetric sinusoidal function of the mean monthly
temperature, with minimum just before sunrise and maximum in the
afternoon C2.ao p.m.3. The mean hourly radiation sums on
horizontal surface of mean day of each month during the heating
season is averaged over the 20yr records period by Institute of
Geophysic C19833 and represented by hourly rectangular waves.
Vertical Ground Heat Exchangers. Ground System Model.
The calculation method applied is the finite difference method
CFDMD. As it was mentioned earlier, one hour time-step was applied
in the calculations. However, even much shorter time steps
are used for the heat transfer phenomena in the ground, because of
the numerical stability criterion for the FDM. The performance of
a ground coupled heat pump that uses heat exchangers in the ground
depends on the thermal processes in ground and heat exchangers
characteristics. In order to obtain the ground temperature field
around heat exchangers and the amount of heat extracted from the
soil, the following assumptions are made:
- the parameters of the soil: thermal conductivity and diffusivity
are assumed to be uniform in space and constant in time;
- no artificial recharging the soil;
- heat transfer in the soil only by conduction;
- no internal heat transfer between the fluids in the tubes Cdown
flow through an insulated inner tube, up through an outer tube};
- calculations are restricted to one tube of the group;
- moisture migration effect is not accounted for.
Heat extraction from the soil is characterized by the rapid heat
flow in the vicinity of every tube and relatively slow heat flow
in surrounding area in the ground. The ground temperature field
around the tube is calculated with FDM assuming cylindrical
symmetry Cwith an axis through the centre of the tubeD. The FDM
grid is generated automatically according to the given number of
nodal points. The step size is small near vicinity of the tube,
increasing with distance Cthe logarythmic transformation of radius
is used}. Taking into account two dimensional cylindrical model of
conduction, the temperature field in the ground utilizes the
following formula:
dT
dt
dZT 1 ST dZT
drZ r dr
-------
1593
RESULTS AND CONCLUSIONS
Figure 2 shows at the left side monthly values of heating load of
the chosen single family house near Warsaw, and, at the right side,
the height of the bars represents the percentage of the monthly
heating requirements met by free energy: solar and energy
extracted from the soil. The balance of the load is supplied with
purchased energy, i.e. heat pump electrical power input C f or
presented computational example, -there is no need to use auxiliary
heaters).
itkWh]
4000
3000
60
solar
2000
ground
1000
Fig. £ Monthly heating load and monthly fraction of solar
and ground energy providing the heating load.
It is evident that values of heating load are opposite to values
of solar energy. During spring CMarch, April} and fall COctoberO
magnitudes of load and solar energy fraction approximate to each
other and solar energy can be effectively used. At that time,
referring to all obtained results, the effect of increasing
collector area on the amount of solar energy supplied to the
system is evident. But in winter the percentage of solar energy is
very small, independent.of the collector area. This disappointing
performance of the system is due to poor insolation Cheavy
clouds), short duration of solar radiation and low ambient
temperature. From November to the end of February the amount of
energy provided by the solar system is fairly small compared to
amount of energy provided by the ground heat pump system. It can
be said that during winter dual source solar assisted heat pump
system is working as monovalent ground heat pump system. During
the heat extraction from the ground, the soil near the vicinity of
the heat exchangers tubes is cooled down. The temperature field in
the ground near tubes is calculated for whole heating season. At
the end of heating season the distribution of the isothermal lines
for the different number of tubes differ significantly For example
for heating load of the studied family house, when 10 tubes are
considered, the soil, temperature in the area surrounding the tube
decrease Cfrom about 11°0 to S C, but when the ground heat
exchanger consists of 4 tubes, the soil at the vicinity of these
tubes begins to freeze C-1°0. COP coefficient of performance of
the heat pump decreases with time and smaller number of tubes,
when the conditions of heat extraction from the ground are worse.
The influence of number of ground heat exchanger tubes and
collector area on seasonal performance factor (ground and solar!)
is shown in Fig. 3. It is evident, that fraction of solar energy
supplied to the system increases with collector area, but this
fraction is never higher then fraction of ground energy. For
example,for large collector area and small tubes number solar
-------
1594
and ground fraction are about- 33 and 42 percent, respectively.
60
50
40
30
20
10
Fig. 3 The effect of increasing number of ground heat exchangers
tubes and collector area on seasonal performance factors.
Referring to all results, it can be said that dual source solar
assisted heat pump system can be considered as a promising
alternative of solar energy utilization in Polish climatic
conditons. It shuld be pointed out, that heating operation in a
bivalent mode Cwithout auxiliary heating} can be accomplished via
the system considered in this study, when solar system has not
less then 10 m of solar collectors area and ground system
consists of four or more heat exchangers tubes Csituated from 1 to
11 meters below the earth surface, each 10 meters length!).
However, it is noticable, that with increase of heat exchangers
tubes number Cthe fraction of solar energy extracted from the
ground increase) during winter, the time of solar system operation
decreases Cit is more efficient for the heat pump to use the
ground as the low temperature source).
REFERENCES
Chwieduk, D. C1990D. A series solar assisted heat pump system for
famiiy house heating in Poland. In A. A. M. Sayigh CEd.D, Energy
and Environment, Vol. 2 , Pergamonn Press, Oxford, pp. 1036-1040.
Estes, R. C. , and W. Kahan CI978!). Analitical selection of
marketable SAHP systems. In T. N. Veziroglu CEd.), Solar Energy
and Conservation, Vol. 3, Pergamonn Press, New York, pp. 455-475.
Hottel, H. C. , and A. Whillier C1955). Evaluation of flat-plate
solar collector performance. Trans, of the Conference on Solar
Energy, Vol. 2, 74-104.
Institute of Geophysic, C19S3). Solar Radiation. D-20/78/83.
Publications of Institute of Geophysic of Polish Academy of
Sciences, Warszawa.
Liu, B. Y. , and R. C. Jordan CI960). The interrelationship and
characteristic distribution of direct, diffuse and total solar
radiation. Solar Energy, 4C3>.
Winter, F. CI 9763. Heat exchange penalties in double loop solar
water heating system. Solar Energy, 17, 335-337.
-------
1595
10 years operation of a large scale solar system ,
Doho park gymnasium
SABURO TAKAMA P.E.
Scientific airconditioning institute, Tokyo JAPAN
ABSTRACT
During 10 years operation , this active solar system with large array of 1912m2 flat plate collector
proved its durability and good thermal performance for energy conservation.
KEYWORD
Long time operation record; Large collector area; The amount of energy saving; The collection
efficiency; The replacement of system parts;
-a.
Fig. 1 SCHEMATIC DIAGRAM OF THE SYSTEM
INTRODUCTION
Active solar heating & cooling systems are now widely in use, however there are very few
systentwhich are measured during long time operation. The operation of the system , as shown in
Fig.l started on August, 1980 and is continuing until today. This system which has Japan's largest
collector area, is used for multi purpose, space heating & cooling , hot water supply and pool water
-------
1596
heating. The purpose of this paper is to report its thermal performance and maintenance record of
over 10 years operation.
DESCRIPTION of The INSTALLATION
Fig.l shows the simplified diagram of the system , which has two collector array s , 870m2 as
the roof of the arena, 1042m2 as those of the pool house. Each collector consists of copper sheet on
tube absorber (1.0 X 2.0m) coated by black chrome. The heat storage tanks (20m3) are made of
stainless steel and insulated by 150m/m glass wool. The utilization system has six loops ; pool house
heating, pool side floor heating, pool water heating, lobby / office room heating and cooling locker
room heating and hot water supply. A plate type heat exchanger is provided between the collector
loops and the utilization loops. By dividing an auxiliary heat loop form the above loops , even when
a high temperature system such as cooling is operated by auxiliary heat, a low temperature system
can be operated by solar energy . Furthermore, thermal losses, such as excessive heating by
auxiliary heat which often occurred with conventional systems , can be prevented. The auxiliary
system is controlled in accordance with prearranged priority order for effective utilization of solar
energy by the computer .
SYSTEM PERFORMANCE
Monitoring of the system's performance is done by the system computer which is connected to
temperature sensors and calories meters. Fig.2 shows the monthly thermal load of the utilization
system . It is recognized that during 10 year operation the thermal load has almost the same pattern,
the load at summer is approximately 1/3 of winter load. The reason fa' the thermal imbalance comes
from the huge load of the pool house heating and pool water heating. Fig.3 shows the collected solar
energy by the collector. The fluctuation of the monthly collected energy is almost 100~150GJ during
1981 ~ 1990 operation. It is also recognized, the maximum point comes in August, in contrast with
the thermal load.
1981 500
1983
1984
1985 400
1989 300
1 2 3 4 5 6 7 8 9 10 1 1 12
month
1 2 3 4 5 6 7 8 9 10 11 12
month
Fig.2. TOTAL THERMAL LOAD (MONTHLY) Fig.3.MONTHLY COLLECTED SOLAR
ENERGY
Fig.4.shows monthly useful solar energy to the system , which is used by the utilization loops.
-------
1597
The monthly fluctuation is much bigger than those of collected energy , comes from time lag between
solar radiation and the thermal load. It is also recognized that there is the wide difference between
collected energy and useful energy in August. This comes partly from the system design, the
absorption unit (30USRT) is drived only by the poolroof collector array, this causes the overheating
of arena collector system in summer time. Only cooling load and rather small hot water load exit in
August. However the ratio of energy savings by solar is very hii |h in August for the small thermal
load, as shown in Fig.5 .
%
100
80
40
1 2 3 4 5 6 7
month
10 11 12
1 2 3 45 6 78 9 10 11 12
month
FigAMONTHLY USEFUL SOLAR ENERGY Fig.5.RATIO OF ENERGY SAVINGS
TO THE SYSTEM BY SOLAR
Fig.6 shows the system proved good contribution to the energy conservation during 10 years
operation.
GJ
7000
6000
5000 .
•4000
3000 ,
I
i
I
I
I
p-|-
I
I
i
i
Auxiliary heat
I—| ENERGY saving
by SOLAR
1981 1982 1983 1984 1985 1986 1987 1988 1989 1990
Year
Fig.6.THE AMOUNT OF ENERGY SAVINGS BY SOLAR
After 5 & 10 years operation we measured the collection efficiency of a single collector, taken
from the arrays , by the procedure of ASHRAE standard. Fig.7 shows the collection efficiency of 5
-------
1598
year old collector, Fig.8 shows those of 10 year old collector.
COLLECTION EFFICIENCY
1.0
0.8
k
20~C
|40«C
] niOYf]
1 1
Hmean water temperature ccir
i_.
¦
11
»
• Apr.25/1985
0 Apr.28/1985
Apr.29/1985
• May. 8/1985
0 02 0 04 0.08 0.0 8 0.10
(m2hx;/Kcal)
Fig.7. MEASURED PERFORMANCE OF
5 YEAR OLD COLLECTOR
COLLECTION EFFICIENCY
1.0
0.8
0.4
20 r:
riox
!n rl'oVi
Hmean water temperature cx:j-
• Nov. 20/1989
0 Nov.21 /1989
A Nov.24/1989
9 Nov.30/1989
0.02 0.04 0.06 0.0 8 0.10
-SEjis (rr^h-c/Kcal)
Fig.8.MEASURED PERFORMANCE OF
10 YEAR OLD COLLECTOR
There is some reduction of collection efficiency at the high temperature outlet, but the
collector has still enough efficiency .
MAINTENANCE RECORD
Only 2 collectors out 1029 were replaced, except 4 collectors taken out for measurement of
efficiency during 10 years operation. However silicone sealant between cover glass and collector
frame were replaced 2 or 3 times . By checking the strength of silicone rubber hose connecting
ghe collector and pipes , it was found that the tearing strength was obviously lowered. But no
trouble has happened from this during its ; long time operation. The pumps of collector loops
were replaced at 1989, utilization loops at 1990. The main reason of this replacement is the
uneasiness of changing the pump seal.
CONCLUSIONS
After 10 years successful operation, the system is at ill operating satisfactorily, and-has
proved %reat contribution to the energy conservation.
REFERENCE
S.TAKAMA: SOLAR SYSTEM of DOHO PARK GYMNASIUMS, ASHRAE
TRANSACTIONS (88PTI) 1982
-------
2.13 Active Cooling I: Advances in Open Cycle Absorption
-------
-------
1601
SOLAR COOLING AND AIR CONDITIONING PROCESSES USING CHEMICAL
REACTIONS
K. Speidel, H.P. Kleinemeier
Dornier GmbH, Dep. MTE, P.O. Box 1420. 7990 Friedrichshafen
ABSTRACT
Properly designed solar cooling and air conditioning systems using
chemical reactions can become economically attractive because they
don't require supplemental electric energy and water cooling.
Different salts, reacting with ammonia and combined to systems
specifically adapted to the environmental conditions and the solar
collector technology available, are conceivable for solar powered
cooling and air conditioning units with high COPvs. The use of
chemical reactions is becoming now attractive,since improvements
could be achieved concerning the reaction velocity and stability.
KEYWORDS
Chemical reactions; solar cooling units; solar powered air condi-
tioning systems; ammonia-salt processes.
INTRODUCTION
To make solar powered cooling and air conditioning systems economi-
cally attractive means to improve the overall COP and the system
reliability. Chemical reactions between aitimonia and different salts
are the basic technology selected to design improved systems. This
technology has fcefollowing general advantages: ammonia itself has a
high evaporation heat,and the reaction heat with the salt is high
too; furthermore,the energy storage within the salt is high: four
to eight moles of ammonia can be stored in one mole of specifically
selected salts.
SYSTEM CONFIGURATIONS FOR COOLING AND AIR CONDITIONING
Different salts, reacting with ammonia, are available in different
temperature ranges (see Fig. 1) enabling the combination of differ-
ent systems as shown in Figure 2:
System A:
The simplest system (A) consists of a reactor, a condenser and an
evaporator. During heating (H) ammonia vapour is generated, con-
denses and will be stored within the evaporator (E). The condensa-
tion heat (CH) is rejected .During the reverse cycle cooling power
(CP) is produced by evaporation of the liquid. The vapour is react-
ing with the salt, producing heat to be wasted (WH). This simple
system is useful for small cooling units, working periodically dur-
ing one full day and producing ice during the night to keep the
cooling box cold during the day. During daytime, ammonia is generat-
'receding page blank
-------
1602
ed by solar energy and stored.
The best suitable salt for this type of cooling boxes is
Strontiumchloride, which has a high storage capacity of 7 mol ammo-
nia in one mol of Strontiumchloride, operating at theoretical
100
NH3 V
XI
w
O
lb
o>
to
o>
ib
Temperature 1/T (10A3/K)
Fig.:1 Different salts reacting with ammonia
generating temperatures of about 80-100 Degree C, even at high am-
bient temperatures; ice production as a cold storage is possible
even at high night temperatures of 30 Degree C,when the absorption
takes place. The system requires the application of vacuum tube
collectors, which are becoming more and more attractive on the mar-
ket.
The overall system efficiency caild be practically measured with a
functional model to about 15%, defined as cooling to solar energy.
The theoretical COP of this periodic cycle is about 60%.
System B:
System B consists of two reactors,e.g. ZnCl2 or SrC12 on the hot
side and BaC12 on the cold side and can be used for simple periodic
air conditioning systems. The generation temperature is about 120
to 130 Degree C where vacuum tube collectors are operating well at
ambient temperatures of about 40 Degree C. The cooling temperature
is in the range of about 5-10 Degree C. Air conditioning is espe-
cially required during daytime, therefore four reactors will be ac-
tive simultaneously, two of them with heating and absorbing,and the
other two to generate cooling power. Using these four reactors, a
quasi-stationary cooling operation can be obtained. The theoretical
COP of the process itself is 1, because the complete reaction ener-
-------
1603
gy can be used for cooling. The practical COP, defined as cooling
energy to solar radiation on the installed collector field is ex-
pected do be about 30%. Practical experiences with such systems are
not yet available; specific test are in preparation.
CH
WH
CP
WH
WH
CP
System A System B
1 CH
WvV>
T CH
R2
E2
- .. %
WH
CP 2
System C
System D
Fig 2: System configurations for cooling and air conditioning
System C
The proposed System C combines System A, operating at higher tem-
peratures with a second System type A, operating at lower one s,
connected together by heat transfer. While reactor R1 (e.g. filled
with NiC12) is heated up to about 27 0 Degree C to generate ammonia
vapour which is condensed and stored in evaporator El, reactor R2
absorbs ammonia from evaporator E2;thus producing cooling energy.
The reverse process is afterwards active,producing cooling energy
in El in connection with Rl. During this process;heat in the tem-
perature range of about 160 Degree C is transferred to reactor R2,
which is probably filled with SrC12, to generate ammonia. This
cooling process is also quasi-stationary. The theoretical COP of
the process is about 1,2 (cooling to heat energy), the practical
COP,inclusive the part of solar trough collectors,is expected to be
about 0,4-0,5 (cooling to solar energy on the installed collector
field). The use of concentrating collectors in the mentioned tem-
perature range is necessary. The systemr seems to be favourably ap-
plied for cooling containers or cooling houses with high cooling
-------
1604
energy demand.
System D
This proposed configuration is similar to system C, but uses four
reactors. Reactor R1 may operate with NiC12 in connection with R2,
filled with Ba.C12, while reactor R3 can be filled with SrC12 com-
bined with resictor R4, filled with BaC12,too. During the heating
phase of reactor R1 (decomposition of NH3 within R1 and reaction
with NH3 in R2) reactor R4 produces cooling energy (CP2); the reac-
tion heat of reactor R3 at low temperatures is wasted (WH). In the
reverse operation mode, cooling is produced in reactor R2; the cor-
responding reaction heat at higher temperature level during this
period can be used to heat reactor R2.
Following Fig.l this system is preferably applied for air condi-
tioning purposes where cooling temperatures in the range of 5 to 10
Degree C are required. The attraction of this configuration is the
high COP, which is theoretically 2. In practice, an overall COP of
0.8 is expectable, defined as cooling energy gained to solar radia-
tion input on the installed collector field. The system can also be
regarded as a quasi-stationary one, using preferably trough col-
lectors in the same temperature range as system C.
IMPROVEMENTS OF REACTION KINETIC,MASS TRANSFER AND STABILITY
The economic application of the selected technology depends mainly
on the improvement of the reaction kinetic. The power output and
the desired coefficient of performance can only be achieve^ practi-
cally, if the reaction speed is quick enough (according the specif -
ic demand of the system), and if the mass transfer and the stabili-
ty of the reaction is guaranteed.
The theoretical power output is a function of the specific salt
content, the selection of the salt(s), the selection of the process
(use of evaporation heat or process heat) and the cycle frequency.
The selection of the salt determines principally the process param-
eters concerning the possible storage of energy and the operating
temperature ranges.
After having determined the process from the temperature ranges by
selecting the salt(s); two important parameters are influencing the
system performance: the heat transfer (described by the thermal
conductivity and the heat transfer coefficient) and the mass the
transfer coefficient. These factors have to be improved.
The salt itself has very low conductivity values of about X = 0,1-
0,3 W/mK and low heat transfer coefficient of about a = 10-20
w/m2K, In practice,it is well known, that the mass transfer, using
the salt without any additives is week and can partially or totally
block the reaction after a certain amount of cycles. The stability
of the cycle has therefore also to be improved.
This is the main reason why this technology could not yet progress
despite the general advantages. Therefore, great activities have
been concentrated to this subject the last years. The results ob-
tained so far are very encouraging:
It could be demonstrated, that specifically pretreated graphite,
handled and prepared with the salt can improve the situation con-
siderably. Graphite is temperature stable,does not react with ammo-
nia and has a high heat conductivity; besides these properties,
graphite is improving the vapour flow and thus contributes much to
ameliorate the mass transfer. Furthermore,the salts can be pressed
-------
1605
in order to enlarge the storage capacity and the heat transfer and
to diminish the specific reaction volume without effecting the mass
transfer in specific limits.
It could be shown that the heat conductivity can be varied between
X = 4 to 25 W/mK, while the heat transfer coefficient is increased
to about a = 500 to 1200 w/m2K. Tests have been performed with good
results concerning the mass transfer. Varying the preparation of
the graphite-salt-mixure, the reaction can be adapted to the process
requirements. In case of high reaction speed with high energy den-
sities, the amount of graphite has to be increased in order to in-
crease the heat flux. If the heat flux is lower,the content of the
salt can be increased, the volumetric energy content is thus in-
creased. Test series have been performed and are continued showing
that the process stability even at high cycling rates can be guar-
anteed ( 5000 cycles with CaCl2 without effects on the reaction ki-
netic; further experiments with other salts are in progress ).
Similar test have also been performed without an additive to the
salt (SrCl2); the reaction degradation was considerable: about 10%
after 500 cycles.
TEST RESULTS WITH SOLAR COOLING UNIT AND IMPROVED REACTOR
A solar powered test facility (working on the principle of System A
in Fig. 2) has been built in 1988 and is running continuously. The
cooing unit has a volume of 200 liters and was designed to operate
without water cooling and supplemental electrical supply. The unit
uses Strontiumchloride without additives and Ammonia and is de-
signed for tropical countries. The design was also taking into ac-
count the possible reduction of the reaction kinetic. Following re-
sults could be obtained:
- the COP of the system is in the range of about 15 % (cooling en-
ergy to solar input)
- the thermodynamic properties of the working principle could be
demonstrated
- to get a marketable product,the production costs have to be re-
duced.
In order to reduce the production costs further,activities are in
progress to improve also the process reliability, within a test rig,
the main components are under investigation, especially the reactor
using the above-mentioned preparation method of the salt. Following
improvements are envisaged: reduction of the reactor volume to 1/3
compared to a reactor without additives, reduction of the operating
temperature of the collector by improvement of the heat flux,and
finally the stability of the process which enables a design accord-
ing the theoretical energy balances (desorption and absorption near
100%).
Typical test results are demonstrated in Fig.3, showing a desorp-
tion cycle (cycle No.60). The salt temperature follows strictly the
heating temperature with a temperature difference of about 4-5
Degree C. This temperature difference includes the temperature of a
liquid to be vaporised and condensed in order to transfer the heat
from the source to the reactor wall. The real temperature between
the wall an the salt is about 1-2 Degree C.
After reaching the theoretical generation temperature, ammonia
vapour is produced and condensed. The further temperature increase
of the salt and the theoretical generation temperature depends on
the condensation temperature of the ammonia. During the desorption
phase, which ends with the crossing of the temperature lines of the
-------
1606
salt and the theoretical generation temperature after 5 1/2 hours,
the temperature difference between the salt and the heat source is
about 10 Degree C.
10C
heating temperature
iheOrettcjeneiaftonji img7
:ure
lenerated amnr >nia
¦a
0
3
6
1
2
4
5
Time (h)
Fig.3: Generation of ammonia
The temperature difference between the reactor wall and the salt is
about 5-6 Degree C. Taking into account actual energy balances,
the conductivity of the salt is about 6-9 W/mK. For this specific
application,the salt content is very high in order to reduce the
reactor volume.
REFERENCES
-Systemes de gestion de l'energie thermique bases sur des reactions
solide-gas. G. Crozat, B. Spinner, M. Amouroux, "Recent progrds en
c(6nie des proc6d4s" Volume 2-1988 No. 5 Diffusion: Lavoisier-
Technique et Documentation Paris.
Autarke Kiihlanlage nach dem Absorptionsprinzip fur
Medikamentenkuhlung und Kiihluhg von Lefcrensmittel,
Forschungsvorhaben 0328714 A, K. Speidel, H.-P. Kleinemeier, Dez.
1989
-------
1607
SIMULATION AND ANALYSIS OF OPEN CYCLE ABSORPTION
SYSTEMS FOR SOLAR COOLING
I. Haim, G. Grossman and S. Shavit
Faculty of Mechanical Engineering Technion - Israel Institute of Technology
Haifa 32000, ISRAEL
ABSTRACT
This paper describes a performance analysis of two open-cycle absorption systems for solar
cooling. The working fluid is LiCl-H20, where the water desorbs into the atmosphere and is
replaceable. Both systems comprise a closed absorber and evaporator as in conventional, single-
stage absorption chillers. The open part of the cycle is in the regenerator, used to reconcentrate the
absorbent solution by means of solar energy. One of the systems under study has employed direct
regeneration in a regenerating collector, exposing the solution simultaneously to the sun and to a
stream of air. The other has employed indirect regeneration by contacting the solution with air
heated elsewhere in a flat-plate collector. The analysis was performed using a code developed for
modular simulation of absorption systems under varying cycle configurations and with different
working fluids. The paper presents the performance curves obtained for both systems. Results
indicate a definite performance advantage of the direct-regeneration system over the indirect-
regeneration one.
KEYWORDS
Open-cycle; absorption; solar cooling; solar airconditioning; solar regeneration; liquid desiccant.
INTRODUCTION
Research has been underway in the past few years on advanced absorption cycles and non-
conventional working fluids which can provide significantly greater efficiency than do the
conventional single-effect absorption systems with water-ammonia and lithium bromide-water. In
parallel with the development of the closed-cycle systems, the past few years saw renewed interest
in open-cycle absorption, particularly for solar applications (Grossman and Johannsen, 1981). The
main thrust in research on the novel cycle systems has been toward the achievement of a better
coefficient of performance by taking advantage of high temperature heat sources as provided, for
example, by natural gas. In contrast, open-cycle systems have been developed for use with a low
temperature heat source, and their major advantage is in their potential for reduction in cost. Low
cost regenerating collectors may be built to replace the generator, condenser and fluid-heating solar
collectors of a closed-cycle system (Johannsen and Grossman, 1983). These collectors, which
may be incorporated in an inclined roof, provide for direct absorption of solar radiation in the
regenerated solution and direct contact mass transfer to the air; no heat-exchange surfaces are
required. Another advantage is that in dry air some regeneration may be obtained "free", in addition
to the solar contribution (Lavan et al., 1982).
This paper describes a performance analysis of two open-cycle absorption systems for solar
cooling. The working fluid is LiCl - H2O, where the water desorbs into the atmosphere and is
replaceable. Both systems comprise a closed absorber and evaporator as in conventional, single-
stage absorption chillers. The open part of the cycle is in the regenerator, used to reconcentrate the
absorbent solution by means of solar energy. One of the systems under study (Wood, 1986) has
-------
1608
employed direct regeneration in a regenerating collector, exposing the solution simultaneously to
the sun and to a stream of air. The other (Lenz et al., 1983, 1985) has employed indirect
regeneration by contacting the solution with air heated elsewhere in a flat-plate collector.
The analysis was performed using a code developed for modular simulation of absorption systems
under varying cycle configurations and with different working fluids (Grossman et al., 1987).
The code, originally written for closed absorption systems, has been modified into an enhanced
version in order to accommodate the features of the open-cycle systems, including the regenerating
collector and the air-solution contactor. Based on specified design features, the code calculates the
operating parameters in each system for a variety of conditions. The present study has attempted to
predict the performance of both systems, and of their major solar components.
DESCRIPTION OF OPEN ABSORPTION SYSTEMS
Both systems considered in this study have been under development for several years, striving to
combine the well-known advantages of the absorption cycle with open-cycle regeneration utilizing
solar energy. Both are single-effect chillers using lithium chloride-water as their working fluid
pair. One system, developed at Arizona State University (Wood, 1986), has employed direct
regeneration of the absorbent by exposing it simultaneously to solar radiation and to a stream of air
removing the desorbed moisture, in a solar collector designed specially for this purpose. The other
system, developed at Colorado State University (Lenz et al., 1983, 1985), has employed indirect
regeneration of the absorbent in an air-solution contactor, using air preheated in a rather
conventional flat-plate solar collector. The two systems will be referred to in this paper as the Direct
Regeneration (DR) and the Indirect Regeneration (IR) systems, respectively.
Figure 1 describes schematically the DR system (Wood, 1986). It consists of five main units: An
evaporator, an absorber, a regenerating collector, a recuperative heat exchanger and an expansion
valve. Each unit is marked by a number in a circle, as shown. Numbers with no circle indicate state
points. Liquid refrigerant (water) at state point 3 is supplied through the expansion valve to the
evaporator, where its evaporation is used to chill an external stream of water entering at state 1 and
leaving at 2. The liquid refrigerant entering the evaporator at state 12 reaches its evaporation
temperature (state 4) and leaves as vapor at state 5. The vapor is absorbed in the absorber by a
strong lithium chloride-water solution entering at state 6, reaching equilibrium at state 10 and
leaving weaker at state 7. The heat of absorption is rejected to a stream of cooling water entering at
state 8 and leaving at state 9. The weak solution is pumped through the heat exchanger to the
regenerating collector where it is brought into contact with the air stream 14-15 and exposed to
solar radiation for regeneration. The regenerated solution at state 13 is returned to the absorber
through the heat exchanger.
Figure 2 describes schematically the IR system (Lenz et al., 1983, 1985). The resemblance to the
DR system is quite evident in all respects but the solution regeneration. The system consists of six
main units, marked as before by numbers in a circle, whereas numbers without a circle indicate
state points. Liquid refrigerant (water) at state point 3 is supplied through the expansion valve to
the evaporator and evaporates, cooling the external stream of water entering at state 1 and leaving at
2. As in the DR system, the liquid refrigerant enters the evaporator at state 12, reaches its
evaporation temperature at state 4 and leaves as vapor at state 5. The vapor is absorbed in the
absorber by a strong lithium chbride-water solution entering at state 6, reaching equilibrium at state
10 and leaving weak at state 7. The heat of absorption is rejected to cooling water entering at state 8
and leaving at state 9. The weak solution is pumped through the heat exchanger to the air-solution
contactor, where it is brought into contact with a stream of air preheated in the solar collector. This
hot air provides both the energy and the sink for the moisture desorbed from the absorbent solution
in the regeneration process.
The diagrams in Figures 1 and 2 have been drawn in terms of components recognizable by the
simulation code (Grossman et al., 1987). As mentioned earlier, the code employs a modular
algorithm which makes it possible to vary the cycle configuration of the simulated system. This is
done by employing unit subroutines in the code, each representing a different component of the
system. The main program calls the unit subroutines and links them together in a form
corresponding to the user specifications. Thus, it was possible to employ the same code to simulate
both the DR and the IR system on the same basis.
-------
1609
17
SKY
AMBIENT
EXHAUST AIR
COLLECTOR
EXHAUST
AIR
SKY
22
INLET
AIR
HEAT
EXCHANGER
AIR SOLUTION
CONTACTOR
REFRIGERANT
MAKE UP
—OUTLET
CHILLED
WATER
^OUTLET
CHILLED
WATER
OUTLET —
COOLING
INLET
ABSOR8ER
EVAPORATOR
•INLET
INLET-
EVAPORATOR
Fig,, 1. Schematic Description oftheDR System. Fig* 2. Schematic Description of the IR System.
Three unit subroutines had to be added to those already in the code to make possible the simulation
of open-cycle solar cooling systems. These three subroutines represent an air heater, an air-
solution contactor and a regenerating collector. The subroutines are generally based on the same
physical laws applied in the subroutines for the closed-cycle components - conservation laws,
thermodynamic equilibrium and transfer laws. An added complexity in the air-solution contactor
and regenerating collector is due to the presence of three working fluids - absorbent, water and air.
This leads, under certain conditions, to inversion in the combined heat-and-mass-transfer process
within the unit, namely a reversal in the direction of the heat or mass flux. The use of a logarithmic
mean difference of temperature or concentration to account for the overall transfer process in the
unit is not possible under these conditions. The unit subroutines for these components were written
in such a way that each component may be broken down into several parts, inside which the
changes in temperature and concentration are small. Within each part, average quantities may be
used. Depending on the conditions, the subroutine may be used to reDresent either a part of, or the
overall component. In the former case, the subroutine is called a number of times equal to the
number of parts, as chosen by the user. The calls are made in such a form that the parts are linked
together to form the total unit. This approach, compatible with the program structure, has proved to
be both flexible and efficient.
PERFORMANCE PREDICTION OF THE DR SYSTEM
The DR system, described in Figure 1, is an open-cycle, single-effect absorption chiller, where
solution regeneration is performed by a regenerating collector. Table 1 describes the design
condition selected for the system. The approach in all the simulation work has been to establish a
design point and calculate all the unknown system parameters for it, then perform a sensitivity
analysis by varying the input parameters one at a time.
The sizes of the components and the various flowrates were selected for the system to produce at
the design point approximately 1 TR of cooling. The chilled water outlet temperature is typical for
air conditioning applications. Air temperature and humidity and other ambient conditions were
taken at the design point for a typical summer day. The solar radiation figure refers to the amount
absorbed in the liquid solution covering the entire collector area of 10 m2, after optical losses in
the passage through the glazing etc. have been deducted. Evaporator, absorber and recuperative
-------
1610
heat exchanger UA's were selected based on a heat transfer effectiveness of 0.85 for those units at
the design point.
The operation of the system has been characterized in terms of two criteria: its coefficient of
performance (COP) defined as the ratio of the cooling produced by the evaporator to the solar
heat input; and its relative cooling capacity defined as the ratio of the actual cooling capacity to that
at the design point. These two measures of performance have been calculated as functions of the
ambient conditions, air and solution flowrates and operating temperatures.
Figure 3 shows the COP and relative capacity as functions of the solar irradiation on the solution
for different cooling water temperatures. It is evident that the capacity increases more than linearly
with the solar input. The increase in COP is more moderate. As expected, both COP and capacity
increase with decreasing cooling water temperature. There exists a minimum solar radiation, for
each cooling water temperature, below which the COP and capacity are zero and the system does
not perform at all. This is a well-known phenomenon in closed absorption systems related to the
values of the high and low concentrations (Grossman and Johannsen, 1981).
Table 1: Design Conditions for the PR system
1.
2.
4.
Chilled water:
Flowrate: 20.0 lbs/min. (0.15 kg/sec.)
Outlet temperature: 45°F (7.2°C)
Cooling water:
Flowrate: 20.0 lbs/min. (0.15 kg/sec.)
Inlet temperature: 85°F (29.4°C)
Refrigerant make-up:
Temperature: 85°F (29.4°C)
Weak Solution:
Flowrate: 2.5 lbs/min. (0.019 kg/sec.)
Air inlet to regenerating collector:
Temperature: 85°F (29.4°C)
Humidity: 0.01 kg I^O/kg dry air
Flowrate: 5.0 lbs/min. (0.038 kg/sec.)
Ambient Conditions:
Ambient temperature: 85°F (29.4°C)
Sky temperature: 80°F (26.7°C)
Solar radiation (net amount absorbed) 450 BTU/min.
Transfer coefficients and areas:
Evaporator UA: 37.52 BTU/min°F
Absorber UA: 75.56 BTU/min°F
Recuperative heat exchanger UA: 6.60 BTU/min°F
Regenerating collector area: 10m2
COOLING
WATER
TEMP.
0 100 200 300 400"
TOTAL SOLAR IRRAOIATlON ON SOLUTION f8^]
LminJ
q. O-P
COOLING
WATER
TEMP.
100 200 300
TOTAL SOLAR IRRADIATION ON SOLUTION
Fig # , 3; COP and Relative Capacity (DR).
The COP and relative capacity have been calculated also as functions of the weak solution flowrate
(which may be varied by controlling the solution pump) for different values of the output chilled
water temperature. It was found that there exists an optimum solution flowrate: with too low a
flowrate the capability of canying the refrigerant through the system drops, and too high a flowrate
leads to excessive circulation losses. The system operates better with higher chilled water
temperatures, as expected.
-------
1611
PERFORMANCE PREDICTION OF THEIR SYSTEM
The IR system is an open-cycle, single-effect absorption chiller, where solution regeneration is
performed in an air-solution contactor. The regeneration heat is supplied to the latter by air heated
in solar collectors. Table 2 describes the design condition selected for the system. The sizes of the
evaporator, absorber and recuperative heat exchanger and their flowrates were selected equal to
those of the DR system, for comparison. The design cooling water and chilled water temperatures
and the ambient conditions were also equal to those of the DR design point. An air-heating
collector area of 10 m2 was selected, as for the regenerating collector of the DR system; the air-
solution contactor was a proportional fraction of the one actually used in the IR experimental unit
(Lenz, et al„ 1983, 1985).
As before, the operation of the system has been characterized in terms of two criteria: its coefficient
of performance and its relative cooling capacity. These two measures of performance have been
calculated as functions of the ambient conditions, air and solution flowrates and operating
temperatures.
Figure 4 shows the COP and relative capacity as functions of the solar irradiation on the solution
for different cooling water temperatures. It is evident that the capacity increases more than linearly
with the solar input. The increase in COP is more moderate. As expected, both COP and
capacity increase with decreasing cooling water temperature. As in the DR system, there exists a
minimum solar radiation, for each cooling water temperature, below which the COP and capacity
are zero and the system does not perform at all.
Table 2: Design Conditions for the IR system
1.
2.
4.
5.
6.
7.
Chilled water:
Flowrate: 20.0 lbs/min. (0.15 kg/sec.)
Outlet temperature: 45°F (7.2°C)
Cooling water:
Flowrate: 20.0 lbs/min. (0.15 kg/sec.)
Inlet temperature: 85°F (29.4°C)
Refrigerant make-up:
Temperature: 85°F (29.4°C)
Weak Solution:
Flowrate: 2.5 lbs/min. (0.019 kg/sec.)
Air inlet to air-heating collector:
Temperature: 85°F (29.4°C)
Humidity: 0.01 kg H20/kg dry air
Flowrate: 10.0 lbs/min (0.076 kg/sec.)
COOL NG
WATER
TEMP.
100 200 300 400 500
TOTAL SOLAR IRRADIATION ON PLATE [siu/min]
< 0-8
<= 0-4
Ambient Conditions:
Ambient temperature: 85°F (29.4°C) Q
Sky temperature: 80°F (26.7°C)
Solar radiation (net amount absorbed) 450 BTU/min.
Transfer coefficients and areas:
Evaporator UA: 37.52 BTU/min°F
Absorber UA: 75.56 BTU/min°F
Recuperative heat exchanger UA: 6.60 BTU/min°F
Solar collector area: 10m2
COOLING
WATER
TEMP
M
100 200 300 400 500
TOTAL SOLAR IRRA0IATI0N ON PLATE [stu/min]
Fig, 4: COP and Relative Capacity (IR).
The effects of weak solution flowrate, chilled water temperature, air inlet and ambient conditions
are all similar to those observed in the simulation of the DR system. Quantitatively, the COP
reached in the IR system is considerably smaller than in the DR system under the same
-------
1612
conditions, by a factor of two or more. This becomes evident by comparing Figure 4 with 3. The
primary reason for this, as explained before, is the direct absorption of solar heat in the
solution in the DR system vs. the indirect transfer of solar heat to the solution through air in the
IR system. The latter method leads to considerably greater losses.
CONCLUSION
The modular computer simulation code (Grossman et al., 1987) has been modified and adapted to
simulate open-cycle solar-powered systems. For this purpose, unit subroutines for three
components - air heating collector, air-solution contactor and regenerating collector - were added
to the code, and its property database was expanded. The code was applied first to study the
behavior of the individual components and then to evaluate the performance of the DR and IR
open cycle, air conditioning absorption systems.
The approach taken in all the simulation work was to describe the system in terms of the units
recognizable by the code, then define a design point and carry out a sensitivity analysis, varying
one of the design parameters at the time while fixing all the others.
Once the behavior of the components was understood, a simulation of the complete DR and IR
systems was carried out. In order to obtain a fair comparison, the sizes (UA's) of the absorber and
recuperative heat exchangers in both systems were taken equal; the same chilled water outlet
temperature and flowrate, cooling water inlet temperature and flowrate, weak solution flowrate and
ambient conditions were taken for both systems; finally, the same solar collector area and total
solar radiation absorbed was assumed in both cases. The simulation yielded values of two
measures of performance - COP and relative capacity - as functions of the above parametes. The
behavior of both systems was observed to be qualitatively similar; however, COP values in the IR
system were considerably lower than those in the DR one. Again, the primary reason for this is
that the DR system employs direct heating of the solution during regeneration whereby solar heat is
provided where required, whereas the IR system heats air which then has to transfer the heat to the
solution, leading to greater losses.
REFERENCES
1. G. Grossman, K. Gommed and D. Gadoth: A Computer Model for Simulation of
Absorption Systems in Flexible and Modular Form. ASHRAE Transactions. 22, part 2,
2389 - 2428, (1987).
2. G.Grossman and A. Johannsen: Solar Cooling and Air Conditioning. Prop-ess in Energy
and Combustion Science. 2, 185 - 228, (1981).
3. A. Johannsen and G. Grossman: Performance Simulation of Regenerating Type Solar Colle-
ctors. Solar Energy Journal. 30. 87-92 (1983).
4. Z. Lavan, J.B. Meunier and W.M. Worek: Second Law Analysis of Desiccant
Cooling Systems. ASME Journal of Solar Energy Engineering. 104. 229 - 236, (1982).
5. T.G. Lenz, G.O.G. Lof, R. Iyer and J. Wenger: Open Cycle Lithium Chloride Cooling
System. Final Reports on DOE Contract for 1982/83, 1983/84, Solar Energy Applications
Laboratory, Colorado State University (1983 & 1985).
6. B. D. Wood: Open Cycle Absorption Solar Cooling. Final Report, No. ERC-R-86020D,
College of Engineering and Applied Sciences, Arizona State University, (1986).
-------
1613
ENERGY BALANCE AND MASS TRANSFER STUDIES FOR A
LIQUID DESICCANT BASED SOLAR COOLING SYSTEM
Terry G. Lenz, Shailesh V. Potnis
Solar Energy Applications Laboratory
Colorado State University
Fort Collins, CO 80523
ABSTRACT
A solar cooling unit using lithium bromide solution as the liquid desiccant
was studied to develop a better understanding of the process of mass transfer
in the system. Energy balance runs were carried out in both coupled and decoupled
modes and the precautions suggested by these runs were observed during the
experiments conducted to obtain the relative magnitudes of mass transfer
resistances. The error in the energy balance for the decoupled mode was 5-10%
while that for the coupled mode was 4-14%. A slope of 0.81 obtained from the plot
of solution flow rate versus condensation rate suggested that the dehumidifier
was operated under well mixed (near turbulent) conditions and a Wilson plot based
on these results indicated that the gas phase mass transfer resistance was
negligible for air/solution contact in the packed bed dehumidifier.
KEYWORDS
Solar cooling, liquid desiccant, mass transfer, energy balance, dehumidifier.
NOMENCLATURE
H = Enthalpy (J/kg)
Aha = (Hout- Hln) for air across
dehumidifier(W)
Ka = Overall mass transfer
coefficient
ML = mass of the water condensed
(kg/s)
Subscripts
D - Dehumidifier EA - Tube side of the economizer
EL - Shell side of the economizer G - Gas
L- Liquid La - Latent heat
R - Regenerator
AH = (Houi-Hin) for the solution(W)
Ahw = (Hout- Hin) for water across the
heat exchanger(W)
k = Mass transfer coefficient
WP = Pump power(W)
-------
1614
INTRODUCTION
The solar cooling of buildings seems to be one of the most attractive
applications of solar energy, particularly for combined heating and cooling
installations. Desiccant cooling systems using solar energy hold much promise
for humid climates where the solar heating equipment can provide useful energy
for cooling, heating and domestic hot water supply for the entire year. In these
systems air is dehumidified to reduce the latent heat load and the sensible load
can be reduced by subsequently passing it through various heat exchangers and
evaporative coolers. The dilute solution from the dehumidifier is concentrated
in a regenerator by solar heated air.
The mass transfer rates achieved in the regenerator and the dehumidifier
determine the capacity of such a system. A considerable amount of research is
being conducted to understand the process of mass transfer in these systems.
Burdukov et al.(1980)carried out experiments to obtain mass transfer coefficients
for absorption of steam in lithium bromide solution in a vacuum chiller
environment. They found that the mass transfer coefficients are a strong function
of percentage of air in the steam. Collier (1979)modelled an open cycle
absorption refrigeration system for five cities namely, Phoenix,Miami,
Albuquerque, New York and Dallas. The mass transfer coefficients were obtained
using an analogy to heat transfer and the COP's ranged from 0.09 to 0.45 for
various conditions. Lof et al.(1984) studied reconcentration of a lithium
chloride solution in an open cycle absorption chiller by passing solar heated
air through a packed column. They found that mass transfer coefficients show
considerable variability and generally lower values than would be predicted by
use of the heat/mass transfer ratio for air and pure water.
Thus, it is imperative to have a clear understanding of the mass transfer
resistances offered by each phase in various portions of an overall cooling
system, which is the primary objective of the present research. Specifically in
the present studies, the variation of condensation rate with the variation of
solution flow rate was studied to obtain a Wilson plot and to gain insight
regarding the mass transfer resistance in an atmospheric pressure dehumidifier.
EXPERIMENTAL APPARATUS
Figure 1 shows a schematic diagram of the experimental system employed in
our studies. The dehumidifier is an 81-cm - diameter,200-cm high fiberglass
tower. Tripack No.1/2 polyethylene spheres are used as the packing material and
the height of the packed bed is 28 cm. The solution is distributed on top of the
bed by a manifold of three spray nozzles. An electric steam humidifier introduces
moisture into the air before it enters the dehumidifier. A mist eliminator is
used to minimize the liquid entrainment before the air exits the dehumidifier.
In a decoupled mode, the solution exiting the dehumidifier is cooled in a cooling
unit before it is sent back to the top of the dehumidifier. In a coupled mode,
this solution is sent to the regenerator through the economizer where it is
heated by heat exchange with the hot solution from the regenerator. The
regenerator is similar in dimension and construction to the dehumidifier unit.
Hot air to the regenerator is provided by a flat plate solar collector array
having a total area of 55.7 m2. Auxiliary heat is also supplied by three electric
strip heaters to increase the air temperature and maintain a high regeneration
rate under varying ambient conditions.
-------
1615
a,f OP-0«v» Point
—* — Cooling Water F-Flow
C-Concentratioft
OP
OP
OP
Heaters
Economizer
Cooling Unit
Fig. 1. Experimental open-cycle desiccant system.
The hot solution exiting the regenerator is cooled in the economizer by
exchanging heat with the solution from the dehumidifier. After passing through
the economizer, the warm solution is further cooled by water in a four pass shell
and tube heat exchanger. A cooling tower with a capacity of 28 kW supplies water
to the heat exchanger.
Figure 1 also shows the instrumentation for the experimental apparatus.
Air temperatures are measured at the inlets and outlets of the towers. The mass
flow rate of the air is measured in the inlet duct with a pitot rake electronic
manometer combination. Solution temperatures are measured at the inlets and
outlets of the towers, the economizer and the cooling unit heat exchanger. The
solution flow rate is measured with rotameters in the inlet lines. The
concentration of the solution is measured by a Paar digital density meter.
RESULTS AND DISCUSSION
Enerov Balance
It is essential to estimate the enthalpy changes across each unit of the
system to obtain the losses and identify those parts of the system which require
more understanding and improvement. Experiments were thus designed and conducted
to obtain energy balances in both coupled and decoupled modes of operation as
described above. Improvements suggested by these energy balance runs were
incorporated in the experiments used to conduct mass transfer studies.
Enthalpy change for air in each tower can be obtained from the measurements
of air flow rate, temperatures and humidities. The enthalpy change for the
solution can be obtained from the measurement of solution flow rate, temperatures
and concentration. An energy balance across the entire system can then be
performed by comparing these enthalpy changes. However, this method is dependent
on a difficult humidity measurement. If sufficient mixing length is not provided
after the introduction of moisture, the humidity measurement can depend on
measurement location and the error in the energy balance can be as high as
60%(Flaherty,1989). Therefore it is essential to base the energy balance on more
-------
1616
reliably measured variables.
The lithium bromide solution undergoes both a sensible heat exchange with
the incoming air and a latent heat effect due to the absorption of moisture from
the air in the dehumidifier. In the decoupled mode, the net energy increase for
the liquid due to these heat effects is lost to the cooling water in the cooling
unit. Thus, an energy balance can be carried out by comparing the energy increase
for the solution in the dehumidifier and its energy decrease in the cooling unit.
The energy balance equation for the dehumidifier in the decoupled mode is
AHd + WP -Ahw + losses =0 (1)
In a coupled mode the energy balance can be performed by comparing the heat
effects across the economizer and the cooling unit. The energy balance equation
for a coupled mode is
AH„ + AH^ + AHr + WP + AHel - Ahw+ losses = 0 (2)
The losses from the towers are found to be negligible(Flaherty, 1989). All the
enthalpy changes used in these equations are obtained at steady state. The
criteria used to identify a steady state are steady temperatures at various
points in the system(+0.1°C), steady flow rates(± 0.01 kg/s) and less than 10%
energy balance error across the cooling unit alone. The last criteria introduces
some dependability on the ambient conditions. High ambient temperature can
increase the temperature of the water in the cooling tower which in turn can
increase the temperature of the solution above 25°C, i.e. above the optimum
temperature for the dehumidifier(Flaherty,1989). However, the effect of high
ambient temperature can be controlled by using a cooling tower fan. Taking this
precaution and using these criteria of steady state, the error in the energy
balance for the dehumidifier was 5-10% and in the coupled mode it was 4-14%.
Mass Transfer Studies
The rate of moisture condensation in the dehumidifier depends on various
parameters such as air flow rate, solution concentration, height of the packed
bed, air and solution inlet temperatures, and the solution flow rate. All these
variables except the solution flow rate were maintained constant (with solution
concentration of 51% by weight) in a decoupled mode to study the dependency of
the condensation rate on solution flow rate. The enthalpy change in solution due
to latent heat of condensation was obtained from the enthalpy changes in the
cooling water and air, and the pump power.
AHLa = Ahw + Aha - WP (3)
The mass of moisture condensed per unit time was obtained from
Ml = AHLa / Hv (4)
A plot of Sherwood number versus Reynolds' number is usually used to
correlate various flow parameters. However, since all other parameters except
the flow rate were kept constant, a plot of condensation rate versus solution
flow rate was constructed to represent the conventional plot. Figure 2 shows that
this plot is a straight line with a slope of 0.81 (R2 of 0.91). This value of the
slope suggests that the dehumidifier Is operating under well mixed, near
-------
1617
turbulent conditions for liquid flow.
The overall mass transfer coefficient and the individual phase mass
transfer coefficients can be related by the equation,
1 = !_+.!_ (5)
Slope = 0.81
Solution Row Ratetko/a)
Fig. 2.Variation of condensation rate with the solution flow rate.
For small change of concentration and temperature over the dehumidifier,
K0c*c Sh and Sh=*5 Condensation rate ;
kLoc Re0 81 and Re^ solution flow rate.
Thus a plot of (1/condensation rate) versus l/(solution flow rate)0-81
should be straight line for a constant air flow rate. Figure 3 shows this graph
with an intercept 0.003(R2=0.93). The small magnitude of the intercept indicates
that the gas phase mass transfer resistance is negligible. Therefore the Lewis
number for these contact operations will be greater than one, unlike that for
air/water operations.
f a
1/(solution flow rate)* 0.81
Fig. 3.Wilson plot.
-------
1618
SUMMARY
Experiments were conducted to obtain an energy balance across an entire
liquid desiccant system, using lithium bromide solution as an absorbent. Enthalpy
changes for the cooling water and the solution were used to calculate the energy
balance and the error found for this balance was 5-10% for the decoupled mode
and 4-14% for the coupled mode.
A plot of solution flow rate versus condensation rate showed that the slope
of the straight line obtained was 0.81 and that the dehumidifier thus operated
under well mixed, near turbulent conditions on the liquid side. A Wilson plot
obtained from these results indicated that the gas phase mass transfer resistance
was negligible for air/solution contact in the packed bed dehumidifier.
ACKNOWLEDGEMENT
The authors appreciate the support for this work provided by the Active
Heating and Cooling Division, Office of Solar Heat Technologies, U.S. Department
of Energy.
REFERENCES
Burdukov.A.P., N.S. Bufetov, N.P.Deriy, A.R.Dorokhov, and V.Kazakov (1980).
Experimental study of the absorption of water vapor by thin films of aqueous
lithium bromide. Heat transfer - Soviet Research. 12, 118-123.
Collier,R.K.(1979). The analysis and simulation of an open cycle absorption
refrigeration system. Solar Energy. 23. 357-366.
Flaherty,M.(1989). Mass transfer studies for a liquid desiccant dehumidifier.
Thesis, Department of Agricultural and Chemical Engineering, Colorado State
University.
Lof,G.0.G., T.G. Lenz and S.Rao (1984). Coefficients of heat and mass transfer
in a packed bed suitable for solar regeneration of aqueous lithium chloride
solutions. Journal of Solar Energy Engineering,106. 387-392.
-------
1619
OPEN-CYCLE ABSORPTION SOLAR COOLING: GLAZED AND UNGLAZED
OPEN FLOW LIQUID ABSORBENT SOLAR COLLECTOR/REGENERATOR
M.N.A. Hawlader, Andrew P. Stack and Byard D. Wood
Center for Energy Systems Research and Mechanical and Aerospace Engineering Department,
College of Engineering and Applied Sciences, Arizona State University, Tempe, AZ 85287-5806
ABSTRACT
The performance of two collectors/regenerators (C/R), one glazed and the other unglazed, was
studied when tested side-by-side under the meteorological conditions of Tempe, Arizona. One half
of an 11m by 11m roof structure was covered with Tedlar, and the other half remained unglazed.
When the performance of the C/R was expressed in terms of the quantity of water evaporated from
it, which in effect was a measure of the cooling capacity of the system, the unglazed C/R
performed better than the glazed C/R under the climatic conditions considered here.
Nondimensional correlations for heat and mass transfer were developed from the experimental data
for subsequent use with a performance model. The glazed C/R was found less sensitive to
changes in independent variablestand its regeneration efficiency was, on the average, 7.6% lower
than the unglazed C/R. The glazed C/R performance approaches that of the unglazed C/R at largt
glazing heights.
KEYWORDS
Solar cooling; liquid absorbent; open-cycle; collector/regenerator, glazed/unglazed; performance.
INTRODUCTION
Absorption cooling systems provide an alternative to the conventional cooling systems which use
refrigerants that are in the process of being phased out due to their detrimental effects on the
environment. For solar cooling, using the open-cycle absorption refrigeration (OCAR) system, the
energy required for the refrigeration process comes primarily from the sun. The OCAR system
can be divided into three parts: the evaporator, the absorber and the solar collector/regenerator
(C/R). The refrigerant is water, and it is sprayed into the evacuated evaporator, where it vaporizes.
The cooling effect is derived from this phase change.in that the energy required for evaporation is
drawn from the environment being cooled.The absorbent absorbs the water vapor from the
evaporator and .therefore, must be reconcentrated in the C/R. The simplest design for the C/R is
the roof of the building envelope being cooled. Extensive reviews of the literature for open-cycle
absorption cooling technology including C/R research were given by Novak (1984), Nelson
(1986) and Siebe (1986). The most extensive experimental work on the unglazed C/R to date was
performed by Novak (1984).
The purpose of this work, which was funded by the US DOE (Contract DE-FG03-86SF16345), is
to carry out a side-by-side comparison of a glazed and an unglazed C/R of exactly similar type
under identical inlet and climatic conditions.
EXPERIMENTS
The 1 lm by 1 lm C/R located at Arizona State University (longitude: 112° W, latitude; 33° 26' N,
elevation; 379m above sea level) consisted of a south facing wooden roof structure mounted at an
-------
1620
angle of 20° from horizontal, it was covered with black asphalt shingles and was divided into two
sections: one glazed and the other unglazed. At the top of the unglazed C/R, a spray header was
installed to evenly distribute aqueous LiCl across the collector width. The total unglazed C/R area
was 45.8m2. The LiCl solution, after it has flowed over the collector, was collected into a PVC
gutter and returned to a storage tank by gravity flow, as shown in Fig. 1. The other half of the
roof structure was covered with an adjustable glazing consisting of seven panels each 1.67m by
4.88m. Each panel was fitted with Tedlar Poly Vinyl Fluoride glazing. The measured
transmittance was 90%. Its aperature area was 44.0m2, while the area for heat and mass transfer
was 45.8m2. The nine independent variables, shown in Table 1, were measured to describe the
heat and mass transfer characteristics of the two C/R's.
FI2 FI3
Til
Vll VV12
filteiy'
spray header
Unglazed
Glazed
C/R
C/R
: Nrs
concentration ,.
sampler
to roof drain H—
V13
concentration \
_q sampler collection
TI2
gutter
Tank#l
V14
s—
Tank #2
—\
V -
V15! [V16. [
Tank #3
V5
city water
v4 V3 ^pp'y
Fig. 1. Collector/regenerator system configuration.
TABLE 1. Independent Variables and Their Ranges
Variable Range Units
Solar Lxadiance
1159
to
794
W/m2
Amb. Humidity Ratio
0.0304
to
0.0062
kg water/kg air
Inlet Sol. Temp.
66.7
to
37.9
°C
Sol. Flow Rate:
Glazed C/R
0.255
to
0.081
kg/s
Unglazed C/R
0.255
to
0.0914
kg/s
Inlet Concentration
0.486
to
0.296
kg LiCl/kg sol.
Wind Velocity
13.5
to
1.11
m/s
Ambient Temperature
40.1
to
28.5
°C
Ambient Pressure
734.1
to
726.2
mmHg
Glazing Height
0.236
to
0.0635
m
Tests were undertaken on days with clear, cloudless skies. Since the overall C/R performance was
sensitive to the solar input (Nelson, 1986), the glazing surface was cleaned early each morning to
ensure a constant,or nearly constant,transmittance of the glazing surface. Tank #3 was filled with
aqueous LiCl solution at the desired concentration. Valves V14-V16 were positioned so that the
LiCl solution leaving the C/R's flowed back into Tank #3. Thus, the inlet concentration was
allowed to vary continuously from a dilute to a nearly concentrated value. Details of the
equipment, instrumentation and data acquisition system are given in Stack and Wood (1989).
-------
1621
NONDIMENSIONAL HEAT/MASS TRANSFER PARAMETERS
Nondimensional parameters, developed from data obtained under specific conditions, enable the
generalization of the results so that predictions can be made for a wide range of circumstances. A
linear least-square curve fitting routine was used to find the correlation that gave "best fit" to the
experimental data. The Prandtl and Schmidt numbers for air were assumed to be constant (Pr =
0.7 and Sc = 0.6). For the unglazed C/R:
Nul = 10-27-32ReL°'302GrL*1-611N*-1-312
ShL = 1022.54ReL0.057GrL-1.616N-0.799
(1)
(2)
Although the actual unglazed C/R was neither a UH/MF surface nor a UWT/C surface, the Nu
number was best correlated as a function of UH/MF nondimensional parameters and the Sh
number best correlated as a function of UWT/C nondimensional parameters (Novak, 1984;
Nelson, 1986). Stated another way, under the conditions imposed by the present experiment, the
heat transfer functioned as if the C/R were a UH/MF surface and the mass transfer operated as if
the C/R were a UWT/C surface. The net solar flux, the difference between the absorbed energy
and that radiated to the sky, was considered as a constant heat flux. Novak (1984) determined the
local temperature profile at low flow rates and demonstrated that the bulk of the sensible heat gain
occurred within first meter of collector length. The rest of the collector was approximately at
constant temperature. Thus, the bulk of the mass transfer may be assumed to occur at uniform
wall temperature.
A regeneration efficiency, T|, was defined as the ratio of energy required to evaporate water from
the solution to the solar energy incident on the C/R,minus the sensible heat gain of the solution.
Typical values of the regeneration efficiency for the unglazed C/R testing ranged from 0.138 to
0.692)and the average was 0.466. The regeneration efficiency correlation was:
r| = 10-2.347ReL-0.057GrL*0.255N* 1.027 (3)
Figure 2 shows that the present results appear generally consistent with Novak's previous results.
1C 5]
3
a
a.
X
a
~ Nu - Present Work
• Nu - Novak's Data
*
103 104
Nu • Predicted
(a)
Sh - Present Work
Sh - Novak's Data
10 4.
10 4
Sh - Predicted
(b)
Fig. 2. Comparison of experimental results with those predicted by Novak's data.
The form of the correlations for the glazed C/R was suggested by Nelson (1986) from a solution of
the governing equations for the local transfer coefficients. The results were:
Nub = 10-8-502ReL0.032z*.31.52Rab*0.224
Shb = 100.157ReL-4.439E-3Z-2.615Rab0.283
¦q _ l()8.203z*30.27Rab*-1.501E-3
(4)
(5)
(6)
-------
1622
The regeneration efficiency for the glazed C/R for the specific conditions of the experiments varied
between 0.30 and 0.48, with an average of 0.39.
PERFORMANCE OF COLLECTOR/REGENERATOR
A simulation model was then developed to identify important variables. This model, validated
with experimental results, may be incorporated into a simulation program for the entire open-cycle
absorption refrigeration (OCAR) system. The performance of each C/R was judged by overall
evaporation rate, since it determines system cooling capacity.
The model predicted the outlet solution concentrations and temperatures with a great deal of
accuracy for both the glazed and unglazed C/R's, as shown in Figs. 3 and 4. Hence, this model
was used to evaluate the influence of each independent variable on the performance of the C/R with
all other variables held constant
0.40
Ed 0.30
330 335 340
Tg Predicted, K
(a) Glazed C/R outlet solution temp. Tt
Fig. 3 Experimental vs. predicted results of glazed C/R performance model.
0.35 0.40 0.45
Cg Predicted, wt. fraction
(b) Glazed C/R outlet solution conc. C
0.50
g
3 340
w 325
0.55
£ 0.45
U 0.30
315 320 325 330 335 340 345 350
Tu Predicted, K
(a) Unglazed C/R outlet solution temp. Tu
0.35 0.40 0.45 0.50
Cu Predicted, wt. fraction
0.55
(b) Unglazed C/R outlet solution conc. Cy
Fig. 4. Experimental vs. predicted results of unglazed C/R performance model.
From Fig. 5, it is seen that an increase in solar irradiance increased the evaporation rate almost
linearly for the glazed C/R. In contrast, the evaporation rate for unglazed C/R went through a'
maximum at approximately 900 W/m2. For the same condition, the fluid left the glazed C/R at a
higher temperature and lower concentration due to low evaporation rate. This could be due to the
design of the g lazing, where air enters the C/R at the inlet and travels a long distance before it
reaches the outlet. As the air enters into the C/R, it becomes fully laden with water vapors within a
short distance and, thereafter, it travels freely without carrying additional water vapor.
The overall evaporation rate for the unglazed C/R increased with mass flow rate while the
evaporation rate for the glazed C/R decreased, as shown in Fig. 6. The unglazed C/R was more
sensitive to changes in the ambient temperature than the glazed C/R, as shown in Fig. 7. The
-------
1623
evaporation rate for the glazed C/R went through a maximum near 30°C [303K] and then
decreased steadily.
Inlet solution temperature affected overall evaporation rate differently, as shown in Fig. 8. The
outlet temperatures for both C/R's increased, as expected. Obviously, a warmer solution had a
greater driving potential for heat and mass transfer. This produced an increasing evaporation rate
in the glazed C/R, but it had a mixed effect for the unglazed C/R. The evaporation rate for the
unglazed C/R went through a maximum near 49.4°C [322.4 K], The driving potential for mass
transfer was directly affected by changes in humidity ratio. The glazed and unglazed C/R's
responded similarly to increases in ambient humidity with the glazed C/R having much less
sensitivity, as shown in Fig. 9. The driving potential for mass transfer decreased as a result of
increase of humidity, and a greater fraction of available energy manifested itself in sensible instead
of latent heat. Figure 10 shows that the glazed C/R performance was essentially unaffected by
•wind velocity.
Like the effect of ambient humidity on performance, inlet solution concentration affected the glazed
and unglazed C/R performance in the same way, as shown in Fig. 11. The glazed C/R evaporation
rate was much less sensitive to changes in inlet solution concentration than the unglazed C/R.
Figure 12 shows the effect of aspect ratio on the performance of the glazed C/R. The results of
unglazed C/R are included for comparison only. The overall evaporation rate increased slightly
with increasing plate height, while the outlet solution temperature decreased. Although Nelson
(1986) predicted a better performance of the glazed collector, this study indicated that the unglazed
C/R performed better, as shown in Fig. 12, than the glazed C/R for the conditions of tests carried
out under the meteorological conditions of Arizona.
fS
a 12
„ 1.0
"a o.i
&
0.6
* 04
w
Fig. 5.
> Unglazed Evaporation Rate
' Glazed Evaporation Rate
800 850 900 950 100 105
Insolation (W/mA2)
Effect of insolation on C/R
performance.
110 115
i Unglazed Hvap. Kate
¦ Glazed Evap Rate
0.01 0.02 0.03 0.04 0.05 0.06 0.07
Inlet Solution Flow Rate (kg/s m)
Effect of inlet solution flow rate on
C/R performance.
1.2
1.0
g 0.8
2 0.6
2 0.4
¦" Unglazed Evag. Rate
• Glzd Evaporation Rate
> 295 300 305 310 315
w
Ambient Temperature (K)
Fig. 7. Effect of ambient temperature on C/R
performance.
Unglazed Evap. Rate
Glazed Evap. Kate
0.014 0.015 0.016 0.017 0.018 0.019 0.020
Ambient Humidity
Fig. 9. Effect of ambient humidity on C/R
performance.
s
68
8-
w
Fig.
1.2.
1.0.
0.8 ¦
0.6.
0.4.
0.2
. Unglazed fcvap. Rate
• Glazed Evap. Rate
Inlet Solution Temperature (k)
8. Effect of inlet solution temperature on
C/R performance.
' ¦ Unglazed Evap. Rate
• Glazed Evap. Rate
123456789 10 U
Wind (m/s)
Fig. 10. Effect of wind velocity on C/R
performance.
<
1.2
B
1.0
u
%
0.8
&
0.6
0.4
a
0.2
Bk
>
w
-------
1624
¦ Unglazed Evap. Rate
• Glazed Evap. Rate
£¦ 0.38
0.39
0.40
0.41
0.42
Inlet Solution Cone
0.43 0.44
Fig. 11. Effect of inlet solution concentration
on C/R performance.
Unglazed Evap. Rate
Glazed Evap. Rate
r*
0.015
(2
Aspect Ratio
Fig. 12. Effect of aspect ration (b/L) on
glazed C/R performance.
CONCLUSIONS
The analysis of the experimental data led to the following conclusions:
1. The unglazed C/R performed better than the glazed C/R for the same solar/meteorological
conditions.
¦2. The nondimensional heat and mass transfer correlations developed from the experimental data
for use with performance simulation model predicted the results, obtained previously at ASU,
well.
3. The performance model demonstrated that the glazed C/R was much less sensitive to changes
in most independent variables. The regeneration efficiency determined from experiments for
the unglazed C/R was, on the average, 7.6% higher than the glazed C/R. The exit solution of
the glazed collector demonstrated a considerably higher sensible heat gain, making it a better
candidate for winter heating in addition to summer cooling.
4. The optimum geometry for the glazed C/R was at larger plate heights, where the performance
approached that of an unglazed collector/regenerator.
NOMENCLATURE
b
Glazing Height, m
Sc
Gr
Grashof number
Sh
L
Length of C/R, m
UH/MF(*)
Le
Lewis number
N
Ratio, mass to heat transfer
UWT/C
Grashof numbers
Nu
Nusselt number
Pr
Prandtl number
z
Ra
Rayleigh number
Schmidt number
Sherwood number
Uniform heat and max flux
boundary condition
Uniform wall temperature and
concentration boundary
condition
Scaling factor, 1/2[1 + N/Le]
REFERENCES
Nelson, DJ. (1986). Combined Heat and Mass Transfer Natural Convection Between Parallel
Plates. Ph.D. dissertation Arizona State University.
Novak, K.S. (1984). Combined Heat and Mass Transfer on an Open Flow Liquid Absorbent
Solar Collector/Regenerator. Master's Thesis, Arizona State University.
Siebe, D.A. (1986). Evaluation of Air-Conditioning Systems Utilizing Liquid Absorbents
Regenerated by Solar Energy. Ph.D. dissertation, Arizona State University.
Stack, A.P. and B.D. Wood (1989). Heat and Mass Transfer From Glazed and Unglazed Open
Flow Liquid Absorbant Solar Collector/Regenerator. Department of Energy - San Francisco
Operations: DOE/SF/16345--4.
-------
1625
EXPERIMENTAL STUDY FOR A GLAZED SOLAR COLLECTOR/
REGENERATOR OF THE OPEN CYCLE ABSORPTION
SOLAR COOLING SYSTEM
R. Yang, W. D. Chang and C. J. Peng
Institute of Mechanical Engineering
National Sun Yat-Sen University
Kaohsiung, Taiwan, R.O.C.
ABSTRACT
This paper presents an experimental study for the performance of a
solar collector/regenerator of the open-cycle absorption solar
cooling' system operated in a humid area. The typical averaged C/R
efficiency obtained is about 10%, although the C/R efficiency can
go up to 20% in some cases, and the typical averaged cooling
capacity for a 36 m2 collector/regenerator is about 0.8 ton. The
correlations for the heat and mass transfer coefficients are
g iven.
KEYWORDS
Solar collector/regenerator; solar cooling; open-cycle; absorption
solar cooling.
INTRODUCTION
The open-cycle absorption solar cooling system has been studied
for more than twenty years (Kakabaev and Khandurdyev, 1969).
According to the previous studies (Wood, et, al. , 1983, 1986),
this type of solar cooling system has been proposed as the
most viable solar cooling system owing to its low system cost and
high solar fraction (ratio of the solar cooling capacity to the
total cooling load). In particular, our previous computer
simulation study (Yang and Yan, 1989) showed that the solar
fraction can be higher than 95% for this kind of solar cooling
system operated in a humid area - Taiwan. Therefore, a three-
year-term project- for the prototype system testing is now in
progress in order to justify the system feasibility. For the
study of the solar collector/regenerator (C/R), some similar
research has been reported by Eokabaev and Golaev (1971),
Kokabaev et. al. (1978), and Folkman et. al. (1989). However, our
major concern in this work is the performance of a solar C/R
operated in Taiwan -- a humid area (RH>80%).
EXPERIMENTAL SETUP
The C/R is constructed facing south with 20 degree tilt angle and
located at southern Taiwan (22.5° latitude). It has a size of 9
-------
1626
meters in length and 4 meters in width. The C/R surface is made
of plywood, and covered firstly by a waterproof plastic layer and
then by a layer of black cloth material which ensures a good
wetting of the aqueous lithium-chloride solution flowing on the
surface. The glazing material is the 5 mm-thick reinforced glass.
The glazing is supported by a 1 stainless steel; frame which is
specially designed so that the glazing gap height can be varied
from 3 cm to 15 cm. Figure 1 shows the C/R appearance. Solution
temperatures are measured by thermocouples at the locations shown
in Fig. 2. Solution concentrations are measured at the C/R inlet
and outlet by a pycnometer. A schematic of the solution loop is
shown in Fig. 3. The solar irradiance is measured and integrated
by Kipp & Zonen pyranometer (CM11) and integrator (CC12). Other
data measured for the data analysis include ambient dry-bulb and
wet-bulb temperatures, the C/R outlet humidity, and the solution
flow rate.
EXPERIMENTAL RESULTS
A typical data set is shown in Fig. 4. The C/R efficiency is
defined as the ratio of the cooling load corresponding to the
amount of evaporated water to the total solar irradiation incident
on the C/R surface. Although the C/R efficiency can go up to 20%
in some cases, the typical averaged efficiency is only about 10%,
and the corresponding cooling capacity is about 0.8 ton for a 36 m2
C/R. The measured results of the C/R performance are much lower
than the previous simulation results (Yang and Yan, 1989) using
Nelson's (1986) numerical model. The previous simulation gave an
averaged efficiency of about 30%.
The experiment has been carried out for the glazing gap height of
3, 5, 7 and 9 cm. The results already indicate that the C/R
performance is the best when the gap height is the least.
Therefore, further increasing the gap height is not necessary.
The data analysis is done following a procedure of a. uniform
temperature model similar to that of Folkman et. al. (1989). The
Nusselt number (Nub) and the Sherwood number (Shb) are correlated
by the Reynolds number (Reb ), the Rayleigh number (Rab ), and a
scaling parameter, Z (Nelson, 1986). The results based on 230
data sets are given as follows:
Nub = 101-36 Reb"0'259 Rab0-362 Z"0•299 (R2log = 0.49}
Shb = 10-2•79 Reb1-19 Rab0-284 Z~0-544 (R2log = 0.94)
Since the mass transfer is of our major concern, the obtained
correlation for the Sherwood number is desirable. For a better
correlation for the Nusselt number, an analysis using uniform heat
flux model (Folkman et. al. , 1989) may be employed. The present
correlation results are shown in Figs. 5 and 6.
CONCLUSIONS
The results of the present study for the C/R operated in a high
humidity area indicate a relatively low (about one third)
efficiency as compared to the results of a C/R operated in a dry
area (Folkman et. al., 1989). Further research for improving the
-------
1627
C/R efficiency must be done before the system could be
economically viable in a humid area.
ACKNOWLEDGEMENT
The authors gratefully acknowledge the financial support of the
Energy Committee in supporting this research under the contract
792J7.
REFERENCES
Folkman, C.C., A.P. Stack, and B.D. Wood (1989). DOE/SF/16345--4.
Kakabaev, A., and A. Khandurdyev (1969). Gelioteknika 5 (4), 69-
72.
Kakabaev, A., and M. Golaev (1971). Go1 i.otoknj ka 7 ( 1), 80-83.
Kakabaev, A., 0. Klyshchaev, S.Tuiliev, and A. Khandurdyev (1978).
Gel ioteknika 8 ( 2), 42-45.
Nelson, D.J. (1986). Combined heat and mass transfer natural
convection between parallel plates. Ph.D. Dissertation, Arizona
State University.
Wood, B.D., D.A. Siebe , M.A. Applebaum, K.S. Novak, and L.M.
Ballew (1983). DOE Research Report, DE-AC03-82SF11691 MOO I .
Wood, B.D., K.S. Novak, and D.J Nelson (1986). DOE Research
Report, DErAC03-84S.F12 2 23 .
Wood, B.D., D.A. Siebe, M.A. Applebaum, and R.K. Collier (1986).
DOE Research Report, I)K-A(H13-8I SK 1 222:1 .
Yang, R., and W.J. Yan ( i 9 8 9 ) . Proceedings of the Solar World
Congress.
""JP*
F1 i £5 * 1. The appearance of the solar collector/regenerator.
-------
1628
Intel —(icl
1"
1.49
4
I.44
4
1.47
4
1.49
4
1.49
*4
GD
CUD
C tc6 )
( tc8 )
QcIO)
tlcl2l_
CZD
(£D
G£D
(jsD
(tcTT)
ZItcl3l_
\
T
1.36
4
i-5 3
f
1.42
i
1.37
f
Fig. 2. The schematic of the
solar collector/regenerator.
solar collector/regenerator
filter
valve
u«-—«-|
-t^Q_
preparing tank
strong solution lank
CL
-•*0 Q_
5-14-1990
t> 800
¦9 600
U 500
5 300
W 200
5-14-1990
time
5-14-1990
14
12
10
>»
o
&
6
e*
4
o
2
0
Fig. 4. A typical experimental
result for May 14> 1990.
weak solution lank
Fig. 3. The schematic of the
experimental loop.
-------
1629
20 40 60 80 100
experimental data
experimental data
Fig. 5. Correlation vs.
experimental data for Nu.
Fig. 6. Correlation vs.
experimental data for Sh.
-------
1630
REGENERATION PERFORMANCE OF AN ADSORBENT IN THE DESICCANT/REGENERATOR
AND IMPROVEMENT OF REGENERATION EFFICIENCY
Y. Saito
Department of Mechanical Engineering
Osaka Institute of Technology
5-16-1, Omiya, Asahi-ku, Osaka, Japan
ABSTRACT
This study discusses the solar desiccant system using adsorbent, especially an
"integrated desiccant/regenerator" using direct solar energy as heat source for
efficient regeneration of adsorbent. The objective of this study is to propose
the desiccant/regenerator in which adsorbent absorbs solar radiation directly
and is heated for regeneration. Another objective is to obtain the regeneration
characteristics in the equipment proposed. In this study, silica gel is used as
an adsorbent. Also, the possibility of increasing regeneation efficiency was
discussed.
KEY WORDS
Desiccant/Regenerator, Regeneration, Adsorbent, Silica Gel, Dehumidification
INTRODUCTION
Header Casing
In the previous solar desiccant cooling system, a high regenerator temperature
was required because of the
need for considerably high h 400bi]1 (ff=200mm)-
temperature air for regenera
tion of adsorbent. In the
results, the regeneration
efficiency decreases,
creating the necessity to
increase the size of the
regenerator area. In order
to solve these problems,
an integrated desiccant/
regenerator has been
designed and constructed,
in which a thin layer of
adsorbent is heated by solar
radiation directly for
regeneration. In comparison
with the dehumidifier in
which regeneration of adsor-
bent is done by hot air
obtained by the solar
col lector, it is clear
that the experimental
apparatus constructed
enables an increase in
temperature of adsorbent
Glazing
7777
/y/////
Silica Gel Layer
(thlckness=25mm)
¦Wire mesh
Aluminum fo!1
Insulation
Glazing
Si 1 lea Gel Layer
O Adsorbent temperature
° Air temperature
® Dew point temperature
• Glazing temperature
— Aluminum foil
Fig.1 Schematic of Desiccant/Regenerator
-------
1631
with the same solar insolation by improving the regeneration efficiency. In
this study, the regeneration performance when silica gel is used as an
adsorbent was obtained. Next, the regeneration performance when carbon particles
are added on the surface of silica gel layer was obtained, and the possibility
of improving regeneration performance was discussed.
METHODOLOGY
Figure 1 shows the integrated
desiccant/regenerator used in
this study. It has a col lector
area 0.08m2(400mmx200mm) and
25mm thick adsorbent layer.
In this apparatus, a single
acrylic sheet(3mm thickness)
is used for glazing and the
adsorbent layer incorporated
in the desiccant bed has a
copper wire on the bottom.
The spherical silica gel is
packed on the wire mesh while
air flows through the silica
gel layer. Figure 2 shows
the experimental methods
and process air flows
through the layer of
adsorbent, exhausting the
vapor desorped from adsorbent
and become regenerated. As
the experiments were carried
out inside the laboratory,
a solar simulator was used.
Bed temperatures, air tempera-
tures and dew point tempera-
tures were measured at various
locations in the desiccant
layer for different solar
intensities and air flow
rates. Bed temperature
distributions were measured
by five scattered samples of
silica gels in the adsorbent
layer, which was equipped with
thermocouples inserted through
small wells. In this study,
the regeneration process time
is two hours. In this study,
other experiments were
carried out when carbon parti-
cles are added on the surface
of si Iica gel layer.
RESULTS
1) Regeneration performance
when silica gel is used
as an adsorbent
dt « 0 0
®-4~ I 1 I I
@
1:Desiccant/Regenerator 7:Humid if ier
2:Lighting bulbs
3 i Pyranometer
4:Nozzle(Flow meter)
5 : Manometer
6:Heater
8:Blower
9:Hygrometer
10:Thermometer
11: Fan
Fig.2 Experimental method
&
0.03
0.02
0.01
0
0.02
-0.01
INITIAL CONDITION
1=600 , (W/»2) W (01=0.4080 (kg Ag")
G=5x6ff2 tkB/3) X,=0.0117 IkgAgM
13 10) =301.0 IK) to =301.0 IK)
60 90 120
Tiie 0x6(j' [sec)
Fig.3 Regeneration characteristics
Figure 3 shows the comparisons
-------
1632
between a typical example of
test data of outlet absolute
humidities of process air,
water content of adsorbent and
regeneration rate with time
together with the simulated
results. In this case, silica
gel is used as an adsorbent.
Where the regeneration rate
of water from an adsorbent
is obtained by using Eq.(l).
Rc= G(x2 — Xi ) (1)
Figure 4 shows the regenera-
tion efficiency of adsorbent
with time. Where, the
regeneration efficiency of
the desiccant/regeneratoris
def ined as follows:
Rc-hs
(2)
In Fig.4, simbol O shows
the regeneration efficiency
when silica gel is used as
an adsorbent, and it is found
that the maximum efficiency
i s around 30 %.
2) Regeneration performance
when carbon particles are
added on the surface of
siIica gel
In the case when silica gel is
used as an adsorbent, there is
a problem that the value of
silica gel layer absorptance
is relatively small. However,
it is possible to improve
regeneration efficiency if
there are any methods to
increase an absorptance of
the adsorbent layer. To
reduce reflectance of an
absorbent, carbon particles
are added on the surface of
the silica gel layer. Figure
7 shows a relation between
an absorptance and area ratio
. Where is an area
ratio of a cross sectional
area of carbon particles and
the regenerator area and
defined by Eq.(3) referfingto
Fig.6.
INITIAL CONDITION
1=600 , [W/.2] W10) =0.4080 IkgAg")
G=5x60 [kg/a) X, =0.0117 Ikg/kg*)
ts (0) =299.0 (KJ _tn=29£LQ IK)
--Predicted(Si I ica gel)
— Predicted(Si I ica gel+Carbon)
(*=0.152. e=7.5(iij])
O Experigent(Si Iica gel)
_L
60
90
120
Time 8x60 (sec)
Fig.4 Regeneration efficiency
Cor re
!at i on
Exoer i Bent
0.7
6
£3 0.8
«.
~ 4 O
R
XoXs
TToOs
^5 0.5
P
.0.4
cc
0.3
-
R
0.2
0.1
1— i A A
XaTk
< <
•A—*—
'
' '
0.1 0.2 0.3 0.4 0.5 0.6 0.7
W
Fig.5 Relation between rediation proper-
ties and water content of silica gel
e *-¦ Carbon
Particle
Dc=3.3mm
Fig.6 Area ratio between regenerator area
and carbon particle cross-sectional
area
-------
1633
K 2
= Dc /e2 (3)
4
In Fig.7, W shows the water
content of silica gel, and
the effect of water content
of carbon particles is not
considered, since the amount
of carbon particles is negli-
gible. From Fig.7, an absorp-
tance increases when carbon
particles were added on the
surface of silica gel layer.
By using the radiative
properties of an adsorbent
layer when carbon particles
are added shown in Fig.7, the
regeneration performance is
simulated. Figure 8 shows a
relation between the regenera-
tion efficiency versus the
process air Reynolds number.
In Fig.8, the regeneration
efficiencies are shown by
dotted lines and were compared
with the efficiencies when
silica gel is used as an
adsorbent. Where, Reynolds
number is defined by Eq.(4).
Ds G'
Re= (4)
where, G' is an air mass flow
rate per unit regenerator area
F, and m is an air viscosity.
Figure 9 shows the effect of
an area ratio 4> on the re gene
ration efficiency, and it is
found that the efficiency
increases with remarkably.
~ &
¦
n on
^8.?8
Correlation
Experiment
OW=0.10
AW =0.30
~ W =0.80
0
0.1 0.2
» i i
(for 0-3
'
2015 10
7 6
e [M]
Fig.7 Relation between absorptance and
Si I i ca ge I
Si Ii ca gel+Carbon
f 4> =0.152 , e=7. 5(mJ )
INITIAL CONDITION
is (01=301.0 (K) Xi =0.0142 (kg/kg*]
tg (01=301.0 (KJ W(0)=0.40 (kg/kg")
to =301.0 (K) RH=60 (%)
I
1.0
2.0
l
3.0
4.0
Re
5.0
6.0
Fig.8 Relation between regeneration
efficiency and Reynolds number
CONCLUSIONS
Figure 9 shows a relation between regeneration efficiency and area ratio defined
by Eq.(4). From Fig.9 ,it is found that the regeneration efficiency depends on
the area ratio . And the maximum regeneration efficiency becomes around 40
% when carbon particles are added on the surface of silica gel layer.
The dotted line in Fig.9 shows the increasing ratio in regeneration efficiency
A 77, and it is found that it is able to obtains the maximum value of & v is 40 %
and it is effective to add carbon particles on the adsorbent layer for improving
regeneration efficiency.
-------
1634
NOMENCLATURE
Dc= diameter of carbon
particle
Ds= average diameter of
s iIi ca gel part i cle
F = col lector area of the
des i ccant/regenerator
G = mass flow rate of
process a i r
hs=heat of sorption of
water
I = insolation on regenerator
surface
R = reflectance
Rc= regeneration rate
Re= Reynolds number in the
adsorbent layer
RH= relative humidity
t = temperature
W = water content of silica
gel
x =absolute humidity
1 [W/pfl
0.5
0.4 -
0.3
0.2 -
0.1
A 7] 11=500 (W/V)
INITIAL CONDITION
13(01=301.0 IK) Xi =0.0142 IkgAgT
tg(0)=301.0 (K) W(0)=0.40 (kgAg")
to =301.0 no m=60 txi
G =3x60"! (kg/sl Rs=1.66
40
30
20
10
0.1
0.2
0.3
Fig.9 Relation between regeneration
efficiency and area ratio
Greek letter Subscript
a = absorptance 1= inlet
V = regenerat ion efficiency 2= out let
0 = time 0= ambient
r = transmittance s= adsorbent
= area ratio defined by Eq.(3) G= glazing
REFERENCES
Johansen, A. and Grossman, G. (1983). Performance Simulation of Regenerating
Type Solar Collectors. Solar Energy, 30, 87-92.
Toei, R. (1975). Drying Equipment, Nikkan Kogyo Press.
Saito, Y. (1987). Regeneration Characteristics of Adsorbent in the Integrated
Desiccant/CoI lector. Solar Engineering-1987 Vol.Two, 861-866. ASME
Test data of the Fuji-Davison Type III silica gel, Fuji-Davison Chemical Co.
-------
2.14 Active Cooling II
-------
-------
1637
EXERGY ANALYSIS OF A SOLAR ABSORPTION REFRIGERATION
SYSTEM FOR EFFICIENT UTILIZATION
OF SOLAR ENERGY
P. Kumar, T.K. Chaudhuri and A. Dasgupta
Metallurgical & Engineering Consultants (India) Ltd
Ranchi - 834002
INDIA
ABSTRACT
Exergy analysis of an absorption refrigeration system has been done. The
refrigerant-absorbent pair used for the analysis is water-lithium bromide.
A flat plate solar collector array has been considered for supply of heat
to the generator of the refrigeration system. The emphasis is on reducing
the required solar collector area for a given cooling capacity. It is shown
that a given solar collector array could be utilized more efficiently by
adjusting the flow rate of the absorbent i.e. aqueous lithium bromide in
the system. The exergy relationship in the generator of the system is
displayed graphically by exergy utilization diagram.
KEYWORDS
Exergy analysis; solar; absorption refrigeration
INTRODUCTION
The present study deals with the use of a solar energy driven water lithium
bromide absorption refrigeration system for airconditioning application.
Hot water from a flat plate solar collector energises the generator of the
refrigeration system. The performance of such absorption refrigeration system
can be improved if it is operated at relatively higher flow ratios. Flow
ratio is defined as ratio of mass flow rate of strong lithium bromide solution
to the mass flow rate of refrigerant in the cycle. Exergy analysis of the
refrigeration system is presented in this paper. The exergy change Ax
for a process can be defined as follows :
AX = AH - Tam.AS (1)
Equation (1) can be rearranged as Equation (2).
A = 41 = 1 " tam* As/ Ah <2)
' A' is called availability factor which represents energy quality. In an energy
transfer process there is an energy doner and an energy acceptor. If AA
and AD are availability factors for energy acceptor and . energy doner
respectively, AD is always greater than AA. An energy utilization diagram
can be obtained by plotting AD and AA against the transferred energy. The
area between the line for AA and AD represents exergy los«?.
The exergy loss in the system can be evaluated if information of all input
and output streams is given. The exergy change for a component of the cycle
can be evaluated by equation (3)(Chuang, 1990):
'receding page blank
-------
1638
" - 0- [1-(TAm' • ta
-------
1639
nEX
= (T6 -T7)/T6 - T4)
(8)
°EX
= nEXm6 (CP>6
>>
be
4) 0.0
a
-200
-400
0 10 20 30 40 50 60
flow ratio (FR)
Fig 2. Variation of exergy change with
flow ratio
-------
1640
as flow ratio is increased. Each of the curves on the figure is plotted
for a constant value of heat exchanger effectiveness. As the heat exchanger
effectiveness increases the minimum of exergy loss curve shifts towards a
higher value of flow ratio as shown in the figure.
0.8 o
8 11
o
in
<
a
0)
Sh
o
O
10 20 30 40 50
flow ratio (FR)
Fig 3. Variation of solar collector area and
coefficient of performance with flow
ratio
Figure 3 shows variation of solar collector area and coefficient of
performance for the ^system with variation in flow ratio for a constant
insolation of 0.7kW/m . The required solar collector area decreases, reaching
a minimum value and then increases as flow ratio is increased. As the flow
ratio is increased, the generator temperature reduces (Kumar, 1984). This
reduction in generator temperature leads to reduction in solar energy collection
temperature, which eventually reduces solar collector area requirement.
However, the reduction of the solar collector area requirement is at a cost
of slight reduction in coefficient of performance of the system. Upjto a certain
value of flow ratio, this reduction in (COP) is immaterial because the cooling
capacity of the system remains the same and required solar collector area
reduces. A relatively lower value of insolation is chosen because lower
insolation values are more critical for operation of the refrigeration system.
As insolation increases not only the solar collector area requirement
reduces, but the minimum of the curve also shifts to a lower value of flow
ratio and thereafter the curve remains flat (Kumar, 1985). The solar collector
area requirement reduces further as solution heat exchanger effectiveness
is increased. It can be noted that the variation of Ag^, with increase in
flow ratio is very similar to the variation of exergy loss with increase in
flow ratio. However, the minimum for both the parameters is not the same,
except when the heat exchanger effectiveness is 0.9.
-------
1641
It may be noted that exergy change of the absorption refrigeration system
is independent of the insolation values. It is because a specific flow ratio
corresponds to a unique generator temperature and insolation change doesn't
affect the parameters of the absorption refrigeration system.
0.2
o
fl}
>,o.i5r
Si
t—i
•H
0.
AD
AA (FR) = 7
AA (FR) = 28
I I I 1
1000 2000 3000 4000
generator heat load, QGE> kJ/h
Fig 4. Energy utilization diagram for
the generator
An Energy Utilization Diagram (EUD) for the generator has been plotted in
Fig.4. It has been assumed that the main generator process proceeds almost
at the exit condition of the equipment because an abundant amount of exit
stream is premixed with the inlet stream. It makes the main system proceed
at the minimum driving force or at the pinch. The EUD has been plotted
for flow ratio values of 7 and 28. The area of each of these diagrams
represent the exergy loss in the generator. Therefore, thermodynamically, the
system should be operated at a flow ratio of 28 rather than 7. This
conclusion matches with the minimum solar collector area requirement at higher
flow ratio.
It can be concluded from the above discussion that for solar energy
operated systems, the flow ratio should be kept at a relatively higher value.
A higher flow ratio enables system to operate with lower temperature
thermal energy source and therefore the solar collector area requirement also
reduces. It is further confirmed by the Exergy Utilization Diagram. This
also means that an absorption refrigeration system, coupled with a given
collector array, would operate for a longer time in a given day if operated
at relatively higher flow ratio.
ACKNOLWEDGEMENT
The authors are thankful to the Management of MECON (India) Ltd for
encouragement and for permission to carry out this work.
-------
1642
REFERENCES
ASHRAE Handbook - Fundamentals (1981). Atlanta : American Society of Heating
Refrigerating and Air-Conditioning Engineers, Inc. USA
Chuang, C.C. and M., Ishida (1990). Comparison of three types of absorption
heat pumps based on energy utilization diagrams, ASHRAE Transactions,
96, Pt.2.
Kumar, P. and S. Devotta (1985). Analysis of solar absorption cooling systems
with low generator temperatures, Int. J. Refrig., f5, 356-359.
Kumar, P., S. Devotta and F.A. Holland (1984). Effect of flow ratio on the
performance of an experimental absorption cooling system, Chem. Eng.
Res. Des., 62, 194-196.
NOMENCLATURE
(COP)
coefficient of performance
specific heat
enthalpy
insolation
mass flow rate
heat load
entropy
temperature
lithium bromide concentration
change in exergy
effectiveness
difference
CP
H
I
m
Q
S
T
X
AX
n
A
Subscripts
AB
AM
CO
EV
EX
GE
OV
SC
WI
1-10
absorber
ambient
condenser
evaporator
heat exchanger
generator
overall
solar collector
hot water temperature at generator inlet
State points in Fig.l
-------
1643
A 2-STAGE LiBr ABSORPTION CHILLER FOR SOLAR
COOLING
Huang Zhi-cheng Xia Wen-hui Ma Wei-bin
Guangzhou Institute of Energy Conversion,
P.O.Box 1254» Guangzhou, China
ABSTRACT
A 2-stage LiBr absorption chilling system,driven by low tempera-
ture hot water, has been designed and tested. To produce 9°C chi-
lled water with 32cC cooling water, the generator input tempera-
ture can be 86-75t'C and the generator output temperature is 60°C,
system COP lies on the range of 0.37 to 0.33. The features of this
system appear advantageous for solar space cooling application.
In this paper, the prototype design and test results are presented
and discussed.
KEYWORDS
Absorption chiller; lithium bromide absorption machine; two-stage
absorption; solar cooling.
INTRODUCTION
The technical possibility of solar space cooling, using solar
medium temperature collectors to drive single-stage LiBr absorp-
tion chilling machine, has been well demonstrated not only in
industrial countries, but also in Chinam. For the single-stage
chiller, the nominal generator input temperature is 88°C. Vacuum
tube solar collector with glass-to-metal sealing is required.
Because of its sophisticated manufacturing technique, vacuum tube
collector costs quite high (3-4 times of the cost for flat plate
collector). So, one of the restrictions of practical use of solar
cooling system in China is an- economical aspect—too high capital
cost of the system. It seems reasonable to lower down the solar
collector cost by using collector models of lower temperature
range, if the generator temperature of the chiller can be lowered
down by using two-stage LiBr absorption chiller instead of single-
stage chiller. In order to search an approach to a more economical
solution of solar cooling, a 2-stage LiBr absorption chiller pro-
totype, working on lower temperature heat source, has been design-
ed and tested. Test results show that the 2-stage chiller seems
desirable for this purpose.
(Initially, the 2-stage LiBr absorption cooling machine was de-
signed for the purpose of low temperature industrial waste heat
recovery, but it seems also suitable for solar cooling applica-
tion) .
-------
1644
PRINCIPLE OP 2-STAGE ABSORPTION CYCLE
The absorption cooling cycle, described in enthalpy-concentration
diagram is shown in Pig.1a. According to the physical properties
of LiBr solution, when cooling watertftemperature is 32®C, it is
difficult to make chilled water of 9 C under low temperature of
heat source. Because for given temperatures of cooling wa,ter and
chilled water, evaporation pressure Po and condensation pressure
Pk and the temperature t* of diluted solution at the outlet of
absorber will be fixed, so, concentration of diluted solution and
the temperature at which solution in generator begins to boil
(point 5) are also fixed. With source temperature decreasing, so-
lution temperature in generator (point 4) decreases,concentration
difference between concentrated and diluted solutions becomes
less and refrigeration capacity falls. If heat source temperature
falls bellow a certain value, point 4 and point 5 will coincide,
resulting in zero refrigeration capacity. Hence, to make chilled
water of 9#C under the, condition of 32'C cooling water temperature
and low, say 75"C, hot water temperature, 2-stage LiBr absorption
process has to be used.
>5
ft
r—I
-P
Concentration
1
7-37 Kfr.
2-73 KPa.
0.93 KPa.
465 50.5 573 60.8 %(**) -j
(a) (b)
Pig.1 Diagram of absorption cooling cycle
As shown in Pig.1a, the cycle is divided into two stages of ab-
sorption. Points 5,4,6, and 2 represent the cycle of high pressure
stage and points 5',4',6', and 2' represent the cycle of low pre-
ssure stage, wit'h an intermediate pressure P' linking the two
stages together. Refrigerant water is made in high pressure stage
and the absorbent—concentrated solution is made in low pressure
stage. So, through the high pressure absorption process, the ge-
neration process in the low pressure generator occurs under lower
pressure, completing a full refrigeration cycle. The flow diagram
of 2-stage absorption cooling cycle is shown in Pig.2.
DESIGN AND TEST OP 2-STAGE ABSORPTION SYSTEM
A prototype of hot water driven 2-stage LiBr absorption cooling
-------
1645
Cooling
water
vapor
Thin
solution
Hot
!water
High pressure
stage
Concentrated
solution
High pressure
absorber
Cooling
water
vapor
Thin
solution
Hot
water
Low pressure
stage
Chilled-
water
Cone.
, vapor
1 Cooling
water -
Condenser
High pressure
generator
Low pressure
generator
Low pressure
absorber
Fig.2 Flowchart of 2-stage absorption process
system of cooling capacity 6 KW has been designed and tested. The
design parameters are given as:
Chilled water temperature: outlet 9°C, inlet 14°C
Heat medium temperature: input 86° C, output 70°c
Cooling water temperature: inlet 324 C, outlet 37°C
Designed points of state and thermocycle are shown in Fig.1b. The
configuration of the system is as shown in Fig.3. Diluted LiBr
solution in high pressure generator 1a is heated by hot water.
Generated water vapor is condensed in condenser 1b. The condensed
water flows into the evaporator 3b to be evaporated, producing
refrigrating effect. Concentrated solution from high pressure
generator enters into high pressure absorber 2a and absorbs water
vapor, generated from low pressure generator 2b, changing back to
diluted solution, and then is pumped back to high pressure genera-
tor, completing a high pressure cycle. The concentrated solution
in low pressure generator goes down into low pressure absorber 3a
and absorbs water vapor from evaporator 3b. The diluted solution
from low pressure absorber is then pumped back to low pressure
generator 2b, completing a low pressure cycle.
The prototype was successfully put in test operation. The refri-
-------
1646
2a 12b
"lira
w 13*1 r
3b
®—
IB6
Fig.3 Configuration of 2-stage
absorption system
1a.High pressure generator; 1b.Condenser;
2a.High pressure absorber; 2b.Low pressure generator;
3a.Low pressure absorber; 3b.Evaporator;
4. Vacuum tank; 5. Low pressure heat exchsnger;
6. High pressure heat exchanger; 7. Pump
geration performance curves are shown in Fig.4.
TEST RESULTS ANALYSIS AND DISCUSSION
Test results of the prototype are compared with the data of the
WFC-2 model 1-stage absorption machine (see Table 1). The 2-stage
one can work on hot water of temperature far lower than 86"C with
cooling water 32°C and with acceptable steady system COP, making
chilled water of 9*C, suitable for space cooling or industrial
cooling process. Temperature drop in generators can reach over
20*C. For the 1-stage one, when heat medium temperature decreases
below nominal, cooling capacity and COP decrease drastically. So
the 2-stage one can use heat Bources of lower temperature to ach-
ieve better cooling effect under more severe conditions, under
which even the 1-stage one cannot work. For solar cooling, these
features of the 2-stage system seem particularly advantageous:
1. Cooling system can work steadily under not so steady solar
input. 2. The lower generator inlet and outlet tempertures in-
crease both instantaneous and daily efficiencies of solar collec-
tor system. 3« Required lower operating temperature provides
possibility to use simpler model of .solar collector, e.g. flat
collectors yrith simple convection'suppressing device**? instead of
vacuum tube collectors, which are ,3~4 times more expensive than
the flat plates, thus reducing theJ,c obstruction cost 'of the 6olar
system. As the collector cost'makes up above one third of the
-------
1647
86"C tw=3Cfc
Cooling water tw=32C
6000
6000
80 "0
5000
75 'c 4600
4400
0.3c-
1.2 14 'C 8 10
Chilled water temperature
Fig.4 Performance curve of 2-stage prototype
TABLE 1 Performance Comparison of 2-Stage
and 1-Stage Cooling Systems
Model
1-Stage(WFC-2)
2
-Stage
Cooling capacity, kw
7
6
Hot water temp., p0
85
80
75
86
80
75
Temp, drop in generator,"C
6-8
6-8
6-8
24-27
17-21
14-17
Cooling water temp., 'C
31
31
31
32
32
32
Chilled water temp., 'C
9
9
9
9
9
9
Cooling effect, kw
4.8
3
0.8
6.6
5
4.9
COP
0.49
0.36
0.09
0.371
0.333
0.327
total system cost, so the extra manufacturing cost of the 2-stage
chiller over the 1-stage one's and the lower COP of the 2-stage
one can be compensated.
-------
1648
CONCLUSION
A prototype o:f hot water driven LiBr 2-stage absorption chiller
has been successfully tested. Results obtained show the superior
features of this system in providing cooling effect at more severe
operating conditions in solar application. It is expected to use
simpler and less expensive models of collector for solar cooling
system, thus to achieve a large scale construction cost reduction,
which is especially essential for solar application in developing
countries. Another bigger 2-stage prototype of cooling capacity
60 kw has been designed and is expected to be constructed. We
hope the further test results can be obtained soon.
REFERENCES
Huang, Z.C., H.S. Ward, and Zheng, Z.H. (1988). Shenzhen solar
cooling and hot water supply system. Advances in Solar Energy
Technology, Vol.3, pp.2876-2880..
Zheng, Z.H. and Huang, Z.C. (1988). Plat plate solar collector
wittj V-corrugated insulator. Advances in Solar Energy Technology
Vol.1, pp.631-635.
-------
1649
PERFORMANCE TESTING OF ft SOLAR ASSISTED HEAT PUMP
J.G. Cervantes*, E. Torres**, J.C. Baltazar**
$
Faculty of Engineering, National University of Mexico,
^ Mexico City, 04510
Scientific Research Institute, University of Guanajuato,
Guanajuato City, L. de Retana 5, 36000
MEXICO
ABSTRACT
A Solar Assisted Heat Pump System, using Freon-22 as the working
substance, was designed, constructed and tested in a semiarid
climate. The system works with a solar flat collector for the
evaporation of the refrigerant and a 740 W hermetic compressor.
The thermal performance of the system was experimentally
determinated by producing hot air and hot water in two properly
designed itQflsThe COP range of the whole system, was from 3 to 6,
while the collector's efficiency ranged from 1 to 3. A simple
model that predicts fairly well the system behavior, is also
presented and compared with experimental results.
KEYWORDS
Heat pump, Solar assisted, Refrigerant filled collector, Water and
air solar heating.
INTRODUCTION
In recent years, heat pumps have been conceived and installed in
various industrial processes with the purpose of saving energy.
In heating applications, solar assisted heat pump systems (SAHPS)
present many advantages over conventional solar systems
(Chaturvedi, 1967; Lin, 1784; Manton, 1982), when operated in
heating mode. Moreover, in certain climates SAHPS have
significant advantages during summer when cooling is required.
The majority of the research conducted to assess the performance
of solar assisted heat pumpsv hgg, -been carried out in northern
zones of the Earth, with the heating operation mode being used in
winter and the cooling operation mode during Summer.
This paper sets out a description of the performance of a SAHPS
using Freon 22 as the working fluid. This system was tested in
Guanajuato City, Mexico where the weather conditions are moderate
throughout the year. In this geographic region the interest for
cooling systems comes from the need for conservation and
preservation of food, while heating systems find their application
in supplying the energy required for drying leather and crops. So
far, the performance of SAHPS has been assessed in the heating
-------
1650
mode only.
The system can work either with a conventional evaporator or, as
was aimed, with a solar flat collector (where the evaporation of
the refrigerant takes place), in a nearly isothermal way.
The main objective of the testing program was to get a better
understanding of the operational aspects of the STAMPS in a
moderate semiarid climate and to verify a theoretical performance
model, which can be used to calculate the amount of thermal energy
that the system can supply by producing hot air and hot water.
MATHEMATICAL MODEL
The theoretical thermal performance of the evaporator is
represented by a steady-state energy balance at the solar
collector. It is assumed that the input comprises a dry saturated
vapour and that temperature and pressure changes between inlet and
outlet are negligible.
The energy balance for the evaporator—collector is then as follows
m (h2-h1) = FR Ac [ G (ax) - UL ( Te - Ta ) D (1)
where the left hand side term is the energy absorbed by the
refrigerant.
The overall coefficient of heat losses is made up of the sum of
the losses which occur at the top and bottom of the collector.
These losses take as a basis the solar energy collection area and
are related directly to the convective effect of the wind over
both sides of the collector (Torres, 1990).
Within a heat pump system, the thermodynamic performance of the
compressor is strongly linked to the evaporation and condensation
temperature levels that the system can handle. The mass flow rate
pumped by compressor, is calculated by the following expression:
Vv VD co
(2)
The coefficient of performance of a heat pump is calculated as the
ratio of the thermal energy rejected by the condenser to the
electrical energy used by the compressor.
m (h - h . )
COP = 5 *2_ (3)
The evaporator—collector thermal efficiency is defined as the
ratio of collector useful energy to the total incident solar
energy on the collection area. This efficiency can be obtained
using the experimentally measured values for refrigerant mass
flow-rate and the temperatures considered in the SAHPS analysis.
DESCRIPTION OF THE PROTOTYPE PERFORMANCE
In Fig. i, a block diagram is presented which describes the
-------
1651
P.T
Collector
Evaporator
T.v
Conventional
Evaporator
Compressor
P,T.
P.T
,P.T v | Air
.Expansion fCa(3mary
valve I | r. '
IF low meter A
>Si)ht glass
P.T.
PJ.
Filterl
Water-Refrigerant
VaterCondenser '
Air
1. Schematic diagram of the solar assisted heat pump system.
prototype that was actually constructed. This consists in a
system assisted by solar energy where the evaporator of a
conventional heat pump (dark line in Fig. 1) was substituted by a
solar collector. In this way two distinct functions to be
performed in the same section were achieved. These are the
collection of solar energy and the evaporation of working fluid.
The system absorbs solar radiation on the evaporator—collector at
ambient temperature. The quality and quantity of the thermal
energy content of the working fluid after passing through the
evaporator—collector are increased by adding energy in the form of
work. The higher thermal energy content of the refrigerant is
rejected in the condenser where is transferred to the cold fluid.
The solar collector area was sized using the same theoretical
calculation employed to evaluate the useful heat in the collector
(Torres, 1990). A conventional system of 1 h.p. of capacity and
the annual average weather conditions around the city of
Guanajuato were used.
The collector plate was construted of copper tubes and aluminum
fins painted with black oil paint. The total area is 4.5 m*.
There are 12 collector tubes, each having a nominal diameter of
0.0127 m and joined forming a flat pipe coil. The collector tubes
are bonded to the fins at a pitch of 0.15 m.
Normally, the superheating that the refrigerant undergoes during
compression is not totally removed in the condensing unit of
commercial equipment (Baltazar, 1790). Therefore, in the actual
prototype this unit was enhanced with a shell and tube heat
exchanger using water as a coolant. The heat transfer area of
this additional unit is 0.116 m2. This exchanger operates between
the compressor and the air-refrigerant condenser (see Fig. 1)
removing a portion of the thermal energy of the superheated fluid.
Condensation of the refrigerant takes place at saturation
temperature in the air-refrigerant condenser whose heat transfer
area is 9.8 ma.
A hermetic compressor of 1 h.p. of capacity was employed during
the test. The thermal expansion valve employed to control the
refrigerant flowrate in the cycle was properly selected based on
oressure drop, evaporator temperature and system capacity for the
-------
1652
working fluid.
Weather conditions like solar radiation, ambient
wind velocity on the one hand, and refrigerant
pressure,at every marked section' in Fig. 1, are
that were experimentally measured.
temperature and
temperature and
those variables
EXPERIMENTAL RESULTS AND DISCUSSION
Due to varing weather conditions, experimental tests were carried
out along different periods of time in a range going from 2 to 9
hours.
Prototype behavior
Various parameters , such as temperature and pressure of
evaporation,6the evaporator-collector efficiency registered during
the experimental runs, were compared with the parameters
calculated by using the mathematical model. The experimental
average pressure drop through the collector was 0.035 MPa.
A comparison between the observed and the calculated pressure
within the evaporator—collector is shown in Fig. 2. The 45°
straight line on this figure represents a perfect match between
calculated and experimental results.
In Fig. 3, expected refrigerant evaporation temperatures are
plotted against the experimental values measured over the testing
period on February 26, 1790. From this figure, it is clear that
the expected evaporation temperature decreases more rapidly than
the experimental one. This is due to the thermal inertia this
equipment possesses. It is worth mentioning that the mathematical
model employed does not account for such effects. In this case the
drop of the registered values was caused by a prolonged overcast.
A comparison between experimental and calculated efficiency over
the testing period is in Fig. 4. Efficiency values larger than
one are characteristic of the kind of the uncovered collector type
working with refrigerants. The foregoing occurs throughout the
year in the geographical region where the test were conducted.
Q7
S„0.6
c «
1 =
c
w V*
£20.5
0.4
* v
>
+* +
+t .
/< *
J* *
V"
^ 1 1
1 1 I
1
0.35 0.45 0.55 0.60
Predicted pressure
0.7
Fig. 2. Experimental and cal-
culated evaporator
pressure.
8 I O 12 14
Experimental time
a Predicted + Experimental
Fig. 3. Experimental and calcu-
lated evaporator tempera-
ture. Feb. 26, 1990.
-------
1653
120*
1.5 2.5 3.5 4.5
Predicted efficiency
0.3 0.5 0.7 0.9 f.l
Solar radiation
(kW/m2)
Fig. 4. Measured and predicted
efficiency of the
col lector.
Fig. 5. R-22 Mass flow rate
vs. solar radiation
measured during experi-
mental period.
The thermal behavior of the uncovered collector using refrigerant
R-22 as working fluid is represented by the curve of performance
below, where all registered experimental data were used
adjusting the best fit.
7> = -17.8 (Te - Ta)/G + 0.78
c
Figure 5 shows the refrigerant mass flow-rate distribution
according to the incident solar radiation registered during the
experiments. Mass flow-rate increases with solar radiation, in a
nearly linear way. The same behavior is obtained when mass flow
rate is plotted against ambient temperature.
The solar heat pump COP is plotted as a function of the incident
solar radiation in Fig. 6 a linear relation is observed. The
range of the COP values is from 3 to 6. This range is consider-
ably higher than the one that a similar system employing an open
compressor can reach. However in a system like the latter,
significant heat losses to the surroundings generally occur.
Heat transfer in the condensation unit during the run on March 2,
1990 is shown in Fig. 7. This energy parameter presents a similar
behavior and follows the coefficient of performance very closely When
ihiiil!'-
-.6.2.
0.3 0.5 0.7 0.9 I.I
Solar radiation
( kW /m 2 )
It 13
E xperimental
Fig. 6. COP vs. Registered solar
radiation.
Fig. 7. Heat released by the
condensation of R-22.
March 2, 1990.
-------
1654
related to ambient temperature and solar radiation, being linearin
this relationship.
CONCLUSIONS
The experimental study carried out shows that the mathematical
model can be improved since there are significant inertial effects
that are not accounted for. However, solar collector area and
compressor capacity for a given thermal load may be calculated
using a programi package developed for a system which utilizes
R-22 as working fluid. Among other refrigerants, R-22 possesses
the remarkable! characteristic of being environmentally acceptable.
According to the performance of a solar heat pump, the COP values
profile as well as the solar collector efficiency obtained are
figures of merit that represent the feasibility of this system to
be used in warm climates with a broad variety of applications.
NOMENCLATURE
Ac Solar collector area Cm*]
Fp Heat removal factor
G Solar flux [W/m*]
hi y h^ Inlet and outlet refrigerant enthalpies [kJ/kg]
h , h Inlet and outlet air or water enthalpies [kJ/kg]
ao' ai
m Refrigerant mass flowrate [kg/s]
ma Air or water mass flowrate [kg/s3
Te, Ta Evaporation and ambient temperatures EK]
U^ Heat, losses overall coefficient [W/m* K]
VD Volume displaced by the compressor [m3/s]
v Specific volume at the compressor inlet
(otr) Absorptance-transmittance product
Y)c Collector efficiency
n Volumetric efficiency
Compressor velocity [r.p.m.]
REFERENCES
Baltazar, C.J.C. and R.E. Torres. (1990). Design, construction and
testing of a solar heating system under the principle of a solar
heat pump. Thesis, University of Guanajuato,
Mexico.(In Spanish.)
Chaturvedi, S.K. and M. Abazeri. (1987). Transient simulation of a
capacity-modulated, direct-expansion, solat—assisted heat pump.
Solar Energy, 39, 421-428.
Lin, S. and K.I. Krakow. (1984). The thermal performance of
refrigerant cooled solar collectors in a solar source heat pump
system. Trans, of the CSME, 8, 40-43.
Manton, B.E. and J.W. Mitchell. (1982). A regional comparison of
solar, heat pump, and solar heat pump systems. Trans. of ASME,
104. 158-161.
Torres, R.E., J. Cervantes de G. and J.C. Baltazar. (1990).
Experimental study of a solar heat pump with freon R-22. Proc.
of Natl. Solar Energy Soc. (Mexico), La Paz, BCS (Mexico),
42-47.(In Spanish.)
-------
1655
A COMPARISON OF HEAT PUMP SYSTEMS USING DIFFERENT TYPES
OF DIRECT EXPANSION SOLAR COLLECTORS
S. Ito* and N. Miura*
*Department of Mechanical Systems Engineering,
Kanagawa Institute of Technology,
Atsugi-shi, Kanagawa-ken, 243-02 Japan
ABSTRACT
A comparison of heat pump systems using different types of direct expansion
solar collectors was made. Using a collector which was made of a copper tube
with many aluminum fins and was 1.09 m2 in the collector area, the performance
of the system was fairly good even when the solar radiation intensity was
small, but solar radiation did not increase COP--; mudh. Using a flat plate
collector without a cover, whose area was 3.24 m2 , COP of 5.3 was obtained
when the solar radiation intensity was high and was very low when it was small.
Using the two types of the collectors arranged in two rows, the weak points
of each collector could be covered by each other under certain conditions,
but the arrangement of the collectors gave negative effects on the performance
when the solar radiation was too much.
KEYWORDS
Heat pump; direct expansion solar collector; COP; evaporator, solar heating.
INTRODUCTION
It is well known that COP of a heat pump increases when the evaporation
temperature is raised. Using a solar collector as the evaporator of a heat
pump, the evaporation temperature can be raised when the solar radiation is
available. Thus, the performance of the heat pump system, which may be used
for home heating or hot water supply, can be improved.
Chatavedi (1982) and O'Dell (1984) showed analytically that a flat plate
collector without a cover presented a better performance as a whole and was
less expensive than a collector with a cover. In experimental studies, finned
tubes were used as collectors by Sigma Research, Inc. (1982), Hino (1984),
Iwanaga (1985) and Shinobu (1985). Flat plate collectors without covers and
with covers were used by Sadler (1987) and Fujita (1983) in their studies,
respectively.
The collectors made by finned tubes has a higher. wind convective
coefficient than a flat plate collector without. fins. In this study, the
performance of a collector made of a tube with many fins (convective type
collector) and a flat plate collector without cover (radiative type collector)
was studied firstly. Secondarily, the case that the two collectors arranged
in two rows was considered in order to get the merits of each collector and
the performance was tested.
-------
1656
EXPERIMENT
Figure 1 shows the experimental apparatus. Either one of the two types of
collectors or both at a time can be used by adjusting the three-way valves.
When the two types of collectors are used, refrigerant 12 flows through the
double pipe condenser, the two thermostatic expansion valves, the two
collectors and the accumulator. Water for cooling the refrigerant flows
through the thermobath, which has a pump to circulate the water, the condenser
and the flow meter. The compressor is an electrical rotary type and the related
power is 350 W.
A convective type collector manufactured by Matsushita Housing Products Co.,
Ltd. was used. The inside and the outside diameters of the copper tube are
7.24 and 7.94 mm, respectively. Aluminum fins with the thickness of 0.25 mm
and thej pitch of 5.8 mm are fixed on the tube. The area of the collector is
1.09 m (0.719m x 1.516m). Thermocouples are soldered on the tube in order
to measure the temperatures and to guess the evaporation temperature. The
details of the performance of the collector were given by Iwanaga (1985).
The bl^ck painted surface of the flat plate collector has the total area of
3.24 m . A copper tube, whose inside and outside diameters are 8.0 mm and
9.52 mm, respectively, is soldered on the back side of a copper plate with
a pitch of 100 mm, which is 1 mm in the thickness. It is mounted on a wooden
plate. The collectors face to the south and the angle between each collector
and the horizontal is 50°.
The solar incident radiation averaged in 30 minutes was measured by pyranometer
which was set with the same angle as the collectors. Thermocouples were used
for all the measurements of temperatures. The coefficient of performance of
the heat pump was obtained from the following equation:
COP = mc (t -t )/W (1)
p w,2 w,l c
where m is the mass flow rate of the water, c is the specific heat, t
and t ^ are the temperatures at the inlet of^the condenser, respectiveY^,
and W^ 'is the electric power consumption of the compressor.
i— Collector
\ (Convective
\ evaporator)
\
rTree-way valve
.Feeler bulb
(
II
1
I'M
-H*
; i
11 *m
! \
1
1
M
]!
!>)
(
1-
¦¦U
Wi.
tf
mulator
Fig.
Thermostatic
expansion valve
Collector
(Radiative
evaporator)
Compressor
I \ sCondenser
/ 'Flow meter • Thermo-
'Thermobath with pump couples
1 A heat pump system which uses one or two
collectors for experiments
-------
1657
Date:
March 16
ua
w, 1
= 40 °C
O
tec,1-
• tec,2
_ Wc=276 • COP
^V" 2 J
12 14
Time h
Fig. 2 Experimental results for the
system using the convective
type collector.
4J
'o -5
0)
-10
. 1 I •[ 1
I 1
Date: March 1 6 ,P"
Time
010:00
- /
011:00 .
a 14:00
~ 16:00
617:00
1 1 1.1 .
• i
Fig.
10
20
30
Temperature distribution
on the tube of the
convective type collector
from the inlet to the exit.
Uw= 290 t/h
tw.i= AO C
Wc = 265
-297 W
COP
Date:
December 25
-10
h
Time
Fig. 4 Experimental results for the system using
the radiative type collector.
RESULTS AND DISCUSSIONS
The experimental results obtained when the convective type collector was used
are shown in Fig. 2. The water flow rate Uw was kept at 290 1/h. The water
temperature at the inlet of the condenser t^i was kept at 40 °C. The wind
velocity V was in the range between 1 and 4 m/s during the time of the
measurements. The ambient temperature ta was 8«3°C at 9:00 p.m. and the highest
temperature was 12.8°C at 1:00 p.m.. The temperature at the exit of the
collector tec 2 was about 3°C higher than the one at the inlet teCf^ . It
can be known that COP did not much depend on the solar incident radiation
I and that COP did not decreased even when the solar radiation intensity was
small. Figure 3 shows the temperature distribution on the copper tube of the
collector from the inlet to the exit. Lc is the distance from the inlet along
-------
1658
the tube. The decrease of the tube temperature tec with Lc is caused by the
drop of the evaporation temperature due to the pressure loss of the refrigerant
along the tube. The largest increase of the evaporation temperature by the
solar radiation was about 4°C.
The experimental result. for the case of radiative type collector is shown
in Fig. 4. F„ in the figure is the mass of the refrigerent enclosed in the
heat pump. A liigh COP of 5.3 was obtained artiund noon. After 4:00 p.m., COP
was about 0.28 even when the ambient temperature was more than 12°C.
llw = 290 1/h tWi, = 40 "C
Fq =1.2 kg Wc = 267-285W
Date:
December 23
¦ COP
Time h
Fig. 5 Experimental results for the system using
the two types of collectors, tw,l=40 °C.
Figure 5 shows the results obtained when the two collectors were used. The
coefficient of performance was more than 4 between noon and 3:00 p.m., but
it was not high before noon when the radiation intensity was high. When the
solar radiation increased at 9:00 a.m., COP decreased, giving negative
effects of the collector arrangement on the performance of the heat pump.
The temperature distribution along the copper tubes of the collectors is shown
in Fig. 6 and Fig. 7. Lr is the distance from the inlet of the tube. The data
taken at noon, 1:30 p.m. and 2:30 p.m. indicate that the evaporation
temperature could not go up beyond a little above the ambient temperature.
When the solar radiation increases, the pressure as well as the evaporation
temperature in the radiative type collector increases. The pressure rise in
the radiative type collector increases the evaporation temperature as well
as the pressure in the convective type collector because the collectors are
connected at both exits through a tube. The temperature rise of the convective
type collector makes it difficult to collect heat from the ambient and this
results in a smaller flow rate of the refrigerant. Eventually, the thermal
expansion valve is closed, and the flow of the refrigerant through the
convective type collector stops. When there was too much solar radiation on
the radiative type collector, some of the super heated vapor produced at the
collector would flow into the convective type collector and condense, at a
little above the ambient temperature. This would cause lack of the refrigerant
in the loop of the system, resulting in the poor performance.
-------
1659
0 9:00
Date:
December 23
Uw = 290 1/h t^AO'C
Fq =1-2 kg Wc = 267-287 W
20
30
Lr
Fig. 6 Temperature distribution on
the tube of the radiative type
collector from the inlet to
the exit.
(S
¦p
i
o
at
¦P
Date:
December 23
20
Lc
Time
9:00
00
Uw =290 l/h
Fq =1.2 kg
•w.i =/<0 °C
Wc =267-287 W
30
Fig.
Temperature distribution on the
tube of the convective type
collector from.the inlet to the
exit.
Date:
January 27
Uw= 270 l/h
Fq = 1.2 kg
s
II
5>
u>
-203 W
d r/\
¦ cop
• ta
A 1
O V
O *w,1
Time h "me n
Fig. 8 Experimental results for the system using
the two types of collectors, tw,l=9.5°C-
12.7 °C.
Figure 8 also shows the results for the case of the system using two
collectors. The temperature of the water at the inlet of the condenser was
about 12°C. In this case, COP was high at a high solar radiation.
-------
1660
CONCLUSIONS
The thermal performance of a heat pump system using the convective type was
fairly good even when the solar radiation intensity was small, but the solar
radiation did not much increase COP. On the other hand, with the radiative
type collector, COP was very high when the solar radiation intensity was high
and very low when there was no or little solar radiation. Not only was the
wind convective coefficient small for the radiative type collector, but the
sky radiation cooling effects would give negative effects. Using the two
different types of collectors, the weak points of each collector would be
improved under certain conditions; but the arrangement of the collectors gave
negative effects on the performance when the solar radiation was too much.
Devices such as check valves at the exit of the collectors for preventing
the refrigerant from flowing into collector from the exit of the other
collector would be expected to be useful to improve the system.
REFERENCES
Chaturvedi, S. K., Y. P. Chiang, A. S. Roberts, Jr. (1982). Analysis of two-
phase flow solar collectors with application to heat pumps. Journal of Solar
Energy Engineering, 104, 358-365.
Fujita, M., H. Kasagi, K. Tsuchiya, M. Otsubo, N. Otsuka, and H. Yoshino (1983).
A solar assisted heat pump system using direct expansion solar collectors.
Proceedings of the Eighth biennial Congress of the International Solar Energy,
Perth, Australia, 1347.
Hino, T. (1984), Solar Engineering, 435.
Iwanaga, S., T„ Watanabe, M. Ohama, T. Sakurabu, S. Hattori, and Y. Fujimoto
(1985). Solar water heater using direct expansion collector heat pump. Proc.
Int. Symp. Thermal Application of Solar Energy, Hakone, Japan, 53.
O'Dell, M. P., J. W. Mitchell, and W. A. Beckman (1984). Design method and
performance of heat pumps with refrigerant-filled solar collectors. Journal of
Solar Energy Engineering, 106, 159-164.
Sadler, G., J. D. Dale, J. Kirdiekis, L. Doskoch (1987), A solar assisted heat
pump. Sol. 87 (USA), 338-342.
Shinobu, Y., K. Matsuki, M. Yoshikawa (1987), A prototype direct-expansion solar
heat pump system. Proceedings of 1987 ASHRAB Winter Meeting, New York, 615-625.
Shinobu, Y. and K. Matsuki (1990). Recent technical trends in direct expansion
solar heat pump systems. Proceedings of the 3rd International Energy Agency
Heat Pump Conference, Tokyo, Japan, 12-15 March 1990, 487.
Sigma Research, Inc. (1982). Direct expansion solar collector and heat pump, US
DOE Rep. DOE/SF/1Q542-T1.
-------
1661
AN EXPERIMENTAL STUDY OF LITHIUM CHLORIDE SOLAR
LIQUID DESICCANT SYSTEM
L. C. Chen, F.K. Kao and T. Tang
Energy and Resources Laboratories, ITRI,
Chutung, Hsinchu, Taiwan, R.O.C.
ABSTRACT
A solar liquid desiccant system using lithium chloride solution
was built. It consists of nine pieces of solar regenerator/
collectors with each collecting area 1.96nf. The dehumidifier
of this system is a packed bed tower with 38cm diameter and
45 cm packing height. The performance of the system was tested
under the summer and fall seasons in Taiwan. The average water
vapor condensation rate in thedehumidifier is 2.4 kg/hr and the
average solar efficiency in regenerator/collectors is 18.IX.
A linear relationship between solar efficiency of regenerator/
collector and inlet concentration of liquid desiccant can be
obtained from experimental data.
KEYWORDS
Liquid desiccant; solar regenerator/collector; dehumidifier;
packed bed tower; lithium chloride.
INTRODUCTION
The dehumidification process in a solar liquid desiccant system
is resulted from the tendency of vapor pressure equilibrium in
its dehumidifier where the air with greater vapor pressure
contacts directly with cool and strong desiccant solution. The
dilute liquid desiccant solution is then pumped to the solar
regenerator/collector to release water vapor and become
concentrated again. The solar liquid desiccant system has
received considerable attention due to its ability to accept low
grade thermal energy and simple system construction.
Recent results (Chen, 1989a, 1989b) indicate that an open type
solar regenerator/collector with glazing was suitable for
-------
1662
regenerating weak desiccant solution while a packed bed tower
with plastic packing worked very well as a dehumidifier. The
purpose of this research is to build an experimental liquid
desiccant system to study its performance under Taiwan's hot and
humid climate.
EXPERIMENT
The layout diagram of the experimental system is shown in Fig.l
Which is composed of a dehumidifier, nine regenerator/
collectors, a shell and tube heat exchanger, a cooling tower,
and a solution storage tank. The packed bed dehumidifier with
130.5 cm high and 38 cm internal diameter was built by 1 cm
thick acrylic plate. Inside the tower are a 45 cm height mesh
type plastic packings above the packing material supporter and a
mist eliminator. Outside the tower, there is a rubber foam
insulation with 3 cm thickness to approach adiabatic operation
condition. The liquid solution is distributed to the packings
uniformly by a nozzle on the top of tower. The solar
regenerator/collector was made of wood with a waterproof rubber
WET AIR- DRY AIR
I DEHUMIDIFIER
SOLUTION TANK
COOLING TOWER
TEAT EXCHANGER,
REGENERATOR/COLLECTOR
Fig.l. Layout diagram of the experimental
solar liquid desiccant system
-------
1663
lining and a glass cover. Its dimension is 1 m wide and 2 m
long, with two 1 cm openings in upper and lower ends. The
regenerator/collector was faced south and tilted 30 degrees from
the horizontal. The weak solution is pumped to the top of the
regenerator/collector, and flows downward forming a thin liquid
film covering the entire surface of the black jute absorber. In
this process water is evaporated from the hot solution to the
air between glass cover and absorber. The warm humid air then
flows out to the atmosphere by natural ventilation.
The hourly average values of temperature, absolute humidity, and
solar insolation were collected by a Yokogawa hybrid recorder.
Temperature was measured by using T-type thermocouple. Humidity
was measured by a Ahlborn psychrometer. Solar insolatin was
measured by an Eppley pyranometer. Solution flow rates were
measured by using rotameters which were calibrated in different
Li CI concentrations and temperatures. Air flow rate was measured
by a turbine flowmeter. The concentration of Li CI solution was
determined by measuring the refractive index in a Milton Roy
refractometer.
RESULTS AND DISCUSSION
The performance of this system was studied under both decoupled
and coupled modes. In a decoupled operation mode, the
regenerator/collectors were operated from 9:00 to 15:00, which
translated solar energy to useful vapor evaporation energy.
Therefore the regeneration efficiency r) was calculated and the
relation between n and desiccant concentration Xg was shown in
Fig.2. A linear experimental equation was correlated as follows,
rt=- 1.39XS+64.41 (1)
where n was expressed in % and Xg was in weight %. The ranges
of experimental climate parameters are shown in Table 1, where
Tg is ambient temperature, V 1 is wind velocity, It is solar
insolation and H is relative humidity of ambient air.
The relation between condensation rate D in g/min and Xg in the
decoupled operation mode of the dehumidifier was also linear as
shown in Fig.3. The correlated equation is
D=3.21XS—68.89
(2)
-------
TABLE 1 Ranges of Climate Parameters
Climate Parameter Range
Ta, °C
V, m/s
It, Whr/nf
H, RH%
20.8-32.5
1.5-6.1
3860-4950
60-80
31 32 33 34 35 36 37
Xs f*t X)
Fig.2 Relation between 77 and Xg
31 32 33 34 35 36 37
Xs (*t X)
Fig.3 Relation between D and Xg
-------
1665
The ranges of the operation parameters for Eq.(2) in this
experiment are shown in Table 2, where Mg is solution flowrate,
Tg is inlet solution temperature, Hg is inlet air flowrate, and
H is absolute humidity of inlet air.
d
In the coupled operation mode, the regenerator/collectors and
dehumidifier worked together. The experiments began at 9:00 and
ended at 15:00. The average values of parameters D and n for
each day calculated from measured data are shown in Table 3,
where H_ is absolute humidity of inlet air in dehumidifier. From
Q
Table 3, the average solar efficiency and average condensation
TABLE 2 Ranges of Operating Parameters
Operating Parameter
Range
Ms, kg/ufmin
57.8-59.0
T_, °C
29.9-32.0
* O
M_, m/min
5.86-5.95
Ta, °C
30.0-30.6
Ha, g/kgDA
18.0-19.5
TABLE 3 Results of the Coupled Operating Mode
Ta
I
Xs
Ha
n
D
Date
~
u
°c
Whr/m2
wt%
g/kgDA
%
g/min
9/26
33.0
4374
37.50
18.96
15.8
50.7
9/27
30.8
3168
37.30
19.27
15.5
54.3
10/9
27.7
4050
35.34
13.04
19.3
31.3
10/11
31.2
4608
35.32
16.92
19.1
46.8
10/30
28.3
4356
35.18
12.25
19.8
32.7
11/2
27.8
3762
35.84
11.31
17.1
27.2
11/3
27.0
3174
36.10
13.99
17.5
40.5
11/6
29.4
3936
34.73
14.08
20.8
32.3
-------
1666
rate of the system are 18.1% and 39.5 g/min. When compared to
the decoupled mode, the same trend can be observed, that is the
higher the desiccant concentration the lower the solar
efficiency and the larger the condensation rate. The absolute
humidity of inlet air also has great effet to condenation rate,
the higher H_ results in the higher D.
O
CONCLUSIONS
An experimental solar liquid desiccant system was built and
tested. The performance of the system was investigated under
decoupled and coupled operation modes. Linear relations between
ri and Xs, and D and Xg were obtained in the decoupled mode.
It shows that Xg has positive effect to D and negative effect
ton • A 35 weight % of LiCl solution is suggested when
considering these two conflicting effects. From the experimental
results in this study, it appears that solar liquid desiccant
cooling system has good potential for hot and humid climate
regi on.
ACKNOWLEDGMENT
The financial support by the Energy Commission of the Ministry
of Economic Affair in Taiwan R.O.C. is sincerely appreciated.
REFERENCES
Chen, L.C., and F.K. Kao (1990). The experimental study of three
types of regenerator/collector for solar liquid desiccant
cooling. Proceedings of the 1989 Congress of the ISES, Kobe,
Japan, 840-844.
Chen, L.C., C.L. Kuo, and R.J.Shyu (1989). The performance of a
packed bed dehumidifier for solar liquid desiccant systems.
Proceedings of the 10th ASME Solar Energy Conference, San
Diego,California, 371-377.
-------
1667
RENEWABLE ENERGY SPACE COOLING
D.R. Neill, M. Bean, T. Ho, and L. Huang
Hawaii Natural Energy Institute, University of Hawaii
2540 Dole St., Holmes Hall 246, Honolulu, Hawaii 96822 U.S.A.
ABSTRACT
This paper presents seven options where renewable energy technologies can be used to reduce
energy costs to cool the interior of buildings, enhance the comfort level for the occupants, eliminate
mildew, and, in some instances, improve the interior building health quality.
KEYWORDS
Desiccant; heat pipe; heat pump; deep cold water, solar power; photovoltaics; hybrid.
INTRODUCTION
The energy used in warm, humid climates, like Hawaii and other warmer sections of the U.S., to
cool the interior of commercial buildings is about 50 percent of the building's energy requirement.
In sub-tropical or tropical areas, like the Philippines and Indonesia, it may be as high as 80 percent
of the total energy consumed in buildings. Yet these areas have abundant solar energy, and some
have an additional resource in cold, deep water within one or two miles of the shoreline that can be
used to dehumidify and cool buildings. Renewable energy space cooling will become even more
important if the world keeps warming, as some predict, with the global warming effect.
There are seven approaches where renewable energy technologies can be used to cool or air
condition a building: desiccant and heat pipe dehumidification systems, heat pump, deep cold water
cooling, solar powered absorption air conditioning systems, PV powered compression systems, and
hybrid energy systems. The first two systems are now under test and evaluation for two separate
rooms in Sinclair Library on the University of Hawaii at Manoa Campus.
The use of these dehumidification systems can not only greatly reduce the energy requirement (by
cooling drier air) but also can improve the building health condition by removing "bad guys" from
the air (curing the sick building syndrome), reduceor eliminate any mildew problems, and improve
the comfort level in the building by providing a less humid or drier condition (which can also allow a
higher temperature and still provide a good comfort level).
The growing concern with the sick-building syndrome emerged in the 1960s when indoor air
pollution was first recognized. Air quality was already being predicted as a major problem for the
1990s (Yaukey 1990). In 1989 the Environmental Protection Agency (EPA) expressed its concerns
about indoor air pollution to the U.S. Congress, stating, "Sufficient evidence exists to conclude that
indoor air pollution represents a major portion of the public's exposure to air pollution and may pose
serious health risks" (Kerch 1990). Both Houses of Congress began considering an Indoor Air
Quality Act to provide funding to the EPA for a five year program to gather more data.
-------
1668
The mission of the Hawaii Natural Energy Institute's and the State of Hawaii's renewable energy
space cooling program is to save energy, improve the health environment in air conditioned
buildings, eliminate mildew, and enhance the comfort level for the occupants. This paper discusses
the results to date of these efforts and provides information on the options that will prove to be
increasingly important alternatives in replacing polluting, fossil fuel powered space cooling systems.
DESICCANT AND HEAT PIPE DEHUMIDIFICATION SYSTEMS
The State of Hawaii funded a demonstration project to install a desiccant dehumidification system
utilizing solar and waste heat recovery water heating to regenerate the desiccant. Baseline data was
obtained (Table 1). The project was expanded in scope to include heat pipes in one 20 ton package
DX air conditioning system and the solar regenerated desiccant unit in an adjacent 20 ton system on
the roof of the Sinclair Library on the University of Hawaii Manoa Campus. The desiccant unit is an
ASK Energymaster, rated at 1200 CFM of supply air. It is being regenerated by four 4 ft by 10 ft
(1.2192 m by 3.048 m) selectively coated solar collectors and a water-to-water heat pump. The
chilled water produced by the heat pump is supplied to the coil originally designated for space
heating in the ASK unit. The unit is used to dehumidify the make-up air for the audio visual storage
room. An identical 20 ton package DX air conditioner on.the same building was retrofitted with heat
pipes across the supply and return air ducts to increase the dehumidification capability of the existing
air conditioner.
The heat pipes are installed in the return and supply air duct work of one air conditioner. In this
manner, the heat pipe in the return air duct precools the air before it gets to the DX coil, which
allows the coil to cool to a lower temperature, decreasing the absolute humidity of the air. This air
then passes through the second heat pipe in the supply air duct, which acts to reheat the air, reducing
the relative humidity. The process of boiling and condensing the working fluid (R-22) in the heat
pipe will normally reduce the sensible capacity of an air conditioning system about two percent while
it doubles the condensate removal across the DX coil.
This particular installation was a learning experience. The geometry of the heat pipes is important.
This system was installed with two separate fin tube assemblies in ducts approximately six feet
apart. The return air heat pipe (evaporator) is approximately six inches lower than the supply air heat
pipe (condenser). The optimum geometry for a heat pipe is to have one vertical tube, with
evaporation taking place in the bottom of the tube and condensation taking place in the top of the
tube. In this optimum configuration, a two row fin tube unit that boils and condenses in the same
tubes will produce approximately a 5° F increase and decrease in air temperature. In this case, we
observed approximately 0.5° F difference. However, the humidity was reduced from 74° F/74%
RH in July 1989, the day before the heat pipes were installed, to 74° F/51% RH in September 1989.
From this limited test it appears that the lower performance installation will gradually pull the
humidity down, where the optimum heat exchanger placement will produce more dramatic results.
Since this is a library, the mass of the books acts to store a large volume of moisture, which requires
some time to reach a new equilibrium with a change in humidity. In any case, this appears to be a
technology that will be of benefit to air conditioning systems in humid climates.
TABLE 1. Baseline Data Acquisition
11/1/88 11/4/88 11/19/88
Weather Condition
Ambient Temperature
Room Temperature
Ambient Relative Humidity
Room Relative Humidity
Electricity Consumption
Cloudy Partial Sunny
26.2° C 26.7° C 26.6° C
22.2° C 23.3° C 21.5° C
74.5% 79.6% 73.1%
56.2% 82.6% 58.6%
549 kWh 572 kWh 541 kWh
An ASK Energymaster was installed to dehumidify the make up air for the other 20 ton package air
conditioner. This unit was originally budgeted to use a desuperheater on the 20 ton compressor to
-------
1669
provide the heat for regeneration along with the solar panels. However, due to the questionable
condition of the compressor, it was decided to assemble a heat pump to provide the auxiliary heat.
Since it was necessary to find a heat source for the heat pump evaporator, a chilled water loop was
installed to utilize the space heating coil in the Energymaster to cool the dehumidified air sensibly
before it was introduced to the building. Start up problems with the ASK have prevented the taking
of data as of yet, but initial results look promising.
Four categories of data will be measured for future analysis and evaluation of the two systems.
1. Human comfort: including indoor air temperature, relative humidity, and air changes or
ventilation quantity (and quality if possible).
2. Related data: Ambient temperature/relative humidity outside the building, solar insolation,
wind speed and direction, and building cooling utilization data (start/end, length of
operation).
3. System performance: energy used, resultant temperatures and relative humidity, heat gain
(infiltration through windows, roof, and walls; insolation; indoor lighting; and heat ejected
by occupants and equipment), and specific items as listed
• outdoor ambient temperature and relative humidity,
• supply and return air flow rate, temperature, and relative humidity,
• working fluid flow rate plus inlet and outlet temperature,
• condition of air before and after the air processing unit,
• energy consumption for pumps and fans,
• energy consumption for refrigeration unit, and
• solar heater/drier performance.
4. Economics: energy used/saved, capital costs, operating and maintenance costs, and
resultant life cycle costs.
The principal approach in this design is to utilize the heat rejected by the compressor to regenerate a
desiccant, which will dehumidify the make up air. The low side of the system then sensibly cools
the air. A survey of users of these two spaces was taken before and after the systems were installed.
Respondents were asked about their level of activity, their clothing, and their comfort level in the
room.
HEAT PUMP
Several years ago the principal author installed a heat pump in a two bedroom condominium unit
above the refrigerator unit. The primary purpose was to heat the domestic water for that unit. The
water heater was located under the kitchen counter. The reported coefficient of performance (COP)
of a conventional heat pump is 3, i.e., for each kWh of electricity used, a three kilowatt equivalent of
thermal or heat value will be gained. This is accomplished by the system taking the heat from the air
around the unit. A second benefit was to remove some of the hot air released by the refrigerator,
which helped make the non-air conditioned unit somewhat cooler.
Besides a small concern with the noise level of the heat pump, there was a major concern raised by
the manager about the disposition of the fairly large amount of water (condensate). Because the
kitchen was located on an interior wall, a small hose was connected to drain the water into the sink.
Because this was unsatisfactory, the project was terminated.
With the awareness of the benefits of dehumidification now identified, plans are to reinstall system.
Since the original test, the unit (along with 41 other units in the complex) has had air conditioning
units installed. The proposed plan is to measure the amount of energy the three air conditioning
units consume, as well as the solar water heater, with and without the heat pump operating. A
significant savings is anticipated both in the air conditioning as well as in the water heating
requirements.
-------
1670
Secondly, the condensate will be collected and analyzed for the main elements in it, especially those
that may affect the health quality of the interior room. Again, this will be measured with and without
the air conditioning systems operating.
Once this is adequately documented, the economics and health gains will be developed and,
hopefully, others will be encouraged to utilized heat pumps for similar applications. If possible, the
condensate may be used to water plants, which will provide another gain in the disposal of the
water, or it will be piped into the drainage system.
Thomas York (1990) wrote a feature article that discusses the value of heat pumps to an electric
utility system. He says that the dual-fuel heat pump (DFHP) incorporates a winning combination for
commercial space conditioning — an electric heat pump for high efficiency and a gas furnace for
economical supplementary heating during cold snaps. The DFHP can be programmed to choose the
optimal mix of electric and gas operation to deliver uncompromising performance at the lowest cost.
The unit's programmability also provides a hedge against fuel price changes and allows the customer
to take best advantage of a utility's time-of-use rate incentives.
The electric utility representatives promote heat pumps as a better alternative to solar water heating
because one can save up to two-thirds of the water heating cost while, at the same time, partially
levelizing the demand on the utility (i.e., the utility demand is not increased greatly during cloudy
conditions). However, the heat pump, which has been called an in-side-out refrigerator, will need
replacing every 8 to 12 years. Although some heat pumps have lasted that long, this is a situation
that should be improved.
DEEP COLD WATER COOLING
HNEI completed a preliminary study of the use of deep cold ocean water (DCOW) to cool
commercial buildings, especially hotels. The DCOW of 6° to 7° C can be pumped from the ocean
(from a depth about '700 meters) to the buildings near the shoreline where it can be used in several
ways. It can cool the chiller water directly through a heat exchanger. Thus the conventional chiller
that consumes much electricity can be replaced (the conventional system can be kept as a backup).
The DCOW can be used to replace the cooling tower, which is the standard for heat rejection, to cool
the condensing water (Fig. 1). This can cut the electricity consumption of the cooling tower and also
save fresh water. The DCOW can also be used to lower the condenser water temperature further
through a heat exchanger to raise the efficiency of the chiller. The study showed that a payback
period of less than four years is possible, depending mainly on the distance between buildings and
the location where the DCOW can be reached (Table 2).
7.0° C
7.2° C
35.0° C
13.3° C
29.4° C
35.0° C
Conventional
Chiller
Cooling
Tower
Building
(Hotel)
Fig. 1. Replacing the cooling tower with DCOW as a heat sink.
-------
1671
TABLE 2. Payback Calculation
Loan Due
Annual Net Saving
Loan Balance
Year 1
629,752
216,950
412,802
Year 2
454,082
216,950
237,132
Year 3
260,845
216,950
43,895
Year 4
48,285
216,950
- 168,665
The cooling tower is the standard for heat rejection in large air conditioning systems. The two
means of rejecting heat in a cooling tower are the sensible cooling of the warm water by the air
stream and the latent cooling by evaporation. The water that was evaporated must be made up by
clean, fresh water, which competes with the population for potable water supplies. The rule of
thumb for cooling towers is one percent of the water flow will evaporate for each 12.6° F of water
temperature change. A 500 ton chiller with 1,500 gallons per minute condenser water flow would
then require about 15 gallons per minute make up water, or 21,600 gallons per 24 hours of
operation. The use of ocean water, or brackish well water, will eliminate both the use of potable
water for cooling towers and the need for the towers. Due to the lower temperature condensing
water, the efficiency of the chillers will increase, which could result in significant energy savings.
Several buildings in Honolulu have had good results with the use of ocean and brackish well water
in lieu of cooling towers.
Of course, the best approach is to eliminate the chiller altogether by bringing up water that is cold
enough to cool and dehumidify the air without extra chilling. Care must be taken in the building's
mechanical design if there is a chance of the water not being cold enough to dry the air properly. An
earlier program of "thermal discomfort" in the name of energy conservation resulted in many
buildings shutting off the fresh air to the air conditioning system and raising the chilled water
temperature. This did result in energy savings, and is also one of the principal causes of the
humidity problems suffered in many buildings. If water cold enough to dehumidify the air properly
cannot be delivered to the building, then extra chilling must be used. In addition, the use of heat
pipe coils should be considered to increase the dehumidification of a system.
SOLAR POWERED ABSORPTION AIR CONDITIONING SYSTEM
Another alternative is a solar powered absorption air conditioning system. Flat plate solar water
heating systems are capable of producing temperatures less than 200° F, which is not sufficient for
powering an absorption system. However, evacuated tube solar, parabolic trough, and combination
trough thermal/electric systems can provide temperatures sufficient for this application. A high level
of direct sunlight is needed for concentrator systems. The evacuated tube system may be better than
a concentrator system under a partly cloudy condition.
PV POWERED COMPRESSION SYSTEM
Photovoltaic (PV) systems are another option, possibly when a concentrator system might also
prove cost effective where both electric and thermal energy are produced to power a space cooling
system. These are also being explored by HNEI and its China and Pacific Area cooperative
programs. For remote, non-utility powered areas, PVs are becoming cost effective today and could
be cost effective for powering small air conditioning systems.
HYBRID ENERGY SYSTEMS
As noted in the discussion of the desiccant dehumidifying system, a solar system is used to
regenerate the desiccant. A heat pump is also used to enable the desiccant system to work more
efficiently during non-sunlight hours. HNEI, under its China program, is having a one-ton heat
pump/heat pipe hybrid system designed and built in China for test and evaluation in Hawaii.
Another concept is a type of co-generation system, i.e., when there is waste-heat or a thermal by-
-------
1672
product, or a possible "combined system" concept, where electrical generation (e.g., solar thermal
electric or PV-thermal system) where water heating, space cooling, and other applications of thermal
might be combined for a more economic overall system. Another concept that has merit is the
making of ice with surplus electricity, either from renewable energy technologies, like wind power,
or off-peak power, to lower the cost of the generating system. The ice is then used as the pre-chiller
in the air conditioning system during the day. The Hawaii State Legislature has provided a
significant tax credit for this, as well as 35 percent credit for solar water heaters and PV systems,
and is currently considering tax credits for the incremental added costs for energy efficient space
cooling systems.
CONCLUSION
There are many ways to save energy and improve the comfort and health level of interior spaces.
This paper highlighted some of the ideas being pursued by HNEI and Hawaii to this end. The plan
is to emphasize the importance of energy efficiency, especially when there are so many benefits.
Each one has merit. A good data base on the results of each system under evaluation is needed.
There is also the need to educate decision-makers to ensure good technologies are utilized.
Space cooling is increasingly needed in many buildings year around in warm areas of the U.S. and
the world, as well as during the summer in many places. The added concern of internal pollution, or
"sick building" syndrome, must also be addressed. Improving the comfort level of the occupants
can improve their productivity as well as their attitudes. Energy efficiency is the best source of
energy, and it is renewable.
REFERENCES
Kerch, S. (1990). Buildings that make people sick: often a problem without a solution. Chicago
Tribune. In The Sunday Star-Bulletin & Advertiser. Honolulu. 19 August, p. B9.
Yaukey, J. (1990). Dust diagnosed as cause of sick-building syndrome. The Ithaca Journal. In
The Honolulu Star-Bulletin. Honolulu. 25 December, p. Dl.
York, T. (1990). Heat pumps, developing the dual-fuel option. EPRI Journal December, 22-27.
-------
2.15 Active Solar
-------
-------
1675
Computing Incidence Angle Modifiers for Advanced Solar Collectors
K. Knappmiller and W. Duff
Department of Mechanical Engineering, Colorado State University,
Fort Collins, Colorado, USA
ABSTRACT
A computer program using 3-D Monte Carlo Ray tracing was created to calculate the amount
of solar radiation absorbed by solar collectors. A second order method was used so that
advanced geometries such as those found in evacuated tube collectors could be modelled
accurately. The results from this program were used to obtain Incident Angle Modifiers (IAMs)
for the entire hemisphere over a collector. These results were compared to those obtained by
the currently accepted approach of a multiplicative combination of the bi-axial IAMs. These
indicate that for highly asymmetrical collectors at angles greater than 30 degrees from normal,
a simple multiplicative combination of bi-axial IAMs may result in unacceptably inaccurate
predictions of the solar energy absorbed.
KEYWORDS
Incidence Angle Modifiers; Evacuated Tube Solar Collectors; Constructive Solid Geometry;
Monte Carlo Method; Parallel Computing;
INTRODUCTION
When constructing a computer model to predict the performance of a solar collector, it is
necessary to have an accurate estimate of the fraction of incident solar radiation which is
absorbed. This fraction is dependent on the incident angle of the radiation. For flat plate
collectors this dependence on incidence angle is relatively easy to calculate, but for the more
complex geometries of advanced collectors it is very difficult to get an accurate answer.
Currently, performance curves estimated from collectors with similar geometries have been used
to get an incident angle modifier curve for each axis which relates the energy absorbed at a
particular angle to the energy absorbed at normal incidence. To get the IAM at a particular
angle that is not coincident with either axis, many modelers have used a simple multiplicative
method (Mclntire 1982) of multiplying the two axes' IAMs. The accuracy of this method was
questionable, especially for complex geometries when absorption at high incident angles may be
significant. A method was developed to compute the fraction of solar radiation absorbed at a
specific incident angle. This method was then applied at various angular grid spacings over the
hemisphere above a collector. Results were computed for a variety of evacuated tube collectors:
Internal Compound Parabolic, 'Sun Family' Internal Storage, and slightly modified Corning and
'receding page blank
-------
1676
Phillips geometries.
COMPUTER MODEL
The collector model is constructed using second order solid primitives, including cylinders,
cylindrical paraboloids, spheres, cones, and first order block primitives. These primitives are
rotated, scaled, translated and then combined with the intersection, difference, or union
operations to achieve the appropriate geometry. A sub-program for the IBM PC in VGA mode
was then used to display the geometry. Each primitive was assigned as a transmissive, reflective
or absorptive object. The transmissive object properties were the index of refraction and
absorptivity per mm of material. The reflective object properties were the absorptivity and
diffuse fraction. The absorbing object property was absorptivity. Because of the geometries
involved, once a ray reached an absorber, it was traced no further. The contribution of rays
reflected from an absorber was assumed to be minimal because, for the geometries considered,
it would have to be reflected at least once before having a chance of being absorbed again.
Because of the high absorptivity (low reflectivity) of the absorbers considered, the contribution
would be much less than 1 %. If different geometries were explored in which this contribution
was significant, the code could be easily modified to include it.
Once the geometrical model with associated physical properties was constructed, ray tracing
could be conducted. Classic optical methods were used at ray-transmissive-surface intersections
to compute the probability of reflection or refraction. All fractional properties of reflection,
refraction, and absorption were treated as probabilities and go or no-go decisions were made
with a random number generator. The ray was traced in only one direction from such an
intersection. This concept combined with the casting of rays from random points in the plane
above the collector at a specific angle, is the essence of the Monte Carlo method. Rays traced
are thought of as photons and are either absorbed by the absorber or not. The absorbed fraction
was obtained to the desired accuracy with a certain associated probability by casting a sufficient
number of rays (Maltby, 1990).
IMPLEMENTATION
The code was developed on an IBM compatible PC using Borland Turbo C 2.0. The code was
written with object based computing techniques in mind, but at the time of implementation an
object based language compiler such as C++was not available on either of the hardware
systems on which the program would need to run.
The PC was used for developing the input file and graphically displaying the resultant geometry.
To compute the IAM to an accuracy of 2 percent for a typical collector geometry such as the
Sun Family requires approximately 10 minutes on a 20 Mhz 80386 machine for an incidence
angle at which the absorbed fraction is approximately 50%. The time required is considerably
greater for smaller fractions. To achieve hemispherical grid spacings of sufficient density such
that there is a maximum angle of 10 to 15 degrees between grid locations requires use of about
100 grids.
This density resulted in computational times of greater than 10 hours on the PC, but the problem
solution lends itself to distributed parallel computing. A Local Area Network (LAN) was
available with many DEC 2000 workstations which were unused at off hours. A UNIX C-shell
-------
1677
script was used which searched the network for computers with no active users, loaded the
program on these machines, each with its appropriate input vector and concatenated all the
results into one file once all runs were completed. Using this technique it was possible to
analyze a specific collector geometry in less than 1 hour at a grid density which would allow
accurate linear interpolation over the entire hemisphere.
Since this run time was acceptable.no attempts were made to optimize the code itself. Rather,
it was attempted to keep the code as simple and expandable as possible. Object based
programming techniques and modularity were used. This was intended to allow additional
capabilities to be added with a minimum of reprogramming. An example of such modularity
is the addition of another geometric primitive. When it was desired to model a collector with
parabolic concentrators it was necessary to add the cylindrical paraboloid primitive. This
addition was easily accomplished. Another example would be the addition of an absorbing
surface with an angular dependence of absorptivity. The style and structure of the code makes
such an addition possible in several hours. The C programming language lends itself to such
code construction much more readily than FORTRAN. Unfortunately, C is inferior to C++ in
this respect. It is believed that recoding in C++ would make further expansions possible in
half the time, especially if such expansions were to be completed by persons not intimately
familiar with the code.
Incidence Angle Modifiers
Sun Family Collector
2.0 r
Latitudinal
Longitudinal
Computed Results at 45 Degrees
Multiplicative Results at 45 Degrees
0.4
0.2
0.0
0
20
30
40
50
70
1 0
60
Angle from Normal in Degrees
Fig. 1. Sun Family IAM
-------
1678
RESULTS
The Sun Family collector module consists of four 2 meter long tubes, each 126 mm in diameter.
Inside each tube is a 110 mm diameter absorber. The 4 tube module is a total of 630 mm wide.
The incidence angle modifiers computed are shown in figure 1. As can be seen, the
multiplicative method yields results which are reasonably accurate. This is due to the nature of
the response of this type of collector. If the IAM along the latitudinal axis grows from unity,
and along the longitudinal axis it decreases from unity, a multiplicative combination of the two
is again close to unity. In cases such as this the IAM value at an off axis incidence angle is
closer to unity than its value at the same angle from normal on the axis. It is very tempting to
use this trend to compute IAM's with the multiplicative method for a wide variety of collectors.
Unfortunately, since the method has no physical basis, it is very easy to generate errors of
unknown magnitude. In many cases these errors may be negligible, especially if a significant
fraction of the incident radiation is at angles within 30 degrees of normal, or at angles close to
one axis, but a method which eliminates the unpredictability of such an error, with little or no
additional cost to the engineer, would be of some value.
Inidence Angle Modifiers
Cosine Collector
longitudinal
latitudinal
Multiplied 45 deg
Computed 45 deg
0.7
Computed
0.6
0.5
Multiplied
0.4
0.3
0.2
0.0
0
10
20
30
40
50
60
80
70
90
Fig. 2. Asymmetrical Flat Plate Collector
In figure 2 is an example of a potential IAM curve, where the longitudinal IAM is a cosine curve,
and the latitudinal IAM is a cosine square curve; this could represent an asymmetrical flat plate
collector. The multiplicative method gives erroneous results when the Biaxial curves both drop
from normal rather than one going up and the other going down. The computed value tends
closer to the average than the product. Having to select an appropriate combinatorial method
adds unneeded uncertainty when a direct computational method is available.
-------
1679
Computing the Bi-axial IAM's for a collector of complex geometry can be very difficult. A
slice of the collector along an axis may be taken and an optical analysis done in this two-
dimensional plane. If the collector surfaces have only a small diffuse component, such a method
will give accurate results, but in many cases the energy contributions from diffuse reflection
is not negligible, especially for collectors that can get dirty. In such a case it is necessary to
conduct a three-dimensional analysis to achieve results to the desired accuracy. A total
hemisphere approach takes only a small amount of additional computation time. For the angular
densities computed here, approximately 150% more time is required, which is usually negligible.
Few collector system modelling programs are capable of using a total hemisphere IAM. For
many collector geometries, the desired accuracy may be obtained with a standard combinatorial
approach, of which the multiplicative method is one example. The results of a complete
hemispherical computation may not be used directly, but they may be used to give a good
estimate of the error incurred in the use of a particular combinatorial method. This program can
be used to compute the two biaxial incidence angle modifiers which are then used in a
conventional collector system modelling program. Only a few off angle IAMs can be computed
to get an engineering estimate of how accurate a chosen combinatorial method is.
CONCLUSIONS
A computer model now exists which can calculate the absorbed radiation at any particular
incidence angle, for any complexity of collector, to any desired accuracy within the accuracy
of specification of surface properties. The technique can be extended to a total hemispheric
computation,if necessary, with no additional programming or model construction. With such
a tool available, it is now feasible to adopt a standard method of using hemispherical incident
angle modifiers in collector system modelling programs.
A three-dimensional field of sufficient density for accurate linear interpolation would be
applicable to any type of collector. Different sub-programs could still be used to generate this
field, depending on the nature of the collector, but the interface would be standard and sufficient
for any type of collector. In most cases this would also result in decreased computation time
as the computation would be linear instead of second order or higher. Shading data could also
be incorporated into the IAM rather than at a later point in the process.
It is recommended that a Total Hemispheric Incident Angle Modifier of 400 elements be used
as the standard interface, resulting in an additional memory overhead of only 1.6 kilobytes at
single precision. This would cover the entire hemisphere with points at 10 degree intervals.
For most collectors, this density would result in an accuracy greater than that of the original IAM
computation. For highly non-linear collectors, it may be necessary to have a greater density.
The interface can be easily modified to use a density of 1600 or 6400 elements.
REFERENCES
Maltby, J.D., (1990) Analysis of Electron Heat Transfer Via Monte Carlo Simulation, Ph.D.
Dissertation, Colorado State University.
Mclntire, W.R., (1982) Factored Approximations for Biaxial Incident Angle Modifiers, Solar
Energy, Vol 29, Number 4, 315-322.
-------
1680
A METHOD TO COMPUTE THE INCIDENCE ANGLE MODIFIER
AND TO ESTIMATE ITS INCIDENCE ON COLLECTIBLE SOLAR ENERGY
C. Armenta-Deu and Boris Lukac (*)
Grupo de Energia Solar. Fac.Fisicas. Univ.Complutense
28040 Madrid (SPAIN)
(*) University of T. aad C. Zilina (Czechoslovakia)
ABSTRACT.
The influence of the incidence of the incoming radiation striking on the
collectible energy by a solar collector is analyzed. The estimation of this
effect permits an accurate prediction of the long-term solar collector
performance and of the energy it could deliver as well. An analytic expression
to determine the monthly mean value of the Incidence Angle Modifier (IAM) has
been developed. The expression is based on monthly mean values of global and beam
radiation (or diffuse) computed from solar radiation data or from geographic and
climatic parameters if solar data are not available. Correlated results obtained
from the different solutions proposed in the computation of the Incidence Angle
Modifier have shown a good agreement with those obtained from meteorological
data.
KEYWORDS
Solar radiation, Collectible energy, Flat-plate solar collectors.
INTRODUCTION.
Efficiency curves of solar collectors are usually reported at normal incidence.
In most of the collectors, however, the efficiency varies with the angle of
incidence. The result is an attenuation of the collected energy; this attenuation
is taken into account in the computation by the Incidence Angle Modifier K.
Reduction in the average value of the collectible energy by a solar collector is
a function of the optic and thermal properties of the solar collector. These
properties should be treated separately to separate both effects and to obtain
a more accurate prediction.
The angle of incidence 8, only affects the optical efficiency^, thus defining
the Incidence Angle Modifier as the ratio of optical efficiencies at a fixed
angle and at normal incidence
K(0) = ?5/?o(0=0) (1)
The method of computation of the Incidence Angle Modifier K, is based on the
determination of an analytic expression depending either on average radiation
(global, beam and diffuse) computed from available radiation data files or on
geographic and climatic parameters for those sites where radiation data are not
available.
-------
1681
MEAN VALUE OF THE INCIDENCE ANGLE MODIFIER.
It has been shown and discussed by a number of investigators (Hottel and Woertz,
1942; Bliss. 1959: Whillier. 1967; Duffie and Beckraan. 1974) that the performance
of a flat-plate collector operating under steady state conditions can be
successfully described by the following relationship:
VAa= It(xa)e" uL(tp_ta) = (m'/Aa)cp(tf,e"tf,i) (2)
from which the efficiency can be determined as
*= FR^a(™>e.n " lH (tti~V)/:t]> <3>
where Kt(J(xa)e n represents the Incidence Angle Modifier previously defined in
equation (1). '
It was suggested by Sofka and Safwat (1966) an expression for the Incidence Angle
Modifier
Kta(xa) = (xa) / (xa)n (4)
being (xa)^ the transmittance-absorption product of the collector at normal
incidence.
The effective value of IAM can be determined through the expression (6),
K(6) = 1 + b0(l/cos9 - 1) (5)
where bQ is a coefficient dependent of the type of glazing coverage. This
coefficient takes negative values as defined in equation (5).
ANALYTIC EXPRESSION OF THE MONTHLY MEAN VALUE OF THE IAM.
Global radiation on the aperture area of a tilted collector can be expressed
according to ASHRAE standards as.
G(a,(J) = Bcos0+ D(l+cos£)/2 + £ 9G( l-cosj3) / 2 (6)
being G, D and B the monthly average irradiance at time t, respectively global,
diffuse and direct normal, p the tilted angle of the plate and f0 the ground
reflectivity.
If we consider now the effect of the angle modifier, the global irradiance adopts
the form,
G*(a,ji) = BKb(0)cose+ DKd(8)(l+cosJ!s)/2 + £ p/6)(l-cos^)/2 (7)
where and represent the IAM for direct normal and diffuse radiation.
Considering the diffuse and ground-reflected radiation to be isotropic, the
average value of by is (1-b, ) assuming an effective average incidence angle of
60Q; therefore, equation (7) is transformed into,
G*(a,Jb) = BKb(&)cos9+ (l-b0){D(&)(l+cos^)/2 + fgG(l-cos^.)/2} (8)
The IAM is considered only if positive,
k(6)>0 for cos(&)>tjj I (l+t(j) (9)
The expressions for the global radiation, given by Lalas and Petrakis can be
averaged over a month assuming 9 to be constant at a given time of the day.
-------
1682
Defining now the IAM as.
K = G'^a.jDlG^a,?) (10)
where and Ga are the average values of G. we have,
K = 1 - b. + b.[2Bj(a,J8) -
B dt ]/G,(«,£) (11)
cosft >b01(l+b0)
Equation (11) enables to compute K from the average daily profile of direct
irradiance B(t) which also can be generalized to account for circumsolar diffuse
radiation.
DETERMINATION OF AVERAGE VALUES OF SOLAR RADIATION.
Average values of solar radiation can be determined either from the radiation
data files source, or from the geographic and climatic characteristics of the
site. Equation (11) can be, therefore, expressed as,
K = 1 + bQ - b0[Bs + 2DI(a,/3)]/Gi(a,j&) (12)
being Gf(a,p) and Da(a,J}) the global and diffuse onto the tilted plane, and the
direct normal onto horizontal plane.
Three different methods have been used to compute monthly mean values of solar
radiation:
a) EUFRAT method. This method is based on computing monthly mean values from
hourly values of solar radiation.
b) HELIOS method. This method uses geographic characteristics of the location
treated to generate solar radiation data from which monthly mean values are
calculated.
c) Analytic method. In this method Bourges proposses an analytic expression is
developed to compute K based on the geometry and the position of the collector.
EFFECT OF K ON THE COLLECTIBLE ENERGY.
If we consider the reduction of the incoming radiation because of the effect of
the angle of incidence the useful global energy is 4>'H', being ' the new
utilizability and H1 the monthly mean collected energy after the reduction. From
the above definition the average value of the IAM is obtained from the ratio,
K = 4>'H'/4>H (13)
which can be simplified for a threshold null,
K = H'/H (14)
DETERMINATION OF THE AVERAGE VALUES OF K.
Monthly values of collected energy (H,H' ) have been computed using radiation data
source (EUFRAT), geographic and climatic parameters (HELIOS), or geometric
variables (analytic method). Twenty-eight locations from seven different
countries of the European Community have been tested. Predictions have shown a
good agreement for all types of correlation. The estimated error of the monthly
mean values of K have not ever surpassed a 3%, except for the cases where
collectible energy is very low. These cases are, however, of low importance
because of the small number of applications.
-------
1683
COMPARATIVE RESULTS OF COLLECTIBLE ENERGY.
To determine the influence, in terms of error, of the IAM on the collectible
energy a comparison between H' and KH should be made. The results of this
comparison showed a good correspondence between all of them with similar error
values in the prediction of the collectible energy at the same time of the year.
In Table I results for a plane facing due south (horizontal, tilted latitude and
vertical) for the location of Athens are presented.
Differences in estimated collectible energy are within a low error (+/- 5%) but
for the months where the collectible energy is very low.
Table I
Experimental and correlated values of the collected energy
by a flat-plate collector
LOCATION : ATHENS
MONTH
* H'
%f H
abs
rel
1
927
970
43
4 . 65
2
1574
1618
44
2.78
3
2444
2515
71
2.93
4
3846
3960
114
2.97
5
4693
4776
137
2 . 96
6
5086
5214
128
2.53
7
5270
5416
146
2.76
8
4616
4786
170
3.7
9
3488
3599
111
3 .19
10
2100
2169
69
3.28
11
1191
1231
40
3.34
12
734
772
38
5.15
1
1932
1984
52
2 . 66
2
2622
2683
61
2.32
3
3262
3344
82
2.49
4
4244
4323
79
1. 85
5
4373
4492
119
2.72
6
4457
4567
110
2.45
7
4787
4927
140
2.94
8
4789
4913
124
2.59
9
4364
4459
85
2 .19
10
3227
3313
86
2.65
11
2301
2364
63
2.73
12
1649
1692
43
2.65
1
1757
1824
67
3.82
2
2049
2137
88
4 .33
3
1966
2088
122
6.18
4
1783
1910
127
7.12
5
1172
1300
128
10. 93
6
874
988
114
13.08
7
1095
1234
139
12.68
8
1719
1866
147
8.57
9
2410
2504
94
3.92
10
2321
2442
121
5 .22
11
2013
2095
82
4.03
12
1553
1612
59
3.82
-------
1684
CONCLUSIONS.
An easy-to-use algorithm has been developed to determine the monthly mean value
of the Incidence Angle Modifier. The accuracy of the method has revealed to be
satisfactory, especially in summer months.
Predictions of the average value of K can be estimated from several sources,
using solar radiation data files, geographic and climatic parameters, or simply
geometric characteristics of the collector. The results from every case are in
good agreement.
The best predictions have been obtained for the cases where more information was
available; the errors in the predictions, however, never surpassed the 2% for the
worst cases.
The estimated collectible energy including the effect of the IAM, obtained
through a correlation method, produces an error of 3% to 5% depending on the case
treated.
REFERENCES.
ASHRAE Standards. (1977). Methods of Testing to Determine the Thermal Performance
of Solar Collectors.
Bliss, R.W. (1959). The Derivation of Several 'Plate-Efficiency Factors' Useful
in the Design of Flat-Plate Solar Heat Collectors," Solar Energy, Vol.3,No4,p.55.
Bourges, B. (1987-89). Monthly mean values of the Incidence Angle Modifier.
EUFRAT Project .
Duffie, J.A. and Beckman, W. (1974). Solar Energy Thermal Processes, John Wiley
& Sons.
EUFRAT Project. (1987-89). Utilizability and Cumulative Frequency Curves.
Hottel. H.C. and Woertz, B.B. (1942). The performance of Flat-Plate Solar Heat
Collectors, ASME Transactions .
Lalas. P. and Petrakis, M. (1987-89). HELIOS Program.
Souka, A.F. and Safwat, H.H. (1966). Optimum Orientations for the Double Exposure
Flat-Plate Collector and its Reflectors. Solar Energy, Vol.10,p.170.
Whillier, A. (1967). Design Factors Influencing Collector Performance," Low
Temperature Engineering Application of Solar Energy, ASHRAE . N.Y.
-------
1685
ISFH - AN ADAPTIVE SIMULATION MODEL
Konrad R. Schreitmiiller, Michael Mack, Gunther Hesse
Institut fur Solarenergieforschung (ISFH), Sokelantstr. 5, D-3000 Hannover, F.R.G.
ABSTRACT
Component oriented models used in computer simulation of complex solar systems are often
cumbersome as generally any variation in the system design necessitates the repetition of the
full-length calculation. Contrary to that "adaptive" simulation models perform detailed cal-
culations on the component level only once in order to derive comprehensive characteristics.
The consequent use of this concept has led to the development of the ISFH program for
hydraulic solar thermal and photovoltaic systems. We compare features and performance of
ISFH with TRNSYS (12.2). While ISFH is not designed to compete with the universal ap-
plicability which TRNSYS offers, it shows substantial savings in calculation time and more
ease in handling extended parameter variations, which is typical in solar system design
problems. ISFH has already been validated with several experiments and shows fair com-
parison with the numerical outcome of TRNSYS.
KEY WORDS
Computer simulation, component models, collector model, synthetical climate, TRNSYS,
solar system output calculation
INTRODUCTION
Computer simulations models are primarily used in order
- to improve the understanding and analyzing of experimental data, to screen incompletely
understood systems, and to identify key parameters,
- to interpolate between results of different experiments, and
- to design both technically and economically improved, i. e. "optimized" systems.
The broad spectrum of applications calls for a hierarchy of simulation programmes on
different levels (e. g. finite element approaches for detailed component investigations,
condensed programmes for determination of long-term performance). Although these
different programmes are strongly interconnected (e. g. f-chart bases on TRNSYS), the
models operate generally independently. The consequences are
- particular effects are often insufficiently modelled and/or
- the determination of long-term performances of complex systems is often so lengthy, costly,
and wearisome, that most systems are designed according to rules of thumb or very simple
and inaccurate design methods, and seldomly reach their best performance.
These disadvantages are due to the fact that standard simulation programmes operate only
-------
1686
on one level, i. e. they do not learn from experience. Certain mathematical procedures are
repeated many thousand of times within one run and the programme exhausts itself with
trifles. Thus "adaptive" programmes are needed, which concentrate on a particular effect,
component, or system for only a limited time, investigating it in detail, but then condense the
gained knowledge to characteristics, which simplify the following investigations substantially.
ISFH (/l, 2, 3/) is such a program. We discuss its structure, some approaches and some
outcomes in the following.
THE ISFH METHOD
The input/output data of many carefully monitored solar energy systems, covering low and
high temperature applications, photovoltaics, and else, prove to be aligned along rather
simple graphs - the performance characteristics (Fig. IV If the basing equations, including
second order effects, are known, an accurate, fast, and easy way of modelling is opened
(/4/). This is performed by ISFH: it investigates given systems in a detailed manner, and
generates then performance characteristics, which are used for the following investigations.
The schematics of the method are summarized in Table 1. The general procedure is
- determination of the behavior of a given entity (effect, component, or system) in a detailed
way for a set of 30.. 150 different operational conditions,
- determination of the respective characteristics by statistical methods,
- use of those characteristics in the further investigations.
As these characteristics are determined for just the particular application, they are very
specific and thus, at least inherently, very accurate.
Table 1 ISFH Simulation Program (Schematics)
entity
example
procedure and results
Effect
incidence angle
pattern & polynomial/multi-
modifier
linear fit
Component
collector, HX, etc.
operational behavior & multili-
near regression fit (if characteris-
tics not available)
"ideal" system
DHW-system
"typical" operational behavior
("ideal climate"), regression fit of
system's characteristics
climate
Denver
synthetical climate from
monthly/daily/hourly values
"real" system
non-ideal collector orien-
collector output, solar to load,
tation, "real" climate,
solar fraction, surplus energy,
shading, tracking (if any),
efficiencies
weekly/monthly profile etc.
COMPONENT MODELS: TRNSYS and ISFH
The best developed, documented, and validated simulation programme on the level of
components is TRNSYS. Thus it seems reasonable to compare some of the ISFH approa-
ches with the respective TRNSYS (version 12.2) ones. In what follows, we give an abridged
synopsis of the most prominent differences in modelling features.
-------
1687
Solarhaus Freiburg
Input-/0utput Diagram
"GOOD" OPERATIONAL DATA
Daily Thermal Output
| kWh/m2d |
3
rf
0 •
—1 1 ! 1 1 1 1
0 1 2 3 4 5 6 7
Daily Incident Energy | kWh/m2d |
SSPS Almeria: Accurex Field
Input-/Output Diagrams
"Good" Operational Data
Daily Thermal Output
kWh
/ms*d |
"ft
0 -
a a T&1
~ AO'*-'
Uono" 1 1 1 1 1
0 2 4 6 8 10 12
Daily Incident Energy
Solar I Barstow, Cal.
Energy Input-/0utput Diagram
"Good" Operational Data
Daily Energy Output
|k
3
tVh/m2d|
" J&Z*
0
. -—i 1 1 1
Daily Incident Energy kWh/m2d|
i Thermal 0 Electrical
Energy Energy
Fig. l.Input-/Output-Diagrams of various Solar Energy Systems
kWh/ma*d
-------
1688
Component
TRNSYS
ISFH
Internal Heat Ex-
changer
exchange factor/-
efficieny
External Heat Ex-
changer
efficiency
Collector
efficiency
from test data
from design data
incidence angle
modifiers
capacity
Storage Tank
# of tanks
stratification,
# of layers
return layer
conductivity
allotment of losses
no type available; HX is simu-
lated by mixing collector back
flow with cold fluid
efficiency either constant or
according to selected HX-type:
cross-/counter- or parallel-flow
two options :
1) linear efficiency fit
r/ = f( A T, I, SMFR )
2) interpolation
17 = f( A T/I, I, vw)
A T refers to Tin, Tout, or Tavg
subprogram calculating
r) = f(F, UL, a, e, material &
design of transparent
cover)
ASHRAE approximation, or
Fresnel's formula, or perfor-
mance map (interpolation, sym-
metric/ asymmetric)
(not available)
optional
two modes to vary layer volume
and # of layers: < 15 with
type 4 (user-defined) or con-
trolled by calculation (type 38)
optional, floating return layer
possible
depending on fluid
even, modifications possible by
altering insulation thickness
along the walls
= 60 operational values depending on design,
geometry, mat. properties, temperatures of
resp. fluids, derivation of characteristic by mul-
tilinear regression
constant efficiency, counterflow type only
collector parameters are derived according to
the following model:
constant optical efficiency ra0
thermal loss factor dependent on absorber
and ambient temperature, and wind veloci-
ty
(U0 +U11*Tabs-U12*Tamb)*( l+vw)
internal conductivity depending on collec-
tor design, fluid temperature, and mass
flow rate
or by performance map (user's data)
A T referring to T)(avg
subprogram calculating
rj = f(a, e, material & design data of absorber,
:over, casing and construction),
parameters derived using multilinear regression
Fresnel's formula or performance map
(polynomial fit, symmetric or orthogonal func-
tions)
absorber design/ fluid properties
0, 1 or 2
<99 layers/constant volume
optional, floating return layer possible, internal
HX within lowest layer
depending on fluid and temperature
uneven, according to wall area of respective
layer (largest losses for top and bottom layer)
-------
1689
THE SYNTHETICAL CLIMATE
TRNSYS allows the input of hourly (or even better time resolved) data of the key meteoro-
logical values. Besides this, it offers two synthetical climate processors (a deterministic and
a stochastic type) in order to generate hourly values from monthly average data. ISFH -
originally developed for regions with only limited availability of meteo data - asks only for
averaged monthly values, which are processed by the following procedure:
- a intercorrelated series of "good", "average", and "bad" days with the respective radiation
sums and mean ambient temperatures is developed for each month,
- the daily diffuse to total radiation ratio as function of the respective clearness index is
determined according to the selected radiation processor (seven processors available)
- the atmospheric extinction model assumes a two-fold extinction process: one part of the
spectrum is completely absorbed and the remainder only partially according to the re-
spective air mass; with this approach a (constant) daily turbidity factor is determined
- the instantaneous beam and diffuse radiation intensities are then determined in order to
fit the daily radiation sum.
The deviations of the synthetical to real climate increase with increasing collector
inclination, as then the differences in the beam radiation densities are most effective. Thus
a comparison between "real" radiation sums (determined with hourly total and beam radia-
tion intensity data) onto a vertical, south-facing plane and the corresponding ISFH data
refers to the worst case. Extended comparisons of this type show that the deviations remain
almost completely within the ±3 % range.
As second option ISFH offers a daily synthetical climate, basing either on daily or hourly
values. The radiation intensities are then chosen to fit both the total radiation sum onto the
horizontal plane and the beam radiation sum on the two-axes tracked normal plane. Thus
good agreement with measured values is ensured.
ADVANTAGES OF THE ISFH METHOD
The ISFH method intends less to compete with the full complexity and wide range applicabi-
lity TRNSYS offers its users, than to allow for a quick and instructive inspection of solar
system design options. When working with I/O-characteristics, we separate the parameters
in two subsets: parameters belonging to the first subset strongly influence the shape of the
I/O-characteristics itself and must be fixed by the user before calculation starts. Variation of
these "primary" parameters like,e.g.,(for a solar system) the collector characteristics, HX
characteristics, piping lay-out and storage tank volume necessitates oyeidamg the full I/O-
characteristics calculation procedure. The effect of variations in system parameters belonging
to the second subset, however, can best be described either as a variation of the input to a
system with fixed I/O-characteristics or as simple changes in the I/O characteristics like
shifting the whole curve up or down. Examples of such "secondary" parameters are collector
area, orientation, inclination, tracking mode, shading obstructions and storage tank/piping
loss values. All changes in output resulting from variations of these parameters can be well
approximated by interpolation procedures and thus there is no need to repeat the full-length
calculation.
The advantages are twofold: first,. there is a substantial saving in calculation time.
While it may take up to 10 minutes to calculate the I/O-characteristics of a complex 100 m2
IPH-system, variations of secondary parameters appear with almost no delay on the screen.
Secondly, this fast inspection of the system's performance under different "real" conditions
helps to guide and focus the design engineering towards successful solutions - without
bothering and exhausting the user by the same full-length procedure for each variation. To
-------
1690
exploit this advantage further, ISFH -nates some effort in presenting user friendly, self-explai
ing input menu's and guiding side information. This allows fctffifrizatim with ISFH within
only one or two days.
SOME RESULTS
Simulation programmes are best validated with carefully monitored experiments to prove
their accuracy and area of validity. ISFH was fairly compared with measured data of several
solar DHW and IPH plants. However, the easiest way of validation is the comparison with
numerical outcome of another, well validated simulation model. TRNSYS is the most logical
choice for such a programme. First comparisons have been performed at the "Models
Simulation Workshop" at Colorado State University, Ft. Collins, Col., USA, in June '89
(/5/). Meanwhile, these investigations are substantially extended (applied TRNSYS version:
12.2). The results of the following "systems" will be presented
- collector with constant inlet temperature with increasing complexity of collector model
(temperature dependent loss values, internal conductivity, incidence angle modifiers, etc.)
- collector field with piping and heat exchanger, with constant or seasonally variable inlet
temperature (district heating system, unlimited demand)
- collector field with piping, heat exchanger, and storage tank(s) (DHW or IPH system,
limited demand).
CONCLUSIONS
The essential ISFH features as compared with those of a well developed "standard" pro-
gramme (here: TRNSYS 12.2) are the following
- ISFH is limited to hydraulic solar thermal and photovoltaic conversion systems, whereas
TRNSYS is a universal calculation program; on the other hand,
- in ISFH, particular effects or components are often modelled in a manner directly
suitable for solar system calculations,
- ISFH generates a "synthetical climate" from hourly/daily/monthly values; TRNSYS may
be based on - certainly more reliable - observed hourly values,
- ISFH shows substantial savings in calculation time, and by this can serve as a real design
and development tool in solar system engineering tasks.
REFERENCES
1. Schreitmiiller, K. R. Modelling Active Solar Energy Systems with Holistic Characteristics.
Advances in Solar Energy Technology, vol. 1, 828 (Pergamon Press, 1988)
2. Schreitmiiller, K. R. Mathematical Compression in Solar Modelling. Clean and Safe
Energy Forever, vol. 1, 547 (Pergamon Press, 1990)
3. Schreitmiiller, K. R. The Development of Mathematically Highly Condensed Computer
Simulation Models. Report of the IEA Solar Heating & Cooling Programme, TASK VI
(1991)
4. Duff, W. S., Boardman, E. Developing Performance Models of Solar Energy Systems
Using Daily Energy Input/Output Curves. Paper submitted to the American Society of
Mechanical Engineers Transactions to Solar Energy (1990)
5. Duff, W. S. Model Testing Workshop. Report of the IEA Solar Heating and Cooling
Program, TASK VI (1989)
-------
1691
A SIMPLIFIED TECHNIQUE FOR ESTIMATING THE ECONOMIC OPTIMUM
TEMPERATURE SWING OF THERMAL ENERGY STORAGE
IN SOLAR HEATING SYSTEMS
Charles A. Bankston
CBY Associates, Inc
5039 Cathedral Avenue, NW
Washington, DC 20016 U.S.A.
ABSTRACT
An analytical procedure for estimating the economic optimum storage temperature swing for a
solar heating system operated on a periodic basis is derived. Using the economic optimum
temperature difference for the storage unit and the principle of unconstrained collection, the
optimum size of the collector array, the thermal energy storage, and the end-use heat exchanger
can be calculated for any desired solar fraction. The procedure is most useful for heating systems
with seasonal storage, but can be applied to any heating system to optimize for a particular
period.
KEYWORDS
Active solar heating; system optimization; thermal storage optimization; seasonal storage; central
solar heating plants with seasonal storage (CSHPSS); collector cost effectiveness.
INTRODUCTION
The design of large-scale solar energy systems, such as central solar heating systems with
seasonal storage (CSHPSS), usually involves the optimization of the system components' size and
operating conditions. For projects involving thermal energy storage systems that are embedded
in the ground, the analysis and optimization can be costly and complex because the thermal
coupling between the storage volume and its surroundings is usually modeled by finite difference
or finite element methods. Optimization is justified by the high cost and the one-of-a-kind nature
of most projects.
Nevertheless, there are times when a simpler, less expensive, and easily implemented approximate
method is needed. For example, CSHPSS should be included as an option in expert systems that
are intended for a broad audience of engineers, consultants, and decision-makers. Therefore,
simpler methods of optimizing CSHPSS systems and estimating their costs relative to other
options are needed. A simplified method also is useful in the preliminary stages of the design
process in order to narrow the scope of the simulations needed to optimize the system and predict
its performance.
Simplified methods should, however, be developed on sound thermodynamic and economic
principles. Recently, Peter Lund (1988) has developed a simplified method of performing the
-------
1692
thermodynamic analysis and optimization (Ac and V) of systems using the principle of
unconstrained collection (Hull, 1991). If the operating conditions of the components of the
system are specified, the thermal optimum determined by Lund's method (SOLCHIPS) will also
be the economic optimum.
In large-scale systems, however, the designer may be free to chose the operating conditions for
all the components of the system, including the end-use equipment. Under these circumstances,
the relative costs of the components enter into the optimization process. For example, if the
collectors are expensive and the storage relatively inexpensive, the system should be designed
to operate at a low temperature (as long as the load can be met). This will result in a relatively
small collector array because the collector energy output is greater at low temperature, but a
relatively large storage volume because the capacity is restricted by the temperature swing.
Likewise, if the collectors are inexpensive and the storage expensive, the lowest-cost system will
employ a larger collector array and a smaller storage volume. Thus it may be possible to find
an optimum temperature swing for the thermal storage that results in the least-cost system.
SIMPLIFIED ANALYSIS
Total Energy Cost Equation
The total system energy cost can be expressed in terms of a few energy cost functions that are
either independent of the specific site, or can be easily localized. For a simple system in which
all the energy from the collector goes into storage, and the load is connected directly to the
storage, the unit cost of solar energy delivered to the load is given by
C . - . - (1)
Ct ~ + cs + c
where: c£ is the collector cost per unit energy on an annual basis,
cs is the storage cost per unit energy delivered annually,
cd is the distribution cost per unit energy supplied to the distribution system
annually, and
Tl is the annual storage efficiency.
Optimum Storage Temperature Swing
The collector and distribution energy cost functions are insensitive to size of the system so long
as the operation is unconstrained. The storage energy cost, however, is sensitive to the capacity,
so we seek an optimum storage capacity, Es. Differentiating with respect to Es
dct = 1 dcc _ Ccj>n_ + + (2)
dEs *1 dEs "n d£s dEs dEs
The storage capacity is a function of the volume, V, and the swing, 5T = Tmax - Tmin, of the
average storage temperature.
The storage efficiency is obviously the most difficult to evaluate and the main reason numerical
simulations are time consuming and expensive. It is, however, independent of latitude and
weather (or nearly so) so that empirical correlations could be developed that would apply in all
-------
1693
countries (the thermal properties of the media and surroundings must be included). However,
here we assume very simple relationships for the heat loss from storage, the cost effectiveness
of the collector and the cost effectiveness of the delivery system. The equations are
E. = pc.VbT (3)
¦n
(Es + Qi)
(4)
= 1/2/3
V^U, AT
5 T
(5)
~ Tmin TA
(6)
cc = A + B
T ¦ +
min
5 T
(7)
(FC + C'J' V)
p cpvsr
(8)
cd = G +
H
5f
(9)
Carrying out the necessary differentiation using the chain rule yields
dc t
BE.
B
2pcpV
1 +
Ul&Ta ^ UL
pcpVlfibT 2p CpV
2FC+c's" y
1/3
4 ULV,
2/3
(pc VSt)
ATA
5 T
H
(pcpVbT)2
pc VST-1
(10)
A cost minima will occur if 3ct/3Es = 0 and 32cT/3Es2 > 0.
It is most interesting to find that 8T heat minimizes the cost. The volume and the collector area
can then be computed from the load although iteration may be required. The general solution
is
Sr - b ± (b - 4ac)
01 opt
1/2
2 a
a = (1 + UL/2pcpViri)
b = (3ATa - 1 IB) / 3pc V
1/3
(11)
BpcaV)
luLVm(ATA) + 2FC + C'J' V + HpcpV
provided B is not equal to 0.
If B = 0 (central receivers, parabolic troughs and evacuated collector outputs are nearly
-------
1694
independent of temperature over the range of interest); there is no minima. The cost decreases
monotonically with Es.
Note that the H term is derived from the term in the distribution cost equation representing the
cost of end use heat exchangers, i.e., H/5T, and that a large value for H will increase the
optimum 8T for a particular collector storage type. If the storage volume is well insulated, or
very large, so that heat losses are small, and the variation of distribution cost is neglected, the
optimum temperature swing simplifies to
Kt
,, III 2FC
S T/
(12)
The question of the optimal solar fraction for the system depends on the economics of alternative
energy sources. As shown by Lund, the collector area is a linear function of the solar fraction,
so the optimum parameters for any desired solar fraction can be obtained by scaling the collector
area.
Comparison with Simulation Methods
The validity of the analysis was confirmed by comparing the results obtained for the optimum
temperature difference from the analysis with results derived from numerous MINSUN
simulations. Figure 1 shows the comparison. The curves are the numerical results of MINSUN
calculations with fixed temperature swings
and the collector area and storage volume
from SOLCHIPS as input. The heavy dots
are the points along the expansion path, i.e.
the locus of optimal systems, determined by
Zinko (1989) from hundreds of MINSUN
calculations using a systematic variation of
parameters. For the collector to storage cost
ratio and location shown in Figure 1, the
predicted optimum temperature swing is about
80°C which is slightly higher than the result
(about 70°C) inferred from Zinko's findings.
This comparison also shows that reasonable
accuracy can be achieved at a small fraction of the computational effort required to determine
the optimum from systematic parameter variation alone if the results of the simplified analysis
are used as the starting point for the simulation study. The technique should be valuable for
preliminary estimates and as a starting point for detailed simulations using MINSUN or
TRNSYS.
APPLICATIONS
CSHPSS Applications
As an example of an application of the method, consider the range of options for a large
CSHPSS system in three locations—Albuquerque, Boston or Stockholm—using one of three
collector options—high-efficiency flat plate, evacuated tube, or E-W parabolic troughs—and one
8ol«r Cost, t/MVh
!5olor Fraction. X
Fig. 1. Comparison of simplified method with numerical
simulation for 1000 residence system in Madison,
Pits storage (50 $/m3) and flat plate collectors
(183 $/m2).
-------
1695
of two storage technologies—a rock cavern or drilled rock. The collectors' cost functions are
determined by regression for the energy output calculated by annual integration of hourly results
for collectors operating at a constant fluid temperature of 10°, 30°, 50°, 70°, and 90° C. (It has
been shown that output calculated for constant temperature operation at a temperature of (Tnmx
+ Tmjll)l2 is very close to that for a system with variable temperatures so long as the storage does
not constrain the collector output (Bankston, 1983). The collector efficiencies are assumed to
be 0.75 - 2.9 AT/I for the high efficiency flat plate, 0.51 - l.3l(AT/I) for the evacuated tube, and
0.0808 - 0.086 (AT/I) - 0.809 (AT/lj2 for the parabolic trough, and the costs are assumed to be
250 $/m2, 300 $/m2, and 350 $/m2 respectively. Assume that the cost of the drilled rock and
rock caverns are 6 $/m3 and 40 $/m3 respectively. (These cost and performance data are typical
of those used in previous system studies of CSHPSS (Bankston, 1986) and do not represent
current or future technology. They are cited merely as examples to show the influence of
location and relative cost.)
Figures 2 and 3 show the collector energy cost functions, and Table 1 shows the estimated
optimum temperature swing for each of the options. Note that optimal temperature swings for
the low cost storage (drilled roek) is much lower than the cavern, and that the optimum
temperature swing is quite sensitive to location, as shown by the difference between Albuquerque
and Stockholm. The evacuated tube collector, which has a smaller B than the flat plate, requires
a larger temperature swing, and the parabolic trough would be optimum at temperatures above
the practical limits.
i, FP Slo, FP
* Alb, FP
-G-- Alb, ET "H- Bos, ET -A- Sto, ET
1--
nT 0.8-
0.6-
O 0.4-
Tempersiure, deg. C
Fig. 2. Collector cost effectiveness functions for flat
plate and evacuated tubes.
Alb, FP Bos, FP Sto, FP
s
Alb, PT -M" Bos, FT Sto, PT
/
/ s
|
—T5
10 20 30 40 50 60 70 80 90 100
Temperedure.deg.C
Fig. 3. Collector cost effectiveness functions for flat
plate and E-W parabolic troughs.
The total energy cost of the system can be calculated from Equation 1 if the storage efficiency
is assumed. The collector area can be found from the temperature swing with the aid of the
collector production data from which the collector energy costs were derived, and the storage
volume can be calculated from the temperature swing and the efficiency. With a little experience
or some empirical efficiency curves, it should be possible to estimate the collector area and
storage volume within a few percent.
DISCUSSION
Two important points are brought out by the example in this paper: 1) optimization of both the
size and operating temperatures of components in solar heating systems is important, and 2)
optimization and characterization can be greatly facilitated by expressing component performance
-------
1696
in terms of their cost per unit of energy. The collector cost functions plotted in Figure 2, for
example, are easy to produce, and provide an unequivocal basis for comparison and selection.
TABLE 1 Optimum Thermal Storage Temperature Swings
Collector Energy Cost Coefficients
Albuquerque
Boston
Stockholm
Flat Plate Collector, B
0.00259
0.00916
0.0145
Evacuated Collector, B
0.00109
0.00352
0.00524
Parabolic Trough, B
0.000107
0.000372
0.00767
Optimum Temperature Swing
Flat Plate, $250/m2
Drilled Rock, $6/m3
Cavern, $40/m3
67
115
36
61
32
55
Evacuated Tube, 300 $/m2
Drilled Rock, 6$/m3
Cavern, 40$/m3
104
178
58
99
47
81
Parabolic Trough, 350$/m2
Drilled Rock, 6$/m3
Cavern, 40$/m3
332
568
178
305
124
212
REFERENCES
Bankston, C.A. (1983) Central Solar Heating Plants with Seasonal Storage: Basic Perfor-
mance. Cost, and Operation of Solar Collectors for Heating Plants with Seasonal Storage.
Report of the International Energy Agency, Task VII, published by Argonne National
Laboratory, Argonne, IL.
Bankston, C.A. (1986) Central Solar Heating Plants with Seasonal Storage: Evaluation of
Concepts. Report of the International Energy Agency, Task VII, United States Department
of Energy, Washington, DC.
Hull, J.R. (1991) "Analytical and Numerical Modeling of Thermal Energy Storage Sys-
tems." In Solar Collectors. Energy Storage and Materials. Francis de Winter, Editor, Volume
5 of Solar Heat Technologies: Fundamentals and Applications, Charles A. Bankston, editor-
in-chief. MIT Press, Cambridge, MA.
Lund, P.D. (1988) A General Design Methodology for Seasonal Storage Solar Systems.
Helsinki University of Technology, Espoo, Finland.
Zinko, H. and H. Walletun (1989) Central Solar Heating Plants with Seasonal Storage:
Evaluation of Water Storage Systems. Studsvik Energy, Studsvik, Sweden.
-------
1697
COMPUTER SIMULATION OF A PHYSICO-MATHEMAIIGAL
MODEL OF SOLAR ENERGY THERMAL STORAGE
V.L.Fara and A.Buour
Dept. ©f Physics, Polytechnieal Institute
of Bucharest, Splaiul Independence^ 313#
Bucharest, Romania
ABSTRACT
The knowledge ©f the time-distance evolution ©f a solar
energy thermal storage system using reek beds Grossed by aft
erganie fluid allows the acquisition of important informatio&
for the optimum dimensioning of thermal solar energy tanks.
Because of the difficulties arising during the analytiaal solu-
tion of the storage equations, we applied the numeriaal solution
to the network method.
KEYWORDS
Storage system; network method; reduoed temperature;
matrix form; temperature distribution.
INTRODUCTION
The study ©f the solar energy thermal storage systems
using rock beds crossed by ga organic fluid represents a very
important research problem for solar installations.
The physical system considered is composed ©f a vertical
cylinder containing the storage medium which is made up of
particles of a very small diameter as compared to the diameter
©f the cylinder. The flow ©f the organic fluid is vertical
with the warai part of the stook at the upper part ©f the
cylinder and the cool part at the base in order to avoid the
natural oonveotion phenomena which lead to the destruction of
the thermal stratification. The thermal transfer equations
for physical system considered (storage equations) will be
(Giaqyiel and others, 1979)
J st
1
v.
where
w(t)
9«<*»*.
3 x
+ b-^(u-v) = 0
+ b2(u-v) = 0
% - *1 .
i^nr; i
u =
V =
TS ~ T1
(1)
are reduced tempera-
tures; b^ and b2 are parameters oharaeterizi&g the thermal
properties of the fluid and the solid; T-^Cxjt) is temperature
of the organic fluid; TQ(x,t) is temperature ®f the solid
-------
1698
mediumj T^ and T^ are temperatures at the lower and upper ends
of the stock, respectively.
The partial-Klerivative equations system obtained is a
hyperbolic system. To solve it, we use an improved version of
the network method.
METHOD FOR THE HUMERIO-AL SOLUTION OF THE
STORAGE EQUATION SYSTEM
Because of the difficulties arising during the analytical
solution of system (5), we applied the numerical solution with
the network method (Ixam 1979; Absi and ..others, 198o).
The principle ©f the method consists in finding solutions
along two families ©f characteristics
dx = w dt fofr the fluid
dx = 0 for the solid
System (1) turns into systems (2) and (2)
fdx = w dt (2) j dx = 0 0^
|du - b-j_(v-u)dt Idv = bgCu-v^t
We suppose the particular solutions already known
f«Ct0,x), v(tQ,x)'
lu(t,x ), v(t,x )
Starting from the vafues of u0aad v at points 1 and 2 we
determine the values of u and v at the point of intersection
of the characteristic curves crossing points 1 and 2j we take
du and dv to vary linearly along these curves.
We also enter two sets ©f data; a set of values for Tq
and T^ measured at the same moment and a set of values of ®
these temperatures at a fixed location in space but a different
time moments.
In order to get easily processable results, the sets of
data have been writtea in matrix form. This procedure was
applied in order to specify the time-distance distribution ®f
temperature in a condensed form. Thus, the variation of the
temperature ®f the liquid is given in matrix form by T^CljJ),,
where the rows I indicate tha behaviour at constant t = const,
as a function of distance x'and columns J the behaviour at
constant distance x = const, as a function of time. This
approach was extended for the functions u(x,t) and v(x,t),
which give Matrices U(I,J) and V(I,J).
The reduced temperatures a and v can thus be computed at
the points where Tg and T^ are knowns
Tx(i,j) - T, ®q(i#;I) - T-,
u(i,o)=—f2 , ^ » ?(i,j)= iji , ij 1 (4)
A numerical solution by means of the network method
requires sampling of input data. In order to get an acceptable
accuracy for theysolution, this sampling should be fine enough
(16 x 16 points). The coefficients ^(i s 1,6) have also been
writteil in matrix form A (I, J) aid have been used to compute the
-------
1699
increments da and dv required by'the network method.
The procedure is iterative. Starting from the givea ini-
tial solution (u0»v0), an attempt is made to verify the system;
if this does not fall -within the required limits of aoouraoy,
then du = duCi^ and dv = dvC^) are computed and henoe:
= Cuo»v0) + (du,dv)
is the new solution with whieh another attempt is made to
verify the system.
The quantities du and dv are also written in matrix form;
DU(I»J)» DV(I,J).
The iteration is performed until the required acouracy is
reached for a set whioh is takes. t§ be the solution of
the system of partial derivative equations.
In order t© get the distribution of the reduced tempera-
tures u and v for a network of points we compute da and dv
as a function of the following coefficients:
h = e"bi ^ = e~h2T5=rw
1-A-, 1 -A 2
a5 = -1 + b L ; = - i + -—,
l (.n-i ;w 2 (s-i; w
A5a(l-A1)[v(3,i)-uQti)] -Aj [vCd+ljD-uCj,!)]
A5=(l-12) fuO+l,i)-v(3+l,i)] -i4[uCjfi)-vCj+l,i)] (5)
A (-""A -yA rl A /-—-A j. A r
ft* = P ? 6 ¦ ; dv = 6 4 ft, (6)
1—A^Aq
(7)
The following recurrence relations are satisfied;
u(j+l, i+1) = u(3,i) + du
v(j+lf i+1) as v(j,i) + dv
If w = 0 we ohose a value for At * (a-l)w' with the
coefficients givea by
-b-. At -bp^t 1-At
A1 ~ ® » a2 ® ® •, Aj = -I + £_¦
X
A/)- " " 1 + ~^TT 5 A5 = ^6 •
(8)
COMPUTER PROGRAM
The program computes the temperature distributions with
double preaisien.
As input data we have chosen the following sets of
parameters:
- the initial temperature distributions
-------
1700
(matrices »d) which correspond to the
storage aediua anm to the fluid, respectively)j
- temperatures Tj_ and 0}^ at the base and upper end ©f the
tank, respectively;
- geometrical parameters of the system j
physical paraaeters of the storage aediua and of the
fluid.
We note the following stages:
- computation of the fluid velocity w, of parameters b-,,
¦b2jof the initial U and 7 Matrices using the initial Tg and
Tt matrices j
- ooaputstion of the set of parameters (i ,=1,6);
- computation of the infinitesimal quantities dU(I,J)
and dV(I ,
»» ¦ (4 ™
-------
1701
TUJSL°C J
2 SO
240
250
220
110
200
m
1&0 -«•
170—'
160
150
-------
1702
I®
Fig.2. !Dwo~dimensional depeadenae ef temperatures (EL on
distance and time.
0 ),, thea
T:
We notice that, if °c2i G (tg p°21
5fjj2j_ w® have a maxiaua thermal transfer.
COlICIiUSIOHS
Ehe main new points ©f this paper are;
- the use of a simple iteration method
- the use of matrix form
- the outline of a correspondence between teaeperatures
and eertain angles which allow a physieal interpretation of the
results obtained.
The method used by the authors to obtain the temperature
distribution has some important advantages. We aotiae the pre-
©ision of the soaputatioas all the relatively high degree of
generality whiah allows the applicability ©f the method t© a
variety of phyaioal systems of interest'for the thermal storage
installations of the solar power plants. Ehe eomparison of these
data with the experimental results will make possible the opti-
mum dimensioning ©f theraal solar energy tanks.
REFERENCES
Griequel. 'R., D. Harang and D. Sohmell (1979). Rev.Gen.Iheria..
Frt, Ko.212-213. 527-555.
Ixaru,T. roserieal Methods for Differential Equations
and Applications. Roumanian Aoademy Publ.House.kuchares'fe
Absi, E.. R. Glowinskx, P. Lasoaux and H. Veysseyre C198o)*
Methodes luia^riques dans lea Sciences de l'insenieur.
Dunod, Paris .
-------
1703
VALIDATION OF A PARAMETER IDENTIFICATION METHOD WITHIN THE
IEA DYNAMIC SYSTEMS TESTING GROUP
A.C. de Geus and H. Visser
TNO-Building and Construction Research
P.O. Box 29
2600 AA Delft
The Netherlands
ABSTRACT
Within the framework of the IEA Solar Heating and Cooling programme a research group
is working on a dynamic system testing method for solar domestic hot water systems. This
group, the Dynamic Systems Testing Group (DSTG), started its activities as follow-up of
IEA-task HI in 1989. During 1991 the activities will result in recommendations for dyna-
mic testing. The method is developed by the German participant and uses mathematical
fitting algorithms in order to determine the main system characteristics. Research is carried
out in order to verify method and model for a large variety of systems being tested in-
doors, outdoors and in-situ. Moreover, a test strategic is developed which should enable
the accurate determination of the system parameters in short time. So far results are prom-
ising both for the accuracy and the test stategy. For the evaluated systems the yearly per-
formance is accurate within ±7%, while the test sequence onlytakes3-4 days (dependent
on prevailing weather conditions). At the end of 1991 recommendations for testing will be
published as main result of the DSTG.
KEYWORDS
Parameter identification; dynamic fitting; yearly performance; solar domestic hot water
systems; test method; test procedure; evaluation method; indoor, outdoor and in-situ tes-
ting.
INTRODUCTION
During the completion of IEA task IE of the Solar Heating and Cooling programme (Per-
formance Testing of Solar Collectors) a promising dynamic test and evaluation procedure
was introduced by the German participants. In order to evaluate and validate this method a
working group of the IEA solar heating and cooling programme was installed, the
Dynamic Systems Testing Group (DSTG). In the DSTG the following countries participa-
te: Canada, Denmark, Germany, the Netherlands, Sweden, Switzerland and the USA, while
Spain and the CEC (Ispra) have the observer status. The working group is organized and
led by the Netherlands.
-------
1704
TEST METHOD
The ideal test method for determining the thermal performance of solar domestic hot water
systems should have the following characteristics:
* it will characterize the system such that short term test data can be used to predict
long term performance for an arbitrary location and for an arbitrary load.
* it can be used for both stationary indoor test conditions and for non-stationary
outdoor conditions;
* it can be used for in-situ testing of both large multi-family systems and compact
single-family SDHW systems;
* it can be used to test a variety of system types, e.g., pumped recirculation, ther-
mosyphon, integrated collector storage, boiling/condensing, etc.;
* it can be used as diagnostic test for optimizing system performance;
* there is no principal restriction to the test sequences. For instance, these can be
chosen in order to give specific information in minimum time, but maybe also in
order not to disturb an operating system in the normal operation mode.
Present day test methods for SDHW systems determine the energy delivered by the system
as a function of the input test variables using mean values over a sequence of steady state
conditions. The new dynamic test procedure recommended by the University of Munich
correlates system output and input variables as a function of time by considering the rate
of change of system output relative to the rate of change of the input variables. Perfor-
mance parameters are identified using a mathematical filter and fitting techniques.
DYNAMIC SYSTEM TEST METHOD
The dynamic testing method is a combination of a parameter identification procedure and
a simplified solar domestic hot water system model. The test conditions are the input
variables of the model. By comparing the measured and calculated output data the parame-
ters are adjusted until a minimum discrepancy is obtained.
The technique uses transient data, a mathematical filter scheme and least squares fit me-
thod to obtain the characteristic parameters: the system parameters are identified. Subse-
quently, these parameters are used to predict the yearly performance under typical referen-
ce conditions. In figure 1 the method is presented schematicly.
The major model parameters are described as follows:
The parameters Ac" and Uc* describe the collector characteristics. These are simular to the
coefficients in the Hottel-Whillier-Bliss equation. Us and Cs describe the overall perfor-
mance of the heat store. Additional parameters characterize the store's stratification, draw-
off mixing and, if included in the system, the auxiliary heater and load side heat exchan-
ger.
Input data
System Gain
Model
Fig', 1. Dynamic system fitting (F) can be interpreted as
inversion of dynamic system simulation (S)
-------
1705
In a paper to be presented at the ISES 1991 conference by the German participant (Spirkl,
1990) more details and fundamentals of the identification algorithm will be discussed.
GOAL OF THE WORK
The goal of the DSTG is the evaluation, improvement and validation of the dynamic test
method.
This has been carried out using simulated and experimental data for a large number of
different domestic hot water systems. Indoor and outdoor tests have been evaluated as well
as in-situ measurements.
An overview of the systems investigated and their test environment is given in table 1.
TABLE 1. Overview of the systems.
System
Indoor Outdoor In situ
Thermosyphon
Pumped, drain back
low flow
boi1ing/condens ing
ICS
Integrated auxiliary
+ + +
+ + +
+
+
+ +
— + +
Moreover a test sequence for indoor and outdoor testing has been developed. From this
sequence the system should be characterized within a short period of time. The criterion
for the test sequence is the prediction of the yearly performance with a sufficient accuracy.
A more advanced characterization is the identification of the system parameters in such a
way that the cross correlation between the parameters is either small or insignificant.
INVESTIGATIONS ON SIMULATED DATA
In order to find good test conditions for SDHW systems computer simulations are used to
generate simulated data. These data are used for dynamic fitting and in this way test stra-
tegies are analyzed. During the study four different SDHW were investigated, varying the
load profile, load volume, measuring time step and the weather conditions.
The objective of this study was to find a suitable method for outdoor testing. From the
simulated test data the system parameters were identified by fitting and compared with
their values used as input parameters for generating the simulated data. Finally, the yearly
performance predicted by the simulation program and by the dynamic fitting package were
compared.
Parameter identification for dynamic testing appears to be a powerful tool in characteriz-
ation of SDHW systems.
However, the model and the test sequence used have to be chosen carefully to get accurate
results. Especially, correlated input variables in the test sequence have to be avoided.
Otherwise, correlated parameters are obtained which are less accurate.
-------
1706
A short 3-day test sequence with non-correlated input variables seems to be sufficient for
an accurate characterization. The weather has to be nice during the first two days. After an
initial conditioning of the system by a high draw-off the store temperature is kept low
during the first test day by the draw-off of 1.5 times the store volume equally distributed
over the day. During the second day there is no draw-off in order to get a high store and
collector operating temperature. After the night time used to identify the store heat loss
coefficient, the store is depleted at the beginning of the third day. For outdoor testing the
test period will be longer until the weather conditions required are encountered.
Among others a SDHW with a primary heat exchanger, 2.4 m2 spectral selective collector
area and a store volume of 120 litres was used for this study.
In table 2 results of the parameter identification for this system are given.
TABLE 2 Comparison predicted and 'real' system parameters
and yearly performance.
Test period
A* U* U C S. F. Solar Gain
c c s s
(m2) (W/m2K) (W/K) MJ (%) (MJ)
expected
fitted
1.8 3.85 1.4 .5 56 4.6
1.75 3.8 1.4 0.5 50 4.1
NOMENCLATURE
*
A =
c
effective collector area
[m2]
*
Uc =
equivalent collector heat loss factor
[W/m2K]
U =
s
heatloss coefficient store
[W/K]
C =
s
thermal capacity of the store
[MJ]
:.F.=
Solar fraction
For this type of system the results of fitting and detailed simulation model are in very
good agreement. Also for most other systems the agreement is well. The test sequence is
not yet optimal for systems without a collector-side heat exchanger.
INVESTIGATIONS ON EXPERIMENTAL DATA
Experimental data from real outdoor tests have been processed using the dynamic fitting
procedure. In the following dynamic fitting results of SDHW systems without auxiliary
heater, tested at participating laboratories are presented:
system 1 with forced circulation and a collector-side mantle heat exchanger,
system 2 with forced circulation and a collector-side coil heat exchanger,
system 3 with thermosyphon flow and a collector-side coil heat exchanger.
For system 1 the test conditions which are presented in table 3 were investigated.
-------
1707
TABLE 3 Test sequences.
Testday
Sequence I
Sequence II
Sequence III
No
Hotwater
(litres)
Wheater
Hotwater
(litres)
Wheater
Hotwater
(litres)
Wheater
360
360
360
' 1
3*60
nice
3*60
reasonable
3*60
nice
2
48
nice
48
bad
48
bad
3
360
nice
360
nice
48
reasonable
4
3*60
bad
3*60
nice
360+48
nice
5
-
-
-
-
3*60
nice
Four different fittings were performed respectively for sequences I, n, III and the com-
bination of 1+ II + IE. The systems parameters are well known from earlier tests.
The parameter values as well as the long term performance can be compared. These yearly
predictions are carried out for the location of De Bilt", the Netherlands and assumes a daily
hot water load of 110 litres heated from 15 °C to 65 °C. The needed energy for heating
the water is thus 8.4 GJ.
In table 4 results are given, both for the parameters and the yearly performances.
TABLE 4 Comparison predicted and 'known' systems parameters
and performance.
Sequence
*
A
c
*
Uc
U
s
C
s
S. F.
Solar Gain
[m2]
[W/m2K]
[W/K]
[MJ]
%
[MJ]
expected
2.0
4.0
2.5
0.5
54
4.5
I
2.15
5.6
3.1
0.6
49
4.1
II
1.8
4.1
3 .2
0.6
47 .5
4.0
III
1.7
3.1
2.0
0.5
50.2
4.2
I+II+III
1.95
4.3
2.9
0.6
49.5
4.2
Though the testing method and conditions were not optimal reasonable parameter identi-
fication results were obtained. The deviation between the annual performance prediction
by the dynamic fitting program and other prediction methods was lower than 15 %. The
other two systems 2 and 3 were tested outdoors for a longer period during 1989 and 1990.
During these extensive measurements the draw-off profile and flow were varied. It became
clear after the first fitting results that application of a certain draw-off profile gave much
better results than a continue draw-off flow. Moreover, the best results, were obtained if
during a sequence various weather conditions were encountered. If the duration of the se-
quence was too short results were less accurate and the parameters were too much correla-
ted. In table 5 results are given for systems 2 and 3.
-------
1708
TABLE 5. FITTING RESULTS FOR SYSTEM 2 AND 3.
System
value
*
A
c
*
u
c
U
s
c
s
no
rm2l
rw/™2Ki
fW/Kl
rM.T]
2
expected
3.8
7.0
3.1
1.3
2
fitted
3 .7±Q .2
7 .7±0 .8
2 .65± . 4
1.15±.1
3
expected
3.0
8.05
3.6
1.45
3
fitted
2 .7±0 .2
6.8±0.6
to
h-1
1+
CO
1.4+0.1
It is obvious that the identified parameters are in good agreement with the expected ones.
Within the DSTG measurements of other systems mentioned in table 1 are being proces-
sed. In this way the dynamic fitting procedure and its accuracy will be thoroughly ana-
lyzed. The recommended test procedure and limitations will become clear during 1991.
FUTURE RESEARCH
At present a number of SDHW systems including auxiliary heater are being tested.
These measurements take place outdoor and indoors under fixed test conditions, but also
for in-situ. The tests concern a series of short test sequences as described above. The
parameters of these systems will be identified from the datasets and annual performance
predictions will be compared with estimations found by means of different methods. In
this way both the dynamic testing method and the parameter identification program is
evaluated.
CONCLUSIONS
Up to now both simulated and experimental data gave quite good results. The identified
parameters are used for yearly performance calculations with a the same simplified and
'widely' applicable model as used for the identification. The uncertainties of the parama-
ters are used to calculate the uncertainty in the yearly performance. With the same model
it is also possible to calculate the yearly performance for circumstances (location, load,
etc.) different from those during the measurements. The DSTG activities will result in a
IEA report at the end of 1991. This publication will include the software (identification
procedure and simplified domestic hot water model).
ACKNOWLEDGEMENTS.
All mentioned activities are carried out within the framework of the IEA Solar Heating
and Cooling programme. The Dutch activities are made possible by financial means from
the Dutch Ministry of Economic Affairs in the framework of the National Solar Energy
Research Programme.
REFERENCES.
Sprikl, W. Dynamic SDHW Testing. J. of Solar Energy Eng.. Transactions of the ASME,
112:98-101,1990.
-------
2.16 Posters: Active Solar I
-------
-------
1711
OPTIMAL COLLECTOR SIDE MASS FLOW RATE
IN DOUBLE-LOOP SOLAR WATER HEATING SYSTEMS
Mazhar Unsal*
"Department of Mechanical Engineering, University of Gaziantep
27310 Gaziantep, Turkey
ABSTRACT
Steady operation of a double-loop active solar heating system consisting of solar collectors, a
heat exchanger and a thermal energy storage tank is considered. Algebraic expressions for the
collector heat removal factor and for the effectiveness of a counter flow heat exchanger are
utilized to obtain an algebraic formula for the collector heat exchanger efficiency factor as a
function of the number of transfer units of fluids in the heat exchanger. For any fixed value of
the load side number of transfer units, the collector heat exchanger efficiency factor is found to
have a unique maximum corresponding to an optimal value of the collector side number of
transfer units. A simple algebraic formula is reported for the optimal collector side number of
transfer units. Graphical results are presented depicting the increase in system thermal
efficiency when the collector side number of transfer units is set at the optimal value.
A fluid with a freezing point below zero is required in the collector loop of active solar heating
systems to affect freeze protection; and most commercial active solar heating systems are built
with a heat exchanger between collectors and the thermal energy storage tank. Energetic
efficiency of double-loop solar water heating systems depends on the mass flow rates of the
fluids in the heat exchanger between the collectors and the thermal energy storage tank. A
double-loop solar water heating system with a counter flow heat exchanger between solar
collectors and the storage tank is analyzed in this study. Analysis presented is based on the
system depicted in Fig.1. A differential temperature controller detecting temperature between
solar collectors and the thermal energy storage tank will either actuate or stop both circulation
pumps according to a preset control strategy. The solar heat collection is a time dependent
phenomenoaSteady operation of the solar heating system has been assumed in most previous
investigations and the same simplification is retained in the present study. Results from the
analysis yield an optimum number of transfer units (or fluid mass flow rate) in the collector
loop for a given load side number of transfer units (or fluid mass flow rate) in the heat
exchanger.
Following deWinter(1975), one can express the useful energy collection rate of the solar
heating system depicted in Fig. 1 by the following formula.
KEYWORDS
Optimal flow rate, double-loop, active solar heating, solar collector, heat exchanger.
INTRODUCTION
ANALYSIS
Qu= FrHtaJ (xa)
(1)
receding page blank
-------
1712
TO LOAD
imsmm
IISTORAGEi
II TANK i
lillfilll
FROM LOAD
Fig.1 Diagram of a double-loop active solar heating system
Considering the case when mcCc is the smaller capacity rate, F'R is given by (Duffie and
Beckman, 1980)
FR
Fr= - (2)
mcCc l £
Effectiveness of a counter flow heat exchanger may be expressed as a function of the load side and
collector side number of transfer units by the following formula:
1 -exp(nL-nc)
e= L c (3)
nL
I exp(nL-nc)
n0
The expression for the collector heat removal factor and the expression for counter flow heat
exchanger effectiveness, when substituted into equation (2), yields:
Fr = r- (4)
| rnc rnc - rnL
II - exp(-F'rnc) exp(nc - nL) -1J
Equation(4)has been obtained under the assumption that the collector side capacity rate, mcCc,
is smaller than the load side capacity rate, mLC|_. It can be shown, by performing an analysis
similar to that given in this section, that equation (4) is also valid when mcCc 5 mi_Ci_- For any
fixed value of the load side number of transfer units,nL , the collector heat exchanger factor
given by equation (4) has a unique maximum value at a finite optimal value of nc. Inspection of
the first partial derivative of F'R, given in equation (4), with respect to nc yields the following
optimal value of nc corresponding to the maximum of F'R.
n°= rF+T <5)
An expression for the maximum value of F'R , F'Rmax, is obtained by substitution of nc given by
equation (5) into equation (4):
-------
1713
1.3
rF'=0.1
rF=0.5
rF=1.0
1.2
1.1
1.0
1 0
1
100
Load side number of transfer units
FV
Fig. 1. Ratio —^,m^x versus the load side number of transfer units, ni_.
Ro
F'rnL
"Rmax
1 -exp
1 + F'r
F'rnL
(6)
DISCUSSION
Heat transfer rate for any counter flow heat exchanger will be maximum when the ratio of
capacity rates is zero. Increasing the magnitude of the larger capacity rate increases the
effective temperature difference in the heat exchanger resulting in a larger heat transfer rate.
This observation is not equally valid for a solar collector-heat exchanger system where an
increase in the collector side fluid capacity rate tends to decrease the collector thermal
efficiency while affecting an increase in the temperature difference within the heat exchanger.
Inspection of equation (4) shows that F'R decreases from its maximum value to zero in the limit
as nc -» oo. F'p, decreases from its maximum value to a finite value F'Ro, in the limit as nc -> 0.
F'Ro is given by the following formula.
FRo I
T=i+. Pr"L <7'
1 - exp(-nL)
Collector side number of transfer units is sometimes set equal to the load side number of
transfer units (deWinter,1975). For this case, letting F'Req denote the value of F'R when nc =
n|_, the following is obtained.
Ffteg = 1 - exp(- PrnL}
F Fr{1 +nL-exp(-FrnL)}
-------
1714
rF=0.1
rF=0.5
rF'=1.0
1.10-
O
<
cc
1.05-
1.00 +
.1
-O-
1
10
1 00
Load side number of transfer units
F nrpny
Fig. 2 Ratio —= versus the load side number of transfer units, n[_.
hReq
The ratio of F'Rmax to F'Ro is presented graphically in Fig. 1. This ratio is asymptotic to unity
for large and small values of the load side number of transfer units. The ratio of F'Rmax to FReq
is presented graphically in Fig. 2. It is observed from Figs. 1 and 2 that the proper choice of the
collector side number of transfer units will yield the best thermal system efficiency. Small
values of rP correspond to large values of Qu- rF' will consequently be small in most practical
applications. As a result, F'Rmax will generally be close to F'Req. Choice of nc given by equation
(5) may upgrade thermal efficiency by a few percent for most
NOMENCLATURE
Ac
collector surface area
Ae
heat exchanger area
f'r
collector heat exchanger factor (defined in equation (2))
F
collector efficiency factor
fr
collector heat removal factor
mcCc
collector side fluid capacity rate
mLCL
load side fluid capacity rate
nc
UeAe/mcCc, collector side number of transfer units
n|_
UeAe/mLCL, load side number of transfer units
r
UcAyUeAe
Ta
ambient temperature
Ti
load side cold fluid inlet temperature to heat exchanger
UL
collector loss coefficient
Ue
overall heat transfer coefficient of heat exchanger
REFERENCES
de Winter, F.(1975). Solar Energy 17. 335.
Duffie, J. A. and W. A. Beckman (1980). Solar Engineering of Thermal Processes. John Wiley
and Sons, New York
Holman.J. P. (1976). Heat Transfer. McGraw-Hill, New York
-------
1715
A PERFORMANCE PREDICTION OF A SOLAR ACTIVE HEATING
SYSTEM AND THE PRACTICE IN NORTH-WEST CHINA
Wei, Yi-Kang
Gansu Natural Energy Research Institute
77 s. Ding-Xi Rd., Lanzhou , 730000 , P. R. China
ABSTRACT
This paper gives a method of thermal performance prediction on the
basis of the local natural condition, meteorological and radi-
ation data, efficiency equations of the solar collectors and the
thermal load of the systemJtalsointroduces the main equipments of
the system, the operation modes and etc.. The experimental results
show that indoor space heated by active systems more comfortable
than the passive solar house.
KEYWORDS
Active solar heating ; passive solar heating ; forced circulation;
flat plate collector; vacuum collector; auxiliary energy.
INTRODUCTION
The mode of passive solar heating rests on the orientation of build-
ing location structure style and material thermal performances of
the building to absorb and store the solar energy for meeting the
house heating need. In recent years, it has developed quickly
as major mode of solar house in China due to the simplicity of the
structure and low cost. Though the solar active system needs more
equipments and much more cost than the passive, it finds wide use
abroad because it can overcome the effects of solar intermitten-
ce, climate and season, and keep stable and comfortable temp. in
the house. The solar active system for building heating and hot
water supply at the Solar Heating and Cooling Demostration Center
in Yu-Zhong county, Gansu Prov. had been built in 1988.
A METHOD OF THERMAL PERFORMANCE PREDITION
The thermal performance prediction is important for the solar
active system and is the base of design.The prediction can be used
this way:
1). Calculating the thermal load
2). Calculating the heat amount of the system from sun
3). Calculating the thermal compensation
To give a specific illustration:to design the active system of our
solar center, we used the above method.
1). The thermal load
-------
1716
d.
2)
to the
data,
vacuum
is the
Based on the indoor heating temp, as 16°c , the heating
thermal load is 40 kcal/hr.m*.
An hour load : The total heating area is 416
hour load is 40*416= 16640 kcal/hr .
Based on the 20 hours of the period of
each day, the heat amount needed is
kcal/day for average day.
The heat amount provided by the active
day is 332800*(1-60%)=133120 kcal/day
energy conservation rate of the solar house.
The collector thermal performance was predicted according
loca:l natural condition, meteorological and the radiation
as well as the efficiency equations of the flat plate and
collectors. Tab. 1 is the local meteorological data. Tab. 2
calculating results of thermal performance.
> so one
heating supply for
16640*20=332800
system for average
Where 60% is the
TABLE t . THE LOCAL,METEROLOGICAL DATUM & CONVERSION VALUES Sep. 1988
Month
Insblation Flux(MJ/day)
Avg. Water
Temp. C
Avg. Envir.
Temp. *C
Level
53°Slope
40°Slope
27°Slope
1
1 10.3
18. 2
17. 3
15.7
0. 0
-3.0
2
12. 5
17. 9
17.8
16. 6
2. 3
-0. 8
3
14. 3
16. 0
16. 6
16. 6
4. 0
6.0
4
18. 5
16. 8
18. 3
19. 1
9. 0
11.7
5
19. 8
15. 4
17. 4
18. 8
12. 0
15. 7
6
21. 3
15. 3
17. 7
19. 4
15. 0
19.5
7
19.9
14. 7
16.7
18. 5
18. 0
21. 4
8
17. 9
15. 0
16. 6
17. 7
21. 0
20. 4
9
13. 8
14. 1
14. 8
15. 0
17. 0
15.7
10
12.5
16.5
16. 4
16. 0
12. 0
10.3
11
10.5
17. 5
16. 8
15. 3
8. 0
4. 0
12
9. 8
18. 7
17.5
15.8
3. 0
-1.9
TABLE 2 . COLLECTOR ARRAY PERFORMANCE PREDICTION Sep. 28. 1988
VIont
Operation points
Efficiency Values
Heat Gd(kcal/mzday
total
(MJ/day)
53 "F. P.
40°V.
27 T • P.
53T\P.
40°V.
27 T- • P.
53"F- P-
40°V.
27T. P.
1
0. 0079
0. 0524
0. 0078
0. 65
0. 40
0. 61
11.8
6. 9
9.6
695. 1
2
0. 0079
0. 0456
0. 0078
0.63
0. 40
0.61
11. 3
7. 1
10. 1
697.9
3
0. 0000
0. 0312
0. 0000
0. 64
.0. 42
0. 65
10.2
7. 0
10. 8
685. 3
4
0. 0000
0. 0149
0. 0000
0. 62
0. 44
0. 66
10. 4
8. 1
12. 6
756. 6
5
0. 0000
0. 0355
0. 0000
0.59
0. 40
0. 65
9. 1
7. 0
12. 2
693. 7
6
0. 0000
0. 0256
0. 0000
0. 57
0. 40
0.65
8. 7
7. 1
12. 6
695. 1
7
0. 0000
0. 0222
0. 0000
0. 57
0. 42
0. 65
8. 4
7. 0
12. 0
668. 5
8
0. 0016
0. 0250
0. 0015
0. 59
0. 43
0. 63
8. 8
7. 1
11. 2
S58. 7
9
0. 0039
0. 0417'
0. 0038
0. 61
0. 40
0.63
8. 6
5. 9
9. 4
586. 0-
10
0. 0042
0. 0203
0. 0045
0. 65
0. 43
0.62
10.7
7. 1
9.9
675.5
11
0. 0108
0. 0360
0. 0108
0. 63
0. 42
0. 58
11. 0
7. 1
8.9
655. 9
12
0. 0127
0. 0491
0. 0127
0. 62
0. 40
0. 56
11.6
7. 0
8. 8
668. 5
Type, area, tilt and efficiency of collectors are as follows:
Type
Area M1
Tilt
Efficiency equation
Flat plate 1
Flat plate 2
Vacuum
28
28
13.9
27°
53°
40°
t =0.7-6.1 (Ti-Ta)/I
1 =0.7-6.1 (Ti-Ta)/I
% =0.5-1.6 (Ti-Ta)/I
-------
1717
Where Ti--inlet temp, of collector array.
Ta--environment temp. .
I --radiation intensity.
The least energy gain by the solar collector array is 586.0 MJ/day,
equivalent to 139957 kcal/day, which we got from the calculating
result, it is more than heat amount needed by the load.
3). The compensation of the electric heater
There is a electric heater connected in the system, if weather is
continuous cloudy and raining,the electric heater can meet the load
needs. On the assumption of operating 20 hours each day and an
energy exchange rate of 93%, the total power of electric heater is
332800/(860*93%*20)=20.8 ( kw ). In fact, we used a 24 kw electric
heater which is enough for the load.
THE MAIN EQUIPMENT AND HEATING LOOP
Fig. 1 is the schematic drawing of active solar heating. It
consists of collectors,heating circulation and hot water storages,
etc.. The solar collecting system is composed of collectors( flat
plate and vacuum collectors), circulating pumps, water distributor
and tank. It includes two independent systems in order to meet dif-
ferent needs in space heating season and hot water supply for other
seasons. Although the installed tilts are different for the two
groups,both of them can operate around whole year. The pressure and
flow rate were adjusted by the distributor between the two collector
groups, and the circulating pump forces the water circulate in the
hot tank and collectors.
| Cold
'Supply
To Cafeteria
r\ Water To
Distributor Bathr°om
Pumpi n
Hump
Cold
Tank
Hot Tank
heating and hot water supply systom
The heating loop contains circulating pumps, insulated pressure
water tank, electric heater, radiators and etc.. In sunshine days,
the hot water comes form the solar system. When it is cloudy or
snowing, thermal energy provided by the solar system is less than
the load need, the hot water in the insulated tank can be compen-
sated, if the water temp, is not high enough, the electric heater
can heat it.For protecting the collectors against freezing in win-
ter, there is a temp, controlled valve in the system for drain down
automatically when the collectors inlet temp, is lower than 2 °c .
-------
1718
measurement of the system performance
1). The measurement content and instruments
The thermal performance of collecting system, the operation situ-
ation of the electric heater and fluctuation of indoor temp.,in
winter,were measured. In the measurement of the thermal performance
of the collecting system, MTZ-280 thermometers were used for mea-
suring water temp. , which were installed in the inlet & outlet of
the main pipes in the collector system. The Li-175 pyranometer was
used for radiation measurement,radiation values on horizontal sur-
face were measured at first,then they were conversed to the diffe-
rent tilted surfaces. The wind speed was measured by the type of
DEM6 three cup anemometer and the environment temp. by a mercury
thermometer.
The data were recorded once each hour, so we can find out
the pattern of water temp, varied with the radiation flux,and
make out the efficiency of whole day.
2). The measurement results
Fig. 2 drafted the water temperature varying with solar radi-
ation flux according to the recorded data of July 28, 1989. The
daily efficiencies are 50.7%, and 49.0% separately, which was cal-
culated based on the reading data. Fig. 3 drafted the indoor temp,
varying with water temp, only heated by electric heater. Fig. 4
drafted the indoor temp, of this active and another passive solar
house for contrast varying with ambient temp, of Feb. 23, 1990.
Temp °c
Ambient lemp-
Temp.
6--40 9-40 I0W tM \ZM l44t 15:4b
summer Tine
Fig. 2,Water temp. as a
function of Ra. & Ta
. -vewc
emp
Tnlet Water
indoor temp.
#30 15:50 mo 1V50 18:30 1930 20:30
Summer Time
Fig. 3.Indoor temp, varying
with water temp.
Temp
°C~
15
10
5
0
-5
-10
T'.-.-mp. 'Passive
'l'emp. Ami
i—i—r~
i'emp. Active
Time.
0 2:00 4m 6:00 m fOW IM (440 1bW 16-00 &:oo22.WMW
Fig. 4.The comparison active room with passive room
-------
1719
ANALYSIS AND DISCUSSION
1). We can know that the water temp. of solar collector system
increases with the solar insolation flux simultaneously from Fig.
2. The increment tendency becomes slow when the water temp, is
raised to a certain temp..This is because , the more water temp,
increase, the more heat loss increase and thermal efficiency dec-
reases. When the absorbed heat quantity is equal to the thermal
loss,it reaches balance state, then the water temp, will no longer
be raised.
2). The energy gain of the calculating result based on the reading
data of July 28, and 29 are 124335 kcal and 139177 kcal. They are
77.9% and 87.2% of the results of thermal performance prediction,
respectively.The prediction could be more identical if we had con-
sidered the heat loss of the system pipes and other details.
3). The recorded maximum water temp. is 82.5 °c actually, it has
reached the predicted point.
4). Comparing the calculating results of the system efficiency, we
find the efficiency value of July 29 is lower than that of July 28.
Reasons for that are, the initial temp, of the July 29 is higher
than that of July 28,so the system thermal efficiency of July 29 is
lower than that of July 28, and another reason is wind influence.
5). We know that from Fig. 3, when the electric heater is on, the
water tempe. in the pressure water tank varies with the heating
time sharply.While the water temp, is getting higher,the heat loss
increases, then the water temp, varies slowly, finally it reaches
the dynamic balance.There is also an increase tendency of the room
temp, following the time extension of the electric heater operation
,The result of the electric heater operation shows that it can keep
a certain room temp, and meet the heating need as auxiliary energy.
6). The indoor temp, due to active system heating is higher than
passive heating because it has auxiliary energy in addition to
solar energy. Therfore it is more comfortable in the indoor space
and less affected by natural ambient condition.
CONCLUSION
1. The result of this experimental project shows that the active
solar system with electric heater as auxiliary achieved more satis-
factory result.
2. Result of the system performance predition in contrast with the
actual measured result is identical, which means that the perfor-
mance prediction method is practical and feasible.
3. The project is helpful for developing solar active systems which
serve both space heating and hot water supply. The writer believes
that active solar heating systems is easy to be popularized and
adopted in China.
-------
1720
STABILITY ANALYSIS OE SOLAR THERMO SYPHON WATER HEATERS
B.J. Huang, S.C. Du and R. H. Yen
Department of Mechanical Engineering
National Taiwan University, Taipei, TAIWAN 10764
KEYWORDS: Thermosyphon Collector; Solar Collector
ABSTRACT
Flow instability of solar thermosyphon water heaters is studied analytically. A system dynamics
model G(s) is derived by using 1-D approach and linear perturbation method. Nyquist criterion is
then used to examine the system stability. It is found that stability curves can be constructed by
using the critical value of the parameter M. It is concluded that flow instability will not occur in
most of solar thermosyphon water heaters commercially available due to high loop friction in design
and low irradiation in field operation.
INTRODUCTION
It has been noted that reverse flow phenomenon could take place in a solar thermosyphon water
heater. In general, reverse flow may result from thermosyphon saturation and hydrodynamic ef-
fect. It was suspected that reverse flow in solar thermosyphon water heaters may be induced by a
hydrodynamic instability which occurs very often in some simple natural circulation loops.
The stability of various simple-geometry natural-circulation loops has been extensively studied by
many researchers (Greif, 1988; Welander, 1967; Huang and Zelaya, 1987). Zvirin et al. (1978)
used linear system theory and assumed linear distribution functions for the steady and perturbed
(transient) temperatures in the collector and the tank or the heat exchanger situated in the tank.
This assumption results in a simple system characteristic equation which can be easily solved.
However, this assumption is not exactly triie. The same assumption was also used by Zvirin and
Greif (1979) to study the natural circulation loop that was studied by Welander (1967). But Zvirin
and Greif failed to verify the flow instability predicted by Welander (1967). They attributed this
error to the assumption of linear distribution for the perturbed temperatures.
In the present study, flow instability of a solar thermosyphon water heaters will be investigated
theoretically. A system dynamics model will be derived analytically by using the 1-D approach and
linear perturbation method. A system transfer function, which treats solar irradiation as the system
input and flow rate as the system output, is derived to represent the system dynamic behavior of
solar thermosyphon water heaters. The characteristic equation is then obtained and the Nyquist
criterion is used to examine the flow instability. Dimensionless parameters related to flow instability
are derived and stability maps are constructed.
GOVERNING EQUATIONS
The solar thermosyphon water heater studied is shown in Fig.l. Several assumptions in modeling
-------
1721
are used in the derivation: (1) The collector consists of a solid phase (absorber plate) and a liquid
phase contained inside the absorber; (2) Negligible heat capacity effects of the connecting pipe walls;
(3) Constant heat transfer coefficients; (4) Constant properties of water except the density in the
buoyancy term of momentum equation, i.e. Boussinesq approximation; (5) Well-mixed model is used
in modeling the energy equation of the tank.
Case 2: Doubled the friction of Case 1
Unstable
Region
Tank Lt
downcomer
^dv ^dh
* Case 1
a Case 2
Stable Region
Practical
Operation Region
=10
W/H=0.4, 7=27°
lc=0.2, It =0.1
uo=0.316, up=0
ut —0.014
rp=750.86, rt =12.01
ru= r„ =45.67
' 11111111111 r 11111111111111111111 r i
50 100 150 200 250 300 350
Fig.l Schematic of solar hot water heater. Fig.2 Stability map for different loop frictions.
Under these assumptions, the continuity equation is m = m(i). The energy and momentum equa-
tions are:
dT
collector plate : ppA„CPiP= Q~ Ua(Tp - Ta) - UW(TP - Tw)
dl* dT
water contained in collector : pwAnCPtW + mCp>w = Uw (Tp - Tw)
(1)
(2)
water in riser or downcomer :
water in tank :
P*KaCp^ ^ + mCPtW ^ - -Up(Ta - Ta) (3)
MCP|„ ~ = rnCP,w[Tu - Tt) - (UA)t(Tt - Ta)
(4)
where
dm
t'dt
: j Twey ¦ e,dy
F = 1 (h- + L^L + *!±\ f
tm pwgp{Aa+ Au + AjJ' F>
P
(5)
(6)
The derivation of eqn(5) uses a friction head relation expressed by a semi-empirical formula Hj =
c-md, where c and d are coefficients determined by a loop friction test suggested by Huang and Hsieh
(1985). The above governing equations can be separated into the steady state equations and the
perturbed equations by linear perturbation and expressed in normalized form. For the perturbed
equations (the dimensionless variables are defined in Appendix A):
d6'
rp~g^ =
-------
1722
39'
r—-—
dr
~g -fl, ( 6' - 0^, where r = rw for collector
+ f'-j^ + = < -up9'a, where r = ru for riser (8)
I —up9'm, where r — rd for downcomer
rt5 = ~ §t) + f {6'u ~ ^ ~ Ut9'' taDk (9)
W is related to the energy transfer
in the collector and the tank, Az represents the relative height between collector arid tank, wfm
represents the loop friction.
TRANSFER FUNCTION
Taking Laplace transformation of the perturbed equations with respect to time and assuming zero
initial conditions, we obtain the perturbed temperatures solutions in the s domain. Then com-
bining with the momentum equation (10), we obtain the transfer function: f'(s)/q'(s) = G(s) —
Gn(s)/Gd(s), where Gd(s) and Gn{s) are:
Gd(s) = Cl • ^w/m(j^-s + d- fd 1^yl-8-c1-c4-cs-c7-(Nc+h'n-hi-h'd)'^ (13)
Gn(s)=c3-(h!qc+h'qa-h'qt-h'qd) (14)
System stability can be determined by examining the unstable roots of Gd(s). The relations of all
the parameters in the above equations are presented in Appendix C.
STABILITY ANALYSIS
The Nyquist criterion is used to examine the unstable roots of Gd (s). A finite region bounded by a
s contour which covers a frequency up to 0.5 rad/s is used to examine the existence of positive real
part roots of Gd{s). A stability map that show stable and unstable regions can be constructed in
terms of some system parameters which can be divided into four groups: (1) geometric parameters:
It, lc, W/H, 7; (2) heat transfer and loop friction coefficients: ut, uOJ up and wfm\ (3) operating
parameters: steady-state heat flux, q\ (4) time constants: rp, rw, rt, r„, rd, and
Since q and Wfm appeared only in the steady state solution, eqn(12), it can be replaced by M.
Therefore, the neutrally stable states Mcr can be written mathematically in the following form:
jT )Ti J ^pi 'V) *"«« J J ^d, ) 0 (1^)
\ a Wjm)
The parameter M is related to the operating conditions. In some similar studies (Welander, 1967;
Huang and Zelaya, 1987), the stability map is expressed in terms of the heat input or flow rate. This
-------
1723
indicates that M can be chosen as the ordinate in the stability maps. This is verified by the stability
curves for two designs with different loop frictions which coincide with each other (Fig.2). Thus,
Mcr can be used to denote the critical values for the neutrally stable states. The stability maps can
be constructed by varying M until neutrally stable state: is reached, for other parameters fixed.
4 Case 1: W/H-0.3
o Case 2: W/H-0.4
* Case 3: W/H-0.5
Unstable
Region
Stable Region
Practical
Operation Region
wtm/wtm=lO
7 -17"
lc=0.2, lt=»0.02
Ug=0.316, Up=0
ut=0.014-
r„=750.86, r<=12.0
rup= rd»45.67
50 100 150
250 300 350
¦ Case 1
1 Case 2
» Case 3
¦» /*»
10
15
20
Unstable
Region
7 -IT, W/H=0.3
lc=0.2, it =0.02
U0=0.316, u„=0
u,=0.014- P
Stable Region rp=750.86
Practical rH =45*57
Operation Region r" —12 01
i i i I i n i 11 i iLT |-rrf i | i i i i | i TTT-fniT'i
50 100 150 200 250 300 350
Fig.3 Stability map for various system geometry. Fig.4 Stability maps for various loop frictions.
From Fig.2, the system will tend to be more unstable if the M value (operating conditions) increases
(by increasing the heat input or lowering the friction). For larger mass in the tank rt, the perturba-
tions in the loop are more easily damped out by the tank, thus the system tends to be more stable.
For a particular design with Ua = 11.0 W/m"C, Uw — 34.8 W/m°C, Up = 0 W/m°C,(UA)t =
5 W/°C,L, = 10 m,Lc/L, = 0.2, U/L, = 0m,WjH = 0.4,7 = 27°,T„ = 20°C,
PpAp = 30 kg/m,Cp,p = 870 J/kg°C,Aw = 0.00157 m2, A„ = Ad = 0.00038 m2,c = l,d= 1,
we obtain M < 0.012, which is far below the neutrally-stable curve in Figure 2. That is, the system
is stable.
The stability is also affected by the geometric design of the solar systems. Figure 3 shows that the
taller system (with lower W/H ratio) is more unstable. This coincides with the conclusion of Huang
and Zelaya (1987). Fig.3 also shows that the systems will tend to be more stable for larger lc or lt
(i.e. for longer collector or tank length). Since the M values for practiced operation of commercial
solar water heaters are far below these neutrally stable curves, solar thermosyphon water heaters
commercially available are usually stable.
It can be seen from Figure 4, that the larger the tratio, the lower the stability curves. That
is, the system will tend to be more unstable for a larger VHm/wjm ratio. In practiced designs, the
ratio is finite and always smaller than this value. This is due to the fact that increasing
wtm will automatically increase wfm. Thus, the M values for the practical operations of solar water
heaters are always far below the neutrally stable curves, and therefore systems are stable. The effects
of r„, r
-------
1724
heaters. The characteristic equation is then found and the Nyquist criterion is used to examine the
flow instability. The stability parameter M is derived and used to define the stability states. By
locating the M value in the stability maps, it is found that flow instability will, in general, not occur
for solar hot water heaters commercially available. This is due to the large water mass in the tank,
high loop friction, and low solar irradiation. The reverse flow that occurred in solar thermosyphon
water heaters must be due to thermosyphon saturation.
Acknowledgment - The authors are grateful for the financial support of the present project from
Energy Committee, Ministry of Economic Affairs, Taiwan, R.O.C. through Grant No. 782J1.
NOMENCLATURE
Ac collector area, w2. / flow rate, dimensionless.
Hf loop friction head, m. h thermosyphon head, dimensionless.
M stability parameter, dimensionless. Mt water mass in the tank, kg.
Q solar energy input per unit collector length, W/m. Q = IA^^a),,/Lc
q heat input per unit length, dimensionless. r time constant, dimensionless.
U heat transfer coefficient per unit length, W/rrfC.
(UA)t heat loss coefficient of tank, W/°C.
u heat transfer coefficient per unit length, dimensionless.
7 collector tilt angle, degree. i[> time constant, s. r time, dimensionless.
Az relative height between exchanger and collector, dimensionless.
steady state value. / perturbed value.
Subscripts:
a ambient c collector cr critical state
d downcomer fm friction head p collector plate, riser or downcomer
ref reference state i tank u riser w water z gravity direction
REFERENCE
Churchill, R.V., Brown, J.W. and R. F. Verhey,R.F., 1974. Complex Variables and Applications.
Third Edition, McGraw- Hill, Inc.
Greif, R., 1988. "Natural circulation loops," ASME J. Heat Transfer, Vol.110, pp. 1243-1258.
Huang, B.J. and Zelaya, R., 1987. "Stability analysis of a thermosyphon loop," ISES Solar World
Congress, Hamburg,1987.
Ong, K.S., 1974. "A finite - difference method to evaluate the thermal performance of solar water
heater," Solar Energy, Vol.16, pp. 137-147.
Welander, Pierre, 1967. "On the oscillator instability of a deferentially heated loop," J. Fluid Mech.,
Vol.29, pp. 17-30.
Zvirin, Y., Shitzer, A. and Bartal-Bornstein, A., 1978. "On the stability of the natural circulation
solar heater," Proc. 6th Int. Heat Transfer Conference, Toronto, Canada, 1978.
Zvirin Y. and Greif, R., 1979. "Transient behavior of natural circulation loops: Two vertical branches
with point heat source and sink," Int. J. Heat Mass Transfer, Vol.22, pp. 499-504.
Appendix A Definitions of Dimensionless Variables
./. PpApCPip ; PwAwCplVJ t MfCPtW , PwAuCPiW , PiuAfiCypui f M
Yp — JJ jVui — P7 I Wt — TT T—;Vu — 7} ",Wd— J} (A.1J
U w Ufu UyjL/s U rjj Uxo
m _Q ._ Q T-Ta y , L Ua (UA)t Up , .
ref 77 ' ^ T TT 1 T ' T 1 T 'Ua ~~ TT ,Ut ~ TT T ~~ TT
Uw IrefUw J-ref L„ L9 Uw UWLS Uw
-------
1725
w *
V'p, , , *!>t .
W' ""W '"W
FtmVw
= r—; rd = r^- (A-3)
Ire/ £re/
, mC,,„ _ .... ..
J~U„L.' Wtm~ Cp,w-trcfTrcf'
wfm =
trcf
*/™(c±-)d
Tre/ • L,
Appendix B Relations of Parameters and Coefficients in Steady-State Solutions
(A.4)
bi = exp
/(!+ «<•).
b2 = exp( y1"); bi - exp(—
_Q_
«a
(5.1)
+ yi(sin7) ; h4f) = 5uMl-bs) (B.2)
«o J **p
HI) = 6th ; **(/) = ^-(i'-M ; i5 = ; be = exP(~^idv) (b.3)
up j J
Appendix C Coefficients in Perturbed Solutions
fc =
/(I + «a)
_9_
«o
; = *4 = to
(C.l)
cx = rtorps2 + [r„,(ui ~ P4) (C.ll)
ru r
-------
1726
USE OF AN EXPERT SYSTEM FOR THE REAL TIME CONTROL
OF A SOLAR BUILDING
J. Nicolas, Ph. Andre, J.F. Rivez, V. Debbaut
Fondation Universitaire Luxembourgeoise
avenue de Longwy, 185
Arlon, Belgium
ABSTRACT
The F.U.L. building in Arlon, Belgium, includes a complex solar assisted heat
punp system, combined with a large seasonal in-ground storage. An experimental
research is now carried out into an expert-system-based control scheme for this
heating plant.
When the management of the plant is complex, such as in many solar buildings,
ah expert system is interesting for several reasons : editing control rules at
random and in natural language, introducing the heuristic knowledge, mixing
technical and economical optimization.
The expert system is written in Prolog on an Apollo workstation and is
organized in three levels: an inner core of pure software, an intermediary
shell, including thermal algorithms, and an outer shell available for user rules
and devoted to the specific application. The expert system must be considered
firstly as a tool that can improve the control flexibility by means of a user-
friendly rules handler. But the performance of the heating system will still
depend on the algorithms used in the rules rather than on the tool itself.
KEYWORDS
Expert system; solar heating control; optimization tool; knowledge-based real-
time management; inference engine.
JUSTIFICATION OF THE EXPERT SYSTEM APPROACH
The F.U.L. building in Arlon, Belgium, includes a hybrid solar-assisted heat
pump system linked to an underground storage (Nicolas and Poncelet, 1988). The
system was designed in 1977 and was finished being built in 1982. It is
intended for the air conditioning of two buildings with a total heat load of
900 GJ per year (Fig. 1).
The main features of the heating system are :
-a total area of 382 m^ of integrated solar collectors,
-a 500 m.3 water storage buried in the soil,
-an earth-storage volume with both horizontal and vertical exchangers,
-two heat puitps,
-a gas boiler as a backup.
The whole plant is supervised by an automatic control system ensuring
regulation, management, work optimization and scientific monitoring. The
control code, in PL.I language, is modular and has been designed so that it
could be improved in the light of experience.
-------
1727
However, the heating system is a complex one, with various energy sources and
many possibilities of heat distribution throughout the buildings. So, during
the 8 years of operation, the control strategy has been modified many times,
according to the experience of the maintenance technician, but also for
research purpose.
•ffifffffi
Fig. 1. Simplified schematic representation of the heating plant.
Each software modification requires a rather long and cumbersome procedure:
full stop of the heating system, search for the subroutine and for the program
line to modify, conpilation, linking and run test.
In general, when a computer is needed for the management of complex heating
systems, many control strategies are tested. To inplement such algorithms,
standard procedural and sequential codes are often developed. In most cases,
not only the algorithm, but also the core of the software are specific and must
be fundamentally recast for another building.
In order to test a more flexible solution to those modifications, we are now
carrying out experimental research into an expert-system-based control scheme
which could subsequently be made compatible with other heating plants (Nicolas
and co-workers, 1990) . The F.U.L. building provides an ideal testing ground for
the design and debugging of this control tool.
The expert system approach allows
-input of management rules to be carried out in natural language (French for
our irrplementation),
-solving the problem of performing the process control, while taking account
of economic optimization, failure diagnosis, forecast, simulation and so
forth,
-incorporating subjective assessments in the management, with the possibility
of using heuristic knowledge in the rules,
-------
1728
-the rules editing, in bulk, almost at random, in the knowledge base, without
changing the core of the computer code,
-learning, by appending observation to a knowledge base.
BASIC OPTIONS
Very few commercial expert systems exist for real time process control and
still fewer for the management of a heating plant. Among all the applications of
artificial intelligence in control and real-time operations collected in
abstracts reviews and in updated bibiography (N.T.I.S., 1990), a great number
deal with robotics and much fewer discuss expert system and knowledge-based
control. Moreover, most works relate to aeronautics and those relating to
energy field are focused on nuclear powerplants,.
We therefore chose to write our own expert system shell, for our specific
purpose.
Prolog has been chosen as the development language, as it is well suited to
first-order logic, needed for our problem. Its intrinsic mechanism corresponds
exactly to the process of an and/or graph path, typical of that used by an
inference engine. Also, backtracking is included as a standard feature of
Prolog : if a dead-end is reached, the language backtracks itself, or re-
explores the current rule occurring (e.g. a variable assignment in the premise
part), discards it, and searches for an alternative one.
The machine is an Apollo-3000 workstation, with 4 Mbytes core-memory. For our
particular application the interface with the process is achieved through the
serial line and is driven by a Pascal code.
We have developed a forward-chaining inference engine. Indeed, it is not
possible in our case to use a goal oriented search. This chaining mode, called
backward chaining, and used in diagnosis, sets a1 hypothesis and searches in
the data base for known facts which would enable this hypothesis to be
confirmed. In process control, it is impossible to set a hypothesis and the
inference engine must progress from known facts towards a " terminal action ",
i.e. a fact which is not used in another rule premise, for example a pump
switching on. At this stage, the system stops the consultation mechanism for
this inference and once more starts to explore the data base for a new one.
SYSTEM ORGANIZATION
The expert system is organized in three levels (Fig. 2).
-An inner core of pure software, including the inference engine itself and its
own codes (drivers, interfaces, ...), inaccessible to the final user.
-An intermediary shell, including typical functions of heat engineering, such
as power corrputation from a temperature difference and a flow rate, energy
integration, estimation of performance coefficient, exchanger efficiency
computation, ...
This shell includes also sane economic functions, such as the calculation of
energy cost on the basis of fuel prices (with a possible discount during
night periods) and consumption data.
All these functions may be called by the user in its rules.
-An outer shell, devoted to the specific application. It includes the user
rules written in natural language, and it is the only part of the system
concerning the user, the rest being transparent.
Around this three-level structure, a tool box is available, chiefly for user
interfacing purposes. It includes: a very simple CAD program to draw up the
installation diagram, rules and messages editors, a curve drawing tool, the
editing of holidays and schedules, the data recording files and the variables
declarations.
-------
1729
user rules
Diagram
display
Data recording
files declaration
¦neat engineerini
functions
CAD.
Messages
editing
Rules
editing
Curve
drawing
inference'
engine ,
Variables
declaration
Holidays and
schedules
editing
Fig. 2. The expert system organization.
MANAGING TIME
One of the most important feature of an expert system for real-time control is
obviously the ability to handle time and to wait for events. Time is managed
through a "lifetime" concept.
A lifetime, expressed in seconds, may be attributed to every item handled by
the expert system (variable, mode of operation, recording file, ...). The user
may write a rule that tests whether this item is "still alive", for example, in
order to trigger a mode of operation for a given period.
A typical application of this lifetime concept, in a set of rules for failure
detection, would be the comparison of the valve opening lifetime and the
waiting time before it reaches its limit stop.
The system presents a special behaviour when the words "wait_for(event)" (where
"event" is any possible event) occur in the conclusion part of a rule.
This case is considered each time an event must be waited for, in the real time
process, before continuing. At this level, the sequence freezes the rule and
will continue the depth-first search mechanism only when the event will occur.
In the meantime, another inference sequence may begin.
SYNTAX
The user rules are written in a file with a standard text editing tool, using a
specific syntax.
A general rule formulation could be as follows :
tag :: if p]_ and P2 or P3 then Cj_ and C2 cf n.
-------
1730
where
tag is a rule identifier, for the purpose of concise reference,
Pi? P2' P3 are Premise terms
ci, C2 are conclusion terms
n is the certainty or confidence factor,which indicates a degree of belief for
the rule, in the range 0 to 100.
The certainty factor is evaluated at each stage of the inference sequence for
new facts entering in the knowledge base as a result of the conclusion of a
rule. If the resulting confidence factor becomes lower than 20, the fact is
considered as false and the search fails.
An elementary term of a premise could take a form like
Element = Value, Element > Value, ...
In such a term, "Element" and "Value" are either the name of a variable of the
heating plant (such as t_out or q2), or a system variable (beginning with a
capital letter and identifiable to any value, depending on the actual knowledge
base), or a number, or a function, defined in the intermediary shell mentioned
above.
IMPLEMENTATION
At this stage of the development, the whole software is written, and real tests
with different set of rules are now carried on for our solar heating system.
The number of males that must be written to perform at least the same control
as the old one on our solar building is about 150. However, as far as the
performance of the heating plant is concerned, it is difficult to conclude in
favor of such an expert system. This knowledge-based real-time control is only
a different approach for the energy management, but the performance of the
control will always depend on the implemented rules.
However, this kind of tool improves the flexibility and the user friendliness
of the management.
For exanple, the user may write a set of "strategy rules", defining either some
periods with different climatic conditions, or some decision making according
to a threshold for tenperature or energy :
climatel :: if month = 10 or month = 4 then climate is medium.
beginl :: if climate is medium and t_out>10 then start_heating = 7.
Then the user may write a set of "technical rules" to specify which valve must
be opened or which pump must be switched on for a given mode of operation :
heating_valve :: if heating then open vl.
heating_pump :: if vl is open and not (thermal_trip pi) then switch_on pi.
It is also possible to write a set of "general rules", using the ability for
Prolog to handle variables. For exanple, the variable V may be identified to
any valve that has been opened as a conclusion of a former rule :
wait_limit_stop :: if opening V then wait_for (V is open) .
Some rules may be written for "failure detection" purpose :
thermal_trip :: if thermal_trip pi and climate is very_cold then
message ('Please verify puirp pi') .
-------
1731
Some other rules may also take into account "economical aspects". For exanple
if a pump consumes more electrical energy than the amount of heat supplied by
the solar collectors to the storage, then some other solutions could be tested:
do_not_worth :: if storage and
power (t_top, t_bottom, q) < powerjpwp(pl) then remove (storage) .
And finally, the "heuristic knowledge" may be included into the rules. For
instance, the experience of the maintenance technician may be taken into
account, even if this kind of rule is not 100 % credible :
experience :: if weather_rather_fine and hour <14 then reduce_heating cf 60.
CONCLUSION
Expert systems can encapsulate the abstract and experimental knowledge
characteristic of this kind of complex control.
Such an approach frees the engineer from the need to know the details of the
software and of the process interface and allows him to concentrate on control.
This research provides a basic framework for future developments in the field
of smart process control.
REFERENCES
Nicolas, J., and J.P. Poncelet (1988). Solar-assisted heat puirp system and in-
ground energy storage in a school building. Solar Energy. 40. 117-125.
Nicolas, J., P. Andre, J.F. Rivez and V. Debbaut (1990). An expert system for
the real time control of a building heating plant. Proceedings 3th BBB
conference. Liege, Belgium.
N.T.I.S. (1990). Published search in artificial intelligence in control and
real time operations. U.S. Department of Commerce, Springfield (U.S.A.).
-------
1732
COVER SYSTEM FOR LARGE ROOF-PLACED COLLECTORS
F. Kristiansen, S. Svendsen
Thermal Insulation Laboratory
Technical University of Denmark
Building 118, DK-2800 Lyngby
Denmark
ABSTRACT
A new cover system for large roof-placed collectors has been
developed and is described. The mechanical strength of the cover
system was tested. A 13 m2 collector with the new cover system
was built on an outdoor test rig on the campus area. The
efficiency of the collector was measured and compared with
existing collector systems.
Finally the price level of the new cover system is specified and
compared with an existing system.
KEYWORDS
Corrugated acrylic sheet, cover glued to laths, low price of the
cover, high solar transmittance, no perforations of the plate.
INTRODUCTION
Solar domestic hot water systems used in multi-story apartment
buildings, will typically have collectors covering the south
facing roof.
For economical and architectural reasons, it is interesting to
use site built collectors covering the whole roof surface instead
of using collector elements.
A new cover system has been developed for such collectors based
on a corrugated acrylic sheet.
The cover is glued to laths in order to avoid perforations of the
plate. Therefore there is a minimum risk of leakage of rain. The
plate is 6-12 m long and covering the roof from the ridge to the
base of the roof, so only sideways connections are necessary. The
sideways connections are made by traditional overlapping.
-------
1733
A special fixture system was developed to enable a short mounting
time of the sheets and secure that there would not be any
problems with the deformations of the cover due to thermal
expansion.
It is hoped that by the use of this new cover system, it will be
possible to significantly reduce the price of large collectors
and at the same time improve the durability of the collectors.
DESCRIPTION OF THE ACRYLIC SHEET
The cover is based on a corrugated acrylic sheet (polymethylmeth-
acrylate). The sheet is of the type "Vedrilser ER" from the
company Montedison. See fig. 1 and table 1.
TABLE 1 Dimensional properties
of "Vedrilser ER".
Geometrical properties
VEDRILSER
ER
Corrugation pitch (p)
mm
76
Corrugation depth (h)
mm
18
Thickness (s)
mm
1,5
Width (a)
mm
1300
Length (b)
mm
6000*
Number of corrugations
17
The acrylic sheet.
In table 2 the acrylic sheet is compared with a sheet of ordinary
glass.
TABLE 2. Data for the curruaated acrylic sheet and a typical
glass sheet.
Acrylic
Glass
Density
kg/m3
1170
2500
Thickness
mm
1.5
4
Weight
kg/m2
2
10
Coefficient of linear
thermal expansion
'c-1
75'10"6
8'10~6
Thermal conductivity
W/m-K
0.19
0.93
Maximum service tempera-
ture without load
°c
70
-
Normal solar transmittance
%
86
85
-------
1734
The acrylic material is known for its high weather durability.
The UV-radiation is not expected to reduce the solar
transmittance significantly.
The price level of the sheet is approximately 10 $/mi.
DESCRIPTION OF THE COVER SYSTEM
The principle of the cover system is to avoid perforations of the
acrylic sheet.
The acrylic sheet is glued to aluminum laths with an elastic
silicone seal, or by an acrylic based glue; The distance between
the laths is 0.5 m and 26 gluing points are used for each square
meter of the cover. See fig. 2.
Fig. 2. Acrylic sheet glued to aluminum lath.
The aluminum laths are hooked into special fixtures mounted on
the roof construction. See fig. 3.
acrylic sheet
silicone seal/acrylic glue
aluminum lath
[
aluminum ]
fixture
wood lath
Fig. 3. Special fixture.
The wood laths are fastened by screws to the roof construction.
-------
1735
DESCRIPTION OF THE COLLECTOR WITH THE NEW COVER SYSTEM.
The construction of the solar collector^ including the cover
system,is planned to take place as follows:
1. The acrylic sheets are cut to the proper length and
aluminum laths are glued to the sheet in the workshop.
2. On the roof wooden laths with the special fixtures are
mounted. See fig. 3.
3. The solar collector is built on the site by stepwise
installing insulation, absorber elements and the acrylic
sheets. See fig. 4.
The absorber elements would preferably have the same width as the
acrylic sheet and be made of strips with an extra distance in
order to leave space for the fixtures.
The acrylic sheet with the aluminum laths is installed by placing
all the laths on the special fixtures and sliding them into the
fixtures. The central lath is fixed while the others are left
free to move, due the relatively large expansion of the acrylic
sheet.
fixture
acrylic sheet
aluminum lath
absorber
insulation
roof cladding
wood lath
Fig. 4. Cross section of the collector.
A 13 m2 collector with the new cover system was built on an
outdoor test rig on the campus area. See fig. 5.
-------
1736
Fig. 5. Photo of the 13 m2 collector.
TESTS OF COVER AND COLLECTOR
The efficiency of the collector was measured and the results are
at the same level as for an existing collector with a similar
cover system and for collector elements with glass covers. See
fig. 6.
100%-
efficiency
Solar irradiance : 800 W/m5
90
80
new cover system
similar cover system
glass covered
70
60
50
40
30
20
mean fluid temperature - air temperature
o
20
40
60
Fig. 6. Efficiency of collectors with the new cover system,
similar cover system and glass cover.
-------
1737
The mechanical strength of the silicone seal joints between the
acrylic sheet and the lath was tested and withstood a stress of
3 kPa based on the area of the acrylic sheet. The acrylic glue
joint has also been tested and withstood a stress of 26 kPa. The
maximum negative wind load in Denmark is about 1 kPa.
Both materials used for joining of the acrylic sheet to the laths
have sufficient strength, but as the acrylic glue obtains its
strength within 20 minutes, while the silicone seal needs 24
hours, the acrylic glue is preferable.
No sign of damage was observed in the 9 months the collector
system has been in stagnation on the outdoor test rig. The system
allowing the thermal expansion of the sheet was observed to work
as planned.
PRICE
The price of the new cover system, including the acrylic sheet,
the aluminium laths, the special fixtures and the wood laths is
approx 40 $/m2. This is one third of the price of a similar
existing cover system.
CONCLUSION
It has been shown that it is possible to benefit from use of a
corrugated acrylic sheet as transparent cover for large roof-
placed collectors.
The cover and the fixture system were tested and can withstand
the maximum wind and snow load in Denmark.
As the sheet is glued to the laths, the sheet is not perforated,
and the risk of ingress of rain water is avoided. The special
fixtures allow the acrylic sheet to move due to the thermal
expansion.
The large sheet elements combined with the fixture system allow
the cover to be mounted very quickly.
The efficiency of collectors with the new cover is at the same
level as for similar collectors covered with glass or special
acrylic sheets.
The price of the new cover system is about 40 $/m2 which is about
one third of the price of a similar existing cover system.
ACKNOWLEDGEMENTS
This work was part of a R&D-project supported by the Danish
Energy Agency.
REFERENCES
Kristiansen, F., Svendsen, S. (1990). Udviklina af taasolfanaere.
Report 215, Thermal Insulation Laboratory, Technical University
of Denmark, (in Danish).
-------
1738
A COMBINED DUAL LIQUID-AIR FLUID TRANSPORT
MEDIUM IN AN OPTIMUM FLAT PLATE SOLAR COL-
LECTOR ENERGjY system
R. A Lift
Pius Parschplatz 10/1/4, A-1210 VIENNA
Austria
ABSTRACT
A performance study of semi-packed serpentines in a flat plate
collector system is presented. The combined use of a packed bed
and an unpacked piping system in the serpentines of a flat plate
collector yield a more effective solar energy collection system.
KEYWORDS
Pebble bed, semi-packed, unpacked, liquid/air, dual, efficiency,
optimization.
INTRODUCTION
The development of an optimally efficient combined dual liquid/
air fluid transport medium in a flat plate collector system is
described in this work. The programm objectives of this solar
system are to expand on the versatile use of a "conventional"
flat plate collector and its ability to meet the energy load when
solar energy is not available.
Solar Model
Generally, the solar system to be installed is directly related
to the load required. The introduction of an unpacked piping
system in a packed serpentine of a flat plate collector not, only
increased the collectors* ability to deliver solar heat in two
independent fluid mediums through one serpentine, but clearly
showed similarities to the behavior of fully packed serpentines
in their ability to store and deliver solar energy during inso-
lation and non-insolation periods.
-------
1739
The principal construction of such a serpentine is illustrated in
Diagram 1:
water inlet
air inlet
outer
Diameter
outer
pipe
inner pipe
unpacked
pipe
if
pebble
inner
Diameter
Diagram 1.
Combined Serpentine construction & cross sectional
view .
The flat plate collector using both water and air as the working
fluid was experimentally set up facing south with a 45 tilt.
The working fluid, being water for the unpacked inner piping
enters the collector at a constant fluid inlet temperature of
10 C with a volumetric flow rate of 0.04 lit/sec. and is mea-
sured hourly. The Pebble packed outer piping with a void frac-
tion of 36% uses air as the transport medium at a constant inlet
temperature of 0 C measured hourly.
-------
1740
RESULTS
Simulated Results &t Discussion
The simulated test took place in Winter at a latitude of 48°
The physical properties of the flat plate collector being:
(fin efficiency) F = 0.99
(collector efficiency) F ' = 0.9
( Collector heat removal factor) F^ = 0.88
(transmittance - absorptance product)fc*. = 0.84
The day length is calculated at 9.2 hrs, with sunrise being
7.27 am and sunset at 4.33 pm.
Figure 1 illustrates the overall efficiency with respect to
the time of day.
Efficiency
%
100
90
80
70
60
50
40
30
20
water
air
10
8
9 10 11 12 13 14 15 16 17 18 19
7
Time (hrs)
Fig. 1.
Overall efficiency with respect to the time of day
for water and air fluid mediums.
-------
1741
The results show that the unpacked inner piping with water as
the transport medium is very similar in behavior to that of a
packed pebble bed . The packed outer piping with air as the
transport medium shou/sa slower working efficiency during the
morning hours till about noon time. Due to the density of the
packing material, a longer time to heat up is required.
Significant is that the unpacked inner piping imbedded in a
packed bed serpentine, yielded about the same results as a
packed bed serpentine .
The combined dual liquid/air transport medium proved promising
in its ability to act as a storage system after sunset. Due to
the semi-packed serpentine, a significantly better capacity of
storing solar energy longer was achieved, and that for about
three hours.
Conclusion
The performance of a combined dual liquid/air transport medium
in a flat plate collector is feasib'le and it optimizes the over-
all performance and versatility of the collector.
The combined system can deliver solar heated liquid/air simulta-
neously or independently from the collector for storage or di-
rect thermal application.
Due to the similar behavior of the combined system to a fully
packed bed serpentine"1", the problems of weight and corrosion
with the use of liquid as the transport medium for packed beds
are overcome.
REFERENCES
+Alim, R. (1990). An Optimum Flat Plate Solar Collector Energy
System. In ISES Solar World Congress, Kobe, Sep. 4-8, 1989
Pergamon Press, New York.
Anderson, E. (1983). Fundamentals of Solar Energy Conversion.
Addison-Uiesley Publishing Company, Reading.
Duffie, J.A., W.A. Beckman (1975). Solar Energy Thermal Pro-
cesses. U/iley Interscience Publisher, New York.
Hottel, H.C., and B.B. UJoertz (1942). The performance of flat
plate solar heat collectors. Trans. ASME, 64. 91-104.
-------
1742
PERFORMANCE MEASUREMENTS ON TWO SELFBUILT
DOMESTIC HOT WATER SYSTEMS
M. Reuss, H. Schulz
Institute of Agricultural Engineering, Technical
University of Munich, Voettinger Str. 36,
8050 Freising, F.R. Germany
ABSTRACT
On a youth educational center about 35 km south of Munich, two
solar domestic hot water systems were installed to cover the
hot water demand of typical 2-5 m3/d;depending on the number
of guests from spring to autumn.
On the main building,a 108 m2 collector with an absorber made
of flexible polypropylene tubes supplies the solar heat to a
4.5 m3 nonpressurized storage tank. The second plant has a
54 m2 glass covered copper collector with selective surface.
The storage consists of three pressurized tanks of 1 m3 each
connected in series.
The project is subsidized by the Bavarian Ministry of Trade
and Commerce as a demonstration project. By monitoring the
systems, several mistakes were detected and could be eliminated.
The system costs (per 1 m2 collector area) are 695,— US$/m2
for the plant with the copper collector that of the plant
with the plastic tube absorber are with 275,— US$/m only
1/3.
KEYWORDS
DHW-system, monitoring of water heating collectors, do-it-your-
self solarsystem
INTRODUCTION
Within the frame work of energy-political seminars,two solar
domestic hot water systems were built and installed by the
participants on a youth educational center near Bad Toelz,
35 km south of Munich. The main objectives of this demonstra-
tion project wese theoretical and practical education of the
participants in solar water heating technology. Both plants
were designed to cover the high hot water demand of the educa-
tional center from spring to autumn. According to the smaller
number of guests in winter, only the system on the main buil-
ding is used. Our institute was responsible for the monitoring
programme following the installation of the plant.
-------
1743
DESCRIPTION OF THE TWO SYSTEMS
The plant at the main building (fig. 1) has a collector area of
108 m2 and a 4.5 m3 nonpressurized storage tank. The typical
hot water consumption is 2 - 5 m3/d,depending on the number
of guests. The collector is integrated in the roof with an in-
clination of 24° facing southwest. The absorber is made of a
black corrugated polypropylene plastic tube covered with a
transparent polycarbonate glazing.
domestic hot water
storage
storage
collector
Tu m
backup heating
system (oil)
©—
winter
summer
fresh water
Solar System Main Building
108 m2 PP-plastic collector, 4.5 m3 storage
Honlinger
LANOTECHNIK-
Fig. 1. Scheme of the DHW-system on the main
building
The solar heat is stored in the 4.5 m3 tank via an integrated
tubular heat exchanger with 44 m2 surface area. From this non-
pressurized storage the energy is transfeared via a second heat
exchanger to a 0.45 m3 buffer storage with domestic water, which
is also connected to the conventional oil heating system as a
backup.
The circulation pump SI is switched on/off by a simple tempera-
ture difference controller. A second one is used to control the
energy transfer from storage I to tank II.
Fig. 2 shows a scheme of the second plant with a 54 m2 glasf co-
vered copper collector with selective surface. This collector,
too,is roof-integrated with an inclination of 20° facing south-
west.
The hot water consumption is with about 1-2 m3/d less than that
in the main building. Additionally,in winter this plant is shut
down most of the time. Three pressurized storage tanks of 1 m3
each are connected in series. The control-unit is connecting the
-------
1744
DESCRIPTION OF THE TWO SYSTEMS
The plant at the main building (fig. l),has a collector area of
108 m2 and a 4.5 m3 nonpressurized storage tank. The typical
hot water consumption is 2 - 5 m3/d,depending on the number
of guests. The collector is integrated in the roof with an in-
clination of 24° facing southwest. The absorber is made of a
black corrugated polypropylene plastic tube covered with a
transparent polycarbonate glazing.
T 10
domestic hot water
storage
storage II
collector
backup heating
system (oil)
O—
4,5
xwinter
.—X 1
.. summer
fresh water
Solar System Main Building
108 m2 PP-plastic collector, 4.5 m3 storage
Hiinlinger
Pig. 1. Scheme of the DHW-system on the main
building
The solar heat is stored in the 4.5 m3 tank via an integrated
tubular heat exchanger with 44 m2 surface area. From this non-
pressurized storage, the energy is transfared via a second heat
exchanger to a 0.45 m3 buffer storage with domestic water, which
is also connected to the conventional oil heating system as a
backup.
The circulation pump SI is switched on/off by a simple tempera-
ture difference controller. A second one is used to control the
energy transfer from storage I to tank II.
Fig. 2 shows a scheme of the second plant with a 54 m2 glass"co-
vered copper collector with selective surface. This collector ,
too, is roof-integrated with an inclination of 20° facing south-
west.
The hot water consumption is withwabout 1-2 m3/d less than that
in the main building. Additionally, in winter,this plant is shut
down most of the time. Three pressurized storage tanks of 1 m3
each are connected in series. The control-unit is connecting the
-------
1745
three tanks to the solar collector depending on the temperature
level of storage.
collector
heating system
^ ~camp ground
domestic hot water
_^j_Jog cabin
backup (gas)
r* fresh water
Solar System Log Cabin
54 m2 copper collector, 3x1 m3 storage
Honlinger
Fig. 2. Scheme of the DHW-system on the log cabin
The energy delivered from the solar system is mainly used for
domestic purposes. Eks ess heat could be used for space-heating
of the log cabin if there is demand.
PERFORMANCE MEASUREMENTS
Both systems were equipped with a computerized data acquisition
system to register the required temperatures and flowrates in
the collector and domestic water -circuite. Additionally, global
radiation is measured in the collector plane at both locations
and ambient temperature in front of each building with a well
ventilated sensor. For all temperature measurements,PtlOO sen-
sors were used to achieve a reasonable accuracy. The data are
registered continuously and energy balances calculated. All
values were integrated over a time interval of 15 minutes and
stored as a sum or mean value on the hard disc. The final eva-
luation is done on the mainframe computer at our institute.
The hot water demand is varying due to the varying number of
guests in the wide interval 2-5 m3/d. The collector with the
plastic tube absorber was found to have a daily efficiency of
about 27 % (198 kWh), 73 % of the solar radiation are thermal
and optical losses on a nice day in spring. Only 84 kWh stored
energy could be delivered to the domestic water.
-------
1746
The major mistakes in the plant found out by the measurements
were the small flowrate through the collector circuit of about
20 1/hm2 and problems of heat transfer from the 4.5 m3 storage
to DHW tank of 0.45 m3. Control problems occured with the simple
-T control unit, the temperature sensors were located nearby
the heat exchanger between the two storage tanks. The heat is
not transfered continuously from tank I to storage II, but only
in the case of hot water demand, when fresh water is passing the
heat exchanger. For a high consumption of hot water - peak load -
the heat exchanger is too small. Therefore,it was suggested to
connect a second)already installed,small storage of also 0.45 m3
in series to the used one and operate the pump S3 for continuous
heat transfer between the solar and the domestic water storage
tanks and thus improving the thermal performance of the whole
system. The losses in piping and the storage tanks are less than
20 % of the useful solar thermal energy.
Fig.3 shows the energy distribution for a typical week in spring.
The solar system could have provided between 1/2 and 2/3 of ener-
gy demand if the losses of the whole DHW could be reduced. These
losses are not only due to the solar system but mainly to the
conventional installation. Especially an improvement of the insu-
lation of the small storage and the piping was proposed.
320
280
240
200
c
o>
160
I
120
$
80
40
0
Solarancrgft
Fr«rd«nargl*
En«rgl«b»dar<
Mon Die Mlt Don Fro San Son
Fig. 3. Energy distribution of the DHW-system of the
main building
The daily collector efficiency of the 2nd plant in the log cabin
was not higher than 21 % but with a mean fluid temperature, which
was about 10 - 20 K above that of. the plastic collector.
The major mistake in this plant was also a too small circulation
pump, which provided only a flowrate of 18 1/h m2.
-------
1747
ECONOMIC CONSIDERATIONS
As not yet a full set of evaluated performance data is available,
it is not possible to make a detailed economic evaluation. The
following table summarizes the costs of the two plants.
main building log cabin
collector area
108
2
m
54
2
m
storage
4.5
m
3
3
m
collector costs
82.—
US$/m2
200.—
US$/m2
system costs
133.—
US$/m2
435.—
US$/m2
labour
60.—
US$/m2
60.—
US$/m2
total
275.—
US$/m2
695.—
US$/m2
The cost comparison shows that the plant with the copper collec-
tor has to perform about 2.5 times better than the simpler pla-
stic collector. The measurements from spring only showed a per-
formance, which was about 10 - 15 % higher. Assuming a lifetime
of about 15 years of the installation with low maintenance,the
costs for thermal energy of the cheaper system will be in the
order of 5 - 10 cts per 1 kWh.
The cost for commercial system in- in the order of 900.— US$/m2
much higher than do-it-yourself solar systems with market avai-
lable components. The price for thermal energy for such commer-
cial plants was calculated to be 10 - 20 cts per 1 kWh based on
measurements carried out by TUEV Munich 1985/86.
CONCLUSIONS
The main objectives of the project was the demonstration that
solar DHW-systems could be built by do-it-yourself techniques
supplying energy almost cost effective.
The measurements showed that the systems had to be designed tho-
roughly. In both plants the circulation pumps were too small and
had to be replaced. Problems in the control unit were detected,
which influenced the solar heat transfeared to the domestic hot
water system.
Due to the widely varying hot water demand,the thermal perfor-
mance of the systems is also varying, and especially standby los-
ses are unexpectedly high.
The investment of both systems are 275.— US$/m2 for the plastic
and 695.— US$/m2 for the copper collector, both are cheaper than
commercialyavailable systems. The energy price will be n the or-
der of 5 - 10 cts/kWh.
ACKNOWLEDGEMENT
The authors would like to express their graditude the Bavarian
Ministry of Trade and Commerce who sponsored this demonstration
-------
1748
project and the monitoring programme. Projects like this are
helpful, in the promotion of saving conventional and increasing
the use of solar energy.
REFERENCES
M. Reuss: Do-it-yourself Solar Water Heating Collectors, FAO CNRE
Workshop on Solar Water Heating in Naxos, Greece, June 1988
M. Reuss, Dr. H. Schulz: Combined Solar Air and Water Heating
for Agricultural Applications, in: Proceedings of the 1989 ISES
Congress, Kobe, Japan, Vol. 2, p. 1505-1509
A. Hoess, et.al.: Sonnenenergie zur Warmwasserbereitung,' Koeln,
Verlag TUEV Rheinland 1987
F. Mittermair, et.al.: Solaranlagen - selbst gebaut, Karlsruhe,
C. F. Mueller Verlag 1990
J. Hoenlinger: Vergleich zweier Solaranlagen zur Brauchwasser-
erwarmung, Diplomarbeit an der Landtechnik Weihenstephan in
Zusammenarbeit mit der Fachhochschule Muenchen, June 1990
-------
1749
THE USE OF PROFILED METAL CLADDING AS A SOLAR COLLECTOR IN AN AIR-
SOURCE HEAT PUMP SYSTEM
D. L. Loveday
Department of Civil Engineering, Loughborough University,
Loughborough, Leicestershire, LE11 3TU, U.K.
ABSTRACT
The performance is estimated of profiled metal cladding as an air pre-heater in a residential-sized
solar-assisted heat pump system. It is shown that the seasonal heating mode coefficient of
performance is enchanced by 11%, that system energy requirements are reduced,and that the likely
number of defrost cycles is halved, compared with the use of an unheated air supply to the heat
pump evaporator. The assessment technique is validated with reference to monitored data from an
installation incorporating a tile roof pre-heater. It is concluded that such solar assistance offers
potential, but that full investigation of cost-effectiveness is warranted prior to larger-scale
application.
KEYWORDS
Solar roof collector, profiled metal cladding, steel-framed buildings, tile roof energy absorber,
heat pump system.
INTRODUCTION
In many industrialised countries, space and water heating in buildings account for a significant
fraction of national energy consumption (U.K. Energy Statistics, 1988). A contribution from
solar energy to satisfy this demand could therefore bring benefits. This often takes the form of
active or passive solar heating, or a cominbation of both. Sometimes, existing sections of a built
structure might be used as the collector (Trombe and others, 1977), or purpose-designed solar
collectors are used for pre-heating a fluid prior to energy extraction via a heat pump (Nicolas and
Poncelet, 1988). The use of conventional tiled roofs and profiled metal cladding as large-area,
low-efficiency solar collectors has been studied and their performance characteristics measured
(Loveday, 1983,1988). Profiled metal cladding, backed with insulation, is a very common
construction material in the U.K., often used for enveloping the steel framed buildings that are
erected as factory units, warehouses and sales outlets. The method is relatively inexpensive
compared with more traditional materials, and additionally offers speed and ease of construction.
Figure 1 illustrates such cladding, which can often be arranged to result in channels through which
air can flow. The air is heated by absorbed solar radiation, and if supplied to the evaporator of a
heat pump, produces an improved heating mode coefficient of performance, COP(H). As an
initial investigation of the potential of such an arrangement, the expected performance of a
residential-size system comprising a profiled metal roof pre-heater, an air-source heat pump, and a
store, is assessed. The study is based on monitored performance data for an occupied test-house
near Basingstoke, U.K., and the specific objectives are:
i) to validate the assessment technique by comparing the performance estimates derived from
the technique with those measured, for the case of a tile roof;
-------
1750
ii) to estimate the effect, on heat pump seasonal COP(H), system energy requirements and heat
pump defrost performance, of solar pre-heating via a profiled steel roof, and to compare this
with the cases of tile roof pre-heating and no pre-heating at all, and
iii) to discuss the implications for larger-scale application.
CORRUGATED METAL
SECTION
AIR CHANNELS
FOAM
BACKING
Fig. 1. Profiled metal cladding
TOE TEST INSTALLATION
This is an occupied three-bedroom detached residence near Basingstoke, U.K., and is
schematically-illustrated in Fig. 2. Living accommodation is on two levels, providing a total floor
area of 142 m2, and the heating system consists of the following units.
i) A conventional Redland 'Delta' tile roof acts as a large area, low efficiency solar collector;
air at temperature ta°C is drawn from outside and passes through channels formed by the
underside of the tiles and a backing of roofing felt, and is raised to a temperature tb°C. Each
side of the pitched roof is of area 55 m2, one side facing South-East, the other North-West
ii) An electrically-driven vapour compression heat pump, a Lennox HP7 split
evaporator/condenser model, is situated in the roof space. Air from the roof 'pre-heater1 at
temperature tb°C is drawn across the evaporator, and after cooling is expelled to the outside
at a temperature tc°C. The heat pump provides for some domestic hot water heating, but
mainly for space heating via a store.
iii) The store consists of three polypropylene bags filled with water (total capacity 21,000 litres)
and is recessed into part of the ground floor. Space heating is provided by a ducted warm
air system which exchanges heat with the store as required, and by supplementary portable
electric resistance heaters.
Relevant variables were logged hourly (temperatures and relative humidities) and weekly
(electrical energy consumptions, kWh, of the heat pump compressor, Pc, and fan, Pf, as well as
-------
1751
ancillary pumps, air blower, immersion heater, cooker, lights and power sockets). Additionally, a
Demand Profile Recorder, supplied by Normalair-Garrett Ltd, provided a half-hourly log of heat
pump electrical demand for more detailed modelling purposes. The installation was monitored
over two heating seasons (October - April) and from the data recorded, the in-situ performance of
the heat pump, together with the system performance, were modelled, and the energy
requirements of the installation evaluated.
South East
Energy Absorber
facing
Direct radiation
North West
facing
Indirect ^
radiation \
Tile or state
Felt
Roof space
Air out
Fibre giass (15 cm)
Heat pump
evaporator
Air in
Domestic hot
water supply
First floor
Heat exchanger
fan
, neax pump
/ condenser
L-Heat exchanger
to store
Ground floor
Thermal store
Fig. 2. The test installation
HEAT PUMP AND SYSTEM PERFORMANCE
Simple linear regressions, using unlagged variables, were found to provide adequate fits to the
recorded data. Based on weekly average figures, the COP(H) may be expressed as a function of
air source temperature at evaporator entry, tb, in °C, by:
COP(H) = 0.095^ + 1.774 (1)
(correlation coefficient, r, of 0.84). The sensible air temperature tc in °C after crossing the
evaporator is given by:
tc = 0.835^ - 3.512 (2)
(r of 0.98), and the air moisture content at evaporator entry and exit, gb and gc respectively, in kg
kg-1 dry air, are related by:
-------
1752
gc = 0.6524gb + 0.0012 (3)
(r of 0.82). Using a method described by Loveday (1983), the enthalpy change in the air crossing
the evaporator may be determined, based on the above correlations. This is used to obtain a value
for Q2, the energy removed from the air, in kWh, for known air mass flow rates and heat pump
operating times in hours.
In order to carry out a seasonal energy analysis, the COP(H) may also be expressed:
COP(H) = (Q2 + Pc)/Pt (4)
where Pc is the electrical energy consumed by the heat pump compressor, kWh, and Pt is given
by:
Pt = Pc + Pf (5)
where Pf is the electrical energy consumed by the heat pump evaporator fan, kWh. Each of these
(02, Pc, Pf and Pt) may be summed to give weekly or seasonal values, as appropriate.
By combining and re-arranging equations (4) and (5), Pc may be found from:
Pc = [Q2 - COP(H)Pf] / [COP(H) -1] (6)
Here, COP(H) may be found from equation (1), and a seasonal value for Q2 may be found as
described above, by adopting an average moist air mass flow rate of 0.9 kg s_1 across the
evaporator, and a heat pump total seasonal operating period of 2093 hours. Since the fan rating is
0.45 kW, the seasonal Pf value (for 2093 hours) is 942 kWh. Hence, seasonal values for Pc and
Pt may be calculated, the former from equation (6), the latter from equation (5). This approach is
necessary when evaluating alternative roof pre-heat systems.
The moisture absorption properties of the installed tile roof pre-heater system were found to be
represented by:
gb = 0.9315ga +0.0006 (7)
(r of 0.95), where the moisture contents are in units of kg kg"1 of dry air, at outside and
evaporator entry positions, ga and gb, respectively.
The seasonal purchased energy requirement for the installation has been estimated by balancing the
house heating load and supply, and involves adjusting the value of supplementary heating
requirement as necessary; incidental and solar gains are, in this analysis, assumed to offset
supplementary heating supply. Operational energy savings are obtained by comparing all cases to
that of the same house but without a heat pump or store, these being replaced by an electrical
resistance heating system. In this way, the effect of the existing tile roof as pre-heater/heat
pump/store system may be evaluated against a dwelling possessing none of these features.
Savings are presented in operating energy terms only, there being no attempt at this stage to
estimate financial benefits nor to conduct any form of life-cycle costing analysis. These are to be
the subject of further investigation.
ASSESSMENT OF COLLECTOR SYSTEMS
Existing Tile Roof Pre-Heating
In order to validate the assessment technique described, the performance of the system with the
existing tile roof as a pre-heater was firstly determined from an analysis of weekly measured data.
The seasonal COP(H), defined as:
-------
1753
, Total heat pump energy output for season
Seasonal COP(H) = ——— - :—: : (8)
Total heat pump electncal energy input for season
was found to be 2.59, with a seasonal requirement for purchased energy of 11958 kWh. This
may be compared with a corresponding estimated figure of 17005 kWh for the same installation
but with the heat pump and store being replaced by an electrical resistance heating system. This
indicates a saving brought about by using the tile roof pre-heater, heat pump and store, of 5047
kWh or 29.7%.
The performance of the existing system was then estimated using the correlations and
methodology presented, based on seasonal average figures for tb and ga. A comparison of
measured and estimated results is shown in Table 1.
TABLE 1 Comparison of Measured and Estimated Results for the Existing Tile Roof System
(with Heat Pump and Store
ITEM
MEASURED
ESTIMATED
AGREEMENT
WITHIN
SEASONAL COP(H)
2.59
2.57
1%
SEASONAL PUR-
CHASED ENERGY
11958 kWh
10975 kWh
8%
ENERGY SAVING*
5047 kWh
or
29.7%
6030 kWh
or
35.5%
20%
*compared with a resistance heating system
It is concluded that the assessment methodology is valid for intercomparing system performances
with differing roof collector systems. The comparisons which follow employ the estimated
figures for the tile roof system.
No Roof Pre-Heating
Here, the seasonal average values for ta (7.0°C) and ga (0.005758 kg kg"1) become the values for
tb and gb in equations (1), (2) and (3). This results in a seasonal COP(H) value of 2.44, a
seasonal requirement for purchased energy of 11762 kWh, and a saving over the resistance
heating system of 5243 kWh, or 30.8%.
Profiled Steel Roof Pre-Heating
It has been shown (Loveday, 1983, 1988) that a 'fin and tube' model adequately describes the
performance of this type of structure as an air-heating solar collector, and that, for an average
value for ta of 7.0°C, the air temperature is raised by 3.0°C by a roof pre-heating system
consisting of dark brown-coloured profiled steel cladding of dimensions and orientation
appropriate to the test house under investigation. The value for tb then becomes 10.0°C, and if it
is assumed that there is no moisture absorption in the trapezoidal channels (since they are steel-
formed, and are often backed by aluminium foil bonded to the insulation, offering high vapour
resistivity), then gb = ga = 0.005758 kg kg"1. The estimated value for seasonal COP(H) is then
estimated to be 2.72, with a seasonal requirement for purchased energy of 10816 kWh and a
saving over the resistance heating system of 6189 kWh, or 36.4%. (Note that tile and steel roofs
-------
1754
have here been assumed to have the same U-values). A summary of results is presented in Table
2.
TABLE 2 Effects of Differing Roof Pre-Heating Systems
TYPE OF
ENERGY
PRE-HEATER
SEASONAL
COP(H)
SAVING IN SEASONAL PURCHASED
kWh %
NONE - USE
OUTSIDE AIR
DIRECTLY
2.44
5243
30.8%
THE EXISTING
TELE ROOF
2.57
6030
35.5%
PROFILED STEEL
CLADDING
2.72
6189
36.4%
DISCUSSION AND CONCLUSIONS
Compared with the no pre-heating case, the use of profiled steel cladding as a large area, solar
roof collector can increase the seasonal COP(H) of the installed aiir-source heat pump by over
11%, and enhance energy savings by 946 kWh. Based on results obtained by Blundell, Heap and
Goodall (1977) for similar air-source heat pumps operated in the U.K. climate, the probable
numbers of defrost cycles per week are likely to be 43, 33 and 21 for the cases of no pre-heating,
tile roof and steel roof pre-heating, respectively. Consequently, there would be less wear on the
unit, together with reduced system energy losses. The comparisons presented are based on the
same fan power requirement for each type of pre-heater. It is estimated that the fan power would
increase from 0.45 kW to 0.78 kW for the profiled steel roof, with consequent reduction in
COP(H) and increase in purchased energy. The author is investigating a method to ameliorate this
drawback. It is concluded that the use of profiled metal cladding as a solar collector offers
potential as regards heat pump performance enhancement, but the overall cost-effectiveness of
such systems requires careful investigation prior to larger-scale application.
REFERENCES
Blundell, C. J., R. D. Heap and E. G. A. Goodall (1977). Heat pumps for space heating in the
U.K. - research and application. 4th Electric Space Heating and Air-Conditioning
Conference, Bordeaux, France.
Department of Energy (1988). U.K. Energy Statistics. Her Majesty's Stationery Office.
Loveday, D. L. (1983). Conventional roofs as collectors in a solar-assisted heat pump system.
Ph.D. thesis. University of Aston, Birmingham, U.K.
Loveday, D. L. (1988). Thermal performance of air-heating solar collectors with thick, poorly
conducting absorber plates. Solar Energy. 41. 6, pp. 593-602.
Nicolas, J. and J.-P. Poncelet (1988). Solar-assisted heat pump system and in-ground energy
storage in a school building. Solar Energy. 40. 2, pp. 117-125.
Trombe, F„ J. F. Robert, M Cabanot and B. Sesolis (1977). Concrete walls to collect and hold
heat. Solar Age. 2. p. 13.
-------
1755
SUMMER OVERHEATING PROBLEMS IN A SOLAR TV COMPLEX BUILDING
IN BRUSSELS, BELGIUM - ONE YEAR EXPERIENCE - RECORDS AND REMEDIES
A.C. GILLET
Atelier "D" S.C., Belle Voie 23, B-1300 Wavre, Belgium
ABSTRACT
The paper presents some observations recorded during the exceptionally hot 1990
summer, in a solar TV complex building situated in Brussels, Belgium.
In spite of some shading and refreshing systems, this bioclimatic low energy
building suffered from overheating inducing complains from the occupants.
Temperature variations were monitored and greenhouse effect was computed.
Extra cooling capacity is proposed, in the bioclimatic mode, through refreshing
the floor networks used for heating purposes during winter.
KEYWORDS
Bioclimatic building; greenhouse effect; solar gains; overheating; radiant
floor cooling.
INTRODUCTION
At our last world meeting in Kobe (1), we have presented a solar TV Complex
Building the heating of which is based on thermodynamic transfer of energy
between an integrated solar roof, various soil network exchangers, retrieval of
studios' artificial lighting heat losses, on the one hand, radiant floor and
pulsed air heating, on the other hand.
With the exception of some local building defects, the results of two
successive winter seasons were very satisfactory (2).
1990 summer was exceedingly hot, and overheating problems were suffered by even
most of the air conditioned buildings in Belgium. Greenhouse effect and
daylighting accentuated the phenomenon in our case.
(1) GILLET A.C. and de LAMINNE J.-M. in Clean and Safe Energy for Ever (1989),
Vol. I, p. 826.
(2) GILLET A.C., Thermodynamic Gestion and Thermal Comfort in a Solar TV
Complex Building in Brussels - First Results, Munich, BHRA International
Conference on "Application and Efficiency of Heat Pump Systems in
Environmentally Sensitive Times" (1-3 October 1990).
-------
underground storage
Fig. i
geothermal
collector
-------
£?R
III
glass cover
aluminium
louver
polycarbonate
underface
insulation
Fig. 2
Ui
-vj
-------
¦ outside
f i ambienl pyramid 1st floor
^ ambient average building
-------
1759
OBSERVATIONS AND MONITORING
The building is equipped with an automatic programmable regulating system which
records, hour by hour, the average temperatures and yields of the appliances as
well as the ambient temperatures outside and inside in 20 different zones.
It is thus possible to monitor the evolution of the situation and to learn from
these observations.
Fig. 1 gives a simplified cross view of the building showing some of the
features of the system. It should be noted that three quarters of the volume
is buried below ground level. The last quarter consists of a square base
pyramidal volume with two storeys above ground level. Only this volume was
affected by considerable daily fluctuations of internal temperatures during
last summer.
The diagonal of the square base is oriented north-south. The roof of the
building is formed by the triangular slopes of the pyramid. Two of them oriented
south-east and south-west constitute the integrated solar roof (1290 m2). The
other two slopes are well insulated conventional black steel roofs.
Fig. 2 gives a cut-away view of the solar roof. It fulfills two functions :
- preheating the ventilation air flowing between the internal and external
skins,
- heating cistern water fed through the hollow aluminium finned rafters,
with "drain down" when out of service, to prevent freezing.
The external skin is made of 8 mm thick tempered glass. One third of the
internal skin is composed of extruded polycarbonate double sheeting, for
natural daylighting of the premises, and two thirds;dsof opaque insulation
material.
Between glass and polycarbonate, specially designed fixed aluminium solar
shading louvers are incorporated in the thickness of the solar roof. The
tilting and the spacing of the aluminium blades are computed to ensure complete
hindrance of direct ray penetration between May and August. Hygienic air
ventilation (0.7 vol/h) is refreshed during office hours, and free cooling is
provided during the night by forced fresh air ventilation, but no air
conditioning is installed.
We made a computer simulation of the heat fluxes through the solar roof during
a typical summer day. It concludes with a maximum net heat gain of 68.6 kW for
the premises and a maximum radiant surface temperature of 43°C for the
polycarbonate area under semi stagnant situation (air flow 1,400 m3/h). A
check of this latter temperature with a contact electronic thermometer gives a
maximum of 45°C, validating the simulation.
Looking at the recorded temperature for that very day, we drew the diagram in
fig. 3. Despite outside temperature above 30°C, thermal mass of the fabrics
helps to keep the average building temperature under the set point of 25°C,
provided free cooling has been actuated during the preceding night. Anyhow,
solar gains in the pyramid volume increase its temperature above this set point
and legitimate complaints of the occupants.
-------
1760
DISCUSSION
It is obvious that the given system is unable to master the temperature
variations during hot sunny summer days. The main reason for this lies in the
greenhouse effect produced by the transparent parts of the solar roof.
Reduction of their area to one third of the roof and the existing shading
device allow a thermal flux superior to the daily buffer capacity of the
pyramid fabrics. Stratification in the open pyramid volume accentuates such
variations in the upper floor of the pyramid.
We were asked to design remedies for this situation and proposed the following
item :
- In the line of the bioclimatic character of the building, to cool the
pyramid by a flow of refreshed water through the existing radiant floor
network. An extra evaporator, added to the existing thermodynamic group,
refreshing by 1.4 K the 42 m3/h supplied to the network can give the 65 kW
cooling power needed.
- To cope with the stratification effect, we also suggest installing an
extra capacity under the second floor ceiling as direct expansion
ventilated units (20 kW cooling capacity).
- In order to evacuate the heat produced by this equipment, an extra
condensing unit will be installed.
CONCLUSION
If these propositions are implemented in due time, we hope to be able to
display further results by the time of the congress.
ACKNOLEDGEMENTS
This operation has been sustained by the Belgian State in order to promote the
rational use of nergy.
-------
1761
COMPARISON OF THE PERFORMANCE OF SOLAR HEATING SYSTEMS
ASSISTED BY HEAT TRANSFORMERS AND BY COMPRESSION OR
ABSORPTION HEAT PUMPS.
G.Galli*, L.Laurent!*, A.Ponticiello*
*Dipartimento di Energetica, Universita di L'Aquila
Facolta dl Ingegneria, Italy
ABSTRACT
The performance of solar heating systems combined with compression or
absorption heat pumps, or with absorption heat transformers are compared; a
double lift heat transformer and a double effect absorption heat pump are
considered as well.
The fraction F of the thermal load supplied by the heating systems, and the
primary energy requirement, PER, are evaluated.
The simple and double lift heat transformers slightly improve the performance
of the conventional solar system; the simple absorption heat pump shows the
highest F values, while the double effect machine requires the lowest amount of
the primary energy.
KEYWORDS
Domestic solar heating; absorption heat pump;compression heat pump; heat
transformer.
INTRODUCTION
Combinations of solar systems and various types of heat pumps have been often
proposed: the performance of solar assisted compression heat pumps (Freeman,
1979; Anderson, 1980; Cesarano, 1985) absorption heat pumps (Lazzarin, 1981),
and absorption heat transformer (Galli, 1990) has been evaluated.
These complex systems are proposed usually to improve the performance of the
machine, as they increase the temperature of the source that supplies heat to
the machine, and to improve the efficiency of the solar collectors, lowering
the inlet temperature of the heat collecting fluid.
In this paper the performance of three types of heat pumps,, combined with a
system of flat solar collectors, for the same climate conditions and thermal
load, are compared: compression heat pump, simple and double effect absorption
heat pumps, simple and double lift absorption heat transformers.
A detached house of 160 m floor area is considered. The thermal load is given
by a space heating load of 23.000 kJ/°C day and a warm water load of 18.3
GJ/year evenly distributed during the whole year.
The heating plaint must keep the temperature of the indoor air at 20°C, except
during the night, when the plant is turned off for seven hours.
Outdoor temperature and solar radiation data that reproduce mean climate
conditions of Rome and Milan are considered; i.e. both Mediterranean and cool
continental moist climate are considered.
-------
1762
Figufce 1 shows a general schematic diagram of the proposed systems. Two
heat storage tanks are provided: S2 supplies heat to the heating plant of the
house; SI is kept at a lower temperature.
A machine that is able to transfer heat from SI to S2, works between the two
heat storage tanks. The solar collectors may give heat either to SI or to S2.
It is assumed that a "perfect" control and regulation system is available, so
that is possible to measure accurately the actual outdoor temperature, the heat
storage tanks temperatures and the solar radiation.
STORAGE
TANK
STORAGE
TANK/
MACHINE
S2
Fig. 1. General scheme of proposed heating system
Only the energetic efficiencies of the proposed systems and of a conventional
solar system are compared; no economic analysis is performed.
The fraction, F, of the thermal load supplied by the combined solar systems,
the ratio, SPE, of the spared primary energy to the thermal load, and the PER
(ratio of the heat supplied by the system and the required primary energy) are
evaluated.
It is assumed that the auxiliary heating system burns the same type of fossil
fuel that is used by the absorption heat pump and by the thermoelectric plant
that supplies electricity.
SYSTEM DESCRIPTION
Compression heat pump J[ CHP 2
The pump works when its COP is such that :
T) COP - V> 0 (1)
tn d
where Ti^is the efficiency of the heat-to-electricity conversion process, and
7} is the combustion efficiency of the auxiliary heating system; they have been
D
assumed : i) = 0.3 and T) = 0.8
th b
The heat collected in the solar panels is given to SI when the following
conditions is verified :
COP i} - 1}
S - Ul( Ts2 - Ta ) < [ S - Ul( Tsi - Ta )] (cop,!)11,, " (2)
otherwise it is given to S2.
-------
1763
In (2) S is the flux of solar energy Incident on the collector surface, Ul is
the heat loss coefficient of the solar panels , Ta is the outdoor air
temperature and Tsi and Tsa are the temperatures of SI and S2 respectively.
Condition (2) evaluates how much primary energy is spared when the solar heat
is given to SI instead of S2.
A water-to-water heat pump is considered; its performance is determined by
interpolating empirical performance data taken from manufacturer's
specifications.
Coefficients of performance between 2.4 and 4.5 are obtained, according to the
values of evaporators and condenser temperatures.
Heating plants requiring utilization temperatures, Tus, either of 40 °C or 503C
have been considered.
When the heat pump is working, the temperature of Tsi can vary only between
13°C and 23"C. The heat pump starts working only when Tsa is less than Tus.
Heat transformer ( HT )
The solar heat is given to SI or S2 according to whether or not the following
condition is verified :
S - Ul( Ts2 - Ta ) < COP [ S - Ul( Tsi - Ta ) ] (3)
When the thermodynamic properties of the working mixture are known, the unitary
heat powers exchanged in the components of the machines can be evaluated
(Ziegler, 1987;Moser,1985).
The working mixture is water-lithium bromide; its thermodynamic properties are
taken from (Mc Neely, 1979).
The efficiencies of all the heat exchangers are assumed equal to 0.75; the mass
flow rates of the heat carrying fluids are such that the temperature
differences in the heat exchangers are not larger than 5"C.
The temperatures and concentrations of the working fluids in the various
components of the machine, and the heat capacity are determined by the heat
transfer efficiencies and by the requirements of thermodynamic equilibrium
between the phases and equality of pressures in absorber-evaporator and
condenser-generator.
The HT works only when the temperatures of the source that supplies heat to the
machine, i.e. Tsi, and the cold source temperature, Ta, are such that the
temperature of heat released in the absorber is larger than Tus.
Absorption heat pump 1 AHP 2
An absorption heat pump working with H20-LiBr is considered; its performance is
evaluated by the same method used for the heat transformer. The convenience of
using a H20-HBr AHP combined with a solar system can be shown by the following
considerations: ^f the temperature of the cold source is not lower than 15°C,
the danger of frost formation in the evaporator is eliminated; besides, in a
machine working with water-salt mixtures, the temperature difference between
absorber and evaporator cannot be larger than 40°C, otherwise salt
crystallization may occur.
Solutions that have no solubility problems, such as NH3-H2O or R22-E181 etc.,
usually have lower COP, often about 1.3, while a H20-LiBr AHP can reach COP
values of the order of 1.5 + 1.7.
in the proposed system, the solar panels supply heat usually to SI; only when
Tsi is larger than a maximum value (assumed equal to 25°C) the solar heat is
given to S2.
-------
1764
Two stage absorption systems
The performance of a double temperature lift heat transformer (DL HT) and of a
double effect absorption heat pump (DE AHP) combined with a solar system have-
been evaluated as well.
The ideal diagrams of the two stage cycles are shown in fig.2.
The DL HT approximately doubles the temperature difference between the heat
supplied by the absorber and the source that gives heat to the machine, i.e.
Tsi; its COP can be calculated by means of the COP of a simple machine that
works with the same solution and has a comparable structure (Ziegler,1987):
COP
COP = — (4)
2 1-COPj+COP ^
where COP is the COP of the two stage machine and COP is the COP of the
2 1
simple machine.
double lift
heat transformer
log p
0
Ta
Ts
double effect
absorption heat pump
log p
0
Tus
Ts
T
Fig. 2. Characteristics of DL HT and DE AHP
The double effect AHP has a high temperature generator, and a second generator,
at the same temperature of the generator of a simple machine, that receives
heat from the higher temperature condenser.
The COP of a double effect AHP can be evaluated by means of the COP of a simple
machine using the following relation (Ziegler, 1987) :
COP = COP + (COP - l)2 (5)
2 1 1
The equations (4) and (5) are non-exact, but, if used for H20-LiBr machines
should be affected by errors not larger than a few percent..
Results and conclusions.
In fig. 3 - 4 the fraction, F, of the thermal load supplied by the combined
.systems and by a conventional solar system and the fraction of the spared
primary energy, SPE, are reported as functions of the collector area for two
values of the utilization temperature. Single and double glazed collectors are
considered.
F and SPE values for double lift heat transformer are not reported, as they do
not differ appreciably from those of the simple machine.
-------
1765
Compression and absorption heat pumps show remarkably higher values of F and
SPE factors than a conventional solar system (CSS); while both simple and
ROME
100
100
1 cover
50 C
1 cover
50 C
80
80
60
60
SPE(%)
40
F(%)
40
20
20
20 40
collector area (m2)
60
60
0
o
20 40
collector area (m2)
100
100
2 covers
50 C
2 covers
50 C
80
80
60
60
SPE(%)
40
F(%)
40
20
20
20 40
collector area (m2)
60
60
0
0
20 40
collector area (m2)
a conventional solar o AHP « DE AHP ¦ HT ~ CHP
Fig. 3.
double lift heat transformers show only slight improvements, when compared with
CSS. Better performance of combined solar heat transformer systems can be
obtained with other system arrangements and for particularly low levels of the
solar radiation (Galli, 1990).
It must be observed that the SPE values are significantly higher for the CHP
than for the CSS only for single glazed collectors, but became comparable for
double glazed collectors.
The double effect AHP has F values lower than the simple AHP, although it has a
better energetic efficiency. But, just because of its higher COP, the double
effect AHP can utilize the same quantity of low temperature heat as the simple
AHP, requiring less high temperature heat, so that the total heat supplied by
the machine is lower, but the consumption of primary energy is considerably
reduced.
-------
1766
MILAN
1 cover
40 C
20 40 60
collector area (m2)
1 cover
50 C
20 40 60
collector area (m2)
100
60
F(%)
40
20
2 covers
50 C
-
60
SPE(%)
40
20
1 cover
40 C
60
SPE(%)
40
20
1 cover
50 C
100
60
SPE(%)
40
20
20 40 60
collector area (m2)
20 40 60
collector area (m2)
2 covers
50 C
20 40
collector area (m2)
20 40 60
collector area (m2)
A conventional solar o AHP • DE AHP ¦ HT a CHP
Fig. 4.
-------
1767
In fig.5 is reported the ratio, PER, of the total heat supplied by the machines
and the required primary energy.
The DE AHP has significantly higher PERs than the other heat pumps: for the
same collectors area and heat storage volume,it gives lower F values, but
requires less primary energy. When the same amount of primary energy is used,
the DE AHP gives higher F values, but that can happen only when larger
quantities of low temperature heat are available, that is, when the collector'
surface and heat storage volume are larger.
MILAN
1 cover
PER {%)
20 40
collector area (m2)
A AHP O DE AHP
Fig. 5.
When the volume of SI storage tank per unit of collector area is increased, the
temperature of SI is reduced so that a larger fraction of the collected solar
heat is stored in SI and a smaller fraction bypasses the heat pump, so that F
values increase (fig.6) while the PER diminishes.
MILAN
so c
40'
F (») <
175
75
125
225
SI storage tank capacity (1/rn1)
~ CHP A AHP O DE AHP
Fig. 6.
It can be concluded that the CHP could match the performance of the AHP and DE
AHP only if its COP would be considerably improved. In principle, this
improvement is possible CMeijer, 1990) and might be reached in a not too far
'future; on the other hand, the AHPs considered in this work have quite good
COPs,and marked improvements in a near future do not seem very probable.
-------
1768
REFERENCES
Anderson, J. V., J.W. Mitchell, and W.A. Beckman (1980). A design method for
parallel solar-heat pump systems. Solar Energy. 25. 155-163.
Cesarano, A., F. De Rossi, and V. Naso (1985). La pompa di calore
.elioassistita La Termotecnica. 1. 89-97.
Freeman, J.L., J.W. Mitchel, and T.E. Audit (1979). Performance of combined
solar-heat pump systems. Solar Energy. 22. 125-135.
Galli, G., L.Laurenti (1990). Absorption heat transformer applications to
solar heating systems. First World Renew.En.Conar. Reading, 1174-1180.
Lazzarln, R.M., (1981). Solar assisted absorption heat pump feasibility. Solar
Energy. 26. 223-230.
Lazzarin, R.M., (1981). Sistemi solari attivi. F.Muzzio ed. ,. Padova.
Mc Neely, L.A., (1979). Thermodynamic properties of aqueous solution of
lithium bromide. ASHRAE Trans.. 85. 413-434.
Meijer, G.J.A.M., (1990). Advances in electrical heat pumps. JEA Heat pump
Centre. Newsletter. 8,2 9-11.
Moser, F., H.Schnitzer (1985). Heat pump in industry. Els.Sci.Publ.
Ziegler, F., G.Alefeld (1987). Coefficient of performance of multistage
absorption cycles. Int.J.Refrig.¦ 10. 285.
-------
1769
THERMOCHEMICAL ENERGY STORAGE WITH LOW TEMPERATURE HEAT
FOR SPACE HEATING
S.Fischer, A.Hauer, S.Hoist, W.Schoelkopf
Physics-Department of the Ludwig-Maximilians-University
Prof. Dr. R. Sizmann
Friedenstr. 18a, D-8000 Munich 80
Federal Republic of Germany
ABSTRACT
Zeolites adsorb water vapor in an exothermic reaction ( Wijsman, 1979). This can be used for
thermochemical storage of exergy with low temperature heat ( Sizmann, 1987). The storage
system is operated as a heat transformer in an open loop attached to the ambience. The
enthalpy content of ambient air can be exploited as a low temperature heat source. Losses
occur during charging (desorption of water vapor) and discharging (adsorption of water vapor
carried by humid air). The maximum attainable energy storage density in zeolite 13X is about
900 MJ/t .
An experimental arrangement has been built for a thermal output up to 10 kW . Energy
densities of 615 MJ/t were measured (charging at 130 °C and discharging with water vapor
saturated air of 25 °C ). The thermal COP of an adsorption-desorption cycle is found to be
0.6. Improved cycles with a partially closed air loop during the charging mode, or with an
internal heat exchanger submerged in the zeolite for providing the heat of desorption, led to
COP values of 1 and higher.
KEYWORDS
Thermochemical energy storage; open air loop; solid adsorbents; zeolites; adsorption heat
pump; space heating;
INTRODUCTION
Heat generating systems cause energy losses and pollution of the environment. The use of
district heating systems with combined power and heat generation saves primary energy and
reduces the emission of pollutants. The fact that local district heating systems are often
operating at the upper limit of their capacity, asks for new techniques for the connection of
additional consumers to the system. Decentral storage devices installed in the heating net
decouple the instantaneous demand for heat from the momentary supply of the combined
power and heat generating system. Therefore, the demand profile of the district heating net
can be arranged to a better power balance and the number of consumers may even be increased.
In particular thermochemical storage systems are useful in the present context. They can be
chaxged off-peak. In times of peak net load they use only low temperature heat extracted from
the return flow of the district heating system. The system can be used as an adsorption heat
pump and thereby helps to achieve a well-balanced net load. By lowering the temperature of
-------
1770
the return flow of the system, the effectively transported power is increased and heat losses
of the net are reduced. In addition thermochemical storage systems offer high energy storage
densities and there is no degradation by heat losses in long term storage. Thermochemical
reactions follow the reaction scheme:
Heat is absorbed when AB (for example the zeolite water complex) dissociates into Bg (water
vapor) and A (dry zeolite)(desorption = charging). The water vapor is carried away by a flow
of an inert gas (air), so that a reverse reaction is prevented. Blowing water vapor as a part
of humid air through the dry zeolite, leads to an increase in temperature on account of the
heat of reaction. This heat can be used, for example, in space heating systems (adsorption =
discharging).
We have been investigating zeolite/water vapor as a storage system operated in an
open air loop attached to the ambience with air as a carrier of heat and water vapor
(D. Jung, 1983)(S. Hoist, 1990). The adsorbent zeolite 13X (Bayer) is nontoxic, inflammable
and easy to handle in a packed bed of spherical pellets of 2 - 4 mm in diameter.
In order to explore the behaviour of a decentral thermochemical storage in a district heating
system, an experimental arrangement for a thermal output of 5-10kW was built. The exper-
imental arrangement (Fig. 1) consists of commercial components and was built in a compact
form for minimizing heat losses (S. Fischer, 1990).
AB -f- heat A -I- Bg
EXPERIMENTAL ARRANGEMENT
HEl >
heat source /
heating system
TCS
HE2
low temperature
energy source
Fig * 1. Scheme of experimental arrangement
HEl : Heat exchanger for heating/regeneration, HE2 : Humidifier/Heat exchanger for low tempera-
ture heat, HE3 : Heat exchanger for recovery of heat, TCS : Thermochemical storage
-------
1771
The central component of the arrangement is the thermochemical storage TCS (volume
0.5 m3 with 351kg zeolite). The heat exchanger HE3 is used for recovery of heat and HE1
for heat input of desorption and heat output of adsorption respectively. A special low tem-
perature humidifier was developed, consisting of a conventional heat exchanger HE2 with a
hydrophilic coating of internal surfaces which are kept wet. The humidifier produced a humid-
ity of 98% at an air temperature of 25 and operated thermodynamically almost perfectly.
The data acquisition system collects samples of 72 temperatures, 3 dewpoints, 4 mass flow
rates, 2 air pressures, 2 heat flow rates and 5 electricity counters with a repetition time of
40 seconds. The horizontal temperature profile is measured at three levels inside the storage
containment. The electric and hydraulic equipment is computer controlled. That allows to run
different heat demand user profiles and to couple the storage to various heat sources (district
heating system, solar collectors, industrial waste heat).
ADSORPTION-DESORPTION CYCLES WITH ZEOLITE 13X
Charging and discharging of the storage under different conditions was investigated. A typical
cycle is with input of low temperature heat at 25 °C and output of heat into a heating system
of 50 °C inlet and 30 °C outlet temperature.
Ambient air of S^C (dewpoint 0 °C ) is available. It is heated to 25 °C and saturated simul-
DESORPTION
ADSORPTION
HE 1
HE 2
TCS
HE 3
Ambient air
15/0
~fm/o
| 30/30
HE 1
. District HE 2
130 heating
TCS
Humidifier
HE 3
Ambience
Ambience
\l5M0
T98/-10
125/25
Heating system
(50/30)
15/0
Ambient air
Low temperature
25 energy source
Fig. . 2. Desorption and adsorption of zeolite 13X
HE1 : Heat exchanger for heating/regeneration, HE2 : Heat exchanger for low temperature heat,
HE3 : Heat exchanger for recovery of heat, TCS : Thermochemical storage, 15/10 : Air temperature
lb"C J Dewpoint 10 *€
taneously. Then it is blown into the adsorption tower containing dry zeolite. The air
leaves the zeolite with a constant temperature of 98 °C and a dewpoint of -10 °C . Heat ex-
changer HE2 delivers useful heat to the heating system. The adsorbens is finally in a second
phase of the cycle regenerated by hot air of 130 °C (dewpoint 0) to a water content of C =
0.09 kg water/kg zeolite. The temperature level of 130 °C is typical for steam-operated district
-------
1772
heating systems. Saturated air of 30 "C leaves the storage. After recovery of heat to the inlet
air it is blown into the ambience.
Experimental data on storage energy density p (pv- per volume,/^ per mass) and thermal
Table 1: Storage energy density and thermal COP of a thermochemical system with zeolite 13X
Heating system: 50/30
Ambient air of S^C (dewpoint Ot! )
Initial water content C = 0.09 kg water/ kg zeolite
Cycle
/>v[MJ/m3 ]
Pw[MJ/t ]
COPth
experimental
430
615
0.6
ideal
490
700
0.75
Improved cycles
part, closed loop desorption
430
615
1.0
isothermic desorption
360
520
1.1
pv '¦ Storage energy density per volume, pm '¦ Storage energy density per mass, COPth : Thermal
Coefficient of Performance
COP (COPth is useful heat produced in adsorption / heat supplied for desorption) for a storage
system with zeolite 13X resulting are shown in Tab. 1. Included are calculated data for ideal
heat exchangers (efficiency = 1.0). The criteria to stop the desorption at the time when 90%
of the maximum energy density of the storage is reached, appeaxs to be an optimum operating
condition. The economically optimal heat exchanger area for heat recovery can be determined.
The thermal COP is below 1, due to the high enthalpy of the outlet air during desorption.
This air contains latent heat of vaporization of the desorbed water at a temperature level of
30 which at present we did not consider for use in a consecutive low temperature heating
system.
IMPROVED CYCLES
To improve the COP of the process we raised the outlet temperature level during desorption.
Then part of the output of the latent heat of the air can be used in the system itself.
Partially closed loop desorption:
A closed air loop is used at 130''C and a high dewpoint of 30 "C for desorption. Then the
temperature of the saturated air at the outlet reached 40 "XU . In a condensing heat exchanger
the air temperature was lowered to 30 "C . The obtained latent and sensible heat is transferred
to the consumer heating system. The dewpoint of the outlet air finally decreases. Then the
desorption process is switched to an open loop in which air with an inlet dewpoint of 0 *C is
heated to 130 °C and used to continue the desorption. This brings zeolite 13X to a final water
content of C = 0.09 kg water/ kg zeolite. The thermal COP of this mode of operation reaches
about 1.0.
Desorption with internal heat exchanger:
A second way to increase the dew point level during the desorption mode is by using an
internal heat exchanger submerged in the zeolite (isothermic desorption). The heat input of
desorption no longer depends on the air mass flow. The air flow through the storage carries
-------
1773
away the desorbed water vapor. The high dew point level of the exhaust air allows an efficient
recovery of heat of condensation at high temperature levels.
Experiments with isothermic desorption of zeolite 13X showed dewpoints up to 60 °C . In this
case, 60% of the heat of desorption can be recovered and used in the space heating system
50/30. A thermal COP of 1.1 was obtained with an storage energy density of 520MJ/t ,
removing 80% of the water which can be desorbed at a 130 °C . These values can still be
improved by further optimizing the isothermic desorption process.
REFERENCES
D. Jung (1983). Absorptive thermische Energiespeicherung. PhD thesis, Fakultaet fuer Physik,
Ludwig-Maximilians- Universitaet Muenchen.
S. Fischer, S. Hoist, W. Schoelkopf (1990). Thermochemische Speicherung von Niederiem-
peraturwaerme zur Hausheizung, Abschlussbericht 90. Technical Report, Fakultaet fuer
Physik, Ludwig-Maximilians- Universitaet Muenchen.
S. Hoist (1990). Thermochemische Speicherung mit festen Adsorbentien zur Raumheizung.
Master's thesis, Fakultaet fuer Physik, Ludwig-Maximilians- Universitaet Muenchen.
Sizmann, R. (1987). Thermal storage systems, present status and trends. Proceedings of the
ISES World Congress, Hamburg.
Wijsman, A.T.Th., R. Oosterhoven, and C. den Ouden (1979). Developement of a thermal
storage system based on the heat of water in hygroscopic materials. Proceedings of the
ISES World Congress, Atlanta.
-------
1774
EVACUATED DISH FLAT PLATE COLLECTOR
A.B. Schaap, W.B. Veltkamp, J. van Berkel
LEVEL Energy Technology, Eindhoven, The Netherlands
ABSTRACT
The evacuated dish flat plate collector is a new concept for conversion of solar irradiance in low and in-
termediate temperature heat. The evacuated dish collector consists of two spherical glass dishes sealed
together to form an airtight envelope. Stress and buckling calculations on the spherical shell are pre-
sented. Experiments demonstrated that the dish shaped envelope is able to withstand the atmospheric
pressure.
KEYWORDS
evacuated collector, flat plate collector, spherical dish, shell buckling, tension ring.
INTRODUCTION
The evacuated dish flat plate collector is a new concept for conversion of solar irradiance in low and in-
termediate temperature heat. The evacuated dish collector consists of two spherical glass dishes sealed
together to form an airtight envelope. In the central plane of the envelope a selective absorber is placed.
To withstand the atmospheric pressure, the dishes are supported by a tension ring at the edge. In this
way, a single collector can be constructed with an aperture of 2 m2 (diameter 1.6 m), while the glass
dishes can have a thickness of 4 to 5 mm (Fig. 1). The performance of the evacuated dish flat plate col-
lector is similar to that of the evacuated tube collector (Bloem, 1982). Because of the very low heat loss,
it is also possible to integrate a storage within the evacuated envelope (ICS).
CROSS SECTION
Glass dish
BACK VIEW
Selective absorber
Tension ring
Absorber suspension" ] \
Glass dish with
heat mirror
Serpentine tube
Fig. 1. Schematic drawing of the evacuated dish flat plate collector.
CONCEPT OF THE EVACUATED DISH COLLECTOR
The aperture of the evacuated tube collector is limited by the cylindrical shape of the envelope. Hence
a solar system is always constructed of more than one evacuated tube collector. These tubes have to be
connected to a header pipe and mounted on a frame. An evacuated collector with a single absorber area
large enough for a domestic hot water system would have no connecting parts, no header and no subse-
quent header losses. The cylindrical form is not well suited to envelop a big absorber area, because the
-------
1775
height (perpendicular to the absorber plane) equals the width. The height should be small enough, to
enable roof integration. The sphere is the ideal shape to withstand external uniform pressure. A com-
plete sphere is also not suitable to envelop a big absorber area (height = width). In order to acquire a
small height, two spherical (thin walled glass) dishes can be used (Fig. 1). The dishes have to be sup-
ported at the edge by a so-called tension ring. The tension ring supports the dishes in such a way that
only compressive stresses in the dish wall remain. Contrary to the evacuated tube collector the size of
the dish collector allows the use of low pressure heavy gas filling, easing the otherwise stringent vac-
uum quality demands needed to suppress the conductive heat losses.
STRESSES IN THE DISH WALL AND IN THE TENSION RING
The compressive stress in the dish wall can readily be calculated. Consider a dish with diameter D,
height H/2, radius R and wall thickness t (Fig. 2). The pressure on the projected area of a half sphere
with the corresponding radius R is supported by the area of the central cross section of the sphere wall.
D
M
Half sphere
Fig. 2. Atmospheric pressure on a half sphere
q _ Pat R2 _ Pat R
2 71 R t 2t
in which:
c = Compressive stress in the sphere wall
Pat = Atmospheric pressure (1 bar = 0.1 N/mm^)
R = Sphere radius
t = Wall thickness (t«R)
This holds for any cross section through the center of the sphere. So a is the compressive stress in the
dish wall in any plane perpendicular to the surface of the dish. Stresses in any plane parallel to the sur-
face of the dish are of the order of magnitude of the atmospheric pressure. These stresses are consider-
ably smaller than the stresses in planes perpendicular to the surface of the dish. The tension ring sup-
ports the dishes at the edge. The dishes have the same shape, so both dishes exercise the same uniform
load F on the tension ring (Fig. 3).
Spherical dishes
Tension ring
Fig. 3. Uniform load F acting upon the tension ring
(1)
[N/min^]
[N/mm^J
[mm]
[mml
-------
1776
The magnitude of this uniform load is: F = a t. F can be resolved into a uniform load in the plane of the
tension ring (Fcos). The uniform loads
perpendicular to the plane of the ring of both dishes are of equal magnitude but of opposite axial direc-
tion. There are no resulting loads in this direction. The uniform loads in the plane of the ring are of equal
magnitude and of equal radial direction. The resulting load in this direction equals 2Fcos. This result-
ing load causes tension in the ring (Fig. 4).
Tension ring
2Fcos
Fig. 4. Tension ar in the ring as a result of the force 2Fcos<))
acting upon the ring.
The tensile stress can be calculated from:
Fr = cr A = r 2 F cos $ (2)
in which:
Fr = Tensile force in the ring [N]
(jr = Tensile stress in a perpendicular cross section [N/mm2]
r = Ring radius [mm]
A = Cross section area of the ring [mm2]
The formulas 1 and 2 are the basic formulas describing stresses in the spherical dishes and tension ring.
For an evacuated dish collector with a diameter D = 1600 mm, a height H = 400 mm (R = 1700 mm) and
a dish wall thickness t = 4 mm, the compressive stress in the dish wall is cr = 17 N/mm2 and the force
Fr = 120,000 N. During evacuation of the spherical envelope the tension ring has to be prestressed fol-
lowing the decreasing pressure inside the envelope.
The basic formulas only describe the ideal situation; the atmospheric pressure equals 1 bar. Because of
fluctuations in atmospheric pressure, stress fluctuations occur. These stress fluctuations can be described
by the bending theory for thin walled shells (Kelkar, 1987). Application of this theory shows that the
stress deviations occur at the edges of the dishes. If a collector with D = 1600 mm, H = 400 mm and a
glass thickness t = 4 mm is supported by a spring steel ring with a cross section A = 200 mm2, no tensile
stresses are present if the deviation from atmospheric pressure stays below + 0.14 bar and above
- 0.18 bar. Tensile stresses stay below the tensile strength of glass (40 N/mm2) if the pressure deviation
stays below +0.42 bar and above -0.51 bar. These values are veiy much dependent upon the rigidity (AE;
E = modulus of elasticity) of the tension ring. The higher the rigidity of the tension ring,the higher the
deviation from atmospheric pressure can be, without exceeding the tensile strength of glass. Stress
deviations can also occur due to temperature differences. Suppose that both dishes have a different but
uniform temperature, we can calculate the allowable temperature difference with the bending theory.
No tensile stresses occur at a temperature difference below 64 °C,and the tensile strength of glass is not
exceeded below a temperature difference of 200 °C.
IMPLOSION (BUCKLING) OF THE SPHERICAL ENVELOPE
Calculations showed that the most important design parameter is not the compressive strength in the
glass dish, but the stability under pressure of the glass dish. If the stability limit is exceeded (the criti-
cal pressure of the envelope), the envelope might implode (buckle), with the risk of injuries by glass
splinters. The risk of implosion exists for spherical shells with a geometric parameter X > 4. The geo-
metric parameter X is defined as (Kaplan, 1974):
------- |