Utfttd States Mustriil Errwrowiwntol Bwwwneft tPA-iOO/i*79-03'lb
&wH»nm»ntit (Vetaefion Laboratory August 1971
%sm
-------
RESEARCH REPORTING SERIES
Research reports of the Office of Research and Development, U.S. Environmental
Protection Agency, have been grouped into nine series. These nine broad cate-
gories were established to facilitate further development and application of en-
vironmental technology. Elimination of traditional grouping was consciously
planned to foster technology transfer and a maximum interface in related fields.
The nine series are:
1. Environmental Health Effects Research
2. Environmental Protection Technology
3. Ecological Research
4. Environmental Monitoring
5. Socioeconomic Environmental Studies
6. Scientific and Technical Assessment Reports (STAR)
7. Interagency Energy-Environment Research and Development
8. "Special" Reports
9. Miscellaneous Reports
This report has been assigned to the MISCELLANEOUS REPORTS series. This
series is reserved for reports whose content does not fit into one of the other specific
series. Conference proceedings, annual reports, and bibliographies are examples
of miscellaneous reports.
EPA REVIEW NOTICE
This report has been reviewed by the U.S. Environmental Protection Agency, and
approved for publication. Approval does not signify that the contents necessarily
reflect the views and policy of the Agency, nor does mention of trade names or
commercial products constitute endorsement or recommendation for use.
This document is available to the public through the National Technical Information
Service, Springfield, Virginia 22161.
-------
EPA-600/9-79-031b
August 1979
Proceedings: Second Conference
on Waste Heat Management and Utilization
(December 1978, Miami Beach, FL)
Volume 2
S.S. Lee and Subrata Sengupta, Compilers
Mechanical Engineering Department
University of Miami
Coral Gables, Florida 33124
EPA Purchase Order DA 86256J
Program Element No. EHE624A
EPA Project Officer: Theodore G. Brna
Industrial Environmental Research Laboratory
Office of Energy, Minerals, and Industry
Research Triangle Park, NC 27711
Cosponsors: Department of Energy, Electric Power Research Institute, Environmental Protection
Agency, Florida Power and Light Company, Nuclear Regulatory Commission, and University of
Miami's School of Continuing Studies (In cooperation with American Society of Mechanical
Engineers' Miami Section)
Prepared for
U.S. ENVIRONMENTAL PROTECTION AGENCY
Office of Research and Development
Washington, DC 20460
-------
ORGANIZING COMMITTEE
Dr. .lohn Neal
Department of Energy
Dr. Theodore C,. Mrna
Environmental Protection Agency
Mr. Frank Swanberg
Nuclear Regulatory Commission
Dr. -John Maulbetsch
Electric Power Research Institute
Mr. Charles D. Henderson
lion"da POWPI- &. Light Company
Dr. Samuel S. Lee
Conference Chairman,
University of Miami
Dr. Subrata Serigupta
Conference Co-Chairman,
University of Miami
ADVISORY COMMITTEE
Dr.' C. C. Lpe ':i-
r.S. Environmental Protection Agency
Mr.. Charles H. Kaplan
U.S. Environmental Protection Agency
Dr. Mostafa A. Shirazi
IT.P. Environmental Protection Agency
Dr. Richard Dirks
National Science Foundation
Dr. Donald R. T. Harleman
Massachusetts Institute of Technology
Dr. Charles C. Coutant
Oak Ridge National Laboratory
Dr. G. s. Rodonhuis
Danish Hydraulic Institute, Denmark
Dr. H. Euchs
Consulting Engineers Inc., Switzerland
Dr. P. F. Chester
Central Electricity Research Laboratory, England
CONFERENCE SUPPORT
Arrangements:
•Tamps Poisant
Ruben Fuentes
The School of Continuing Studies
Special Assistant:
Sook Rhee
-------
ACKNOWLEDGEMENTS
The Conference Committee expresses its gratitude to
the Keynote Speaker, Dr. Eric H. Willis. It also greatly
appreciates the help of the Banquet Speaker, Dr. William
C. Peters.
This Second Conference on Waste Heat Management has
been shaped with help- from the Advisory Committee members
and the Session Chairmen. Their help is gratefully ack-
nowledged .
The numerous students and faculty who have helped as
Co-Chairmen of sessions and other organizational matters
were invaluable to the Conference Committee.
The sustained interest of sponsoring organizations
made this conference possible. The scientists and admini-
strators who have provided a leadership role in nurturing
this growing field of waste heat research deserve our sin-
cerest gratitude.
The participating scientists, engineers and admini-
strators have made this conference achieve the planned
objectives of technical interaction and definition of
future goals.
Conference Committee
Miami, December, 1978
iii
-------
FOREWORD
Tim first conference on Waste Heat Management and Utilization held in
Miami during May 9-12. 1977 was a success in terms of participation,
comprehensive technical representation and quality. A questionnaire
submitted to the sponsors and participants at the meeting indicated a
strong interest in an annual or biannual meeting. Tn responsp to this
the second comprehensive conference in the subject area is being held during
December 4-6, 1978. This wi 11 estabish a biannual frequency and allow
significant progress during meetings.
A perusal of the table of contents will indicate that causes, effects.
prediction, monitoring, utilization and abatement of thermal discharges are
represented. Utilization has become of prime importance owing to increased
awareness, that waste heat is a valuable resource. Sessions on Co-generation
and Recovery Systems have been added to reflect this emphasis.
This second conference has working sessions covering important topics
in the subject area. .This provides an interactive forum resulting in
relevant recommendations regarding researcli directions.
A well balanced Organizing Committee with an Advisory Board with
international composition has brought this conference to fruition. The
sponsoring organizations include governmental and private organizations
who are active in waste heat research and development.
Samuel S. Tee
Subrata Sengupta
IV
-------
CONTENTS
WASTE HEAT MANAGEMENT AND UTILIZATION CONFERENCE
OPENING SESSION
OPENING REMARKS
Samuel S. Lee, Conference Chairman, University of Miami
WELCOMING ADDRESS
Norman Einspruch, Dean of Engineering and Architecture,
University of Miami
KEYNOTE ADDRESS 1
Eric H. Willis, Deputy Assistant Secretary for Energy
Technology, Department of Energy, Washington, D.C.
PROGRAM REVIEW
Sub'rata Sengupta, Conference Co-Chairman, University of Miami
GENERAL SESSION
A WASTE HEAT UTILIZATION PROGRAM 13
J. Neal, Department of Energy, Washington, D.C.
W.F. Adolfson, Booz-Allen & Hamilton Inc,. Bethesda, MD
EPA PROGRAMS IN WASTE HEAT UTILIZATION 25
T. Brna, EPA, Research Triangle Park, NC
REVIEW OF EPRI PROGRAM 38
Q. Looney, J. Maulbetsch, Electric Power Research Institute,
Palo Alto, CA
THE ENERGY SHORTAGE AND INDUSTRIAL ENERGY CONSERVATION 39
E.H. Mergens, Shell Oil Company, Houston, TX
UTILIZATION I
USE OF SOIL WARMING AND WASTE WATER IRRIGATION FOR FOREST
BIOMASS PRODUCTION 66
D.R. DeWalle, W.E. Sopper, The Pennsylvania State University
POWER PLANT LAND AVAILABILITY CONSTRAINTS ON WASTE HEAT
UTILIZATION 76
M. Olszewski, H.R. Bigelow, Oak Ridge National Laboratory,
Oak Ridge, TN
v
-------
Page
COOLING PONDS AS RECREATIONAL FISHERIES - A READY MADE 86
RESOURCE
J.H. Hughes, Commonwealth Edison Company, Chicago, IL
HEAT RECOVERY AND UTILIZATION FOR GREEN BAY WASTE WATER 96
TREATMENT FACILITY
R.W. Lanz, University of Wisconsin, Green Bay, WI
MATHEMATICAL MODELING I
WHY FROUDE NUMBER REPLICATION DOES NOT NECESSARILY ENSURE 106
MODELING SIMILARITY
W.E. Frick, L.D. Winiarski, U.S. Environmental Protection
Agency, Corvallis, OR
A CALIBRATED AND VERIFIED THERMAL PLUME MODEL FOR SHALLOW 114
COASTAL SEAS AND EMBAYMENTS
S.L. Palmer, Florida Department of Environmental Regulation,
Tallahassee, FL
FARFIELD MODEL FOR WASTE HEAT DISCHARGE IN THE COASTAL ZONE 129
D.N. Brocard, J.T. Kirby, Jr., Alden Research Laboratory,
Worcester Polytechnic Institute, Holden, MA
THERMAL CHARACTERISTICS OF DEEP RESERVOIRS IN PUMPED STORAGE 139
PLANTS
J.J. Shin, N.S. Shashidhara, Envirosphere Company, New York,NY
ALGORITHMS FOR A MATHEMATICAL MODEL TO PREDICT ENVIRONMENTAL 150
EFFECTS FROM THERMAL DISCHARGES IN RIVERS AND IN COASTAL AND
OFFSHORE REGIONS
J. Hauser, Institut fur Physik, Germany
F. Tanzer, Universitat Giessen, Germany
EFFECT OF SALT UPON HOT-WATER DISPERSION IN WELL-MIXED 161
ESTUARIES - PART 2 - LATERAL DISPERSION
R. Smith, University of Cambridge, United Kingdom
MATHEMATICAL MODELING II
COST-EFFECTIVE MATHEMATICAL MODELING FOR THE ASSESSMENT OF 179
HYDRODYNAMIC AND THERMAL IMPACT OF POWER PLANT OPERATIONS
ON CONTROLLED-FLOW RESERVOIRS
A.H. Eraslan, K.H. Kim, University of Tennessee,
Knoxville, TN
HEAT LOAD IMPACTS ON DISSOLVED OXYGEN: A CASE STUDY IN 187
STREAM MODELING
A.K. Deb, D.F. Lakatos, Roy F. Weston, Inc.,
West Chester, PA
vi
-------
Page
A STOCHASTIC METHOD FOR PREDICTING THE DISPERSION OF THERMAL 199
EFFLUENTS IN THE ENVIRONMENT
A.J. Witten, Oak Ridge National Laboratory, Oak Ridge, TN
J.E. Molyneux, University of Rochester, Rochester, NY
A TWO-DIMENSIONAL NUMERICAL MODEL FOR SHALLOW COOLING PONDS 214
S. Chieh, A. Verma, Envirosphere Company, New York, NY
UTILIZATION II
WASTE HEAT FOR ROOT-ZONE HEATING - A PHYSICAL STUDY OF HEAT 225
AND MOISTURE TRANSFER
D. Elwell, W. Roller, A. Ahmed, Ohio Agricultural Research
and Development Center, Wooster, OH
BENEFICIAL USE OF REJECTED HEAT IN MUNICIPAL WATER SUPPLIES 236
R.W. Porter, R.A. Wynn, Jr., Illinois Institute of Technology
Chicago, IL
SUPER GREENHOUSE PROJECT UTILIZING WASTE HEAT FROM ASTORIA 6 246
THERMAL POWER PLANT
R.G. Reines, Cornell University, Ithaca, NY
EXPERIENCE WITH THE NEW MERCER PROOF-tDF-CONCEPT WASTE HEAT 247
AQUACULTURE FACILITY
B.L. Godfriaux, Public Service Electric and Gas Company,
Newark, NJ. R.R.Shafer, Buchart-Horn: Consulting Engineers,
York, PA. A.F. Eble, M.C. Evans, T. Passanza, C. Wainwright,
H.L. Swindell, Trenton State College, Trenton, NJ.
UTILIZATION III
WASTE HEAT RECOVERY IN THE FOOD PROCESSING INDUSTRY ,-• " 266y
W.L. Lundberg, J.A. Christenson, Westinghouse Electric *v „.....--"'
Corporation, Pittsburgh, PA. F. Wojnar, H.J.Heinz Company,
Pittsburgh, PA.
GENERATION OF CHILLED WATER FROM CHEMICAL PROCESS WASTE HEAT 277
J. Entwistle, Fiber Industries, Inc., Charlotte, NC
THE SHERCO GREENHOUSE PROJECT: FROM DEMONSTRATION TO 286
COMMERCIAL USE OF CONDENSER WASTE HEAT
G.C. Ashley, J.S. Hietala, R.V. Stansfield, Northern States
Power Company, Minneapolis, MN
ANALYSIS OF ECONOMIC AND BIOLOGICAL FACTORS OF WASTE HEAT 296
AQUACULTURE
J.S. Suffern, M. Olszewski, Oak Ridge National Laboratory,
Oak Ridge, TN
vii
-------
Page
ECOLOGICAL EFFECTS I
A QUALITATIVE/QUANTITATIVE PROCEDURE FOR ASSESSING THE 319
BIOLOGICAL EFFECTS OF WASTE HEAT ON ECONOMICALLY IMPORTANT
POPULATIONS
J.M. Thomas, Battelle Pacific Northwest Laboratories,
Richland, WA
A REVIEW OF STATISTICAL ANALYSIS METHODS FOR BENTHIC DATA 329
FROM MONITORING PROGRAMS AT NUCLEAR POWER PLANTS
D.H. McKenzie, Battelle Pacific Northwest Laboratories
Richland, WA
FURTHER STUDIES IN SYSTEMS ANALYSIS OF COOLING LAKES: ^4
HYDRODYNAMICS AND ENTRAINMENT
K.D. Robinson, R.J. Schafish, R.W. Beck and Associates,
Denver, CO. G. Comougis, New England Research, Inc.,
Worcester, MA.
SYNTHESIS AND ANALYSES OF EXISTING COOLING IMPOUNDMENT 353
INFORMATION ON FISH POPULATIONS
K.L. Gore, D.H. McKenzie, Battelle Pacific Northwest
Laboratories, Richland, WA
COOLING TOWER PLUMES
A SIMPLE METHOD FOR PREDICTING PLUME BEHAVIOR FROM MULTIPLE 357
SOURCES
L.D. Winiarski, W.E. Frick, U.S. Environmental Protection
Agency, Corvallis, OR
MODELING NEAR-FIELD BEHAVIOR OF PLUMES FROM MECHANICAL DRAFT 377
COOLING TOWERS
T.L. Crawford, Tennessee Valley Authority, Muscle Shoals, AL
P.R. Slawson, University of Waterloo, Ontario, Canada
MECHANICAL-DRAFT COOLING TOWER PLUME BEHAVIOR AT THE GASTON 38g
STEAM PLANT
P.R. Slawson, University of Waterloo, Ontario, Canada
CRITICAL REVIEW OF THIRTEEN MODELS FOR PLUME DISPERSION
FROM NATURAL DRAFT COOLING TOWERS
R.A. Carhart, University of Illinois, Chicago, IL
A.J. Policastro, Argonne National Laboratory, Argonne, IL
W.E. Dunn, University of Illinois, Urbana, IL
EVALUATION OF METHODS FOR PREDICTING PLUME RISE FROM
MECHANICAL-DRAFT COOLING TOWERS
W.E. Dunn, P. Gavin, University of Illinois, Urbana, IL
G.K. Cooper, Mississippi State University, Mississippi
viii
-------
Page
ECOLOGICAL EFFECTS II
ENVIRONMENTAL COST OF POWER PLANT WASTE HEAT AND 461
CHEMICAL DISCHARGE IN TROPICAL MARINE WATERS
J.M. Lopez, Center for Energy and Environment Research
Mayaguez, Puerto Rico
THEORY AND APPLICATION IN A BIOLOGICAL ASPECT 468
T. Kuroki, Tokyo University of Fisheries, Tokyo, Japan
OCCURRENCE OF HIGHLY PATHOGENIC AMOEBAE IN THERMAL 479
DISCHARGES
J.F. De Jonckheere, Laboratorium voor Hygiene,
Katholieke Universiteit Leuven, Belgium
RELATION BETWEEN ZOOPLANKTON MIGRATION AND ENTRAINMENT 490
IN A SOUTH CAROLINA COOLING RESERVOIR
P.L. Hudson, S.J. Nichols, U.S. Fish and Wildlife Service
Southeast Reservoir Investigations, Clemson, SC
EFFECTS OF A HOT WATER EFFLUENT ON POPULATIONS OF 505
MARINE BORING CLAMS IN BARNEGAT BAY, NJ
K.E. Hoagland, Lehigh University, Bethlehem, PA
R.D. Turner, Harvard University, Cambridge, MA
COOLING TOWERS I
COLD INFLOW AND ITS IMPLICATIONS FOR DRY TOWER DESIGN 516
F.K. Moore, Cornell University, Ithaca, NY
AN IMPROVED METHOD FOR EVAPORATIVE, CROSS-FLOW COOLING 532
TOWER PERFORMANCE ANALYSIS
K.L. Baker, T.E. Eaton, University of Kentucky, Lexington, KY
THE IMPACT OF RECIRCULATION ON THE SITING, DESIGN, 535
SPECIFICATION, AND TESTING OF MECHANICAL DRAFT COOLING
TOWERS
K.R. Wilber, Environmental Systems Corporation
A. Johnson, Pacific Gas & Electric Co.
E. Champion, Consultant
AN INVESTIGATION INTO THE MINERAL CONCENTRATION OF 547
INDIVIDUAL DRIFT DROPLETS FROM A SALTWATER COOLING TOWER
R.O. Webb, Environmental Systems Corporation, Knoxville, TN.
R.S. Nietubicz, State of Maryland, Department of Natural
Resources. J.W. Nelson, Florida State University, Tallahassee, FL
COGENERATION
COGENERATION TECHNOLOGY AND OUR TRANSITION FROM 548
CONVENTIONAL FUELS
J.W. Neal, Department of Energy, Washington, DC
ix
-------
Page
COGENERATION: THE POTENTIAL AND THE REALITY IN A 553
MIDWESTERN UTILITY SERVICE AREA
D.M. Stipanuk, Cornell University, Ithaca, NY
W.J. Hellen, Wisconsin Electric Power, Milwaukee, WI
ALTERNATIVE APPROACHES IN INDUSTRIAL COGENERATION SYSTEMS 572
J.C. Solt, Solar Turbines International, San Diego, CA
THE ENVIRONMENT FOR COGENERATION IN THE UNITED STATES 532
F.E. Dul, Envirosphere Company, New York, NY
FUEL COST ALLOCATION FOR THE STEAM IN A COGENERATION 595
PLANT
K.W. Li, and P.P. Yang, North Dakota State University,
Fargo, ND
COOLING SYSTEMS
APPLICATIONS OF MATHEMATICAL.SPRAY COOLING MODEL 619
H.A. Frediani, Jr., Envirosphere Company, New York, NY
THE DEVELOPMENT OF ORIENTED SPRAY COOLING SYSTEMS 638
D.A. Fender, Ecolaire Condenser, Inc. Bethlehem, PA
T.N. Chen, Ingersoll-Rand Research, Inc., Princeton, NJ
ONCE-THROUGH COOLING POTENTIAL OF THE MISSOURI RIVER IN 651
THE STATE OF MISSOURI
A.R. Giaquinta, The University of Iowa, Iowa City, IA
T.C. Keng, Jenkins-Fleming, Inc., St. Louis, MO
A MODEL FOR PREDICTION OF EVAPORATIVE HEAT FLUX IN LARGE 663
BODIES OF WATER
A.M. Mitry, Duke Power Compnay, Charlotte, NC
B.L. Sill, Clemson University, Clemson, NC
WORKING SESSIONS - WORKSHOPS
(1) MANGEMENT AND UTILIZATION 677
(2) ENVIRONMENTAL EFFECTS 6gl
(3) MATHEMATICAL MODELING 6g2
(4) HEAT TRANSFER PROBLEMS IN WASTE HEAT MANAGEMENT AND 534
UTILIZATION
COOLING TOWERS II
.riE CHALK POINT DYE TRACER STUDY: VALIDATION OF MODELS AND 686
ANALYSIS OF FIELD DATA
A.J. Poliscastro, M. Breig, J. Zieharth, Argonne National
Laboratory, Argonne, IL
W.E. Dunn, University of Illinois, Urbana, IL
x
-------
Page
COOLING TOWERS AND THE LICENSING OF NUCLEAR POWER PLANTS 720
J.E. Carson, Argonne National Laboratory, Argonne, IL
A DESIGN METHOD FOR DRY COOLING TOWERS 732
G.K.M. Vangala, T.E. Eaton, University of Kentucky,
Lexington, KY
EVAPORATIVE HEAT REMOVAL IN WET COOLING TOWERS 742
T.E. Eaton, K.L. Baker, University of Kentucky,
Lexington, KY
COMPARATIVE COST STUDY OF VARIOUS WET/DRY COOLING CONCEPTS 772
THAT USE AMMONIA AS THE INTERMEDIATE HEAT EXCHANGE FLUID
B.M. Johnson, R.D. Tokarz, D.J. Braun, R.T. Allemann,
Battelle Pacific Northwest Laboratory, Richland, WA
UTILIZATION IV
ENVIRONMENTAL ASPECTS OF EFFECTIVE ENERGY UTILIZATION 805
IN INDUSTRY
R.E. Mournighan, U.S. EPA, Cincinnati, OH
W.G. Heim, EEA, Inc., Arlington, VA
WASTE HEAT RECOVERY POTENTIAL FOR ENVIRONMENTAL BENEFIT 817
IN SELECTED INDUSTRIES
S.R. Latour, DOS Engineers, Inc., Fort Lauderdale, FL
C.C. Lee, EPA, Cincinnati, OH
WASTE HEAT UTILIZATION AND THE ENVIRONMENT 830
M.E. Gunn, Jr., Department of Energy, Washington, DC
THERMAL STORAGE FOR INDUSTRIAL PROCESS AND REJECT HEAT 855
R.A. Duscha, W.J. Masica, NASA Lewis Research Center,
Cleveland, OH
PERFORMANCE AND ECONOMICS OF STEAM POWER SYSTEMS 866
UTILIZING WASTE HEAT
J. Davis, Thermo Electron Corporation, Waltham, MA
COOLING LAKES
A ONE-DIMENSIONAL VARIABLE CROSS-SECTION MODEL FOR THE 878
SEASONAL THERMOCLINE
S. Sengupta, S.S.Lee, E. Nwadike, University of Miami,
Coral Gables, FL
HYDROTHERMAL STRUCTURE OF COOLING IMPOUNDMENTS 908
G.H. Jirka, Cornell University, Ithaca, NY
HYDROTHERMAL PERFORMANCE OF SHALLOW COOLING PONDS 909
E.E. Adams, G.H. Jirka, A. Koussis, D.R.F. Harleman,
M. Watanabe, M.I.T., Cambridge, MA
xi
-------
TRANSIENT SIMULATION OF COOLING LAKE PERFORMANCE UNDER
HEAT LOADING FROM THE NORTH ANNA POWER STATION
D.R.F. Harleman, G.H. Jirka, D.N. Brocard, K.H. Octavio,
M. Watanabe, M.I.T., Cambridge, MA
RECOVERY SYSTEMS
COMPARISON OF THE SURFACE AREA REQUIREMENTS OF A SURFACE 931
TYPE CONDENSER FOR A PURE STEAM CYCLE SYSTEM, A COMBINED
CYCLE SYSTEM AND A DUAL FLUID CYCLE SYSTEM
M.H. Waters, International Power Technology
E.R.G. Eckert, University of Minnesota
UTILIZATION OF TRANSFORMER WASTE HEAT 960
D.P. Hartmann, Department of Energy, Portland, OR
H. Hopkinson, Carrier Corporation, Syracuse, NY
THE APPLICATION OF PRESSURE STAGED HEAT EXCHANGERS TO 980
THE GENERATION OF STEAM IN WASTE HEAT RECOVERY SYSTEMS
M.H. Waters, D.Y. Cheng, International Power Technology
HEAT RECOVERY FROM WASTE FUEL 1000
Y.H. Kiang, Trane Thermal Company, Conshohocke, PA
AQUATIC THERMAL DISCHARGES I
SURFACE SKIN-TEMPERATURE GRADIENTS IN COOLING LAKES 1011
S.S. Lee, S. Sengupta, C.R. Lee, University of Miami,
Coral Gables, FL
FOUR THERMAL PLUME MONITORING TECHNIQUES: A COMPARATIVE 1027
ASSESSMENT
R.S. Grove, Southern California Edison Company, Rosemead, CA
R.W. Pitman, J.E. Robertson, Brown and Caldwell, Pasadena, CA
EXPERIMENTAL RESULTS OF DESTRATIFICATION BY BUOYANT PLUMES 1028
D.S. Graham, University of Florida, Gainesville, FL
THREE-DIMENSIONAL FIELD SURVEYS OF THERMAL PLUMES FROM 1047
BACKWASHING OPERATIONS AT A COASTAL POWER PLANT SITE IN
MASSACHUSETTS
A.D. Hartwell, Normandeau Associates, Inc., Bedford, NH
F.J. Mogolesko, Boston Edison Company
SHORT-TERM DYE DIFFUSION STUDIES IN NEARSHORE WATERS 1057
D.E. Frye, EG&G, Environmental Consultants, Waltham, MA
S.M. Zivi, Argonne National Laboratory, Argonne, IL
EFFECTS OF BOTTOM SLOPE, FROUDE NUMBER, AND REYNOLDS 1Q69
NUMBER VARIATION ON VIRTUAL ORIGINS OF SURFACE JETS:
A NUMERICAL INVESTIGATION
J. Venkata, S. Sengupta, S.S.Lee, University of Miami
Coral Gables, FL xii
-------
Page
ATMOSPHERIC EFFECTS
METEOROLOGICAL EFFECTS FROM LARGE COOLING LAKES 1095
F.A. Huff, J.L. Vogel, Illinois State Water Survey, IL
COMPUTER SIMULATION OF MESO-SCALE METEOROLOGICAL EFFECTS 1104
OF ALTERNATIVE WASTE-HEAT DISPOSAL METHODS
J.P. Pandolfo, C.A. Jacobs, The Center for the Environment
and Man, Inc., Hartford, CT
A NUMERICAL SIMULATION OF WASTE HEAT EFFECTS ON 1114
SEVERE STORMS
H.D. Orville, P.A. Eckhoff, South Dakota School of Mines
and Technology, Rapid City, SD
ON THE PREDICTION OF LOCAL EFFECTS OF PROPOSED COOLING 1124
PONDS
B.B. Hicks, Argonne National Laboratory, Argonne, IL
AQUATIC THERMAL DISCHARGES II
MEASUREMENT AND EVALUATION OF THERMAL EFFECTS IN THE INTER- 1131
MIXING ZONE AT LOW POWER NUCLEAR STATION OUTFALL
P.R. Kamath, R.P. Gurg, I.S. Bhat, P.V. Vyas, Environmental
Studies Section, .Bhabha Atomic Research Centre, Bombay, India
RIVER THERMAL STANDARDS EFFECTS ON COOLING-RELATED POWER 1146
PRODUCTION COSTS
T.E. Croley II, A.R. Giaquinta, M.P. Cherian, R.A. Woodhouse,
The University of Iowa, Iowa City, IA
THERMAL PLUME MAPPING 1160
J.R. Jackson, A.P. Verma, Envirosphere Company, New York, NY
THERMAL SURVEYS NEW HAVEN HARBOR - SUMMER AND FALL, 1976 1167
W. Owen, J.D. Monk, Normandeau Associates, Nashua, NH
BEHAVIOR OF THE THERMAL SKIN OF COOLING POND WATERS 1191
SUBJECTED TO MODERATE WIND SPEEDS
M.L. Wesely, Argonne National Laboratory, Argonne, IL
OPEN SESSION
ALTERNATE ENERGY CONSERVATION APPLICATIONS FOR INDUSTRY 1201
L.J. Schmerzler
MINERAL CYCLING MODEL OF THE THALASSIA COMMUNITY AS 1202
AFFECTED BY THERMAL EFFLUENTS
P.B. Schroeder, A. Thorhaug, Florida International University
Miami, FL
xiii
-------
SYNERGISTIC EFFECTS OF SUBSTANCES EMITTED FROM POWER
PLANTS ON SUBTROPICAL AND TROPICAL POPULATIONS OF THE
SEAGRASS THALASSIA TESTUDINUM: TEMPERATURE, SALINITY AND
HEAVY METALS
A. Thorhaug, P.B. Schroeder, Florida International University,
Miami, FL
WASTE HEAT MANAGEMENT AND UTILIZATION: SOME REGULATORY 1240
CONSTRAINTS
W.A. Anderson II, P.O. Box 1535, Richmond, VA
xiv
-------
APPLICATIONS OF MATHEMATICAL SPRAY COOLING MODEL
H A Frediani Jr, Senior Engineer
Envirosphere Company
A Division of Ebasco Services Incorporated
New York, New York USA
ABSTRACT
A mathematical model to analyze the performance of large-scale spray
cooling systems [l] , has been applied to several different types of spray
systems. This model is unique in that the basic heat and mass transfer
mechanisms are modelled accurately over a wide range of parameter values
that cooling systems encounter.
The model was first used to verify which vendors' systems would be
expected to meet the design conditions and then to predict which vendors'
performance curves were accurate. The model showed some sensitivity to
wind direction, and to dry bulb temperature, which the manufacturers had
assumed negligible.
Another application of this model has been to develop an optimum
(economical) configurationof a spray system that met the design operating
condition. To complete the economic optimization, average operating cold
water temperatures and required spray motor horsepower were calculated
and used to predict capability penalties.
A third application of this model was to design a fixed spray pond and
predict its performance under the most severe operating conditions.
Parameters optimized included nozzle separation, nozzle flow rate, and
spray height. Performance was predicted, including evaporation rate and
the effects of increasing solids concentrations during extended periods
of operation without makeup.
Qualitative conclusions have also been drawn from the results of these
applications concerning fogging and drift associated with closed loop
spray systems. Further development of this model could lead to quantita-
tive predictions in this regard.
INTRODUCTION
Spray cooling systems have been utilized to dissipate heat rejected by
electric generating stations. The cooling water includes both condensing
water for the main turbine steam and cooling water for auxiliary and/or
emergency heat exchangers. The spray systems include arrays of either
fixed or floating nozzles. The traditional method of designing a spray
619
-------
system has been to Interpolate or extrapolate predicted performance from
previously recorded data. Recently, a mathematical model was developed to
predict performance from the physical characteristics of the given spray
system and the basic principles of heat and mass transfer. The model has
since been utilized in various applications and some of the results are
described herein.
MODEL DESCRIPTION
The model has been described in detail in the literature. Briefly, the
continuity and energy equations were developed for a cellular model repre-
senting a single spray in a system of sprays. The equations were solved
using a finite difference solution along a drop trajectory, for both water
and air parameters. The results of the cellular analysis are incorporated
into a system model in which the interaction between sprays for both the
water and air is considered.
The model incorporates the following features:
1. A finite mass flow rate of air, as well as of water is calculated.
2. The amount of heat transferred from the water is added to the air.
3. The amount of mass transferred from the water is added to the air.
4. The air temperature, enthalpy, moisture content and density are
calculated to reflect the heat and mass transferred to the air.
5. Physical parameters such as cooling water salinity, spray height,
pattern, and droplet sizes, air dry bulb, and wind direction are
specifically entered as data and incorporated in the basic equations.
APPLICATIONS
Proposal Evaluation
The first application of the model was to independently check the credi-
bility of various proposed spray systems being considered for a proposed
coal-fired plant. The design was a closed cycle, salt water cooling
system. At that time, no successful closed cycle systems had been
documented. Each manufacturer predicted system performance using
empirical spray performance correlations, a method that is basically an
extension of the "number-of-transfer-units" (NTU) concept, as applied to
cooling towers [2]. An independent check was deemed necessary because of
the following reasons:
1. Closed cycle systems typically operate at higher water temperatures
than open cycle systems. Thus, the operating water temperatures of
the proposed system would be outside the range of data available
to any manufacturer.
2. The basic premise of the NTU concept that the amount of cooling is
solely dependent on the wet bulb and wind speed is an approximation.
620
-------
In actuality, the percentage of heat transfer in spray cooling which
is sensible (i.e. non-evaporative) will approximate twenty percent
when the air dry bulb temperature is significantly below the cooled
wa te r tempera ture.
3. The increase in enthalpy, through increased temperature and moisture
content, of the air as it flows past a spray was not predicted
rationally. Thus, for a large system of sprays, most of which are
downwind from other sprays, the major ambient input to the perfor-
mance correlation was estimated without benefit of either data or
theoretical considerations.
Each manufacturer proposed a U-shaped spray canal containing an array of
floating sprays (Figure 1). In this configuration, the hottest water is
sprayed immediately upwind of the coolest water. Thus the air which has
undergone the largest enthalpy increase will then pass the sprays with the
lowest incoming water temperature. This is where the greatest potential
exists for the incoming air dry bulb temperature to exceed the incoming
water temperature for a particular spray. Under this condition, the
sensible heat transfer is actually reversed and total cooling reduced.
Under extreme conditions of very large systems, the incoming air wet bulb
temperature can approach the cold water temperature closely enough that
all heat transfer is stopped.
Each manufacturer had submitted performance curves plotting system cold
water temperatures versus ambient wet bulb temperatures for several
different wind speeds. The design condition was at an 81 degree wet bulb
and a 5 mph wind speed, perpendicular to the canal axis. Runs were made
at the design conditions for each proposed system. The results indicated
that one manufacturer's system was slightly conservative (i.e. the desired
cold water temperature would be achieved before the last pass of sprays).
A second proposed system was estimated to be approximately 30 percent
deficient. At this point, the first manufacturer's system was selected
and examined further.
The next step was to synthesize the entire performance curve for the
selected system. At each given point, the ambient wet bulb and the
desired cooling range were known, but the equilibrium hot and cold water
temperatures were unknown. Using the model to determine the latter, for
a U-shaped canal, would have been a trial and error process. A hot water
temperature would be assumed, the model run, and the cooling range
obtained. Such a technique would have required a great deal of computer
time and an alternative method was derived.
It was hypothesized that, as a heat dissipation system, the straight canal
shown in Figure 2 would perform identically with the U-shaped canal shown
in Figure 1. There are the same number of sprays, in the same locations.
There are the same water flow rate in and air flow rate across. This
hypothesis was tested, using the model, at various wet bulbs, hot water
temperatures, and spray configurations. Close agreement was predicted,
as can be seen in the typical case presented as Figure 3. The difference
621
-------
In predicted cold water temperature, over a cooling range of 12.1 degreesF,
was only .26 degrees, approximately two percent.
Utilizing a straight canal, only one computer run is required for each
data point. The model is run for a longer canal than the one desired, for
an initial hot water temperature known to exceed the desired result. An
example run is illustrated in Table 1, in which the design number of spray
passes is 19 and the total number of passes run is 30. From the Table, a
plot of cooling range vs cold water temperature can be constructed, from
which the cold water temperature corresponding to the desired cooling
range can be estimated. Such a plot for this example is shown in Figure 4.
In this example, the desired range of 16.6 degrees F, with a corresponding
cold water temperature of 64.3 degrees F, represented a power plant
operating at 80 percent of load capacity. The same plot can be used in
synthesizing performance curves at other desired plant load capacities by
simply varying the cooling range.
The performance curves were synthesized from the model runs for two
different wind directions, parallel and perpendicular to the spray canal
axis. The manufacturer had not specified any relative humidity for its
performance curve. Comparative runs were made, at a given wet bulb, for
relative humidities of 20%, 60%, and 100%. It was found that, as relative
humidity was increased (i.e. air dry bulb temperature was decreased), the
equilibrium cold water temperature decreased. This improvement in heat
transfer is attributable to improved sensible heat transfer. For compari-
son, 60 percent relative humidity was used.
The curves for the parallel wind case were virtually identical over the
full range of wet bulb examined. In the perpendicular wind case, the
model predicted better performance. The improvement ranged from about
4 degrees F cooler equilibrium temperature at a 40 degree wet bulb, to
about 2 degrees F at an 80 degree wet bulb, for 100 percent load. Since
the manufacturer had guaranteed its performance curves for any wind
direction, it was concluded that the model had verified those performance
curves.
Alternative Cooling System Studies
The model has been utilized several times to provide input to aIternative
cooling system studies. This input is normally generated in two stages.
The first stage involves sizing a system to meet a given design point, in
order to estimate the system installation cost. The second atage is to
predict the system operation in order to estimate annual operating costs.
The total cost can then be estimated and compared to other types of
alternatives.
For a typical new power plant, the floating type of spray device has a
decided economic advantage over the fixed type, because of the large size
of such a system. There are two opposing economic factors in optimizing a
floating spray cooling system. Minimizing the number of spray devices
requires minimizing the number of sprays arrayed across the canal.
622
-------
However, this would maximize the length of the canal, and thus the costs
for canal excavation, diking, and auxiliaries such as wiring for the spray
motors. An example of these effects is given in Table 2, in which two
systems with identical performance at a design point are described. Case 1
has approximately 74 percent of the number of sprays required in Case 2.
However, the Cese 1 canal is about twice as long. The economic optimum
depends both on the spray device cost and the canal construction cost.
Results from nurerous model runs indicate that a width of 10 floating
sprays (or 5 on each side of a U-shaped canal) is about the practical
limit. The expense of additional sprays across the canal cannot be
justified by the small increase in performance.
Once the particular system is sized, operating cold water temperatures are
predicted. The gross output power from a generating station, at a given
load factor, is a direct function of the condenser inlet temperature. A
representative curve of this function is shown as gross power output in
Figure 5. As the cold water temperature is increased, gross power out is
decreased. For a particular spray system, at a given set of ambient
meteorological conditions, the cold water temperature is a function of the
number of sprays operating. By operating less than the full number of
sprays, the power consumed by the spray pump motors can be reduced. Table
3 summarizes this effect for a spray system designed for the turbine
generator of Figure 5. The net power output (gross power putput minus
spray motor power consumption) is plotted on Figure 5. The optimum
operating point for this system, at this particular meteorological condi-
tion, is with 79 percent of the sprays operating. This optimization
increases plant output approximately 700 kilowatts over running 100 percent
of the spray system.
In the preceding manner, monthly or seasonal average net power generation
is predicted for the given spray system. The capitalized cost of the
differential power outputs from each alternative are added to the estimated
installation costs to obtain the total costs for comparisons.
Fixed Spray Applications
The same model has also been utilized to evaluate fixed spray ponds.
These traditionally employ smaller nozzles and are used for smaller flow
rates. One example is the design of a reactor coolant spray cooling
system for a nuclear power plant. This system was to function under both
normal and emergency shutdown conditions. Thus two independent sets of
design criteria had to be met by the same design.
Fixed spray design requires pumping the heated water through a distribu-
tion system of pipes to the spray nozzles. The total system flow is
sprayed and collected. Should the cooling be insufficient, a second stage
of sprays can respray the collected water. Manufacturers typically recom-
mend one nozzle pressure and flow rate for a given nozzle. To size a
system, the total flow rate is divided by the nozzle flow rate giving the
number of nozzles required. These nozzles are then arranged in a rect-
angular array. For the example mentioned above, this process indicated
623
-------
that two stages of sprays would be required to meet the desired design
conditions.
Utilizing the model, an alternative design was considered. By increasing
tue pressure to the nozzles, the flow rate and spray height per nozzle can
be increased. This requires fewer nozzles spaced further apart. However,
the spray performance is improved for two reasons. Firstly, the water
drops experience a higher velocity and travel time through the air.
Secondly, the ratio of water to air mass flow is decreased. An example
for a commonly used nozzle (approximate orifice diameter of 1.5 inches) is
given in Table 4. In this example, a wind speed of 10 feet per second and
a total cooling water flow rate of 30,000 gpm were used. By increasing
the nozzle pressure from 7 to 15 psi, the spray height was increased by
40 percent, and the spray pattern diameter was increased by 23 percent.
This resulted in sufficient improvement in cooling range so that one stage
of sprays were adequate when two stages were required for the conventional
layout.
Under shutdown conditions, this spray pond must operate without makeup for
an extended period of time. This results in an increasing dissolved solids
content and a decreasing total volume of cooling water in the system, both
because of system evaporation without replacement water. The increased
solids content decreases the water's ability to carry heat. This causes
the cooling range and hot water temperatures to increase, without signifi-
cant change to the cold water temperatures. The reduced water volume in
the system causes lowering of the water level in the spray pond collection
area, from which the circulating pumps draw suction. This increases the
static head on the pumps, resulting in a reduced pumping rate and nozzle
pressure. This in turn increases both the hot and cold water temperatures
of the operating spray pond. Finally, the heat load of the system varies
with time during this mode of operation.
The model is run, at the design meteorological conditions and the initial
heat load to predict the evaporation rate. The change in solids content
and water volume is calculated from this evaporation rate. A new pumping
rate and cooling range are selected to provide the model input for the
second heat load. This process is repeated for all the heat rates until
plots of hot and cold water temperature, and solids content, versus time,
are produced. The operating temperatures are then checked to insure
satisfactory equipment operation. The solids content is used to estimate
scaling tendencies. If necessary, spray system design is altered and the
process repeated.
One item of interest in Table 4 is the predicted ratio of water to air
mass flow in each system. All other spray models assume this quantity to
be zero, i.e. that the air supply is infinite. Thus they are unable to
quantitatively estimate performance for any spray system without a vast
array of operating data. They are also unable to predict differential
performance resulting from modification of a spray system. It should be
noted that the ratios in Table 4 are for a fixed spray nozzle and that,
for the floating nozzles available, higher ratios are typical.
624
-------
FUTURE APPLICATIONS
Two of the environmental impacts associated with evaporative cooling
systems are fogging and drift. Examination of the psychroraetric conditions
the model predicts for the air leaving a spray canal leads to certain
qualitative conclusions. With respect to fogging, it has previously been
assumed that because a spray system discharges air over a relatively large
area, in comparison with a comparable cooling tower, that the potential
for fogging is much less for the spray system. Model predictions indicate
that the air leaving spray canals is not as close to saturation as that
from cooling towers. Thus, the previous assumption would appear to be
correct. Since the air flow rate, temperature, and moisture content are
all predicted, it should not be too difficult to construct a quantitative
fogging prediction model in the future.
Intuitive assumptions with drift have not been so accurate. As recently
as 1974, the lack of experience with closed-loop, salt water spray systems
precluded the availability of any useful drift data. Impact analyses
typically were based on limited data from a few sprays operating under
insignificant heat load. It was predicted that all drift, from one
proposed installation, would deposit within 600 feet of the spray canal [2j.
In a large spray system the heat transfer causes an increase in moisture
content and temperature in the air passing through the sprays. Thus the
air leaving the system has an upward velocity component due to its relative
buoyancy with respect to the ambient air. The resultant path of water
droplets entrained in the air as drift is thus a parabolic arch (Figure 6).
The path of a drift particle from a spray without heat load is a straight
line (Figure 6). Thus the mechanism is different and predictions of one
based on the other are invalid. The distance of 3400 feet for the heat
load case in Figure 6 was reported for an actual system of the same type
sprays which, without heat load, drifted within 600 feet [3].
As in the case of fogging, it should not be too difficult to predict, from
the model output, the magnitude and direction of the air velocity leaving
a spray canal. From this, a model could be constructed to predict drift
deposition. It should be kept in mind that the system which attained
drift deposition at 3400 feet is a relatively small system, and that
greater ranges could be attained in the future.
CONCLUSIONS
A mathematical spray cooling model has been used for a variety of applica-
tions, some of which are described herein. It was found that the model
gives reasonable quantitative performance estimates over a wide range of
configurations and operating regimes for both fixed and floating spray
canals.
625
-------
REFERENCES
1. "Mathematical Model for Spray Cooling Systems", H A Frediani, Jr, and
N Smith, Trans. ASME Journal of Engineering for Power, April 1977,
pp 279-283.
2. Draft Environmental Statement for Surry Power Station, Units 3 & 4,
USAEC, February, 1974, pp 3-11 through 3-15.
3. "Measured and Predicted Salt Deposition Rates, Closed Cycle Water
Cooling Duty", PSM-SD-6A, Ceramic Cooling Tower Company, 1977.
626
-------
TABLE 1
PERFORMANCE CURVE GENERATION
Input Data
Desired Canal Length - 19 Passes
Wind - Perpendicular to Canal Axis at 19 FPS
C W Flow - 711348 GPM
Flow Per Spray - 10000 GPM
Inlet Water Temperature - 89.61 Deg F
Ambient Dry Bulb - 47 Deg F
Ambient Wet Bulb - 40 Deg F
Initial Drop Diameter - .0165 ft
Initial Salinity - 0.5 ppt
Spray Height - 17 ft
Spray Width - 160 ft
Straight Canal of 30 Passes with 10 Sprays per Pass
Desired Cooling Range - 16.6 Deg F
627
-------
TABLE 1 (Cont'd)
Output Data
Pass
Number
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
Flow
^Rprn)
710528
709751
709011
708307
707634
706993
706393
705818
705267
704739
704232
703746
703278
702829
702407
Temperature
(dej? F)
88.0
86.5
85.0
83.6
82.2
80.9
79.7
78.5
77.4
76.3
75.3
74.3
73.3
72.4
71.5
Pass
Number
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
Flow
(fcpm)
702000
701607
701227
700861
700507
700165
699843
699530
699226
698933
698648
698378
698116
697861
697613
Temperature
(des F)
70.7
69.9
69.1
68.3
67.6
66.9
66.2
65.6
64.9
64.3
63.7
63.2
62.6
62.1
61.6
628
-------
TABLE 2
Number of Sprays Number of Sprays Canal Length
Case Across Canal Required (Number of Spray Lengths)
1 3 273 91
2 8 368 46
629
-------
TABLE 3
POWER OUTPUT VS CONDENSER INLET WATER TEMPERATURE
Percent Spray
Canal Operating
58
63
74
79
84
o\
o 89
95
100
105
Cold Water
Temperature - Deg F
109.3
106.6
102.8
100.6
98.7
97.0
94.6
92.7
91.3
Cross Power
Out -MW
819.1
820.3
822.1
823.0
823.4
823.8
824.3
824.5
824.6
Spray Motor
Power - MW
6.2
6.7
7.8
8.4
9.0
9.5
10.1
10.6
11.2
Net Power*
MW
812.9
813.6
814.3
814.6
814.4
814.3
814.2
813.9
813.4
*Net Power = Gross Power - Spray Motor Power
-------
TABLE 4
SPRAY NOZZLE PARAMETERS
Pressure (psi) 7 15
Spray Height (ft) 10 14
Spray Diameter (ft) 26 32
Drop Airborne Time (seconds) 1.6 1.9
Maximum Vertical Velocity 20.2 22.0
(feet per second)
Ratio of Water to Air Mass Flow .058 .048
Required Number of Nozzles 380 272
Per Stage
Required Area Per Stage (acres) 5.9 6.4
Required Number of Stages 2 1
631
-------
WIND
DIRECTION
HOT
WATER
IN
o o
o o
o o
o o
COLD
WATER
OUT
f
o o
o o
o o
o o
FIGURE ! U-SHAPED SPRAY CANAL PLAN VIEW
632
-------
HOT
WATER
IN
WIND
DIRECTION
1
o o o
o o o
o o o
o o o
1
COLD
WATER
OUT
o
o
o
o
FIGURE 2 STRAIGHT SPRAY CANAL PLAN VIEW
633
-------
to
cr
o
2
>
UJ
QC
OC
UJ
a.
2
UJ
108-
106-
104-
102-
100-
98-
96-
94-
92
0
27
54
81
—T
108
135
162
189
216
U-SHAPED
243
270
297
CUMULATIVE NUMBER OF SPRAYS
FIGURE 3 COLD WATER TEMPERATURE COMPARISON ALONG CANAL
-------
20
16.6'
UJ
UJ
a:
o
LJ
O
UJ
O
oc
e>
o
o
15-
10
5-
61
62
63
64
64.3
65
66
67
68
69
COLD WATER TEMPERATURE -DEGREES F
FIGURE 4 COOLING RANGE VS COLD WATER TEMPERATURE
-------
830 -
825 -
820 -
ID
0.
I-
O 815 -\
810 -
805 -
800 -
70
GROSS POWER OUTPUT
NET POWER OUTPUT
80
90
I I I I I 1
too no
120
COLD WATER TEMPERATURE - DEGREES F
FIGURE 5 POWER OUTPUT VS. COLD WATER TEMPERATURE
INTO CONDENSER - 100% LOAD
636
-------
X
o
Ul
X
PATH WITH HEAT LOAD
PATH WITHOUT HEAT LOAD
0 200 400 600 800 1000 1200 1400 1600 1800 2000 2200 2400 2600 2800 3000 3200 340O
DISTANCE - FEET
FIGURE 6 DRIFT DROPLET PATHS
-------
THE DEVELOPMENT OF ORIENTED SPRAY COOLING SYSTEMS
D. A. Fender
Ecolaire Condenser, Inc.
Bethlehem, PA U.S.A.
T. N. Chen
Ingersoll-Rand Research, Inc.
Princeton, New Jersey U.S.A.
ABSTRACT
The historical and theoretical development of Oriented Spray Cooling Sys-
tems (OSCS) from conception to its current state is described. Originated
by the Thermosciences Research Group of Ingersoll-Rand Research, Inc. in
1968, and being further developed after its purchase by Ecolaire Condenser,
Inc. in 1977, OSCS was conceived as a method to induce air flow through a
low stack cooling tower. Following numerical modeling and the laboratory
testing of several scale systems, a full scale, two dimensional model
cooling tower was constructed at South Carolina Electric and Gas Company's
Canadys Power Station. Extensive performance testing of this demonstration
model proved that system performance was unaffected by the tower enclosure.
As a result, the design of spray pond systems having comparable performance
to natural draft cooling towers without the typical spray pond dependence
on ambient wind conditions was established. OSCS development culminated
with the installation of an oriented spray cooling system for an industrial
turbine condenser application located at Phillipsburg, New Jersey. Heat
and mass transfer relationships are described, and performance curves are
presented for these viable new industrial and utility heat rejection
systems.
INTRODUCTION
OSCS is a method of evaporative heat rejection proprietarily owned and
patented by Ecolaire Condenser, Incorporated. These systems combine the
low cost and environmental acceptability of conventional spray systems
with the consistent and efficient performance of natural draft cooling
towers. OSCS is intended for use in utility plant condenser cooling,
industrial process cooling, or nuclear plant safety system cooling. It
can be adapted to a wide variety of configurations, depending upon plant
site topography, meteorology, and load requirements. This adaptability
makes OSCS particularly appropriate for retrofit and supplemental appli-
cation.
The standard OSCS arrangement consists of an annular array of vertical
spray tree modules (refer to Figure 1). These spray trees each consist
of a vertical riser pipe to which horizontal branch pipes are attached
at equal height intervals and at right angles to each other such that
the spray nozzles attached to the branch ends comprise a counter-
clockwise helix with increasing height. The number of nozzles per tree
638
-------
is selected according to the thermal performance characteristics required
for the specific application. OSCS provides consistent performance for all
ambient wind conditions, including no-wind periods, through its ability to
induce air flow through the fill area by momentum exchange with the sprayed
water droplets. Hence, the major task of the research effort expended on
OSCS was to develop the ability to predict this performance for various
meteorological and configurational conditions.
EXISTING COOLING SYSTEMS
In 1968, the Condenser Division' of Ingersoll-Rand conducted a market exam-
ination of the cooling requirements for commercial power generating plants.
This survey indicated that closed system type condenser cooling would be
showing a dramatic increase in usage. As environmental concern and activism
began to restrict the use of once-through cooling systems fed from natural
lakes and rivers, high efficiency cooling towers were being specified to
provide the lowest condenser circulating water temperature possible. The
survey led to the determination to investigate the feasibility of develop-
ing a viable evaporative cooling tower product line.
Natural draft cooling towers can be arranged for either counterflow,
parallel flow, crossflow, or combinations of these. Typically, maximized
air to water contact, necessary for efficient evaporation, is achieved by
droplet formation from cascading trays, plates, or slats. Air movement
through the "fill" section can be provided by the natural draft effect
caused by the heated air buoyancy within a tall stack section. In some
tower designs, spray nozzles are used to augment droplet formation and
subsequently increase the evaporative heat transfer.
Preliminary fluid dynamics analysis of the various possible fill designs
conducted by Ingersoll-Rand Research, Inc. (IRRI) indicated that in a
crossflow tower independent hollow cone nozzle spraying offered the great-
est potential for reducing tower fill pressure drop. Suggesting that a
smaller tower stack could be used, this finding led to the decision to
pursue the project as a natural draft crossflow spray cooling tower.
SPRAY TOWER PERFORMANCE AND FEASIBILITY
Until that time, no known significant theoretical analysis had been con-
ducted on spray towers. One spray cooling tower was found, and it was
operated by the Electricity Supply Commission in Johannesburg, South Africa.
In late 1969, IRRI undertook the development of an analytical model and
computer simulation program of the proposed spray cooling tower. This
model was to provide operating characteristics which would be subsequently
used in determining the technical and economical feasibility of spray
towers. A complete analytical model of the system was not possible due to
the vastly complex trajectory and thermal history of the spray droplets.
Therefore, the first step of the analytical study was to develop a simpli-
fied model of the actual process while retaining its essential physical
description.
639
-------
Vertical Rain Model
By considering each hollow cone spray within the tower to be comprised of
numerous pairs of spray droplets, the average spray cooling performance
was shown to reasonably approximate that of those droplet pairs moving
transverse to the air flow. This approximation is called.the "vertical
rain model" and was determined valid by a finite element computer anal-
ysis of both transverse and parallel spray models within a spray tower.
In this analysis, the heat and mass transfer was calculated based on the
average air and water properties entering each element, and air drag was
assumed negligible. The air temperature and humidity variation was deter-
mined using a trial and error approach to satisfy the overall energy
balance. Since each droplet trajectory was limited within a vertical
plane perpendicular to the air flow, the analysis was greatly simplified.
Tower Performance Simulation
The vertical rain model was subsequently used in the development of a spray
tower computer simulation program. Again, a finite element grid was
devised to allow numerical solution of the governing equation for heat
and mass transfer, draft, pressure loss, and water surface area and dis-
tribution. The program also allowed the optional simulation of a draft-
inducing fan to augment air movement through the fill.
The results which this program generated suggested that a crossflow spray
tower could be designed to produce comparable cooling performance to a
conventional state-of-the-art tower. This suggested that spray cooling
towers were, indeed, technically feasible. The analysis also indicated
that performance was strongly dependent upon the effective droplet size.
Therefore, not only is the droplet size distribution important, but
droplet collisions or "interference" have a strong effect.
Spray Nozzle Orientation
Finally, the pressure drop through the spray fill was confirmed as the
largest single pressure loss in the tower. To reduce this pressure loss,
and thus reduce the tower stack cost, Dr. T. N. Chen of the IRRI proposed
that the fill spray nozzles be oriented towards the air flow direction.
By reducing the horizontal relative velocity between the water drops and
the air, fluid dynamic theory indicates that the viscous drag, and conse-
quently the pressure drop, would be reduced. Hence, the stack height
required for draft would be reduced as the water droplet horizontal velocity
in the air flow direction is increased. Taken further, the principle
indicated that the tower stack could be completely eliminated if the spray
velocity were sufficient to drag the air through the spray section. This
proposal has become the essential principle of the Oriented Spray Cooling
System concept.
The spray tower computer model was consequently revised to allow the study
of a spray nozzle orientation other than vertically upward. Parametric
studies of the resultant oriented spray system confirmed both its technical
feasibility and drop size dependence.
640
-------
Since the droplet interference is a function of nozzle arrangement and
operating pressure, and it has a strong effect on cooling performance,
it was concluded that a full scale test program was necessary to establish
the performance of OSCS.
LABORATORY MODEL TESTING
Prior to engaging in a full scale prototype testing program, two labora-
tory scale model tests were devised. Using appropriate scaling analysis,
much valuable data concerning various design parameters could be gathered
at greatly reduced expense. The tests were primarily intended to examine
air flow and recirculation phenomena, and to provide criteria for the design
of the full scale prototype.
Laboratory Flow Testing
Beginning in early 1970, this laboratory test program was intended to
experimentally confirm the oriented spray principle of inducing air to
create the draft necessary for cooling performance, and to determine the
effects of various design parameters on this air flow. The test apparatus
consisted of a long, narrow box with an array of horizontal spray manifolds
at the inlet end and a roof opening at the outlet end to direct the sprayed
induced air upward. The manifolds were drilled in a special multiple
orifice arrangement to simulate hollow cone spray nozzle effect. This
type of module was used because it could be scaled to the dimensions of
a finite "slice" through an annular OSCS unit.
Thermal data were taken of air and water conditions, and smoke traces were
recorded for air flow pattern determination. The test results clearly
demonstrated the effectiveness of the oriented spray principle in inducing
air movement. The test also confirmed the strong effects of droplet
interference and spray distribution. The smoke traces of air motion through
the spray apparatus did indicate that air flow deviated from the horizontal
flow assumption used throughout the theoretical study. This is understand-
able because of the parabolic trajectory of the spray droplets inducing the
air. In addition, by removing the apparatus roof, it was confirmed that a
tower enclosure was unnecessary for performance. The results suggested
that further testing of a full scale system would be needed for accurate
performance prediction.
Recirculating Testing
The recirculation of the exhaust plume effluent is a common problem of all
evaporative cooling systems, causing a negative effect on cooling perfor-
mance. Beginning in mid-1972, a laboratory model was constructed and
tested to determine the general magnitude of the recirculation which an
OSCS would exhibit (refer to Figures 2, 3, and 4).
The models consisted of pairs of small open boxes arranged in two different
configurations in a horizontal air stream. The first arrangement consisted
of two linear rows of these box pairs separated by a variable space.
641
-------
The second consisted of an annular arrangement. A wind velocity profile
was simulated by controlling the air velocity profile with a variable
opening screen. The outer box of each set removed a quantity of air cal-
culated to be equivalent to the air induced by the spray. This air was
ducted to a heating device which raised its temperature as desired,
returned it to the inner box, and discharged it to atmosphere (refer to
Figure 2). Recirculation could then be determined for each of the downwind
boxes by the increase in the air inlet temperature over the ambient.
This procedure yielded a reliable quantitative recirculation allowance
factor for the OSCS design.
FIELD DEMONSTRATION SYSTEM TESTING
Full scale testing of an actual prototype OSCS for performance determina-
tion would be substantially cost prohibitive. Hence, it was decided that
these necessary tests would be performed on £. full scale replica of the
laboratory model. Again, this module represented a finite slice of a com-
plete OSCS unit. The comparison between performance factors of this linear
model and" an annular section was determined to be equivalent by the use of
appropriate flow area.
Construction of this 10,000 gpm capacity test module was begun in 1973
at South Carolina Electric and Gas Company's Canadys Power Station. Field
tests were initiated that year, and were continued over an 18 month period.
The tests were intended to provide full scale verification of the oriented
spray principle. They also allowed determination of the effects of various
design parameters such as nozzle size, orientation distribution, pressure,
and exhaust area, as well as the effects of the operating heat load and
ambient meteorological conditions on the overall cooling performance.
Finally, the tests provided extensive data upon which a thermal performance
model for the entire operating range could be accurately based.
During the testing period, 486 separate test runs provided specific point
data over an extensive range of operating pressures, water loadings, nozzle
arrangements, and meteorological conditions. The data were then correlated
into empirical relationships.
Parallel to the test program, a design analysis was conducted to optimize
the method for piping and distributing the water. This analysis led to the
design of the helical pattern of the vertical riser spray tree, a major con-
tributor to the OSCS patent.
DEVELOPMENT OF PERFORMANCE MODELS
The introduction of OSCS into the commercial marketplace also required the
development of several mathematical models for various performance criteria.
These models are essential to the proper application of OSCS to specific
design conditions. They have enabled OSCS to be engineered to specific
applications with a precision and efficiency not found in other spray ponds
or spray towers.
642
-------
Thermal Performance Model
The thermal performance model was formulated from the conservation laws of
mass and energy within a control volume within the spray zone. The energy
balance is:
L C (T. - T ) = KAV(MhD) (1)
f Wll Wl_
where:
L = total water loading rate
C = specific heat of water at constant pressure
TW = average temperature of all water drop at.a point in
their falling period
•^wh = temperature of hot water entering spray
Twc = temperature of cold water after spray
K = mass transfer coefficient
A = average water surface area per unit volume
V = total volume of spray zone
hws = total enthalpy of air-water vapor mixture in equilibrium
with the water drop at a surface temperature Tws and a
bulk temperature Tw
ha = average total enthalpy of air in contact with the drops
MhD, the mean enthalpy difference between the vapor film at the water surface
and the bulk air, is defined from:
MhD = Twh ~ Twc /2)
r Uiw
J ^ Njg - ha'
T
we
so that Equation (1) can be written in the familiar form:
643
-------
The dimensionless KAV/L parameter, as it is for conventional cooling towers,
is the performance factor for a given spray system. Examination of the
left hand side terms of Equation (3) indicates that the KAV/L variation
with temperature is small (±5%). Hence, KAV/L is primarily a function of
nozzle configuration and operating pressure, and not a significant function
of temperature. Given a specific OSCS design, then Equation (3) becomes
a simple relationship between Twc, Twft, and the ambient air enthalpy, hai-
Since KAV/L cannot be analytically calculated, it must be determined
empirically. Thus, the Canadys test data provided the KAV/L value for
each different nozzle arrangement and pressure. By substituting the
various temperature and enthalpy relationships for a variety of load and
ambient conditions into Equation (3), the statistical KAV/L value could
be determined. A typical OSCS performance curve for a specific arrange-
ment or KAV/L is shown in Figure 5. It should be noted that an OSCS KAV/L
value should not be compared with a conventional cooling tower KAV/L,
because the KAV/L"s in the two cases are not defined exactly the same.
Furthermore, because the experimentally determined KAV/L is effectively a
no-wind factor, and since inclusion of wind effects should improve cooling
performance, we can deduce that these KAV/L values reflect the worst-case
(no-wind) performance.
Theoretical Drift Models
All evaporative cooling systems require special attention to the problem
of water droplets which become entrained in the air stream. This phenomenon
is called "drift", and is, in many areas, a substantial environmental and/
or economic problem.
Two separate numerical drift models were developed for computer simulation.
First, the "low-wind" drift model considers the wind to be sufficiently low
to allow the formation of a buoyant plume. By treating the notion of the
water droplets carried upward in the buoyant plume in a cross wind, a finite
difference computer model was generated. This program produces droplet
fallout distances for various wind orientations, drop diameters, and
operational/ambient conditions. Total drift is determined by using the
drop size distribution curves for the nozzle used and the fraction entered
and lifted by the plume to calculate the total mass of droplets which will
fallout beyond the spray basin boundaries. Hence, the basin boundaries can
be established to catch all but th* allowable percentage of drift.
The second model developed was the "high-wind" drift model, which assumes
that the wind forces are strong enough to prevent plume formation. Similarly
to the low-wind model, the high-wind model uses numerical techniques to
resolve the drag forces into particle trajectories. Again, the allowable
percent drift can be met by appropriate design of the basin boundaries.
The worst case model for the specific application is used to determine the
most sonservative basin dimensions for a particular drift requirement.
644
-------
Theoretical System Model
Because the actual cooling accomplished by OSCS is complicated by the tran-
sient system response due to the thermal capcitance of the pond water mass,
a computerized finite difference system cooling model was developed for
predicting the temperature of the pond water being supplied'to the service
system. This model requires consideration of the supplied hot water load,
ambient conditions, pond volume, surface radiation and convection, drift
and evaporation loss, and mixing effects. While this system is helpful
for condenser and industrial process applications, it is essential to
nuclear power plant ultimate heat sink (UHS) design.
PROTOTYPE INSTALLATION
The development of OSCS culminated in 1976 with the installation of an
actual working system at Ingersoll-Rand Company's Phillipsburg, New Jersey,
Turbo Division, plant. The plant cooling pond provides cold circulating
water to various turbine test stand condensers according to a variable
testing schedule. An approximately sixty-year-old flatbed spray cooling
system was converted to OSCS to provide the most adequate and economical
cooling for an increased thermal loading following test facility expansion.
After removal of the old flatbed headers, a new annular header system with
32 spray trees was installed. A helicopter was used for member placement,
facilitating assembly without requiring the pond to be drained. The total
cost of the 20,000 gpm installation, including pumps, was $330,000. This
Phillipsburg OSCS has been in service since its initial startup (refer to
Figure 6), and has provided more than adequate cooling performance.
FOLLOW-UP TESTING AND DEVELOPMENT
Ecolaire's OSCS development program is primarily aimed at further verifica-
tion of the thermal performance model generated from the Canadys data by
thermal testing of the Phillipsburg installation. Such results are expected
in the near future.
Other development work includes computer optimized design methods to
provide the user with the maximum efficiency cooling system at the minimum
cost for the specific application. And, »s with any other newly introduced
technology, OSCS will require an ongoing development program to further
reduce manufacturing and installation costs.
CONCLUSIONS
The above historical review has shown how a unique large scale water cooling
system utilizing water sprays to induce cooling air flows was successfully
developed and expanded into a viable product line. OSCS offers users the
advantages of low cost and environmental acceptability of the conventional
spray pond and the consistent and efficient performance of natural draft
645
-------
cooling towers. The development was conducted through a concerted and
illustrative combination of analytical, numerical, and experimental pro-
cedures which should be insightful to all research and development
engineers concerned with similar tasks. The resultant addition of OSCS
to evaporative cooling technology represents a major contribution to the
industry.
646
-------
\l //
Relation to
Spray AnnuI us
B-nozzle spray tree rise
water •— -.
droplet /
trajectory—sy X"
' ^ Inlet
,t .Air
nozzles
base of tree rise
flange connection
SECTION
'A-A'
Figure 1: Oriented Spray Cooling System Cross Sectional View
CELL PAIR
THERMOCOUPLE
Dl SCHARGE
MANIFOLD
FLOW VALVE
LOCATION
AIR HEATER -•-
3UC-
^J\
ORIFICE
BLOWER
NLET MANIFOLD
-*£
THERMOCOUPLE
THERMOCOUPLE
Figure 2: Schematic of Recirculation Testing Apparatus
647
-------
00
PLAN:
-*- u-
ELEVATION:
Area Constraint : A 3
"o"
Ae 4
h • height of eel 1s
A. » total area of the Inlet to the test apparatus
A - total exhaust area of the test apparatus
Figure 3: Reclrculation Testing Apparatus
Annular Arrangement
L
X.
« overall length of OSCS Installation
• width • depth - height of spray system • h
aspect rat Io - L/h
Inlet
exha
et area » 3 • h * L x 2 • h
aust area V D x L x 2 ff
angle between wind, L-Lo , and centerllne of model
Figure 4: Recirculation Testing Apparatus
Linear Arrangement
-------
ECOLAIRE CONDENSER, INC. - 21 Aug. 1978
VO
120 f;
u.
o
Uj
cc
>~
cc
I
UJ
cc
UJ
I-
i
o
too
VVE7 SUiS TEMPERATURE
Figure 5: A Typical Oriented Spray Cooling System Performance Curve
-------
Figure 6: OSCS Installation At Phillipsburg, New Jersey
650
-------
ONCE-THROUGH COOLING POTENTIAL OF THE
MISSOURI RIVER IN THE STATE OF MISSOURI
A.R. Giaquinta
Institute of Hydraulic Research T'C' Keng
The University of Iowa and Jenkins-Fleming, Inc.
Iowa City, Iowa U.S.A. St- Louis' Missouri U.S.A.
ABSTRACT
The reach of the Missouri River bordering or crossing the state of
Missouri is studied with regard to its potential for use in once-through
cooling of steam-electric power plants. Based on the existing thermal
standards of the state regulatory agencies, the remaining heat assimi-
lation capacity of the river is computed, and sites and capacities of
future permissible once-through-cooled power plants are determined.
The existing and future thermal regimes of the river are computed from
a heat balance equation relating the rates of convective heat transfer,
surface heat exchange between the river and the atmosphere, and heat
inputs from power plants or other artificial sources. Streamwise temp-
erature distributions with existing, future proposed, and future per-
missible heat loads are shown for average river flow conditions and for the
7-day, 10-year low flow condition.
It is demonstrated that this reach of the Missouri River can accomodate
additional once-through-cooled power plants with a total capacity of
several thousand megawatts at average flow conditions. These new plants
must be properly sited to avoid the cumulative effects of upstream thermal
loads. If thermal standards were based on the low flow condition, the
total permissible capacity would be significantly reduced.
INTRODUCTION
I
The continuing increase in demand for electrical energy and the resultant
growth of the electrical power industry in the United States have given
rise to certain environmental problems related to the siting and design
of new power plants. Once-through (open-cycle) cooling is known to be one
of the most efficient methods for condenser cooling. This method is
efficient economically, thermodynamically, and, if the cooling water out-
fall structure is designed properly, it is efficient ecologically.
However, the U.S. Environmental Protection Agency has mandated that in the
near future new power plants will not be allowed to utilize open-cycle
cooling and some older plants will have to backfit closed-cycle cooling
systems. These regulations will incur tremendous expense and a great
increase of energy consumption. As more studies are completed, it is
being found that thermal pollution is not as ecologically harmful as
651
-------
originally thought [1]. Therefore, in light of an expanding power in-
dustry, it is important to consider the future open-cycle cooling poten-
tial of our nation's major rivers.
In this study attention is focused on the reach of the Missouri River
bordering or crossing the state of Missouri. Along this reach which passes
through or near the two major population centers of Kansas City and St.
Louis, the river is used for open-cycle cooling by several power plants,
and more once-through-cooled units are proposed for construction within the
next decade. To determine the once-through cooling potential of the river
it is important to consider the cumulative effects of the existing and
future power-plant discharges. It also is necessary to consider the avail-
ability of water for use in open-cycle cooling and the amount of evapora-
tive water loss. It was shown in [2] that consumptive use is no problem.
The steady-state version of the Iowa Thermal Regime Model (ITRM), a
numerical model for the calculation of streamwise temperature distribu-
tions in rivers, is used to determine the existing and future thermal
regimes of the Missouri River downstream from the southern Iowa border.
The basic equation governing the conservation of thermal energy in a
free-surface flow is reviewed, and the numerical model is presented.
The steady-state ITRM is used to determine the thermal regimes of the
Missouri River along the study reach corresponding to average meteorologi-
cal and hydrological conditions for the months of February, May, August,
and November (representing the four seasons of the year) . The natural
thermal regimes and the modified thermal regimes resulting from the im-
position of external heat loads from power plants and other sources are
calculated. Results are shown in the form of temperature distributions
along the river for the cases of existing power plants and future power
plants proposed for installation within the next decade. Based on the
existing thermal standards of the state regulatory agencies, the remaining
heat assimilation capacity of the river is computed, and sites and capa-
cities of future permissible once-through-cooled power plants of reasonably
large size are determined. The resultant temperature distributions corres-
ponding to these future permissible plants are presented.
It is shown that there is no remaining heat assimilation capacity of the
river in the vicinity of Kansas City. No additional future power plants
using open-cycle cooling are permissible upstream from Kansas City because
they would cause violations at downstream locations. The total capacity
of future permissible plants at average flow conditions is about 6000 MW
for fossil-fueled plants or about 4100 MW for nuclear-fueled plants, based
on allowable increases above the natural temperature base. Thermal regimes
at the 7-day, 10-year low flow hydrological condition also were studied by
Giaquinta and Keng [2] , and some results are presented herein. At the low
flow condition, some existing and proposed future power plants are seen to
violate the excess temperature limitation if they operate at full load.
These plants would require derating.or auxiliary cooling at this extreme
condition. Based on the low flow the total future permissible plant capa-
city is about 1300 MW for fossil plants or about 900 MW for nuclear plants.
652
-------
THE MISSOURI RIVER SYSTEM
The source of the Missouri River is in the state of Montana, and it flows
generally southeasterly 2315 miles to its junction with the Mississippi
River about 15 miles above St. Louis, Missouri. River miles along its
channel are measured upstream from the intersection of the thalwegs of
the Missouri and Mississippi Rivers.
The river reach of concern in this study, starts at the Iowa-Missouri
border (Mile 553) and continues to the confluence with the Mississippi
River (Mile 0). This reach borders the states of Nebraska and Kansas
and crosses the state of Missouri. The major tributary streams entering
the river in the downstream order are the Kansas River (Mile 367) , Grand
River (Mile 250), Chariton River (Mile 239), Osage River (Mile 130), and
Gasconade River (Mile 104). The general layout of the river system is
shown in Fig. 1.
The climatic conditions are represented by data from five Class-A weather
stations located along or close to the Missouri River in the study area.
Monthly mean values of daily weather data for the 20-year period from
1954 through 1973 were used in the analysis. The locations of the weather
stations and detailed tables of data are given in reference [2].
The Missouri River flow rate is regulated by six reservoirs upstream from
Sioux City, Iowa. Because the present study reach is far downstream from
the reservoirs, the thermal effects of the reservoir control are negligible,
and only the obvious consequences of the reservoir regulation on flow rate
are considered.
TABLE I gives a summary of the monthly mean values of daily flow rates for
a 19-year period (1956-1974) at seven gaging stations along the study
reach. Detailed flow-rate tables and a map showing the locations of the
gaging stations are included in reference [2].
The thermal standards for the Missouri River are governed by the water pol-
lution control agencies of the states bordering the river. The maximum
allowable temperature rise is 5°F (2.78°C) and the maximum water temperature
is 90°F (32.2°C) for the entire study reach. The allowable temperature
increase is the governing standard for all the cases considered herein.
Eleven power plants utilizing open-cycle cooling with a total of 40 units
are located along the study reach; their locations are shown in Fig. 1.
Only steam-electric power plants with capacities greater than 50 MW are
considered. TABLE II summarizes the loading and cooling system character-
istics of these power plants. The major data source used in preparing
the table was the FPC Form 67. For most of the power plants along the Mis-
souri River, the forms were provided by the utilities for the year ended
December 31, 1974. For those not supplied by the utilities reference [3]
which covered the year ended December 31, 1973, was used. The list of
utilities and their abbreviations are given in the appendix.
653
-------
The total installed plant capacity is about 5690 MW, which consists of
about 4880 MW fossil-fueled and 810 MW nuclear-fueled plants. All of
these plants use once-through cooling systems. There are four future power
plants proposed for construction through 1990, as listed in TABLE III. The
total plant capacity proposed for installation is about 6620 MW, of which
4240 MW is planned for once-through cooling, and 2380 MW for natural
draft cooling towers [2]. Heat loads from industrial and municipal sources
were considered and found to be negligible compared to the heat loads due
to power plants.
COMPUTATIONAL MODEL
The general differential equation that describes the conservation of heat
in an elemental volume of water in a river is three-dimensional and unsteady.
However, in most streams large temperature gradients in the transverse and
vertical directions occur only in the near-field regions of sites where ther-
mal loads are imposed. In considering the overall thermal regime of a river,
the zones of the three-dimensional effects usually are small compared to
the lengths of the river reaches, and, therefore, a one-dimensional formu-
lation may be employed.
Also, in examining the thermal regime of a river, it frequently suffices
to determine the steady-state temperature distributions corresponding to
average meteorological, hydrological, and thermal loading conditions. Based
on these simplifications, the one-dimensional, steady convection-diffusion
equation expressing the conservation of thermal energy in a free surface
flow may be expressed as
dT = B_ (fr*(T) + TI .
dx Q P cp QPC
in which T = cross-sectional average river temperature, x = streamwise
distance along the channel, B = top width of the river flow section, Q
= river discharge, * = rate of surface heat exchange between the water
and the atmosphere, TI = rate of heat input from power plants and tribu-
tary inflows per unit length along the stream, p= density of water, and
c = specific heat of water. The terms of the equation represent the heat
aavected by the current, the heat transferred by the air-water interfacial
transport processes, and the rates of artificial and tributary heat inputs
to the river. This equation can be solved to obtain the steady-state longi-
tudinal distribution of temperature in a river.
The computational technique to solve Eq. (1) is a finite-difference method
based on the steady-state Iowa Thermal Regime Model (ITRM) developed by
Paily and Kennedy [4]. This method was employed to study the thermal re-
gimes of the Upper Mississippi and Missouri Rivers [5] and was validated
in that study by comparing numerical results with measured field temperature
data along both rivers.
To compute the temperature distribution along a river the total river length
is divided into a convenient number of reaches. Each reach is subdivided
654
-------
into a finite-difference grid defined by a number of meshpoints. The solu
tions for adjacent reaches are linked by the common conditions at the junc
tion node points connecting them. If the temperature at any meshpoint,
Xi' iS Ti' the temPerature at the next meshpoint, X. , which is at a
distance Ax downstream is given by x
in which 4>*.+3^ is the surface heat exchange rate corresponding to T. •>,
the temperature at the middle of the mesh space, Ax. The temperature^2
T
i+% is determined by
in which *. corresponds to the known temperature, T. . If the temperature
at the upstream boundary (i=l) is known, Eqs, (2) and (3) can be used to
calculate the temperature at downstream meshpoints i=2,3,...N, where N is
the total number of meshpoints for the entire length of the reach under
consideration .
In the above equations, the rate of surface heat exchange, * (T) , is one
of the principal factors influencing the thermal regimes of the river .
It depends upon several climatic factors, including solar radiation, air
temperature, wind speed, relative humidity, atmospheric pressure, and
cloud cover. The most important processes included in the mechanism of
heat transfer between the water surface and the atmosphere are the net
short-wave radiation entering the waterbody, ; and the melting
of snow, A . The heat transfer process is expressed as
s
-di.-d,- (4)
y VR VB ^E VH VS
Equations for computing the components of surface heat exchange are
given elsewhere [2,5,6],
The input data required for the model include (1) river mile at the up-
stream boundary; (2) river temperature at the upstream boundary; (3) number
and spacing of meshpoints in each reach; (4) top widths of the river at
selected stations along the river,- (5) main stem river flow rates at selec-
ted locations and tributary inflows; (6) weather conditions at suitable
locations along or close to the stream including air temperature, wind
speed, relative humidity, cloud cover, cloud height, visibility, atmos-
pheric pressure, and solar radiation; and (7) thermal discharges into the
river from power plants and other sources. It was assumed that flow rates,
climatological data, and top widths varied linearly- between measuring sta- _.
tions, and linear interpolation was used to distribute these variables from
the measuring stations to each meshpoint.
655
-------
THERMAL REGIMES
The thermal regimes of the study reach for the months of February, May,
August, and November were computed by the ITRM. In addition to the
assumptions mentioned previously, additional assumptions were made for
the interpretation and use of the Missouri River data as follows:
1. The upstream boundary condition temperatures at the Iowa-Missouri border
(Mile 553) were obtained from reference [5].
2. Thermal impacts of tributary streams were neglected.
3. For the determination of future permissible power plants, temperature
increases are considered relative to the natural-temperature base.
4. All existing, future proposed, and future permissible power plants
operate at full load, and load capacity factors are not considered.
5. As noted in reference [5], Cooper No. 2 and 3, future units at Brown-
ville, Nebraska, will be required to use closed-cycle cooling.
Four temperature profiles corresponding to average meteorological and
hydrological conditions for each study month were predicted. (1) natural
thermal regime of the river; (2) temperature distributions with existing
heat loads (as of January 1975); (3) temperature distributions with exis-
ting and future proposed power plants; (4) temperature distributions with
permissible new power plants which could be installed without violating
present thermal standards.
The locations and sizes of future permissible plants were determined from
the natural-temperature base. The sites were chosen arbitrarily, avoiding
reaches already heavily loaded thermally, and spaced approximately 100
miles apart.
PRESENTATION AND DISCUSSION OF RESULTS
The temperature distributions corresponding to average flow and weather
conditions for the months of February, May, August, and November with the
permissible future plants determined from the natural-temperature base were
computed. The temperature distributions for the month of November are shown
in Fig. 2. These curves are typical of the thermal regimes for all the
study months which are given in reference [2]. It is seen that there is
no more heat assimilation capacity of the river in the vicinity of Kansas
City. Upstream from Kansas City no additional future power plants using
open-cycle cooling are possible because they will cause temperature viola-
tions at downstream locations. Three sites are chosen well downstream from
Kansas City for permissible new plants. At each site the maximum allowable
plant capacity is found for each of the study months, and the future per-
missible capacity is taken as the lowest value for the four months. The
permissible capacities of fossil-fueled power plants at the three sites are
656
-------
(1) Mile 230.5 - 1980 MW, (2) Mile 113.1 - 2860 MW, and (3) Mile 19.1 -
1150 MW. If the plants were nuclear fueled, the permissible capacities
would be reduced, because of efficiency differences, by about 31 percent
to (1) 1370 MW, (2) 1970 MW, and (3) 790 MW.
The temperature distributions corresponding to the 7-day, 10-year low flow,
combined with average weather conditions for the months of August and Novem-
ber and with the permissible future plants based on natural temperature is
shown in Fig. 3. The figure shows that existing power plants in the vici-
nity of Brownville, Kansas City, and Sibley will violate the 5°F excess
temperature limitation; the future proposed plants at Brownville, latan,
and Nearman will violate the criterion also if it is assumed that they operate
at full load. Closed-cycle cooling systems will be required during low
flow conditions at these plants unless they operate at less than full
capacity.
Owing to the rapid decay of temperature during low flow conditions some
new plants may be permitted downstream from Kansas City. The allowable
capacity at Mile 230.5 is 290 MW for a fossil-fueled unit or 200 MW for a
nuclear-fueled unit; at Mile 19.9 the permissible capacity is 1000 MW -
fossil or 690 MW - nuclear.
CONCLUSIONS
Steam-electric power plants will continue to play an important role in the
power industry because of the increasing need of energy in the future.
However, more stringent environmental regulations concerning waste heat
released from power plants demand a better understanding of the thermal
regimes and the heat assimilation capacity of rivers, particularly for the
planning of future plants employing open-cycle cooling. The Iowa Thermal
Regime Model can predict temperature distributions in a river downstream
from imposed thermal loads provided that specific hydrological, meteorolo-
gical, and geometrical parameters describing the river are known. This
model was employed to show that the reach of the Missouri River bordering
or crossing the state of Missouri can be used for once-through cooling of
future power plants based on present thermal standards.
The total permissible future plant capacity for this reach of the river is
about 6000 MW for fossil plants at average flow conditions and about 1300
MW for fossil plants at the 7-day, 10-year low flow condition. Permissible
capacities of nuclear plants are reduced by about 31 percent to about 4100
MW for average flow conditions and to about 900 MW for low flow conditions.
The vicinity of Kansas City already is heavily loaded thermally. If the
natural-temperature base is used, no additional future once-through-cooled
power plants are permissible upstream of Kansas City because they would cause
temperature violations at downstream locations.
If thermal standards were based on the 7-day, 10-year low flow, the total
permissible plant capacity would be reduced by about 78 percent.
657 ARC
-------
ACKNOWLEDGEMENT
This project was financed in part by a grant from the U.S. Department of
the Interior, Office of Water Research and Technology under Public Law
88-379 as amended, and made available through the Iowa State Water Re-
sources Research Institute. Funds for computer time were provided by
the Graduate College of The University of Iowa.
APPENDIX
List of Utilities and Abbreviations
CEPC Central Electric Power Co-op.
KCBPU Kansas City (Kan.) Board of Public Utilities
KCPL Kansas City (Mo.) Power & Light Co.
MPS Missouri Public Service Co.
NPPD Neb'raska Public Power District
SJLP St. Joseph Light & Power Co,
UEC Union Electric Co.
LIST OF REFERENCES
1. Utility Water Act Group , "Biological Effects of Once-Through Cooling,"
Vol. 1, Introduction: Principles of Quantitative Impact Assessments,
Vol. 4, Rivers and Reservoirs, submitted to U.S. Environmental Protec-
tion Agency, June 1978.
2. Giaquinta, A.R. and Keng, T.T.C., "Thermal Regimes of the Mississippi
and Missouri Rivers Downstream from the Southern Iowa Border," IIHR
Report No. 211, Iowa Institute of Hydraulic Research, The University
of Iowa, Iowa City, Iowa, January 1978.
3. Federal Power Commission, "Steam-Electric Plant Air And Water Quality
Control Data"(for the year ended December 31, 1973 based on FPC Form
No. 67) Summary Report, FPC-S-253, Federal Power Commission, Washington,
D.C., January 1976.
4. Paily, P.P. and Kennedy, J.F., "A Computational Model for Predicting
the Thermal Regimes of Rivers," IIHR Report No. 169, Iowa Institute of
Hydraulic Research, The University of Iowa, Iowa City, Iowa, November 1974.
5. Paily, P.P., Su, T.Y., Giaquinta, A.R., and Kennedy, J.F., "The Thermal
Regime of the Upper Mississippi and Missouri Rivers," IIHR Report No.
182, Iowa Institute of Hydraulic Research, The University of Iowa, Iowa
City, Iowa, October 1976.
6. Giaquinta, A.R. and Keng, T.T.C., "Thermal Regimes of the Middle and
Lower Mississippi River During Low Flow Conditions," to be presented
at the ASME Winter Annual Conference, San Francisco, California, December
10-15, 1978.
658
-------
TABLE I
SUMMARY OF MONTHLY MEAN VALUES OF DAILY FLOW RATES
Gaging
Station
Nebraska City
Rulo
St. Joseph
Kansas City
Waverly
Boon vi lie
Hermann
River
Mi 1 P
562.6
498.0
448.2
366.1
294.4
196.6
97.9
Mean Daily Flow Rates in cfs
Averaging „_,
Period
1956-74
1956-74
1956-74
1956-74
1956-74
1956-74
1956-74
reo.
22,272
23,910
25,209
31,696
32,040
39,613
53,621
May
42,101
45,184
47,544
59,134
59,884
71,237
99,149
Aug.
39,494
40,843
42,459
49,047
49,108
53,546
61,151
Nov.
34,033
35,733
36,486
42,935
43,506
50,576
63,073
TABLE II
EXISTING POWER PLANTS ALONG THE LOWER MISSOURI RIVER
POWER PLANT
Utility
NPPD
SJLP
SJLP
KCBPU
KCBPU
KCPL
KCPL
KCPL
MPS
CEPC
UEC
Name
Cooper
Edmond St.
#4,5,7
Lake Road
#1
#2
#3
#4
Quindaro
#2
#3
Kaw
Grand Ave.
Northeast
Hawthorn
Sibley
#1,2
#3
Chamois
Labadie
LOCATION
City, State
Brownville, NE
St. Joseph, MO
St. Joseph, MO
Kansas City, KN
Kansas City, KN
Kansas City, MO
Kansas City, MO
Kansas City, MO
Sibley, MO
Chamois , MO
Labadie , MO
River
Mile
532.5
449
446
374
367.5
365.7
364.4
358.3
336.4
117
57.6
INSTAL.
Rated
Capacity
MW^
810 (N) a
51
15
20
12.5
90
94.5
239.1
161.3
126.7
88
910.1
100
418.5
67.7
2482
CONDENSER FLOW
Quantity
cfs
1455
129.2
47.8
62.3
44.6
114.6
262
340
273
145.6
ft
101. 8C
C1
1045.7
133
393
106.9
1676
Temp.
Rise
op
18
13
18.3
12.2
14.5
17.2
14
14.3
15.9
c
18
c
18
c
18
19.2
17.5
15
29.7
a N=Nuclear, all other-units are fossil
b Plant located on the Kansas River close to Missouri River
c Assumed condenser temperature rise and calculated condenser
flow rate
659
-------
TABLE III
FUTURE PROPOSED POWER PLANTS ALONG THE LOWER MISSOURI RIVER
POWER PLANT
Utility
NPPD/OPPD
KCPL/SJLP
KCBPU
UEC
Name
Cooper
#2
#3
latan
#1
#2
Nearman
#1
#2
Callaway
#1
#2
LOCATION
City, State
Brownville, NE
latan, MO
Nearman, KN
Fulton, MO
River
Mile
532.5
411
380
128
INSTALLATION
Rated
Capacity
MWP
1150
1300
630
630
235
300
1188
1188
Fuel
Type3
N
N
F
F
F
F
N
N
Cooling
System*^
OTF
OTF
OTF
OTF
OTF
OTF
NDCT
NDCT
CONDENSER
FLOW
Quan.
cfs
2066
2335
746
746
270=
345C
1220
1220
Temp.
Rise
oF
V»
18b
18°
18.7
18.7
18°
18C
30
30
SCHEDULED
IN-SERVICE
DATE
/85
5/89
4/80
4/81
4/79
4/83
10/81
4/83
a F = Fossil; N = Nuclear
b Assumed same condenser temperature rise and efficiencies as Cooper #1
c Assumed condenser temperature rise and calculated condenser flow rate
d OTF = Once-through fresh; NDCT = Natural draft cooling tower
-------
OK.
Figure 1. Location of Existing Thermal Power Plants Along the Lower
Missouri River
661
-------
LOCATIONS OF EXISTING MO PROPOSED POWER PLANTS
LOCATIONS OF EXBTINO AND PROPOSED POWER PLANTS
1
%
tn
1
I
K
tt
z
1
VI
<
x
w
_j
•
I
X
O
1
at
1
i
MISSOURI RIVER MILES
300 200
11< MISSOURI RIVER
ii; AVERAGE NOVEMBER CONDITIONS
•I*
• SITES FOP NEW PERMISSIBLE PLANTS
-NATURAL {NO PLANTS)
-WITH EXISTIH8 PLANTS
-WITH EXISTING AND PROPOSED PLANTS
I
-WITH EXISTING, PROPOSED, AND PERMISSIBLE PLANTS ; -T"
180 240
320 400 480 S80 840 720
DISTANCE DOWNSTREAM IN KILOMETERS
3s
:s
u •«
s«
MISSOURI RIVER MILES
ZOO
_J
100
I
NATURAL (NO PLANTS!
WITH EXI9TINQ PLANTS
WITH EXISTING AND PROPOSED PLANTS
WITH EXIST!M. PROPOSED, AND PERMISSIBLE PLANTS
MISSOURI RIVER
LOW FLOW NOVEMBER CONDITIONS
• »TE9 FOR HEW PERMISSIBLE PLANT*
240 920 400 480 560 MO TZO
DISTANCE DOWNSTREAM IN KILOMETERS
Figure 2. Temperature Distributions for Average
Conditions with Permissible New Plants
Determined from the Natural-Temperature
Base
Figure 3. Temperature Distributions for Low Flow
Conditions with Permissible New Plants
Determined from the Natural-Temperature
Base
-------
A MODEL FOR PREDICTION OF
EVAPORATIVE HEAT FLUX IN LARGE BODIES OF WATER
A.M. Mitry
Duke Power Company
Charlotte, NC 28242, U.S.A.
B.L. Sill
Department of Civil Engineering
Clemson, SC 29631, U.S.A.
ABSTRACT
In earlier papers [1,2], one and two dimensional analytical models have been
developed for the prediction of seasonal variation of the temperature dis-
tribution in large bodies of water. The one dimensional model, along with
field data are used to evaluate various wind speed functions (used in calcu-
lating Tg and K' and evaporative heat flux) to determine their overall
effect on temperature profiles in stratified lakes. Results indicate that
temperature prediction is very insensitive to the particular wind speed
function employed. Based on this conclusion, a direct and straightforward
method which utilizes the model, but completely independent of wind speed
function, is proposed to calculate evaporation from large, natural bodies
of water. The method is applied to predict a daily evaporative heat flux
from Lake Belews, N.C. The results agree well with the field data
available.
INTRODUCTION
Evaporation from large bodies of water such as lakes is a topic of much
current interest. Despite a large amount of work regarding the prediction
of evaporation, the current state of practical predictions unfortunately
is not completely satisfactory. Many predictive techniques assume that the
evaporative flux is proportional to the vapor pressure difference between
the water surface and the air, that is qe = f(es - ea). The proportionality
coefficient is usually a function of the wind speed only. Many different
wind speed functions, f, have been proposed; these vary surprisingly in
functional form and more importantly, in value. In the present work we
first study the effect of wind speed function on the temperature profile of
the body of water. Different wind speed functions are used to compute
surface heat exchange coefficients and surface equilibrium temperatures.
Such results are used in a previously reported analytical model. [1] to
predict the seasonal variation of temperature distribution in a natural
stratified lake (no artificial heat load). It is found that lake tempera-
ture prediction is quite insensitive to the ultimate choice of wind speed
function. This conclusion leads to the second portion of the study in
663
-------
which a direct energy balance method is proposed to calculate evaporation
from bodies of water.
ANALYSIS
Water Temperature
The analytical model used here has been presented in earlier papers [1,2]
and verified with the field data from Cayuga Lake, New York. The model has
been developed for the prediction of the seasonal variation of the tempera-
ture distribution in large stratified bodies of water. The analysis used
in developing the model will be briefly outlined. The two-layer model
consists of:
(i) A well-mixed upper layer in the region 0 <_ z <_ h where the vertical
temperature distribution is considered uniform and taken as Ts.
(ii) A lower layer in the region h £ z £ H where the temperature varies
from Ts at z = h to a constant value TH at the bottom of the lake
where z = H.
The governing equations are taken as
pC — = - -— (q + q ) in 0 ^ z _< H (1)
where
T = T (t) in 0 < z < h(t) (la)
s — —
T = T(z,t) in h(t) £ z <_ H (Ib)
with the boundary conditions
T = Tu at z = H (Ic)
hi
q=q =0 at z=H (Id)
In addition the net heat transfer flux at the surface, q is represented
in the form [3]
qs = K'(Te-Tg) (le)
Here T is the temperature, t is the time, q and qr are the turbulent and
radiative heat fluxes respectively, z is the vertical coordinate measured
from the surface, p is the density, Cp is the specific heat, K' is the
surface heat exchange coefficient and Tg is water surface temperature.
The annual variation of the equilibrium temperature TC, is represented [4]
as — ~ • f
where T is an average value, ATe is half of the annual variation,
664
-------
ft = 2ir/365 day and depends upon the conditions from which the computa-
tions begin.
A dimensionless temperature 8(n) is defined for the lower layer as
T (t) -T(t)
8(n) = T f^ T , in h(t) < z < H (3)
Tg(t)-TH - -
where
n = I—TTTT , in h(t) < z < H
The temperature profile 6 is represented by
e(n) = 3n- 3n2-n3 , in 0 <_ n l 1 (5)
which satisfies the boundary conditions
2
6 = 0 , ^-| = 0 at n = 0 (5a)
dn
2
6(n) =1,^- = 0 and ^-|= 0 at n = 1 (5b)
dn
clearly if T (t) and h(t) are known the corresponding r\ and 6(n) are de-
termined from Eqs . (4) and (5) respectively, and the temperature distribu-
tion T(t,z) in the lower layer from Eq. (3).
Two equations that are needed for the determination of Tg(t) and h(t) are
then derived from the energy Eq. (1) . The turbulent heat transfer q can be
related [4,5,6] to the wind stress acting on the water surface by making
use of the fact that thermal stratification in a lake acts as a barrier to
mixing while the wind stress creates turbulence that acts against the
buoyancy gradient. Therefore, a mechanical energy balance in the water
relates the kinetic energy input from the wind directly to the transforma-
tion of the potential energy into kinetic energy by convection within the
layer if the turbulent energy dissipation due to viscosity is neglected;
the kinetic energy input into the water is then related to the wind stress
at the water surface [5,6]. With an analysis based on these considerations
it can be shown that the turbulent heat flux q is related to the wind
stress TS at the surface by [4,5-8]
H n ~ W*
/ -7*- dz = f— (6)
where W* = /T /p is the friction velocity, 6 is the coefficient of vol-
umetric expansion of water, g is the gravitational acceleration and i|i is
von Karman constant (4» = 0.4). The determination of the radiative heat
flux, q , however, requires the solution of the equation of radiative trans-
fer over the entire depth of the lake. The radiation part of the problem
665
-------
to account for the bulk heating of the water due to the penetration of the
solar radiation is treated by considering a plane parallel, absorbing, emit-
ting, isotropically scattering gray medium with azimuthal symmetry. The
P^-approximation for the spherical harmonics method is used to solve the
radiation problem. In this method the equation of radiative transfer takes
the form [9]
2
^-f - K2(T) = - 4K2 aT(T,t) in 0 < T < TQ (7)
dT
where „
K = 3(l-co)
to is the single scattering albedo, T = 3z is the optical variable, a is the
Stefan Boltzmann constant and T(r,t) is the temperature distribution in the
lake. Once the function G(T) is known from the solution of Eq . (7) subject
to appropriate boundary conditions, the net radiative heat flux qr(x) is
determined from
(8)
For most lakes the source term on the right hand side of Eq . (8) is very
small compared to the solar radiation energy incident on the lake surface.
Then Eq . (7) is simplified as
_ K2G(i) =0 in 0 < T <_ T (9)
_ _ Q
dr
The boundary conditions for this equation are established assuming that the
solar radiation incident on the lake surface is specified and that no radi-
ation is coming from the bottom of the lake. With this consideration the
boundary conditions for Eq. (9) are taken in the Marshak boundary condition
approximation as [9]
G(T) - — - = 4Tr[I + AI sin (fit + cj> ' ) ] , T = 0 (lOa)
G(T) + -~ = 0 , T = TQ (lOb)
The boundary condition (lOa) assumes that the annual variation of the in-
tensity I of the solar radiation incident on the water surface is specified
[10,17] as
I + AI sin (fit + <{>')
where I is an average value and AI is half of the annual variation of the
solar radiation intensity, fi = 2ir/365 day~l and the value of ' depends on
the conditions at the start of computations.
The solution of Eq. (9) subject to the boundary conditions (10) is
straightforward. Knowing G(T), the net radiative heat flux qr(T) is ob-
tained from Eq . (8) as
666
-------
(
—
^ 7rK[I + AI sin
- - - -
j K cosh(KTQ)+(l + | K )sinh(KTo)
{[cosh(KTQ) + - K sinh(KTo)]cosh(K-r)-
- [sinh(KTQ) +- K cosh(KTQ)]sinh(KT)} (11)
Noting that T = 3z and T = 3H we can write
H .[I + AI sin
; q dz = - - - - - - -- (12)
o gK + (-^ + -j K )g tanh(KBH)
Taking zeroth* and first** moments of Eq. (1) over the entire depth of the
lake and making use of Eqs. (le) , (3), (4), (5), (11) and (12) and after
some manipulation the desired two equations respectively become,
dT ,
(H-aH+ah) ^f+a^-V ^ = JL_ (TS - Tg)
dh W*
[2(a-Y>h+2(Y^)H].(Ts-TH) jj = ^
.[I+AI sin (- t-M,')l.{tan(KBH)+f K[l- c
+ -- - --- 3 - £ — o ------- ( }
pC L3K+ (|- + ^ K )6 tanh(KBH)
= 0.75
where u ~ J
o
and Y = / n0 dn = 0.45
o
* Integrate Eq. (1) from z = 0 to z = H.
** Multiply Eq. (1) by z and then integrate from z = 0 to z = H.
667
-------
Equations (13 and (14) provide two coupled, first order nonlinear ordinary
differential equations for the determination of the temperature Ts(t) in
the upper layer and the depth h(t) of the thermocline. A computer program
based on a Runge Kutta method was developed to solve those two equations
numerically. Then the temperature distribution is determined by Eqs. (3),
(4) and (5) and the net heat flux, qs at the surface by Eq. (le).
Equilibrium Temperature and Heat Exchange Coefficient
It is obvious from the above analysis that expressions for the equilibrium
temperature Te and the heat exchange coefficient K1 are essential to the
prediction of water temperature and the net heat flux at the surface.
Various techniques for calculating Te and K' have been presented in the
literature and most of the approaches are similar. The procedure given by
Ryan and Harleman [11] is used to yield the following expressions.
(q +n ) + f(bT,+0.142 T )- 73.3
T = sw ^iw7 \ d a' o (15)
e 1.30+ f(b +0.142)
K1 = 1.30+f(b +O.U2) , cal/m2/s/°C (16)
where q is the short wave solar radiative heat flux specified from
sw
meteorological data or calculated [11]. The long wave solar radiative
heat flux q and the constant b are respectively expressed as [11],
q .. = (1.24x 10~13)(T^ + 273)6 (1 + 0.172 C2) , cal/m2/sec (17)
b = 134,000 exp(17.62. 1300) f ^ (18a)
(T*) T* C
U4,uuu /. ., ,,, 5300, nimii£ M OKI
b = exp(17.62 - ) , ~~ klob)
s (T*)2 T*
s s
T* = -S 0 d + 273 , °K
°K
and Tj are dry bulb and dew point temperatures, C is the cloud cover in
tenths and f is a specified wind speed function discussed in the following
te
section
668
-------
Effect of Wind Speed Function on Predicted Water Temperatures :
The wind speed function f, is usually experimentally derived and taken as
dependent on the wind speed W, over the water surface. A number of wind
speed functions have been proposed by several investigators [11, 12, 31]
and all yield somewhat different values of Tg and K' . The following, chosen
for their widely differing forms, are the particular functions used here
f = 0.91 W (ref. 11)
f = 2.20 + 0.11 W2 (ref. 12)
f = 0.78 (e - e)+- (ref. 13)
2
In these expressions, f has the units of cal/s/m mmHg, W is the wind speed
in m/s measured at 8 meters, and (Re^ ) is the air Reynolds number based on
fetch L and the free stream velocity W.
As a first investigation, the analytical model described above for the
prediction of water temperature [1,2] was used to perform a sensitivity
analysis of the effect of the choice of wind speed function on the ultimate
prediction of lake temperature (coupled through K' and Te) . Monthly average,
meteorological data at Greensboro, N.C. for the period 1941 - 1970, along
with Ryan, Brady and Goodling wind speed functions [11, 12, 13] were used
in the procedure above to calculate the equilibrium, dry bulb, dew point
temperatures, the incident solar radiation, and the surface heat exchange
coefficient. These were then accurately expressed in the following simple
forms [4, 10],
Te = Te + ATe sin (fit + cf^) (20)
K1 = K1 -f AK' sin (fit + 4) (23)
I =1 + AI sin (fit + 4>5) (24)
where fe , K' , fd , and I are average values, AT£ , AK1 , ATa , and AI
are half of the annual variation, fi = 2rr/365, t is the time in days and
§1 > 'f'o > 'fo > $r > 'f'c; are Pnase angles dependent upon the conditions from
which the computations begin.
Three computations, each corresponding to a different wind speed function,
were presented for input conditions that correspond to Lake Belews , N.C.
with input data taken as
First run: „
Te = 16.21 + 13.45 sin (— ^ t - 0.994) , °C
K' = 293.10 + 86.67 sin (-~ t - 0.762) , kcal/m2/day/°C
669
-------
Second run:
Third run:
T = 16.06 + 13.22 sin
K1 = 299.77 + 94.13 sin
t - 0.999) , C
t - 0.793) , kcal/m /day/°C
T = 15.97 + 12.83 sin
e
t - 0.985) , C
K1 = 321.92 + 144.58 sin C^nf t - 0.815) , kcal/m /day/°C
The dry bulb temperature, the dew point temperature and the incident solar
radiation are
T =14.50 + 10.69 sin
a
t - 1.141) ,
T, = 8.19 + 10.56 sin (-~ t - 0.683) , °C
d
I = 3645.73 + 2058.86 sin
t - 0.683) , kcal/m /day
The semi-emperical relation between the wind stress Ts at the surface and
the wind speed given by Munk and Anderson [14] was used to evaluate the
friction velocity W*. The minimum temperature during spring homothermy was
assumed to be 5.83°C and calculations were started for this temperature on
Julian day 45(t = 0). The absorption and scattering coefficients for water
and the particles in suspension were assumed to be 1.017 m and 2.06 m"1
respectively , and the depth of the lake was taken as 30.48 m. Figure (1)
shows a comparison of the predicted water temperatures corresponding to the
various wind speed functions. Examination of the values indicated that
temperature predictions are very insensitive to choice of wind speed func-
tion. Thus it is felt that in most situations, model predictions for tem-
perature in a non-heat loaded body of water will depend only very slightly
on the choice of f, a conclusion of importance to those researchers who
employ mathematical models for temperature analysis of large bodies of
water.
Calculation of Evaporation
Now after having demonstrated that lake temperature predictions are in-
sensitive to the wind speed function, we are in a position to propose a
straightforward method for calculating evaporative heat flux from an un-
heated lake without the necessity of selecting a particular form of f.
An energy balance for the water body yields
*sw
^br
(25)
670
-------
The conductive heat flux q is related to the evaporative heat flux, q by
[11],
qc = Bqe (26)
where the Bowen ratio B is defined as
T> _ /" * 1 • ( —\ ( 9 7 ~)
b Ts ~ Td
Combining Eqs. (25) and (26) yields the required equation to calculate eva-
porative heat flux.
q = (T^—) • (q + q - qb - q ) (28>
A first order expression of the Stefan-Boltzmann relation for back radiation
from the water surface is [11]
o
qbr = 73 + 1.3 Ts(°C) , cal/m /s (29)
The surface temperature Tg is determined by using the analytical model as
indicated in section 1., qs and q, are computed by Eqs. (le) and (17) and
q is specified from meteorological data or calculated as given in Ret". 11.
Using this method, daily evaporation rates were calculated for the previous
test case. The evaporations are plotted in Figure (2) and give an annual
evaporation of 1.18 m compared with values of 1.02 m and 0.96 m in Ref. 15
and 16 respectively. Such favorable agreement gives confidence in the use
of this technique for predicting evaporation rates from unheated large bodies
of water.
CONCLUSIONS
A previously developed [1] analytical model for lake temperature prediction
was used to evaluate the sensitivity of wind speed function choice on pre-
dicted temperatures. Results indicate that the predictions were very in-
sensitive to the particular wind speed function used. Next, this result
was utilized to allow evaporation calculations via the energy budget
approach. This technique is satisfactory only because the insensltivity
of calculated lake temperatures to wind speed function allows proper calcu-
lation of energy budget terms (such as long wave back radiation) which
depend on the water temperature. An application of this approach was pre-
sented and the agreement of predictions with field data is encouraging.
NOMENCLATURE
b, defined by Equation (18);
B, Bowen ratio defined by Equation (27);
C, cloud cover;
671
-------
q, , back radiative heat flux;
q , conductive heat flux;
q , evaporative heat flux;
£w, long wave solar radiative heat flux;
q , radiative heat flux incident on the water surface;
q , net heat flux at water surface;
s
q , short wave solar radiative heat flux;
w*, friction velocity= /r /p ;
S
z, vertical distance measured downward from the lake surface.
/
Greek Symbols:
1
a, / 6(n)dn = 0.75 ;
o
<5, ' coefficient of volumetric expansion of water;
41 > von Karman constant - 0.4;
3, extinction coefficient;
<(>, phase angle;
Y, /1n6dn = 0.45;
o
n» dimensionless variable defined by Equation (4)
u), single scattering albedo;
p, density of water;
a, Stefan-Boltzmann constant;
T, optical variable = 3z;
T , optical depth of the lake = 3H;
T , surface shear stress induced by the wind;
O
6(n), dimensionless temperature defined by Equation (3).
672
-------
C , specific heat;
e , saturated vapor pressure at the dry bulb temperature;
3.
e , saturated vapor pressure at the water surface temperature;
s
f, windspeed function;
g, acceleration due to gravity:
H, depth of lake
h(t), depth of upper layer;
I, intensity of solar radiation incident on the water surface;
I, average value of the solar radiation intensity;
Al, half of the annual variation of solar radiation intensity;
K', heat exchange coefficient at water surface;
K1, average value of the heat exchange coefficient;
AK', half of the annual variation of heat exchange coefficient;
^rt, L-- time;
T dry bulb temperature;
cL
T(z,t), temperature of the lower layer;
T , average value of the dry bulb temperature;
3.
AT , half of the annual variation of dry bulb temperature;
3.
T,, dew point temperature;
AT,, half of the annual variation of dew point temperature;
T^, equilibrium temperature defined by Equation (2);
T , an average value of the equilibrium temperature;
AT , half of the annual variation of the equilibrium temperature;
TH, temperature at the bottom of the lake;
T (t), temperature of the upper layer (epilimnion);
S
q, turbulent heat flux;
673
-------
REFERENCES
1. A. M. Mitry and M. N. Ozisik, "One-Dimensional Model for Seasonal
Variation of Temperature Distribution in Stratified Lakes," Inter-
national Journal of Heat and Mass Transfer, JL9, 201-205 1976.
2. A. M. Mitry and M. N. Ozisik, "A Two-Dimensional Time Dependent Model
for Seasonal Variation of Temperature Distribution in Stratified
Lakes," Letter in Heat and Mass Transfer, _3, 475-484 1976.
3. J. E. Edinger and J. C. Geyer, "Heat Exchange in the Environment,"
Sanitary Engineering and Water Resources Report, Johns Hopkins
University, Baltimore, Maryland 1967.
4. T. R. Sundaram and R. G. Rehm, "Formation and Maintenance of
Thermoclines in Stratified Lakes Including the Effect of Plant
Thermal Discharges," AIAA Paper, No. 70-238, 1970.
5. T. Y. Li, "Formation of Thermocline in Great Lakes," Paper presented
at the 13th Conference on Great Lakes Research, Buffalo, New York,
1970.
6. E. B. Kraus and J. S. Turner, "A One-Dimensional Model for the
Seasonal Thermocline, II. The General Theory and its Consequences,"
Tellus, Ij? JCJO, 98-105 1967.
7. 0. M. Phillips, The Dynamics of Upper Ocean. pp. 198-243 Cambridge
University Press, Cambridge, 1966.
8. A. S. Monin and M. M. Obukhov, "Basic Regularity in Turbulent
Mixing in The Surface Layer of the Atmosphere," U.S.S.R. Acad.
Sci. Works Geophys. Met. No. 24, 163, 1954.
9. M. N. Ozisik, Radiative Transfer. John Wiley, New York, 1973.
10. Unpublished field data, Lake Belews, N.C., Environmental Section,
Duke Power Company, Charlotte, N.C., 1973.
11. P. J. Ryan and D. R. F. Harleman, "Analytical and Experimental
Study of Transient Cooling Pond Behavior," Ralph M. Parsons Laboratory,
Dept. of Civil Engineering, Report No. 161, Massachusetts Institute of
Technology, Cambridge, Mass., Jan. 1973.
12. Edinger, J. E., D. K. Brady, and J. C. Geyer, "Heat Exchange and
Transport in the Environment," Report No. 14, Electric Power Research
Institute, Research Project RP-49, Johns Hopkins University,
Baltimore, Maryland, November, 1974.
13. J. S. Goodling, B. L. Sill, and W. J. McCabe, "An Evaporation
Equation for an Open Body of Water Exposed to the Atmosphere,"
Water Resources Bulletin, Vol. .12, No. 4, August, 1976.
674
-------
14. W. H. Munk and E. R. Anderson, "Notes on the Theory of Thermocline,"
Journal of Marine Research, ]_, 276-295, 1948.
15. J. J. Geraghty, D. W. Miller, F. Van der Leeden, and F. L. Troise,
Water Atlas of the United States, Water Information Center Publication,
Port Washington, K.Y., 1973.
16. W. L. Yonts, G. L. Giese, and E. F. Hubbard, "Evaporation From
Lake Michie, North Carolina, 1961-1971," U.S. Geological Survey
Water-Resource Investigation, 38-73, 1974.
17. T. R. Sundaram, C. C. Eastbrook, K. Piech and G. Rudinger, "An
Investigation of the Physical Effects of Thermal Discharges into
Cayuga Lake," Report VT-2616-0-2, Cornell Aeronautical Laboratory,
Buffalo, New York, 1969.
675
-------
25 -
MEASURED (BEF. 10]
—"— COMPUTED
120 180
TIME , DAYS
360
Fig. 1 Predicted surface water temperatures
using three different wind speed func-
tions (refs. 4, 5, 6) as compared with
measurements.
Fig. 2 Calculated evaporative heat flux
for Lake Belews, North Carolina.
676
-------
Second Conference
on
Waste Heat Management and Utilization
WORKING SESSION Wl - MANAGEMENT AND UTILIZATION
December 5, 1978
Co-chairmen: Theodore G. Brna, U.S. Environmental Protection Agency
John Neal, U.S. Department of Energy
This session was divided into two parts: low grade (Part 1) and high
grade waste heat (Part II). Low grade waste heat was defined to apply to 93°C
(200°F) and lower temperatures, while the high grade concerned waste heat
available at temperatures above 93°C (200°F). Consequently, separate work
session summaries, prepared by the co-chairmen in the order listed above, are
presented below for these classifications.
PART I - Low Grade Waste Heat
The beneficial uses of this quality of waste heat include hot water
district heating (with and without augmentation by heat pumps), agriculture,
and aquaculture. Waste heat suppliers may be industries and power plants,
while users could encompass the industrial, commercial and residential sectors.
The major constraint to low grade waste heat utilization was felt to be
the lack of favorable economics. Tt s, demonstrations of waste heat applica-
tions which are profitable and independent of government subsidies are needed.
One mechanism for progressing toward successful demonstrations would be govern-
ment-guaranteed loans, such as for waste heat aquaculture.
A suggestion that consideration of waste heat utilization alternatives be
required as part of the permitting/licensing process for power plants received
little comment. One utility representative opposed such a requirement on
economic and scheduling grounds.
677
-------
Relative to waste heat from steam-electric generating plants, the use of
this resource as it is normally available seemed to be favored. Such use was
seen as highly site dependent. Modification of plant operation to accommodate
greater utilization of waste heat was viewed by some as enhancing a secondary
benefit while lowering electrical output, the primary product. Use of this
cogeneration concept could also adversely affect plant reliability because of
two different outputs with variable demands.
Governmental regulation of fuel prices was pointed out to be an inhibitor
to the use of heat pumps. Deregulation of fuel prices was suggested as neces-
sary for making heat pumps economically feasible in low grade waste heat
applications, particularly district heating.
Further consideration of governmental regulation concerned several areas.
The concensus of those present was that environmental regulations should
provide offsets for the beneficial uses of waste heat. One approach would be
to permit exceptions to environmental standards to encourage the overall
reduction of pollutants in a region via a less than proportionate increase in
pollutants in the locale of the waste heat source. State rate regulatory
agencies in not recognizing waste heat utilization as an energy conservation
measure impede beneficial uses of waste heat. Some of the utility participants
indicated that the support of waste heat research and development projects by
utilities may adversely impact requests by utilities to these agencies as
these projects are non-income generating and non-electrical generation activi-
ties.
678
-------
PART II - High Grade Waste Heat
High Grade Uasto Heat Recovery was in general defined to include
process heat cogcneration as opposed to the low grade heat definitisn
which emphasized both beneficial uses cf waste heat and district heeting-
Concern was expressed by various workshop participants that district
heating should be categorized in high grade heat and separated for the
purpose of management and planning, from the broader area of beneficial
uses of waste heat. It was acknowledged by the chairman that for future
meetings or policy decisions on this subject, it would be better tc
consider both process heat industrial cogeneration and residential/
comaercial district heat-ing cogeneration together.
Three problem areas, or issues, vere brought up for discussion by
the group. These were: 1) should cogcnerators be exempted from coal
conversion, 2) should industrial coge.r.erators be regulated as utilities,
and 3) what is the best way of permitting excess power to be sold back
to the utilities from industrial cogenerators?
With regard to the first of these, concern was expressed that if
exemption were granted unilaterally for cogeneration, many units "called
cogeneration" would be installed just to promote the exemption. As a
result the nation would not necessarily benefit from the potential
national savings afforded by well planned cogeneration systems. After
some discussion it was concluded by the chairman that great care should
be taken in formulating the language for any such exemption since it
is one of our primary goals in energy planning to move toward the use
of more readily available domestic resources, such as coal and uranium.
Secondly, the group discussed the issue of regulation as to whether
cogenerators should be regulated as utilities. Utility members of the
679
-------
nudlcncc were very specific in their views that PUC Control should
not extend to process steam. Examples were also given where co-
generators were considered municipal utilities and therefore not
subject to TUG regulation.
The third topic concerned the best way of permitting excess
power to be sold back by a cogenerator to the utility grid. In
general the utility members of the group thought this would be a
minimal problem. They were of the opinion that the net transfer
of electricity would remain predominately from the utility to the
industrial. It was more or less unanimously felt that the utility
Bust remain in control of dispatching power. Eonneville Power in
the Pacific Northwest cited their experience with cogeneration.
They indicated that utility systems made up of a significant percen-
tage of cogeneration, such as theirs, could be dispatched, power sold
back and forth, etc., in a very acceptable manner. It was noted that
Bonneville had large hydro capacity which eased the problem by in
effect providing storage.
A more important problem than ownership or sellback was judged
by the group to be the issue of standby power and its attendant
charges. Several examples were given ranging from increased costs
for standby due to cogenerators, to reduced cost of standby because
of multiple cogenerators creating greater redundancy.
The conclusive remarks which seemed to receive overall endorsement
by the group were to the effect that arrangements cen be made without
additional government or local regulations.
Additional discussions were then held on how the government could
help. It was suggested by a group member who is involved now in co-
generation that the EPA could help the nost by allowing an overall
fuel efficiency credit in cogeneration emissions regulations. The entire
group endorsed this concluding remark.
680
-------
Second Conference
on
Waste Heat Management and Utilization
WORKING SESSION W2
ENVIRONMENTAL EFFECTS
December 5, 1979
Co-Chairmen: C. Coutant, Oak Ridge National Laboratory,
Oak Ridge, Tennessee
R. Wilcox, Florida Power and Light Company,
Miami, Florida
The working session on environmental effects were co-chaired
by J. Ross Wilcox and Dan McKenzie on Tuesday, December 5, 1978.
Approximately 41 persons were in attendance during all or part
of the discussion period.
There were lively discussions centering around environmental
effects of cooling water intake and thermal discharge systems.
Most people felt that entrainment of zooplankton and phytoplankton
was a non-problem; however, more review is required to determine
the effects of entrainment on icythoplankton. Concerns were raised
about the suitability of any baseline data set so that a natural
perturbation or seasonal variations could be distinguished from
a man-made perturbation. The question of mitigation was discussed
as a means to soften an environmental impact. Many people were
concerned about the continued standardization of techniques because
one data set may be difficult or impossible to compare with another
data set due to sampling gear Differences. Suppression of data
and its exchange was of concern to some individuals, but others
countered that exchange of data among professionals is good and
will improve when data banks are computerized and collected under
the auspices of two or three national data centers.
681
-------
Second Conference
on
Waste Heat Management and Utilization
WORKING SESSION W3
MATHEMATICAL MODELLING
December 5, 1979
Co-Chairmen: D.R.F. Harleman, M.I.T., Cambridge, Massachusetts
S. Sengupta, University of Miami, Coral Gables,
Florida
Mathematical modelling is an attractive tool for predictive
and diagnostic analyses of environmental effects of waste heat.
The session was attended by approximately 30 people. The dis-
cussions were primarily related to physical effects and numerical
techniques. However, some concerns regarding the state-of-the
art in biological modelling were expressed. A summary of the
discussions is presented.
1. The merits of rigid-lid and free surface numerical-differen-
tial models for aquatic discharges were discussed. The rigid-
lid models are appropriate for cooling lakes whereas free-
surface models are more suited to coastal and estuarine domains.
The discussion was in relation to three-dimensional models.
More effort in calibration of 3-D models should be directed
as computational costs become less prohibitive with more effi-
cient numerical techniques and better computers.
2. Numerical matching between "near-field" and "far-field" regions
of thermal discharges was perceived as a problem that needs
more extensive investigation. Complete field models with hori-
zontal stretching and variable diffusion coefficients is a di-
rection of research that might avoid the problem of matching.
3- Open boundary conditions is a problem for almost all classes
of mathematical models. The present techniques rely quite
extensively on measured knowledge regar ding flow fields and
temperature and/or salinity distributions. Sensitivity analy-
sis of existing models as a function of open boundary condi-
tions is an useful direction for further research.
4. The state of the art in cooling tower plume models is somewhat
more primitive than aquatic discharges. Integral models are
being developed to include more complex physical effects.
Basic research in determination of entrainment coefficients,
mixing mechanisms and condensation processes is essential.
5. Lack of reliable data bases for verification of cooling tower
plume models is also a problem. Extensive data bases under
682
-------
diverse meteorological conditions are needed for evaluation
of existing models and to provide bases for future model
development.
6. The gap between the groups working in mathematical modelling
of physical effects and biologists (the user community for
physical data) is wide. Multidisciplinary demonstration pro-
jects may be a route to enhance greater exchange of infor-
mation. It will also develop greater appreciation for cross-
cisciplinary needs.
683
-------
Second Conference
on
Waste Heat Management and Utilization
WORKING SESSION W4
HEAT TRANSFER PROBLEMS IN WASTE HEAT
MANAGEMENT AND UTILIZATION
December 5, 1978
Co-Chairmen: W. Aung, National Science Foundation
F. K. Moore, Cornell University
Research in heat transfer and related areas in necessary for
the economic realization of various schemes for waste heat manage-
ment and utilization. In recent years a number of symposia have
been held in which heat transfer problems in waste heat technolo-
gies have received attention. These discussions were continued
in one of the four workshop sessions held during the present con-
ference. We list beljw a summary of the perceived research needs
in heat transfer and related tupi. :s which if carried out would con-
tribute towards effective waste heat management or utilization.
Our sources are the recommendations provided at the latest workship
session. In addition we have also included some of the recommen-
dations made at previous meetings of a similar nature.
(1) There is a need to develop and understand the behavior of new
heat exchanger surfaces for cooling tower applications. Cost
and size reductions are important considerations in cooling
tower designs and efforts in these directions are now limited
by the existing heat exchanger technology. This field is in
need of innovative new design concepts. New surfaces that
are developed should be characterized not only in terms of
thermal performance but also in terms of fluid flow fields
and pressure drops.
(2) More accurate information is needed on near-field plume
behavior including the interaction of multiple plumes,
the effect of atmospheric stratification, the factors
leading to re-entrainment, etc. Better understanding in
this field could lead to substantial savings in real estate
and in transmission costs by making it possible to position
towers closer together.
(3) Methods of achieving flow uniformity over heat exchangers
are needed especially for dry cooling towers. These methods
should account for the movement of the ambient air since the
dynamic pressure there is often of the same order of magni-
tude as the pressure drop across the heat exchanger.
(4) There is a need to control and eliminate regions of separated
684
-------
flow in cooling towers, sometimes evidenced as "cold inflow".
(5) In relation to soil warming in waste heat utilization,
improved methods for characterizing the transport of heat
and moisture are needed. Laboratory experiments and theo-
retical simulation studies are both needed. Theoretical
studies should not only focus on modelling through the use
of the full transport equations but also son developing sim-
plified mathematical models that are capable of elucidating
important mechanisms.
(6) The fouling properties of heat exchanger surfaces in cooling
towers are in need of further understanding.
(7) Novel experimental techniques should be exploited to provide
detailed heat transfer information on new and existing surfaces
of complicated design. Promising experimental methods include
those based on the analogy of heat and mass transfer, such as
the naphthalene sublimation technique used in the past for
heat transfer in complex geometries.
(8) Simulation studies are needed to facilitate power plant siting
and design that include waste heat utilization and management
options.
(9) Research should be conducted to identify new fluids for
refrigeration or heat engine application involving low temper-
ature thermal energy.
(10) Studies are needed to convert low temperature thermal energy
into a form more suitable for practical utilization.
(11) Research is needed to clarify the limitations of extended
Reynolds analogies among heat, momentum and mass transfer,
especially in turbulent flow. Current basic turbulence
studies of "scalar transport" should be encouraged and
applied to heat-exchange problems.
(12) In general, increased understanding is needed in the areas
of thermal discharges, spray cooling and in transport pheno-
mena in cooling ponds.
685
-------
THE CHALK POINT DYE TRACER STUDY: VALIDATION OF MODELS
AND ANALYSIS OF FIELD DATA
A. J. Policastro*
W. E. -Dunn0
M. L. Breig*
J. P. Ziebarth0
ABSTRACT
Predictions of ten models are compared with field data talon during the
Chalk Point dye tracer study of June 1977. The ESC/Schrecker,
Hosler-Pena-Pena, and Wigley-Slawson models compared most favorably with
the deposition data from the cooling-tower alone and are generally
within the error bounds of the data. Most models predict larger drop
diameters at deposition than were measured. No model predicted each of
the deposition oarameters consistently within a factor of three.
Predictions of stack deposition compared rather poorly with the stack
deposition data probably due to the lack of good information on exit
conditions.
A comparison of Johns Hopkins University (JKU) and Environmental Systems
Corporation (ESC) ground-level drift data showed that the JHU data had
larger drop counts in both the smallest and largest drop size ranges yet
both sets of data agreed quite well in the intermediate drop size range.
The JHU methodology appears superior since their data were more
internally consistent and their technique of using large sensitive paper
samplers and counting all drops on the paper yields a greater
statistical accuracy.
INTRODUCTION
Drift refers to the small droplets of liquid water released from a
cooling tower along with the warm, moist plume. These droplets, ranging
in size from a few to more than 1000 microns in diameter, are
transported through the atmosphere eventually evaporating totally or
being deposited on the ground. If the droplets contain large
concentrations of dissolved solids, as is particularly the case when
brackish cooling water is used, then the drift deposition may damage
vegetation and/or accelerate the corrosion and deterioration of
structures.
* T-._
Engineer, Div. of Environmental Impact Studies, Argonne National Lab,
DAsst. Professor, Dept. of Mech. & Ind. Engr. , Univ. of 111., Urbana.
•Visiting Scientist, Div of Environmental Impact Studies, Argonne Nat,
Lab.; Perm. Add.: Dept. of Physics, Eastern 111. Univ., Charleston.
OSngineer, Div. of Environmental Impact Studies, Argonne National Lab.
686
-------
Therefore, predictions of anticipated drift-deposition rates are
essential to an informed estimate of the environmental impact of a plant
for which cooling towers are planned.
Once emitted from the tower, a drift drop moves under the combined
influences of gravity and the aerodynamic drag force produced by the
vector difference between the drop and local air velocities.
Simultaneously, the drop experiences both heat and mass transfer. As a
result, the drop temperature will approach the drop wet-bulb temperature
and evaporation will occur as long as the vapor pressure at the drop
surface exceeds that of the local ambient. For a drop containing salt,
evaporation will increase the concentration within the drop and thus
lower the vapor pressure at the drop's surface. The salt concentration
will continue to increase until either (a) the droplet vapor pressure
exactly equals that of the local ambient after which evaporation will
cease or (b) the salt becomes saturated within the drop after which salt
particles will begin to precipitate out as evaporation proceeds. In the
latter case, the drop will eventually become a dry particle, although it
may strike the ground before reaching its final state. The purpose of a
drift model, then, is to predict the number, size, and character of
drops and/or particles striking the ground at any riven location with
respect to the emitting tower.
Numerous mathematical models have been formulated to predict drift
plumes and drift-deposition patterns. Although each of these models has
a number of theoretical limitations, good quality field data have been
lacking to determine the limits of reliability of these models. Field
data taken at the Chalk Point Power Plant in 1975 and 1975 by the
Environmental Systems Corporation suffered from several inherent
deficiencies: ground samplers were too small in size and few in number,
no separation of cooling tower and stack drift was made, etc. Those
data provided a rough test of the models, yet the limitations of those
data did not allow definitive conclusions to be made about the field
performance of the models tested.
The field data taken at Chalk Point in June, 1977 by the Environmental
Systems Corporation <"SSC) [11 and independently by the Johns Hopkins
University (JHU) r2,3l represent a significant improvement in the data
collection methods and the culmination of more than three years of
experience in drift data collection. The data,, taken as a whole, are of
good quality and sufficient to provide a true test of the models'
capability. In fact, these data are presently the only good-quality
field data on drift deposition available in the literature. The purpose
of this paper then is to evaluate the performance of 10 drift models
T4-111 with respect to these data and to provide an analysis of the data
themselves to uncover special trends. Moreover, the ground-level data
taken simultaneously by the two groups CESC and JHU) will be
intercompared as a test of their measurement and data reduction methods.
It is important that such data be studied in detail due to the
uniqueness of these good-quality data as well as the difficulty and
expense of acquiring new data.
687
-------
It must be noted that while these data are the best available and were
obtained only through a very carefully executed measurement program, the
data were obtained at only two radial distances from the tower. Thus
the data encompass only one of several possible regimes of droolet
behavior.
THE FIELD EXPERIMENT
The Chalk Point Unit Mo. 3 cooling tower and stack effluent scrubber
produce salt water drift because of the saline Patuxent River water used
for the cooling tower circulating water and the stack particulate
scrubbing agent. Previous drift measurements at Chalk Point have used
sodium as a tracer and consequently separation of cooling tower and
stack drift was not possible. To provide a positive identification of
the drift deposition from the individual sources, JHU used a water
soluble fluorescent dye ^Rhodamine WT) as a tracer in the cooling tower
circulating water. The Dhotolytically unstable dye required that the
experiment be performed at night. The drift dye tracer experiment was
conducted during a four-hour period on June 15 and 17, 1977.
The instrumentation used by JHU consisted of 10.5 inch diameter modified
deposition funnels for sodium and dye concentration measurements and
10.5 inch diameter Millipore HA type filter papers for measurement of
total chloride and dyed drift droplet deposition. Three filter papers
per sampling station were used for the deposition measurement of all
water droplets (water sensitive filter paper), chloride containing
droplets (plain filter paper"! and dyed drift droplets 'plain filter
paper). A sketch of the sampler is shown in Fig. 1. The sampler
consisted of a po'st with rectangular and triangular brackets for holding
the funnel and sample bottle, and a filter paper holder plate with a can
type candle heater. Filter paper heaters were required because of night
time condensation which could affect the drop size measurements. The
filter papers were photographed for fluorescent droplets using
ultraviolet light. In this way, droplets deposited from the cooling
tower could be identified. The water sensitive filter papers were used
to define total drops depos-ited from all sources (stack and cooling
tower). A calibration curve for droplet sizes was used to relate drop
deposit size to falling drop size. The funnel samples were corrected to
a standard volume (after being washed with distilled water 1 and split
into two parts. One part was analyzed for sodium using an atomic
absorption soectrophotometer while the other part was concentrated by
boiling and analyzed for dye by fluorometry. The funnels could then
give sodium deposition rate from all sources (tower and stack) by
analyzing total sodium Content, of the sample. The funnels could also
determine the part contributed by the tower alone by pro rating the dye
deposited in the funnel to the ratio of the sodium to dye concentration
in the basin water.
Fig. 1 also shows the Chalk Point power plant area and the JHU array of
3 stations on the 0.5 km arc (40 m apart'l and 14 stations on the 1.0 km
688
-------
arc (40 m apart ^ . Each sampling station consisted of three samplers
(see Fig. 1) to ensure at least one good saraole in case of accidents or
contamination during sample collection and for good statistics. A total
of 25 sampling stations were used by JHU on the night of the dye test.
Each sampling station used during the experiment by JHU is identified
with a number.
A number of drift parameters were measured at ground level downwind of
the cooling tower by SSC. Typically, ESC uses four or five stations to
measure the following ground-level drift quantities.
1. Sodium concentration in the air fmicrograms-Na/m ) using a
rotating tungsten mesh.
2. Liquid droplet concentrati
(g-water/m ) using a rotatin
ration as a function of droplet size
ng a rotating sensitive paper disk.
3. Liquid droplet deposition flux as a function of droolet size
f kg-water/km -month"* using a stationary sensitive paper disk.
2
4. Sodium mass deposition flux (kg-Na/km -month"! using a
stationary funnel and bottle assembly.
The ESC sampling stations for the dye study are also located in Fig. 4
(denoted E1-E4). Some of the ESC ground-level stations were fixed in
location and thus received drift only when the wind was blowing in the
proper direction. Other stations were located beneath the cooling tower
plume, being moved as the wind direction changed. For the purpose of
model-data comparisons with the ESC data, we used the droplet number
deposition flux measurements obtained using sensitive paper disks fixed
to a petri dish and the sodium mass deposition flux obtained using the
stationary funnel and bottle assembly. In addition to the ground-level
measurements, source and ambient conditions were also measured by ESC.
Drift rates from the cooling tower were determined by ESC using an
instrument package suspended in a plane approximately 13.5 m below the
tower exit. The following measurements were made:
1. The drift droplet size spectrum was measured using sensitive
paper and with a device based on scattering of infrared laser
light (PILLS II-A, Particle Instrumentation by Laser Light
Scattering) .
2. The drift mineral mass flux was measured with a heated glass
bead isokinetic (IK) sampling system,
3. The updraft air velocity (from which droplet velocity was
determined) was measured using a Gill propeller-type
anemometer .
4. The dry-bulb and wet-bulb exit temperatures of the plume were
also measured.
689
-------
The IK system sampled continuously during the traverse and yielded the
sodium and magnesium mineral flux at the measurement plane. Updraft air-'
velocities were acquired and averaged for each point. Grab samples of
circulating water were also taken for chemical analysis of sodium and
magnesium content. These two cations, which are present in the highest
amounts in the water, were chosen as tracer elements for the IK
measurements. No source measurements were made for the stack however.
Ambient meteorological measurements were made using the Chalk Point 100
meter instrument tower which has wind and temperature instruments at
three levels (7 m, 50 m, and, 92 nO and dew point sensors at two. levels
(7 m and 92 m). Ten minute averages of dry bulb and dew point
temperature and wind speed were taken. Due to the location of the
meteorological tower on a hill, the 92 meter level on the meteorological
tower was at the same vertical elevation, as the cooling tower exit
plane. To supplement the,meteorological tower measurements, rawinsonde
flights were conducted at intervals of 1 hour by JHU in order to
establish the short-term history of diurnal stability characteristics.
Measurements of pressure 'elevation), dry-bulb temperature, relative
humidity, and wind speed ''and direction1* were made every 10 to 20 meters
vertically.
ANALYSIS OF FIELD DATA AMD COMPARISON OF JHU AMD ESC DATA
The published presentation r2,3"1 of the JHU data revealed several
interesting facts. A histogram plot of the total water and fluorescent
droplet size distributions for the approximate cooling tower plume
centerline sampling stations, 0.5 km/355 deg. and 1.0 km/350 deg.,
indicates a bimodal distribution Csee Fig. 2). One peak occurs at
about the 40-50 micron droplet size and the other between 200 to 400
microns. The second peak is expected from model calculations while the
first one is not. Meyer and Stanbro r2,31 suggest that the source of
these droplets is most probably blowoff from the cooling tower fill.
The droplet distribution data for the other 22 sampling stations in the
JHU net has yet to be reduced. Figure 3 presents the above droplet
distribution data as percent mass fraction. The smaller droplets with
their greater number contribute less than 1* to the total mass fraction.
Note also that the fluorescent droplet distibution. peak is separated
from the total water peak by approximately 30 microns. The shift in the
peaks between fluorescent and total drops is probably due to larger
droplets originating in the stack. Also shown in Fig. 3 is a
comparison of salt deposition contributions from the cooling tower and
stack at near centerline locations 0.5 km and 1.0 km downwind of the
tower. Mote that each distribution is nearly bell-shaped and due, we
believe, to the variation in wind direction with time during the
measurement campaign. Also, the distinction between the contributions
of the two sources is clearly seen at the 0.5 km distance and gets less
distinct further from the tower as may be seen by 'the comparison at the
1.0 km location.
69U
-------
Figure 4 shows the placement on the ground of the four ESC and the two
JHU samplers which have data reduced in the form of droplet size ranges.
J1 and J2 indicate the two samplers of JHU, and E1 through E4 represent
the locations of the appropriate ESC samplers.
The first parameter we studied for each of the six samplers was the drop
size spectrum measured at particular sampler locations. Figure 4 gives
the drop size distributions reported for the JHU and ESC data. The JHU
spectra are clearly bimodal with a large peak of small drops Cup to 100
microns) and a second peak of larger drops (approximately 250-280
microns). The ESC spectra also show bimodal tendencies, but the small
drop count is smaller for samplers E2, E3, and E4.
Figure 5 shows the game data replotted in terms of mass . distribution.
Here, we see that very little mass is contributed by drops less than 100
microns in diameter. The largest drops also contribute very little
except for ESC sampler E1 in which one drop contributed 8$ of the total
liquid mass. Problems with a few large drops contributing a significant
fraction of the mass were evident in the 1975 ESC data as well.
It is instructive to examine next the average drop size measured at each
of the S3C and JHU samplers. Defining an average drop size poses some
interesting questions as several alternatives are possible.
1 . Mass Mean Diameter - d,
d,™ =(Z C. d3 /I C ^1/3
MM i i i
where Ci is the number of drops in an interval and d. is the
corresponding drop diameter.
2. Mass Median Diameter - d
d is selected such that 50? of the total mass is contributed
by drops larger than d and 50? by"drops less than d.
3. Count Mean Diameter - d
4. Mass Peak Diameter - d^
d.^ is the diameter at which the greatest mass contribution
occurs.
691
-------
5. Count Peak Diameter - dcp
d is the droo diameter with the highest recorded count,
CP
Listed in Table 1 are the values of these characteristic diameters
computed from the JHU and ESC drop size distributions shown in Fig. 5.
The mass mean and mass median diameters are fairly representative of the
corresponding distribution with the mass mean being roughly UO to 50
microns smaller than the mass median. The mass peak diameter is
intermediate between these two. The count mean is much smaller
reflecting the large counts of small drops. The count oeak diameter is
not unique.
Among these, either the mass mean diameter or mass median diameter is
preferable; however, neither of these is totally satisfactory. The mass
mean diameter can be greatly affected by errors in the small drop data
(large count, small aiassl. In contrast, the mass median diameter is
sensitive to errors in the large drop data 'small count, large mass"i.
Since the greater uncertainty appears to be in the small drop counts for
the 1977 data, we have chosen to use the mass median diameter to
characterize these data.
Figure 5 shows how mass median diameter varies with distance from the
tower. A trend of decreasing drop size with increasing distance from
the tower -is evident, but Sampler E1! does not follow the trend. This
may be due to a greater influence of the stack. Recall that the JHU
investigators found that the stack distribution has a greater number of
larger drops. A.s shown in Fig. H, Sampler E4 experiences a stronger
stack influence than do the other samplers.
A fourth test of the data concerns the consistency between the four
independent measurements: sodium deposition flux, liquid deposition
flux, sodium concentration and liquid concentration. We can calculate
from the data apparent droplet salt concentration and deposition
velocity.
1. Apparent Droplet Concentration
"DD
CD
Apparent concentration
from deposition data _
Apparent concentration
from concentration data
Sodium deposition flux
Liquid deposition flux
Sodium concentration
Liquid concentration
692
-------
Apparent Deposition Velocity
SD"
["Apparent velocity"
|_from sodium data
Apparent velocity
_from liquid data_
Sodium decosition flux
Sodium concentration
Liquid deposition flux
Liquid concentration
Table 1 summarizes the comparison of these calculated quantities. ''Note
that the basin-water salt concentration (for the tower) was 0.014 g/g.)
The agreement here is within a factor of 2 with one execution,
suggesting scene consistency among the ESC data. Also, the magnitudes
given are not unreasonable. Notably, C is consistently larger than
CDD, and VLD is consistently larger than V
a suitable explanation is presently lacking.
SD'
This may be fortuitous as
As it happens, Samplers J1 and E3 are within 25 meters of one another.
Thus, we nay compare almost directly the measurements obtained
independently by these two different groups. Figure 6 compares the
count and mass distributions as functions of drop diameter. The
following observations can be made. First, the JHU sampler shows a
greater droplet count both below about 100 microns and above about 300
microns, 'although agreement above 500 microns is good). Second, the
JHU mass distribution
clearly shifted toward greater diameters,
although agreement above 550 microns is good. Despite this discrepancy,
the mass median diameter computed from the JKU distribution is 400
microns whereas that computed 'from the ESC distribution is 336 microns,
which is less than a 25* difference. It is possible, although unlikely,
that the JKU sampler received a larger contribution of drops from the
stack than did the ESC sampler.
MODEL VALIDATION WITH JHU DATA
Critical reviews of the 10 models tested appear in References 12 and 13.
Described below are the major features of the methodology used to make
the model/data comparisons in this study.
1. Model predictions were made using the 10-minute averages of
meteorological conditions acquired at the time of the dye study
in order to better account for the variability of these
conditions on deposition predictions. Predictions were made
for each 10-minute period and the results summed over the
four-hour duration of the study.
2. A 15 degree sector was chosen over the more common 22 1/2
degree sector due to the short duration of the averaging
period.
693
-------
3. For nine models the 92-m level on the meteorological tower was
used co orovide the needed incut. For one model
-------
predictions of sodium deoosition rate are too large to fit on the scales
of Figures 7 and 3. Second, the models tend to underpredict sodium
deposition at the left and right end of the 0.5 km and 1.0 km arcs.
This underprediction may be due, in part, to our choice of a 15 degree
sector. A. larger angle for sector-averaging may partially mitigate the
discrepancy. Third, two predictions were made for both the
ESC/Schrecker and Wigley-Slawson models in order to study the effects of
changes in the input data on predictions. The second prediction of the
ESC/Schrecker model, labeled "ESC/ Schrecker (limited)", was made by
reducing the measured drop size spectrum from 25 to 3 intervals.
Clearly, this modification of the spectrum led to a significant
overprediction of sodium deposition in this case. The first prediction
of the Wigley-Slawson model, labeled "Wigley-Slawson ''profiles/", was
made using the full ambient profiles as recorded by radiosonde flights
with wind direction obtained from the meteorological tower. The second
prediction was made using the met-tower data alone as was done for each
of the other models. Here again, performance is degraded as the detail
of the input data is degraded.
A few models, notably ESC/Schrecker, Wolf II, and Wigley-Slawson appear
to be most accurate over the range of comparisons in Figs. 7 and 8.
Observations on the performance of individual models will now be
presented.
The ESC/Schrecker model ''full spectrum) is rather good in its prediction
of sodium deposition except at angles between 3^0 and 3^5 degrees on the
1 km arc, where the prediction is rather low. The predictions at 0.5 km
are excellent. However, the prediction of number drop deposition (Table
2) at 1 km from the tower is too small by a factor of 3- This
underprediction is carried through to the liquid mass deposition rate
which is also too small by a factor of about 3. The prediction of final
droplet diameter is quite good at both the 0.5 and 1.0 kin distances from
the tower.
The Wolf I and II model predictions are very similar at 0.5 km from the
tower since evaporation is rather insignificant due to the high ambient
relative humidity and the short time to deposition. Wolf II provides
excellent predictions of sodium deposition except between angles of 330
and 335 degrees where low predictions occur. A larger difference
between the predictions of the two Wolf models occurs at 1 km, where
Wolf II now predicts noticeable evaporation; the effect of evaporation
is to distribute the drift at the ground further downwind from the
tower. The Wolf II predictions of final droplet diameter and liquid
deposition rate give results that are low compared with the data
probably due to excess evaporation predicted owing to the omission of
solute effects in the model. Although the Wolf II predictions of sodium
deposition are quite good at 0.5 and 1 km from the tower, the Wolf I
model (which assumes no evaporation"1 overpredicts deposition.
The MRI model predicts sodium deposition reasonably well at both the 0.5
km and 1.0 km distance from the tower. However, the model underpredicts
695
-------
number droplet deposition flux by a factor of U at 0.5 km from the tower
and a factor of 5 at 1.0 km. No final droo size or liquid mass
deposition is computed since the model is based on the equilibrium
height conceot which does not allow the computation of the final state
of the drop. The model permits only two categories of relative
humidity, greater than 50* and less than or equal to 505. The case here
of high relative humidity, approximately 93*', is perhaps not well
represented by the formulas. The prediction of the Wigley-Slawson model
with full ambient profiles is, overall, superior to the prediction of
the model without profiles.
The Slinn I and II models were develooed to provide uoper and lower
bounds on deposition. Clearly they do so. The Slinn II model predicts
deposition .just beginning to occur at 1 km. The prediction at 0.5 km is
nearly zero. The Slinn I prediction for sodium deoosition varies
between a factor of 3 to 7 too large rsee Table 2^. Interestingly, the
Slinn I orediction of average diameter at deposition is too small
perhaps because the larger droplets have already deposited closer to the
tower.
The Hanna model underpredicts sodium deposition probably due to the
overprediction of evaporation in the model r 1 <4] . Predictions of number
drop deposition rate and liquid mass deposition rate are also too low.
The Hosler-Pena-Pena model has in our previous model/data comparisons
r12,13"i underpredicted salt deposition rates ''near the tower"1 but
usually provided larger values than predicted by the Hanna model. Here,
it does predict larger deposition rates than Manna's model and performs
quite well with the sodium ground flux data. The model, however,
continues to underpredict the number drop deposition flux, here by
factors of 2.5 and 3.5 at 0.5 and 1.0 km, respectively.
The Overcatnp-Israel model underpredicts sodium deposition flux at 0.5 km
from the tower. Tn addition, the deposition peak is shifted to the
right. There is underprediction also at 1 km but only slightly. There
is an underprediction in droplet number deposition rate and an
overprediction in droplet size. In total, there is a consequent
underprediction of liquid mass deposition flux by factors of 2.1 at 0.5
km and 2.7 at 1 km.
A few general comments should also be made. First, from Table U, the
models generally overpredict droplet diameter at deposition. Second,
the peak deposition for sodium predicted by the models is generally
coincident or nearly coinident with the observed peak along the two
arcs. Third, it should be recalled that the sodium flux predicted and
measured included droplets of all sizes, whereas, our droplet number
deposition flux, average diameter, and liquid mass deposition rate
include droplets only above 100 microns in size. We would expect the
observed sodium deposition rate to be slightly larger than the predicted
deposition rate since it includes some sodium coming from blow-off from
the tower fill. The 100 micron cutoff for other deposition quantities
696
-------
was set because it is difficult to accurately count drops less than this
value and also it eliminates most of the blow-off droplets which are not
considered by the models.
The models have also been run for the stack input data with results
given in Figures 9 and 10 and Table 3. Combined results of model
predictions from cooling tower and stack appear in Figs. 11 and 12 and
Table 4. Field data taken from the water sensitive paper were used for
comparison with model predictions. Some observations follow.
1. In the angular range fat 0.5 and 1.0 km distances^ where the
tower has a predominant effect, the models perform in a
reasonable manner. However, in the angular range 350 to 355
degrees, the stack contribution becomes important and the tower
contribution becomes insignificant (at 0.5 km). At 1.0 km, the
stack contribution is about 3-^ times the tower contribution.
From Figs. 11 and 12 and Table 4, the models overpredict by a
factor of 5-15 in the angular range of 350-355 degrees. The
poor comparison of models with stack plus tower data may be due
to the use of average stack parameters from the year before.
Among the unknowns for the stack exit were: (a) drop size
spectrum, (b^ liquid mass emission rate, rc^ drop concentration
at exit (we assumed saturated drops following ESC r15n, 0.26
g/g) , and (dl stack exit velocity and temperature.
2. The model predictions for the stack are quite consistent among
themselves. One of the reasons may be our assumption that the
drops are saturated with salt and evaporate only little out to
the deposition samplers.
3. The cooling tower contribution to total deposition can be
easily distinguished from the stack contribution at the 0.5 kin
distance but not as easily for the 1.0 km distance. Perhaos
our assumed drop spectrum had too large a mass fraction in the
large drop sizes.
4. In terms of total deposition there is less discrepancy between
model predictions and data for the 1.0 km distance than for the
0.5 km distance. Here, the stack contributes 2-3 times more
drift than does the cooling tower; in total, the predictions
are about four times larger than observed. As expected from
the earlier tower comparisons, the ESC/Schrscker (Limited") and
Slinn I and II models perform very poorly.
5. It is interesting that the Wolf I and II predictions for the
stack are very similar at both 0.5 and 1.0 km in contrast to
the increasing effect of evaooration from 0.5 to 1.0 km seen
for drift drops from the cooling tower. The similarity in
predictions for Wolf I and II for the stack is due to the
slower rate of evaporation which occurs for the larger size
stack-omitted droos which fall from the stack Dlume to the
nearby samplers at 0.5 and 1.0 km downwind of the tower.
697
-------
VALIDATION OF MODELS WITH ESC DATA
The loostions of the SSC sensors ar° given in Fig. 4. Unfortunately,
the data for only >d of the 9 samplers ESC placed at the site were
reduced. Tables 5 and 5 provide a comparison of the modal predictions
with the data in terms of sodium deposition rate, number deposition
rats, liquid mass deoosition rate, and average diameter ''mass averaged"'..
Clearly, significant discrepancies exist between the model predictions
and the data. Notably the predicted averaged deposited diameter is
50-1005 larger than that measured. Clearly then, the mass of salt in
the predicted drop should then be about 2.3-4 times that in the observed
drops. Also, the droplet deposition flux is predicted to be about twice
as large as observed ''considering only drops of size greater than 100
microns). In total, the deposited sodium mass should be ore-dieted as
5-3 times observed. Actually an average value of overpre-diction of salt
deposition flux is more like 10-15- The overprediction of deoosition at
these near-tower sensors may be due in Dart to the questionable
assumptions we had to make concerning the conditions at the stack exit.
However, in view of the fact that the models ovsroredict deposition due
to the tower contribution alone fcompared to the total observed
deposition from tower and stack \ the problem is much more disturbing.
ESC uses a smaller sensitive pape" '122 cm2) than the JHU samoler r700
cm 1 leading to a less statistically significant sample. Moreover, ESC
does not count all drops on the paper. In their method of data
reduction, two squares are drawn on the 122 cm2 area, the larger one to
size the larger drops- and the smaller one to size the smaller drops.
JHU, on the other hand, sizes all drops on the full area of their
sampler. This difference in data reduction methods may be at the root
of the difference between SSC and JHU measurements. It would be
advisable for each group to count the droplets on the other's samolers
to judge the potential differences in data reduction methods.
CONCLUSIONS
The field data acquired in the Chalk Point Dye Study represent the best
thus far available for validation of salt-drift deposition models.
Sodium deposition measurements taken on the ground along arcs 0.5 km and
1.0 km from the tower showed a bell-shaped profile. This shape was also
evident in the model predictions when 10-minute average meteorological
data were used and total deposition predictions were obtained by summing
predictions made for each 10-minute period. Variation in wind direction
thus appears to be a satisfactory explanation of the lateral
distribution seen along arcs on the ground.
Comparison of JHU and SSC data yielded interesting results. The JHU
measurements of drop size spectrum at ground locations yielded a clear
bimodal distibution while the SSC measurements were at best weakly
bimodal. The JHU measurements yielded a large peak of small drops 'up
to 100 microns and a second peak of larger drops ''approximately 250-230
microns). The peak of small drocs is thought to be due to blow-off from
698
-------
the fill section of the tower and reoresents only a small fraction of
the total mass deposited at an;- sampler. For a JKU and ESC sampler
located close together (25 m apart1;, the following observations were
made. The JHU samoier showed a greater droplet count both below 100
microns and above about 300 microns with reasonable agreement in
between. In addition, the JHU mass distribution is clearly shifted
toward greater diameters although agreement above 600 microns Is good.
The median diameters were only 25* different 'JHU: UOO microns; ESC: 336
microns) . Consistency checks on the ESC data revealed a factor of 2
difference between different methods of calculating droplet salt
concentrations and droplet settling velocity at the ground sampler
locations. In general, the JHU measurements were of better quality in
terms of methodology of measurement, data reduction, and internal
consistency. The general trends in ESC and JHU measurements agree
although they differ in details. These details may be important in
specific cases.
Ten drift-deoosition models are compared with the JHU and ESC field
data. For the cooling tower taken alone, a wide ranee in predictions
occurs for sodium deposition flux, number drop deposition flux, liquid
mass deposition flux, and average diameter. A number of models
predicted very poorly; most, however, were not far off from the data, at
least in terms of the sodium deposition predictions. The ESC/Schrecker,•
Hosler-Pena-Pena, and Wigley-Slawson Models compare best with the sodium
deposition flux measurements and are generally within the error of the
data. Those models which degrade the level of input data (e.g., use
readings from one location on a meteorological tower rather than full
profiles, or degrade the spectrum from 25 to 3 bins) lose accuracy in
their predictions. Most models predict larger drop diameters at
deposition than were measured. This may be due to an incorrect
treatment of breakaway in which, in reality, smaller drops are breaking
away from the plume sooner. The wind moving past the tower causes a
wake or cavity effect with a resultant downdraft on the plume; this
effect combined with complex internal circulations within the plume may
be causing earlier breakaway. It should be noted that the comparative
levels of performance of the models aoply only to this special case:
high relative humidity, moderate to large wind speed, very stable
atmosphere. One cannot a priori extend the specific accuracy of any
model to more general environmental conditions without further testing.
For the stack, calculations were made with average June conditions of
the previous year since no stack parameters were measured on the date of
the dye test. Average values from measurements on the previous June had
to be used instead for model input; they were: droplet size spectrum,
liquid mass emission rats, exit temperature and velocity. Also, the
drops were assumed to be saturated at exit. Model/data comparisons
yielded large overprediction of deposition by the models at 0.5 km but
more realistic predictions at 1.0 ka. In any case, the stack parameters
need to be measured on any particular date calculations are required;
this is due to the fact that a significantly larger discrepancy existed
between stack plus tower predictions and data than with just tower
699
-------
predictions and data. An important unknown is the salt concentration of
droplets leaving the stack. Such exit conditions for the stack need to
be measured because the impact of the stack can be as great as the
tower, at least in terms of salt emitted.
ACKNOWLEDGMENTS
This work was funded by the Electric Power Research Institute. The
authors also wish to express their appreciation to the modelers whose
work was utilized for their cooperation.
REFERENCES
1. Environmental Systems Corporation. Cooling Tower Drift Dye
Tracer Experiment. Chalk Point Cooling Tower Project,
PPSP-CPCTP, August, 1977, pp.92-95.
2, J. H. Meyer and W. D. Stanbro. Cooling Tower Drift Dye
Tracer Experiment. Johns Hopkins University Applied Physics
Laboratory. Chalk Point Cooling Tower Project, PPSP-CPCTP-15,
Volume 2, August, 1977, pp 6-13 through 6-26.
3. Meyer, J. H. and Stanbro, W. D., "Separation of Chalk Point
Drift Sources Using a Fluorescent Dye." IN: Cooling-Tower
Environment - 1973, A Symposium on Environmental Effects .of
Cooling Tower Emissions, May 2-4, 1978. Chalk Point Cooling
Tower Project Report PPSP-CPCTP-22. WRRC Special Report No.
9. Baltimore, Maryland. May, 1978.
U. Hosier, C., Pena, J., and Pena, R., "Determination of Salt
Deposition Rates from Drift from Evaporative Cooling Tower," J,
Eng. Power, Vol. 96, No. 3, 1974, p. 283.
5. Wolf, M., Personal Communication, Battelle Pacific Northwest
Laboratory, Richland, Washington, July, 1976.
6. Slinn, W. G. N,, Personal Communication, Battelle Pacific
Northwest Laboratory, Richland, Washington, February, 1977.
7. Hanna, S. R., "Fog and Drift Deposition from Evaporative
Cooling Towers." Nuclear Safety. Vol. 15, Mo. 2.
March-April, 1974. pp. 190-196,
8. Slawson, P. R. and Kumar, A., "Cooling Tower Drift Deposition
Program ENDRIFT II," Envirodyne Ltd., Tennessee Valley
Authority Air Quality Branch, April, 1975.
9- Maas, S. J., "Salt Deposition from Cooling Towers for the San
Joacquin Nuclear Project," MRI 75-FR-1361, September 15, 1975.
700
-------
10. Schreckar, G. and Rutherford, D., Personal Communication,
Environmental Systems Corp., Knoxville, Tennessee, 1975.
11. Overcame, T., "Sensitivity Analysis and Comparison of Salt
Deposition Models for Cooling Towers." Paper presented and
> published in Proceedings of the Conference on Waste Heat
Management and Utilization. Miami Beach, Florida. May 9-11,
1977.
12. Policastro, A. J., Dunn, W. E., Breig, M., Ziebarth, J., and
Ratcliff, M, Evaluation of Mathematical Models for the
Prediction of Salt-Drift Deposition from Natural-Draft Cooling
Towers (in preparation's. Division of Environmental Impact
Studies, Argonne National Laboratory, Argonne, Illinois. 1973.
13- Policastro, A. J., Dunn, W. E., Breig, M., and Ratcliff, M.
"Evaluation of Theory and Performance of Salt-Drift Deposition
Models for Natural- Draft Cooling Towers." IN: Environmental
Effects of Atmospheric Heat/ Moisture Releases, presented at
the Second AIAA/A3ME Thermophysics and Heat Transfer
Conference. Palo Alto, California. May 2-4-25, 1973.
(available from ASMS, Mew York City}.
1U. Dunn, W. E., Boughton, B., and Policastro, A. J. "Evaluation
of Droplet Evaporation Formulations Employed in Drift
Deposition Models." IN: Cooling-Tower Environment - 1973, A
Symposium on Environmental Effects of Cooling Tower Emissions,
May 2-4, 1973. Chalk Point Cooling Tower Project Report
PPSP-CPCTP-22. WRRC Special Report No. 9. Baltimore,
Maryland. May, 1973.
15. Environmental Systems Corporation. Chalk Point Cooling Tower
Project. Comprehensive Project Final Report for the Period
October 1, 1975-June 30, 1975. Volume 2. PPSP-CPCTP-12.
October 1975.
701
-------
Table la. Comparison-of Average Diameter (by Several Definitions) for
the ESC and JHU Samplers
Sampler d.,,.
JHU-J1 320
JHU-J2 240
El 353
E2 268
E3 291
E4 289
a
360
280
500
326
336
344
dCM
207
152
199
195
237
238
d^ dcp (micron)
360 60,280
280 40,180,240
375 80,375
285 80,285
285 65,285
285 35,225
Table Ib. Comparison of Apparent Droplet Concentration and Droplet
Settling Velocity at the ESC and JHU Samplers.
Sampler CDD
JHU Jl 0
JHU J2 0
ESC El 0
ESC E2 0
ESC E3 0
ESC E4 0
(gm/gm)
.029
.019
.006
.011
.018
.031
CCD
-
-
0.009
0.022
0.020
0.052
VSD '(m/s) VLD
-
-
1.29 1.76
0.69 1.47
1.41 1.53
1.57 2.63
702
-------
JHU
Dve Data
1 June 16-17, 1977
Samoler
Sodium Deposition Flux
Tower
mg/m'-4 hour-;
Distance Dir.; CBS.
SOO 330 1.9 * .5 0
500 533 ;,7 * .7 1
500 340 4." * 2.1 1
500 34S 3.9 * 2.0 3
500 350 10.9 - 2.- 5
500 355 7.- * 2.5] 5
500 0.0 6.1*2.4:3
5'~-" 5.0 1.9 * .3 0
! : j 3 4:5 6 - j 3 i 9 : 10
oo i o.noj o.ooi o.ooi o
j 11 12
j
oo o o.ooi o.oo o.oo: o.ooi o.oo o.oo
11 3.22 ).08 O.OOj 10-3 0 1.66J 1.43 3.30 1 3
98 3.bSJ 0.39J 3.50| 18.9 oi 9.24 3.66 6.221 3.9
97 10.0 O."6i 3.65J 32.3 0 13." 12.3 10.8] 6.'.
41 11.2! 2.49J 11.04 35.3 0 13.3 11.' 12.1 9.2
25 9.38] 3.13 10.53J 29.5 0 12.9 11.0 10.4! -.3
! 2.19 5.59
1 4.22 10.2
".46 18.3
• 9.06 21.5
3 -.-S 18.3
07 4.32! 3.13 6.34! 10.7 Ol 3.38 4.93; 4.77 4.501 3.31 3.32
93! 1.37] 1.23J 3.55! 1
64 Ol 0.39 O."4i 1.35 1.1
2! 0.95 2.52
1
JHU
Dye Data
1 June 16-17, 1977
Sodium Deposition Flux
Sampler
Distance Dir. 03S. ]
',ml i
1000 340 1.4 * .4 C
1000 342.5 3.6 * .9 f
Tower
mg/il
2543
1
.6! 1.71 1.29 0.61J 4.
1*^-4 hours
6 - 3 9 10
-IJO.10 4.28 2.52 0.40 i 1.45
.7 1.72 1.46 l.flsl 5.2410.13 4.96 2.72 0.53 L.o"
1000 345.0 2.4 * .4 i 0.9 2.35 1.99 I. Ill 8.
1000 34". 5 j 3.5 * .3
1000 350.0 2.4 *_ 1.2
1000 352.3 2.4 * 1.2 f.
6 0.19 6.55 3.33 1.04 2.23
1 I
.0 2.34 2.02 2.30} 9. 50J 0. 22 i 6. 30 } 3. 12 } 1. 40 2.44
.0 2.19J 1.63 2.32 11
.8 1.73 1.42 1.74 10
1000 333.0 1.2 *_ .3 0.3! 1.291 1.161 2.11 9.t
1000 35". 3 ' 1.2 - .3 <
.6 1.21 0.301 1.51 3.
3 0.20 6.25 2.95 1.99 2.26
6 0.19 6.59 2.90 3.00 2. 09
6 0.17 5.79 2.75 2.28 1.32
4 0.11 3.92 1.86 2.09 l.=0
1000 0.0 1.4* .4 G.4J 0.04J 0.40 l.Joj 3.34H1.05 2.63J 1.08 1.73JO-92
1000 3.0 .51 * .1 L
1000 7.5 0.0 C
1000 10.0 .55 * .2 t
.l| O.Ol! 0.10 1.65! 1.5210.01 0.36 0.16 0.45|!i.2fi
.oj o.ooi o.oo o.ooi o.ooi o.oo o.oo Q.OO o.oo o.oo
.oi o.ooi o.oo o.ooi o.ooio.oo o.ooi o.oo o.oojo.oo
'
; Sampler
Distance Dir. OBS. 1
500 355 6300 S90|
1000 350 7203! 231!
JHU Dye Data
June 16-17. 1977
Tower
234 5
67 3 9 10 1
' Drops/m2-houv
2475 408 4793 55"37
2019J 537 j 2738 1 100240
- 6?93 47Q6 4066 149b J,
15100 4432 2113 1505 1:
Average Diameter tymj
500 353 JlOi 319
1000 350 24li 354i
i
607 . 424 . 262
411 307 | 157
Liquid Mass
- 376, ^4| JW - | 4
11 12
i 1.29 j l.K
1 1.49] 2.36
2. 14J i 3.75
; 2.20 ! 5.09
2.18 6.25
1.94 -.67
i
l.'O ".51
i 1.41 -.11
0.36 5.98
i Q. 22 i 1.65
; 0.68 O.OOJ
• 0.00 0.00
1 12
33 '14j
30 30661
34 ~i
2251 119 j 241 i 367 23l)
Jeoosition Kux
mg/m2-4 hours !
500 355 393 1731
1000 350 ' 234 21
191 763 210«|
73 169 806
! i -
706 j 367 526 ] 367 j 945
360 IS i 52 ! i 159 209J
i
LEGEND
1. Hanna
2. Kosler-Pena-Pena
3. Overcamp-Israel
4. Wigley-Slawson (orofilesl
5. 51 inn I
6. Slinn II
7. Wolf I
3. Wolf II
9. ESC/Schrecker
10. MRI
11. Wiglev-Slawson
12. ESC/Schrecker (limited)
Table 2. Comparison of Predictions o£ 10 Drift Deposition Models to Ground-
Level Measurements of Sodium Deposition Flux, Number Drop Deposition
Flux, Average Deposited Diameter, and Liquid Mass Deposition Flux
. . . Cooling Tower Contribution at JHU Samplers.
703
-------
Sanpler
Distance Direction
'.•nl
589.5 319
5SO." 323
571.4 3"
561.3 332
551.1 536
540.3 340
529.1 345
51". 6 350
OBS.
1 2 3
1.0 i).o| 1.0
3.0
0.0
0.0
13.3
24.3
31.2
54.4
0.0
0.0
0.0
13.0
22.5
46.7
0.0
0.0|
o.o|
24. 0|
29.01
63.1
56.3 ! 54.1
I
Sampler
Distance Direction
(m)
1064.8 335
10S9.D 335.6
1054.3 33S
1043.8 340.3
1043.3 343
1037.6 345
1031.9 347
'1025.1 350
1020.2 352.2
:1008.3 357.1
1002.3 359.6
992.3 2.1
Sampler
Distance Direction
Cm)
540.2 340
1043.3 343
500 355
1000 350
540.3 540
1043.3 343
OBS.
1
0.74
2.59
5 . 32
3.t>5
4.96
5.47
7.34
3.12
7.65
5.55
4.53
2 . 32
•>
2.15
7.19
9.C3
9.99
15.1
19.0
3
0.67
> i-
2.95
3.12
4.74
6.29
21.2] -.52
21.3
19.5
U.4
3.02
3.44
3.18
6.15
s.ool
5.2S| 2.93
OBS.
1
2
3
S'S
572
1166
4994
324
353
JHU Dve Data
June 16-17, 1977
Sodium Deposition Flax
Stack
mg/m;-4 hours
4 3
6 7
I
0.0| ,1.0
D.O
0.0
0.0
3.0
O.D
0.0
0.0
0.0
0.0
21.2
26.6
53.4
O.J
0.0
0.0
0.0
0.0
0.0
0.0
S
5
1 10 11 L2
0.0
0.0
0.0
2.41
13.0
45.5
SO. 4
31.6 "3.3 O.al "0.6
1 i
0.0
0.0
0.0
2.14
15.6
44. .)
'].0| OJ| 0.0
O.i) 0.0
0.0
0.0
26.3-
31.6
S7.l! 61.-
0.0
0.0
13.9
25. 0
0.0
0.0
0.0
3.26
11.0
0.0
0.0
0.0
0.0
24.2
50.7
48.1 21.3 SI. 3:
"5.6i 7l.S| 52. n 29.0
nl."
I
JHU Dve Data
June 16-17, 1977
Sodium Deoosition Flux
Stack
•ng/m^-4 hours
4
0.0
0.0
0.0
0.0
0.0
5
5.49
12.3
15.0
17.1
24.1
0.0| 31.4
0.0
4.25
5.99
5.23
5.44
5.99
33.4
31.1
26.3
16.9
10.1
4.99
6
0
0
o
0
')
0
0
0
0
7
0 . 39
2. U
3.15
5.99
'.59
10.1
12.0
11.0
11.2
0 7.96
0
0
5.33
4.28
8
0.13
1.94
5.34
5.96
•>.09
9.57
10.7
9.54
9.60
6.00
9
1.03
2.33
3.65
4.17
6.45
3.39
9.10
3.34
9.19
6.06
10
0.7-
2.79
5. "2
4.40
6.19
3. "5
10.1
10.7
9.95
'.37
4.11J 4.57 3.9S
2.86
4.96J 3.23
11
0.36
1.36
12
4.62
16.5
1.84| 21.0
2.52 23.1
2.93
3.98
4.79
T! ,\
33.3.
42.5
5.05 39.1
5.23
4.65
4.13
2.96
35.6
T> 7
16.0
4.5;
JHU Dve Data
.June 16-17, 1977
Stack
4
; 6
7
3
9
10
11
•L1
* Drops/m--hour
0
0
1390
9952
.
362 300
327
585
762
1327
555
620
404
525
389
4599
Average Diameter (urn
699
431
-
-
630
419
310
. 72
-
525
64
1 Kanna
2 Hosler-Pena-Pena
j Overcamp- Israel
0
0
463
233
605
339
556
!•>?
631
508
351 534
580
329
Liquid Mass Deposition Flux
mg/m2-4 hours
0 217
0
19"7
-
•
300 216 | 300
60
10
61
106
292
23 :S7
LEGEND
7. Wolf t
8.. Wolf t!
9. ESC/Schrecker
4 Wigley-Slawion (profiles)
5 Slinn T
6 Slinn II
10. MRI
11. Wijley-Slavson
12. ESC/Schrecker (limited)
Table 3. Predictions of 10 Drift Deposition Models of Ground-Level Sodium
Deposition Flux, Number Drop Deposition Flux, Average Deposited
Diameter, and Liquid Mass Deposition Flux . . . Stack Contribution
at JHU Samplers.
704
-------
Sampler
Distance Di.recti.oi
(m)
JHU Dye Data
June 16-17, 1977
Tower and Stack
Sodium Deposition Flux
mg/m*-4 hours
OBS.
1
500(589.5) 330(319)
SOOCSao.7) 333(323)
500(571.4) 340(327)
508(561.5) 345(332)
506(551.1) 330(336)
500(540.3) 355(340)
500(529.1) 0.0(345)
500(517.o) 5.0(350)
1.96 i .26
5.16 : .43
2.38 ± .96
5.44 i .75
8.91 £ .44
7.99 ! .45
3.63 : .73
12.6 : .98
1
0.0
1.1
1.98
3.97
23.7
29.6
54.3
55.5
2
0.0
3.22
3.65
10.0
29.2
32.4
51.0
58.2
3
0.0
0.03
0.39
0.76
26.4
31.3
66.1
55.4
4
0.0
0.0
3.50
S.6S
11.0
10. S
6.54
35.2
5
0.0
10.8
18.9
32.6
56.4
56.1
64.1
77,9
6
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
7
0.0
1.66
9.24
16.1
26.8
58.4
S5.8
"1.5
3
0.0
1.48
8.66
14.4
27.3
55.9
92.1
T6. 3
9
0.0
3.50
5.43
10.3
58.9
42.0
66.5
73. 2
10
0.0
1.S7
3.95
6.74
28.2
32.9
52.4
53.7
11
0.0
2.19
4.22
7.48
17.3
18.3
25.6
50.0
12
0.0
5.59
10.2
IS. 3
45.7
49.5
70.7
64.2
Sampler
Distance Direction
On)
1000(1064.3) 340(333)
1000(1059.6) 342.5(335.6)
1000(1054.3) 345(358)
1000(1048.3) 347.5(340.3)
1000(1043.3) 350(343)
1000(1057.6) 352.5(345)
1000(1051.9) 355(347)
1000(1026.1) 357.5(350)
1000(1020.2) 0.0(352.2)
1000(1008.3) 5.0(357.1)
1000(1002.5) 7.5(359.6)
|lOOO(996.3) 10.0(2.1)
Sampler
Distance Direction
(m)
JHU Dye Data
June 16-17. ]
977
Tower and Stack
Sodiun Deposition Flux
mg/!ni-4 hours
OBS.
2.00 : .32
2.93 ± .26
2.98 ; .34
3.67 - .07
3.71 t .15
3.18 t .30
4.31 i .12
4.31 i .14
4.98 - .09
4.72 I .22
2.37 r .51
5.49 ± .68
1
1.34
3.29
4.22
4.63
5.96
7.27
3.64
3.72
3.05
5.65
4.55
2.32
2 3
3.84
8.91
11.4
12.3
17.3
21.3
22.5
22.5
20.1
11.4
3.02
3.23
1.96
5.73
4.94
5.14
6.39
7.71
3.68
9.34
3.58
6.25
5.06
2.93
4
0.61
1.08
1.12
2.30
2.32
1.74
"2.11
5.76
7.45
6.93
5.44
5.99
5
8.03
17.5
23.3
26.6
35.4
42.0
43.3
39.6
32.6
18.2
10.1
4.99
6
0.10
0.13
0.19
0.22
0.20
0.19
0.17
0.11
0.05
0.01
0.0
0.0
7
4.67
7.07
9.70
12.5
13.3
l(i.7
17. S
14.9
13.8
3.32
5.58
4.28
8
2.70
4.66
6.67
9.08
10.0
12.5
13.5
11. 0
10.7
6.16
4.11
2.36
9
1.43
3.36
4.67
5.57
8.44
11.4
11.4
11.9
10.9
6.51
4.57
4.96
10
2.22
4.46
6.00
6.84
3.45
10.3
11.9
12.2
10.9
8.04
5.98
3.28
I
11
1.65
2.35
3.98
4.52
5.11
5.92
6.49
6.46
6.09
4.37
1.31
2.96
12
6.51
13.9
24.8
28.2
33.3
46.0
50.0
46.2
39.6
24.4
16.6
4.52
JHJ Dye Data
June 16-17, 19
77
Tower and Stack
OBS.
1
2
3
4 5
6
7
3
9
10 11
12
» Drops/mJ -hour
500(540.3) 355(340)
1000(1043.3)350(343)
7595
7311
1168
803
5641
7013
1232
1090
4793 57127
2788 110192
-
-
Average Diameter
500 355
1000 350
553
280
582
581
-
584
396
• 42S 269
307 163
-
-
7255
15927
5506
5017
4328
3440
20501 2792
2125 1355
3032
12665
(ura)
405
233
370
134
434
258
486
364
419
260
Liquid Mass Deposition Flux
mg/mz-4 hours
500(540.3) 355(340) 723
1000(1043.3)350(343)
538
183
93
-
-
514
142
7681 232]
169 1003
•
-
1006
420
583
25
327
123
675
137
1237
466
LEGEND
1. Hanna
2. Hosler-Pena-Pena
3. Overcamp-Israel
4. Wigley-Slawson (profiles
S. Slinn I
6. Slinn II
7. Wolf I
3. Wolf II
9. ESC/Schrecker
10. MRI
11. Wijley-Slawson
12. F.SC/Schrecker (limited)
Table 4. Comparison of Predictions of 10 Drift Deposition Models to Ground-
Level Measurements of Sodium Deposition Flux, Number Drop Deposition
Flux, Average Deposited Diameter, and Liquid Mass Deposition Flux . .
Contribution of Cooling Tower and Stack at JHU Samplers.
705
-------
Sampler
Distance Direction
(m)
230(261) 131(212.9}
300(346) 357(334)
400(461) 347(330)
500(547) 352(338)
750(773) 358(543)
750(800) 348(339)
1050(1110)342(333)
980(1023) 550(343)
1740(1756) 0(356)
ESC Dve Data (Evening)
June 16-17, 1977
Sodium Deposition Rate
Tower and Stack
;ng/m2-4 hours
OBS.
0.02
6.58
1.54
4.24
NR
MR
NR
NR
NR
1
0.0
62.0
12.2
29.1
17.5
3.93
2.61
5.19
3.05
2
0.0
75.5
15.9
32.4
21.4
12.4
7 . ; b
14. 28
8.91
3
0.0
42.6
4.56
31.4
16.9
7.73
3.54
5.31
3.54
4
0.0
24.2
17.3
7.56
1.01
3.56
0.44
2.52
2.44
5
0.0
105
36.4
52.8
37.6
26.1
13.9
31.1
11.0
6
0.0
0.0
0.0
0.0
.002
.005
0.17
0.15
0 . 39
7
0.0
65.6
21. S
37.7
16.4
16.1
5.79
14.3
6.41
8
0.0
62.0
20.2
37.3
12.3
12.2
3.98
9.35
4.57
9
0.0
3t r
J . 0
35.4
42.9
29.1
11.6
4.81
T . 15
3.70
10
0.0
*" A 1
3D . 4
14.4
29.1
17.5
10.5
5.65
".90
4.98
-i
11
0.0
QA **
BU . _
'.24
19.3
13.1
9.93
1.92
4.79
0.04
12
0.0
CO C
JO . 3
40.3
49.6
30.8
5.68
24. 5J
35.5
0.32
-
Sampler
Distance Direction
(m)
230(261) 131(212.9)
300(346) 35/(354)
400(461) 347(jjO)
500(547) 352(333)
750(773) 358(348)
750(800) 348(339)
1050(1110)342(335)
380(1023) 350(543)
1740(1756) 0(556)
ESC Dye Data (Evening)
June 16-17, 1977
Tower and Stack
* Drops /mz-hr.
OSS.
0
10766
2101
3630
NR
NTC
NR
MR
:MR
1
0
2454
1237
1164
1208
812
424
699
605
-)
0
3020
2430
3643
4278
4150
4025
5557
8255
3
0
5039
584
1213
1392
326
744
788
1127
4
0
3064
537:
3312
98;
3214
51£
2844
120:
S
0
14178
55 508
53317
67137
90883
51377
101393
37684
6
0
0
0
0
0
0
0
0
0
7
0
5055
7294
7319
6011
S398
11368
14006
8531
3
0
4632
6485
5575
5139
4425
5239
4925
1572
9
0
4460
4058
4862
5610
5630
3610
5223
5500
10
0
2833
1953
1837
1359
1371
1873
1523
11
0
4672
1650
2748
2654
2364
392
1777
44
12
Q
4600
4463
7958
1735
1067
15921
13146
590
1. Hanna
2. Hosler-Pena-Pena
3. Overcamp-Israel
4. Wigley-Slawson (profiles)
LEGEND
. Slim. I
. Slinn II
. Wolf I
8. Wolf II
NR - Not Reduced by ESC.
9. ESC/Schrecker
10. MRI
11. Wigley-Slawson
12. ESC/Schrecker (limited)
Table 5. Comparison of Predictions of 10 Drift Deposition Models to Ground
Level Measurements of Sodium Deposition Flux and Number Drop Deposition
Flux . . . Contribution of Cooling Tower and Stack at ESC Samplers.
706
-------
Sampler
Distance Direction
Cm)
300(346) 357(334)
400(461) 347(330)
500(547) 352(338)
ESC Dye Data (Evening)
June 16-17, 1977
Average Diameter (urn)
Tower and Stack
OBS.
360
306
310
1
652
678
610
2
-
-
-
3
732
622
563
4
650
482
431
S
509
334
274
6
-
-
-
7
658
475
410
8
638
461
377
9
679
652
437
10
.
-
-
11
663
535
482
12
662
659
422
Sampler
Distance Direction
(m)
230(261) 181(212.9;
300(346) 357(334)
400(461) 347(330)
500(547) 352(338)
750(778) 358(348)
750(800) 348(339)
1050(1110)342(335)
980(1023) 350(343)
1740(1756) 0(356)
ESC Dye Data (Evening)
June 16-17, 1977
Liquid Mass Deposition Flux
Tower and Stack
mg/m2-4 hours
OBS.
0
1047
126
226
NR
NR
NR
NR
NR
1
0.0
1422
806
552
298
213
52
79
25
2
-
-
-
-
-
-
-
-
-
3
0.0
2485
294
504
308
187
102
111
44
4
0.0
1762
1261
551
74
259
32
169
36
5
0.0
3914
2604
2301
1196
1146
353
967
183
6
0
0
0
0
0
0
0
0
0
7
0.0
3022
1639
1056
491
471
276
516
144
8
0.0
2542
1327
608
147
173
101
106
4
9
0.0
2966
2352
851
498
391
145
114
92.3
10
-
-
-
-
-
-
-
-
~
11
0.0
2848
528
660
361
370
79
183
3
12
0.0
2794
2675
1253
299
57
696
476
6
Hanna
Hosler-Pena-Pena
Overcamp-Israel
4. Wigley-Slawson (profiles)
LEGEND
5. Slinn I
6. Slinn II
7. Wolf I
8. Wolf II
NR - Not Reduced by ESC.
9. ESC/Schrecker
10. MRI
11. Wigley-Slawson
12. ESC/Schrecker (limited)
Table 6. Comparison of Predictions of 10 Drift Deposition Models to Ground
Level Measurements of Average Deposited Diameter and Liquid Mass
Deposition Flux . . . Contribution of Cooling Tower and Stack at
ESC Samplers.
707
-------
o
00
FILTER PAPER
HOLDER AND
CANDLE HEATER
o
o o
7777777777777777777,
DEPOSITION
FUNNEL
SAMPLE
BOTTLE
SAMPLER
POST
ATTACHMENT
STAKE
m
•
©-
COOLING
TOWER
RADIUS
-SAMPLERS
30° SAMPLING ARRAY (3 FUNNELS AND 3 FILTERS PER STATION)
* DDTE ACTIVATED 30° ARRAY SAMPLING STATIONS
» OTHER SURVEYED SAMPLING STATIONS
Fig
•ig. 1. (upper left) Sketch of Position of JIIU Samplers at Typical Sampling Station. (Lower left) Relative
Position of Duplicate Samplers at a Sampling Location. (right) JHU and liSC Sampling Arrays at
/••"u ,. T i, n^ -;«-*- ^ A ,1 .->«-*-,-., i r..™^ T^^ r o~\ ». *- " *
Chalk Point. (Adapted from Ref. 2).
-------
JUG
26U
24U
2/U
JIK)
tou
IbU
14(1
1 *>U
IUO
HO
WJ
4(i
/U
0
"T"
_
-
~
-
T
-
-
it
-7
i
—
i
—
j
1
i i
HI
-
~
';
1
n
_
i
_
r1
"
-
3
V
f
•
w*
,H
U<
c
f
b
>-l
1
k,
1.1
"•'
(>1
iH
1 "I 1 1 I I 1 1TT
.t lilh-, I..,,"
•"•• bO' 111
MO.,,,,.., >!«,
-
-
:
* l 8tM ,-M.a.oc.i." -
• ] 1CW-J ^T, oijj.iei
— .
•ijn '.Hfi r,';(i i.i.n f/w ,
^uo
?bli
?40
2,(,
?OU
IWJ
p Ibtl
"» 14U
i
100
HO
60
40
1
:
..
-
-
.,
I
P
~
"
-,
~
—
—
•
"
~
~
_
-
i i i i i i i i i i i i i i r_.
Jlji 0 t fc (t.
Fillci Ivalif, tli-a tfU""' L "'•
^, :<,.,' l.ltc, ..C«(0... .!,-.! «>•%
"'""""""""' ""' '9'J'
-
-
1
1
1 — 1
— '1 /61I */',. O, upltrl
* 1 1 190 Mr" Jiuplel
Ihi^^^-^ _ i r
til) UXI MO 1HO ?
40 iiHO G>0 (.1.0 /(K
140 IBU 220 2bO 301, J40 380 4?0 460 MX) i
2ttor-
260
240
220
200
„ 1BO
g
° 140
4
£ 120
z 100
80
60
40
20
Q
-
-
„
-
-
-
-
-
-
„.
1 1 1
I 1 I
\
n i i i i i i rn n i ri i_
• fcr.ri
° 3bO 1 Okm
H; Wdlr, it-nsitwt filtei WIH.-I
lilu-r |u|wi jre» .- tiO'J I".'
% t,lit( SUM couniwl 72'V
loljl {Jru[jlt'li cuunied 2^41
1
"
1-
~
-,
r~
— i
-
-
-
Tl -
lh
rn~n i i j i _i i M i
280
2tiO
200
240
220
ISO
|
•o
"° 140
t
E 120
2 100
BO
60
40
70
0
'J '
_
—
-
-
-.
-
_
—
~
-
_
•~
rr
i
I M 1 1 1 1 1 1 1 1 M 1 1 I 1 l 1
40t> 3'jl1 1 0m,
flu II -iLt'Ml l>ld,h liKrl ltain-1
tihi'i jk.^-1 ared GUI. "
lOllf tl.Up"liU.UHH'< 20 jjm ime(v>i
-------
m n ii i iiii111 M ii i
• A Single droos for given di;
355 0.5 km water sensitive
(«) 881 and 1044
0.70
0.60 -
£ 0.50 -
0.40 -
- •= 0.30 -
0.20 -
0.11 -
II I I I I I I I I I I i I I I I I I I I I I I I I I I I I I M I I I
350° 1.0 km water sensitive
/, 350° 1.0 km fluorescent
Mean drop size (jim x 10)
22 30 38
Mean drop size
1 1 1
•— — -« Cooling tower
Unit No. 3 stack
0
325 330 335 340 345 350 355 0 5
Station location (degrees)
10 15 20
15
14
13
12
_ 11
w
T 10
s
6
I I I I I
*• — -* Cooling tower
* * UnitNo. 3 stack
325 330 335 340 345 350 355 0 5
Location (degrees!
10 15
Fig. 3. (top) Percent Mass Fraction as a Function of Mean Drop Size at TVn
JHU Samplers, (bottom) Separation of Tower and Stack Sources"of
Sodium Deposition at the 0.5 km and 1.0 km Arcs. (Adapted from Re£. 2)
710
-------
GROUND - LE v'EL SAMPLE?
LOCATIONS FOR 6/-0/"7? DYE TES~
• J2
>• E4
'N tO M:"-.
MEAN A.NO
E3 ••
\ •' \
, • El i
i 0 100
COOLING ",
\V/-STACK
{
4 HP. MEAN •
WIND DISEC
~>
z
p
r
; o/ ^7 ESC
1
^ j » :0
DROP S.ZE urn ,0
DROP SIZE DISTRIBUTION
E3 6/16/77 ESC
I 3 5 7 II 15 21 27 16 45 60 SO
•z.
3
o
n
u
1
D3CP SIZE
2.0
8
0.5
DROP SIZE DISTRIBUTION
E2 6/16/77 ESC
i 5 7 II ;S 2l 27 3S 45 SO SO
DROP SIZE pm . ,0
1.0
o
o
0.5
DROP SIZE DISTRIBUTION
E4 6/16/77 ESC
3 5 7 :l 15 21 27 36 45 SO 30
DROP SIZE m « 10
Fig. 4. (upper left) Location of 4 ESC and 2 JHU Samplers at which Drop Size
Distributions were Measured, (lower left and right) Droplet Count
as a Function of Droplet Size for All 6 Samplers.
711
-------
MASS OlSIKIHUIION
Jl b/lb/ff JHIJ
Z
o
§0,
a:
in
3
3.
MASS DISTRIBUTION
J2 eM/'/f JHU
A A
MASS OlSIRIBUIION
El 6/16/77 ESC
MASS DISTRIBUTION
E2 6/16/77 'ESC
MASS DISTRIBUTION
E3 6/16/77 £SC
; II 15 ?l
DROP SIZE
01
E
X
g"
b
Fig. 5. Percent Mass Fraction as a Function of Droplet Size for the 4 ESC and 2 JHU Samplers.
-------
-0
h-1
U>
6OO
i>OO
CE 40G
Ul
S
< 400
a
z
<
a 20O
100
0.5
^0.4
2
O
o
n;
u_
0.2
O.I
FINAL DROP SIZES vs DOWNWIND DISTANCE
Q--.
~- JHU 77
WAItNStNSlTlVE
™-JHU 77
ore TKACER
ESC 77
cr
"E
E
a..
o
o i3o i3o 300 400
?3o eoo
ooo
DOWNWIND DISTANCE m
COMPARISON OF MASS
FRACTIONS AT TWO NEARBY SAMPLERS.
r-\
Jl JHU 77
E3 ESC '77
200 40O
DROP SIZE r
600
eoo
COMPARISON OF DROP
COUNTS AT TWO NEARBY SAMPLERS.
Jl JHU 77
WA1EH! btNS.
— E3 ESC 77
200
400
60O
aoo
DROP SIZE
(upper left) Variation of Mass Median Diameter
with Distance from the Tower, (lower left)
Comparison of Mass Fractions at Two Nearby
Samplers, (above) Comparison of Drop Counts
at the .Same Two Nearby Samplers.
-------
Q
1C-
SODIUM DEPOSITION RflTE
JHU DYE DflTfl -- 0.5 KM
JUNE 16-17,1977
TOWER
SODIUM DEPOSITION RflTE
JHU DYE DflTfl — 0.5 KM
JUNE 16-17,1977
TOWER
OBSERVED
HOSLER-fENR-ra«
OVERCttf-13RflQ.
HISLET-SUWSON UfWILESI
MIOLEK-SLHUON
SLMi-ll
- • EStVSCmECKEH
ESC/SOHRCCKZR
MU
310-0 335-0
345-0 350-0
flNBLE (DEGREES)
355-0
0-0
S-0
OBSERVO)
MOLF-I
MDLF-II
330-0
335-0
M5-0 3SO-0
flNGLE (DEGREES)
355-0
0-0
~f~
5-0
Fig. 7. Comparison of Predictions of 10 Drift Deposition Models to Sodium Deposition Flux Measurements .at
8 Locations Along the 0.5 km Arc . . . Cooling Tower Contribution*Alone.
-------
SODIUM DEPOSITION RflTE
JHU DYE DflTfl — 1.0 KM
JUNE 16-17,1977
TONER
SODIUM DEPOSITION RflTE
JHU DYE DflTfl — 1.0 KM
JUNE 16-17,1977
TOWER
OBSERVED
HOSLER-PENft-PENH
- • DVERCflMP-lSWCL
— MtBLO-SLrtGt* (PROFILES)
MIflLEJ-SLftOON
SLIW-II
Ul
en
a; o
0>J
I
OBSD^VED
MCLF-I
HOLf-II
ESC/SCrtiECXER
ESC/SORECKER (LIH1IEDJ
- NRI
MO-0 MS-0 533-0 355-0 0-0
flNGLE (DEGREES)
10-0
1S-0
345-0
J5D-0
355-0
flNCLE
0-0
(DECREES)
5-0
10-0
15-0
Fig. 8. Comparison of Predictions of 10 Drift Deposition Models to Sodium Deposition Flux Measurements
at 8 Locations Along the 1.0 km Arc . . . Cooling Tower Contribution Alone.
-------
SODIUM DEPOSITION RflTE
JHU DYE DflTfl ~ 0.5 KM
JUNE 16-17,1977
STflCK
SODIUM DEPOSITION RflTE
JHU DYE DflTfl ~ 0.5 KM
JUNE 16-17,1977
STflCK
— HflNNfl
HOSLER-P£WH>a«
- • ovERow-isiwa.
UIGLEI-SUMSON [FHFILESJ
UIGLEI-SUUSON
SLlm-II
38
X
§'
§'
33S-0
3iO-0
•H5-0 3SO-0
flNBLE (DEGREES)
3SS-0
0-0
B-0
HOLF-!
WlF-tl
ESa/SOIHECKER
CSO/SCHCCKCR (LlnlTEDI
ItRI
330-0
335-0
3-45-0 3SO-D
UNCLE (DECREES)
355-0
0-0
5-0
Pig. 9. Comparison of Predictions of 10 Drift Deposition Models of Sodium Deposition Flux at 8 Locations
Along the 0.5 km Arc . . . Stack Contribution Only.
-------
O '
l"
CM
X. C)
* <•
X
(S
r.
SODIUM DEPOSITION RflTE
JHU DYE DflTR ~ 1.0 KM
JUNE 16-17,1977
STflCK
imtMt
t«3U«-f£NrH>£Nfl
OVEKCfW-lSRflCl.
MIttLEI--SLf*(3UN CCKDFILES)
Hia£I-SU*SON
310-a
34S-0
350- D
355-0 0-0
flNGLE (DEGREES)
5*0
10-0
15-0
C)
-c.
I
CM
X o
X J.
r: S
CD
CJ '
o
SODIUM DEPOSITION RflTE
JHU DYE DflTH — 1.0 KM
JUNE 16-17,1977
STflCK
/\
HCLT-I
MKI
51.K1N--I
.. \
\ \
\ \
\ \
310-0 MS-O
3SS-Q 0-0
(Ut:GREf:S)
6-0
10-0
1S-0
Fig. 10. Comparison of Predictions of 10 Drift Deposition Models of Sodium Deposition Flux at 8 Locations
Along the 1.0 kin Arc . . . Stack Contribution Only.
-------
SODIUM DEPOSITION RflTE
JHU DYE DflTfl ~ 0.5 KM
JUNE: 16-17,1977
TOWER RND STRCK
SODIUM DEPOSITION RRTE
JHU DYE DflTfl ~ 0.5 KM
JUNE 16-17,1977
TOWER RND STRCK
sr
?
,n
S-
0)
oc ^
"^ VJ "
CJ '^
M'
X rj
£s-
•'0
1~"
S S*-
o
c.1
rtj
ir
o •
< j
S"
CJ
OBStKVEO £-
HM^H-WTW-PITIH
_. DVEKtm>-13«€t.
.— - Hie.'.£I-SLf«SQH ,.
-- — SLIttt- II :|_
'•• . 83"
/ .-•**- 1~
/ ' .^-"* ^*
/ /' c^'
/{' ="-
//'' '' '^ ^
•' ' /-''"' x- ^
.'" ' / / tN
/ ./'
///••"" /
^JJ''-'-'' "-"< •••-../ ?
••- ' '• -..-:-:-'-;;: • "~^-~?.'%'--""" / cj
,«ii*-----:7-" ..,,-,-: . . . . . Q.
OBSEKVfD
KiU' II
- - -— — • FSG/'iGIHtCKEK
-..-.. ESQ/SOHRL'OiER
- MKI
SLIM*- 1
/
/
•••'
/ .•';.
2^5^" """^ :
•i-SO-11 33.S-0
'iifl-ll H.S-0
"SSt)-D
0-0
5-0
Ti'J-O T3S-0
'HU-0
r —
•Jti-0
0-0
B-0
Pig. 11. Comparisoji ol: Predictions of 10 Di'ift Deposition Models to Sodiiun Deposition Vl\.\x Measurements
at 8 Locations Along the 0.5 Ion Arc . . . Cooling Tower and Stack Contributions.
-------
SODIUM DEPOSITION RRTE
JHU DYE DRTR -- 1.0 KM
JUNE 16-17,1977
TOWER RND 5TRCK
OBSERVED
- • ovEKCfirr-isfina.
WIGLEr-SUUSON
(PROFILES)
CM
*
s
CO
1 J"r~"'""r'~ r— 1 r—-— i
HO-0 3i5-0 150-0 3SS-D 0-0 6-0 10-0
flNSLE [DECREES)
1S-0
SODIUM DEPOSITION RRTE
JHU DYE DRTR ~ 1.0 KM
JUNE 16-17,1977
TOWER RND STRCK
OBSEKVU)
WOLF-II
_ . ESC/SCtKtUQJt
ESC/SCtKKCKER (LIHITIDJ
UK I
a. INK-i
•sin-o 3iS-o
33.1-0 3SS-0 0-0
flNBLE (UECKLTS)
E-O
10-0
Pig. 12. Comparison of Predictions of 10 Drift Deposition Models to Sodium Deposition Flux Measurements
at 8 Locations Along the 1.0 ton Arc . . . Cooling Tower and Stack Contributions.
-------
COOLING TOWERS AND THE LICENSING
OF NUCLEAR POWER PLANTS
J. E. Carson
Division of Environmental Impact Studies
Argonne National Laboratory
Argonne, Illinois 60439, U.S.A.
ABSTRACT
One provision of the National Environmental Policy Act of 1969 requires
quantitative estimates of the effects of effluents from cooling towers
used by nuclear power plants on the local air environment. Meteorologists
were required to make these predictions even though adequate quantitative
observational data at operating power plants were not available and for
which accurate, proven models did not exist. Many of the environmental
questions raised concerning the use of wet cooling towers in the early
1970's have been sihown to be, in fact, non-problems: acid rain, plant
damage due to salt drift from fresh water cooling towers, fogging and
icing from natural-draft units, offsite fogging and icing from mechanical-
draft towers, etc.
The procedures used in the environmental review process are discussed.
Examples of the types of questions raised at environmental hearings, for
some of which good answers are not available, will be discussed. Obser-
vations at hundreds of operating cooling towers in the United States and
in Europe show that, except for the visual impact of the towers and their
visible plumes, wet cooling towers are effective, economical heat sinks
that are environmentally acceptable if properly constructed, maintained
and sited.
INTRODUCTION
The National Environmental Policy Act of 1969 (NEPA) completely altered
the method of licensing of many facilities, including the issuance of con-
struction and operating permits for nuclear power plants. The NEPA
review process, as outlined in the Act and expanded by court decisions,
requires a much more thorough, expensive and systematic review of the en-
vironmental impacts of the proposed facility than was previously required.
Among other items, NEPA requires an analysis of all alternatives to the
proposed action. For nuclear power plants, these include not building any
new generating capacity, using other than nuclear fuels, locating the plant
on other sites, and using other types of cooling systems. The benefit/
cost ratio is one of the methods to be used to determine whether or not
the license should be granted.
The bottom line of the NEPA process is the decision by the licensing agency
(in this analysis, the Nuclear Regulatory Commission, NRC), whether or~not
to issue a permit for the construction or operation of the facility. A
720
-------
"no" verdict is made if one or more of the environmental impacts is not
acceptable for that site; for example, using mechanical-draft cooling
towers next to a major highway. Another example would be a threat to an
endangered species. Or the licensing agency, using a number of criteria
including the benefit/cost analysis, may require that one of the alter-
native sites, cooling systems or fuels be used.
It should be remembered, and many opponents of nuclear power seem to
ignore this fact, that it is not possible to generate energy from any
source with creating some negative impacts on the environment. The NEPA
review process is the method used to insure that the total environmental
impact of the proposed power plant is low and acceptable.
As a result of the National Environmental Policy Act of 1969 and the
Calvert Cliffs court decision, the Directorate of Licensing of the United
States Atomic Energy Commission (AEC, now the U.S. Nuclear Regulatory Com-
mission) entered into a crash program to write environmental impact state-
ments (EIS's) as a step in the licensing of nuclear power plants and other
facilities. Argonne National Laboratory (ANL) is one of the three national
laboratories that have served as consultants to the AEC/NRC in the pre-
paration of EIS's. I was assigned to write the meteorological sections
of the EIS's prepared by ANL. In this paper, only those items related to
the environmental impacts of waste heat on the atmosphere are discussed.
As part of the program to prepare EIS's, several meteorologists, Includ-
ing consultants for the utilities and the author, were forced to become
"instant experts" on the effect of cooling-system effluents on the atmos-
phere. We were (and are) required to make quantitative oredictions of the
effects of the cooling system on the local air environment—effects such
as fogging, icing, and drift. Recent trends indicate that EIS's may be
required for fossil-fueled plants and many other types of facilities.
Unfortunately, the state-of-the-art in atmospheric understanding and model-
ing is such that meteorologist are not able to make accurate, quantitative
predictions on how the atmosphere will react to the large amounts of heat
and water vapor generated from limited areas per unit of time from closed-
cycle cooling systems.
A survey of the literature in the early 1970's indicated lots of generali-
ties but very little factual information on cooling-tower effects. State-
ments with as little usefulness and authority as "cooling towers have the
potential to cause fogging and icing" were found all too frequently. In
addition, some of the facts presented were wrong (for example, the drift
rates from cooling towers were quoted as about 0.2%). Also, the time
available to become an "expert" was quite short. Nevertheless, meteoro-
logists both for the NRC and the utilities are required to make these
estimates and publish them in Environmental Reports (ER's prepared by the
Utility) and Environmental Statements (ES's prepared by the NRC). We later
have the dubious privilege of defending our analyses and conclusions in
public hearings. These calculations and analyses must be made even though
the complex processes involved are not understood and for which adequate,
proven models are not available. Even in 1978, the amount of high quality
721
-------
observational data on cooling tower drift and plumes is small. A large
number of mathematical models have since been developed; however, none
of the models has so far been shown to accurately simulate nature over a
wide range of tower and atmospheric conditions with the degree of pre-
cision needed for the NEPA process.
STEPS IN THE PREPARATION OF ENVIRONMENTAL IMPACT STATEMENTS*
It should be remembered that an EIS is a legal document whose primary
function is to provide factual, quantitative information on the environ-
mental impact of the proposed installation (and alternative plant
designs and sites) to both the public and to licensing and regulatory
agencies, who in turn use this information in their decision-making roles.
The key word to be considered in writing a section of an EIS is impact
rather than effect; the agencies and the public are more concerned with
how the plant will affect people, fauna, flora, and the environment than
with processes or effects. For example, if it can be shown that plumes
from natural drift cooling towers never causes fog or that drift from a
freshwater cooling tower with state-of-the-art drift eliminators never
casuses problems to the biota due to salts or wetting, no model or study
should be required to prove it for each plant. Thus, it is not sufficient
to only predict the frequency, extent, and severity of a specific event
(such as fogging and icing from a mechanical-draft cooling tower or cool-
ing pond); some effort must be made to estimate how these changes will
affect people, traffic, flora, etc., which is usually a much more diffi-
cult problem. Frequent fog in winter from a MDCT or cooling pond over a
vacant field owned by the utility is quite acceptable whereas fog only a
few hours per year over a busy highway is not acceptable. The atmospheric
effects of a cooling system depend primarily on the type of cooling system
selected and on the local climate; the impact of the cooling system will
be controlled to a considerable degree by the location of the cooling de-
vice with respect to roads, homes, trees, etc., and on the height of
release.
A brief summary of the NEPA review process for the atmospheric effects
of waste heat from a nuclear power plant is presented below (similar pro-
cedures are taken for the other environmental impacts as well). First,
the utility (or its consultant meteorologist) prepares its assessment of
the impacts of the cooling system selected for the plant (plus all of
viable alternative cooling systems for that site) on the air environment.
This massive document (up to 8 thick volumes), the ER, is sent to state and
local agencies, the NRC, EPA and other federal agencies, and is made avail-
able to the public and to potential opponents. NRC meteorologists use
this report, plus information from other sources (such as literature re-
ports, field studies, models, personal observations of cooling towers in
action, and the known environmental impacts from cooling systems) to pre-
pare its own independent analysis. Independent is a key word in the
description of the NRC analysis.
*In this paper, the term Environmental Impact Statement includes both ER's,
prepared by the applicants, and ES's, prepared by the regulatory agencies.
722
-------
The NRC's environmental review for the plant is published in the Draft
Environmental Statement (DES), which is circulated to local, state and
federal, regulatory agencies, potential intervenors and the public for
comments and criticisms. After a 45-day comment period, the responses
to the DES are collected, studied, and responded to by the NRC staff.
The NRC then issues a Final Environmental statement (FES), which in-
cludes the staff's responses to the comments and questions generated by
the DES: a reproduction of all written messages received is included in
the FES.
The FES is then recirculated to all interested parties and, after a suit-
able waiting period, a public hearing before the Atomic Safety and Licens-
ing Board (ASLB) is held. These hearings are conducted in an adversary
environment. Statements in the FES and ER are used by the ASLB in its de-
liberations as to whether or not the facility should be licensed. All
parties to the legal process are free to challenge the accuracy or
adequacy of the statements made in the FES and ER, present new information,
or raise new issues: the NRC and the utility may use Supplemental Testi-
mony (a written, signed, sworn document) or present expert testimony at
the environmental hearing.
The rules-of-evidence at an ASLB hearing are quasi-legal; that is, they
are not as strict as those in a civil or criminal case. Hearsay evidence,
which includes observations and opinions not put in writing by plant
personnel or others as to, for example, the actual extent and frequency of
fogging and drift at an operating power plant, may be admitted into the
record; such testimony is usually given "low probative value"--legal talk
for "we hear you but we really don't believe it." An oral statement at
a hearing that, for example, no one has reported damage from acid misting
as a result of the merging of an S02 plume with that of a NDCT cannot
be used as proof that it does not occur. The regulatory agencies are like
other legal bodies; they want documentary evidence that can be placed
into the record. Many of those participating in the review process and
those in the hearing room have never seen a big cooling tower in operation.
It is my experience that the most useful type of evidence is a written
report published in a quality, refereed journal. This report is then
referenced in the ER, DES, FES or supplemental testimony, docketed and
made available to the public by being placed in public reading rooms.
Thus, all parties to the hearing can determine the accuracy and validity of
your sources of information and the basis for your conclusions. The
accuracy or validity of such references is rarely questioned. Big reports
describing field studies are also very useful, as are theoretical (model)
studies and generic reports.
It should be pointed out that the burden-of-proof that what is said in the
ER and ES is true and complete rests with those who prepared them, and that
intervenors, the utility or ASLB members can challenge any statement made
or make their own independent assessment. It is of course hard to prove
that postulated long-term or subtle effects will rarely or never happen, or
are truly insignificant. A frequently heard phrase at these hearings is
"if you cannot say with 100% confidence what will happen and then prove it,
don't build the plant."
723
-------
The objective of most intervenors is to either prevent the construction
of the power plant, force a change in location, or force the utility to
make a major change in plant design (for example, change from once-through
cooling to cooling towers). (Some intervenors want to stop the construc-
tion and/or operation of all nuclear power plants; others are trying -to
prevent the construction of all new electric generating stations.) The
best way for an intervenor to attain his objective is to present a positive
case at the ASLB hearing; that is, present written evidence and/or expert
testimony that some aspect of proposed nuclear power plant is in fact
environmentally not acceptable. He can also attempt to show that the NRC
staff's analyses are wrong or inadequate; the best way to do this is to
present a positive case. The mere assertion that the NRC has not provided
a satisfactory analysis is usually not sufficient.
The ASLB, which consists of three members, then makes its decision, based
on the ER, FES and all of the other documents introduced into the "record"
during the lengthy review process, plus transcripts of the hearing.
The ASLB's decision can be appealed within the NRC structure to the Atomic
Safety and Licensing Appeal Board, or the Commission itself, which may
reverse all or part of the decision, or call for further testimony and
evidence on specific points.
Appeals to the civil courts have also been used to delay or suspend the
issuance of a construction or operating licenses.
A large number of atmospheric effects of cooling systems can probably be
answered by "too small to be measured" or "too small to be significant."
However, merely saying it doesn't prove that the effect is too small — the
conclusion must be proved by actual measurements or a validated model to
become acceptable evidence. Mesoscale weather changes, such as the genera-
tion of clouds, additional precipitation and severe storms, should be
items of major concern. An unfortunate consequence of the NEPA review
processes is that it encourages the formation of energy parks — it takes
very little additional effort, money and time to license multiple-unit
power stations than a single unit. Power centers containing a nuclear
capacity of 6500 MWe, are now being reviewed; even larger ones are being
discussed. There must exist a critical heat load for a given site which,
if exceeded, can create its own mesocirculation or heat island and thus
create inadvertent weather changes. However, no one knows where this
limit is.
MATHEMATICAL MODELS AND EIS's
One of the unfortunate features of the EIS work is the emphasis placed on
quantitative estimates; the numbers generated by models which do not
accurately simulate nature tend to be more acceptable and given higher
probative value than are observations made at operating power plants. In
one case, a utility spent money to hire a meteorological consultant"to de-
velop a computer simulation program for a NDCT at a proposed nuclear plant,
724
-------
but has never taken a single plume measurement from his own operating NDCT's
located 20 miles from the proposed site, or tried to compare actual plume
behavior with his model.
Since the primary use of cooling tower models in EIS work is to determine
the environmental acceptability of that tower for a specific heat load at
a known location, only a model that has been shown to accurately predict
Illume and drift parameters for that tower size and for that climate (that
is, a fully validated model) should be used. Unfortunately, none of the
models now available has been proven to be sufficiently accurate. For
example, a model was used to predict the frequency and extent of fogging
from various types of cooling towers for a nuclear power plant which may
be forced to convert from once-through to closed-cycle cooling. The model
for mechanical draft cooling towers predicted a few hours per year of fog
over a nearby (300-350 m) major highway. If we had complete faith in the
model, the MDCT could have been listed as an acceptable cooling system.
However, the lack of proven accuracy of the model for predicting the
extent of surface fog, the irregular terrain at this site, and the con-
sequences of a wrong decision forced NRC to reject this type of cooling
tower as an acceptable alternative for this plant. The recommended cooling
system for this power plant, natural-draft towers, would be more expensive
and would have a much greater aesthetic impact.
In the past decade, a large number (more than 50) mathematical models have
been developed and used to make quantitative predictions of plume rise,
plume length, drift deposition, local changes of temperature and humidity,
and other effects that are created by the heat and moisture discharges
from cooling towers.
Although meteorologists have found that mathematical models are very useful
in studying and understanding a wide range of atmospheric processes, the
primary use for mathematical models of cooling tower plumes and drift
has been to provide quantitative predictions of cooling tower effects at
proposed power plants for use in environmental reports and environmental
impact statements. Therefore, mathematical models for this use should be
simple and easy to apply with available data, be inexpensive to run on the
computer, and have been shown by tests with independent^data to accurately
simulate nature for the range of atmospheric conditions expected at the
new location. Research-type models can and should be more complex and may
require specialized data—such as vertical profile of wind speed and direc-
tion, air temperature, humidity, etc.--not readily available at other sites.
QUESTIONS RAISED AT ENVIRONMENTAL HEARINGS
Most of the issues raised at the public hearings concerning cooling tower
impacts are valid ones that must be addressed. Typical of the type of ques^
tion that does have an answer is one posed at a recent public hearing for a
nuclear power plant in Indiana: "Will the heat, humidity, icing, water
droplets (due to both fog and drift) and salts (due to drift) added to the
atmosphere by a large group of mechanical-draft cooling towers less than
725
-------
43
one kilometer away decrease the yield of fruit or increase the incidence of
fungus diseases in a peach/apple orchard located within one kilometer of the
towers? If yes, by how much? What can be done to mitigate or lessen the
damage? If no, prove it in an adversary environment in a court of law."
Models plus observations of cooling tower plumes in a similar climate and
mathematical models were used to provide answers to this question. Un-
fortunately, because of the lack of demonstrated proof of the accuracy of
the model used and the shortage of quantified observations at operating
cooling towers of (comparable size in a similar climate), 100% confidence in
the predicted changes of temperature and humidity in the orchard is not
possible. Compounding the uncertainty of the conclusion is the fact that
biologists cannot state how large the temperature and humidity changes
would have to be in order to affect the fruit trees, lower crop yields,
increase the incidence of fungus and other plant diseases, increase insect
populations, etc.
There are a number of meteorological questions that are raised durinq the
environmental review process that do not have provable (in either the
scientific or legal sense) answers. These questions, which are valid ones,
relate mostly to mesocale effects, such as local climatic changes, generation
of clouds, snow showers, thunderstorms, and tornadoes. Unfortunately, the
state-of-art in cloud physics and other phases of meteorological knowledge
do not permit us to predict with any degree of certainty what will be ob-
served downwind of a group of wet cooling towers. Snowfall from cooling
towers has been reported many times under very cold winter conditions; in
one case, 140 mm (5.5 in) of snow was measured downwind off a complex of
natural draft cooling towers in West Virginia [1].
Given below are samples of questions, raised by the invenors and the ASLB
related to mesocale weather effects raised at the ASLB hearings for a pro-
posed two-unit nuclear power plant in an area of high frequency of tornadoes:
"The waste heat released to the atmosphere could also increase the
incidence of turbulent weather, fog, icing, inversions and possibly
climate changes and tornado incidence."
"The PSAR* recognizes the climatic effects of the thermal plume. Obviously
tremendous amounts of energy are present as a result of the thermal plume,
plus the effect of the up-drafts. Intervenors believe that, given the
proper climatic conditions (conditions which are not unusual in this area)
the energy and up-draft will contribute to the existence of additional
precipitation and spawn tornadoes. The PSAR appears to admit that there are
uncertainties in this area but dismiss the implications without conducting
the necessary experimental basis for rejecting the consequences. As part
of their answer, Intervenors request and require that Staff and Applicant
test the effects of the thermal plume under appropriate climatic conditions
and furnish the results thereof. If these tests were conducted, and it is
the responsibility of Staff and Applicant to do so, the issue could be
resolved.11
*Preliminary Safety Analysis Report, prepared by the utility.
726
-------
"Intervenors contend that the applicant and regulatory staff have
inadequately considered the effect of the plume of the (proposed
facility) cooling towers in the following areas: increased precipi-
tation: spawning tornadoes.
"Although we have the many days of sunshine and wind we also have
many tornadoes and earthquakes. If you think you know someone who
can predict those things are what will happen in them then you have
fools for advisors." How would you like to try to answer this type of
question, under oath, in a adversary environment?
Sometimes, intervenors use "overkill" in their questions and comments in
an attempt to stop the licensing of the facility. These contentions usually
are relatively easy to answer. Given below is the sworn testimony of a
highly respected professional meteorologist given under oath at the environ-
mental hearing for a proposed two-unit nuclear power in the northern part
of the country using a cooling pond. "The second feature of this winter
situation is that the fogging will almost'certainly occur as liquid drops
at sub-freezing temperatures. All exposed surfaces will be subject to rim-
ing and glazing. Roads within a mile or so downwind of the pond are likely
to be impassable throughout the winter. How far out occasional espisodes of
hazardous driving conditions are likely to extend is speculation." Under
cross examination, the witness was forced to admit he had never seen a
cooling pond in operation, and that his estimates of distance of fogging
and the duration of icing were too high.
If all or part of the written or . i testimony of a witness at the hearing
is shown to be incorrect or otherwise unsatisfactory, the remainder of his
testimony is (and should be) given low probative value. The above state-
ment is especially true for witnesses for the utility or the NRC. In other
words, one incorrect bit of information can destroy the creditability of
the witness for all other issues as well.
Some contentions are false, and can easily be refuted. One example was made
at a recent hearing (I have altered the wording of the question slightly):
"local observation indicated that when the plant was shut down on (date)
for three months, it was because so much snow and ice had deposited
on (a local major highway), about 0.8 km from the plant that the utility
feared suit in case of accidents."
This nuclear power plant was shutdown for refueling before heavy natural
snowstorms created this traffic hazard.
A question that has been asked of several locations that is hard to answer is
that of acid rains caused by the merger of S02 and other gases from fossil
smoke stacks and cooling tower plumes:
"the interactions of the plume and the vapors from said plant with emis-
sions of oxides of sulfur and particulates from other existing fossil fuel
plants in the area, including a fossil fuel plant located within one mile
of the city of and with temperature inversions, common to the area,
727
-------
will produce unacceptable adverse effects on the historic building and
property in the city of _ and to the health of the citizens of the
city of ."
This particular contention was easy to answer, as wind conditions favorable
for plume merger carried the merged plumes away from the city in question.
But proving that significant impacts due to a merger of fossil chimney
eflluents and cooling tower moisture will not occur remains.
Fortuantely, one comment heard frequently a few years ago is rarely used now:
"Build a cooling tower and all the thermal problems will disapear." This
is simply a false statement. While certain thermal effects are reduced or
eliminated, others are created which, for a specific location, could be worse.
This is especially true if a plant is forced to retrofit to another cooling
system either during construction or after the start of operation.
A problem of communications with the boards and the public is the lack of
precise meaning of certain words. "Salt" means "NaCl" to most people; there
is little of this material in the blowdown and drift from cooling towers at
inland sites. The effects of NaCl on plants, metal, etc., are quite dif-
ferent from that of the CaS04 and other materials in drift. "Fog" and "ice"
are other poorly understood terms. Most people at these hearings feel that
fog is only present when "one cannot see his hand in front of this face"
and ice is hard and dense like ice cubes. Cooling-tower fog that restricts
visibility to 300 m (1000 ft) is not a traffic hazard but is "fog" to meteor-
ologists. Ice produced by cooling tower plumes is light, friable rime ice
of little strength and very low density; such ice does not cause damage to
structures or vegetation. Cooling tower fogs rarely cause ice on clear
road surfaces. At one hearing concerning a cooling pond, I showed photo-
graphs of steam fog over the pond; the reaction was "Is that what we are
talking about? Forget it. Go on to the next topic." I have used movies
and color slides at public hearings; they did a much better job of explain-
ing what happens than several pages or days of testimony. I strongly urge
all of you to document your observations with photographs and movies (be
sure to include date, time, weather conditions, etc.), as they are very
effective pieces of evidence. The old saying that a picture is worth a thou-
sand words is very applicable in EIS work.
WAYS TO IMPROVE THE EIS PROCESS
In my opinion, which is shared by most people in the field, the primary
reasons meteorologists are not able to make accurate, quantitative estimates
of the atmospheric effects of cooling-system operation required by the NEPA
review process is the lack of systematic detailed observations made at operat-
ing power plants. Therefore, there is a need for a series of major field
experiments at power plants with mechanical-draft and natural-draft cooling
towers, spray canals, once-through cooling, and cooling ponds One result
of these field observations would be to clearly identify and quantify the
environmental problems caused by cooling systems, and to indicate which of
728
-------
the postulated issues are in fact nonproblems and need not be considered
further. Another result of equal importance would be the construction of a
suitable data base that would al>ow mathematical and physical models to be
developed and adequately tested. These models could than be used to predict,
with accuracy and confidence, conditions at proposed power plants in other area
areas. As a result, multimillion-dollar design decisions, which are'now
being based on very poor information, would be supported on a more accurate
and complete assessment of cooling-system effects.
The observations would also be used to formulate "rules-of-thumb" that could
be used in determining the environmental acceptability of a specific cooling
system on a given site. For example, if more thorough observations show that
fog from MDCTs and cooling ponds does, in fact, always or almost always
evaporate or rise above the surface within a short distance, then no model
would be needed for acceptance of such a cooling system on another site.
But "how far is far enough" remains a valid question requiring a quantita-
tive answer that can only come through observations over a wide range of
meteorological conditions at operating cooling systems. Such "rules" could
be used by the decision-making agencies without the need to model each and
every proposed plant. These agencies could also dismiss with confidence those
environmental concerns that have raised at public hearings but which are
known to be insignificant (such as temperature and humidity changes downwind
of NDCTs) or do not in fact occur (fog downwind of NDCTs).
Finally, the studies would lead to a series of generic reports that would be
very useful to the regulatory and licensing agencies and for educating the
general public. If such reports were now available, much of the wheel-spin-
ning that is going on in the EIS procedure would be eliminated, with a con-
siderable savings of time, effort and money for all parties involved. These
generic reports should summarize and evaluate our present knowledge of cool-
ing tower effluents, and include a critical comparison of the models now
available. If such generic reports or "rules" were now available, each
utility would not be required to "reinvent the wheel" each time it generated
an ER.
There is a large amount of good factual data and information locked up in the
files of cooling tower manufacturers, power companies and their consultants.
This data, if properly summarized and published, would demonstrate the actual
impact of cooling towers on the air environment, would shorten the time and
effort to complete a NEPA review, and could lead to a better selection of
cooling system for a specific power plant. The dollar and time savings
would be very large. The legal and other values of such studies buried in
classified files are zero.
SUMMARY
Due both to shortages of cooling water and to regulatory actions, future
power plants—both fossil and nuclear—will use evaporative closed-cycle
cooling systems to dissipate their waste heat directly to the atmosphere.
729
-------
Unfortunately, the state-of-the-art of atmospheric knowledge and modeling is
such that meteorologists are not now able to predict quantitatively how the
atmosphere will react to the large amounts of heat energy and water vapor
that it will be forced to absorb from limited areas fo cooling towers, cool-
ing ponds, and spray canals. Conceivably, critical heat release rates may
exist which, if exceeded, may lead to considerable effects, such as the
formation of extra precipitation, severe storms and/or tornadoes.
Closed-cycle cooling methods reduce but do not eliminate chemical and thermal
discharges into the aquatic medium; they transfer the primary area of impact
from hydrosphere to the atmosphere. These cooling systems do create adverse
atmospheric effects (such as fogging and icing, noise, drift, greater evapora-
tive loss of water, esthetics, etc.) which may be environmentally unacceptable
at some sites.
Because evaporative or wet cooling towers provide a convenient, dependable,
economical, and well-understood method of rejecting heat directly to the
atmosphere, they are usually chosen as the means of heat rejection for
power plants and large industrial plants. ' Where sufficient level land is
available near the plant at moderate (i.e., farmland) prices, cooling lakes
or spray canals may be utilized. Occasionally, to meet some stringent con-
dition such as the lack of cooling water at a mine-mouth plant, a dry cooling
system will be installed, even for a large heat-load plant.
The primary impact of the operation of natural-draft cooling towers is their
visual bulk and the formation of visible plumes that remain aloft. Plumes
as long as 80 km may be generated; under certain weather conditions, snow
does fall from these plumes. Most of the postulated adverse impacts—such
as fogging, acid mist formation, noise, and the wetting, icing, and salt
deposition due to drift with towers using fresh water for makeup and state-
of-the-art drift eliminators—do not, in fact, occur.
Aerodynamic downwash frequently brings the plume from mechanical-draft cool-
ing towers to the ground next to the tower. The plumes will evaporate or
lift due to their buoyance to become a cloud deck within a short distance
(of the order of 0.4 km). Thick deposits of light, friable rime ice may be
generated in this zone. Most of the drift that does fall to the ground will
do so within the same distance. Thus, areas of adverse impacts are limited
to those quite close to the cooling towers—the exclusion area required for
nuclear power plants. Observations at operating facilities indicate that the
rate of deposition of salts from drift is very low and below the threshold of
injuring to vegetation. Recent improvements in drift eliminator design reduce
the already low drift rates by an addition order of magnitude or more.
Salt-water towers equipped with these devices could become environmentally
acceptable even in areas of salt-sensitive crops.
The question of the degree of mesoscale weather modification by cooling tower
plumes, either wet or dry, cannot be satisfactorily answered at this time
due to our lack of understanding of the atmospheric processes involved and
the inadequacy of available modeling techniques. Very little information is
currently available on the possible effects of large plumes on severe weather
730
-------
events such as thunderstorms, hail, severe rainstorms, and tornadoes. Some
observers think that severe thunderstorms, and even tornadoes, can be caused
by cooling tower effluents during very unstable weather situations. This
question remains unanswered.
Cooling ponds and spray canals will cause frequent fogging over the w-ater
surface; this fog may move inland several hundred meters before lifting,
becoming very thin, or evaporating, Because of the larger area of heat
release from ponds and canals, fogging and icing conditions are less severe
near these cooling options than near mechanical-draft cooling towers. There
is no drift from a cooling pond; drift and icing near spray canals can be
heavy but restricted to a hundred meters or so from the canal.
When I was a graduate student at the University of Chicago many years ago,
I was furtunate in having as one of my professors Dr: Carl-Gustaf Rossby,
the great Swedish/American meteorologist. One day in class in the late 1940's
1940's (this is in the B.C. era: before computers), he had a long, complex
set of partial differential equations on the blackboard. His comment, which
still has relevance in the A.C. (after computers) era and should not be
ignored, went something like this: "I cannot solve these equations. Nature
can and does solve them, with no approximations or assumptions. All I need
to do to get nature's accurate solution is to carefully read the weather map."
Nature can and does solve the complex set of equations related to cooling
tower emissions; all we need do to find the exact solutions to these complex,
not fully understood physical processes—with no approximations, no assump-
tions, no errors due to finite grid sizes and time steps—is to make careful,
detailed observations at operating cooling towers. The answer is that
modern cooling towers in good repair and properly sited, have a low and ac-
ceptable impact on the air environment.
REFERENCE
1. R. E. Otts, "Locally heavy snowfall from cooling towers," NOAA Tech.
Memo. NWS ER-62, December 1976.
731
-------
A DESIGN MliTMOD FOR DRY COOLING TOWERS
G.K. Vangala and I.E. Eaton
Mechanical Engineering Department
University of Kentucky
Lexington, Kentucky, USA 40506
ABSTRACT
The cumulative cost of a large dry cooling tower is minimized by simul-
taneously optimizing the heat exchanger area and air friction losses; the
optimization parameters are sensitive functions of the initial temperature
difference (ITD).
For a given value of ITD, the air temperature rise varies from its opti-
mum value because, even though the air friction losses are minimal at the
optimum air temperature rise, the heat exchanger area decreases monoton-
ically as the normalized air temperature rise decreases.
INTRODUCTION
The continuing growth in demand for electric power requires planning, sit-
ing, and construction of large central generating stations. These stat-
ions use a regenerative Rankine energy conversion cycle which typically
rejects two-thirds of the energy input as waste heat and which requires
low heat sink temperatures for high efficiencies. Most power stations
reject heat using cooling towers which transfer heat from water to air.
There -are two basic cooling tower types, i.e., wet and dry; see Parker
and Krenkel [l] for a. detailed introductory discussion. However, few
dry towers have been used for cooling central electric generating stat-
ions. The wet or evaporative cooling tower has the advantages of a lower
cost and a lower, sink temperature (the wet bulb temperature) than a dry
tower. A major disadvantage of wet towers is the high water consumption
and the associated environmental effects due to fog and drift. While
the dry cooling tower approaches the dry bulb temperature, it neither
consumes water nor releases moisture into the atmosphere.
Typcia.1 schematics of dry cooling towers are shown in Figure 1. The nomen-
clature used in this work is listed at the end of the text and illustrated
in Figure 2.
OBJECTIVES
Because of problems associated with power plant siting, the dry cooling
tower is receiving serious consideration as a heat rejection technique
732
-------
for large central generating stations. However, large dry cooling towers
are inherently expensive,and a careful cost optimization results in large
capital savings.
Although the optimization of various dry cooling tower design parameters
has received considerable attention, little has been reported on the over-
all design of a cost optimized dry cooling tower. This paper presents a
method for dry cooling tower design which uses the existing literature
and standard sourcebooks on heat exchanger design and performance to ar-
rive at a cost optimized dry cooling tower for a specific application.
The general method presented incorporates a detailed analysis of heat
exchanger performance and sizing based on the results of intensive work
reported by others which pertains to dry cooling tower design.
DRY COOLING TOWER DESIGN METHOD
The design of a dry cooling tower is determined by the ambient dry bulb
temperature, the hot fluid temperature, the heat rejection rate, the
hot fluid properties, and the optimum cooling system cost.
The design method presented proceeds by determining the air friction losses
and heat exchanger area based on the optimum air temperature rise across
the heat exchanger. Next, a specific heat exchanger type is selected, and
the number of tube columns and fluid passes, as well as the tube length
are determined. The best heat exchanger type is selected from the alter-
natives considered based on an evaluation of cost and performance as is
illustrated in Figure 3A.
With the type of heat exchanger selected, the optimum tower design is
established based on cumulative cost while the air friction losses and
air tempei-ature rise are varied as is illustrated in Figure 3B.
Air Friction Losses
The basis of the air friction loss analysis for the design method present-
ed is essentially adapted from work by Moore [2,4]. The tower is treated
as a duct, and expressions for air friction losses, i.e., either tower
height or fan power, are obtained.
Moore [3] has defined an air friction loss function as:
^ is minimized for various values of normalized approach, and correspond
ing values of normalized air temperature rise were found. From Table 1,
it may be seen that the normalized air temperature rise is not a strong
function of the normalized approach; therefore, in this work, iterations
to establish the optimum tower cost were initiated with a normalized air
733
-------
temperature rise of 0.83, irrespective of the value of the normalized ap-
proach.
Heat Exchanger
With this, the heat exchanger is designed such that it operates at the
predetermined optimum conditions. Knowing the heat rejection requirement
and the optimum operating conditions, the total heat transfer area may be
established [3,4].
Next, using heat exchanger design sourcebooks, cf. , references [5], [6],
and [7]) and work on the Rugley dry tower [8], a specific heat exchanger
type and its design parameters, i.e., K, E, and F, are selected simultan-
eously. K is the ratio of friction coefficient to Stanton number (K =
f/St) [6]; E is a coefficient combining water side resistance and air
side effectiveness, see reference [7]; F is a coefficient expressing
counterflow equivalence [5].
A particular category of heat exchanger which has small values for K and
hydraulic radius must next be chosen. The weight, of heat exchanger per
unit air side heat transfer area is minimum for a device in which the fin
surface area dominates. The plate-fin type heat exchanger typically will
have the ideal surface-to-mass ratio. Reference [5], Figures 10-52 through
10-64, illustrates various plate-fin heat exchanger designs. One must
select and examine several heat exchanger designs in order to determine
the design which optimizes cost and performance.
The selection of a specific heat exchanger type establishes the geometric
parameters of the heat exchanger. Then, computations are made to evaluate
water travel distance, number of water passes, and the number of water tube
columns. An iterative procedure which varies K, E, and F is used to op-
timize cost and performance for the heat exchanger type selected. The
optimized cost and performance results for several heat exchanger types
are then compared to establish the optimum heat exchanger type.
Cumulative Cost Optimization
Next, the total tower cost is optimized by varying the air temperature
rise and the air friction losses.
The optimum air temperature rise for which calculations have been made.
only optimizes air friction losses. In order to minimize true heat ex-
changer area, the air temperature rise must be smaller than the air fric-
tion optimized value [9], see Figure 4A. The dashed lines show the var-
iation of tower size with air temperature rise while the solid lines show
the variation of the heat exchanger area.
Thus the air temperature rise value is varied, and the corresponding
variations of air friction losses and heat exchanger area are evaluated.
This procedure will determine the optimum air temperature rise based on
734
-------
a cumulative cost criterion.
From Figure 4B, it may be seen that the air friction losses (tower height)
approach their minimum value at large values of heat exchanger area. In
order to conclude the cost optimization, the ratio of the design tower
exhaust area to the tower exhaust area corresponding to the optimum air
friction losses is iteratively determined.
APPLICATION OF THE METHOD
The application of the dry cooling tower design method presented will be
illustrated for the case of a 200 MWe electric generating station with
seasonally varying load and ambient conditions. Typical variations of both
the system load and the ambient dry bulb, temperature during 1976 are il-
lustrated in Figure 5. The peak load may be seen to occur when the ambient
temperature is the lowest thus improving the attractiveness of a dry
cooling system.
The complete dry cooling tower design optimization procedure will be repeat-
ed for each month and the solutions compared. Ultimately, a tower design
will be obtained that includes the optimum air temperature rise, the opti-
mum ratio of exhaust area to its minimum value based on air friction losses,
and the optimum type of heat exchanger which will minimize the cumulative
tower cost.
The numerical results for this case will be presented in detail at the
Conference.
735
-------
REFERENCES
[l] J.L. Parker and P.A. Krcnkel, Plr/jn^n^nnM^^
Thermal PoUution_ (Cleveland, CRC Press,, 1970).
[2] F.K. Moore, "On the Minimum Size of Large Dry Cooling Towers with
Combined Mechanical and Natural Draft," Journal^ ofJieat Transf er,
Vol. 95. Series C, August 1973, pp. 383-389.
[3] B.M. Johnson and D.R.' Dickenson, "On the Minimum Size for Forced
Draft Dry Cooling Towers for Power Generating Plants," Dry and Wet/Dry
Cool-ing Towers for Power Plants (New York, ASME, 1973), pp. 25-34.
[4] F.K. Moore, "Dry Cooling Towers," Advances in Meat Transfer, Vol. 12
(New York, Academic Press, 1976), pp. 1-75.
[5] F. Kreith, Principles of Heat Transfer, 3rd Edition (New York,
Intext PressT 1973).~'~
[6] W. Kays and A.L. London, Compact Heat Exchangers, 2nd Edition (New
York, McGraw-Hill, 1964).""
[7] F.K. Moore, "Minimization of Air Heat-Exchange Surface Are;js in Dry
Cooling Towers for Large Power Plants," Dry and Wet/Dry Cooling Towers
for Power Plants (New York, ASME, 1973) .^PpTT^Z^
[8J P.J. Christopher and V.T. Forster, "Rugeley Dry Cooling Tower System,"
Proceedings of the Institution of Mechanical Engineers, Vol. 184,
Part I, No. 11, 1969-70, pp. 197-221T"
[9] F.K. Moore and T. Hseih, "Concurrent Reduction of Draft Height and
Heat Exchange Area for Large Dry Cooling Towers," Journal of Heat
Transfer, Vol. 96, Series C, August 1974, pp. 279-"285~!
[ 10] M.W. Larinoff, "Dry Cooling Power Plant Design Specifications and
Performance Characteristics," Dry and Wet/Dry Cooling Towers for
Power Plants (New York, ASME, 1973), pp. 57-83"
[ll] J.P. Rossie and E.A. Cecil/'Research on Dry-Type Cooling Towers
for Thermal Electric Generation: Part I," Water Quality Office,
U.S. EPA Report 16130EES11/70.
736
-------
NOMENCLATURE
A = Total Heat Exchanger Area
a
A = Heat Exchanger Free Flow Area
A = Cooling Tower Exhaust Area
E = A Coefficient Combining Water Side Resistance and Air Side
Effectiveness
f • Friction Coefficient
F = A Coefficient of Counterflow Equivalence
I = Initial Temperature Difference
K = Ratio of Friction Coefficient -to Stanton Number
p = Ratio of A to A -Minimum, Based on Air Friction Losses
Q = Approach
St = Stanton Number
THX = Type of Heat Exchanger
a ' Air Temperature Rise
»
a = Optimum Air Temperature Rise (Optimized for Air Friction Losses)
T = Tower Size
*
¥ = Optimum Tower Siie (Optimized for Air Friction.Losses)
Superscripts
$ = Cost Optimized Parameters
* = Parameters Optimized for Air Friciton Losses
Subscript
I = Normalized with Respect to Initial Temperature Difference
Table 1
VARIATION OF OPTIMUM, NORMALIZED AIR TEMPERATURE RISE WITH
CHANGES IN NORMALIZED APPROACH
Normalized'Approach (Q.) Optimum, Normalized Air
i *
Tcmpcrnturc Risn (a. V
0.2 0.800
0.3 0.313
0.4 0.822
0.5 O.S29
0.6 0.835
0.7 0.840
0.8 0.844
0.9 O.R4S
1.0 0.857
737
-------
NATURAL-
DRAFT TOWER
COOLING COILS
V
If
cu
AIR
FLOW
v
\
\i
FIGURE 1: TYPICAL DRY COOUNG TOKER SCHT-."AT1CS
(Taken from Reference [ll])
CO
MECHANICAL-
DRAFT TOWER
COOLING COILS
WATER FLOW
Fir.URE 2; TFHI'RRATURr DIAGRAM OF A TYPICAL DRY
COOLING TOWER HRAT FXCHANQ-R
738
-------
FTGURl: 3A- I^.Y_cnju,r;r. TOWER n'-.r.Tr.N MF/n
HEAT EXCHANGFiJ Di'SHIN ni'TIMl/./VnON
DATA: AMBIENT TF.MFEKATURE . HEAT REJECTION KATE
PHYSICAL PROPERTIES
EVALUATE ,-ty* o( * [1,21
(1) SETtf=oC", (2) INITIALIZE P, (3) EVALUATE A , A [4]
CHOOSE A HEAT EXCHANGER TYPE (THX) [1,7]
USING [3,4], CHOOSE HEAT EXCHANGER PARAMETERS - E, K, F
EVALUATE WATERSIDE INFORMATION: (I) MW'F.ER OF TUBE COLUMNS,
(2) 'NUMBER OF WATER PASSES, (3) VJATFR TURF. LENGTH.
EVALUATE HEAT EXCHANGER PERFORMANCE AND COST
IF NO /
( OPTIMUM
\
IF YES
OPTItnjM VALUES OF K, E, and F.
CHOOSE ALTERNATE HEAT EXCHAMGER TYPE (THX)
EVALUATE WATERSIDE INFORMATION: (1) NUMBER OF TUBE COLUMNS
(2) NUMBER OF WATER PASSES, (3) WATER TUBE LENGTH.
EVALUATE HEAT EXCHANGER PERFORMANCE AND COST
IF NO /
/ OPTIMUM \
IF YES
OPTIMUM HEAT EXCHANGER DESIGN
(OPTIMUM THX, K, E, F)
•PROCEED TO TOWER CUMULATIVE COST OPTIMIZATION
739
-------
FICUJUT- 7>
.NG
MI'-TIIOD :
TOTAL TOWKR COST OPT I'M T.7ATION
*XfP- CONSIDER OI'TIMIIM HKAT EXC',!1A!1CF.R DF.STGfl ^jj^
(OPTIMUM TUX, K, 1C. K)
1
INITIALIZE P
1
CHOOSE "C (SF.T c<=oC")
1
1
EVALUATE Afi AMD Ag
!
EVALUATE WATERSIDE INFORMATION: (1) NUMBER OF TUBE COLUMNS
(2) NUMBER OF WATER PASSES, "(3) WATER TUBE LENGTH.
1
EVALUATE COST OF HEAT EXCHANGER
EVALUATE CUMULATIVE COST OF TOWER AMD HEAT EXCHANGER
]
1
IF MO / \
/ nPTTMTTM \
1
, IF YES
OBTAIN OPTIMUM VALUC OF o<
i
fc_
** CHOOSE A VA
'
LUF FOR P
EVALUATE A^ nnd Aa
1
r
EVALUATE WATERSIDE INFORMATION; (1) NUMBER OF TUBE COLUMNS,
(2) NUMBER OF WATER PASSES, (3) WATER TUBE LENGTH.
1
1
EVALUATE COST OF HEAT EXCHANGER
EVALUATE CUMULATIVE COST OF TOWER AND HEAT EXCHANGER
'
r s
It' NO / V
-a / OPTIMUM \
IF YES
OBTAIN OPTIMU1! VALUE FOR P
OPTIMUM DESIGN OF DRY COOLING TOWER
OPTIMUM K, E, F, Tt!X, <=( and P
r~—: -i
11£»- STOP -«ai
-------
25
02 04 06 00
Air Temp-eroiure Rise, a:
FIG. 4A: TOWER SIZE FUNCTION - * (Dashed
Lines) AND HEAT, EXCHANGER AREA
FUNCTION - a^2 (Solid Linos)
(Taken from Reference [7])
Ire;-Flo.. .VIM. Sc
FIG. 4B: INFLUENCE OF FREE-FLOW AURA - A
ON TOWER SIZE - A (Solid Lines)
AM) HEAT EXCHANCER AREA - A^
(Hashed Lines)
(Taken from Reference [/])
90
80
£ 70
J I I I I
I I I I
100 '^
80 «
60 ±
J F M A M J J A S 0 N D
MONTH OF 1P76
FIGURP: S: §_F.ASOS;AI;_VV\_RIAl'lO\_nF_ PI.AVI' 1 .0 A 0 .VS' P .VI 1! 1 1 • NT_ I ) R V
_ i-nu TIII: _IM:MI;N C.ASI- IISF.II TO
II.I.USTRATH Till; PKY COOI.INn TOWI-R PI-SKIN MITIIOD
741
-------
EVAPORATIVE HEAT REMOVAL IN WET COOLING TOWERS
by
Thomas E. Eaton
Mechanical Engineering Department
University of Kentucky
ABSTRACT
The ratio of evaporative-to-total (sensible plus evap-
orative) heat transfer in a wet, cross-flow, mechanical
draft cooling tower was analyzed. The ratio was found to
vary from 60% to 90% during typical operating conditions.
The evaporative heat removal fraction increased as temper-
ature (either wet-bulb or dry-bulb) increased and as re-
lative humidity decreased. Similar results were obtained
for a counter-flow, natural draft tower.
INTRODUCTION
Wet or evaporative cooling towers are commonly used to provide for
the cooling of water by direct contact with air. Two heat removal
mechanisms dominate in an evaporative cooling tower: evaporative
heat removal and sensible heat transfer. Sensible heat transfer
refers to heat transfered by virtue of a temperature difference
between the water and air. Evaporative heat removal refers to the
energy removal from the water as latent heat of evaporation; this
heat removal is the result of the evaporation of water into air
during the direct-contact cooling process.
The cooling tower industry typically quotes the fraction of energy
removed from the water by evaporative cooling as three-fourths or
about 757o. As will be shown in this work, the fraction of the heat
removed by evaporative cooling in wet cooling towers varies between
60% and 90% during typical operating and ambient conditions. It is
of interest to note that under certain conditions, both the water
and air are cooled by evaporation so that the evaporative cooling
exceeds 100% of the cooling effect on the water alone.
The water evaporation losses from wet cooling towers determine make-
up water requirements. Although the literature discusses the cal-
culation of evaporative losses (see Hamilton [1], for example), the
specific topic of the fraction of heat removed by evaporation and by
sensible heat transfer has not been quantitatively evaluated [2].
742
-------
w* cri » M In Hq
EVAPORATIVE
HEAT
TRANSFER
SENSIBLE
HEAT
TRANSFER
FIGURE 1: TYPICAL PSYCHOMETRIC CHART
WATER
ELE>5ENTAL VOLU11E (GRID) DETAIL
FIGURE 2: SCHEMATIC OF GRID LAYOUT USED TO ANALYZE A
CROSS-Fl'OW. MECHANICAL DkAFt COOLING
743
-------
This paper investigates the influence of variations in ambient con-
ditions and changes in cooling tower design parameters on the evap-
orative cooling-to-total cooling ratio in wet cooling towers. Par-
ticular emphasis is given to the commonly-used, mechanical draft
cross-flow cooling tower design. Results for a typical natural draft
cooling tower of counter-flow design are also given.
Fundamental Cons i derat ions
The enthalpy H of a mixture of air and water vapor is given by
Hm = 0.240 Td + W (1041 + 0.444 Td)
where W is the humidity ratio (Ibm water vapor/lbm dry air) and T,
is the dry bulb temperature of the mixture. From the psychometric
chart, see Figure 1, it may be readily determined that the enthalpy
remains nearly constant at constant wet bulb temperature TW.
Sensible heat transfer involves an increase in the dry bulb tempera-
ture of the mixture but evaporative heat transfer involves a change
in the humidity ratio of the mixture. Thus, a sensible heat trans-
fer from water to air inside a cooling tower involves a horizontal
change on the psychometric chart while evaporative transfer involves
a vertical movement as is illustrated in Figure 1.
In a wet cooling tower, where the tower-on temperature is greater
than the ambient wet bulb temperature, the air humidity always in-
creases as the air passes through the tower. However, as will be
demonstrated later, sensible heat transfer may be either positive or
negative. When the tower-on temperature is less than the ambient dry
bulb temperature, the sensible heat transfer may be negative and the
air dry bulb temperature will be lowered as the air passes through
the tower; under these circumstances, the air as well as the water is
cooled by evaporative transfer in the cooling tower.
In this paper, the total heat transfer will be taken as the evapora-
tive plus the sensible heat transfer to the air as it passes through
the tower.In cases where air cooling occurs in addition to water
cooling, the ratio of air-side evaporative and sensible heat trans-
fer to the water-side heat transfer will be greater than 10070.
Consider a counter-flow natural draft cooling tower for example; in
this case the exhaust air conditions are usually saturated. If the
ambient conditions are known, say 72 F and 507,, relative humidity,
and the exit conditions are determined as sav 96 F (10Q70 RH) , then
the air dry bulb temperature increases by 24 F (from 72 F to 96 F),
the humidity ratio increases by 0.030 Ibm-WV/lbm-DA from 0.0084 to
0.0380), and the mixture enthalpy varies by 38.5 BTU/lbm (from 26.5
to 65.0). Based on the air-side information only, the fraction of
heat rejected by evaporation can be estimated: AW(h,- )/AH, or (0.030)
(1040)/38.5 = 807o. f§
However, because the exit air conditions vary with fill height in a
cross-flow tower, the average exhaust conditions must be determined
before the evaporative cooling fraction can be estimated.
744
-------
CROSS-FLOW COOLING TOWER ANALYSIS
The computer program used for the analysis of cross-flow, evaporative
cooling towers was developed using the enthalpy-difference driving
force model to calculate the combined effects of heat and mass
transfer in the cooling tower. The basic equations are similar to
those presented by Kelly [3] or Hallett [4].
The cross-flow cooling tower packing is divided into a grid as
shown in Figure 2. In the upper, air inlet corner of the packing,
the air and water inlet conditions to the grid are known. With
this, the water outlet and air outlet conditions may be calculated
for the first grid. The air inlet coridition for the next grid
element is then known, and the program analysis proceeds across the
tower fill. At the end of the grid row, the calculation proceeds
to the air inlet of the next row down. In this manner, the entire
fill is analyzed. Both air outlet and water outlet conditions along
the fill are predicted by the code.
Because the code uses the enthalpy-difference for the driving force,
little information about the air inside the cooling tower is known.
Typically, only the air enthalpy is calculated. The wet-bulb
temperature is nearly constant with air enthalpy over a wide range
of relative humidities so that the wet-bulb temperature is also
known to a good approximation.
From routine observations, it is apparent that the air entering the
tower is usually not saturated and the air leaving the tower is not
always saturated. The enthalpy-difference driving force model is
only capable of predicting the local enthalpy and wet-bulb temp-
erature inside the tower fill; local air humidity and dry-bulb
temperature are not predicted when using this analytical method.
Calculation of Humidity Inside the Tower Fill
Because the calculation of evaporative cooling was of primary interest
in this work, considerable effort was devoted to attempting to
calculate the humidity at each grid inside the tower fill. The gov-
erning equations based on a "humidity-difference" driving force
inside the tower were developed in a manner similar to that of Park
and Vance [5,6], However, the humidity-difference driving force
method was abandoned because the appropriate empirical correlations
for the tower characteristic (Ka) were not available.
Tower characteristics developed using enthalpy-difference methods to
analyze data cannot be used to predict cooling tower performance
using a humidity-difference driving force.
In order to determine the humidity variations inside a wet cooling
tower, the humidity ratio change between successive grid points
must be known, i.e., the amount of water evaporated inside the grid
745
-------
must be calculated. Conventional cross-flow cooling tower analysis
methods can not do this. An improved method for cross-flow cooling
tower analysis has been developed by Baker and Eaton [7], but was
not available for this work.
For this work, the air humidity inside the cooling tower was estimated
assuming that the humidity ratio change within a grid was that pre-
dicted by the humidity ratio change along the air-water vapor mixture
saturation line. That is, the entering and leaving air enthalpies^
are calculated so that the humidity ratio change along the saturation
(100% relative humidity) line could be determined. Since the air
humidity at the tower inlet was known and the change in humidity
ratio in any calculational increment was determined in the manner
described above, the humidity at any point inside the tower could
be estimated.
Although this method was not exact, it was felt to be more accurate
than the assumption that the air inside the tower is saturated.
Occasionally, the above internal humidity calculations predicted
supersaturated air conditions; if such calculations occurred,
the air was assumed to be saturated at the calculated enthalpy.
The basic equations used to analyze the performance of the cross-
flow cooling tower are discussed in detail in Appendix A.
Other Computer Code Information
For the analysis of cross-flow cooling towers, the tower characteristic
equation developed by Hallett [4] was used throughout this work:
Ka
0.120 G°'410 L°'525 (Eq. 2)
As may be seen in Figure 3, this correlation was found to predict
the performance of the reference design cooling tower with +2 F
and -0 F of the tower manufacturer's performance curves. The details
of the reference design cooling tower are given in Table 1. This
tower design was the basis for all results reported on cross-flow
towers.
Properties of air-water vapor mixtures were calculated using the ideal
gas equations, see the ASHRAE Brochure on Psychrometry [8] for details.
Water properties were calculated based on data from the 1967 ASME
Steam Tables.
Outlet Air and Water Temperature Distributions
For the case of a cross-flow cooling tower, hot water enters the top
of the tower fill at a uniform temperature. This water is cooled by
air admitted from the side of the tower. Because of the cross-flow
cooling arrangement, the temperature of the water at the bottom of the
746
-------
TABLE I
DESIGN PARAMETERS FOR A TYPICM.
INDUSTRIAL CROSS-FLOW COOLING TOWER
Cooling Tower Type Cross Flow Design,
Double Flow Arrangement
Draft . Mechanical, Induced Draft
Number of Tower Cells 12
Dimensions
Overall Length 432 ft
Air-Travel Distance 18 ft
Water-Travel Distance 36 ft
Fan Diameter 28 ft
Stack Exhaust Diameter 31.5 ft
Cell Length 36 ft
Design Operating Conditions
Total Air Flow (G*) 1,410.000 CFM/CELL
Total Water Flow (L*) 190,000 GPM
L*/G* 1.44
Wet Bulb Temperature 76°P
Tower-On Temperature 117°F
Tower-Off Temperature 90°F
Range 27°F
Approach 14°F
Liquid Loading (I.) 6100 Ibm/hr-ft2(12.3 GPM/ft2)
Air Loading (G) 2150 lbm-DA/hr-ft2
L/G 2.88
Velocity at Fan 2400 ft/sec
Velocity at Stack Exhaust 1800 ft/sec
Tower Characteristic
KaX/G 2.26
tfaY/L 1.60
Ka 270
Fan Power 175 HP
Note: Relative Humidity of Inlet Air
Does Not Effect Cold Water Temp
•This work using Hallett-J
Correlation
A ' 50 ' 60 ' '
WET-BULB TEMPERATURE (°F)
70
80
FIGURE 3; COMPARISON OF COOLING TOWER PERFORMANCE CURVES SUPPLIED
BY* MANUFACTURER VS. CALCULATED IN THfS WORK FOR THE
REFERENCE DESIGN COOLING TOTTER
-------
fill as well as the temperature and humidity of the air at the inside
of the fill varies with position.
It was necessary to determine the fill outlet variations of water
temperature, air dry-bulb temperature, and relative humidity in order
to adequately assess the heat removal from the water and the total and
evaporative transfer to the air. The evaporative heat removal fraction
was determined as the average of the ratio of evaporative heat removal-
to-total air heat removal for each row of fill calcualtions:
Q N
IX = 1 **? h (W . - W. .) / (H . - H. .) (Eq. 3)
Q*. *. N ,4-r fg oj ij' oj ij
tot j=1
The average outlet water temperature was calculated by averaging the
outlet water temperatures:
T^ — T —
iwo ~ ioff ~ H .
J=J-
With this, the total heat removed from the water is
cpw Ton * *wo cpw Toff (E<*' 5)
The correction for water evaporation loss has been included in the
heat rejection equation.
Typical results of air and water outlet variations are shown in
Figures 4 and 5. The cooling tower design parameters are given in
Table 1; the results plotted are for a hot water temperature of
119°F with 76°F wet-bulb and 81°F dry-bulb temperatures (80% relative
humidity) as the ambient conditions.
748
-------
6-
w
OT
g
i.o
0.8
0.6
0.4
u
a
Cd
N
3 0.2
0.0
Inlet
Conditions
80
Dry Bulb Temp. -
—Wet Bulb Temp.
———Relative Humidity
90
100
110
TEMPERATURE (°F) , RELATIVE HUMIDITY (7.)
FIGURE 4: VERTICAL VARIATION OF AIR-SIDE CONDITIOMS FOR THE REFERENCE
120
DESIGN COOLING TOWER
120
110 -
hi
O
g 100 -
I
0.0
0.2 0.4 0.6 0,8 1.0
NORMALIZED HORIZONTAL POSITION
FIGURE 5: WATER TEMPERATURE DISTRIBUTION IN THE
REFERENCE DESIGN COOLING TOWER
749
-------
THE EVAPORATIVE HEAT REMOVAL FRACTION
Cross-Flow Cooling Tower Analysis
The evaporative heat removal fraction as influenced by the variation of
typical cooling tower operating conditions or design parameters was
evaluated using the computational methods described earlier for a cross-
flow, mechanical draft tower design.
Cross-Flow Cooling Tower Parameters
The cross-flow, mechanical draft cooling tower is common throughout
the United States. For this reason, the influence of the variation
of several parameters on the evaporative cooling fraction was invest-
igated for a typical, large industrial cross-flow tower. The goal
in studying the parameter variations was to determine their effect
on a cooling tower of fixed design. The parameters varied are listed
below:
(A) Ambient Air Wet-Bulb Temperature
(B) Ambient Air Relative Humidity
(C) Hot Water (Tower-On) Temperature
(D) Cooling Tower Range
(E) Liquid Loading
(F) Air Loading
(G) Cooling Tower Characteristic (Ka)
(H) Elevation Above Sea Level.
Typically, the cooling tower parametric variations were examined with
the total water flow to the cooling tower and the velocity of the
water vapor/air mixture specified at the tower exhaust. The water
loading and the mixture exhaust velocity were varied only in the
above items (E) and (F), respectively.
Recall that the specific cooling tower design parameters "used for the
reference design cooling tower (cross-flow design) are given in Table 1.
Parametric studies on cooling towers can, in general, be quite involved
because many design parameters can be varied. Indeed, there are many var-
iablies which can be varied which have not been considered here, e.g.,
fill height, fill width, and packing design - to name just a few!
Typical Behavior
The variation of the evaporative heat removal fraction in the reference
design cooling tower operating at the design heat rejection rate
(2.6 Billion BTU/hr) under varying ambient conditions is shown in
Figure 6. The evaporative cooling effect varies from 60% to over 90%
750
-------
OS
Ed
Cb
a
100
90
•J 80
70
o
a.
i
Cb
O
6C-
s -t
I r
RANGE - 27°F
40 50 60
WET-BULB TEMPERATURE (°F)
1007.
FIGURE 6: INFLUENCE._OF AMBIENT CONDITIONS ON THE RATIO
Of EVAPORATIVE-TO-TOTAL HEAT TRANSFER IN THE
REFERENCE BEStCH COdLiMg TOWER"
FIGURE 7: EVAPORATIVE HEAT REMOVAL FRACTIOHVS._AMBIENT
CRT-BULB TEMPERATURE FOR VARIOUS RELATIVE HUMIDIES
10Q
S 8C-
$
a
60
a
e.
w ,20
25%
100%
75% Rel. Hum.
REFERENCE DESIGN COOLING TOWER
RANGE = 27°F
40
50
60
70 80
DRY-BULB TEMPERATURE
90
100
110
751
-------
during normal changes in ambient conditions. The fraction of heat re-
moved by evaporative cooling increases as the wet-bulb temperature
increases and increases as the relative humidity decreases (at constant
wet-bulb temperature).
The variation of the evaporative heat removal fraction in the reference
design tower operating at the design heat rejection rate (range = 27°F)
is shown plotted versus ambient dry-bulb temperature in Figure 7.
Note that for a fixed dry-bulb temperature, the relative humidity has
little influence on the fraction. The evaporative cooling fraction
increases about 1% for each 3 F rise in dry-bulb temperature.
Tower-On Temperature Effect
The evaporative heat removal fraction plotted versus tower-on (hot
water) temperature is shown in Figure 8 for wet-bulb temperatures of
40 F, 60 F, and 76°F at 80% relative humidity. For a given hot-water
temperature, the evaporative cooling fraction increases as the wet-
bulb temperature increases; also, a minimum fractional value can be
identified for a given wet-bulb temperature (at 80% relative humidity).
Figure 9 shows the effect of relative humidity on the evaporative
fraction versus tower-on temperature for a fixed wet-bulb temperature
of 60 F. The evaporative cooling fraction increases with decreasing
relative humidity. The minimum evaporative cooling fraction depends
on relative humidity at a fixed wet-bulb temperature.
Cooling Tower Range
The effect of the cooling tower's operating range (hot water temper-
ature minus cold water temperature) is shown in Figure 10A-D for ranges
of 30 F, 20 F, 10 F, and 5 F, respectively. As the range increases,
the variation in the evaporative cooling fraction due to. changes in
the wet-bulb temperature becomes less significant. During typical
operating conditions, i.e., a range above 20 F and a relative humidity
above 50%, the evaporative cooling fraction varies from 60% to 90%.
In general, the evaporative cooling fraction increases over all oper-
ating ranges as the wet bulb temperature increases and as the relative
humidity decreases (with wet bulb temperature constant). At low ranges,
i.e., less than 10°F, the fraction increases rapidly as the relative
humidity decreases. On hot dry days with ranges below 10 F, the evap-
orative cooling fraction will exceed 100%.
Recall that when the evaporative heat removal fraction is greater than
100%, the cooling tower is physically cooling both the water and the
air (by evaporative cooling) as it passes through the fill. That is,
the sensible heat transfer to the air is negative.
The effect of range variations for relative humidities of 80% and 2070
are shown in Figure 11.
The effect of relative humidity on the evaporative heat removal fraction
versus range is shown in Figure 12. At ranges below 20°F, the relative
humidity has a marked effect on the evaporative cooling fraction.
752
-------
100
--J
Ul
10
(1(1
o
s
I'
;.'i
ji;
!•••
o
PH
6
60
70
/6"K He I bull- Teni|>.
KKFKKKNCK UF.:i(RN
nrrf)F.l MO TOWKH
Relative Humidity
- 80%
60 90 100 110 120
TOWER-ON (HOT WATER) TEMPERATURE (°F)
FICUItK 8: EVAPORATIVE J1KAT KEtlOVAL RATIO
VARTobsn3F.r-BUr£TTEI1POATURLS~
B!
l-l
H
'KfL Relative Humidity
REFERENCE DESIGN COOLING 'I'JWLfc
60 F HET-BULfa 'fEfff-Eh>.TUk£
40
20 -
90 100 110
TOWER-ON HATEk TEMPERATURE
120
FIGURE 9: IHFLUENCE OF RELATIVE HlfttlDITY ON EVAF-OP^.TIVF.
' MATER
-------
140
3 loo
C
u 40
Relative Humidity 20V
LOOT.
RANGE - 30 F
50 60 70 80
WET-BULB TEMPERATURE ("n
2 100
.
2
u 40
20
FIGURE 108
RANtlE - 20°F
toor.
50 f.0 70 30
WET-BULB TEMPERATURE (°F)
140
120
s
I 100
Eb
<
o 80
I
I
« 60
1 I 7 T
Relaclv« llunidir.
FIGURE 10D
507.
RANGE - 5°F
50 60 70 80
WET-BULB TF.MPERATURF. (°F)
SO 60 70
WET-BULB TEltPERATURE (°
FIGURE 10: EVAPORATIVE HEAT REMOVAL FRACTION
yERSUS_WET-BULB TEMPERATURE FOR
VARIOUS RANGES
754
-------
Ul
Ul
140
120
100
H
O
2
1
a
I
I
80
60
40
20
207/ Rel. Humidity
REFERENCE DESIGN TOWER
30
40
50
60
,70
80
WET-BULB TEMPERATURE (°F)
FIGURE 11 : TMF. INFLUENCE OF RANGE ON THE EVAPORATIVE HEAT
REMOVAL FftACTreTTAT'm'and 201 RFXATiyE~TlWfDm
o
M
H
I
Q
I
I
140
120
100
80
60
40
20
20% Relative Humidity
_ 100%
WET-BULB TEMPERATURE - 60°F
REFERENCE DESIGN TOWER
0
10 20
RANGE
30
40
FIGURE 12: VARIATION OF EVAPORATIVE HEAT REMOVAL
FS5?f ?£..""' "
TEMPERATURE
-------
Cooling Tower Range (Dry-Bulb Temperature)
The effect of cooling tower range on evaporative cooling fraction
versus dry-bulb temperature is shown in Figure 13A-D for ranges of
30°F, 20 F, 10°F, and 5°F, respectively. The effect of relative
humidity variations is shown in each figure.
For the typical large, industrial cooling towerQconsidered (operating
under routine conditions, i.e., a range over 20 F) , the evaporative
cooling fraction increases linearly with dry-bulb temperature at a
rate of about 1% / 3 F. The fraction varies from 65% at 40 F. to about
90% at 100 F; changes in the ambient relative humidity have only a
slight influence on the evaporative cooling fraciton during normal
operating conditions (at constant dry-bulb temperature).
As the range decreases below 20°F, the relative humidity has an in-
creasingly important influence on the variation of the fraction with
dry-bulb temperature. At ranges below 10 F, the fraction increases
rapidly as the humidity increases. This sensitivity increases as the
dry-bulb temperature increases.
The same results of Figure 13A-D, are shown in Figure 14A-C; however,
the latter figure shows the effect of range variations at constant
relative humidity. Note that the effect of low operating ranges (i.e.,
less than 20°F) on the evaporative cooling fraction changes markedly
as the relative humidity decreases.
Liquid Loading - L
For the reference design tower, the effect of liquid loading (L)
variations on the evaporative-to-total heat removal ratio is shown
in Figure 15. Increasing the liquid loading produces a small decrease
in the evaporative heat removal fraction. The exhaust air mixture
velocity was held constant for this case.
Gas (Dry Air) Loading - G
The influence of varying the gas or dry air loading in the reference
design tower fill is shown in Figure 16. With the liquid loading held
constant, the evaporative heat removal fraction slightly decreases as
the gas loading increases.
Cooling Tower Characteristic - Ka
The effect of varing the cooling tower characteristic in the reference
design cooling tower is shown in Figure 17. Over the range of interest
i.e., Ka = 100 - 350, there is a small effect on the evaporative heat
removal fraction.
Elevation Above Sea Level
The barometric pressure varies with elevation and ambient conditions.
-------
Ul
-J
Relative Humidity
40 SO
RANGE - 30 F
I I I 1
\ L
60 )0 BO
DRY-BULB TEMPERATURE <°F)
90 100 110
£ 20
-1 1 T
Relotiv. Humidity ^._
40 50
-To"
70 80 90
DRY-BULB TEMPERATURE ( F)
100 110
I - 1 - 1
Relative Humidity
40 50
RANGE - 10°F
60 70 80 90
DRY-BULB TEhtPERATURE ( F)
100 no
S
I
1 I I / I
Relative Humidity s'201
RAHCE - 5°F
40 50 60 70 BO 90
60 70 80 90
DRY-BULB TEMPERATURE <°F)
100 110
FIGURE 13: EVAPORATIVE HEAT REMOVAL FRACTION
VERSUS DRV-BULB TEMPERATURE FOR
CARIOUS RANGES
-------
120
100
30
Range <"F)
Ul
oo
Relative Humidity - 1001
FIGURE UA
60 eo
DRY-BULB TEMPERATURE (°F)
100
Relative Humidity - 20%
FIGURE 14C
60 80
DRY-BULB TEMPERATURE ( F)
100
120
100
60 80
DRY-BULB TEMPERATURE (UF)
FIGURE 14: EVAPORATIVE HEAT REMOVAL FRACTION
VERSUS DRY-BULB TEMPERATURE FOTT~
7AETOUS RfeLATlVE
-------
U1
H
3
W
H
EC
2
o
Si
w
100
80
60
40
20
RANGE = 27°F
1
1
1
3000
FIGURE 15:
4000
5000
6000
LIQUID LOADING (Ibm/hr-ft )
7000
EFFECT OF LIQUID LOADING VARIATIONS ON
EVAPORATIVE HEAT REMOVAL FRACTION
100
80
60
40
2°
REFERENCE DESIGN COOLING TOWER
±
1
1600 1800 2000 2200 2400
Gas (Air) Loading - G (lbm-DA/hr-ft2)
1 1 1 1
1400 1600 1800 2000
DIFFUSER (STACK) FXHAUST VELOCITY (FT/MIN)
FIGURE 16: - INFLUENCE OF GAS LOADING VARIATIONS ON
EVAPORATIVE HEAT REMOVAL RATIO
-------
4.UU
Co
v> 80
2
M
H
O
g
-) 60
>
o
S
E^J
PH
3 4°
K
|_|
H
o 20
i
w
o
1 1 1
RH - 20% _. —
• Tw 76 P
76°F
^^ 1 \j ¥
RH = 807. • ~~~
60°F _
— —
- -
_ ^
_
_ _
1 1 1
100 200 300 400
Ka - COOLING TOWER CHARACTERISTIC (-
100
80
t~\
b
,
^
? 60
o
s
g
^
40
>
'
O
§ 20
H
n
1 | | 1 1 1 1 1 1
—
• ~~^
-
— ~
_
_ -
_
REFERENCE DESIGN COOLING TOWER
CONSTANT EXHAUST VELOCITY
i i I 1 I I I I I
FIGURE 17: INFLUENCE OF COOLING TOWER CHARACTERISTIC
ON EVAPORATIVE HEAT REMOVAL
-1000
1000 3000 5000
ALTITUDE (FEET)
7000
9000
FIGURE 18: EFFECT QF ELEVATION ON EVAPORATIVE HEAT
REMOVAL
-------
The effect of barometric pressure (altitude) variations on the evap-
orative cooling fraction is shown in Figure 18. The barometric pres-
sure may be seen to have a small effect on the fraction.
For this particular study, the air exhaust velocity was held constant
so that the effect on a fixed tower design could be examined. Because
the air density decreases as the elevation increases, a practical tower
design would provide for an increase in the air velocity as the elevat-
ion increased.
Barometric pressure as a function of elevation above sea level was
calculated using an empirical formula from the ASHRAE Brochure on
Psychrometry [8].
Natural Draft Cooling Tower Analysis
The evaporative heat removal fraction in a typical natural draft
cooling tower is shown in Figures 19 and 20. The cooling tower
analyzed was of the counter-flow type and used parallel-plate packing.
The cooling tower design details are given in Table 2 below.
Table 2
NATURAL DRAFT COOLING TOWER DESIGN PARAMETERS
Overall Height 480 ft.
Fill Diameter 340 ft.
Air Inlet Height 30 ft.
Fill Design , Counter-Flow
Packing Type Parallel-Plate
Packing Height 12 ft.
Plate Spacing 1.0 in.
Plate Thickness 0.19 in.
Heat Rejection Rate 5.5 X 109 BTU/hr
Water Flow Rate 450,000 GPM
The analysis was performed using a computer code which was a modified
version of a code written by Winiarski, Tichenor, and Byram [10].
The fraction of heat removed from the natural draft tower by evaporation
is shown (versus wet-bulb temperature) in Figure 19 for the case of
constant heat rejection. The only parameters varied in the natural
draft tower studies were the ambient conditions. The evaporative heat
removal fraction varied from 60% to 90% over the range of typical
operating conditions. The evaporative heat removal fraction was
761
-------
§
O
a
o
w
H
u
w
>->
3
w
fe
o
n
H
P-,
100
90
80
70
60
50
40
30
20
10 h
NOTE:
Constant Heat Rejection Rate
(5.8 Billion BTU/HR)
Counter-Flow Operation.
Parallel-Plate Packing
480 FT. Total Tov/er Height.
40
50
60
70
80
FIGURE 19:
AIR WET-BULB TEMPERATURE - (°F)
EVAPORATIVE HEAT REMOVAL FRACTION VERSUS WET-BULB
TEMPERATURE: TYPICAL NATURAL DRAFT COOLING TOWER
1
H
U
100
90
80
70
60
50
40
30
20
10
Relative Humidity 303
NOTE:
Constant Heat Rejection Rate
Counter-Flnw Operation
Parallel-Plate Packing
480 Ft. Total Tower Height
40 50 60 70 80
AIR DRY-BULB TEMPERATURE (c
90
100
FIGURE 20: EVAPORATIVE HEATREMOVAL VERSUS DRY-BULB TEMPERATURE
FOR A TYPICATTNATURAL DRAFT COOLING TOWER
-------
increased by about 12% of the total heat rejection as the relative
humidity was decreased from 100% to 30% at a constant wet-bulb
temperature.
Figure 20 shows the evaporative removal fraction plotted versus
dry-bulb temperature. The relative humidity is seen to have a more
significant effect on the evaporative cooling fraction in the natural
draft tower than was the case in the mechanical draft tower. At any
dry-bulb temperature, decreasing the relative humidity from 100% to
307o increases the value of the fraction by 6%. As a good approximation,
the fraction increases by 1% for each 3 F increase in the dry-bulb
temperature.
Because of the difficulties associated with converging the air flow
rate and heat rejection rate simultaneously, it was not convenient
to examine the effects of varying the natural draft cooling tower
design. Nevertheless, the results given are believed to be indicative
of natural draft cooling towers using counter-flow, parallel-plate
packing.
EVAPORATIVE WATER LOSSES
Evaporative cooling results in moisture release into the atmosphere;
it is the evaporative water loss that is responsible for the
consumption of water in wet cooling towers, i.e., evaporation and
the blowdown it necessitates. If all cooling were by evaporation
the associated water loss due to evaporative heat removal would
be 1% / 10°F range.
Figure 21 shows the water loss due to evaporation in the reference
design mechanical draft, cross-flow cooling tower plotted versus
dry-bulb temperature. A similar plot for the reference design
natural draft cooling tower is shown in Figure 22. Except for
the vertical scale, the plots are quite similar to those presented
earlier for the evaporative cooling fraction.
The water evaporation rate in wet cooling towers (both mechanical
and natural draft) was found to vary typically from 1% / 15°F range
at 35°F to 1% / 11°F range at 100°F dry bulb temperature; this
behavior reflects an increase in sensible cooling as the ambient
dry-bulb temperature decreases. The evaporative water loss (WL)
in large wet cooling towers may be estimated as follows for ranges
above 20°F and relative humidities above 50%.
WL = { 0.061 + 0.0004 ( Td - 35 ) > X Range (Eq. 6)
In this equation, the water loss WL is in per cent (7=) of the
circulating water flow, T, is ambient dry-bulb temperature in
degrees Farenh«it, and the range is in degrees F. This eauation
underpredicts the water evaporation if the range is below 20 F
or if the relative humidity is below 507=.
763
-------
2.5
tr.
O
as
w
s
I
w
1.5
1.0
0.5
0.0
40
50%
Relative Hunidity - 257.
1007.
Reference Design Mechanical Draft Tower
Design Heac Rejection Rate
50
60
70 80
DRY-BULB TEMPERATURE
90
(7.)
100
FIGURE 21: EVAPORATIVE WATER LOSS VERSUS DRY-BULB TEMPERATURE IN-A
TYPICAL MECHANICAL DRAFT COOLINO TOWER
2.5
2.0
1.5
OL
W
g 1.0
i
£ 0.5
0.0
1007;
Relative Humidity
307=
757.
50
40
50
Reference Design Natural Draft Tower
Constant Heat Rejection Rate
60 70 80
DRY-BULB TEMPERATURE
90
100
22: EVAPORATIVE WATER LOSS VERSHF DRY-BITLS TEMPERATURE
IN A TYPICAL NATURAL DRAFt COOLlNC TOWER ' '
764
-------
RECOMMENDATIONS FOR FUTURE WORK
Analytical Method
The details of the enthalpy-difference driving force method used for
this have been discussed earlier. Importantly, however, for future
work, it is recommended that a more exact analysis by Baker and Eaton
[7] be used to evaluate evaporative heat removal in wet, cross-flow
cooling towers.
Other Tower Designs
This work has reported the evaporative heat removal fraction in a
typical mechanical draft, cross-flow cooling tower and in a typical
natural draft, counter-flow tower. Although it is expected that the
results for a mechanical draft, counter-flow tower and for a natural
draft, cross-flow tower will not differ significantly from those re-
ported herein, it would be of interest to do the detailed analysis
for those tower designs not considered here, i.e., mechanical draft
counter-flow towers and natural draft cross-flow towers.
The analytical method developed by Navahandi, et al. [11] is recom-
mended for evaluating mechanical draft, counter-flow tower designs.
It is expected that the results for mechanical and natural draft count-
er-flow towers will be similar though the draft in a mechanical draft
tower is nearly independent of the operating conditions.
The variation of the air humidity and dry bulb temperature at the out-
let of the fill of a cross-flow natural draft tower might alter the
evaporative cooling fraction somewhat from that in a counter-flow
natural draft tower. It will be necessary to develop an iterative pro-
cedure to establish the air loading under natural draft conditions when
using the Baker and Eaton method for cross-flow tower analysis.
Field Measurements
In counter-flow tower designs and natural draft cross-flow towers,
simple field measurements of inlet and outlet air humidity and dry
bulb temperature will rapidly establish the fraction of heat removed
by evaporation. Such measurements over an extended period of time will
demonstrate the effect of several parameter variations on the evaporative
cooling fraction.
For cross-flow mechanical draft towers, vertical air temperature and
humidity profiles at the fill outlet would be required; however, five
to ten measurement positions should be adequate. Here again, extend-
ed measurements would establish the effect of variations in tower
operating conditions of evaporative heat removal.
765
-------
SUMMARY
Wet or evaporative cooling towers are commonly used to provide for
the cooling of water by direct contact with air. Two heat removal
mechanisms dominate in an evaporative cooling tower: evaporative
heat removal and sensible heat transfer. Sensible heat transfer
refers to heat transfered by virtue of a temperature difference
between the water and air. Evaporative heat removal refers to the
energy removal from the water as latent heat of evaporation; this
heat removal is the result of the evaporation of water into air dur-
ing the direct-contact cooling process.
The ratio of evaporative-to-total (sensible plus evaporative) heat
transfer in a wet, cross-flow, mechanical draft cooling tower was
analyzed. The ratio was found to vary from 60% to 90% during typical
operating condtions. The evaporative heat removal fraction increased
as temperature (either wet-bulb or dry-bulb) increased and as relative
humidity decreased.
Typically, the fraction of the heat removal from a wet cooling
tower due to evaporative cooling increases about one per cent
per three degrees F ( 1% / 3°F ) increase in the ambient dry-
bulb temperature. For large wet cooling towers (either mechanical
or natural draft), the evaporative heat removal fraction can be
estimated as follows for operating ranges above 20°F:
Qev/ Qtot = 0.65 + 0.0038 (Td - 35 ) (Eq. 7)
The theoretical maximum evaporative water loss is one per cent
of the circulating water flow rate per 10 F operating range,
i.e., 1% / 10°F range. For ranges above 20 °F, the evaporative
water loss WL (as % of circulating water flow) in large wet
cooling towers can be estimated by
WL = (0.061 + 0.0004 ( Td - 35) ) X Range (Eq. 8)
where the ambient dry-bulb temperature T, and the range are in
degrees F. The above equations apply at relative humidities above 50%.
Kesults similar to those obtained above were obtained for a
counter-flow natural draft cooling tower. For the mechanical
draft, cross-flow tower, it is of interest to note that under
certain conditions, both the water and air are cooled by evap-
oration so that the evaporative cooling exceeds 1007» of the
cooling effect on the water alone.
766
-------
REFERENCES
(1) Hamilton, Thomas H., "Estimating Cooling Tower Evaporation
Rates," Power Engineering, V. 81. No. 3, 1977, pp. 52-54.
(2) Eaton, Thomas E., "Evaporative Cooling Tower Performance: A
Comprehensive Bibliography," Industrial Heat Rejection Project
Report, Mechanical Engineering Department, University of Kentucky,
December 1978.
(3) Kelly, Neil W., "A Blueprint for the Preparation of Cross-Flow
Cooling Tower Characteristic Curves," Cooling Tower Institute
Technical Paper TP-146A, 1976, 30 pp.
(4) Hallett, G.F., "Performance Curves for Mechanical Draft Cooling
Towers," Journal of Engineering for' Power, Trans, of the ASME,
Vol. 97, October 1975, pp. 503-508, also ASME Paper 74-WA/PTC-3.
(5) Park, J.E., J.M. Vance, K.E. Cross, and N.H. van Wie, "A Computerize
Engineering Model for Evaporative Water Cooling Towers," Proceed-
ings of the Conference on Waste Heat Management and Utilization,
May 1976, pp. IV-C-180 to 199; also U.S. DOE (ORNL) Report
K/CSD/INF-77/1.
(6) Park, J.E., and J.M. Vance, "Computer Model of Cross-Flow Towers,"
Cooling Towers. Vol. 1, AIChE-CEP Technical Manual. 1972, pp.
122-124.
(7) Baker, Kenneth L., and Thomas E. Eaton, "An Improved Method for
Evaporative Cross-flow Cooling Tower Performance Analysis,"
Second Conference on Waste Heat Management and Utilization,
Miami, December 1978.
(8) ASHRAE Brochure on Psychrometry (N.Y., American Society of Heat-
ing, Refrigerating, and Air-Conditioning Engineers, 1977),
167 pp.
(9) Croley, Thomas E., II, V.C. Patel, and M.S. Cheng, "The Water and
Total Optimizations of Wet and Dry-Wet Cooling Towers for Electric
Power Plants," Iowa Institute of Hydraulic Research, The University
of Iowa, IIHR Report No. 163, January 1975, 290 pp.
(10) Winiarski, Lawrence D., Bruce A. Tichenor, and Kenneth V. Byram,
"A Method for Predicting the Performance of Natural Draft Cooling
Towers," U.S. Environmental Protection Agency, Water Quality Off-
ice, Water Pollution Control Research Series Report 16130 GKF
12/70, December 1970, 69 pp.
(11) Nahavandi, Amir N. , and Johann J. Oellinger, "An Improved Model
for the Analysis of Evaporative Counterflow Cooling Towers,"
Nuclear Engineering and Design. Vol. 40, 1977, pp. 327-336.
767
-------
NOMENCLATURE
a - Interfacial air-water contact area per unit
fill volume (ft2/ft3)
c - Specific heat of water at constant pressure (BTU/lbm°F)
G - Gas (Air) loading (Ibm-DA/ft2)
H - Enthalpy of an air-water vapor mixture (BTU/lbm-Dry Air)
H - Same as H
H' - Air-Water vapor mixture enthalpy at a specified water
temperature (BTU/lbm-DA)
2
K - Overall mass transfer coefficient (lbm-WV/(hr-ft -Ibm
h« - Latent heat of vaporization of water (BTU/lbm)
water/Ibm DA)
Ka - Cooling tower characteristic (BTU/hr-ft3)/BTU/lbm-DA)
L - Liquid loading (Ibm-water/ft )
11^ - Total water flow to cooling tower (Ibm/hr)
M - Number of horizontal grid points (-)
N - Number of vertical grid points (-)
Q - Totrl evaporative heat removal (BTU/hr)
Qtot - Total evaporative plus sensible heat removal (BTU/hr)
O^gj - Heat rejection rate of cooling tower (BTU/hr)
T^ - Air dry bulb temperature (°F)
T - Water temperature (°F)
TQn - Hot Water Temperature (°F)
T ££ - Average cold water temperature (°F)
W - Humidity ratio (Ibm-water vapor/lbm-dry air)
5X - Horizontal grid width (ft)
SY - Vertical grid height (ft)
- Relative humidity (7.)
Subscripts
i - inlet
o - outlet
768
-------
ACKNOWLEDGEMENTS
The author would like to express his sincere gratitude to the Institute
for Mining and Minerals Research and to the Mechanical Engineering
Department of the University of Kentucky for their partial financial
support of this project,
Mr. C.F. Hsu's computational efforts which established the preliminary
estimates of the evaporative cooling effect in mechanical draft towers
are gratefully acknowledged.
Also, the author would like to thank Ms. Paulette Montross for prepar-
ing the final draft of this paper. Professor O.J. Hahn of the Uni-
versity of Kentucky is acknowledged for his assistance in obtaining
financial support for this work.
Finally, Professor Norman C. Rasmussen of the Massachusetts Institute
of Technology is gratefully recognized for supervising the author's
first project on cooling towers which inspired this paper.
769
-------
APPENDIX A
CROSS-FLOW COOLING TOWER ANALYSIS METHOD
Basic Enthalpy-Difference Equations
The performance analysis of cross -flow mechanical draft cooling towers
was accomplished using the enthalpy- difference driving force method.
This method has been used by others, e.g., Kelly [3], Hallett [4], or
Croley [9 ], to analyze cross-flow cooling towers. Further, Hallett
has developed an empirical correlation of the unit volume coefficient
(Ka) for a typical cross-flow cooling tower packing. Hallett fs corre-
lation for Ka is intended to be used in an enthalpy-difference driving
force calculational model for cooling tower performance.
The enthalpy- difference driving force model provides a reasonably
accurate prediction of water temperatures without calculating the de-
tails of the air conditions in the tower. This method is successful
principally because the enthalpy of a mixture of air and water vapor
is nearly constant for constant wet bulb temperature. The enthalpy-
difference driving force model then estimates the heat transfer between
the water and air assuming that the air is saturated at the wet bulb
temperature and that the air-water vapor mixture near the liquid water
surface is saturated at the water temperature; with this, the enthalpy
of the air and water vapor near the water surface is readily deter-
mined .
The basic equation for the enthalpy change of the air in a differential
volume of the tower fill is
Ho ' Hi
where H] is the enthapy of saturated air at the entering water temper-
ature and H.^ is the enthalpy of the entering air. While this equation
is convenient for an initial estimate of the outlet air enthapy, the
following equation, which averages .the entering and leaving enthalpy
difference, is more accurate:
H = H. + [H! - H. + H1]
° -i - ^EZ - 1 - i - 2l_ (Eq. A2)
[1 + Ka <5X ]
~~ZG
The outlet water temperature T may then be estimated using the follow-
ing equation:
Two - Twi - CTTTCT (Ho * V
-------
proved estimate of H This iterative procedure converges rapidly;
typically the values of HQ and TWQ will converge in less than five
iterations .
Estimation of the Relative Humidity Inside the Tower
The use of the enthalpy difference driving force method allows one
to evaluate the local water temperature and air enthalpy inside
the cooling tower without regard to the local air relative humidity.
Because the outlet air relative humidity was important in evaluating
the evaporative heat removal in the cooling tower, the following
procedure was used to estimate the local humidity ratio inside the
cooling tower fill.
Since the objective was simply to determine the change in the humidity
ratio as the air crossed a differential volume element of fill, and
since the change in enthalpy across the volume element was known, the
humidity ratio change was estimated as the change in the humidity ratio
along the saturation line between the inlet and outlet air enthalpies.
This change in the humidity ratio was added to the entering humidity
ratio to determine the outlet humidity ratio , i.e.,
WQ = W. + [Wsat(H0) - Wsat(H.)] (Eq. A4)
At this point the enchalpy and humidity ratio leaving the grid
were known so that the wet-bulb and dry-bulb temperatures could
also be determined.
This method for determining the humidity ratio distribution in the
tower is not exact; however, for the purposes estimating the outlet
air humidity, this method was believed to be better than assuming
saturated outlet conditions. Further, this exhaust humidity calculation
technique is in better agreement with observed cooling tower exhaust
conditions than is the assumption of saturated conditions - particular-
ly on mechanical draft, cross-flow towers.
With the local humidity ''variation known, the small correction to
the liquid loading do to evaporation was made and used in the next row
of calculations:
Lo " Li - G jg
-------
Prepared for Presentation at a
Waste Heat Management &
Utilization Conference
Miami Beach, Florida
December 4-6, 1978
Comparative Cost Study of Various Wet/Dry Cooling
Concepts that Use Ammonia as the Intermediate
Heat Exchange Fluid
B. M. Johnson, R. D. Tokarz, D. J. Braun, R. T. Allemann
1. PURPOSE OF THIS WORK
Dry cooling of thermal power plants, by which the heat from the
power cycle is rejected directly to the air, has been used in a few
isolated instances throughout the world for the past 15 years Very few
installations are in operation in the U.S. although i~ is ceinn
given increased consideration for new large ocwer stations. Dry cooling
is a more costly option than once-through or evaporative cooling, but
there are a few locations now, and there will be far more in the future,
at which once-through and all-wet evaporative cooling towers cannot be
used because of the increased competition for existing water supplies
among growing populations, agriculture and industry. Earlier studies
at the Pacific Northwest Laboratory, have shown that considerable
incentives exist for development of an advanced concept which makes use
772
-------
of arrcnonia as an intermediate heat transfer fluid in a process which
provides augmented cooling by evaporation.
This paper summarizes the conceptual desian and costs of four
different configurations for such a system and compares them to a state-
of-the-art integrated dry/wet circulatino water system.1 All are mechanical
draft systems.
These studies are part of an ongoing effort, supported by both
the U. S. Department of Energy and the Electric Power Research Institute,
to increase the flexibility of plant siting and reduce the break-even cost
of water at which power companies would choose to conserve water through
use of some dry cooling.
1.2 Incentives for Or//Wet Cooling '
Providing seme capability for augmented cooling via water evapora-
tion to dry cooled heat rejection systems has been shown to be highly
cost-effective. It is probable that in this country most dry-cooled
systems for large power plants will have some evaporative cooling capability
included in the system to avoid either of the costly alternatives of (1) buildi
excessively large systems to provide adequate heat rejection for oeak power
projection during the hottest summer days, or (2) buying power from other
sources during peak demand periods on the hottest days. With some evaporative
cooling capability the dry/wet system can be built so as to use whatever water
is available for cooling and thus minimize the required size of the high-
priced dry cooling system. How to best provide this evaporative cooling
capability with the ammonia system was one of the purposes of this work.
1. R. D. Tokarz, et. al., "Comparative Cost Study of Four Wet/Dry Cooling
Concepts that use Amnonia as the Intermediate Heat Exchange Fluid,"
PNL-2661, Battelle Pacific Northwest Laboratories, May 1978.
773
-------
at a specific ambient temperature, which was established so as to require
a predetermined amount of water each year for augmented cooling.
Capital costs included all engineering, construction and material
costs associated with the cooling towers, condenser, water treatment
equipment and related piping and pumps. Construction costs included the
contractor's profit and overhead, but excluded any escalation or contingencies.
Operating costs included the cost of auxiliary power for pumps and fans,
maintenance, and water treatment.
Credit was taken for improvements in plant heat rate associated with
lower back pressures made possible by the advanced designs. However, no
credit was taken for increases in load that would be made possible
by back pressures lower than the design values of the reference plant.
To assure the validity of the comparisons, every effort was made to
use uniformity in the conceptual designs and-cost estimation of each concept.
All estimates were prepared by an architect-engineer subcontractor ^
from prsconceptual design descriptions prepared for each concept. All
design descriptions used a common reference plant location, the San Juan
Unit 3 of the Public Service Company of flew Mexico. This plant was selected
as the reference plant for this study because a plant with integrated dry/
wet cooling towers is currently under design and construction at this location.
As a result, adequate site data were already available on which to base the
preconceptual designs of the advanced alternatives , including
meteorology,
fuel costs,
(a) SSQ Engineering Corporation
774
-------
1.3 Incentives for Using Ammonia
The use of ammonia as a heat transfer medium between the-steam
condenser of the turbine-generator and the air-cooled heat rejection system
has been shown to be cost effective in earlier studies which were reported
in the previous conference. The use of ammonia offers at least four
advantages leading to reduction in system costs. These are:
1. Reduced pumping power in the transport loop;
2. Elimination of the temperature range of the transport
loop as a temperature increment between the ambient
dry bulb and the condensing steam temperature;
3. The ability to use high performance surfaces on the
ammonia side of the steam condenser/ammonia reboiler
to reduce the condenser terminal temperature difference,
and lastly,
4. No need to prevent freeze up.
2. BASIS OF COMPARISON
The comparisons of the various concepts were performed on the
basis of "comparable capital cost" defined as the sum of the estimated
capital cost of the installation plus the capitalized operatino cost.
This latter term is just the operating cost divided by the annual-fixed-
charge-rate of 18 percent. The designs have not been optimized in the
sense that they would yield the lowest bus bar cost of electricity.
At the time the study was initiated the dry/wet design optimization code
was not completed. Instead, each design satisfies a set of
design parameters, particularly with respect to heat rejection capability
^"Dry/Wet Cooling Towers with Ammonia as an Intermediate Heat Exchange Medium,"
Paper 4C-4, Waste Heat Management & Utilization Conference, Miami Beach Florida,
May 9-11, 1977.
775
-------
at a specific ambient temperature, which was established so as to require
a predetermined amount of water each year for augmented cooling.
Capital costs included all engineering, construction and material
costs associated with the cooling towers, condenser, water treatment
equipment and related piping and pumps. Construction costs included the
contractor's profit and overhead, but excluded any escalation or contingencies.
Operating costs included the cost of auxiliary power for pumps and fans,
maintenance, and water treatment.
Credit was taken for improvements in plant heat rate associated with
lower back pressures made possible by the advanced designs. However, no
credit was taken for increases in load that would be made possible
by back pressures lower than the design values of the reference plant.
To assure the validity of the comparisons, every effort was made to
use uniformity in the conceptual designs and-cost estimation of each concept.
All estimates were prepared by an architect-engineer subcontractor ^a)
from preconceptual design descriptions prepared for each concept. All
design descriptions used a common reference plant location, the San Juan
Unit 3 of the Public Service Company of New Mexico. This plant was selected
as the reference plant for this study because a plant with integrated dry/
wet cooling towers is currently under design and construction at this location.
As a result, adequate site data were already available on which to base the
preconceptual designs of the advanced alternatives , including
• meteorology,
• fuel costs,
(a) SSQ Engineering Corporation
776
-------
• water availability and quality,
• onsite construction costs,
• transportation costs,
• power costs, and
• site characteristics.
Each of the dry/wet systems was conceptually designed and estimated
using the same procedure, including the integrated wet/dry concept used in
the reference plant. No cost or detail design information was obtained
from the utility about the integrated wet/dry concept, so it, too, was
designed and costed on the same basis as the other four. This
cost comparison study applies, only to the reference plant design conditions.
Other sites would have different conditions that could markedly affect the
resulting comparison.
2.1 Conceptual Design Bases
The conceptual designs of the three cooling tower concepts were
based on performance requirements established by the Public Service Company
of New Mexico for the San Juan Unit 3. These requirements are listed below.
1. The heat rejection capability of the cooling system shall
be about 2.5 x 10^ Btu/hr over a yearly cycle.
2. The cooling system shall accommodate the meteorological profile
of Farmington, New Mexico (Table 1).
3. The turbine shall be operated at or below a back pressure of
4.5 in. Hg at an ambient temperature of 95 F or below. Above
95°F, the turbine back pressure shall be allowed to increase
to a maximum of 5.0 in. Hg.
777
-------
Table 1. Meteorological Profile at Fanni'noton, NM
Dry Bulb Air
Temperature, °F
•?
t
12
17
22
27
32
37
^2
47
52
55
57
62
65
67
70
72
75
77
80
82
87
92
97
102
Wet Bulb Air
Temperature, °F
7
n
16
20
25
29
33
36
39
41
44
45
47
49
50
52
53
54
54
55
56
58
61
62
63
Hours per
Year
55
98
198
336
553
698
688
708
678
648
388
259
704
411
274
351
234
295
197
245
164
331
179
34
1
778
-------
4. The maximum amount of water available annually for consumptive
use is 1900 acre-ft or 5.12 x 109 Ib, which is about 2Q°'° that
consumed by all-wet tower of similar rating.
5. The maximum instantaneous flow rate of consumptive water due
to evaporation shall be 2.0 x 106 Ib/hr (4000 gpm).
The San Juan River was assumed to be the source of water to the
plant. Water treatment requirements for closed-loop recirculating systems
associated with dry towers were assumed to include demineralization, vacuum
deaeration, corrosion inhibition,and pH control (pH 8.5). Open loop systems
used in wet and wet/dry towers were assumed to require lime-soda softening
(side stream), scale inhibition and biofouling control. Delugeate treat-
ment to maintain a Langlier saturation index of zero or slightly negative
was assumed.
3. ALTERNATIVES CONSIDERED
The four cooling concepts principally studied utilize the ammonia
liquid-vapor phase change to transfer heat from the steam turbine outlet
to the cooling towers. These concepts are compared with the conceptual
design of the integrated dry/wet cooling tower of a configuration similar
to that being constructed at Farmington, New Mexico. This design and
cost estimate were developed without obtaining design details or costs
from either the owner or manufacturer of that system. Consequently all
systems were estimated by the same method and from similar data base.
However, the ammonia systems' designs had not undergone the extent of
engineering optimization studies inherent in a commercial system.
779
-------
3.1 Ammonia Heat Transport System
The following is a brief description of the salient features of the
ammonia heat transport system.
The ammonia heat transport system for power plant heat rejection
is functionally similar in many respects to the "direct" system in which
the exhaust steam from the last stage of the turbine is ducted directly
to an air-cooled condenser. The principal difference is the existence of
a steam condenser/ammonia reboiler in which ammonia is "substituted" for
steam as the medium for transporting heat from the turbine to the tower
(heat sink). In all respects the ammonia system, with vapor moving from
the reboiler to the air-cooled condenser and liquid returning to the
reboiler, will function and respond to load changes in the same manner
as tne direct system. Figure 1 is the process flow sketch.
Exhaust steam from the last stage of the turbine is condensed
in the condenser/reboiler located directly below the turbine. Instead of
water circulating through the tubes, liquid ammonia is boiled as it is
pumped through the tubes under pressure, set by the operating temperature
in the condenser. The flow rate of ammonia is set to yield a vapor quality
emerging from the tubes varying from 50 to 90". This two-phase mixture is
passed through a vapor-liquid separator from which the vapor is sent to the
air-cooled condenser, while the liquid is combined with the ammonia condensate
from the dry tower and recycled back through the condenser/reboiler.
780
-------
Table 2. Design Parameters
To-.er
Tower 51 Z2 (ft)
Tower Design Tcmo.
Design [TO, degrees
Numoer of Towers
Nurnoer of Bundles
Dimensions
Total Surface Arei,
ft2
Frontal Area, ft^
Tube 00, incties
Fin Oasign
Fin Dimensions
Fins Per Incn
Tuoe Material
Fin 'Material
Tuoe Geometry
Verti cal
.'OTERV
259 dia < 56
niyn
55°F
57
3
238
47.5 ft x 8 ft
x 6 in.
9.71 x 10°
1.072 x 105
0.78
Rectangular
Plate
5 in. deep
7.37 ft (11 gn
9
Aluminum
Aluminum
Staggared Sows
HOT£=V
2C5 x 230 x 57
55'F
67
2
238
47.5 ft x 3 ft
x a in.
9.71 x 10°
1.072 x 105
0.78
Rectangular
Plate
6 in. deep
7.37 ft hign
9
Aluminum
Al umi n um
Staggared Rows
SCAT Tower
225 Jia x 56
55°F
57 dry, 32 «»t
2
122
50 ft x 12 ft
x 1 ft
8.91 x 10s
0.732 x 105
0.3
Integral
0.707 in. hign
12 in. deeo
10.5
Al umi n um
Al 'jminum
Rectangular
Augmenting '1113
Conacnser
170 dlj x. 56
hign
35'F
37 ary. 32 net
2
38
50 ft < 12 ft
x 7.2 in.
5.41 x 10°
0.522 x 105
0.3
Integral
0.707 in. high
12 in. deep
10.5
Aluminum
Al uninum
Sectancular
Intenrnc-: '..'et/Drv
4G2 < 133 < 35 -ii yn
3S°F
90 day, 30 wet
2
320
48 ft x 72 ft x
10 in.
7.205 x 10°
9.216 x 104
1.37
Single Leg
Wrapoed
2.25 round
10
Admi ral ty
A 1 urm n um
Houi lateral
Transfer TuBe Pitch,
Incnes
Heat Transfer
Coefficient Stu/hr-
7.57
2.35
7.57
NA
,'IA
9.1
2.35
6.41
Frontal Velocity/
Internal Velocity.
ft/sec.
Air Flow, Ib/hr - Cry
4ir Flow, !b/hr - Wet
3/13.3
1.33 x 103
1.3 x iG<3
Air MISS "ow Rate, - Orv i ;q t -Q3
ib/hr-ft* - w.jt 3.9 -, !0J
Cooling water Flow, G?H 7.000 (80 TOH)
Airside Heat -.xcnanae/AP/ 0.356/0.464
Total iP, !ncnes rij^
Fans - Orv
Wet
Fan Diameter, ft - dry
Wet
HP ?er Fan -
Dry
Wet
of Slaaes - Dry
wet
Pi ten, degrees - Qry
We t
57
23
32.3
6
12
3/13.3
1.93 x 1C5
1.3 x 103
1.79 x Id3
3.9 i 103
12/16.4
1.97 x 108
O.iS x 103
2.59 ^ i03
14.0/19.1
1.70 x Id8
j.47 x IG3
3.:s x io3
2.195 x :C3
3.24 x 1G5
2.:a « 'o3
J.5 >. 10-1
21,300 (40 TTJ11) 170,000 (35 TOH) 200,OCO CS TOH) 219,COO (77 "OH)
0.356/0.464 0.2E1/0.384 0.243/0.353 0.345/0.533
So
23
82.3
5
12
48
23
50
150
10
22
25
3
28
105
''50
16
22
40
10
CO
24
115
90
3
6
14
16
781
-------
The vapor from the vapor-liquid separator flows to the dry tower
under the driving force of the pressure difference between these two
components created by the temperature difference and the associated vapor
pressure of the ammonia.
The steam condenser is composed of horizontal tube bundles, with
steam condensation on the shell side, and anhydrous ammonia evaporation
on the tube side. Design tube side maximum pressure is 350 psig, 1350F.
Tubes are aluminum with the following dimensions:
tube length 50 ft
tube 00 1 in.
tube gauge 12 BWG
tube pitch 1.5 in.
total number of tubes 15,100
The tube is enhanced on the outside for condensation and on the inside
for boiling with proprietary Union Carbide Ccmpany/tinde enhanced condensation
surface. The tubesheets are aluminum and the condenser is eauipped with
impingement protection where necessary.
The air removal section of the condenser is stainless steel.
The performance and cost of this component are significant uncertainties
in this study. The cost algorithms developed for computer optimization studies
on the basis of laboratory data indicated its cost would be very nearly similar
to that of a conventional turbine condenser. The estimate developed by the
architectural engineer and used in this study, reflected the lack of firm
data from similar equipment. The architectural engineer estimated the equip-
ment to be 50% more costly than a conventional condenser.
782
-------
The piping for the system consists of vapor transport piping,
vapor distribution piping, condensate collection piping, and condensate
return piping. Associated with this system are pumps for condensate
return and reboiler circulation, a combination vapor separator/reboiler
supply tank, and ammonia storage tanks. The vapor separator/reboiler
supply tank is located as close to the steam condenser/ammonia reboiler
exit as possible. The upper portion of the tank acts as a cyclone
separator to remove liquid ammonia carried over in the vapor leaving
the reboiler. The lower portion of the tank acts as a reservoir for
supplying the reboiler injection pumps and also provides system surge
capacity. The lower portion of this tank has sufficient volume to contain
the inventory of two tower quadrants if it becomes necessary to evacuate
them for maintenance or in case of leaks. The material for all piping and
tanks is carbon steel.
Each of the two condensate return pumps would have a capacity of
10,000 gpm at 27 ft NH, TDH. Each of the two condenser recirculation
pumps would have a capacity of 20,000 gpm at 30 ft NH- TDH. The drain
and fill pump would have a capacity of 2500 gpm at 50 ft NH- TDH.
Excess storage capacity would be provided by ten 7750-gal pressure
tanks. These tanks will store the entire quantity of ammonia if it becomes
necessary to evacuate the system for maintenance or in case of emergency.
Provision is made for a nitrogen purge system to flush the air from
the system before filling it with ammonia to prevent the possibility of
stress corrosion cracking of the steel components. The total volume of the
system is approximately 50,000 ft3. Vents are located at the highest
point in each quadrant from which ammonia vapor can be evacuated after
the quadrant has been drained and isolated. The vents are piped to a
flare station on top of the tower.
-------
3.2 HOTERV Plate Fin Heat Exchanger with Deluge Augmentation,
The initial cooling tower arrangement using the HOTERV plate fin
exchangers was round towers with fans across the top and the heat
exchangers around the periphery. Previous studies had shown this to
be a cost effective configuration for long fin-tube exchangers. However,
as the result of the ensuing cost estimate for the towers, it was con-
cluded that it was not a good arrangement for the HOTERV bundles arranged
horizontally to accommodate deluging. A second configuration was scoped
out and estimated in which the heat exchangers were arranged as A frames
on a plane below the fins. Figures 2 and 3 show these two arrangements.
With the vertical peripheral arrangement, three towers 260 ft in
diameter and 56 ft high (to the fan deck) are needed. The cooling tower
is designed to operate as a completely dry system when the ambient temper-
ature is below 55°F. Above this temperature, a portion of the heat
exchanger surface is deluged with water on the outside of the plate fins
to increase heat rejection capability. In this way sensible heating of
the air is augmented by heat transfer to the air through evaporation of the
deluge water. The tower design temperature is based upon the maximum
use of available water for augmentation (1900-acre ft) with minimum amount
of heat exchange surface area.
The airflow through each tower is induced by 19 fans (28-ft diameter)
mounted at the top of the tower structure. The heat exchanger bundles
(240 tubes/bundle) are arranged around the periphery of the towers. No
louvers for air control to prevent freezing of the ammonia are required.
However, passive louvers are located beneath each fan to prevent back
flow Q-f air when a particular fan is off. For airflow control, one or
784
-------
more fans can be started or stopped. Protection of the heat exchanger
surfaces is provided by hail screens mounted directly to the face of each
bundle. Table 2 gives specific information on all of the cooling tower
systems at design point conditions.
The HOTERV heat exchangers are 47.6 ft (15 m) long, 7.8 ft (2.3 m) high,
and 5.9 in. (15 cm) deep in the direction of airflow. There are 16 bundles/
tower quadrant , 96 bundles/tower, and 288 bundles total. The bundles are
sloped at a 5 degree angle to promote drainage of the condensed ammonia.
They are also canted forward to promote uniform deluging of the plate fins
during wet operation.
All of the vapor transport piping lies above grade. The main vapor
line transports ammonia vapor from the vapor separator to the general area
of the cooling towers through a 48-in. diameter pipe aporoximately 1000-ft
long. The piping then splits into successively small pipes leading to
each tower and subsequently to tower quadrants, bundle groups and eventually
individual bundles. The condensed ammonia liquid drains to a collection
header running around the inside periphery of the tower. The main return
line is 18 in. in diameter.
The deluge system is capable of augmenting the entire heat exchanger
surface, although the maximum design wet area is probably less than 67%.
Augmentation of the plate-fin surfaces is accomplished by allowing an
approximate water flow rate of 2 gpm per lineal ft of heat exchanger to
run down the plate fins. A small perforated pipe header adds water above
each bundle to make up for the deluge water evaoorated in the previous
bundle. The deluge piping for each tower consists of
785
-------
• two deluge pumps (vertical sump),
• deluge storage sump,
• distribution piping, and
• deluge distribution headers and splash plates.
The deluge pumps will have a capacity of 1200 gpm at 80 ft of HLO (two
pumps per tower). The suction side of the vertical sump pump will be
immersed in a circular concrete channel that catches all the water falling
from the tube bundles and serves as a storage sump when the tower is
operating dry. Polyethylene or PVC is used throughout. Maximum instan-
taneous consumptive use rate is approximately 4000 gpm. The maximum
recirculation rate to the top of the towers is 7200 gpm, although the
maximum anticipated is about 4500 gpm, with additional makeuo being added
at each of the five layers of heat exchanger in the vertical arrangement.
Water treatment will consist of sulfuric acid addition to control pH to
7.6-7.8 and blowdown (800 gpm) to maintain a sufficiently low dissolved
solids concentration. The blowdown will undergo lime softener treatment,
and the effluent from the treatment plant will be recycled into the deluge
system. Sludge from the softener (85 gpm) will be discarded to the effluent .pond,
The horizontal arrangement of the HOTERV heat exchangers differ
essentially only in the tower configuration. Each of the two required
towers is 205 ft by 230 ft and 57 ft to the fan deck. The horizontally
arranged bundles are 35 ft above the ground to provide adequate area
for air flow. Twenty eight fans (28 ft diameter) are used.
The A frames of the heat exchanger bundles are tilted at 5° as in
the vertical design to promote drainage of the ammonia.
786
-------
The total recirculation flow of the deluge system is higher, about
20,000 gpm because the bundles are not vertically stacked to provide a
means of water flow down the stack.
The savings in this arrangement accrue from the need for only two
towers. Table 2 lists the significant design parameters which are very
similar to those of the vertical arrangement.
3.3 Separate Channel Augmented Tower
The heat exchanger in this concept is an adaptation of the Curtiss-
Wright surface comprising integral fins chipped from an extruded multi-
port aluminum tube. Additional cooling is provided by the separate channel
augmented tower (SCAT) system, which uses selected channels within each
multichannel tube as water channels. (Figure 4) VJhen water is pumped
through these channels, increased cooling of the ammonia occurs by heat
transfer to the water. The heated water is piped to a wet cooling tower,
located either inside the dry tower (this design) or outside. The basic
design parameters for the SCAT system are the same as the previous two concepts
The tower can reject the design heat load without the use of any water at a
turbine exit temperature of 130°F, an ambient temperature of 55°F, and an
&OF temperature drop across the condenser/reboiler and the ammonia transport
lines.
Each of the quadrants of the two towers can be operated all dry or with
additional SCAT cooling using the wet tower. The airflow through each tower
is induced mechanically with 34 fans (28 ft diameter) mounted at the top
of the tower structure. The 50 ft by 12 ft x 1 ft (in the direction of
air flow) heat exchanger bundles (80 tubes/bundle) are arranged vertically
around the periphery of the towers.
787
-------
Ammonia vapor enters at the top and saturated liquid ammonia
emerges at the bottom. Figure 5 shows the cross-section of the tubes.
For the purpose of sizing the tower, the fins over the back portion of
the tube where the water channels are located were not included in the
calculation of heat transfer to the air during wet operation but were
included in the calculation for pressure drop. For enhanced cooling,
water is run through five alternating channels in the rear (relative to
airflow) of the SCAT tube and then through the wet tower for cooling.
The temperature ran^c oi' this water, the o'verall heat trancfar coefficient,
and the effectiveness of this section of the bundle for heat transfer are
calculated independently of any interaction with the airflow over the
tube. This is justified by the fact that the air and the cooling water
would be at approximately the same average temperature in this part of the
bundle and the presence of the air would neither add nor subtract from the
cooling action of the circulating water. Table 2 lists the significant
design parameters of the tower.
The wet tower which provides cooling of the circulating SCAT water
is located concentric and within the dry tower structure. A portion of
the air drawn through the heat exchanger is taken on through the wet
packing and exhausted by the wet tower fans. The rest of the air is
exhausted by the dry-only fans arranged in the annular region between
the respective peripheries of the wet and dry towers. When the tower is
operated at less than fully'enhanced cooling capacity, sections of the
wet packing are not wetted; none are wetted during all-dry operation
(below 55°F). The tower is designed for a 67.5°F wet bulb and 113.60F
788
-------
dry bulb for air inside the dry tower. A water range of 17.8°F and an
approach of 22.5°F is used with inlet water at 107.8 and outlet at 90°F
Up to 170,000 gpm of circulating water through the SCAT channels is
provided by 16 pumps (2 per quadrant in each tower). Very close coupling
exists between the heat exchangers and the wet tower. Eight inch poly-
ethylene lines carry water up through heat exchanger and then to the tower.
Water treatment is the same as for the integrated wet/dry system although
a smaller quantity is needed.
3.4 Augmenting NH., Condenser
The concept of using a water-cooled ammonia condenser for augmented
cooling, located at the dry tower and close coupled with a wet tower, was
selected for the following reasons:
1. Less design uncertainty than with a turbine condenser cooled
by both water and ammonia;
2. Close-coupling the ammonia condenser and wet tower was believed
to more than offset the increased equipment size and cost
resulting from the loss in temperature difference.
The condensers (four for each tower) function in parallel with the
dry tower to maintain the pressure in the ammonia. Since the operation
of the dry tower is unaffected by the operation of the condenser (unlike
the deluge approach), evaporative cooling is rot substituted for d"*y
cooling and the dry tower can be somewhat smaller for the same water
allotment. Like the integrated tower, described in Section 3.5, it is
designed for an ambient temperature of 35°F (ITO=87°F) rather than 55°F
(ITD=67°F) for the other three systems.
789
-------
Placement, spacing, and general configuration is similar to the SCAT
towers. However, the higher design ITD and simpler tube configuration
result in a smaller tower. The heat exchanger bundles (SO tubes/bundle)
are arranged around the periphery of the towers as shown, and the water-
cooled condenser (four in each tower) are hung within 'the annular space
between the periphery of the dry and wet towers. The enhancement cooling
water is pumped from the center basin to the top header of each of the
condensers, passes down and back up through the cooling tubes and out the
top header to the wet tower inlet distribution box.
The heat exchangers are bundles of multiport finned channels 50 ft x
12 ft x 7.2 in. (in the direction of airflow) of the integral chipped-fin
type manufactured by Curtiss-Wright. Each bundle consists of 80 tubes
50 ft long.
Table 2 summarizes the design parameters for the dry tower.
The eight water-cooled condensers are tube-in-shell pressure vessels
designed for 350 psi at 150°F with ammonia on the shell side. Each is 8 ft
in diameter and contains 875 U-bend aluminum tubes 1 in. -in diameter 50 ft
long. Maximum flow through each is 25,000 gpm. The wet tower which provides
cooling for the circulating water is integral with the dry tower and is
located concentric within this structure essentially the same as with the-
SCAT concept. The water system is closely similar to SCAT except that
slightly more water is used.
3.5 Integrated Dry/Wet Cooling Tower Concept
This heat rejection concept is currently planned for use in the San
Juan Unit 3. It was included in this study to provide a basis for comparing
790
-------
these alternative concepts to previous design concepts and to each other.
To assure the validity of the comparisons, this system was conceptually
designed and estimated using the same bases as the other concepts, i.e.,
without reference to actual cost figures and design details.
The condenser cooling water is transmitted to the cooling towers via
a 96-in diameter concrete piping system and circulated by three 73,000 gpm
(77 TDH) vertical well pumps. Two rectangular cooling towers house both the
air-cooled heat exchange surface, which is composed of spiral-wrapped
finned tubes tilted 25 degrees from horizontal and the wet tower packing.
The hot water from the condenser passes first through the dry section and then
flows directly into the wet towers. The cooling tower is designed to operate as
a completely dry system at temperatures below 35°F by turning off the fans
above the wet portion of the tower. A sketch is shown as Figure 6.
There are 10 heat exchanger units per tower, two units in each bay.
The spiral-wrapped fin tubes are 1 in. in diameter, of Admiralty metal,
with the thin (0.018-in.) aluminum fin wound as a single leg wrap around
the tube. They are arranged in a staggered equilateral close-packed
spacing three rows deep.
A total of 25 induced draft fans were specified for each tower, 20
in the dry section and 5 in the wet section. Louvers have not been
specified although they may be required for airflow control during high
winds and for equalizing flow into each bay to avoid local freezing.
The conventional steam condenser contains 28,500 admiralty metal
tubes 1 in. in diameter and 35 ft long.
791
-------
4. BASES FOR COST ESTIMATES
All cost data developed by the architect/engineering firm reflect
construction as of mid 1976 and thus include no contingency or escalation.
The preconceptual designs which have been developed and evaluated in
this report are not the optimal design for each system, i.e., they are not
designs which have been coordinated with the design and sizing of the
steam supply to give the lowest cost of busbar electricity. Instead,
they are designs which fit a stipulated set of conditions with respect to
ambient temperature and heat rejection capability.
Optimization studies of dry (and dry/wet) systems generally compare
the "operating costs" of alternative systems in terms of several "penalty
costs" which represent the incremental increases in plant operating costs
resulting from the use of the dry cooling system in relation to a reference
system with a once-through flow of cooling water. Included in the list
of penalty costs may be those for (a) an energy penalty because during
hot weather the plant cannot export as much energy as the reference plant,
(b) a capacity penalty because reserve generating capacity must be
available to make up for the deficiencies of the dry cooled plant, (c)
a make-up water penalty which reflects the cost of any water treatment
unique to the subject plant.
An "optimized" design represents a trade-off between a larger-sized
cooling system which has small energy and capacity penalties and a smaller-
sized cooling system which has larger penalties.
With the present comparison of five systems there were no such trade-
offs involved in the designs. All were sized to meet stipulated parameters
792
-------
Table 3. Operating Cost Summary
(dollars)
Vertical Horizontal SCAT
Hb'terv Hb'terv Tower
Hours of (Dr*
Operation |Wet
6 640
6 640
4 426
6 640
4 426
Augmenting
Ammonia Integrated
Condenser Wet/Dry
6 640
5 146
5 640
5 146
Circulation Pump 38 500
Primary Fluid
38 500
38 500
38 500 694 000
Circulation Pump 10 100
Augmenting Cool ing
Water
19 600 110 900
130 900
Water Treatment 79 000
79 000
89 000 105 000 105 000
Fan Power
900 571 600 435 400 433 300
798 000
Capacity Penalty 83 300
(Annualized)
84 700
95 600
98 000 184 000
Fuel Saving Credit-235 000
Due to Reduced
Back Pressure
-235 000 -235 000 -55 000 - 55 600
TOTAL
564 800 558 400 534 400 750 100 1 725 400
793
-------
of inlet temperature difference,'heat rejection capability and annual water
rate. Thus, the gross plant output is approximately the same from a plant
equipped with each cooling system. However, the total penalties would
differ with each design, depending on the characteristics of each with
respect to: 1) the power required for fans and water/ammonia recirculating,
2) the capacity penalty for this power, and 3) water treatment and pumping
power required for the enhanced (evaporative) cooling system.
In addition, there is a negative energy penalty which arises from
increased output in cold weather which differs in each case. All plants
have been designed to reject the stipulated heat load at a particular
design ambient temperature. At temperatures below this, the plant is
capable of operating at rated output with lower fuel consumption because
of higher turbine efficiency (lower back pressure). Credit is taken for
fuel savings from the higher plant efficiency. The three alternatives
using deluge cooling are designed to a higher ambient temperature (55QF
vis-a-vis 35°F for the other two) because a large dry tower is required
to compensate for the dry capability taken out of service as it is
converted to evaporative cooling by deluging. This has the comoensating
effect that the plant can operate at a lower back pressure in the winter
and thus use slightly less fuel.
In summary, in this study the differences in "penalties" among the
various alternatives are accounted for by evaluating five "operating"
cost terms and a sixth capital cost term. Those six cost terms are:
794
-------
• power for the main circulating system,
• power for the fans,
• water treatment operating costs,
• power for pumping deluge water,
• fuel savings resulting from the capability to operate at lower
than the reference turbine back pressure at temperatures below
the design ambient, and
• capital cost of peaking reserve capability to provide auxiliary
power to the cooling system.
To combine these "operating" and other penalty costs with the capital
costs of the plant, the former are "capitalized" by dividing by an
annual fixed-charge rate of 18% and adding them to the capital cost.
Water treatment would include scale inhibition by pH adjustment
and biofouling control with chlorine. Slowdown would be treated by lime
softening to remove dissolved solids with return of the effluent and
drying of the sludge. In addition provision would be made to suoply
demineralized water from zeolite softeners to flush the deluged surfaces,
The main differences in cost are due to the cost of biofouling control.
All operating costs are summarized in Table 3.
5. RESULTS
To facilitate comparison of the total costs of the five dry/wet
systems a "comparative capital cost" was used which is defined as the
sum of the estimated basic capital cost (i.e., the estimated cost
795
-------
Table 4. Capital Cost Estimates
(Thousands of Do!lars
Vertical
HOTERV
STEAM CONDENSER 2,653
200LING TOWER
Dry Tower
Structures 1,905
Piping-NH3 308
Heat Exchangers 5,374
Pumps/Piping- ,,„,
H70 JUb
Augmented
Ammonia Cond.
Wet Tower
Fans 1,205
Vents & Flair 15
Subtotal 9,612
Horizon
HOTERV
2,653
998
102
5,387
423
— —
--
1,186
10
8,105
SCAT
Tower
2^653
858
213
4,982
678
—
263
1,070
10
3,073
Aug.
Ammonia
Cond.
2,653.
600
152
2,734
703
1 ,562
309
880
10
5,950
Integral
Jje.t/Dry
L2£Q
2,182
—
1,215
1 ,078
2,646
—
10,307
IAIN COOLING SYSTEM
Pumps/Piping
Vapor
Pumps/Piping Liq. 422
Pumps-Reboiler
Yapor/Liq.Sep.
Subtotal
;TORAGE/FILL/DRAIN
:OVER GAS
iATER TREATMENT
ELECT/INST
"WILDINGS
OMPLETE SUBTOTAL
ONTRACTORS OH &
.NGRG & COST MGMT
OTAL CAPITAL COST
(Without Contingency
309
. 422
159
310
1,200
730
211
562
2,263
52
17,334
ROFIT 3,329
4,132
24,795
344
264
159
310
1,077
780
143
628
1,554
52
14,994
2,998
3,598
21,590
342
273
159
310
1,084
780
143
545
1,633
52
14,963
2,948
3,582
21,493
342
273
159
310
1,084
780
143
545
1,633
52
13,840
2,531
3,394
19,765
""
1,951
--
—
1,951
86
--
451
1,994
52
16,584
2,921
_2J02
23/01
796
-------
without escalation and contingency) and the capitalized annual operating
cost. This latter cost is just the estimated annual operating cost (sum-
marized in Table 3) divided by the annual fixed charge rate of 18 percent-.
5.1 Capital Costs of Alternatives
The subsystem capital costs of the four ammonia systems and the
integrated wet/dry system are listed in Table 4. The integrated dry/wet
concept is considered a baseline for comparison because it represents
current practice. The San Juan Plant Unit 3 is currently being constructed
with a heat dissipation system of this type. The estimate of the integrated
dry/wet concept was performed without the benefit of prior knowledge of
actual construction costs of San Junit Unit 3 to put all estimates on the
same relative basis and may or may not correspond to actual costs.
5.2 Comparative Costs
The comparable costs of the four concepts plus the state-of-the-art
integrated Wet/Dry are listed in Table 5.
Table 5. Summary of Comparative Capital Costs
(dollars)
Basic Capitalized Comparable
Cooling Tower Concept Capital Cost Operating Cost Capital Cost
Integrated dry/wet 23,407,000 9,586,000 32,993,000
Vertical HOTERV tower 24,795,000 3,138,000 27,933,000
Horizontal HOTERV tower 21,590,000 3,102,000 24,692,000
SCAT tower 21,493,000 2,969,000 24,462,000
Augmenting Ammonia
Condenser 19,765,000 4,167,000 23,932,000
797
-------
The costs presented in this paper are approximate in nature.
None of the concept designs were fully optimized from the standpoint
of all parameters involved. However, all designs and estimates were
arrived at utilizing the same bases and uniform procedures. It is not
anticipated that exhaustive optimization would change the relative
ranking of the concepts with regard to comparative capital costs.
The ammonia systems were found to have potentially lower capital
and operating costs than comparable capital cost for the integrated
concept considered in this base study. Although the ammonia systems require
(1) an aninonia reboiler, which may be somewhat more complex and expensive
than a simple- condenser, and (2) a complex pressurized ammonia fill and
drain system, the ammonia systems have a number of important
cost advantages associated with the evaporation-condensation heat transfer
system. Among these advantages is the enhanced heat transport from the
reboiler to the cooling tower. Only small pumps are required to return
the ammonia to the reboiler and to provide forced recirculation. Water treat-
ment costs are also less because of the need for treating smaller quantities
of water. Moreover, the cost of the ammonia condenser/reboiler was conserva-
tively estimated to be significantly greater than a conventional turbine
condenser but there is reason to question that estimate.
Operating costs for the ammonia systems are substantially less than
the integrated concept because 1) less power is required to operate recir-
culation pumps and fans, and 2) the capacity penalty is lower because
less generating capacity must be provided in reserve.
798
-------
LAYOUT OF STEAM PLANT WITH AMMONIA
AMMONIA
VAPOR
DELUGE
WATER
TURBINE GENERATOR
CONDENSATE
CONDENSER
REBOILER i:
VAPOR LIQUID
SEPARATOR
HOT WELL
AMMONIA
STORAGE
PUMP HOUSE
FOR DELUGE WATER AND
AMMONIA CONDENSATE
FIGURE 1.
-------
VERTICAL HOTERV ARRANGEMENT
28'
-------
HORIZONTAL HOTERV ARRANGEMENT
205
28'<£FAN, TYP FOR 30
-230'
n I-H-
TOWER *1
(TYPICAL)
\ L
iA
AIR OUT
t t t
B
PLAN
AIR IN -*>
ELEVATION "A-A"
ELEVATION "B-B"
FIGURE 3.
801
-------
SCAT TOWER USING CURT1SS-WRIGHT HEAT
EXCHANGERS
TUBE BUNDLES
TYP FOR 61
FAN SUPPORT
COLUMNS
WET TOWER CELL
TYP FOR 16
28'
-------
SCHEMATIC OF CURT1SS-WRIGHT EXCHANGER
ADAPTED FOR THE SCAT TOWER
SLOTTED
FINS
12.3'
FIGURE 5.
803
-------
INTEGRATED DRY/WET COOL ING TOWER
00
o
FAN CYLINDER
ELECTRIC MOTOR DRIVER
GEAR
LREDUCER
MULTI-BLADE
FAN
DRY
SURFACE
DRY/WET
EXCHANGER
SEPARATION
WALL
ULL LENGTH
FLUME
WET
SECTION^
BYPASS
SECTION ,
CONCRETE COLLECTION
BASIN
CROSS SECTION
FIGURE 6.
-------
ENVIRONMENTAL ASPECTS OF EFFECTIVE ENERGY
UTILIZATION IN INDUSTRY*
Robert E. Mournighan
U. S. Environmental Protection Agency
Industrial Environmental Research Laboratory
Office of Research and Development
Cincinnati, Ohio
ABSTRACT
This paper presents some of the energy conservation program in
which the Power Technology and Conservation Branch, of ..the EPA's Office
of Research and Development, is involved. Initial results -of hardware
research and development projects are presented.
Examples of combined energy conservation-pollution control projects
concerning the glass, steel and textile industries are given. These are
research programs funded under a coordinated federal program.
The tentative results of these studies indicate that 30-40% of
energy is wasted in industrial manufacturing processes. Effective
utilization of energy could provide at least a partial solution to our
energy supply problem. At the same time, effort must be made to reduce
the pollution associated with the waste streams, or else a great oppor-
tunity will pass by, resulting in a waste of economic resources and
unnecessary pollution.
Three specific projects are the principal focus for our discussion:
(1) preheating of glass furnace batch with waste furnace emission gases;
(2) dry quenching of coke in the steel industry; and (3) reverse osmosis
recovery of hot textile dye wastes.. The technologies being investigated
have the combined advantages of improved water or air pollution control
with energy, water and/or chemicals recovery.
INTRODUCTION
The U. S. Environmental Protection Agency (EPA) is conducting a
research effort in assessing the environmental aspects of efficient
energy utilization technologies. A major goal of this program is to
*Presented at the "Second Conference on Waste Heat Management and Utilization,'
December 1978. Mr. Mournighan is with the Power Technology and Conservation
Branch, Energy Systems Environmental Control Division.
805
-------
foster a reduction in environmental pollution, particularly air pollution,
while at the same time making use of energy currently wasted.
The program is managed by the Power Technology and Conservation
Branch (PTCB) of EPA's Industrial Environmental Research Laboratory,
Cincinnati, Ohio, under EPA's Office of Research and Development.
PTCB's activities are concerned with the environmental control problems
and benefits of a broad range of energy technologies. Branch interests
include: solar and geothermal energy conversion; advanced conversion
systems such as high temperature turbines, magnetohydrodynamics and fuel
cells; energy management, i.e., conservation and energy recovery in all
sectors; and indoor air quality control in homes and public buildings.
The results of the PTCB program will provide useful data to other EPA
functions, such as, air and water standards development, and will be
useful to other federal, state and local agencies concerned with energy
environmental issues in developing the most environmentally sound energy
technologies.
This paper presents, principally, the results from three industrial
projects on improved energy management which have positive energy savings
and environmental impacts and whose economics indicate a moderate chance
for success.
BACKGROUND
The EPA industrial energy research program has three primary areas
of activity: (1) assessment of the amount of industrial energy which
can be recovered, including, where possible, the amount and types of
associated pollutants; (2) assessment of the technologies available; and
(3) support of technology research and development.
In the first area, an assessment of the amount of waste heat available
is being done by DSS Engineers with field measurements by KVB, Inc.
(The project will be reported in the following paper.)
In the second area, an evaluation of technologies for recovering
waste heat is the objective of a project now being conducted by Energy
and Environmental Analysis, Inc. The results of this project are expected
to define the areas of greatest potential and the technologies which
have the best energy/environmental tradeoffs.
In the third area of interest, research and development of technologies
which have significant energy savings associated with a positive environ-
mental impact, are three projects applicable to the glass, steel, and
textile industries, being conducted by PTCB in conjunction with EPA
industrial groups and the Industrial Energy Conservation Division of the
Department of Energy. On the textile project, the Department of the
Interior also has an interest.
806
-------
DISCUSSION
Table I shows the estimated energy consumption, wasted energy, and
estimated energy savings possible for the container glass industry
(Largest segment of glass industry), the coking operation in the steel
industry, and in the textile dying and finishing industry.
Table I. Estimates of Energy Consumption, Wasted
Energy and Possible Energy Savings
Energy Possible
Consumption Energy Waste Energy Savings
1015 Btu/yr. 1015 Btu/yr. IQ1^ Btu/yr.
Glass .22 .140 .018 (.052)
Steel (Coking) 1.44 .44 .14
Textile Dying
and Finishing 0.54 .37 .23
The table shows that of the 2.2 quads (10 Btu) consumed, the
waste energy for these three categories is 0.95 quads. The third column
shows that an estimated 0.4 of the 0.95 quads can be saved by effective
and economic technologies. To put the possible savings in perspective,
it represents about 14% of the yearly oil needs of the electric utility
industry or about 65 million barrels of oil per year.
GLASS INDUSTRY PROJECT - PREHEATING WITH STACK GASES
The major reasons for undertaking a project in the glass industry
are the declining reliability of natural gas as an energy source, the
need for reducing stack emissions and a need to find an environmentally
and economically acceptable solution to both problems.
Under contract to EPA, Battelle Memorial Institute and Corning
Glass Works initiated a bench scale project to study the feasibility of
preheating pelletized glass batch raw materials with waste furnace gas.
It was theorized that the batch pellets could also capture some of the
pollutants from the stack gas. Figure 1 shows a simplified diagram of
the process.
The study was recently completed^7), and it was found that without
impairing glass quality:
• Soda lime glass batch can be pelletized successfully;
807
-------
The. polJeU: can be heated with the waste gas Lo 800° C
v.vi t.hout stjck.iug together or deformi ng excessively;
o The w.'ltinp, furnace temperature may be. reduced up t:o
50° C, reducing tin- furnace energy requirement by about
'/OZ - throughput: may also be', increased sij;n:i f icantly to
bring the overall savings t:o about ^0%;
o Up to 85% of the SO,, can be absorbed from the waste
gases, by the pclletix.ed material;
o NOj, f7:0:11 the f,r,s conibustion zone cnn be reduced nbout 65%
by instituting the above, mentioned furnace temperature
drop and fuel reductions;
&
o The pelle.tized biitcli may capture, as uiuc.li as 75-80% of the
furnace particulate ernir.sions. Further worl: is necessary
to deteruiine exactly how accurate this figure is.
Referring back to Table I, the item in parenthesis under "Possible
Energy Savings" represents the energy savings, if this technology could
be applied to the v;hole indu.stry.
There are other advantages to this process modification in that it
can be: retrofitted on existing furnaces and, through energy efficiency
and pollution control, it should prove to be of economic benefit to the
industry. Also, it has flexibility and efficiency that are not found in
the use of v;aste heat boilers and a large scale switch to electric
melting. Currently, a proposal is under consideration for the construction
and operation of a pilot plant for the evaluation of this technology.
STEEL INDUSTRY - DRY QUENCHING OF COKE
Most people are aware that the steel industry uses a large amount
of energy in the manufacturing processes. Most of the time we picture
the white-hot heat, of the furnace with sparks of white-hot metal flying
through the air; or the incandescent flowing liquid being poured into
molds or transfer cars.
There is, however, a less spectacular place in the. integrated steel
mill where there is a tremendous waste of energy and a considerable
source of pollution: the coke batteries and quench tower.
Currently, hot coke (at 1000 to 1100° C) is transported from the
coke ovens to a tower where tons of water are sprayed onto the incandescent
coke to cool it below ignition temperature. In this process, large
quantities of steam and pollutants are vented to the atmosphere. A
significant amount of polluted water i
-------
As shown in Figure 2, a dry quenching system, hot coke is cooled in
a closed system by recirculating inert gas. The hot gas, in turn, is
used to generate steam in a waste heat boiler or is treated by other
heat removal techniques.
In 1977, the Department of Energy (then ERDA) and EPA initiated a
contract with National Steel Corporation to determine the economic and
technical feasibility and the design for installing a dry quenching unit
at one of their locations. Recently, the project was completed(^ with
the following results:
• Dry quenching of coke, while a huge energy saver would
not be economically attractive because of the huge initial
investment for equipment - $21,000,000 per 1,000,000
ton/yr. plant;
• If the pollution abatement benefits must be figured into
the economics, the process becomes much more attractive.
The energy benefits are considerable. Referring to Table I, the
possible energy savings as seen by Streb(^) could be as much as 0.14
quads. Even at that, the economics are poor, with a projected return on
investment between 7 and 16%.
However, emissions to the atmosphere are projected to be less than
0.01 grans of particulate/SCF gas at 1 million SCF/hr. or an emission
rate of about 1.5 pounds per hour, about 10% that of a large industrial
boiler.
Water pollution with this system will be at a minimum. The volume
of heavily contaminated water is orders of magnitude less than with the
wet quenching system. The only problem, however, is that we don't know
what toxics, if any, are present and how much there may be.
In any case, in making a comparison of wet vs. dry quencing, the
pollution control estimates for wet quenching must be figured in, and
the technology should not be evaluated on energy savings alone.
TEXTILE INDUSTRY PROJECT - ENERGY AND WASTE RECOVERY BY REVERSE OSMOSIS
The textile industry in this country consumes about 300 million
gallons of process water per day. Approximately 50% of this water is
hot (>40° C), and is not reused, thereby wasting a tremendous amount of
energy and at the same time causing thermal pollution problems^-'.
EPA, the Department of Energy and the Department of Interior have
formed a joint effort to develop technology which would (1) conserve as
much waste heat as possible, (2) conserve the water resources, and (3)
solve the problem of toxic chemical discharges from these plants. A
dying and finishing plant with a continuous processing unit was chosen
for the project.
809
-------
Figure 3 is a simplified diagram of the washing process, the largest
consumer of hot water in the process. Before this project was begun,
each section received a separate supply of fresh hot water and the
discharge sent to the waste treatment plant.
To meet all the goals of the project, reverse osmosis (RO) was
chosen as the water cleanup process. A high temperature membrane,
recently developed and tested (serviceable up to 100° C), seems to be
the best choice available. Membrane systems are very expensive, so a
water conservation program was undertaken to minimize the size and cost
of the RO unit. The water conservation effort was very successful,
showing it possible to reduce the process flow to a washing section from
300 gpm to about 50 gpm.
The RO unit has been sized to treat the 50 gpm stream and will in
effect reduce the discharge to the waste treatment system to a small
blowdown of about 4 gpm. Most of the hot water will be directly recycled
back to the process. Figure 4 is a diagram of the proposed installation.
Currently, the project is the stage where equipment is being purchased,
with construction starting in a few months. After the system is complete,
the unit will undergo shakedown testing for a full year to determine the
economics, energy savings and pollution control benefits.
The projection is that there will be a return on investment for
this system, but will be below 20% energy and water recovery considered
alone. In comparision with separate energy and pollution control alternatives,
the system looks very attractive indeed, especially when toxics substance
control is brought into the picture.
A WORD ABOUT RETURN ON INVESTMENT
In a recent article in the Harvard Business Review^-*-"', the point
was made that most manufacturers set a high rate of return for energy
projects, actually higher than their regular capital projects. This
adds to the energy conservationists woes. In the article, the authors
point out that if firms would accept lower rates of return and fund
energy recovery projects, they would find waste heat recovery systems
providing energy at about one-half the cost that the utility itself
would charge if it were the supplier.
The point is that if an energy conservation project or waste heat
recovery project has any pollution control benefits, it should be looked
on by management in this light. We should not look upon these projects
as quick expense-cutting programs, but as an investment, guarding against
future higher energy and pollution control costs. Conservation projects
are probably the least capital intensive way to save our energy resources' '.
810
-------
CONCLUSION
In summary, I have given three examples of energy conservation
projects that involve our Agency. We are. generally trying to encourage
the idea of adding in pollution control benefits in the decision-making
process in energy conservation projects. -We feel that, with all this
considered, energy conservation alternatives can help our country through
the energy squeeze, maintain the environment and still permit a healthy
economy.
311
-------
REFERENCES
1. Brandon, C. A., et al. "Hot Textile Process Effluent Recycle by
Membrane Separation." Presented at the 85th National Meeting of the
American Institute of Chemical Engineers, Philadelphia, PA, June 4-8,
1978.
2. Lee, C. C. "Potential Research Programs in Waste Energy Utilization,"
Proceedings of Second Waste Heat Management and Utilization Conference,
Miami Beach, Florida, May 9-11, 1977.
3. Mournighan, R. E. and Bostian, H. E. "Energy Conservation and Improvement
of the Environment," proceedings of Fifth Conference on Energy and the
Environment, Cincinnati, Ohio, November 4-7, 1977.
4. Sternlicht, B. "Capturing Energy from Industrial Waste Heat," Mechanical
Engineering, p. 30, August 1978.
5. Streb, A. J. "Priority Listing of Industrial Processes by Total Energy
Consumption and Potential for Savings," ERDA CONS/50151, 1977.
'6. Draft Final Report, DOE Contract EC-77-C-024553. "Dry Coke Quenching,"
Phase I Engineering.
7. Draft Report, EPA Contract 68-02-2640. "Technology for the Conservation
of Energy and Abatement of Emissions in Glass Melting Furnaces."
8. "Environmental Considerations of Selected Energy Conserving Manufacturing
Process Options," EPA report, EPA 600/7-76-034k.
9. Steam Electric Plant Factors, 1976, National Coal Association, Washington,
DC.
10. Hatsopoulos, G. N., et al. Harvard Business Review, March 1978.
812
-------
PELLETIZER
CHARGE
HOPPER
00
M
CO
PREHEATER
GULLET
TO STACK
MELTING'
FURNACE
Figure 1 : Diagram of Glass Melting Furnace With Preheat
MELT
-------
COKE OVEN
Co
QUENCH
CHAMBER
CHARGE CHAMBER
INERT
GAS
COKE CONVEYOR
STEAM
TO PROCESS
WASTE HEAT
BOILER
Figure 2: Dry Coke Quenching Process
-------
TO
SEWER
WATER
00
H
(Jl
tr.
WATER
EACID
SOAPING
sa
WATER
NEUTRALIZE
TO
SEWER
Figure.3: Continuous Dyeing Process - Before changes
-------
CLOTH-
00
r
ACID
SOAPING
5 a e
HOAC
MAKEUP
t_J
NEUTRALIZE
T
10
PRODUCT
T
CONCENTRATE
RECYCLE
CONCENTRATE
Figure A: Continuous Dyeing Process - Washing Section Flow Schematic.
with recycle
-------
WASTE HEAT RECOVERY POTENTIAL
FOR
ENVIRONMENTAL BENEFIT
IN
SELECTED INDUSTRIES
Prepared By
S. R. Latour
J. G. Menningmann
DSS Engineers, Inc., Ft, Lauderdale, Fla.
Dr. C. C. Lee
U. S. Environmental Protection Agency
Industrial Environmental Research Laboratories
Cincinnati, Ohio
817
-------
The power Technology and Conservation Branch of the EPA's Industrial Env-
ironmental Research Laboratory in Cincinnati, Ohio, is currently conducting
a program intended to assess the relative economic/environmental impacts of
waste heat utilization. The reasons for the EPA's involvement in this area
are twofold:
1) First, increasing the efficiency of energy utilization may be
considered a pollution control alternative in that the resulting
decrease in fuel consumption will also result in a corresponding de-
crease in quantity of pollutants discharged.
2) Secondly, it is necessary to insure that as these more efficient
systems are developed, new pollutants are not generated which would
adversely affect the environment.
As a result, the EPA has funded the study title, "Waste Heat Recovery
Potential for Environmental Benefit in Selected Industries." The objective
of this study is to identify the points, quantity and quality of heat dis-
charged by Energy Intensive Industries and Emerging Technologies for Energy
development. Energy Intensive Industries were selected on the premise
that those industries which consumed the greatest quantities of energy
offered the greatest potential for discharging substantial quantities of
waste heat to the environment. Consideration was also given to the thermal
intensity and diversification of each industry. Table #1 lists those 4-
digit SIC classification included in the study.
TABLE #1
SELECTED INDUSTRIES
SIC #
2611 Pulp Mills 2911 Petroleum Refineries
2621 Paper Mills (ex Bldg Paper) 3211 Flat Glass
2631 Paperboard Mills 3221 Glass Containers
2812 Alkalies and Chlorines 3229 Pressed & Blown Glass
2813 Industrial Gases 3241 Cement Hydraulic
2819 Industrial Inorganic Chemicals 3274 Lime
2865 Cyclic Crudes and Intermediates 3312 Blast Furnace & Steel Mills
2869 Industrial Organic Chemicals 3321 Grey Iron Foundries
2873 Nitrogenous Fertilizers 3331 Primary Copper
3334 Primary Aluminum
818
-------
For each of these industries, a study was conducted to document the points.
quality, and quantity of all waste heat discharges to the environment. The
major source of data collected on flue gases was from the National Emmissions
Data System's ( NEDS ) Point Source Listings. This data was then verified
by discussion with various industry officials and by correlation with other
related studies conducted both by the EPA and the DOE.
Data on wastewater and non-contact cooling waters containing significant
quantities of waste heat were also identified, when possible, from EPA
Development Documents for Effluent Limitations, correpondence with Industrial
Pollution Control Officers ..literature surveys,and various U.S. Government
sponsored R & D Reports,
Since it is not possible, within the scope of this presentation, to present
the data collected in this study in the detail contained in the final report,
and since considerable variations in the accuracy of the data on wastewater
discharges exist between each SIC classification, it was decided to present
a summary of only the flue gas emmission of each industry. This data
accounts for about 99% of the total waste heat discharged, from industries
such as cement production, to approximately 50% for such industries as
petroleum refining and steel production which utilize considerable quantities
of both non-contact cooling and process wastewaters.
For this presentation it was decided that four graphs would be adequate to
summarize the main findings of the original report.
In Fig. 1, the annual waste heat discharged by flue gasses versus the stan-
dard industrial classifications is represented. As mentioned previously,
the industries such as petroleum refineries and steel mills (incl blast
furnaces) are only represented by about 50% of there total waste heat, the
other 50% was contributed by wastewaters, whereas cement and paper-mills are
819
-------
8.0-
7.0 •
6.0
5.0-
4.0
3.0
2.0
.0
-FIGURE #1
*ANNUAL WASTE HEAT DISCHARGED
BY
STANDARD INDUSTRIAL CLASSIFICATIONS (SIC #) (1977 EST.)
n
n
i — CM CO
IO VO U)
CM cMCM
CM oo en m en n i — •— • — 01 • —
i — i — « — U3 VO f^-i — i— CM CM^d-
00 -OO CO 00 00 COCTiCVJCM OJOJ
csicMCMCMCM cMOJcoco roco
CM
CM i — i —
r — C\JCO
COCOCO
fO
ue Gasses Only
SIC #'S
820
-------
represented by roughly 99% and 96% of their total waste heat because of the
minimal amounts of non-contract cooling and process wastewaters utilized
within these industries. From these figures and the wastewater data it was
determined that approximately 50% of the total waste heat discharged, with-
in these SIC groups, is discharged by petroleum refineries and steel mills.
With this in consideration it seems apparent that petroleum refining and
steel production should be prime candidates for further research into the
potentials for energy recovery.
The percentage of waste heat discharged above 350 F is the subject of Fig. 2.
Waste heat streams above this temperature were termed BTU's available be-
cause of three primary reasons: 1) this is the approximate dew point of
sulfuric acid, which is present as an acid gas in most combustion processes,
and can deteriorate equipment materials such as baghouse fabrics and stack
liners when condensed out of the flue gas; 2) heat recovery tends to reduce
the buoyancy of stack plumes thereby reducing plume height and causing an
increase in ground level concentrations of sulfur and nitrogen oxides; and
3) temperatures lov/er than 250 F do not prove advantageous for "heat ex-
changer devices'^ however there are systems such as heat pumps, conventional
and direct contact organic rankine cycles that do operate in this range.
Fig. 2 illustrates the industries that have the greatest percentages of
"waste heat available". These industries show more potential for energy
recovery via conventional "heat exchanger devices."
In Fig. 3 we see the percent of purchased fuels and electricity discharged
as waste heat by standard industrial classifications. This data was also
generated with only flue gas discharges.
It should be noted that in Fig. 3, for example petroleum refining, discharges
about 62% of their purchased fuels and elec. by flue gas. This may -not
821
-------
FIGURE #2
* PERCENT WASTE HEAT DISCHARGED
ABOVE 350'F
BY
STANDARD INDUSTRIAL CLASSIFICATION (SIC #)
60 -i
50 -
o
t/0
UJ o
IE O
in
LU OO
oo LU
3 o
u_
o
20
10
r— CM
LjO
i — ID
co co
CM CM
CTt
co
CM
CO
co
CM
CT> CM
CM 00
CM
CM
00
cr>
CM
CM
OO
CM
oo
CNJ
oo
00
OO
CM OO OO
OO OO OO
OO OO OO
*Flue Gasses Only
STANDARD INDUSTRIAL CLASSIFICATIONS
(SIC #'S)
822
-------
FIGURE #3
PERCENT OF PURCHASED FUELS AND ELECTRICITY
DISCHARGED AS WASTE HEAT BY
STANDARD INDUSTRIAL CLASSIFICATION (SIC #)
(1977 EST.)
801
70
60
o
I—I
OL \-
l-<
O Ul
ui :c
_i
Ul Ul
O CO 50
?g
co co
-i<
Ul
=> Q
U Ul
Ul
O CO
ct: i-i
ZD Q
D-
40
30
20
10
Total BTU's Discharged
Total BTU's Discharged
Above 350°F
n
r— CM CO
V£> VO «3
CM CM CM
CM CO
CO CO
CM CM
cr> ur>
r- <£>
CO CO
CM CM
01
co
CM
*Flue Gasses Only
CO r—
r-- i —
co en
CM CM
SIC #•
r— r— Cn r- <=T CM 1— i— Si-
i— CM CM «^-r~-i— CM roro
CMCMCM CMCMCOCO COOO
rococo rocoroco coco
823
-------
seem reasonable at first glance because flue gas only represents about 50%
of the waste heat discharged in refineries. This means that 124% of pur-
chased fuels and electricity is discharged by flue gases and waste waters
as v/aste heat. The explanation for this is that 50% or more of the petroleum
refineries energy needs are supplied with byproduct refinery gas and coke.
Then, with these points considered, petroleum refineries would only discharge
roughly 62% of the total energy consumed by that industry. This is the case
for steel production and to a much lesser extent for the other industries.
Considering Fig. 3 one can see that some industries which have a high % of
purchased fuels and electricity discharged as waste heat do not necessarily
have a high % of "BTU's available", the % black area.Keeping in mind the total
BTU's discharged annually for each Sic #, this Fig. differentiates industries,
which have close to the same percentages of waste heat discharged, by
there potential for energy recovery with conventional "heat exchanger
devices."
The next graph, Fig. 4, gives annual waste heat discharged in the 10 EPA
regions. For the individual Sic #s, we assessed the BTU's discharged in
each region. With this data we could attribute the high percentage of
waste heat in region 5 to primarily steel, petroleum and industrial inorganic
chemicals N.E.C. productions; region 6 to petroleum, industrial organic chem-
icals N.E.C., cement, industrial inorganic chemicals N.E.C. and papermills
(excl. Bldg. paper) productions; region 3 to industrial gases, steel,
petroleum and cement productions; region 4 to papermills, petroleum, cement
paperboard mills and steel productions.
This data reflects the regional potentials for commercial use of waste heat
in the fields of space heating, soil warming, aquaculture farming and
other potential uses of low grade waste heat.
824
-------
10
UJ
CD
O
h-1
a
8
U:H
JC O
^"~
t-x 6
3 vi
—i :o
3 ca
=E
<
2-
FIGURE #4
ANNUAL WASTE HEAT DISCHARGED
BY
EPA REGIONS
(1977 EST.)
n
n
1 23456789 10
EPA REGIONS
*Flue Gasses Only
825
-------
FIGURE 15
EPA REGIONS
00
ro
(incl. ALASKA)
(CALIF *
(incl.HAWAII) ^ j
-------
POSSIBLE ENVIRONMENTAL IMPACTS
Waste heat discharged by industry can create undesirable thermal loading of
the local environment. The impact of these heat additions is dependent
upon both the concentration and route by which this heat energy is discharged.
While a variety of pollutants are often directly related to waste heat dis-
charge, this discussion will focus primarily on the impact of the thermal
discharges to the environment.
Heat is unlike most "pollutants" which can be readily collected, concentrated,
and disposed of under controlled conditions. Conversely, heat energy must be
disposed of/through retention with controlled dissipation so as to minimize
its effects on the surrounding biosphere.
Heat rejection by cooling and process waters can be a significant amount of
the overall waste heat discharged in some industries. Until recently, a
common method of waste heat disposal was to discharge once through cooling
water into a nearby waterway, this has proven to be both efficient and
economical. However, public concern over the potential adverse environ-
mental impacts caused by the addition of this waste heat to natural water-
ways has promoted considerable research in an attempt to define these im-
pacts in order to determine "acceptable" levels of thermal pollution.
Temperature is one of the most important single factors governing the
occurance and behavior of life. The discharge of waste heat into a
ifr
natural body of water can cause a number of physical, biological and chemical
effects. Raising the temperature of water reduces the oxygen retaining
capacity of the water, reduces the reaeration rate, changes the density
which may inturn result in stratification, increases the rate of evaporation,
increases the rate of many biological, chemical and physical reactions,
827
-------
and decreases the viscosity thereby reducing the sediment transporting
ability of the water.
The impacts of these changes can be detrimental, beneficial, or insignificant
depending upon the extent of these changes and the intended use of the re-
ceiving body of water. Heat imputs into a receiving body of water increases
the rate of BOD exertion which, when coupled with the accompanying reduced
reaeration rate, may reduce its organic waste assimilation capacity. On
the other hand, the addition of waste heat during winter months may signifi-
cantly lengthen the shipping season of a waterway by shortening the period
of ice cover in the shipping lanes.
However, the greatest potential impact of waste heat discharges to natural
bodies of water is upon the aquatic ecosystem. Although a large number of
studies have been and are being conducted in attempts to further define
these cause and effect relationships, considerable data is still lacking.
Some of the known and reported effects associated with temperature increases
of natural waterways are: decreasing gas (oxygen) solubilities; changes in
species diversity, metabolic rates, reproductive cycles, digestive and re-
spiration rates, behavior of the aquatic organsms; and increasing the para-
sitic bacterial populations. All of these have the potential of creating
an unbalanced, unchecked aquatic ecosystem.
Although the potential adverse environmental impacts of discharging waste
heat into the aquatic environments are quite numerous and diversified the
direct release of this waste heat to the atmosphere is not without its
own potential for adverse environmental impact.
It is becoming increasingly apparent that man affects the climatic condi-
tions of the earth by the release of heat and materials to the atmosphere.
828
-------
For this reason, stack gases and cooling tower plumes are of considerable
concern to investigators from a environmental and energy standpoint.
Several studies have suggested that possible intensification of convective
activity and associated concentration of vorticity may be caused by the
release of large quantities of heat in relatively small areas, resulting
in severe thunderstorms and tornadoes. On a smaller scale, the release of
this heat, and contained moisture, can increase or change the spatial and
temporal pattern of precipitation, cloud cover, and mean temperatures.
Cooling towers, either wet-or-dry, are frequently used to dissipate waste
heat to the atmosphere. The major complaint from the public concerning
these towers has been the appearance of these devices and, at close range,
the noise generated by them. They are, however, several environmental
impacts directly related to the operations of these towers. Some of
these are 1) the restriction of sunlight caused by visible plumes
("shaddowing") 2) restriction of visibility when plumes reach ground level
(fogging) 3) deposition of detrimental chemicals contained in cooling
waters onto surrounding areas ("drift") 4) atmospheric changes. For
most sites these impacts are rather small and local, and usually environ-
mentally acceptable.
Since proven mathematical models are not yet available for accurately
predicting the extent and frequency of these atmospheric effects for a
particular site and heat-dissipation system, considerable field research
will be required to develope these models before accurate determinations of
"critical heat loads" may be projected.
829
-------
WASTE HEAT UTILIZATION AND THE ENVIRONMENT
M.E. Gunn, Jr., Program Manager
Division of Fossil Fuel Utilization
U.S. Department of Energy
Washington, D.C. U.S.A.
ABSTRACT
One way of reducing the national energy needs and conserving
valuable fossil fuels in the near and long term is the recovery
of waste heat energy. Industrial processes, residential and
commercial heating, transportation systems, and electric power
generation efficiently utilize only a small percentage of the
energy fed into them. Much of this energy can be recovered by
using the new technologies now being developed. This recover-
able energy amounts to 20 to 30 percent of the forecast national
energy consumption.
Waste heat recovery also has a beneficial impact on the environ-
ment when the rejected energy is harnessed and used; thermal as
well as air pollution is significantly reduced. In particular,
thermal pollution is reduced to the atmosphere and to waterways,
as applicable, and the pollution to the atmosphere is reduced
when the use of waste heat recovery decreases the quantity of
fossil fuel that would be burned to achieve given performance
levels.
In late 1978, a significant number of preproduction prototype
waste heat recovery systems will be in operation at selected
electric utility and industrial generation stations located
throughout the U.S. They represent modern versions of tech-
nology dating back to the 1930's and more recently used in
aerospace applications. DOE is supporting the development of
several unique concepts of packaged Rankine cycle systems using
different working fluids ranging from steam to Freon. Each
system will be rated under one megawatt and will be ideal for
recovering waste heat from the diesel engines used by many small
municipal electric utilities as well as other waste heat recovery
opportunities in industrial applications. The systems will
recover the waste heat from the prime mover exhaust streams and
convert it to additional useful shaft power at efficiency levels
of 18-20 percent.
This paper will describe the technology as it is characterized
in the DOE sponsored concepts and will address the impacts, pro
and con, on the environment as a result of their implementation.
The conditions of using the organic fluids as working fluids
will be discussed and an attempt will be made to quantify
selected thermal and air pollution improvements.
830
-------
INTRODUCTION
A most important principle of the President's National Energy
Plan (NEP) is seen as the "Cornerstone of National Energy
Policy." This_principle states that the growth energy demand
must be restrained through conservation and improved energy
efficiency. Conservation represents practice that is cheaper
than production of new energy supplies, and a most effective
means for protecting the environment. See Figure 1.
The level of imported fuels can be substantially reduced by
pursuing attractive energy efficient technologies. Studies
have estimated that overall economy of the U.S. operates at
less than 10 percent energy efficiency. A recent analysis
determined that the U.S. could expend between 20 and 40 per-
cent less energy and still maintain overall economic growth
into the 1990's. Much of the required energy is lost as waste
heat and can be recovered by using new technologies. This
recoverable energy can amount to 20 to 30 percent of the fore-
cast national energy consumption.
If the waste heat recovery program currently underway via
Federal support and sponsorship were carried through to com-
pletion, potentially, the Nation's annual expenditure for im-
ported oil can be reduced by $15 billion in 1985 and by $48
billion in the year 2000. Full utilization of recoverable
waste heat energy would result in potential savings of $57
billion in 1985 and by'$68 billion in 2000. See Figure 2.
By reducing the need for additional oil imports by recovering
and making use of waste energy, conservation and improved
efficiency in the use of energy can contribute to national
security and international stability. This leads to the
possible reduction of the need for additional domestic energy
production, thereby contributing to environmental protection.
To achieve these savings and ultimate improvements for the
environment, DOE has been supporting waste heat recovery and
utilization projects since 1975.
In the High Temperature Heat Recovery program at DOE (Figure 3)
a goal was established to develop heat recovery technology as
an alternate source of energy by developing the necessary
technology base for recovering and using waste heat and by
demonstrating the technical and economic feasibility of the
technological components. As shown in Figure 4, specific
projects of this program include the development of several
unique concepts of organic Rankine cycle systems. These
systems are ideal for recovering waste heat from diesel
engines used by many small municipal electric utilities and
831
-------
are suitable for recovering waste heat in industrial processes.
The shear implementation of these systems will significantly
reduce thermal and air pollution typically characteristic of
the respective prime movers, and the impact on the environment
may even be considered negligible when one takes into account
the system performance and reliability aspects.
THE TECHNOLOGY
The organic Rankine cycle system technology being developed in
the DOE program represents modern versions of technology dating
back to the 1930's and more recently used in aerospace applica-
tions. Basically, a Rankine cycle system, depicted in Figure 5,
is a thermally driven engine that converts heat energy into the
mechanical energy by alternately evaporating a working fluid at
high pressure and producing shaft power which operates at low
pressure.
The Rankine cycle system can be readily identified as the
thermodynamic cycle that characterizes a steam generation
system used to produce electricity. The use of organic fluids
instead of steam offers advantages as well as the disadvantages
listed in Figure 6. As indicated in Figure 7, organic fluids
typically have low heats of vaporization, thereby allowing for
sensible heat use at the lower temperature conditions. There-
fore, the systems being developed are suitable for low or
middle temperature heat utilization.
Before the end of 1979, at least four preproduction prototype
organic Rankine cycle waste heat recovery systems will be in
operation at selected electric utility generation stations in
the United States. The units will produce additional electric
power from the exhausts of stationary diesel engines.
Three of the units are products of a DOE/Sundstrand Energy
Systems Cooperative Agreement. Under this agreement, Sund-
strand has designed and developed a system which uses toluene
as a working fluid to generate 600 Kw of electric power. The
fourth unit was developed by Mechanical Technology Incorporated
(MTI), and employs two safe and well accepted power fluids,,
steam and Freon to generate 500 Kw of electric power.
As illustrated in Figure 8, the Sundstrand system uses a single
stage supersonic high-work impulse turbine, and a vaporizer
which uses a compact centrifugal separator to remove liquid
from the vapor stream at the outlet of its natural circulation
boiler. A modular packaging concept (Figure 9) is employed so
that the power conversion system may be easily transported and
set up without special foundations. One organic Rankine cycle
loop is entirely sealed with exception of the turbine output
832
-------
shaft. The regenerator, condenser and hotwell are combined in
a single vessel. Sundstrand units will be in operation at
municipal utility plants located at Beloit, Kansas, Easton,
Maryland, and Homestead, Florida.
The MTI unit is characterized by the cycle shown in Figure 10.
In this system basically, two Rankine cycles are employed.
The steam topping cycle buffers the Freon bottoming cycle
enabling the system to be applicable over a wider range of
gas temperatures. The machinery arrangement consists of two
radial in-flow turbines that drive a common output gear. The
system is designed to use exhaust gases at 520°F to generate
steam at 430°F which is expanded across the Freon turbine.
Each turbine is independently optimized. The system is low
pressure in character and conventional process and refrigera-
tion industry heat exchange components have been adapted for
use. The system is neatly packaged for simplicity in trans-
portation and installation. The MTI unit will be recovering
exhaust gases from two turbocharged diesels in operation at
the Municipal Power Plant in the Village of Rockville Centre,
New York (Figure 11)
INSTALLATIONS AND ENVIRONMENTAL CONSIDERATIONS
Unlike nuclear and fossil fuel cycles, the basic fuel cycle
for these waste heat power conversion systems is located at
the source of fuel, which is, in this case, exhaust gases
from stationary diesel engines. Therefore, environmental
effects occur mainly during the operation phase of these
systems, and are very site specific. This operating phase
consists of power generation under specific load conditions
and constraints that might be imposed by the system users.
Typically, the environmental factors normally considered for
power generation plants include but are not limited to land
use, noise, seismic effects, thermal discharges, and gaseous
and liquid effluents. Figures 12a through c provides illus-
trations of the site plans for the Sunstrand installations.
It is readily apparent that adequate land area is available
at each installation. This is also true for the MTI installa-
tion shown in Figure 13. Perhaps, the most significant
aspect of this installation is the location of the steam
boiler. This component is mounted between the exhaust
stacks of the two diesel engines supplying the recovery
system "fuel", on the top of the diesel engine building.
These initial installations are retrofits to existing facil-
ities. The main disturbance of the land area takes place at
the Sundstrand installations where toluene sumps are made
available to contain any major leakage of the toluene inven-
tory. The system contains ~900 gallons of toluene when fully
charged. The tank is designed to prevent any leakage into
833
-------
into the earth and is buried well above water table levels at
each site. Aside from this, no significant land modifications
are required, i.e., mining, well digging, etc.
Seismic problems are not seen to pose any significant concerns.
Structural integrity for each installation will be consistent
with the existing powerplant requirements and will be at least
as safe as the primary systems.
Noise problems are centered around the turbomachinery or power
conversion components of each unit. Considering the noise
level of the muffled diesel engines operating in the existing
facilities, the turbomachinery noise cannot be detected during
operation. Since the waste heat recovery system only operates
when the diesels are running, noise pollution can be considered
negligible for the additional systems.
The discussion of thermal pollution is concentrated mainly at
two interfaces — the vaporizers at the heat source and the
cooling towers at the heat sink. The Sundstrand system is
designed for vaporizer application in heat sources between
800°F and 330°F; the MTI unit is designed for exhaust (heat
source) temperatures of 520°F with the heat source exit
temperature at 333°F. Considering that typical exhaust
temperatures for large stationary diesels range as high as
1200°F, it is readily apparent that when the entire exhaust
streams are captured by the systems in question, or when the
available heat source 'exceeds ~10xlO^ Btu/hr., above the 330°F
temperature, there is.a substantial reduction in thermal
pollution as a result of waste heat recovery system implemen-
tation. Even if the systems recover only a portion of the
available waste heat, the thermal impact on the environment is
reduced. Although no thermal discharge measurements have been
recorded as of yet for either the Sundstrand or MTI unit
under actual operation, one can expect that the above drawn
conclusion will be substantiated.
For the installations discussed in this writing, cooling
towers are utilized to reject the energy transferred at the
condensing systems for each unit. The MTI unit is designed to
reject energy at 67°F to the cooling tower. The Sundstrand
system has a liquid (cooling water) side condenser exit
temperature of 100°F. Of course, the cooling water temperature
from each system is reduced via the cooling towers by some 6°F
to 15°F. Therefore, thermal pollution at the cooling towers
of these waste heat systems is minimal.
When considering gaseous effluents to the environment, basically
two areas of concern come to mind. The first area is at the
diesel exhause stacks. A question raised here is with respect
834
-------
to any change in the exhaust stream of the diesel engines as a
result of heat extraction. The mere fact that implementation
of these organic systems to generate power from expended
energy sources leads to the analogous situation that would
exist to, say, generate that same power using a prime mover,
such as another diesel. Figure 14 shows a plot depicting the
impact on emissions from fuel combustion relative to the
efficiency of utilization of fossil fuel. The 20 percent
efficient bottoming plants improve fuel utilization efficiency
by up to 10 points. This corresponds to significant reduction
in emissions and represents the emissions impact on the
environment that does not occur as a result of using waste
heat as a fuel source.
The extraction of heat from the diesel exhaust stream does
raise possible concern for sulphuric acid formation in the
stacks and subsequent acid mist introduced to the atmosphere.
In each installation the fuel for the prime movers is rela-
tively clean #2 fuel oil. Thermal conditions, however, are
related to the formation of sulphuric acid. In a recently
completed report that included diesel exhaust gas analysis,
the exhausts from five large diesel engines were sampled over
a range of engine operating conditions using fuels with sul-
phur contents varying from 0.05 percent to 1.8 percent
(Figure 15). The exhausts were characterized via measurements
of S02, S03, CO, H20, NO, chloride, acid dew point, peak rate
temperature of acid deposition, particulate loading, particle
sizing, particulate composition and smoke number. The results
of the analysis were used to determine that the temperature
where the peak acid deposition rate was approximately 20°F
lower than the determined acid dew point temperature of ^240°F.
The peak acid deposition rate corresponds to the point of
maximum corrosive environment for the vaporizer. Therefore,
if acid formation is avoided, problems with regard to acid
mist and corrosion can be mitigated.
Recall that in the Sundstrand system the lowest allowable
exhaust temperature after heat extraction is 330°F, while in
the MTI system the steam boiler is designed so that the diesel
exhaust temperature never drops below 333°F. Both are safely
above the acid dew point limit suggested by the analysis.
The other area of concern when discussing gaseous effluents is
the possible leak of organic vapors or liquids into the
environment. The use of organic fluids raises serious concerns,
at times more emotional than actually hazardous. As indicated
in Figure 16, characteristics of organic fluids typically
include toxicity and flammability limits. The designs employed
by Sundstrand and MTI have taken these limits into considera-
tion, but despite this, one might speculate that leakages may
835
-------
occur that could prove to be hazardous to the health and
safety of workers, and toxic substances may escape.
As mentioned before, Sundstrand uses toluene as a working
fluid which is moderately toxic. It has a National Fire Pro-
tection Association (NFPA) health hazard rating of 2. A
threshold limit value (TLV) of 200 ppm (750 mg/m3) has been
assigned to toluene. The recommended average TLV is 100 ppm
with a peak of 200 ppm for no more than 10 minutes.
The operation of the Sundstrand units at each installation
will be without the need of an operator, and each installation
will have adequate ventilation to guard against excessive
accummulation of toluene leaks intp the atmosphere. Sufficient
fire protection is also included. As mentioned before, toluene
sump tanks are supplied with each system installation. These
tanks are designed to protect the environment from leaks of
the fluid. Since the temperature and pressure of the toluene
in the system is never expected to exceed 465°F and 200 psia
respectively, there is no apparent concern for autoignition.
Toluene decomposes at 750°F.
The MTI unit employs Freon-11 in its bottom cycle. A TLV of
1000 ppm has been assigned to Freon-11 (CFCls). In animal
tests, closely related chemical species such as Freon-112
(CGC12 CFCL2), choloroform (CHCL3), and carbon tetrachloride
(CCl4> have been shown to be carcinogenic. However, no such
conclusion has been drawn regarding Freon-11. In the MTI
design, Freon-11 will be heated to 190°F at 90 psia, thereby
mitigating the possibility of decomposition. Decomposition of
R-ll occurs between 350°F and 400°F.
Freon-11 has been reported to catalyze the breakdown of the
ozone layer, Design conditions will not permit leaks of
Freon-11 during normal operation and barring any unforeseen
failures, it is not expected that Freon-11 leaks will be a
problem. The mere fact that there is significant handling
experience via the refrigeration industry will enhance the
acceptability of the fluid.
SUMMARY AND CONCLUSION
After considering some of the typical environmental effects
pertinent to power generation plants, it can be safe to assume
based upon this somewhat simplified assessment that the
implementation of organic Rankine cycle waste heat recovery
systems in municipal utilities will benefit rather than impair
the environment. The critical area of concern will continue
to center around the organic fluids themselves and the
character of the respective Sundstrand and MTI designs. Each
836
-------
has taken into account the seriousness of catastrophic failures
and has taken the necessary precautions in design to mitigate
their occurrence.
It can therefore, be concluded that implementation of waste
heat recovery devices can, in fact, serve to protect the
environment from adverse influences.
837
-------
Heat Engine and Heat Recovery R&D
Supports MEP and Supply Strategy Policy
CD
OJ
CO
• Enhance Conservation and Lower the Rate of Growth of
Total U.S. Energy Demand
• Shift Industrial and Utility Consumption of Natural Gas and
Oil to Coal and Other Abundant Resources
• Develop Synthetic Substitutes for Oil and Gas
• Reduce Dependence on Oil Imports and Vulnerability to
Interruptions of Foreign Oil Supply
FIGURE 1
78-112B9M/14-34
-------
00
Potential Savings in Oil Imports
'i^^ i.y.:
Category
Total Savings of Oil (MBDOE)
Total Savings of Oil (Quads)
% Reduction in Oil Imports
$/Yr. Savings on Oil Imports*
Estimated Energy Savings
Ongoing Projects
1985
2.7
5.4
23
14.8B
2000
8.7
17.4
76
47.GB
Total Recoverable
Energy
1985
10.4
20.8
GG
5G.9B
2000
12.5
25.0
79
G8.4B
"B.isocl on an cstimntccl vnlun of $15 per linrrcl, winch appears quite consorvntivn for tlte 1985-
2000 time frame.
FIGURE 2
-------
00
£>
o
To Develop Heat-Recovery Technology as
an Alternative Source of Energy by:
• Developing the Necessary Technology
Base for Recovering and Using Waste Heat
• Demonstrating the Technical and Economic
Feasibility of the Technological Components
FIGURE 3
7I-112MM/3-34
-------
Bottoming Cycle Systems for Waste
Heat Recovery and Degeneration
00
Four Unique Concepts in Organic Rankine
Cycle System Technology
• Mechanical Technology Inc. — 500 KW Binary-Rankine
Cycle System
• Sundstrand Energy Systems — 600 KW Toluene
Rankine Cycle System
• Thermo Electron Corporation — 440 KW Ruorinol
Rankine Cycle System
• Biphase Energy Systems — 400-600 KW Two Phase
Heat Engine Cycle System
78-58717-32
FIGURE 4
-------
Rankine Bottoming Cycle Concept
00
4^
to
RANKINE
BOTTOMING
CYCLE
EXPANSION
TURBINE
SOURCE
(WASTE
HEAT)
AAA/
ELECTRICTY
VAPOR
GENERATOR
REGENERATOR CONDENSER
FIGURE 5
78 11289M/18-M
-------
00
^
CO
Organic Vs Steam Comparison
Advantages Of Organics
• High Efficiency With Single-Stage Turbine
e High Efficiency In Small Sizes
• Little Or No Superheating And/Or Desuperheating Required
• Condenser Pressures Near Atmospheric Reduces Leakage Problems
• Compact And Lightweight Turbomachinery
• Low System Cost
• Wide Variety Of Fluids Possible
Disadvantages
• Maximum Temperature Limited By Chemical Stability
e Fluids Can Be Expensive
• Expensive Materials Required To Avoid Decomposition At Elevated
Temperatures
* Lower Heat Transfer Coefficients Require Larger And More Expensive
Heat Exchangers
• Fluids Can Be Toxic, Flammable
• Very Limited Availability Of Off-The-Shelf Hardware Specifically
Designed For Fluids
FIGURE 6 OSTtOOTOt/l-IB
-------
For Same Pinch Temperature Organic Fluids Can
Extract More Heat Than Steam
CO
Steam
Organic
u.
oT
^
13
CL
W
H
600
480F
400
200
0
"""• ~~ _^_ Sen«jjjj Pinch
- ~-.__/e«m 1
\ Steam ' — ~ j^
-
Superheat
If required Preheating
i i i i
X
I N
n- 300F
Pinch
20 40 60 80
Percent of recovered heat
100
20 40 60
Percent of recovered
200F
80 100
heat
FIGURE 7
78-11289M/10-34
-------
600 KW ORC Schematic
200 PSI System
Waste Gas
Out
Toluene Liquid
Condenser Regenerator
Cooling
Tower
00
^
Ul
2.70 PS I A
140° F
Boost Pump
Preheater
Separator
Superheater
.Waste Gas
In
Feed Pump
(Two Stage)
^ 900 Shaft HP (600 KW}
9,300 RPM
FIG
IBS*
e
Seal
-------
00
EXHAUST
GASES
600 KW ORC Bottoming System
COO Li NO TOWtR
SHUTOFF VALVE
CONTROL VALVE
CONDENSER
GENERATOR
OtARBOX
TURilNE ft
FEEDPUMP
START PUMP
BOOST PUMP
VACUUM PUMP
DIVERTER
VALVE
CONTROL
CONSOLE
FIGURE 9
087WOnfl/10-1B
-------
Cycle Schematic For Binary System
00
Stem Loop
81*F
Freoq Loop
FIGURE 10
OSTMOtM/l-H
-------
General Arrangement of MTI Binary Rankine Cycle System
for Waste Heat Recovery/Electric Power Generator
00
*>
00
ipal Power Plant
Rockville Centre, New York
FIGURE 11 (13)
Q87WOHI/1B II
-------
Installation Schematic
600 KW ORC at Beloit, Kansas
Municipal Utility
COOLING
TOWER
CO
PCM
ENGINE
RADIATORS
ENGINE
COOLING
TOWERS
VAPORIZER
CITY WATER
TREATMENT
TOLUENE
SUMP
DIVERTER
VALVE
DIVERTER
VALVE
. 11
^
1 350
1 LSV
Tnmrai? 1 9o
W ^
s
QKW
'-16
s
X"
4100 KW|
LSV-16 I
G
QBTMXMBt/12-U
-------
ORC
COOLING
TOWLK
ENGINE
COOLING
TOWER
00
ui
o
Installation Schematic
600 KW ORC at
Easton, Maryland
Municipal Utility
TOLUENE f—I
SUMP LJ
OPTIONAL
^DIVERTER
VALVE
RV16-4
6600 KW
FIGURE 12b
0*7MMM/11tt
-------
00
Ul
Installation Schematic
600 KW ORC at
Homestead, Florida
Municipal Utility
DIVERTER
VALVE
COOLING
TOWER
OPTIONAL
OIVERTER
VALVE
GOO
KW
ORC
FIGURE 12c
D
TOLUENE
SUMP
PAD FOR
PCM A VAPORZIER
Q87tOOMt/1l-1l
-------
The Impact of Advanced Cogeneration on Emissions
CD
U1
to
.s
o
c
Ul
I
4""
CO
1
ft
0
1
t 5
Ul
•o
0)
*••
1 4
o
O)
a
I 2
11
o
I
1 0
Ul
Range for Conventional Engines
Making Only Electrical Power
\*
Range for Advanced Engine*
in Advanced Cogeneration
Making Both
•Electrical Power
•Process Heat
I
0.2 0.4 0.6 0.8 1.0
Effieieney of Fossil Fuel Utilization
FIGURE 14
-------
Diesel Exhaust Analysis Summary
00
U1
CO
•H*
1.
FmlNo. 2
A
-------
00
Ul
Fluid Study
Bottoming Cycle
Characteristics
Toxlclty
TLV
OSHA/NIO8H Class
Availability
Quantity
Co»t-»/Oel
Material*
Vessel*
Seels
nemmeblllty
Flesh Point
Flra Point
Auto Ignition Point
Product! of Combustion
Vapor Pressure
PSIAOTO'F
Max. Oparatlng Tomp°F
System Efficiency
Haat Sourca Exh.
Tamp *f
Turbina Inlet
Temp "F
Condenser Temp"!6
"») - Elect P°wtr
By* mCphJTM 2401*
Total Heat Ixehenger
Volume- Cu. ft.
2 Methyl
Pyrldlne
6PPM
Toxic
Large
4.M
No copper
mixture
EPR. ECD-006
M°F
146160-F
900°f
Non-tonic
1.59
675-760
248
675
167
.192
200
Fluorine!
85
10PPM
Toxic
Medium
(22,000 lbt/yr|
35.00
EPR
105'F
160"F
900'F
Toxlo
1.0
RBOB7B
248 300
650 S50
181 140
.188 -1S1
166 228
Fluoroiene
M
N.A.
Toxic
Large
25.00
Fluorocarbon
110°
227°F
Toxic
0.28
650700
980
600
180
.161
288
Pentafluorobenzene
Hexafluoro benzene
<20PPM
Toxic
3500 Lb
233.00
, Stainless Steel
Fluorocarbon
None
None
None
Toxic
3.28
900-1000
315
600
169
.178
177
Toluene
(Methyl Benzene!
200PPM
Toxic
Unlimited
2.50
Fluorocarbon
40"F
40°F
1025°F
Non-toxic
0.4588
750-800
300 300 342
650 465 550
160 140 140
188 1621 1638
255 225 271
•••eod en •p*tntl»a4 ttngte stage turbine: 1] Feed Pump-0.6. T) Generator-0.98. •») Qeerbox-0.96
FIGURE 16
087100611/14-16
-------
THERMAL STORAGE FOR INDUSTRIAL PROCESS AND REJECT HEAT
R. A. Duscha and W. J. Masica
NASA Lewis Research Center
Cleveland, Ohio U.S.A.
ABSTRACT
Industrial production uses about 40% of the total energy consumed in the
United States. The major share of this is derived from fossil fuel.
Potential savings of scarce fuel is possible through the use of thermal
energy storage (TES) of reject or process heat for subsequent use. Re-
sults of study contracts awarded by the Department of Energy (DOE) and
managed by the NASA Lewis Research Center have identified three espe-
cially significant industries where high temperature TES appears attrac-
tive - paper and pulp, iron and steel, and cement. Potential annual
fuel savings with large scale implementation of near-term TES systems
for these three industries is nearly 9 x 10^ bbl of oil.
INTRODUCTION
One of the many responsibilities of the Department of Energy (DOE) is
administering the Voluntary Business Energy Conservation Program. This
program, under the guidelines of the 1975 Energy Policy and Conserva-
tion Act, requires major energy consuming firms within industries for
which energy efficiency improvement targets have been set to report
directly to DOE on their energy efficiency. The fact that industrial
production uses about 40% of the total energy consumed in the United
States indicates the tremendous potential that exists for significant
energy savings through a concerted effort by all concerned.
Major energy consuming industries, arranged by the two-digit Standard
Industrial Classification (SIC) Code, were assigned 1980 goals for 1m-
'provement in energy efficiency over their 1972 base level. As of the
first six months of 1977, the index of energy efficiency was at an esti-
mated 9.2 per cent above the 1972 base level Qj. Although very encour-
aging in regards to the overall energy savings implicit in this index,
the decline in the use of natural gas was offset by an increase in the
use of fuel oil.
As with every major problem, the solution for achieving maximum energy
savings lies in many approaches. One approach, known for decades but
relegated to the sidelines because of the past availability of rela-
tively cheap energy in the United States, is the recovery and use of in-
dustrial waste heat. Recognizing the increased importance of waste heat
855
-------
recovery and use, the former Energy Research and Development Administra-
tion (ERDA) funded a study to determine the economic and technical fea-
sibility of thermal energy storage (:TES) in conjunction with waste heat
recovery QQ. This study was directed toward identifying industrial
processes characterized by fluctuating energy availability and/or
demand, a key criterion for TES applicability.
At least 20 industries were identified as areas where thermal energy
storage had potential application to some degree. Responses to a Pro-
gram Research and Development Announcement (PRDA) issued by ERDA shortly
after the conclusion of the feasibility study program resulted in con-
tract awards to study three industries in the high temperature
(>250°C) TES area with potential significant energy savings. These
industries were paper and pulp, iron and steel, and cement. DOE's
Division of Energy Storage Systems awarded the contracts, and the NASA
Lewis Research Center provided the technical management. Major empha-
sis was given to TES systems and applications that have potential for
early commercialization within each specific industry.
PAPER AND PULP
The forest products industry, as a whole, is one of the largest users
of fossil fuels for in-plant process steam generation. Boeing Engineer-
ing and Construction, with team members Weyerhaeuser Corp. and SRI
International, investigated the application of process heat storage and
recovery in the paper and pulp industry \_3j. For this investigation,
Weyerhaeuser's paper and pulp mill at Longview, Washington f4j was se-
lected to assess the potential energy savings and to evaluate the effec-
tiveness of thermal energy storage in achieving these savings.
The paper and pulp operation at Longview consists of process systems
and a power plant which supplies steam to the processes and the power
generation turbines. Figure 1 shows schematically the energy supply
characteristics without energy storage. The recovery (liquor-kraft
black and sulfite from conventional chemical wood pulping) and waste
(hog fuel-wood waste produced by the various machining processes)
boilers provide a base load of steam generation while the oil/gas
boilers provide the time dependent load. The primary goal of using
thermal energy storage at Longview (and similar paper and pulp mills
throughout the industry) is to substitute usage of more hog fuel for
the oil/gas fossil fuels.
The inability to follow rapidly changing steam demands with hog fuel
boilers requires the reduction of hog fuel firing in favor of increased
fossil fuel firing. However, this can be overcome by the use of thermal
energy storage. The hog fuel boiler would be operated at a higher base
856
-------
load, the excess steam would be stored when the demand is low, and stor-
age would be discharged when the demand is high. The economics of steam
swing smoothing in the paper and pulp industry depends on the capacity
of the swing smoothing system and the number of hours per year the sys-
tem will allow hog fuel substitution for fossil fuel.
Daily operational data from the Longview plant was used to evaluate the
effectiveness of thermal energy storage. This plant was considered re-
presentative of paper and pulp mills where the potential exists for the
economic use of thermal energy storage. The analyses using this typical
mill data indicated that for a system as shown on Figure 2, a storage
time of about 0.5 hours with a steaming rate capacity of 100,000 Ib/hr
would result in 60,000 Ib/hr of steam load transfer from fossil fuel
boilers to the hog fuel boiler. This corresponds to about a 50% reduc-
tion in fossil fuel consumption for load following.
Initial sizing and cost estimates for storage system concepts were gen-
erated for a range of steaming rates and storage times. The results in-
dicated that for storage times less than one hour, direct storage of
steam using a variable pressure steam accumulator was more economically
attractive than indirect sensible heat storage using media such as rock/
oil or rock/glycol combinations.
Figure 3 shows the variable pressure accumulator TES concept. Steam
used for charging storage from either the high pressure or intermediate
pressure header bubbles through the saturated water contained under
pressure in the vessel. The steam condenses and transfers energy to the
water, raising the water's temperature and pressure. Upon discharging
to the low pressure header, the steam pressure above the water surface
is reduced causing the water to evaporate, supplying steam but lowering
the water's temperature and pressure.
Oil savings estimated for the Longview plant is 100,000 bbl/yr based on
the transfer of 60,000 Ib/hr of steam load from the fossil fuel boilers
to the hog fuel'boiler. A survey performed using data supplied by the
American Paper Institute indicated that there are 30 candidate mills
that either have now or will have by 1980, operating characteristics
similar to the Longview plant. Therefore, potential near-term (1985)
fossil fuel savings are projected as being 3 x 106 bbl/yr.
Energy resource and environmental impact studies completed by SRI Inter-
national indicates potential long-term (2000) fuel savings of 18 x
106 bbl/yr based on a 10* shift in steam generation from gas and oil
to hog fuel and coal due to TES use. This also takes into account the
additional cogeneration accompanying this shift and the resultant
decrease in purchased electricity. This displacement of gas and oil
will decrease the.-national sulfur dioxide emissions but will result in
an increase in the nation's particulate emissions - roughly two pounds
of S02 removed for each pound of particulate added.
857
-------
Preliminary economic evaluation shows a potential return on investment
(ROI) for this TES system in excess of 30% over a 15-year return and de-
preciation period. The conceptual system using a steam accumulator
appears technically and economically" feasible. Because of the avail-
ability of all the required technology, implementation would not require
technology development or a reduced scale technology validation. In-
stallation at full scale in one of the candidate mills utilizing commer-
cially available equipment could be accomplished within a two-year time
period for a cost of less than one million dollars.
IRON AND STEEL
The primary iron and steel industry accounts for about 11£ of the total
national industrial energy usage. Rocket Research, with team members
Bethlehem Steel Corporation and Seattle City Light, investigated the
use of thermal energy storage with recovery and reuse of reject heat
from ste^l processing in general and electric arc steel plants specifi-
cally HO. Thermal analysis of the complex heat availability patterns
from steel plants indicates significant potentially recoverable energy
at temperatures of 600 to 2800°F.
A detailed assessment for Bethlehem's Seattle scrap metal refining
plant was made of the energy sources, energy end uses, thermal energy
storage systems, and system flow arrangements. This plant is typical of
electric arc furnace installations throughout the United States, allow-
ing results of this site-specific study to be extrapolated to a national
basis.
The hot gas in the primary fume evacuation system from a pair of elec-
tric arc steel remelting furnaces was selected as the best reject energy
source. Presently, the dust laden fume stream is water quenched and
then ducted to the dust collection system prior to discharge to the at-
mosphere. The new flow arrangement shown in Figure 4 would have the un-
quenched fume stream flowing through the energy storage media prior to
discharge. The solid sensible heat storage media would have to be able
to withstand the hot gas temperature which could be as high as 3000°F
while averaging about 1750°F. Potential materials are refractory
brick, slag or scrap steel.
Two energy storage beds are required. The operational storage bed
serves to time average the widely fluctuating temperature of the energy
source. The peaking storage bed serves to hold energy until the demand
arises. During charging, all of the furnace-gas discharge flow goes
through both storage beds and is exhausted through the baghouse, the
dust collection system.
During peak demand periods, the combined streams from the furnace
(through the operational storage) and the peaking storage (in a
858
-------
reversed flow direction) would flow through the heat exchanger to
create steam to drive the turbogenerator. Upon initial discharge of
the peaking store, ambient air is drawn in through the lower fan/valve
arrangement. When the required flow rate through the peaking bed is
established, the ambient air valve is closed. At the exit of the heat
exchanger, gas flow is divided, with a portion going to the baghouse
and the rest providing the peaking storage discharge gas stream.
To complement the assessment, Seattle City Light provided data on
electricity costs. The economic benefits to be derived from the use of
energy storage to provide peak power generation is a direct function of
either a demand charge, time of day pricing, or a combination of both.
The conceptual system proposed for the Bethlehem plant would result in
a payback period of about five years depending on the combination of
electricity costs and size of the power generation equipment. For
example, a system providing a four-hour peak storage capability and gen-
erating 7MW of peak demand electricity would result in a five-year
payback period if it were displacing peak power at a cost of 10
-------
use of thermal energy storage in conjunction with reject heat usage in
the cement industryTej. Thermal performance and economic analyses were
performed on candidate storage systems for four typical cement plants
representing various methods of manufacturing cement. Basically, plants
with long, dry-process kilns and grate-type clinker coolers offer the
best choice for reject energy recovery.
An assessment of potential uses of the recovered energy determined that
the best use for it would be in a waste heat boiler to produce steam
fordriving a turbogenerator to produce electricity for in-process use.
Approximately 75% of a plant's electrical requirements could be met
with on-site power generation. However, this reject heat source for
the steam boiler is not available when the kiln is down for maintenance
of either the clinker cooler grate or the kiln. At this time, the power
demand for other cement plant operations must be obtained from a
utility. This would require demanding large amounts of utility power
for short periods of time, e.g. 5 to 10 MW for 2 to 24 hours. The cost
to the plant in peak power rates and to the utility in maintaining ex-
cess peaking capacity is significant. The other alternative is to cur-
tail other plant operations such as raw or finish milling.
This problem could be alleviated by using thermal energy storage to re-
duce the utility load demand. By charging the storage unit while the
kiln is operating, the stored thermal energy would be available when the
kiln is down. The storage concept proposed in conjunction with dry-
process kilns uses a solid sensible heat storage material such as mag-
nesia brick, granite, limestone, or even cement clinker. The storage
system would use two separate thermal stores as shown on Figure 5. One
would store high temperature (1500°F) reject heat from the kiln exit
gas. The other would store low temperature (450°F) heat from the
clinker cooler excess air. These two separate storages would be charged
independently but discharged in series. Ambient air would be passed
through the low temperature TES units and heated to about 400°F. It
would then be heated to about 1200°F while passing through the high
temperature TES units. The heated air would then flow through the waste
heat boiler and generate steam to produce electricity.
Storage system sizing for typical cement plants indicates that provision
for 24 hours of power production at about 10 MU would be a beneficial
size in relation to normal plant operation. During kiln operation 80-
90% of the kiln exit gas would go directly to the waste heat boiler to
produce electricity while the rest would pass through the high tempera-
ture storage unit. Therefore, it would take roughly one week to charge
the system to its full 24 hour withdrawal capacity.
An economic evaluation of the system indicates that a 10 MW waste heat
boiler/power plant/TES installation would cost about 10 million dollars.
A 90% ROI was calculated for a 30-yr system life and an average energy
cost of 2.8
-------
system. Again, assuming fossil fuel is originally required to produce
this waste-heat derived power, a potential energy savings of about 4 x
10b bbl of oil per year is projected. This is based on utilizing the
cement industry's current installations of about 120 long dry kilns. As
with the steel industry storage/generation systems, this represents a
potential direct reduction of sulfur dioxide emissions.
There is another similarity between the cement plant and steel plant
systems. The necessity for a phased technology development and valida-
tion program through full scale demonstration also exists for the cement
plant system. Estimates of 8 years and 5 to 10 million dollars appear
to be valid for such a program.
SUMMARY
From the response to ERDA's FY 77 Industrial Applications PRDA, three
attractive industries which could utilize high temperature thermal
energy storage were selected for study. These industries are paper and
pulp, iron and steel, and cement which account for 25% of the total
national industrial energy usage. Potential annual fuel savings with
large scale implementation of near-term thermal energy storage systems
for these industries is nearly 9 x 106 bbl of oil. This savings is
due to both direct fuel substitution in the paper and pulp industry and
reduction in electric utility peak fuel use through in-plant production
of electricity from utilization of reject heat in the steel and cement
industries. Economic analyses for all of these systems indicate
potential return on investments from 30 to
90%.
CONCLUDING REMARKS
The results of these three studies appear to be so attractive that the
question immediately arises - "If it looks so good, why aren't the in-
dustries involved already doing it on their own?" Perhaps the answer
to this question can be found in a recent article on energy related cap-
ital investment m. The point being made in this article is primarily
that most companies set the rate of return from energy-saving invest-
ments at a level about twice as high as that for mainstream business in-
vestments. Discretionary investments that do not increase productivity
have a low priority. In addition, paper studies without the visible
proof of a working demonstration do not stimulate the flow of working
capital that is already in limited supply.
The ultimate objective of the effort summarized in this paper 1s the
demonstration of cost-effective thermal energy storage systems capable
of contributing significantly to energy conservation. To achieve this
the Department of Energy's role is that of a catalyst to bring these
861
-------
systems to the point that they will be accepted and widely implemented
throughout the various industries. This effort has shown that a full
scale working system for the paper and pulp industry could be available
in the very near term at moderate co.st. Other systems, although depen-
dent on further technology development and significant capital invest-
ment, appear capable of being implemented within the next eight years.
REFERENCES
1. Voluntary Business Energy Conservation Program, Progress Report No.
6. U.S. Department of Energy (DOE/CS-0018/6), April, 1978.
2. Glenn, D. R.: Technical and Economic Feasibility of Thermal Energy
Storage. General Electric Co., (COO-2558-1), 1976.
3. Carr, J. H.: Application of Thermal Energy Storage to Process Heat
Storage and Recovery in the Paper and Pulp Industry. Boeing
Engineering and Construction (CONS/5082-1), 1978.
4. Nanney, W. M.; and Gustafson, F. C.: Large Bark and Wood Waste-
Fired Boiler - A Case History. Tappi, Journal of the Technical
Association of the Paper and Pulp Industry, pp 94-97, Vol. 60,
No. 8, August, 1977.
5. Katter, L. B.; and Peterson, D. J.: Applications of Thermal Energy
Storage to Process Heat and Waste Heat Recovery in the Iron and
:>teel Industry. .Rocket Research Co. (CONS/5081-1), T978.
6. Jaeger, F. M.; Beshore, D. G.; Miller, F. M.; and Gartner, E. M.:
Applications of Thermal Energy Storage in the Cement Industry.
Martin Marietta Aerospace. (CONS/5084-1), 1978.
7. Hatsopoulos, G. N.; Gyftopoulas, E. F.; Sant, R. W.,; and Widmer,
T. F.: Capital Investment to Save Energy, pp 111-122, Harvard
Business Review, March-April, 1978.
862
-------
Wood
Pulping
liquor
ncovtry
boiten
H'
Ji
Hog
tut!
boilen
hob Mppry up » 40% «f
0,000 tb/hr
ITS ,000 Ib/hr
,000 to > 200,000 tb/hr
IS XXX) Ib/hr
Pulp ft ftp* mil)
•Digmtn
• Evtporvton
• Biueh plarra
• Chiohn* pUnt
/
inn
Figure 1, Paper and Pulp Energy Supply Characteristics
FUcowry
boiten
I!
• Energy nor^gt on rwduo* fuvO fud
byom-tulf
850,000 Ib/hr
far
n
Hog
boilcn
; 435.000 X-
(•60,OOo'VDI
1 Ib/hr) N-
•**
^•*» ,
1
;;! 10,000 to
-v 975,000
1**/*a5-000
•^ Ib/hr
- r J
-
> 100 ,000 Ib/hr
6.000 Ib/hr
(40,000 tb/hr)
Figure 2, - Energy Supply With Thermal Energy Storage
863
-------
CO
(^
*=•
Hog
Fuel
Boiler
High pressure header
Intermediate pressure header
Adjust charge
rate to maintain
intermediate
pressure
Adjust HFB
ma ttain
target inventory
Storage
inventory
Low pressure header
Adjust dis-
charge rate
to maintain
low
pressure
Figure 3. - Variable Pressure Accumulator TES Concept
-------
Figure 4. - Steel Arc Furnace Energy Recovery and Storage System
Oust Clinker
l|. Separator Cooler
\J
ID
Fan
raj
I;1*
Air
Kiln
Waste Heat
Feeder Boiler Separator
C
Rock
Bed
Rock
Storage
Condenser
Ft«tf«rater Pumps
•no Heaters
lurtinc Generator
Rock Hock
Bed Bed
Storage Storage
Figure 5. - Cement Plant Energy Recovery and Storage System
865
-------
PERFORMANCE AND ECONOMICS OF STEAM POWER
SYSTEMS UTILIZING WASTE HEAT
J. P. Davis
Thermo Electron Corporation
Waltham, Massachusetts U.S.A.
ABSTRACT
The performance and economics of steam systems for electric power
generation from waste heat sources are discussed. A simple method
for determining after-tax discounted return-on-investment is presented.
General performance data for steam power systems utilizing waste heat
are shown.
INTRODUCTION
The majority of steam turbine systems utilized in industry for in plant
electric power generation are of the back-pressure type, 600-1200 psig
inlet steam, with or without intermediate pressure extraction as shown
in Figure 1. Where condensing is employed, it is often for the purpose
of affording some degree of flexibility over the power/steam ratio rather
than a desire for substantial continuous power generation from condensing
steam. This approach is usually correct when the steam is being gen-
erated by purchased fuel. The portion of the system which is in the
condensing mode is essentially duplicating what the electric power utility
is doing - and less efficiently than the utility.
However, when the heat source is combustion of waste materials, or lower
temperature waste heat from process operations, or low pressure waste
steam itself, the economics of condensing power are altered dramatically.
Whereas combustion of purchased fuel to generate power exclusively, i.e.
no process steam, is almost never competitive with purchased power; use
of waste energy almost always results in a positive return-on-investment.
Of course, whether or not that positive return is sufficiently high to
warrant the investment is another story.
ECONOMICS
Simple payback, i.e. the ratio of initial investment to pre-tax annual
savings, is often used as a criteria for investment. Paybacks of 3 years
or less are generally considered by industry to be reasonably attractive
and worthy of further consideration. While this rule-of-thumb is a
rough indication of economic desirability, it obviously does not give a
true picture of worth for comparison to various other investment opportunities.
866
-------
However, using this readily calculable parameter of payback based on
first years's pre-tax savings it is possible to calculate equivalent
after-tax discounted return-on-investment for a specific set of assump-
tions. Figure 2 shows the results for the following set of assumptions.
509a tax rate
20 year plant life
IS year sum-of-digits depreciation
10% investment tax credit
6% savings escalation rate
Continuous cash flow model
It is interesting to note that, for this set of assumptions, the after-
tax discounted return is approximately equal to the reciprocal of the
pre_-_tax_ simple payback based on first year's savings.
Maintaining the above assumptions except for investment tax credit and
assumed escalation rate, additional calculations yield the results shown
in Figure 3.
The example shown in Figure 4 \vill show how these results are utilized.
Installed costs for steam power systems vary depending on the plant -
specific situation, particularly power level, waste heat temperature,
and retrofit installation requirements. Typical installed costs for
[a] a 600°F gaseous waste heat source (with waste heat boiler), and (b)
15 psig waste steam (without waste heat boiler) are shown in Figure 5
for condensing non-extraction systems.
For those situations where waste heat can be utilized for both electric
power and process steam, in either of the configurations shown in Figure
1, the economics can be extremely favorable, in some cases yielding
paybacks in the 1-2 year range. Installed costs for such systems are
highly application-specific and cannot readily be generalized.
PiiUFORMANCli
Typical steam rates for a 1500 kWe system condensing at 3" HgA (115°F)
are shown in Figure 6. Performance is improved for higher power systems
and/or lower condensing temperatures, and conversely for lower power
systems and/or higher condensing temperatures.
The range of frame sizes and maximum output capabilites available are
shown in Figure 7. A typical condensing system, fully integrated and
skid mounted, is shown in Figure 8. All systems can be substantially
derated with small loss in efficiency for lower power applications,
although 500 kWe is roughly the lower limit for reasonable turbines.
867
-------
Smaller hack-pressure turbines are available from others down to
outputs of under 100 kWe.
For gaseous waste heat sources, approximate power generation capa-
bilities for condensing non-extraction systems are estimated and shown
in Figure 9. Calculations assume a 3" HgA condensing pressure and a
fixed waste heat exhaust temperature of 350°F from the heat recovery
boiler.
SUMMARY
The utilization of waste heat for electric power generation or com-
bined process steam/electric power often results in attractive after-
tax rates of return, particularly when anticipated escalation of costs
of power and fuels is included in the analysis. In particular,
condensing steam systems not competitive with purchased utility power
when fired with conventional fuels become highly competitive in many
situations. Economics generally dictate the lower limit of power
output for these condensing systems in the range of 500-1000 kWe.
868
-------
BACK-PRESSURE STEAM TURBINE
INLET
(EXTRACTION)
GEN.
PROCESS
PROCESS
CONDENSING STEAM TURBINE
INLET
(EXTRACTION)
PROCESS
GEN.
CONDENSER
BOILER
Pig. 1. Back-Pressure Steam Turbine
869
-------
h-
o
<
(T
00
-J
o
OL
i
U.
O
LJ
a:
_i
<
z
-------
§
a:
LJ
2:
£C
UJ
h-
u.
<
_J
Z
<
UJ
o
U!
U.
U.
UJ
).7
0.6
0.4
z °5
o
I
1ZL
tr
H
UJ
OL
Q
UJ
h-
"2.
O
O
52 0.3
Q
X
0.2
O.I
0.0
• 50 % TAX RATE
20.YEAR PLANT LIFE
±ttttt± '
•15 YEAR SYD; DEPRECIATION
CONTINUOUS CASH FLOW MODE if
ti
ANNUAL SAVINGS ESCALATION
T,RATE -AS SHOWN
INVESTMENT TAX CREDiT
HOWN
6%.ESCALATIONt
130% IT C
4-
±10% ESCALATION/
HlO % ITC
4 L
• !0% ESCALATION,
30%ITC
6% ESCALATION;
410% ITC
o
8
10
SIMPLE PRE-TAX PAYBACK BASED ON INITIAL INVESTMENT
8 INITIAL ANNUAL SAVINGS RATE, YEARS
Fig. 3. Effective Annual After-Tax Discounted Return-On-Investment
vs. Simple Pre-Tax Payback
871
-------
Thermal Source
Initial Investment
Hlcctric Power Savings
First Year Full Power Hours
First Year Local Taxes, Insurance, Main-
tenance, Incremental Operating Labor
Waste Heat
$650/kW
6000 hrs/yr
5% of initial investment/yr
First Year Pre-tax Savings
Payback Based on First Year Savings
= .03 x 6000 - .05 x 650
= 180 - 32.50
= 147.50 $/kW-yr
650
147.50
=4.41 years
Assuming the assumptions shown in Figure 1 apply:
After-tax Discounted Return-on-Investment = 24%
Fig. 4. Sample Analysis
872
-------
00
^J
OJ
1200
1000
•w-
I-
co
o
o
LJ
800
600
cn
2 400
200
T—~T
WITH WASTE HEAT BOILER
WITHOUT WASTE HEAT BOILER
500 1000
1500 2000 2500
POWER LEVEL ( KW)
3000 3500 4000
Fig. 5. Installed Power System Costs
-------
Si
UJ JQ
J- C-
14
12
500 600
TURBINE INLET PRESSURE
Fig. 6. Performance Data for Nominal 1500 kwe Thermo Electron Steam
Power System (TCondenser - 115°F)
874
-------
Thermo Electron Corporation
ENERGY SYSTEMS
MAXIMUM CAPABILITIES OF MULTI-STAGE TURBINE FRAMES
GEARED BACK PRESSURE TURBINES
Frame
No.
9
14(A)
14(8)
3
4B
4A
1B
1A
Size
(inches)
12
18/20
18/20
18
18
18
24
24
Power
(MW)
3
3
5
Speed
(rpm)
12,000
10,000
10,000
5 ; 10,000
7
10
11
15
8,500
8,500
6,600
6,600
Inlet Steam
Conditions
(psig/'F)
900/950
300/750
450/750
900/950
350/650
900/950
350/650
900/950
Exhaust Steam
Conditions
(psig)
60
30
50
350
150
150
50
150
BACK PRESSURE/EXTRACTION TURBINES
Frame
No.
2
Size
(inches)
18
Power
(MW)
8
Speed
(rpm)
8,500
Inlet Steam
Conditions
(psig/°F)
900/950
Exhaust Steam
Conditions
(psig)
100
Extraction
(psig)
250
GEARED CONDENSING TURBINES
Frame
No.
7A (1C)
7B(SC)
15A(IC)
15B(SC)
12(SC/DF)
Size
(inches)
18/22
18/22
18/22
18/22
18
Power
(MW)
2
2
3
3
5
Speed
(rpm)
10,000
10,000
10,000
10,000
10,200
Inlet Steam
Conditions
(psigTF)
500/650
500/650
900/950
900/950
900/950
Vacuum
(in. Hg)
1V2
1'/2
11/2
1Vz
11/2
DIRECT DRIVE CONDENSING TURBINES
Frame
No.
16(SC)
12A(SC/DF)
Size
(inches)
18/22
18
Power
(MW)
3
5
Speed
(rpm)
10,250
10,200
Inlet Steam
Conditions
(psig/°F)
900/950
900/950
Vacuum
(In. Hg)
11/2
1'/2
GEARED CONDENSING/EXTRACTION TURBINES
Frame
No.
13(DF)
17(DF)
17A(DF)
Size
(inches)
18
24/30
24/30
Power
(MW)
5
15
i!>
Speed
(rpm)
10,000
6,600
6,600
Inlet Steam
Conditions
(psig/°F)
900/950
900/950
900/950
Vacuum
(in. Hg)
1VZ
11/2
11/2
Extraction
(pslfl)
100
100
250
1C = Integral Condenser
SC - Separate Condenser
OF = Double Flow last stage
Fig. 7
875
-------
Fig. 8. Packaged Steam Power System
876
-------
OO
Waste Heat
Inlet Temp.
(°F)
1000
800
600
400
Steam
Pressure
(psig)
600
250
125
50
Steam
Temp.
(°F)
900
700
500
298 (sat)
Power
Heat Recovered
(kWh/106 Btu)
75.7
55.3
47.5
39.1
Heat Recovered
Heat Available
0.69
0.61
0.47
0.15
Heat Recovered _ n no /To - 350
~ u. yo
Heat Available u'"u \Jo - 80
(Heat available above 80°F - no condensation)
Fig. 9. Power Generation Capability
-------
A ONE-DIMENSIONAL VARIABLE CROSS-SECTION
MODEL FOR THE SEASONAL THERMOCLINE
S. Sengupta, S. S. Lee and E. Nwadike
University of Miami
Coral Gables, Florida 33124
ABSTRACT
A 1-D model which assumes lateral uniformity is developed
to study the seasonal temperature variations in a lake. The
model includes the effects of variation of horizontal cross-
sectional area with depth. The surface heating due to solar-
radiation absorbed at the surface layer and the internal heating
due to the transmission of the unabsorbed solar radiation to the
deeper layers of the lake are also included. The exchange of
mechanical energy between the lake and the atmosphere is accounted
for through the friction velocity and eddy diffusivity under neu-
tral conditions. The effects of power plant discharge and intake
are also considered.
The equations, describing the above model were solved by
explicit finite-difference methods. The effects of thermal dis-
charges on the turbulent diffusivity and thermocline formation
are studied. The effects of the non-linear behavior of the eddy
diffusivity on the overall stratification are also studied quali-
tatively and quantitatively. Model simulations have been compared
to data acquired to Lake Cayuga. It is demonstrated that the in-
clusion of area change with depth has significant effect on
temperature distributions at mid-depth. All prior models neglect
this parameter.
878
-------
A ONE-DIMENSIONAL VARIABLE CROSS-SECTION
MODEL FOR THE SEASONAL THERMOCLINE
S. Sengupta, S. S. Lee and E. Nwadike
University of Miami
Coral Gables, Florida 33134
INTRODUCTION
In temperate regions most deep bodies of water develop a thermo-
cline during their annual heating cycle. A warmer epilimnion
at the top is isolated from a cooler hypolimnion below by severe
stable thermal gradients. This stratification has a seasonal
cycle, and is an important natural characteristic of a water
«
body. The formation time phasing and the depth and severity
of the thermocline are crucial factors affecting the bio-chemical
processes in an aquatic ecosystem. The nutrient levels, species
spectra and physical characteristics are quite different in the
two distinct domains below and above the thermocline.
r
Convective transport and heat addition caused by power plant dis-
charges result in disturbances in the thermocline. The seasonal
phasing of thermocline formation and decay is affected by thermal
discharge.
The formation of this stratification is caused by non-linear inter-
action between the wind generated turbulence and stable buoyancy
gradients. While being heated from above, a basin forms stable
stratification thereby inhibiting wind generated turbulence. The
thermocline is a region of very stable buoyancy gradients and con-
sequently low turbulence levels. Therefore, turbulent diffusion
through the thermocline is minimal. Further heating merely
879
-------
accentuates the warming of the upper layer and enhances the thermo-
cline gradients. .The temperature of the hypolimnion, therefore,
remains almost constant. With the beginning of the cooling process
in winter, unstable buoyancy gradients in the epilimnion augment
the turbulent mixing caused by wind stress. Thus, the thermocline,
recedes downwards as the epilimnion cools finally resulting in
overturn of the water in the lake. Near homothermal conditions
result. The present paper presents a one-dimensional numerical
model that simulates thermocline behavior and the impact of thermal
discharges.
Previous Thermocline Studies
Numerous attempts in modeling the thermocline have been attempted.
Most of these are one-dimensional time-dependent models. One of
the earliest theories was presented by Munk and Anderson (19^8).
They formulated the vertical transport of heat and momentum as
functions of shear generated turbulence and buoyancy effects.
They proposed functional relationships between eddy transport
co-efficients and Richardson number. They used these co-efficients
in the steady state Ekman spiral formulation. The time-dependent
features of the thermocline could not, therefore, be investigated.
Kraus and Rooth (1961), studied the well-mixed layer above the
thermocline for oceanic problems. They assumed exponential
radiative flux with depth. They concerned themselves with the
steady state energy balance in this layer. The surface tempera-
ture and the depth of the surface layer were numerically predicted
with variations in atmospheric conditions. Some qualitative
880
-------
transient analysis was also presented.
Kraus and Turner (1967), developed a one-dimensional model of
the seasonal thermocline. They accounted for interaction of
stratification and wind-generated turbulence by using the zeroth
and first moments of the one-dimensional, time-dependent conduc-
tion equation and the equation for global conservation of turbulent-
energy. They assumed two well-mixed layers below and above the
thermocline. They assumed that the temperature profiles could
be represented by two parameters, namely the depth of the upper
well-mixed layer and its temperature. Detailed analysis of the
formation and destruction of the temperature profile over the
season could not be adequately studied.
Dake and Harleman (1969), developed a theory for the thermocline
based on exponentially decaying absorption of solar radiation
with depth. Adequate representation of the turbulent transport
and interaction with buoyancy field was not modeled. They
predicted the formation of a thermocline only after the onset
of cooling of the upper layers and consequent static instability
and rapid mixing. In reality the thermocline forms sometime
after the start of heating in spring and before the peak heating
periods of summer.
Sundaram et al (1970, 1971, 1973), in a series of papers have
presented a theory for the formation and sustenance of the seasonal
thermocline. They also investigated the effects of thermal
discharges. They solved the one-dimensional energy equation of
the form:
881
-------
(K S^T
Z 3z
with K = K (1 +a, R. )-
Z ZO 1 1
and R. = a gz29T/3z
IV
where T is the temperature, t is the time, z is the vertical dis-
tance from the surface, K is the vertical eddy transport coeffi-
Z
cient. K is the eddy diffusivity without stratification, a is
the volumetric coefficient of thermal expansion of water, g is the
acceleration due to gravity and W* =/(T , ) the friction velocity
due to surface wind stress T , p is the density and a is an
S _1_
empirical constant. They compared their numerical results with
observations in Lake Cayuga. The agreement was good. The positive
feature of this model is the adequate formulation of the shear-
generated turbulence and buoyancy effects. However, there are
two aspects where improvement is essential. The surface boundary
condition is taken similar to that suggested by Edinger and Geyer
(1967), where the surface heat flux
*s ' Ks '(TE - V
q is the surface heat flux (downwards) K is the surface heat
s s
transfer coefficient, T is the surface temperature and T_ is
s is
the equilibrium temperature, or the temperature of the surface
at which no heat flux occurs. The differential absorption of
solar radiation with depth has been completely ignored.
Moore and Jaluria (1972), have studied the effects of thermal
discharges on the vertical temperature profiles in lakes. They
882
-------
assumed two well-mixed layers with the upper layer having a
linear temperature gradient. This model is not adequate to
study the temporal variation of temperature profiles during the
formation of the thermocline.
More recently Roberts et al (1976), has used a higher order tur-
bulent closure to study the effect of discharges on the oceanic
thermocline. They developed a two dimensional model for an
ocean thermal power plant. They ignored solar radiation absorp-
tion and were primarily concerned with the effect of discharges
on a developed oceanic temperature profile.
Mitry and Ozisik (1976) have developed a two layer model for
the thermocline. They applied their model to Lake Cayuga.
No single model to date includes all the pertinent effects viz.
a) The effects of area change with depth.
b) Nonlinear interaction of wind generated turbulence and
buoyancy.
c) Absorption of radiative heat flux, below the surface.
d) Thermal discharges.
e) Effect of vertical convection caused by discharge.
The model presented in this paper includes all these effects.
Model Formulation
The basic balance equations of mass and heat are:
^ = -V .PV (!)
3t
3 (pC T) = V -pC K «V T - V -pC T V + H (2)
Tt P P P
883
-------
where
P is the density
t time
J velocity of flow
C the heat capacity
T the temperature
K heat diffusivity tensor (including turbulent
diffusivity)
H source of heat per unit volume.
There are at least two reasons for the existence of horizontal
divergence in real lakes.
a) The variation of horizontal cross-sectional area of the
lake with depth.
b) The existence of sources of heat and matter efflux at
depths above the deepest point.
The need to include these in the diffusion equations of lakes
was already felt by Lerman and Stiller (1969)» Button and Bryson
(1962) and Tzur (1973). Only Tzur (1973) formulated a corrected
diffusion equation. The effects of area change with depth are
included by the following treatments of equations (.1) and (2),
Integrating equation (1), over the volume of water below height h.
measured from the deepest point in the lake;
/ 8_p_ dV = - y V • pVd¥
V 3t
Using Gauss theorem on left hand side;
/3£ d¥ = -p /ft - VdS
at
884
-------
where S is a surface completely sorroundlng the volume V, hence
dS=dc + dA
where c Is the surface area of the part of the bottom of the lake
that is bounded by the contour at height z. As s subscript it
marks a variable at the contour.
Using, d¥=Adz in equation (3),
/h 3p Adz = - p/ n • Vdc -p{ n ,• VdA
0 9t A
or
h h
/ 8_p_ Adz = - pV A(h) - /n pV dc (4)
0 8t Z on
Integrating equation (2), over the volume of water below height h,
measured from the deepest point of the lake,
' 9_ pC Td¥ = * (V -pC K -V T)d¥ - * V • pC TVd¥
¥ p ¥ p ¥ p
Applying the divergence theorem to the first two terms on the
right ,
/h A(z) 9 pC Tdz = / ft »(pC K • VT)dS - /(n • pC TV)dS
0 9t p S P S P
A(z)H(z)dz
o
Using dS=dA + dc
fhA(z)9 pC Tdz = /h (pC K -VT) dA + /h (pC K • VT) dc
o -JTZ- p op z op n
'n (pC TV) dA - /"(pC TV) dc
o p z o ^ p n
885
-------
i.e.
/hA(z)a PC Tdz = PC A(h) (K • W + /hp C (K • VT) dc
o-g^-p p zocpc n
- PC A(h)TV - /h p C T V dc + /hA(z)H(z)dz (5)
p z ocpcn o
C
where
z Is the vertical coordinate, measured upward from the
deepest point of the lake. As a subscript it marks the
vertical component of a vector.
n subscript, marks the component of a vector that is
perpendicular to the lake-bottom, positive outwards.
A(z) is the horizontal cross-section of the lake at height z.
Differentiating equations (4) and (5) with respect to the height,
a set of 1-D equations are obtained,
A9p _ 3 Ap V - p V 9c
~ "~ Z c n
or A9p = - a ApV + IA' , ^
It ^ z (6)
where I = the bottom-surface source of mass per unit area.
and A' = dA -n dc,
dz dz
Where n=l, is the average of the cos-arc-tan(gradient ) of the
bottom surface at that depth
i.e. A' = dc
dz
From equation (5),and noting that A=A(z) and H=H(z);
pc ° n dz
AH (7)
886
-------
The terms in the square brackets are the heat addition terms.
Because the horizontal gradients vanish, equation (7) can be
simplified further by noting that:
(K • VT) = K 9_T
A9_ (pC T) = 3_ (pC AK 9T) - 3_ (pC ATV ) + QA' + AH (8)
9t p 3z p z 9z 9z P z
where, Q = the bottom-surface source of heat per unit area.
Equations (6) and (8) are the equations to be solved. Before
attempting to solve these equations numerically, the relevant
terms and parameters will be discussed.
The numerical values represent the Lake Cayuga application pre-
sented in subsequent sections.
1. Density, P is assumed to vary with temperature in the form
P = A' + B'T + C'T2
where A' = Density at 0°C
= 1.02943 gm/cc
B1 = -0.00002
CT = -0.0000048
This is the form given by Sengupta and Lick (1974).
2. Eddy diffusivity, K is a function of both thermal and
Z
current structure of a lake. The form used in this study
was deduced by Rossby and Montgomery (1935).
Kz = Kzo(1 + a!Ri) (9)
Where R. is the Richardson number which characterizes the
interaction between the mechanically generated turbulence and
the thermal structure is defined as
887
-------
w*
a = •!, is an empirical constant, estimated in this study
by comparing the values used by Sundaram et al (1971), the
original value of Monin and Obukhov (195*0.
3. a is the volumetric coefficient of expansion of water and
varies as shown below
ay = AI + B1(T-4) + C1(T-4)2 (11)
where A, = 0 is a at 4°C
B = 1.538 x 10~3
C± = -2.037 x 10~7
4. W* is the friction velocity given by the surface shear stress,
T , induced by the wind, and the density
s
w* - ^W (12)
An empirical form of W* which has been widely used is also used
in this study;
W* = A0 + B Sin ( 2u t + C ) (13)
d 355 2
where A2 = Average value of W*
= 3.048 cm/sec
Bp = Half annual variation of W*
= 0.762 cm/sec2
C = Phase angle (chosen in such a way that at time,
t=0, W*=initial value of the friction velocity).
= 2.61
888
-------
5. It has also been assumed that the eddy diffusivity under
neutral conditions, K varies as
zo
K = A + B_ Sin ( 2u t + C_) (14)
zo 3 3 3
where A = Average value of K =0.21 cm2/sec
j zo
BQ = Half annual variation of K = 0.052 cm2/sec
j zo
C., = Phase angle (chosen such that at t=0, K =initial
3 SO
value of K )
zo
= 2.61
6. The heat source, H, is that part of the solar radiation
transmitted exponentially through the depths of the lake.
(In this study H, is not included in the equilibrium tempera-
ture estimation since H, is not absorbed at the surface).
H = n(l-B)A(z)* exP(- n(Z-h) (15)
B = 0.5, is the fraction of the solar radiation absorbed
at the surface
n = 0.75, is the solar radiation absorption coefficient
4> , is the net solar radiation reaching the water surface.
An empirical relation has been used in this study to describe
j> = A,. + B,, Sin (2ir t + C,.) (16)
O 4 4
where Aj. = Average value of
= 6.14 x 10~3 cal/cm2S
B. = Half the annual variation of <|>
= 3.52 x 10~3 cal/cm2S
889-
o
-------
C^ = Phase angle (chosen in the same way as C~ or C^)
= 0.049
7. The discharge from the power plant is included in two ways.
(i) The heat flux Q in equation (8) is defined as
Q = (PC ATQ )/A(z) (17)
Where Q is the volumetric discharge from the power plant.
Q -j
In this study'Q = 1.508 x 10 cm /sec, this value is chosen
to correspon to Sundaram et al pumping velocity of % ft/day.
AT=10°C is the assumed temperature change through the conden-
sers of the power plant.
(ii) The pumping velocity term is O C A(z)TV }
p z
where V = Qp (18)
2 A(z)
The pumping velocity term effects are only felt between the
intake and the level at which the heated effluent becomes
neutrally buoyant (effective discharge level).
Numerical Integration, of the Governing Equations
A forward time - Dufort Frankel scheme is used to solve the gov-
erning equations. The solution of equations (6) and (8) requires
one initial condition and two boundary conditions.
The temperature of the lake at spring homothermy is taken as the
initial temperature. For Cayuga Lake the spring homothermy occurs
around March and the temperature at spring homothermy is 2.9 C.
(i) The first boundary condition is
|
z = h
• K*!ii „ " VW (19)
890
-------
The equilibrium temperature, T£> surface heat exchange coef-
ficient, KS are both functions of wind speed, air temperature
and humidity, and net incoming (sky and solar) radiation.
Methods of evaluating T and K are fully described by
ili S
Edinger and Geyer (196?) and Sundaran et al. In this study
T~ is defined as
h
T = A + B x Sin( 2TT t + Cc) (20)
*- 5 5 1^5 ' 5
where the constants Ap-, B and C,_ are chosen in such a way
that at spring homothermy T = initial temperature, Ac=ll°C,
ij D
B =16°C and C =0.531.
o ?
(ii) The second boundary condition is at the bottom of the
lake which is assumed to be perfectly insulated,
3T
=0 (21)
z=0
During the heating portion of the annual cycle once the thermo-
\
cline is formed, the values of K below the thermocline do not
z
represent a correct thermal diffusivity since nonlinear effects
are now dominant. A cut-off procedure is used to eliminate this
problem. After thermocline formation (defined by the condition
that _3T reaches a minimum value which is not at the surface of
9z
the lake z=h), the minimum value of KZ is determined. This value
is denoted by K and the position is z . . Similarly during
•* z . mm
mm
cooling, convective mixing becomes important within the epilimnion,
again another cut-off procedure is used, the local maximum K
max
and z are calculated.
max
Thus the following limits on diffusivity are used
891
-------
K = K for all z < z .
z z — mm
heating
K = K for all z > z .
zmin ~ min
K = K for all z > z
z z — max
K = K for all z < z
z z — max
max
cooling
The conditions applicable to Lake Cayuga are taken from Sundaram
and Rehm (1973) .
The depth of the lake (Cayuga) = 200 ft (60.96 m. )
K = 180 Btu/ft2 day °C (5.65 x IQ~^ cal/cm2 - S - °C)
s
= 6.14 x 10~3 + 3.52 x 10~3 x Sin( 2 t - 0 . 049)cal/cm2S .
TE = 11 + 16 _
Initial Temp. = 2.9°C
For a postulated 3500 mW plant for Cayuga Lake a 8.79 x
-4 2
10 cal/cm S of waste heat will have to be rejected.
AT = 10°C
The intake to the power plant is fixed at 125 ft (38.1 m) from the
surface of the lake.
Two topographies were considered for Cayuga Lake:
1. Cylindrical Topography
The area of the lake is constant throughout.
The term A' or dA = 0 (see equations (1) and (2)).
dz
2. Circular Paraboloid Topography
The lake is assumed to be a circular paraboloid, with surface
radius, 6=7.38 x 10 cm. (Surface area of Cayuga Lake 66 sq.
mis.) The area at any depth z (measured from the deepest point
892
-------
of the lake) is given by
, x
(22)
A = TiB2z
where h is the depth of the lake.
Thus A' becomes a constant:
A' = TiB2 (23)
h
Results
Computations for a yearly cycle for Lake Cayuga are presented.
The verification data base consists of vertical temperature pro-
files compiled by Henson et al (1961). The comparison of simu-
lated and observed vertical temperature profiles are shown in
Figs. 1, 2, and 3- Each figure shows five profiles representing
observed, and the four cases of discharge, no-discharge, parabolic
and cylindrical domains. The no-discharge simulations are in
good agreement with the data. (The data was for no-discharge
condition). The parabolic case has somewhat better agreement since
It represents qualitatively, the decrease in area with depth.
However, the closeness of the simulated results for the two cases
Is surprising. Most lakes have the rate of decrease in area with
depth greater than a paraboloid, which has a linear decrease.
Thus, when realistic area changes are used a greater difference
between cylindrical and paraboloid cases can be expected.
The discharge from the power plant is treated as a plane source
and is injected into the lake at the level where the discharge
temperature equals the local level temperature. The effects of
the pumping velocity term are applied from this level to the intake
893
-------
level (also considered as a plane sink) which for this study
is 125 ft. (38.1 m) from the surface of the lake.
A pumping velocity V of( 1.62 x Id5) cm/sec
z A
z
was assumed corresponding to the value of \ ft/day assumed by
Sundaram et al in one of their calculations for Cayuga Lake.
A temperature rise of 10°C through the condensers was also assumed
between the intake and discharge levels, a situation which calls
for the use of density as a function of temperature.
The effect of discharge is significant only in the top layers
until July. This is because the heated discharge rises to the
surface. For the later months the discharge temperature is lower
than the surface temperature causing the discharge to reach static
equilibrium somewhere below the surface. Thus significant thermal
effects of discharge are seen at mid-depths until December. The
temperatures were higher at these depths for the paraboloid topo-
graphy. In general, a temperature difference of the order of 3 C
over no-discharge case, can be seen. At the end of the annual
cycle a residual temperature increase of 1.75 C is detected.
Figs. 4 and 5 show the annual stratification cycle. It is evident
that the surface temperature difference between the four cases is
less than 2 C over the yearly cycle. However, at mid-depth the
paraboloid discharge case shows a 5 C difference compared to no-
discharge case. The cylindrical-discharge case at mid-depth
shows a 3 C difference from observed no-discharge data and
simulation. The highest surface temperatures are reached after
150 days. It is noted that the highest equilibrium temperature
894
-------
occurs after 120 days. Thus there Is approximately a 30 day lag
in surface temperature response. The maximum temperatures at
mid-depth occur after 240 days for the no-discharge case. For
the discharge case maximum temperatures at mid-depth occur after
210 days. No significant phase lag between cylindrical and
paraboloid cases are observed.
Figs. 6 and 7 show the eddy diffusivity variation with depth and
time for cylindrical and paraboloid cases. It is observed that
thermal discharge causes increase in eddy diffusivity in the
epilimnion owing to increased mixing. No significant changes
are seen in the hypolimnion. The difference between discharge
and no-discharge cases increase with time. The effect of dis-
charge is also seen as an increase in epilimnetic depth or low-
ering of the thermocline. These observations apply to both the
cylindrical and paraboloid cases. Comparison of cylindrical
and paraboloid cases indicate that the diffusivity values are
larger for the paraboloid cases. Also at any given time the para-
boloid case shows deeper thermoclines.
Conclusions
A one-dimensional model which includes area-change with depth,
vertical convection, varying diffusivity, thermal discharges,
and internal absorption of radiation has been formulated. Its
application to Lake Cayuga indicates excellent performance. A
comparison of cylindrical and paraboloid cases indicate that
significant differences in thermocline depth, eddy-diffusivity,
and temperature at mid-depths are observed. This indicates
895
-------
the effects of area change with depth are not negligible. These-
-effects will be more pronounced in real basins where decrease
in area with depth is more severe than the linear variation for
the paraboloid case.
Acknowledgement s
This work was conducted under funding from National Aeronautic
and Space Administration, Kennedy Space Center and Environmental
Protection Agency.
896
-------
Nomenclature,
z Vertical coordinate measured upward from deepest point of
the lake. As a subscript it marks the vertical component
of a vector.
h Depth of lake
A(z) Horizontal cross-sectional area at height Z.
I(z) Bottom-surface source of mass per unit area.
Q(z) Hot torn-' surf ace source of heat per unit area,
T Temperature (°C)
p Density of water
V Vertical velocity
z
K Eddy diffusivity
z
K Eddy diffusivity under neutral condition
zo
*
, ) Friction Velocity
s/ p
o Empirical constant
R. Richardson number
a Volumetric coefficient of expansion of water
T Surface shear stress
s
Cp Heat capacity
H(z) Heat source/unit vol.
A1 Density at 0°C
BJC' Density variation constants
A1 Volumetric coefficient of expansion at 4 C
B ,C Volumetric Coefficient of expansion variation constants
#
A~ Average value of W
*
B Half of the annual variation W
897
-------
C2>C3>C4>G5 Phase angles
A- Average value of K
zo
B0 Half the annual variation of K
3 zo
<}> Solar radiation incident on the water surface
A^ Average value of 4>
B^ Half the annual variation of $
n Extinction coefficient (equation 10)
6 Absorption coefficient (equation 10)
Qp Volumetric discharge (equation 12)
AT Condenser Temperature change (equation 12)
Tn Discharge temperature
q Surface heat flux
s ,
K Surface heat exchange coefficient
S
T_ Equilibrium temperature
iii
A(- Average value of T
B.- Half annual variation of T^
t> £•
T Surface temperature
S
q_ Bottom surface heat flux
D
B Lake surface radius
dA_ Area variation with depth,
dz
898
-------
61
Depth
(m)
M
L
PN
PN PD
61
PD
(March)
Deptl]
(m)
(April)
61
Depth
(m)
8 12 16 20
.__JTejnp.°C
PN CN
.M
8 12 16 20
Temp.°C
(May)
Depth
(m)
8 12 16 20
Temp. C
8 12 16
Temp.°C
Fig,l
Vertical Temperatrue Profile (0 to 90 Days)
(PN=Paraboloid, No-Discharge; CN=Cylindrical
No-Discharge; PD=Paraboloid + Discharge;
CD=Cynndrical + Discharge; M=Meat>ured) 899
-------
61
Depth
(m)
61
Depth
(m)
Temp.C
61-
Depth
(m)
CD
(September)
Depth
(m)
i i i i
0 4 8 12 16 20 24
Temp.°C
(August)
0 4 8 12 16 20
Temp.°C
M.
0 4 8 12 16 20
Temp.°C
Fig, 2 Vertical Temperature Profiles (from 120
to 210 days)
900
-------
bl
Depth
(m)
M
8 12 16 20
Temp. C
Depth
(m)
(January)
Depth
(m)
8 12 16 20
Temp.°C
PN ,PD
61
M'
Depth
(m)
k_
-
>•
^ CD
'^N
(February)
8 12 16 20
Temp. C •
048 12 16 20
Temp. C
Fig.3 Vertical Temperature Profile (from 240
to 330 days)
901
-------
o,
28
25
20
15
10
5 .
-5 u
30 60 90 120 150 180 210 240
TE
M
CN
CD
CNM
COM
Equilibrium temperature
SURFACE TEMPERATURES
Measured
Cylindricalj no discharge
Cylindrical + discharge
MIDLAYER TEMPERATURES
Measured
No discharge
Discharge
Fig.4 Stratification Cycle (Cylindrical Domain)
902
-------
28 r-
25
20
15
10
-5
30 60 90 120 150 180 210 2*40 \?6 300 330/J60
Days
TE Equilibrium temperature
SURFACE TEMPERATURES
M Measured
PN No discharge
PD Discharge
MIDLAYER TEMPERATURES
MM Measured
PNM No discharge
PDM Discharge
Fig.5 Stratification Cycle (Parabolic Domain)
903
-------
cmVsec
30
27
24
21
18
15
12
9
6
3
r
No Discharge
With Discharge
300 days
240 days
90 days
"*~
6 12 18 24 30 36 42 48 56 60
Depth(Meters)
Fig.6 Variation of Eddy Diffusivity with Depth
(Cylindrical Domain)
904
-------
No Discharge
With Discharge
crnVsec
36
33
30
27
24
21
18
15
12
9
6
r
300 days
240 days
days
0 6 12 18 24 30 36 42 48 56 60
Depth (.meters)
Fig,7 Variation of Eddy Diffusiyity with Depth
(Parabolic Domain)
905
-------
REFERENCES
Dake, J. M. K. , and D. R. F. Harleman, An Analytical and Experi-
mental Investigation of Thermal Stratification in Lakes and' Ponds,
MIT Hydrodynamics Lab. Tech. Kept. 99, Cambridge, Massachusetts,
September 1966.
Dake, J. M. K. and Harleman, D. R. F., "Thermal Stratification
in Lakes: Analytical and Laboratory Studies," Water Resources
Research Vol. 5, No. 2, April 1969, PP 484-495-
Button, J. A., and Bryson, R. A. 1962, Heat Flux in Lake Mendota,
Limnol Oceanog. 7,80.
Edinger, J. E. and Geyer, J. C., "Heat Exchange in the Environ-
ment." Sanitary Eng. and Water Resources.Report, 196?•
Henson, E. B., Bradshaw, A. S, and Chandler, D. C., "The Physical
Limnology of Cayuga Lake, New York," Memoir 378, 1961, Agricultural
Experimental Station, Cornell University, Ithaca, New York.
Kraus, E. B., and Rooth, C,, 196l, Temperature and Steady State
Vertical Heat Flux in the Ocean Surface Layers. Tellus, 13,
pp. 231-238.
Kraus, E. B. and Turner, J. S,, "A One-Dimensional Model for the
Seasonal Thermocline II. The General Theory and Its Conse-
quences," Tellus, Vol. 19, No. 1, 196?, pp 98-105.
Lerman, A. and Stiller, M. 1969 Vertical Eddy Diffusivity in Lake
Tiberias Verh. Internat. Verein. Limnol. 17, 323.
Mitry, A. M. and Ozisik, M. N., A One-Dimensional Model for
Seasonal Variation of Temperature Distribution in Stratified
Lakes, International J. Heat Mass Transfer Vol. 19, pp. 201-205,
1976.
Monin, A. S., Obukhov, A. M., Basic Regularity in Turbulent
Mixing in the Surface Layer of the Atmosphere, USSR Acad. Sci.
Works of Geophys. Met. No. 24, 163 (1954).
Moore, F. K. & Jaluria, Y, 1972, Thermal Effects of Power Pla
on Lakes Journal of Heat Transfer, Transactions of the ASME,
pp. 163-8,
Munk, W. H, and Anderson, E, R., "Notes on the Theory of the
Thermocline, "Journal of Marine Research 1, Vol. 7, No, 3,
March 1948, pp 276-295.
Roberts, G. 0,, Piacsek and Toome, J,; Two Dimensional Numerical
Model of the Near-Field Flow for an Ocean Thermal Power Plant.
Part I, The Theoretical Approach and a Laboratory Simulation,
906
-------
NRL-GFO/OTEC.5/76, Naval Research Laboratory 1976,
Rossby, C. C.; and Montgomery, B. R., "The Layer of Frictional
Influence In Wind and Ocean Currents, Papers in Physical Oceano-
graphy, Vol. 3, No. 3, 1935, p. 101,
Sengupta, S. and Lick, W., A Numerical Model for Wind-Driven
Circulation and Heat Transfer in Lakes and Ponds. FTAS/TR-74-98.
Sundaram, T. R., Rehm, R. G., Rudinger, G., and Merritt, G. E.,
"A Study of Some Problems on the Physical Aspects of
Thermal Pollution," VTV2790"A«1, 1970, Cornell Aeronautical
Laboratory, Buffalo, New York,
Sundaram, T, R,, and Rehm, R, G,, Formulation and Maintenance of
Thermoclines in Stratified Lakes Including the Effects of Power
Plant Thermal Discharges, AIAA Paper No. 70-238, 1970.
Sundaram, T, R, Easterbrook, C, C,: Piech, K, R, and Rudinger,
G., "An Investigation of the Physical Effects of Thermal Dis-
charges into Cayuga Lake," Report VT-26l6-0-2, Cornell Aeronauti-
cal Laboratory, Buffalo, N.Y. (Nov. 1969).
Tzur, Y. "One-Dimensional Diffusion Equations for the Vertical
Transport in An Oscillating Stratified Lake of Varying Cross-
Section, Tellus XXV, 1973,
907
-------
ABbTKACT
Hydrothermal Structure of Cooling Impoundments*
by Gerhard H. Jirka :
School of Civil and Environmental Engineering
Cornell University
Ithaca, NY 14853
Cooling impoundments, such as on-stream reservoirs and off-stream perched
cooling ponds, can exhibit a highly complex and variable temperature and
circulation structure. The understanding of this structure and its dependence
on governing parameters is of crucial importance for the formulation and appli-
cation of predictive mathematical models for cooling pond design and impact
prediction. Proceeding from an analysis of two-layer stratified flow with
variable density, a characteristic "pond number" is defined which accounts for
the effects of pond shape, depth, condenser flow rate and temperature rise,
entrance mixing, and internal friction. Use of the "pond number" allows to
distinguish cooling ponds into the vertically well stratified, partially mixed
and vertically fully mixed type. The partially mixed and fully mixed types
can be further classified in terms of their internal circulation pattern as a
recirculating pond or a dispersive pond. Comparisons with available field
and laboratory data are given. The application of mathematical models to these
pond types is discussed.
* A revised version of this paper has been submitted under the title
"Thermal Structure of Cooling Ponds" by G.H. Jirka and M. Watanabe
for publication in the Journal of the Hydraulic Division, American
Society of Civil Engineers.
908
-------
HYDROTHERMAL PERFORMANCE OF SHALLOW COOLING PONDS
E. Adams
A. Koussis
M. Watanabe
G. Jirka
D. Harleman
R.M. Parsons Laboratory for Water Resources and Hydrodynamics
Massachusetts Institute of Technology
Cambridge, MA 02139, USA
ABSTRACT
The hydrothermal performance of shallow-dispersive cooling ponds is analy-
sed for the purpose of facilitating pond design. In the first part of the
paper, plant performance is simulated with a transient mathematical model
for a variety of pond configurations including variation of surface
area, depth, length-to-width ratio, condenser flow rate and temperature
rise. In the second part, a quasi-steady model is developed and compared
with the results of the transient simulation. Together with pertinent
cost information, these models should be useful in establishing trade-offs
among the various parameters which characterize pond design.
INTRODUCTION
Cooling ponds are large, artificially constructed bodies of water used for
closed cycle cooling of steam power plants. In regions where land use per-
mits, ponds offer a number of advantages over other forms of closed cycle
cooling (e.g. mechanical or natural draft evaporative towers) including
lower operation and maintenance cost and higher thermal inertia.
One disadvantage, however, is the relative difficulty in predicting pond
performance. Unlike wet towers, ponds respond to a complex combination of
meteorological parameters, and because of their heat capacity, their re-
sponse is transient with a time constant on the order of days rather than
minutes as is the case with towers. This thermal inertia helps filter out
peak temperatures caused by fluctuating meteorology and plant operation but
requires that some sort of transient analysis be adopted to obtain the
correct response. Further difficulty lies in the complex circulations,
both lateral and vertical, which may result from discharge momentum, buoy-
ancy or surface shear stress from wind.
In order to learn more about pond behavior, an effort has been made to
classify ponds with respect to their basic hydrodynamic circulation.
Jirka [1] for example, has described a classification scheme based
on relative depth of the pond and the extent of horizontal circulation.
909
-------
It was found that the relative depth of a cooling pond is dependent on the
pond number
where L, W and H are the pond length, width and depth, Q and T are the
condenser flow rate and temperature rise, D is the dilution produced by
entrance mixing, f. is an interfacial friction factor, g is a coefficient
of thermal expansion and g is acceleration of gravity. Ponds for which
IP <0.3 are classified as deep and exhibit a definite two layer structure
with a warm surface layer and a cooler, horizontally uniform, lower layer.
Ponds for which IP > 0.3 are classified as shallow and do not possess a
distinct surface layer. For 0.31.0 only horizontal temperature gradients are present. The
tendency for horizontal circulation depends on the relative pond depth and
its length to width ratio (computed along the flow path). In shallow ponds
for which L/W < 3 to 5 horizontal circulation takes place in the form of
large eddies while for L/W > 3 to 5 the flow is essentially one- dimensional,
and dispersive in character. For deep ponds, density-induced spreading
promotes utilization of the entire pond area, thereby decreasing the
tendency for horizontal eddies.
Of the three classes of ponds - shallow-dispersive, shallow-recirculating,
and deep-stratified - it appears that the shallow-dispersive pond offers
a number of general advantages 'for artificially constructed ponds. These
advantages include avoidance of short-circuiting associated with adverse
wind conditions, reduction of destructive entrance mixing due to the
absence of horizontal or vertical circulation, and relatively low diking
costs due to shallow depth.
The object of this paper has been to study the performance of shallow-dis-
persive ponds with the airm of facilitating" pond design. This effort has
included two parts. In the first, a transient mathematical model has been
used to simulate the annual performance of various pond configurations.
The results, in terms of plant intake temperature and associated power pro-
duction can be used to help evaluate the cost effectiveness of various pond
designs (pond area, baffle density, etc). In the second part a quasi-
steady model has been developed and compared with the transient analysis.
Because of its simplicity, it can serve as a design tool in the preliminary
screening of cooling pond designs.
TRANSIENT SIMULATION
A transient, mathematical model for shallow-dispersive cooling ponds has
been developed by Watanabe and Jirka [2]; the essential features are
indicated in Figure 1. The pond is characterized by the variables L, W, H,
Q and AT . The jet entrance mixing region is a small fraction of
the total pond area with the major throughflow portion of the pond being
characterized by a longitudinal dispersion process.' Temperatures within
the pond are governed by a one-dimensional bulk diffusion equation with
cross-sectionally averaged variables;
910
-------
+ n = F
8t 3x L Ib? " pcH (2)
where T is the cross-sectional mean temperature, U is the cross-sectional
mean velocity QQ/WH, x is longitudinal distance, t is time, E is longi-
tudinal dispersion coefficient, is net heat flux across the surface and
pc is the heat capacity of water per unit volume. E is based on Fischer
[3] and is given by
(3)
where K is von Karman's constant (0.4) and f is a bottom friction factor.
The surface heat transfer includes short and long wave net radiation,
evaporation, conduction, and back radiation and is given by Ryan and
Harleman [4]. Boundary conditions are specified at either end of the
pond to ensure conservation of thermal energy, and the equation is
solved with an implicit numerical scheme. A comparison of predicted and
observed temperatures at the Dresden cooling pond (Watanabe and Jirka, [2])
indicated good agreement.
This model was used to simulate pond performance for a generic-shaped pond
under variation of a number of its parameters. For a base case, a pond
of area A=1000 acres, length to width aspect ratio L/W=12, and depth H=9
feet was used for a plant with condenser flow rate Q =1800 cfs and tempera-
ture rise AT =20° F (heat rejection J =8.09 * 10 Btu/hr, corresponding
roughly to a 1200 MWe nuclear power plant). 'Note that these parameters
result in relatively high loading (approximately 1.2 nuclear MWe per acre)
in order to highlight the sensitivities. For sensitivity, values of
A=750, 1500, 2000 and 3000 acres, L/W=36, H=6 and 12 feet and AT =10
and 30°F (Q =3600 and 1200 cfs) were considered. In addition, an extra
test with L/w=3 was performed with a shallow recirculating pond model
assuming that D=2 for entrance mixing. Only one variable was changed for
each run and J was assumed constant at 8.09 x 10^ Btu/hr. The pertinent
information for each run is summarized in Table 1. Also tabulated are the
dimensionless dispersion coefficient (see below) and the pond number IP.
Values of the latter indicate that, while some vertical temperature
gradient would exist, each pond is well within the shallow category.
Simulations were performed for the year 1970 using three-hour 'time steps
with three-hour meteorological data (air temperature, wind speed, relative
humidity and cloud cover) obtained from the NWS station at Moline, Illinois.
The 2920 values of plant intake temperature (pond outlet temperature) for
each run were compiled into cumulative distribution functions for the year
as shown in Figures 2a-d. Associated with eac.. temperature is the gross
poxrer, which could have been produced from a conventional 1200 MWe nuclear
turbine-generator. The cumulative distribution function for power produc-
tion shown in Figures 3a-d allows easier comparison of the cost effective-
ness of various pond designs. In order to illustrate more of the short
range performance of the ponds, the mean and standard deviations of the
911
-------
intake temperatures were computed for the month of July and are listed
in Table 1. Since the governing meteorology was more or less stationary
over this period, this table allows one to identify differences in gross
efficiency (given by variation in the mean temperature) and thermal inertia
(given by variation in the standard deviation) .
It is clear from the figures and table, as well as from a steady state
analysis, that surface area and condenser flow rate (temperature rise) are
the primary variables affecting pond performance. Increasing pond area
or flow rate both result in higher plant efficiencies at the expense of
greater land purchase and preparation costs on the one hand, and greater
pumping and condenser costs on the other. The effect of baffling (aspect
ratio) and pond depth show secondary effects. The aspect ratio influences
performance through its effect. on the dispersion coefficient. (The non-
dimensional coefficient E *=£, /UL decreases with the 3/2 power of L/W) .
Thus as E * decreases (L/w increases) plant intake temperatures decrease
towards trie ideal limit of plug flow, while as E* increases (L/W decreases)
fully mixed conditions are approached. In addition, by comparing the
results for L/W=12 and 36 with those for L/W=3, it is clear that the
achievement of one-dimensional flow (suppression of horizontal eddying)
results in a significant decrease in intake temperature. The fourth
variable, pond depth, shows modest sensitivity within the parameter
range tested. E * is inversely proportional to H, implying some improvement
in steady state performance as depth is increased. Furthermore, increasing
depth slows the response to fluctuating meteorological conditions
leading to a decrease in variation (standard deviation) of the plant
intake temperatures (see Table 1) .
QUASI-STEADY MODEL
In order to evaluate the performance of a cooling pond, it is necessary to
cover a wide range of meteorological conditions which might occur during
the pond's lifetime. One way to do this would be to run a transient
model with time-varying meteorological conditions for a number of years.
A disadvantage of this type of simulation, however, is the considerable
computation and effort which is involved; at the design stage, where a
number of alternative designs must be evaluated, such a simulation is
impractical. Therefore it is desirable to develop a simpler, approximate
model to be used for the purpose of initial pond design. In particular it
is desirable to use a steady state model so that, as with the design of
cooling towers, a frequency distribution of meteorological data, rather
than a long time series, can be used as model input. The more accurate
transient simulation can then be used to evaluate the chosen design.
The quasi-steady model uses the following differential equation
where T is the equilibrium temperature. Equation. (A) requires the same
boundary conditions as those used with Equation Q.) and differs from
912
-------
Equation (2) only in the use of a linearized excess temperature represen-
tation for surface heat transfer and the fact that the time-dependent
term is missing. The model is quasi-steady in the sense that the input
parameters governing pond performance (plant operating condition and
meteorology) are assumed to be constant over a period of time and the
pond temperature is assumed to be in instantaneous equilibrium with these
parameters. • The constant input parameters are derived by averaging
the real time .parameters over the time interval. Clearly this prpcedure
is an approximation of true pond behavior. By averaging the input data
one is filtering high frequency fluctuations and by assuming "instant
response" one is ignoring the "thermal inertia" known to characterize
ponds. The intent is to adjust the averaging interval such that the
two effects cancel as much as possible in their influence on the cumulative
distribution of intake temperatures.
The solution to Equation (4) was obtained by Wehner and Wilhelm [5], The
resulting plant intake temperature is given by
T.-T 4a exp{l/2E*}
,r\
v '
>z exp{a/2EL*} - (1-a)^ exp{-a/2EL*} - 4a exp{l/2EL*}
where a =
In order to predict the intake temperature T., the equilibrium temperature
T and the heat exchange coefficient K have to be" defined. The former
is defined as the water temperature at which the net heat flux $ =0 and
can be found by an iterative procedure. The linearized surface heat
exchange coefficient is defined by <(> =K(Tg-T ) where Tg is a characteristic
surface temperature. For this analysis T =T + ATQ/2r, where r is
determined by iteration.
Cumulative distributions of predicted intake temperatures using both the
quasi-steady and the transient models were compared using the base case
pond described in the previous section. The transient model was run for
one year using three-hour time steps with three-hour meteorological data.
The cumulative distribution of predicted intake temperature and power
using this model are shown in Figures 4 and 5 as solid lines. Quasi-
steady calculations were also made for the same pond and time period by
averaging the meteorological data over different averaging intervals,
computing values of K and T for each time interval, and then using
Equation (5) to compute intake temperatures. Distribution of intake
temperature and power production are plotted in Figures 4 and 5 for
averaging intervals of 1, 3, 5, 10 and 30 days.
913
-------
Comparison of the various curves suggests that reasonably good agreement is
obtained between the transient model and both the 3 day and 5 day averaged
model. By contrast, results for one day averaging show greater extremes
in temperature suggesting that the averaging has not adequately filtered
the high frequency fluctuations, while the distributions resulting from
the 10 and 30 day averaging are the flattest, suggesting that the averaging
of input data provides more filtering than the transient model. These
results indicate that , for this site and pond, an averaging of between
3 and 5 days is appropriate. This figure is reasonable because it
corresponds approximately to the time constant pcH/K which governs the
response of a shallow water body to a step change in T .
E
REFERENCES
[1] Jirka, G., "Hydrothermal Structure of Cooling Impoundments," presented
at this conference.
[2] Watanabe, M., and G. Jirka, "A Longitudinal Dispersion Model for
Shallow Cooling Ponds," Proc. of First Conference on Waste Heat
Management and Utilization, Miami Beach, May 1977.
[3] Fischer, H., "The Mechanics of Dispersion in Narrow Streams," Proc. of
ASCE, Vol. 93, No. HY 6, 1967.
[4] Ryan, P., and D. Harleman, "An Analytical and Experimental Study of
Transient Cooling Pond Behavior," R.M. Parsons Laboratory for Water
Resources and Hydrodynamics, Technical Report No. 161, Dept. of Civil
Engineering, MIT, January 1973.
T51 Wehner, J., and R. Wilhelm, "Boundary Conditions of Flow Reactor,"
Chemical Engineering Science, Vol. 6, 1956.
Entrance Mixing Region 1
-> -r-^1 •
a) /
^Longitudinal Mixing Region
W
Plant Discharge
Flow: 0 , Temp: T =T.+ AT
xo r o a. o
Plant Intake
Flow: Q , Temp: T.'
o i
Throughflow: DQ
b)
W
Return Flow:/(D-l)0
"o
Figure 1 Mathematical Schenatization for a) Shallow-Dispersive Cooling
Pond b) Shallow-Recirculating Cooling Pond (Plan Views)
914
-------
Ul
Area A
(acres)
750
10001
1500
2000
3000
1000
10001
1000
10002
10001
1000
1000
10001
1000
Depth H
(ft)
9
9
9
9
9
6
9
12
9
9
9
9
9
9
STUDY CASES
Aspect Temperature
Ratio Rise AT
L/W • (°F) °
12
12
12
12
12
12
12
12
3
12
36
12
12
12
20
20
20
20
20
20
20
20
20
20
20
10
20
30
Flow
Rate Q
(cfs)
1800
1800
1800
1800
1800
1800
1800
1800
1800
1800
1800
3600
1800
1200
IP3
0.53
0.51
0.49
0.47
0.44
0.77
0.51
0.38
0.51
0.51
0.77
0.86
0.51
0.38
v4
0.36
0.41
0.51
0.58
0.72
0.62
0.41
0.31
0.83
0.41
0.08
0.41
0.41
0.41
STATISTICS OF INTAKE TEMPERATURE
Month of July
Mean Standard Deviation
(°F) (°F)
94.6
90.7
86.4
84.1
81.7
91.5
90.7
90.1
97.3
90.7
87.4
92.6
90.7
89.2
3.0
2.9
2.8
2.8
2.7
3.5
2.9
2.5
2.8
2.9
3.0
2.8
2.9
3.0
1 base case
2 computed as shallow-recirculating pond using entrance dilution of 2.
3 based on f± = 0.01, and 3 = .00018 op"1
4 based on f - 0.02
TABLE 1: SUMMARY OF SENSITIVITY STUDY
-------
40
30
20
.10
0
100C
750
1500
v/x
X '' 2000
/ , '
' 3000
0 .5 time
a) Variation of A (acres)
100
(°F)
80
60
40
1.0
4U
30
20
10
0
(
;
(3600,10) ,...->^'.
(1800,20) y/^'^
,'/^\
^Z^'''' (1200.30)
'
3 . 5 time 1
100
80
60
40
0
b) Variation of Qo(cfs), ATo(°F)
40
30
20
10
100
80
60
40
'0 .5 time
c) Variation of L/W
1.0
30
20
10
0
9\ ^
\/
12
100
80
60
40
0 .5 time
d) Variation of H (feet)
1.0
Figure 2 Cumulative Distributions of Predicted Plant Intake Temperature
-------
P(MHe)
1200
1190
1180
1170
1160
0 .5 t ime
a) Variation of A (acres)
1200
1190
1160
0 .5 N time
c) Variation of L/W
1200
1190
1180
1170
1160
3600,10)
(1800,20)
(1200,30)
1.0
0 .5 time 1.0
b) Variation of Q (cfs), AT (°F)
1200
1190
1180
1170
1160
0 .5 time
d) Variation of H (feet)
Figure 3 Cumulative Distributions of Predicted Power Production
-------
0 .5 time
a) Plant Intake Temperature
1200
P(MWe)
x Q.S. 1 day ave.
1190
1180
1170
1160
b) Power Production
Figure 4 Cumulative Distributions of Plant Intake Temperature
and Power Production using Transient and Quasi-Steady Models
918
-------
TRANSIENT SIMULATION OF COOLING LAKE PERFORMANCE
UNDER HEAT LOADING FROM THE NORTH ANNA POWER STATION
D. R. F. Harleman, G. H. Jirka, D. N. Brocard,
K. Hurley-Octavio, and M. Watanabe
M. Parsons Laboratory for Water Resources and Hydrodynamics
Department of Civil Engineering
Massachusetts Institute of Technology
Cambridge, Massachusetts U.S.A.
ABSTRACT
The North Anna Power Station of the Virginia Electric and Power Co. (4 nu-
clear units with a combined capacity of 3800 MWe) is located North-West of
Richmond. The heat dissipation system includes a Waste Heat Treatment Fa-
cility consisting of a series of lagoon cooling ponds with attached dead-
end side arms, discharging into Lake Anna, on which the intake is located.
An experimental and analytical study of the buoyancy-driven circulations in
long, dead-end side arms of cooling lakes was carried out. Results were
utilized in the subsequent simulations to demonstrate the relative effect-
iveness of cooling lake side arms in dissipating heat.
A transient, segmented cooling pond model was developed which links the
mathematical models applicable to the components of the Waste Heat Treat-
ment Facility (WHTF) and the main lake. To provide additional information
on the isotherm and velocity patterns in the main lake, a finite element
model for the surface layer in a stratified cooling ^Lake was developed.
The above models were utilized in long-term (10 year) simulations to evalu-
ate the effect of the power plant operation on Lake Anna. The natural
(ambient) temperature regime was predicted using the MIT Lake and Reservoir
Model. The segmented cooling pond model was used to simulate one-, two-,
three- and four-unit power plant operation.
INTRODUCTION
This paper summarizes the development of mathematical models for predict-
ing the performance of a cooling lake used as the condenser heat dissipa-
tion system for the North Anna Nuclear Power Station of the Virginia Elec-
tric and Power Company. The performance of a cooling lake is determined
by both its effectiveness and its thermal inertia. "Effectiveness" re-
lates to the ability of cooling lakes to dissipate the artificial heat
load with the lowest possible intake temperature. This is governed by the
geometric configuration of the lake and by the design of inlet and outlet
structures. One of the interesting features of Lake Anna is the existence
of several long, isolated side arms which required a detailed investigation
919
-------
of their role in the heat dissipation process. "Thermal inertia" is the
ability of cooling lakes to damp out meteorological transients and fluctua-
tions in the power plant operation. Cooling lakes are practically never in
a steady-state condition, hence an evaluation under realistic transient con-
ditions is necessary.
The optimal approach to assess cooling lake performance - whether environ-
mental impacts or technical parameters - is to consider long-term behavior,
over a period of the order of ten years, so as to form representative sta-
tistical, measures, such as average or extreme thermal conditions.
A portion of the lake, into which the condenser water is discharged, is
separated from the main body of the lake by dikes. A major fraction of the
waste heat is dissipated in this portion, known as the Waste Heat Treatment
Facility (WHTF). Its purpose is to minimize the thermal impact on the main
lake and on the stream below the dam forming the impoundment.
The results of the analysis are presented as induced temperatures or tem-
perature rises above natural conditions at various points of interest.
During operation of the power station it will not be possible to measure
"natural" lake temperatures. Therefore, it is necessary to predict tran-
sient natural lake temperature under the meteorological conditions prevail-
ing during operation. The predictions of the natural lake temperature mod-
el were compared with pre-operational observations during several years of
record. Surface isotherms and longitudinal temperature profiles have been
prepared for operating conditions of one to four units. The results are
intended to be used in subsequent analyses by biologists and engineers to
assess the potential environmental impact of the North Anna cooling system
and to compare to applicable thermal regulations. However, no such assess-
ments and/or comparisons are made in this paper.
PLANT CHARACTERISTICS
The North Anna Power Station is located in central Virginia, between Rich-
mond and Charlottesville. The station is situated on the south bank of
Lake Anna formed by a dam on the North Anna River (Fig. 1) which was closed
in January 1972.
The station will ultimately consist of four nuclear units Of a combined
capacity of 3760 MWe (about 940 MWe per unit). The waste heat load rejec-
ted in the condensers is 6.5 x 109 BTU/hr per unit or 25.9 x 109 BTU/hr
total. The condenser cooling water flow rate is about 2,100 cfs per unit
(8,400 cfs total) and the temperature rise through the condensers is about
14°F.
Lake Anna and Waste Heat Treatment Facility
The construction of three dikes and the dredging of channels formed a sep-
arate series of ponds, the Waste Heat Treatment Facility (WHTF). Both the
WHTF and the lake participate in the dissipation to the atmosphere of the
920
-------
waste heat loading, but the WHTF dissipates the major portion. Lake Anna
has a surface area of 9,600 acres, a volume of 10.6 x 109 ft3, and an aver-
age depth of 25 ft. The maximum depth at the dam is 70 ft. The lake re-
ceives an average annual inflow of about 270 cfs. The lake elevation is
maintained by radial gates at the dam. The outflow rate equals the inflow
minus the rate of evaporation from the lake surface (estimated at about 60
cfs average).
Q -3
The WHTF has a surface area of 3,400 acres, a volume of 2.66 x 10 ft and
an average depth of 18 ft. The maximum depth is 50 ft in the vicinity of
the dikes. As shown in Fig. 1, Dike I forms Pond 1 of the WHTF which re-
ceives the cooling water via the discharge canal from the power plant. Con-
necting channels have been dredged between Pond 1 and Pond 2 (formed by
Dike II) and between Pond 2 and Pond 3 (formed by Dike III). These channels
have a constant trapezoidal cross-section of 25 ft depth and 160 ft aver-
age width. After passing through Ponds 2 and 3, the cooling water is dis-
charged into the main lake through a submerged discharge structure in Dike
III. After residence in the main lake, cooling water is withdrawn through
near-surface intakes in the vicinity of the station. A major characteris-
tic of the system is the existence of the long narrow side arms in the
WHTF. These arms comprise about 1530 acres or 45% of the area of the WHTF.
The North Anna heat dissipation system has a low heat loading per unit sur-
face area. Using the combined surface area for the main lake and the WHTF,
the loading ranges between 0.15 and 0.6 MWt (waste heat) per acre for 1 and
4 units, respectively. Frequently, cooling pond designs have a much higher
thermal loading, between 0.5 and 3 MWt per acre. The loading on the area
of the WHTF along corresponds to 2.2 MWt/acre for four units.
CHARACTERISTICS OF cuOLING LAKES
Experimental and theoretical studies on cooling lake behavior have been
conducted by Ryan and Harleman [1] and Watanabe, Harleman and Connor [2],
The major results of earlier studies have been summarized by Jirka, Abraham
and Harleman [3] and a detailed report on the North Anna cooling lake study
has been prepared [4].
The temperature differential which exists in a cooling lake between the
discharge and the intake of the power plant is, if transient fluctuations
are averaged, equal to the condenser temperature rise. As density changes
are associated with temperature changes, buoyant forces arise which tend
to cause spreading of lighter (warmer) water over heavier (cooler) water.
The paramount role of these density currents has been observed in labora-
tory experiments [1]. In deep ponds it was found that density currents
effectively spread the heated water over the entire surface of the pond,
even if there are distinct backwater ("dead") areas. Thus, deep cooling
ponds are characterized by a heated, thin surface layer, in which pre-
dominately horizontal temperature variations occur due to cooling to the
atmosphere, and an underlying subsurface layer, in which only vertical
temperature variations occur due to the gradual advective flow to the
921
-------
submerged intakes. Ryan and Harleman [1] also established the importance
of discharge channel design to minimize entrance mixing and of a submerged
skimmer wall intake structure and formulated a transient predictive model
consisting of two parts:
a) the surface layer model, which assumes a constant surface layer thick-
ness and computes transient areal temperature distribution resulting from
heat loss to the atmosphere; account is also taken of an entrance mixing
region,
b) the subsurface model, which calculates the vertical temperature varia-
tion due to downwelling from the end of the surface region; the subsurface
region may be weakly or strongly stratified. This model is an adap'tation
of the M.I.T. Deep Reservoir Model [5],[6], The cooling pond model has
been applied to several field cases of deep cooling lakes and excellent re-
sults have been obtained in the calculation of the transient annual behav-
ior.
An accurate prediction of temperatures induced by heated discharges hinges
on the correct specification of the heat transfer from the water surface to
the atmosphere. The heat dissipation of artificially heated water surfaces
has been addressed by Ryan et al [7] and heat dissipation formulae have
been developed which specifically account for the evaporative heat transfer
due to free buoyant convection which arises from the virtual temperature
difference between the moist air at the water surface and a certain dis-
tance above the water surface.
DEVELOPMENT OF PREDICTIVE MODELS FOR THE NORTH ANNA COOLING SYSTEM
The preceding discussion has stressed that the applicability of available
mathematical models for cooling lake prediction is strongly tied to the
thermal structure of a cooling lake. In turn, the thermal structure de-
pends on geometric features of the lake and discharge and intake struc-
tures. The North Anna cooling system, consisting of a series of ponds in
the WHTF with attached side arms and connecting channels and of the main
lake, is expected to have a particularly complex thermal structure. For
example, while the individual ponds of the WHTF will be distinctly strati-
fied, there is a tendency for destratification in the connecting channels
of the WHTF. Also, the role of buoyant convective circulations into the
isolated side arms of the WHTF is expected to be important. None of the
existing models encompass all of these features.
The following approach was taken in the development of predictive models
for the North Anna cooling system:
Side Arm Heat Dissipation
An experimental and analytical study of the buoyant convection which occurs
due to surface cooling in long side arms of cooling ponds was carried out
by Brocard, et al. [8]. As shown in Fig. 2,the salient features are the
922
-------
length and depth of the side arm, the thickness and temperature of the
stratified layer at the entrance to the side arm and the surface cooling
rate. In addition, the special features of the lateral constructions with-
in the side arm and of bottom slopes were investigated. The results of the
side arm investigation are represented in design graphs, which give the
flow rate and associated temperature drop as a function of the governing
parameters.
In order to analyze the complex structure of the North Anna heat dissipa-
tion system, a "segmented model" was developed which links different mathe-
matical models applicable for each of the components of the WHTF and the
main lake. A schematic diagram of the segmented model is shown in Fig. 3.
For the three ponds of the WHTF which are of shallow average depth, a two-
layer model was developed in which each of the layers is assumed to be ver-
tically uniform and which includes the inflows into and outflows from the
side arms. Stability criteria describe the mixing of the layers at the
connecting channels between the individual ponds. The residence times in
each WHTF pond is of the order of two days and thus larger than the compu-
tational time step of one day. Therefore, the transient characteristics
were accounted for through a delay in the computed temperature at the end
of each pond which is equal to the residence time of each pond.
WHTF Model
Pond 1 of the WHTF does not include any major side arm and is schematized
as shown in Fig. 4. The condenser discharge Qo at temperature TQ under-
goes some mixing at the entrance. The dilution ratio Ds = (Qo + Qe)/Do is.
obtained from the buoyant surface jet model of Stolzenbach and Harleman [9]
corrected for the interference of the jet with the bottom of the receiving
water [3], Qe, the entrained flow, is a function of the densimetric Froude
number of the surface jet and the geometry of the discharge channel. T^ is
the temperature after mixing and ^ is the temperature in the canal leading
to pond 2. The heat flux to the atmosphere, cf>n, is linearized in the usual
way, <|>n = - K(T - Tg) where K is the surface heat transfer coefficien4- and
T and TE are the surface and equilibrium temperatures. The temperatur^
distribution in the reach is treated in a one-dimensional fashion with re-
spect to surface area. Since Ds depends on the entrance densimetric Froude
number which itself depends on T2 (the temperature in the counterflowing
lower layer), the solution involves iterations.
Pond 2 of the WHTF has two side arms and is shown schematically in Fig. 5.
Three possible flow configurations must be considered: (a) the entrance
jet entrainment flow is greater than the sum of the flows entering the side
arms, (b) the entrainment flow is smaller than the sum of the side arm
flows and (c) the entrainment is smaller than the flow entering the first
side arm. Figure 5 shows the counter flow conditions for case (b).
The model for pond 3 is similar to that for pond 2. The final considera-
tion for the WHTF is the submerged discharge of the condenser flow, at tem-
perature T^, into the main lake through dike III (see Fig. 3). As shown in
Fig. 6, the lake in front of the dike is rather shallow and is constrained
923
-------
laterally. Therefore, the lake water for entrainment and mixing with the
dike III jet must come through a restricted section. It is assumed that
the limiting entrainment flow, Qe, is reached when the lower layer flow at
section "A" is critical.
Main Lake Model
The maximum depth of the main lake is 70 ft near the downstream dam and
50 ft near the plant intake, while the computed upper layer depth is 'of the
order of 15 ft for all cases of plant operation. The main lake can there-
fore be considered as a deep cooling lake for which the model of Ryan and
Harleman [1] is applicable. The lake is separated in two regions: - a
surface layer assumed to be of uniform temperature vertically. Its hori-
zontal temperature distribution is solved as a function of surface area on
a transient basis. The shape and location of the isotherms is therefore
not determined, but the model gives the surface area inside each isotherm;
- a subsurface pool assumed to be vertically stratified, but of uniform
horizontal temperature. The vertical temperature profiles in this region
are computed following the assumptions and method of the M.I.T. Lake and
Reservoir Model described in [10],
A side arm reach is attached to the end of the main lake. This reach rep-
resents the regions of Lake Anna (4231 acres) which are located upstream
of a lateral construction about two miles to the northeast of the power
plant. The amount of side arm flow, its temperature drop and the return
flow in the lower layer are calculated using the techniques discussed for
the WHTF.
Finite Element Model for Velocity and Temperature
Distributions in Surface Layer of Lake Anna
A transient finite element model has been developed by Watanabe, et al. [2]
which predicts the two-dimensional temperature and velocity distributions
in the surface layer of the main lake. The FEM model is an extension of
the main lake model which predicts surface temperatures only as a function
of area fractions. Because of the expense of running the two-dimensional
FEM (215 mesh points), it was used only for short period (2 week) studies
for detailed temperature distribution in the main lake. The FEM grid is
shown in Fig. 1 and representative velocity and temperature distributions
are shown in Figs. 7 and 8.
Prediction of Natural Lake Temperatures
Any lake temperatures which are induced by the power plant operation must
be considered relative to the naturally occurring conditions (in the ab-
sence of plant operation). The M.I.T. Lake and Reservoir Model (Octavio
et al. [10] was used to provide the predictions on a long-term basis. The
model was verified using meteorological and hydrological input data col-
lected on the North Anna site during 1974-76 by comparing the predictions
with measured pre-operational lake temperatures.
924
-------
Many lakes and reservoirs exhibit horizontal temperature homogeneity and
thus a time-dependent, one-dimensional model which described the tempera-
ture variation in the vertical direction is adequate to describe their
thermal structure. The M.I.T. Lake and Reservoir Model is a time-dependent,
one-dimensional, variable area, discretized mathematical model based on the
absorption and transmission of solar radiation, convection due to surface
cooling, advection due to inflows and outflows and wind mixing. The model
contains provisions for simultaneous or intermittent withdrawal from multi-
level outlets and time of travel for inflows within the reservoir. Turbu-
lent entrainment at the thermocline is treated by the wind mixing represen-
tation. The wind mixing algorithm is based on the rule that the rate of
change of potential energy of the water column due to entrainment is equal
to the rate of input of kinetic energy by the wind. An iterative procedure
minimizes the accumulation of errors in the computation of the heat input.
The1 model inputs include daily averaged values of air temperature, relative
humidity, wind speed, cloud cover, and total short wave solar radiation.
The model time-step is one day. The absorption coefficient for short wave
solar radiation, T], was computed from Secchi disk depths.
A comparison of measured and predicted water surface temperature during
1976 is shown in Fig. 9, good agreement is obtained with respect to both
absolute value and transient behavior. Vertical temperature distributions
for two days in May and July for which detailed measurements were performed
are compared with predicted values in Fig. 10.
LONG TERM SIMULATIONS
A long term simulation of the natural surface temperature of Lake Anna was
made for a ten year period using meteorological data for 1957-1966. The
ten year average and standard deviation above and below the mean are shown
in Fig. 11. Similar ten year simulation runs were made for operational
conditions corresponding to one through four units. A comparison of sur-
face temperatures at the dam for meteorological conditions corresponding
to year 1962 (an average year in the 10 year sequence) with 2 units opera-r
tional is shown in Fig. 12. Changes in the vertical temperature structure
of the lake were also computed. Figure 13 shows the ten year average for
3 operating units in comparison with the natural temperatures. The rela-
tive heat losses in the WHTF and in the main lake are indicated by the
longitudinal temperature profile shown in Fig. 14 for 4 units in operation.
Seventy percent of the total induced temperature change of 14°F occurs be-
tween the plant discharge into the WHTF and the end of the jet mixing zone
downstream of dike III.
CONCLUSIONS AND ACKNOWLEDGEMENTS
The North Anna heat dissipation system is an example of an effective com-
bination of a highly loaded, stratified cooling pond (the WHTF) and a
lightly loaded cooling lake. It has also been demonstrated that in strat-
ified systems, dead-end, side arms are effective in dissipating heat
925
-------
through buoyancy induced circulation. The hydrotherraal model developed for
North Anna is computationally efficient, thereby making possible long-term
simulation runs covering a wide range of meteorological and plant operating
conditions.
This study was supported by Virginia Electric and Power Company, Richmond,
Virginia, and by Stone and Webster Engineering Corporation, Boston, Mass.
We gratefully acknowledge the close cooperation and assistance of the fol-
lowing individuals: Morris Brehmer, Carl Pennington and Robert Rasnic at
VEPCO and David Knowles, David McDougall, Fred Mogolesko and Robert Taylor
at Stone and Webster.
REFERENCES:
1. Ryan, P.J. and Harleman, D.R.F., "An Analytical and Experimental Study
of Transient Cooling Pond Behavior", M.I.T., Department of Civil Engi-
neering, R.M. Parsons Laboratory for Water Resources and Hydrodynamics
Technical Report No. 161, Cambridge, Massachusetts, 1973. (Subsequent
reports of this laboratory are referred to as M.I.T., R.M. Parsons T.R.
No. .)
2. Watanabe, M., Harleman, D.R.F. and Connor, J.J., "Finite Element Model
for Transient Two-Layer Cooling Pond Behavior", M.I.T., R.M. Parsons
T.R. No. 202, 1975.
3. Jirka, G.H., Abraham, G. and Harleman, D.R.F., "An Assessment of Tech-
niques for Hydrothermal Predictions", M.I.T., R.M. Parsons T.R. No.
203, 1975.
4. Jirka, G.H., Brocard, D.N., Hurley Octavio, K.A., Watanabe, M. and
Harleman, D.R.F., "Analysis of Cooling Effectiveness and Transient
Long-Term Simulations of a Cooling Lake (with application to the North
Anna Power Station), M.I.T., R.M. Parsons T.R. No. 232, 1977.
5. Huber, W.C. and Harleman, D.R.F., "Laboratory and Analytical Studies
of Thermal Stratification of Reservoirs", M.I.T., R.M. Parsons T.R.
No. 112, 1968.
6. Ryan, P.J. and Harleman, D.R.F., "Prediction of the Annual Cycle of
Temperature Changes in a Stratified Lake or Reservoir: Mathematical
Model and Userfs Manual", M.I.T., R.M. Parsons T.R. No. 137, 1971.
7. Ryan, P.J., Harleman, D.R.F. and Stolzenbach, K.D., "Surface Heat Loss
from Cooling Ponds", Water Resources Research, Vol. 10, No. 5, 1974.
8. Brocard, D.N., Jirka, G.H. and Harleman, D.R.F., "A'Model for the Con-
vective Circulation in Side Arms of Cooling Lakes", M.I.T., R.M. Par-
sons T.R. No. 223, 1977.
9. Stolzenbach, K.D. and Harleman, D.R.F., "An Analytical and Experimental
Investigation of Surface Discharges of Heated Water", M.I.T,, R.M. Par-
sons T.R. No. 135, 1971.
10. Hurley Octavio, K.A., Jirka, G.H. and Harleman, D.R.F., "Vertical Heat
Transport Mechanisms in Lakes and Reservoirs", M.I.T., R.M. Parsons
T.R. No. 227, 1977.
926
-------
to
Fig. 1
North Anna
Cooling System
Fig. 3 Schematic of Segmented Model
for North Anna
Fig. 2 Schematic of Side Arm Convective Circulation
-------
00 T
to
00
Fig. 4 Schematic of Pond 1 of WHTF
Os Qo
DsQo-Osi
DS00-0S,-
1 / // f / I H H i H /1II1111' f " ' ' /ff f ''I II
Fig. 5 Flow Configuration of Pond 2 of WHTF,
with Two Side Arms
0.1 ft/sec
Fig. 7 Velocity Distribution in the Surface
Layer for 50% Downwelling Flow at
the End of the Lake and the Remain-
ing 50% Distributed Along Both Sides
,-,.. 87.0°
87.5'
85.5
IIL
Fig. 8 Temperature Distribution in the
Surface Laver Computed from FEM
Fig. 6 Cross Section Along Axis of the Dike III Jet
-------
32
28
IE
5
1
F H
J
B-
&- 5 •
&ti
i IB-
?B
R-[1S -
c=1
an .
r J
•^
•
" . .
*
_ i — i — i — t
Meas . •
-Pred.
HAY 7. 1974
-4 — 1 — 1 — 1 — 1 — h— 1 — t—
12
Temperature °C
> -.—•—_ 'n _ T^* UJ
Fig. 9 Measured and Predicted Natural Surface Temperature', IB..
NORTH RNNfl
NRTURflL TEMP
SIMULATION
10 YRS flVERRGE
I—I—i—I—H—I—I—I—I—II I
e \z
16 SB
El 28
Fig, 10 Measured and Predicted Vertical
Temperature Profiles
uoo
RPR'MBY'JUN'JUL'nuc'sEP'OCT'NOV'OEC JRN'FEB HBR
Fig. 11 Ten Year Average Natural Surface Temperatures
and Standard Deviations
-------
to
o
*-
NORTH RUHR
COQLING POND
SIHULflllOU
196a 2 UNITS
2 units
W 100 ISO JOT «0J03 ISO HI B
oca 'hint JUN ' JUL 'BUG 'SEP 'OCT 'not DEC 'JON 'FCB 'r.m '
Fig. 12 Loaded and Unloaded Temps.
in 1962, with 2 Units
Operational
NORTH BNNfl
COOL IWC PONO
SlnuiflTION
DVCRflCC 3 UNI I!
d
den '«sr 'JUN 'JUL ouc SEP ucr NOV DEC JBM F£» «nn
Fig. 13 10 Year Averaged Loaded and
Unloaded Temperatures, with
3 Units Operational
101
100
99
98
97
96
95
94
93
92
91
90
89
88
87
86
,,P1ant Discharge
\Canal B (end of reach 1)
Canal C (end of reach 2)
Dike III
of Jet mixing zone
Main dam
2000
4000
6000
Fig. 14
Average Temperature Profile for North Anna Cooling
System during Month of July 1962, with 4 Units
Operating
-------
COMPARISON OF THE SURFACE AREA REQUIREMENTS OF A
SURFACE TYPE CONDENSER FOR A PURE STEAM
CYCLE SYSTEM, A COMBINED CYCLE SYSTEM AND A
DUAL-FLUID CYCLE SYSTEM
M.H. Waters
VP Engineering
International Power Technology
California, U.S.A.
Dr. E.R.G. Eckert
Regents Professor Emeritus
University of Minnesota, U.S.A.
ABSTRACT
A recently issued patent to International Power Technology (IPT)
on the Dual-Fluid cycle (DFC) and analysis by Kinney, et.al., on
steam injected gas turbine cycles has demonstrated significant
benefits for engines which use steam as a second working fluid
in ,a gas turbine engine. These benefits include high thermal ef-
ficiency which is comparable to or better than that for combined
cycle powerplants, and reduced system mechanical complexity and
initial cost when compared with combined cycle powerplants.
The objective of this paper is to provide a quantitative evalua-
tion of the condenser requirements for DFC engines. A very im-
portant feature of such an engine is that the exhaust gas at the
condenser inlet is approximately at atmospheric pressure and is
between 300°-400°F depending upon the cycle parameters. This
contrasts dramatically with the condenser for either a pure steam
system or a combined cycle system which is at low vacuum pres-
sures (0.5-1.5 psia) and thus low temperature (80 -115 F).
INTRODUCTION
A recently issued patent to International Power Technology (IPT)
on the Dual-Fluid Cycle (DFC) and analysis by Kinney, et. al.,
on steam injected gas turbine cycles has demonstrated significant
benefits for engines which use steam as a second working fluid
in a gas turbine engine (References 1 and 2). These benefits in-
clude high thermal efficiency which is comparable to or better
than that for combined cycle powerplants provided the mixture of
steam-air in the gas turbine is carefully controlled. Figure 1
is a schematic drawing of a DFC engine. A detailed description
of the cycle is given later in the report, but the main feature
is that the steam generated in the waste heat boiler is injected
into the gas turbine ahead of the turbine section. In contrast,
931
-------
a combined cycle powerplant uses a separate steam turbine system.
There is an excellant potential for reduced system mechanical
complexity and thus initial cost for DFC powerplants because of
the single shaft output and no requirement for a separate steam
turbine.
An obvious concern for the DFC is the performance of the conden-
ser since the exhaust gas is a mixture of water vapor and non-
condensable gases (typically the steam-air ratio is .15). A
very important feature of a DFC engine is that the exhaust gas
at the condenser inlet is approximately at atmospheric pressure
and is between 300 -400 F depending upon cycle parameters. Con-
densing will begin at 130-180 F depending upon the amount of vapor
in the exhaust and thus relatively high temperature differences
exist across the heat exchanger surfaces. This contrasts drama-
tically with the condenser for either a pure steam system or a
combined cycle system which is .at low vacuum pressure (0-5-1.5psi)
and thus low saturation temperatures (80°F-115 F).
The condenser surface area requirements for steam cycle power-
plants are relatively easy to compute since the condensing fluid
is a pure vapor and the equivalent heat transfer coefficient is
a constant throughout the condenser. However, if the fluid is
a mixture of water vapor and non-condensable gases, the heat
transfer coefficient varies from point to point in the heat ex-
changer as the composition of the gas mixture changes due to re-
moval of water vapor as condensate. The flow of vapor towards
the condensing surface is diffusion controlled in a-pure steam
condenser since the vapor migrates to the cool tube surface as
a sink. In a mixture of gases which contains vapor and non-con-
densable gases, both migrate to .the cool tubes but the non-con-
densable gases would have to diffuse away from the surface. This
can severely reduce the condensation heat transfer coefficient.
In general for a DFC engine, there is a much larger quantity of
non-condensable gases than there is water vapor, and the rate of
condensation is a function of the vapor concentration in the mix-
ture. Thus, the surface area calculation becomes a step wise
process through the condenser as both the bulk fluid temperature
and the fraction of water vapor in the gas mixture is reduced.
The condensation process is also controlled by convection because
of the large fraction of non-condensable gases. Higher velocities
through the tube banks thins the boundary layer thus increasing
the condensing heat transfer coefficient. However, this also in-
creases the pressure dr5p through the condenser which degrades
engine performance thus creating a trade off situation.
The objective of this paper is to compute the condenser surface
area requirements for a given DFC engine and make direct size
comparisons with condensers for power equivalent steam and com-
bined cycle powerplants. A condenser for a DFC engine can either
be a contact type as in boiler scrubbers or be a surface type
932
-------
with a bundled tube arrangement. The latter type is used in this
naoer as a basis for comparison.
NOMENCLATURE
F Equivalent temperature driving force (defined by equation 9)
g Gravity constant
G Mass flow of air
a
G Mass flow of steam
s
h Equivalent heat transfer coefficient of condenser vapor film
Q>
hf Heat transfer coefficient in cooling water film
hf Enthalpy of condensation
h Heat transfer coefficient in flowing gas film
g
h Conductivity of the tube well
w
ID Tube inside diameter
k Gas film mass transfer coefficient
g
k. Thermal conductivity of the tube well
k Thermal conductivity of liquid film
W
MTD Mean temperature difference (defined by equation 5)
n Number of tube rows in a bank
Nu Nusselt number of flowing water
W
OD Tube outside diameter
P .Total pressure
p Steam partial pressure at tilm interface temperature
Pr Prandtl number of water
w
p Steam partial pressure at steam bulk temperature
o
Q Heat flow rate
Re Reynolds number of flowing water
S Heat exchanger surface area
T Film interface temperature
T Bulk temperature of vapor or vapor-gas mixture
S
T Water temperature
U Overall heat transfer coefficient defined by equation 1
U Overall heat transfer coefficient defined by equation 11
t*
AT1 Temperature difference (T -T )
S W
/w
Density of water
u w
Dynamic viscosity of water
933
-------
METHODOLOGY
Condenser surface areas are designed by general procedures of
heat exchanger design with the particluar problem being the cal-
culation of heat transfer coefficients with condensation. In
the case of steam systems - either a conventional steam cycle or
a combined cycle - the overall heat transfer coefficient is cal-
culated from straight forward formulas and is essentially con-
stant throughout the exchanger. However, for the Dual-Fluid
Cycle condensation proceeds from a mixture of gases - the com-
bustion products of air and steam. For this reason, the flow of
vapor towards the condensing surface is diffusion controlled only
across a thin boundary layer. This can be designed as a trade-
off against the pressure drop across the tube bank. The non-con-
densable gases also flow towards the condensing surface and then
diffuse away after they cool to preserve a local mass balance at
the condensing surface. Thus, the heat transfer coefficient varies
from point to point in the condenser as the composition of the
mixture varies due to the removal of water vapor.
The cooling medium in the condenser is specified to be water, and
the following two sub-sections describe the computational methods
in some detail.
Condensing from Pure Vapor
The sketch below demonstrates the mechanisms of heat transfer
that take place in the pure vapor condenser;
* Forced convection between the flowing saturated vapor
and the condensed vapor film
* Conduction across the condensed vapor film
* Conduction across the tube well
* Forced convection in the cooling water film
Tube Wall
Condensed Vapor
Film
-------
The overall heat transfer coefficient is given by:
1_ _ (OP/IP) _^ OP ^ A. ^ A
•f w V.•[/ c
For pure vapor, 1/h = 0, so that the last term disappears from
equation (1) and thi outside film temperature, T , is equal to
the saturated vapor temperature, T . The therma? conductivity
of the tube, kfc, is constant, and thus the heat transfer coeffi-
cient, h , is constant.
The equivalent heat transfer coefficient of the condensed vapor
film, h , for a bank of tubes is given by the formula from refer-
ence 3 :
The cooling water film heat transfer coefficient is found from
MacAdams Formula:
The temperature difference MTD, to be used with U, is the aver-
age between the temperature difference at inlet and at outlet.
We shall use the arithmetic average:
^-rVcT
(5)
The total exchange surface required is:
(6)
935
-------
Condensing From a Mixture of Gases
The heat transfer coefficient U is not constant throughout the
exchanger so that the previous method cannot be applied. A pro-
cedure developed by Colburn and Hougen is used, in thfe form pre-
sented by Votta (See references 4 and 5).
The condenser is divided in several sections for each of which
U is calculated and assumed constant. The exchange surface for
each section is:
The total exchange surface is:
In this procedure, instead of calculating U and AT separately,
their product is used, as determined from a local heat balance
(reference 5) :
4 = U (T, -TC)
The steam partial pressure is computed from G the steam flow
rate, G the air flow rate, and P the toatl pressure of the mix-
ture. a
In the modification by Votta, the first equality in the string
of equations (7) is written as:
where F(T) is a function of temperature and represents a single
potential for both heat and mass transfer at the film interface.
Table 1 is taken from reference 5 for a steam air mixture.
936
-------
Table 1
Temperature F at 1 ata F at 2 ata
°F °F °F
32 0
40 15
50 37
60 62
80 132 90
100 239 154
120 400 242
140 679 381
160 1138 597
In eqns (9) the interface temperature T to be used in finding
F is determined by trial and error froifi the second equality in
tne string (9) :
For each condenser section, first the interface temperature is
found, then F and F from the table, then equation (9) to find
the area dS. Then tne surfaces of the different sections are
added up to give the total exchange surface S. The equivalent
heat transfer coefficient between the cooling water and the outer
face of the condensed steam film, U , is given by the equation:
\*-
where the calculation of h , h , and h. is as given in equation
(2), (3) and (4) respectively.
Pressure Drop Across a Tube Bank
The pressure drop across a tube bank having a staggered arrange-
ment is as follows (reference 6):
where n is the number of tubes, f is the friction factor ( a
function of Reynolds number), /g is the density and V the velo-
city of the gas.
937
-------
POWERPLANT DEFINITION
For purposes of comparison a 10,000 horsepower powerplant for
marine propulsion is assumed. The steam cycle powerplant is
typical for modern marine powerplants, whereas both the combined
cycle and Dual-Fluid Cycle powerplants are based on the cycle
of the General Electric LM 5000 gas turbine.
Sea water is the condenser coolant, and to provide for opera-
tions in tropical seas, it is assumed that the inlet tempera-
ture of the sea water to the condenser is 85°F. The condenser
cooling water discharge temperature is assumed to be 88°F.
Steam Cycle
The steam cycle power selected for this study is a regenerative-
reheat single unit which is typical of steam powerplants for
marine application. The engine has the following character-
istics taken from Babcock and Wilcox (reference 7).
Throttle Pressure (psia) - 1465
Throttle Temperature (°F) - 1000
Reheat Temperature (°F) - 1000
Condenser Pressure (psia) - 0.7
Boiler Efficiency (%) - 90
The system has a best heat rate of 7460 Btu/hp-hr, and thus the
overall efficiency is given by:
IboIUr = (2545) (Q.90) = .307
7460
Therefore, the heat rejected by the condenser is given by:
Rejected Heat = 2545 (-^y - D(0.90) = 5170 Btu/hp-hr
At 0.7 psia condenser pressure, the latent heat is 1043 Btu/lb.
Thus, the flow rate of steam for a 10,000 hp engine is given by
(WJ)(10'000) = 49/569 lb/hr
The condensing temperature at 0.7psia condenser pressure is
quite low (90°F). Thus, the temperature differences in the
condenser are quite small since the cooling water is 85-88°F.
This results in very large surface areas. Higher condensing
pressures should also be considered even though the thermal
efficiency of the steam system will suffer. Table 2 summarizes
the pertinent data for three condensing pressures 0.7, 1.0
and 2.0 psia.
938
-------
Table 2
Steam Cycle 10,000 hp Engine
Condensing Pressure (gsia) 0.7 1.0 2.0
Condensing Temperature ( F) 90 102 126
Overall Efficiency (%) 30.7 30.3 29.3
Rejected Heat ' (Btu/hp-hr) 5170 5269 5526
Steam Flow (Ib/hr) 49,569 50,859 54,07C
These data will be used in computing the condenser surface areas
later in the report.
Combined Cycle
The combined cycle powerplant used in this study is based on
the Curtiss Wright Mod Pod 35A gas turbine which uses the LM5000
gas generator. This cycle has high efficiency, and it was se-
lected because it would represent a very high performance com-
bined cycle powerplant. The heat rate and the exhaust tempera-
ture of the Mod Pod 35A are as follows:
Heat rate = 7050 Btu/hp-hr
Turbine exhaust temperature = 786 F
Although it is not in production, the Curtiss Wright Corporation
has proposed a combined cycle powerplant with the LM 5000 as
the gas turbine (reference 8) . The stated heat rate for this
combined cycle conversion for the LM 5000 is 5513 Btu/hp-hr
(thermal efficiency = 46.2%).
A brief evaluation was made for both single pressure and two
pressure steam turbine systems based on the LM 5000 cycle. De-
tails of this evaluation are given in the Appendix. For the
computation of condenser requirements of a 10,000 horsepower
combined cycle powerplant, the following engine characteristics
are defined:
Table 3
LM 5000/Combined Cycle 10,000 hp Engine
Condensing Pressure
Condensing Temperature F 90 102 126
Single Pressure System
Thermal Efficiency (%) 46.9 46.3 45.5
Power Split (steam/gas) .301 .287 .275
Rejected Heat Flow (Btu/hp-hr) 5910 6322 6710
Steam Flow (Ib/hr) 13106 13608 14161
939
-------
Two Pressure System
Thermal Efficiency (%) 49.0 48.2 47.0
Power Split (steam/gas) .357 .334 .301
Rejected Heat Flow (Btu/hp-hr) 4583 5074 5910
Steam Flow (Ib/hr) 11560 12263 13379
Although the thermal efficiencies shown in the table are greater
than that of the LM 5000 combined cycle quoted above, the dif-
ference is not great and is probably the result of not accounting
for any degradation in the gas turbine performance due to back
pressure from the waste heat boiler. For comparison purposes,
the results in Table 3 are representative of the best in com-
bined cycle engine performance, and these data will be used in
computing the condenser surface areas later in this report.
Dual Fluid Cycle
As with the combined cycle, the Dual-Fluid Cycle (DFC) makes
use of two separate working fluids. In either cycle, each fluid
is compressed separately; but in the DFC the two fluids are
combined in a single mixture for expansion through the tur-
bines and heat regeneration in the waste heat boiler.
The Dual Fluid Cycle essentially combines a regenerative Bray-
ton cycle and a regenerative Rankine cycle system in parallel.
Thus, both cycles are operating at the specific turbine inlet
temperature. In contrast, combined cycle engines combine the
Brayton and Rankine cycles in series and the maximum tempera-
ture of the Rankine cycle fluid is the turbine discharge temp-
erature .
To describe the operation of a DFC engine, the following sub-
paragraphs outline the thermodynamic cycle in more detail.
1. Compression of the two fluids takes place separately. Air
is compressed from atmospheric pressure up to the maximum cycle
pressure in a conventional axial or centrifugal flow compressor.
Water is pumped at ambient temperature to a pressure somewhat
greater than the compressor discharge air pressure.
2. Combustion takes place in a mixture of air and a suitable
fuel in a conventional gas turbine combustor. Water, in the
form of steam, is then mixed with the combustion products of air.
This steam is the result of water being preheated by the regen-
erative heat exchanger (see No. 4 below) and is at a somewhat
higher pressure than the combustion gas to promote proper mixing.
3. The resultant mixture of combustion products of air and steam,
hereafter called the gas mixture, is at a specified maximum tur-
bine inlet temperature which dictates the combination of water-
air ratio and fuel-air ratio selected. Expansion of this gas
mixture takes place in two conventional axial flow turbines.
940
-------
The first or high temperature turbine drives the air compressor
through a connecting shaft. The second turbine is a free tur-
bine which provides the useful output work.
4. The gas mixture discharging from the power turbine is then
passed through a counter flow regenerative heat exchanger. This
heat exchanger uses steam which is then injected into the combustor
(see No. 2 above). Thus the heat in the cycle is being "regen-
erated".
5. The gas mixture leaves the heat exchanger at or above the
saturation temperature of the steam in the gas mixture as deter-
mined by the partial pressure of the steam, and it then passes
through a condenser. In general, no condensation is desirable
from a design consideration in the heat exchanger, but in the
condenser steam condenses to water and is separated from the
mixture. The remaining products of the combustion of air are
exhausted to the atmosphere. The condensed water is purified
pumped to high pressure and recycled to the regenerative heat
exchanger.
It must be pointed out that steam injection into gas turbine
engines is not a new concept. However, almost without excep-
tion the objective has been to increase power for short per-
iods of operation. Extensive work by IPT over the past 5 years
has demonstrated that extremely high efficiencies can be achieved
in the DFC described above. However, peak efficiency can only
be obtained at a particular balance of air-steam ratio and air-
fuel ratio. Reference 1 describes in some detail how this bal-
ance is linked to the cycle pressure ratio and the turbine in-
let temperature of the gas turbine engine.
The Dual Fluid Cycle does lend itself to a retrofit of existing
engines. Mechanical design modifications must be carefully
assessed for a particular engine, but for purposes of this paper
which is to compare condenser surface area requirements - it is
presumed that a LM 5000/DFC engine can be accomplished. Thermo-
dynamic cycle calculations give the following peak performance
characteristics for such an engine.
Table 4
LM 5000/DFC Engine
Overall Thermal Efficiency (%) 48
Power per Unit Air Flow (hp/pps) 291
Steam Flow/Air Flow .113
Air Flow/Fuel Flow 43
As described in the Methodology section, condensing from a mix-
ture of gases poses a different computational problem because
the heat transfer coefficient is not constant through the ex-
941
-------
changer.
The gas mixture temperature at the discharge of the waste heat
boiler is 400 F, and thus the condenser is a gas cooler from
this temperature down to the temperature at which vapor begins
to condense. This will be outlined in more detail in the fol-
lowing section on condenser surface areas.
CONDENSER SURFACE AREA REQUIREMENTS
Steam Cycle and Combined Cycles
Using the methodology outlined in a previous section, the over-
all heat transfer coefficient is relatively easy to compute for
pure steam systems - both the steam cycle engine and the com-
bined cycle engine. Several assumptions were made concerning
the heat exchanger:
Incoming Temperature of the Cooling Water ( F) 85
Cooling Water Temperature Rise ( F) 3
Outside Diameter of Condenser Tubes (in) .75
Inside Diameter of Condenser Tubes (in) .62
Water Flow Velocity in the Tubes (fps) 5
The following equations are used to derive the the number of
condenser tubes:
Cooling Water Flow Rate = Heat Exchanger Duty/Temperature Rise
Total Cross Sectional = Cooling Water Flow Rate/Density/
Area for Water Flow Velocity
TT 2
Number of Tubes = Cross Sectional Area/KID )
Tables 5, 6 and 7 summarize the pertinent data for the steam
cycle engine, the single pressure combined cycle engine and the
two pressure combined cycle engine respectively. Recall that
each powerplant is rated at 10,000 hp and that the combined cycle
steam system is relatively small because of the split in power
between the gas turbine and the steam turbine.
Table 5
Steam Cycle System
Condensing Pressure (psia) 0.7 1.0 2.0
Condensing Temperature (°F) 90 102 126
Cooling Water Flow Rate (Ib/hr x 1.72 1.76 1.84
10"7)
Number of Condenser Tubes 7319 7491 7830
942
-------
Overall Heat Transfer (Btu/2
Coefficient hr ft °F)
Mean Temperature Difference (°F)
Duty (Btu/ 7
hr x 10 )
Condenser Surface Area (ft )
Table 6
Single Pressure Combined
Condensing Pressure (psia)
Condensing Temperature (°F)
Cooling Water Flow Rate (Ib/hr x
10~7)
Number of Condenser Tubes
Overall Heat Transfer (Btu/?
Coefficient hr ft F)
Mean Temperature Diff. (°F)
Duty (Btu/ 7
hr x 10 )
2
Condenser Surface Area tft )
Table 7
Two-Pressure Combined
Condensing Pressure (psia)
Condensing Temperature (°F)
Cooling Water Flow Rate (Ib/hr x
10-7)
Number of Condenser Tubes
Overall Heat Transfer (Btu/-
Coefficient hr ft F)
Mean Temperature Difference ( F)
Duty (Btu/ 7
hr x 10 ')
2
Condenser Surface Area (ft )
543
3.5
5.17
27923
Cycle
0.7
90
.593
2522
444
3.5
1.78
11454
Cycle
0.7
90
.547
2326
440
3.5
1.64
10649
674
13.5
5.27
5791
1.0
102
.605
2575
577
13.5
1.81
2328
1.0
102
.565
2403
570
13.5
1.69
2201
742
39.5
5.53
1887
2.0
126
.615
2618
677
39.5
1.85
690
2.0
126
.593
2522
674
39.5
1.78
668
943
-------
The condenser surface areas for the three different systems are
shown in figure 2. it is fairly obvious that the need for low
condensing pressures to achieve a high thermal efficiency will
result in a very large condenser. Note that the data in figure
1 are plotted on semi-log paper and the reduction in condenser
surface area with higher condenser pressures is quite signifi-
cant.
The combined cycle condenser exhibits an identical trend, but
at significantly lower areas since the steam turbine handles
only a fraction of the total engine power.
Dual-Fluid Cycle
As outlined in the Methodology section, the calculation of con-
denser surface area for a mixture of gases is a step wise process.
The heat exchanger may be considered to be formed of two sections
with different heat transfer regimes:
Section 1 - Single phase heat transfer, from inlet to
point of saturation
Section 2 - Two phase heat transfer from the point
of saturation to the outlet.
The exhaust temperature from the waste heat boiler of an LM 5000/
DFC engine is 400 F, and the temperature in section 1 drops to
133 F before condensation begins to take place. For section 2,
the temperature difference from inlet to outlet is divided into
intervals and the heat exchange duty and surface is computed
for each interval. Table 8 summarizes the pertinent data for
each of the intervals. The total condenser surface area for
the 10,000 hp engine is calculated to be 3823 ft .
The final temperature (not shown on table 8) is 93 F, and it was
chosen so that the water remaining as vapor in the gas mixture
was balanced by the water formed in the combustion process, i.e.,
there is no water lost or gained by the system. If it is accept-
able to provide makeup water, then the final condensing tempera-
ture can be higher, and thus the condenser will be smaller. This
is shown in figure 3. Also shown in the figure is the effect of
water vapor from the atmosphere. For example if the water break-
even point was designed for an ambient temperature of 80 F with
a relative humidity of 0.6 the final condenser temperature would
be approximately2101 F and the required condenser surface area
would be 3080 ft .
It should be noted that a larger portion of the condenser areas
(1436 ft ) is actually a gas cooler - not a condenser. It is en-
tirely feasible to use this waste heat for other purposes in a
separate heat exchanger. For example, it can be used for stack
gas heating or for feed water heating if necessary or desirable.
944
-------
Table 8
Calculation of Required Exchange Surfaces
for the LM 5000/DFC 10,000 hp engine
vO
Section 1
(gas cooler)
Section 2
(condenser)
Gas *
temper-
ature
°F
400
133
125
120
115
110
105
TOTAL
Heat Gas heat Interface Overall Heat
exchange transfer temper- heat trans, exchange
coeff. „ ature coeff. 2 surface
Btu/h Btu/hFft °F Btu/hFft ft
9^1 x 106 38.2 38
3.559 x 106 34.9 105 216
1.773 x 106 34.4 101 210
1.514 x 106 34.3 98 205
1.353 x 106 34.2 96 200
1.160 x 106 34.1 95 195
2.292 x 106 34.0 91 190
2.135 x 10 TOTAL
1436
383
240
252
260
272
980
3823
* at the beginning of the interval
-------
CONCLUDING REMARKS
The objective of this paper was to provide a quantitative eval-
uation of the condenser surface area requirements for a Dual-
Fluid Cycle engine and to compare these results with those for
pure steam condensers - both conventional steam cycle and com-
bined cycle. The evaluation summarizes the choice of parameter
tradeoffs for all three engine systems, and it is safe to state
that at a constant rated power the condenser for a DFC engine will
have a surface area which is smaller than a steam cycle engine
and comparable to the combined cycle engine.
The original concern was that the DFC engine condenser would
be significantly larger because the heat transfer rate from
the gas mixture was greatly reduced due to the presence of non-
condensable gases. However, the heat transfer is no longer com-
pletely diffusion controlled. The temperature and concentration
gradients can be controlled by the forced convection process in
the flow of gases over the condenser tubes. Although the heat
transfer coefficient is much reduced, the large temperature dif-
ferences available through most of the condenser compensate for
this deficiency.
It is worth noting that the DFC engine condenser is essentially
operating at atmospheric pressure. Thus, light weight materials,
even plastics, and units having rectangular cross sections can
be employed. In contrast the low vacuum pressures typical of
a steam system requires heavier metals with cylindrical cross
sections to overcome buckling stress. Thus, the DFC engine con-
denser offers the potential for design flexibility and cost re-
duction .
946
-------
There are many aspects to consider in optimizing either of these
systems, and they should be evaluated on the basis of powerplant
operating costs taking into account the heat rate of the total
combined cycle powerplant and the initial cost and required main-
tenance of the steam cycle equipment. This is beyond the scope
of this paper, but these considerations should be kept in mind
when evaluating the tradeoffs. Reference 9 is a good example
of a tradeoff study for a combined cycle steam system.
Single Pressure System - Power Output
It is assumed that the state-of-the-art in steam turbine isen-
tropic efficiency is approximately 80%. Obviously the power
rating of the turbine will have an effect on isentropic efficien-
cy. Smaller turbines result in a reduced Reynolds number in the
steam turbine and thus reduced efficiency. However, this effect
is considered to be a second order effect and is ignored here.
For the single pressure system, the condensing pressure (and thus
temperature) is specified and expansion is presumed to take place
down to the Wilson line (96% quality for a given pressure). In
addition, the maximum steam temperature is assumed to be 736 F
(50 degrees below the LM 5000 turbine discharge temperature).
Thus, the only unknown in the system is the maximum steam pres-
sure, and this is easily determined from the steam tables.
The results shown in figures Al and A2 demonstrate the expected
result. The maximum steam pressure is reduced for reduced con-
densing pressures; however, the thermal efficiency based on heat
input to the waste heat boiler and the power per unit steam flow
both increase at the reduced condensing pressures. The sketch
below demonstrates how the reduced steam pressure is beneficial
to the waste heat boiler.
Temperature
947
-------
The lower pressure results in a lower pinch point temperature,
and thus more heat is transferred from the gas turbine exhaust.
This results in a greater quantity of steam flow and an increase
in steam turbine power per unit flow of the gas turbine exhaust.
This has a direct impact on the overall thermal efficiency of
the combined cycle powerplant as discussed later in this Appendix.
Two Pressure System -- Power Output
A different approach is taken for the two pressure system. To
specify the system, both the condensing pressure and the maxi-
mum steam pressure of the system are specified. Expansion through
the first turbine proceeds to the Wilson line. This determines
the pressure for the second turbine and a second pass through
the boiler elevates the steam temperature back to 736 F. The ex-
pansion through the second turbine is down to the specified con-
densing pressure or to the Wilson line if it is encountered
at higher pressures. More typically the specified pressure will
be reached while still in the superheat region.
Results for the two pressure system are shown in figures A3 and
A4. Based on the heat input from the waste heat boiler to the
steam, the results in figure A3 suggest - that increased steam
pressure for the first turbine is desirable since both the steam
thermal efficiency and the power per unit steam flow increase
as this pressure increases. Again low condensing pressures im-
prove the system performance significantly. However, the quanti-
ty of steam which can be generated is increased at reduced maxi-
mum steam pressures just as it is for the single pressure system
(see figure A4). Thus, there is a tradeoff, and the upper por-
tion of figure A4 shows there is a well defined optimum in terms
of power from the steam turbine per unit flow in the gas turbine
exhaust.
Overall Thermal Efficiency of the Combined Cycle Powerplant
The Curtiss-Wright Mod Pod 35A mechanical drive gas turbine is
derived from the LM 5000 gas generator, and its thermal efficiency
and power per unit air flow as stated in reference 10 are 36.1%
and 153 hp/p"ps respectively.
Integration of a waste heat boiler for the combined cycle will
affect this performance somewhat due to the back pressure at the
gas turbine. For purposes of comparison, this effect is ignored
for the results shown in figures A5 and A6.
The data presented in figures A5 and A6 summarize the results
for the single pressure and two-pressure systems respectively.
As expected, a reduced condensing pressure increases the power
split (steam turbine power/gas turbine power) and thus the
overall efficiency of the combined cycle powerplant. For the
two pressure system, note that the peak efficiency shifts to
948
-------
maximum steam pressures as the condensing pressure is
reduced.
These results are those used in the main body of the report
to compute condenser surface areas.
949
-------
1. Anon., Regenerative Parallel Compound Dual Fluid Heat Engine,
U.S. Patent 4,128,994 , (12 December 1978).
2. Fraise, W.E. and Kinney, C.; Effects of Steam Injection on
the Performance of Gas Turbine Power Cycles, ASME Paper No.
78-GT-ll, April, 1978.
3. Chen, M.M., J. Heat Transfer, Trans ASME, C 83:55 (1961).
4. Colburn, A.P., and Hougen, O.A., Ind. Eng. Chem., 26:1178
(1934) .
5. Votta Jr., F., Condensing from Vapor Gas Mixtures, Chem.
Engrg., 71:223 (1964).
6. Eckert, E.R.G., Analysis of Heat and Mass Transfer, McGraw
Hill Book Company (1972).
7. Anon, Steam Its Generation and Use, The Babcock and Wilcox
Company, 37th Edition.
8. Anon, "Mod Pod LM 5000's offer lowest plant heat rates",
Gas Turbine World, September, 1976, pp26-30.
9. Berstein, E., and Cashman, J.; The Energy Saver Combined
Cycle, ASME paper 78-GT-127, presented at the Gas Turbine
Conference, London, April 9-13, 1978.
10. Anon, Gas Turbine World Handbook, 1977-78, Pequot Publica-
tions, Inc.
950
-------
Ul
AIR
CONCEPT OF THE DUAL-FLUID CYCLE
ENGINE
fIf F W^) ^^^^ EXHAUST TO
CONDENSER/SEPARATOR
WASTE HEAT BOILER
SUPERHEATED
STEAM INJECTION
INTO COMBUSTOR
CLEAN
WATER
SINGLE
SHAFT
OUTPUT
Figure 1
-------
CN
-P
CM
CD
0)
O
(0
i-i
(1)
OT
c
0)
T3
C
O
u
Steam Cycle
r~\ Combined Cycle - Single Pressure
Combined Cycle - Two Pressure
50000 r-
10000
5000
1000
500
100
0.5 1.0 1.5 2.0
Condenser Pressure - psia
Figure 2 Effect of Condenser Pressure on Vapor Condenser
Surface Area
952
-------
Ambient Temperature = 80 F
4000
•i
4J
(d 3°°°
2
Q)
U
M-l
£ 2000
CO
OJ
CO
c
0)
•o
g 1000
U
•
\ *
\ \
\
\
"
0*
4 * *
»
\
\ \
k\ \.
>S*xs^ ^S*""^^^ ^S^**^_
^•x-, ^^"~* ^"**
100 60 0
Relative Humidity (%)
ii • «
02 0 .02
Water Makeup/Airflow
.04
water gain by system
water lost to system
Figure 3 Relation between Condenser Size and Water Gained/Lost
to the System for a LM 5000/DFC Engine Rated at 10000 hp
953
-------
500 i-
475
O
iH
fn
e
0)
d)
4-)
CO
\
^
Oi
s
O
0)
w
£)'
iH
\
cx
450
425
u
c
0)
•H
O
•H
U-I
14-1
w
to
E
J-l
cu
28
24
01
-U
20
Q)
i-i
0)
S-i
0)
-p
CO
03
•H
tn
CX
250
200
150
0.5 1.0 1.5 2.0
Condensing Pressure - psia
Figure A-l
Effect of Condensing Pressure on the Performance
of a Single Pressure Steam Turbine System
954
-------
O
o
50
o 46
<\)
CO
42
38
o
rH
Pn
w
s
fO
0)
-p
.100
095
090
.085 ..
0.5 1.0 1.5 2.0
Condensing Pressure - psia
Figure A-2 Effect of Condensing Pressure on the Performance
of a Single Pressure Combined Cycle System
955
-------
800 h-
O 700
t— i —*
t, o
-P rH
a)
5
O
fil
x:600
500
35 h
u
c
0)
•H
o
-H
Vn
0)
CJ
-P
30
25
20
0.5
1.0
2.0
Condensing Pressure
(psia)
.- 0-5
1.0
2.0
_L
250 5"lTD 750 1000
Maximum Steam Pressure - psia
JL
1250
Figure A-3 Effect of the Maximum Steam Pressure and the Condensing
Pressure on the Performance of a Two-Pressure Steam
Turbine System I
956
-------
o
en
CO
J-i
S
0
0
rH
En
cn
(0
o
rd
0)
•P
00
55
50
u
0)
tn
A_
45
40
.09
.08
.07
06
250
500
Condensing Pressure
(psia)
_L
750
1000
1250
Maximum Steam Pressure - psia
Figure A-4 Effect of the Maximum Steam Pressure and the Condensing
Pressure on the Performance of a Two~Pressure Combined
Cycle System
957
-------
4_)
•H
i-H
a
CO
M
OJ
*j
b
cu
•jj
o
u
T3
as
c
•r-<
*G
g
o
u
o
c
•H
rQ
h
3
^H
U)
!u
f n
CD
c
•H
M
E-"
e
03
Q)
CO
.35
.30
.25
u
c
•H
o
•H
w
M
(1)
EH
-------
H C
H -H
a &
m
(T3
in
QJ
O
CM
QJ
rH
O
u
T3 EH
QJ
C £
•H ra
^3 (U
e -u
o c/i
U
35
30
0.7
u
C
0)
•H
u
•iH
14-1
M-l
w
(0
Vl
Q)
V
H
u
u
0)
C
•H
J2
E
0
U
0\°
50 u
48
46
44
Condensing Pressure
(psia)
0.7
1.0
2.0
42 -
Figure A-6
250 500 750 1000 1250
Maximum Steam Pressure - psia
Performance of a Two-Pressure Combined Cycle System
959
-------
UTILIZATION OP TRANSFORMER WASTE HEAT
D. P. Hartmann
Bonneville Power Administration
U.S. Department of Energy
Portland, Oregon
H. H. Hopkinson
Energy Systems Division
Carrier Corporation
Syracuse, New York
ABSTRACT
Bonneville Power Administration has installed a Carrier Corporation
specially designed heat pump system to utilize transformer waste heat
for heating a substation control house at the J. D. Ross Substation in
Vancouver, Washington. The source of waste heat is a 250 MVA transformer.
It has 90 kW of iron losses present whenever the transformer is energized.
It also has variable or load dependent copper losses which range up to
300 kW at rated load. Because of the heat sensitivity of the oil/paper
electrical insulation, the hottest spot in the transformer is limited to
less than 55 C rise above ambient.
Because only a fraction of the available waste heat from the transformer
can be used at the control house, and because of possible problems with
interruption of the existing oil to air cooling system, an innovative air
to freon heat recovery system was implemented. This allowed installation
with minimum equipment connection to the transformer, thus preserving the
high electrical reliability required of the transformer service. Preon
22 serves as the transformer medium because it can go directly into the
heat pump eliminating heat exchanger losses and potential freezing and
water line heat losses which would have resulted in a lower coefficient
of performance for the heat pump.
This paper describes the heat pump size in terms of the original building
load, and the building modifications; insulation and storm windows, which
reduced the required heat pump size. The components and system along
with its controls are also described. A special accumulator is required
to separate crankcase oil from the vaporized freon as it comes to the
heat pump.
This project was initiated several years ago. The heat pump was installed
in October 1977* This paper summarizes the test results and operating
data collected to date, along with the anticipated annual performance,
which is expected to yield a quite satisfactorily high coefficient of
performance for the heat pump.
The paper concludes with recommendations for future transformer heat
retrieval systems.
960
-------
INTRODUCTION
The impetus for the Prototype Energy Retrieval and Solar (PERS) system
arose in the autumn of 1973- Rainfall and snowfall in the Columbia
Basin was very low due to a drought. The conservation measures instituted
at that time were successful in limiting electric power demand (l).
Also the boycott of oil, while not directly influencing the Pacific
Northwest's Hydro based electric generating system did reemphasize the
need to conserve energy. The PERS was developed in an attempt to use
locally available energy other than electricity for heating and cooling
control houses at BPA substations. This first system is intended as a
test bed to try various alternatives for retrieving energy for the
substation environment.
Figure 1 is a block diagram showing the overall system. During the winter
heating season, the power transformer has its highest load, and losses.
This is because widespread air-conditioning is not required west of the
Cascade Mountains but electric heating is widely used. Thus the peak
load occurs in winter in the Pacific Northwest. Utilization of transformer
waste heat looked attractive since the potential source is near the
control house (See Figure 2). Electric power transformers have been
designed and built for nearly 100 years so they are very efficient; above
99.98 percent. However, even a 250,000 kVA transformer, 0.15 percent
losses are significant in that nearly 400 kW are lost when the transformer
is at rated load. By connecting a Freon 22 evaporator into the transformer
oil-cooling air stream, a portion of the transformer's waste heat is
captured in vaporized freon to give a boost to the electric heat pump
which heats the control room.
During the summer cooling season, 840 tubular glass vacuum insulated solar
collectors, in 35 panels, inclined at 30° facing south, for a total of
89 sq. meters (959 sq. ft), in four rows, (Figure 3) gather the sun's
heat to drive a lithium bromide absorption chiller (Figure 4) to provide
chilled water to cool the control house space. In spring and fall,
intermediate heating is obtained from the solar energy stored in an
insulated 16,200 liter (4300 gallon) storage tank (Figure 5). The lithium
bromide absorption cooler is described in another paper (2). The heat
pump, absorption chiller, and associated pumps and controls are housed
in an auxiliary building adjacent to the control house (Figure 6).
DESIGN PARAMETERS
The transformer, whenever it is excitable by voltage on the primary, has
constant iron losses of 90 kW. This is due to hysteresis losses in the
core material. Copper losses are a function of load and range up to
300 kW at a rated load of 250 MVA. Figure 7 shows a typical month's
analysis, from existing records, of how the average daily load on the
961
-------
transformer, the ambient temperature and the top oil temperature are
related. The temperature of the cooling air coming out of the transformer
oil cooler averages about 13 C warmer than the ambient temperature for
this particular transformer during the winter months. Figure 8 shows a
typical actual days temperature performance for November 9, 1978.
Table I lists the heating design conditions for the Ross Control House.
Our first winter's experience shows that with the building's new storm
windows, if no outside air makeup is provided, the indoor humidity
becomes too low for comfort; well into the lower 20 percent R.H. range.
Preliminary heat loss evaluation of the building showed an estimated
27 kW heat loss. In 1938, 35 kW of electric resistance heaters had been
installed during original construction. After application of storm
windows (Figure 9) and roof insulation plus removal of the original glass
skylight (Figure 10) and the replacement by precast concrete and an
insulated roof reduced the control room heat needs to about 6 kW.
EQUIPMENT LOCATION
The option of placing the heat pump at the transformer vs in the control
building were explored. The transformer area is under 230-kV power lines
and access to the transformer for maintenance precluded placement of the
heat pump at the transformer. The control house, upon casual inspection,
appears to have space available, but most of this space is required for
instrument carts used during maintenance. Thus, it was decided to erect
a separate Armco type building on the north side of the existing control
house. In a new construction program, before power lines are energized,
better location of the heat pump nearer to the transformer would be
included in the construction plan. As it is, the separate location eased
construction since all of the equipment was in one place away from the
hazards of construction near the 230-kV power lines. The result is that
a loss in booster temperature due to extra line length was traded for ease
during construction and for continuing operation and maintenance. Service
personnel can enter the separate auxiliary building without access to the
secured electrical substation area, and without interference with substation
control room operations and maintenance.
HEAT PUMP COMPONENT AND SYSTEM DESCRIPTION
EVAPORATOR
The evaporator is a specially constructed 1.22 m (4 foot) square, 7.6cm
(3-inch) thick plate fin coil arrangement manufactured by Carrier
Corporation. It is placed directly above the existing transformer to
air cooler and extracts heat from the cooling air stream by vaporization
of Freon 22. The evaporator is shown in Figure 10. Two evaporators are
962
-------
required since alternate weekly operation of the two existing transformer
cooling units is required to assure reliable availability of both
cooling units should the transformers partner bank be out for maintenance.
At that time rated load would be placed on the transformer and
simultaneous operation of both cooling units would be automatically
started by control thermostats.
COMPRESSOR
Figure 11 shows the compressor as installed in the auxiliary building.
It is a Carrier Model MD24/3HP with special oil level indicators and a
welded shell which is hermetically sealed. This is an air to water
heating unit.
CONTROLS
The heat pump is set to operate in a heating mode, a cooling mode for
backup for the solar lithium-bromide absorption chiller and a heat pump
down mode.
Refrigerant gas will condense at the coldest point in the refrigerant
system causing a local low pressure area which induces more refrigerant
gas to flow towards the system cold point. Since this system has tow
remote refrigerant evaporators, it is necessary to prevent these
components from collecting the total refrigerant charge leaving
insufficient charge in the operating circuit. In the cooling mode, the
100 meters of (328 ft) of return gas refrigerant line, up to 35-7 kg
(78.7 Ibs.) of refrigerant could be held.
Check valves have been added at the outlets of each evaporator and also
at the outlet of the return gas line to prevent this refrigerant migration.
However, since solenoid and check valves can leak refrigerant long
periods of time, a heat pump down mode was incorporated into the controls
for the heat pump. This mode is controlled by the liquid level in the
accumulator (receiver).
ACCUMULATOR (RECEIVER)
This is a tank in the heat pump suction line which separates oil and any
liquid freon from the evaporator generated freon vapor. A small hole in
the suction line pickup is provided to return oil to the compressor.
Initially, a small accumulator of about ^94 liter (l qt.) was used.
During cold start, the capacity was exceeded by liquid freon in the return
from the evaporators. One compressor required replacement due to failure
induced by liquid freon entry to the compression chamber. Replacement of
the original accumulator with a 15 liter (3 gallon) unit solved this
problem. Also, a small heater is incorporated into the accumulator to
provide some vapor during cold start conditions.
963
-------
DOWNSTREAM PRESSURE REGULATOR
This pressure regulator located in the return gas line to the compressor
senses pressure entering the compressor and is set to restrict refrigerant
flow to the compressor such tat a suction pressure above 71?kPa
(104 Psig) is not reached. This is to protect the compressor from over-
loading caused by excessive energy absorbed by the remote transformer
evaporators.
PERS HEAT PUMP HEATING SYSTEM
PAN COIL
The fan coil unit is shown in Figure 12. This is a 15-ton carrier
packaged chilled water air handling unit of the 408 RS series. It
delivers heated or cooled air to the room through plenum space ducting
at 200 CFM. Adjustable dampers are available to modulate this air flow
from control room requirements, i.e., eliminate drafts, but assure
adequate air circulation for delivery of heating and cooling.
WATER PIPING - CONTROL VALVES
Water pipes are connected from the heat pump to the fan coil unit for use
in heating or as a backup cooling for the lithium bromide absorption
chiller. There also is a connection which allows use of solar heated
water to be sent to the fan coil unit for solar heating. All water piping
is simulated with a 11.4 cm (4.5 inch) radial thickness of performed
urethane foam insulation. The heat pump cooler uses water cooling. To
prevent water from freezing in the cooler under certain conditions of
the "pump down" cycle, a solenoid valve has been added to the heat pump's
refrigeration circuit. This solenoid valve is de-energized any time the
pressure in the cooler drops below 400 kPa (51Psig). It will remain off
until the pressure increases to 448 kPa (65 Psig). The cut off point
of the pressure switch (58 Psig) corresponds with a refrigerant saturation
temperature of 0 C (32 P).
INSTRUMENTATION
The overall system has 49 analog and 9 discrete (contact closure) inputs
to gather thermal insolation wind condition fluid flow, pressure and one
control room relative humidity data for the PERS system. In order to
obtain a heat balance for the heat pump precision, several thermopiles
capable of measuring temperature to within +_ 0.1 C were installed. Fluid
flow is measured with Brookes Instrument turbine flow meters which
utilize frequency to analog voltage converters to provide inputs to the
BPA supplied PPP1135 data formulating computer. A microwave channel sends
964
-------
accumulated data to Portland headquarters where the CDC CYBER is
utilized for data reduction calculations. Professor Gordon Reistad
of the Mechanical Engineering Department of Oregon State University
of Corvallis, Oregon, has a contract for system data evaluation.
HEAT' PUMP HEATING SYSTEM PERFORMANCE
There were problems initially with too small a suction line accumulator.
Upon replacement of the accumulation with a larger size, the heat pump
unit has worked well. In the fall of 78, after one year's service,
a freon leak was detected. After refilling the heat pump system with
additional freon, and repairing the leak, the heat pump heating system
has again been providing excellent performance. The data for the
78-79 heating season is being gathered and analyzed. A report will be
made available for NTIS Distribution in 1979. If the transformer were
near rated load capacity for most of the heating season, the expected
heat pump COP might rise to near 6. However, the transformer usually
shares load with a sister unit and is only operating at about one-half
of normal name plate rating. Thus, the heat pump COP on November 9,
1978, was calculated at 2.8 average for the hour ending at 1700 PST, and
the COP was calculated at 2.3 overage for the hours ending at 2200 PST.
Use of a higher fraction of the transformer's waste heat would lead to
larger, better insulated freon tubing with a better resultant COP for
the heat pump.
OTHER APPLICATIONS OP TRANSFORMER WASTE HEAT
Commonwealth Edison of Chicago (3) has a number of substations inside the
Sears tower in downtown Chicago. Portable service water is used to
cool these transformers with double wall oil to water heat exchanges.
Thus the Sears' building waste transformer heat is used to preheat the
building's service water year around.
Hydro Quebec (4) in their downtown headquarters building use the air
from their transformer cooling system on the building makeup air source
for the winter heating season. In summer, the transformer waste heat is
vented to the environment.
Seattle City Light (5) in cooperation with the Electric Power Research
Institute (EPRl) and a contractor, Rocket Research, are surveying the
United States utility industry as to the extent of the available resource
in the transformer waste heat. The contractor is also studying the
feasibility of heating the Pacific Science Center in downtown Seattle
with waste heat from a large transformer of Seattle City Light located
about 1,000 feet from the center.
965
-------
RECOMMENDATIONS POR FUTURE TRANSFORMER HEAT RETRIEVAL SYSTEMS
The method of reclaiming waste energy from the transformer described
in this paper is based on a direct expansion evaporator coil mounted
above the transformer oil cooling coil. This method transfers energy
from the cooling oil to the air and in turn energy is transferred from
the cooling air to the refrigerant system by means of the evaporator
coil. Using the air side transfer approach removed any potential
problems of coolant oil contamination or restrictions to oil flow rates.
Figure 13 is a diagram showing the application of a specially designed
"three-way" coil. This "three-way" coil makes it possible to obtain
waste energy from the transformer or if waste energy is not available
from the transformer, then energy can be obtained from the outdoor
cooling air. Under certain operational conditions, energy for the
heat pump may come from both sources. Modulation of either the oil or
air flow rates or both provides a means of optimizing the transformer
and heat pump operation.
This approach has all the advantages of the present PER System plus an
improvement in heat pump COP. The improved COP results from the
elimination of air as the energy transfer medium and transferring energy
directly to the refrigerant by way of the coil fins. It maintains a
separation of transformer oil and refrigerant - an oil leak would not
contaminate the refrigerant system; either system can operate independently
of the other; and installation requires only that the oil cooling coil
be changed - no modifications to the transformer proper are required.
In addition to an improved COP, a savings in heat transfer surface costs
can be realized on new installations. Two individual coils are replaced
by one three-way coil.
While this approach can be utilized to provide space heating in accessory
buildings associated with a power distribution center, as was done in the
PER program, the total potential of the available waste energy will
usually exceed the requirement for space heating. Therefore, to be
successful in fully using the available energy, some combination of a
power distribution center matched with a manufacturing facility requiring
large amounts of heated process water and space heating should be
investigated. A combination of this type offers economies for both the
utility and its customer.
A conventional heat pump such as used on the PER system can successfully
be used to heat water to a 50 C level which is satisfactory for space
heating. For the higher temperature levels required for process water,
a different type of heat pump is required.
966
-------
Refrigerant R-12 and R-22 are "dry" compression types and result in
discharge gas temperatures higher than those obtained by using so-called
"wet" compression types of refrigerant. Refrigerant R-113 is an example
of a "wet" compression refrigerant. A centrifugal compressor using
R-113, for example, could supply process water heated up to 90°C without
exceeding the discharge gas temperature limits for reliable operation.
The COP for all heat pump systems falls as the condensing temperature
level increases; therefore, any pricing structure for heated process
water should increase with the required temperature level.
Figure 14 is a line diagram for a system which could meet the energy
requirements for both heating process water and space heating. In a
system of this type, it would be possible to operate the heat pump only
during "off-peak" hours storing energy for later "peak" time use. This
helps to balance the load with available power generating capacity, but
would substantially reduce the amount of waste energy available.
In certain areas of the country, it may be economically feasible to add
a solar collection system to operate in parallel with the heat pump
system.
967
-------
TABLE I
ENERGY RECOVERY SYSTEMS
INPUT MTA FOR DESIGN POINT
Ambient Air Temperature
Transformer Load
Oil Temperature
Heat Rejection, Oil Radiator
Air Temperature, Leaving Radiator
Air Flow Rate
21 °F
67$
68.5°F
361,500 BTU/HR
54.5°?
10,000 CFM
AIR COIL MTA
Coil Dimensions
Rows
Spacing, Face
Row
Fins
Saturated Suction Temp.
Capacity
58' x 58' x 3 1/2
2
2 Inches
1.5 Inches
12.95 Inch'1
42.1°F
55,000 BTU/HR
968
-------
REFERENCES
1. D. Davey, "The Pacific Northwest Energy Conservation Program"
9th Interagency Energy Conversion Engineering Conference,
San Francisco, August 1974, IEEE publication Mo. 74CH08T4-8 pp 560-566
2. Dr. Wendell Biermann, "An Absorption Machine for Solar Cooling"
to be presented at the 1979 ASHRAE meeting on January 28 in
Philadelphia, PA.
3. Private Communication with Mr. Aldo Zanona of Commonwealth
Edison Company, Maywood, Illinois.
4. Private Communication with Dr. Jacque Bonneville of Hydro Quebec,
Electrical Research Institute, IBEQ, Varannes, Quebec.
5, Private Communication with Mr. Eldon Ehlers of the Electric
Power Research Institute, (EPRl).
969
-------
LITHIUM BROMIDE-LOW
TEMPERATURE-ABSORPTION UNIT
AGSORPTIOr
NIT PUMP
ENERGY RECOVERY It AUXILIARY-
COOLING SUBSYSTEM
"1 (SANITARY ||SANITARV SANiTARy|f JJSANITARY
JORAIN DRAIN ORAINJ| ||DRAIN
PROTOTYPE ENERGY RETRIEVAL SYSTEM BLOCK DIAGRAM
l-t
X
w
I
t->
O
O
Figure 1 System Diagram
-------
~1IIIIIIT
AVERAGES FEBRUARY 1975
T I I I I I I I I I I I \ I T
LOAD - ROSS BANK NO. 1
u
o
UJ
CE
D
Ct
UJ
a.
5
ui
40
30
20
10
TOP Q1L TEMP. ROSS BANK NO.
- PORTLAND
I I I I I I I
10 12 14
TIME OF DAY { PST )
16
18
20
J L
120
100
80 O
O
60
40
20
22
24
FIGURE 7 HOURLY AVERAGE TEMPERATURES AND TRANSFORMER LOAD PROFILE FOR FEBRUARY, 1975
FOR A 24-HOUR PERIOD.
-------
N)
20
u
Q
U
U
Q
bl
IS
10
u
K
2 *
o. 5
ui
i-
0
-1
• Ambient
* Evaporator No. 1 Air Temperature
• Evaporator No. 2 Air Temperature
* *
* •
•
* *
* *
*****
* * * *
1
I i I I
l
I
I
• •
•
•
I
100 200 300
500 600 700 BOO 900 1000 1100 1200 1300 1400 1SOO 1600 1700 1800 1900 2000 2100 2200 2300 2400
PACIFIC STANDARD TIME
FIGURE 8, J. D. ROSS SUBSTATION, TEMPERATURES FOR NOVEMBER 9, 1978
-------
COOLING OIL
TO TRANSFORMER
THERMO
EXPANSION
VALVE
OIL FROM
TRANSFORMER
OUTDOOR
COOLING AIR
REFRIGERANT
TO HEAT PUMP
COMPRESSOR
LIQUID REFRIG
FROM HEAT PUMP
CONDENSER
THREE-WAY-COIL
1. Oil cooled by energy transfer from oil to outside air (oil to fins to air),
heat pump off.
2. Oil cooled by heat pump evaporator by energy flow from oil to refrigerant
(oil to fins to refrigerant), heat pump on.
3. Transformer off, energy flow from air to refrigerant (air to fins to
refrigerarrt), heat pump on.
Figure 13
973
-------
THREE-WAY COIL
TRANSFORMER
IV
1OTOR
k
_J~C
c
CENTRIFUGAL
COMPRESSOR
D
CONDENSER
PUMP
MANUFACTURING
PROCESS
I
HOT WATER
STORAGE TANK
PUMP
SPACE HEATING
FAN COIL
Figure 14
974
-------
975
-------
976
-------
977
-------
°7P
-------
979
-------
THE APPLICATION OF PRESSURE-STAGED
HEAT EXCHANGERS TO THE GENERATION
OF STEAM IN WASTE HEAT RECOVERY SYSTEMS
Dr. Dah Yu Cheng
President
International Power Techonolgy
California, U.S.A.
Mark H. Waters
VP Engineering
International Power Technology
California, U.S.A.
ABSTRACT
The objective of this paper is to describe an improved system
of transferring heat energy from a high temperature fluid to a
low temperature fluid which undergoes a thermodynamic transition
from the liquid phase to the vapor phase. A counter current
heat exchanger is employed and the cool fluid may undergo thermo-
dynamic transition at more than one pressure. This requires ad-
ditional mechanical components. It will be shown that with this
heat exchanger either a greater amount of heat energy can be
transferred per unit surface area or z. greater amount of fluid
will undergo the thermodynamic transition than is possible by
conventional techniques.
INTRODUCTION
Heat exchangers are employed in various chemical engineering
processes such as powerplants, heating and cooling systems and
energy retrieval systems. Generally, heat exchange design has
focused upon means to transfer the greatest amount of heat per
unit surface area of the exchanger. However, with the recent
interest in co-generation, which makes use of gas turbine waste
heat to generate steam, there is also incentive to increase the
amount of steam generated from a fixed quantity of waste heat.
In conventional heat exchangers, the fluid to be heated is sup-
plied at a certain pressure. The temperature of the fluid be-
gins to rise generally under a continuously smooth temperature
profile unless thermodynamic transition occurs. If such a tran-
sition does in fact occur, the heated fluid would for a period
have a constant or flat temperature profile until all of the
liquid has been converted into vapor. The limiting variable is
the temperature difference between the heating and heated fluid
980
-------
for when this temperature difference is small, very little energy
is transferred between the two fluids. If the temperature dif-
ference between the two fluids is small, a heat exchanger must
have a correspondingly large surface area in order to transfer
a given amount of energy. An, optimum situation would exist if
one could maintain the temperature difference between the fluids
at a maximum so that the heat exchanger surface area could be
kept at a minimum and thus reduce the equipment costs involved
in the energy transfer.
The objective of this paper is to describe a novel heat exchanger
design which maintains a greater temperature difference between
the two fluids by having multiple evaporators which operate at
different pressures. This paper is abstracted from the patent
described in reference 1.
DESCRIPTION OF A PRESSURE-STAGED HEAT EXCHANGER
A pressure staged heat exchanger has at least two evaporators
which are separated by staged pumps. As the fluid to be heated
enters the heat exchanger, it increases in temperature until it
reaches its thermodymanic transition point at a pressure below
the final exit pressure of the fluid. During thermodynamic
transition, the fluid is partially vaporized. The two-phase
iluid is then pressurized to a pressure which is substantially
equivalent to the exit pressure of the heated fluid. At this
point, the thermodynamic transition temperature is raised and
the heated fluid begins to increase in temperature until it
reaches a second, higher, thermodynamic transition point. The
fluid enters into thermodynamic transition in a second high
temperature evaporator and continues thermodynamic transition
intil the fluid in a liquid state is vaporized. Once vaporized,
the temperature of the vapor begins to increase and, in the case
of water, superheated steam exits the evaporator.
The pressure staged heat exchanger described in the previous
paragraph was described as having two evaporators separated
by a single stage pump. As will be explained later, however,
a pressure staged heat exchanger can be designed with a multi-
tude of evaporators separated by a multitude of stage pumps.
The number of such stages depends upon design characteristics
such as energy costs in operating multiple pumps, the surface
area of the exchanger, and the nature of the fluids employed
in the energy transfer and the payoff in weight cost and energy
recovery efficiency.
Conventional Heat Exchanger Constraints
Figure 1 represents conventional heat exchangers in which the
heated and heating liquids travel in countercurrent paths.
981
-------
IX-B-113
If thermodynamic transition occurs, the liquid to be heated
enters the heat exchanger at temperature T,, and progresses
to T,,- at which time spontaneous thermodynamic transition occurs.
If trie heated liquid which has been converted into a vapor
exits the heat exchanger during or'after thermodynamic transi-
tion without further heating, this fluid would thus exit at
temperature T,5. If, however, the heated fluid remains in
the heat exchanger after thermodynamic transition has occurred,
then the vapor becomes superheated and exits at temperature T,.,.
As stated previously, a limiting factor in conventional heat
exchangers is the temperature difference between the heating
and heated liquids. As this temperature difference becomes
smaller, a greater surface area is needed to transfer a specific
quantity of heat energy. Thus, for a heat exchanger of a given
surface area, the amount of heat that can be transferred is
directly affected by the temperature difference T which is
called the temperature "neck". Referring again to figure 1,
one would like a maximum T,^ or T,, However, the limiting
factor is T . As T grows smaller,"the heat transfer per unit
area becomes less, thus limiting the amount of heated fluid to
reach T,,- or T,.,.
Tradeoffs in Using the Pressure-Staged Heat Exchanger
The thermodynamic transition temperatures of a fluid can be con-
trolled by the pressure to which the fluid is exposed. To pres-
surize a fluid in liquid form requires relatively little pump
work, but a much larger amount is required to compress a vapor.
Thus, a pressure staged heat exchanger requires more pump work
than conventional techniques in which only a fluid in liquid
state is compressed. The tradeoff is between increased pump
work and the savings in heat exchanger surface area.
The pressure-staged heat exchanger can better be appreciated by
studying figures 2 and 3. The fluid to be heated enters the
heat exchanger at temperature T?. and is heated in a section
called a preheater shown in figure 3. At this point, the fluid
is at a pressure lower than the final exit pressure and thus has
a lower thermodynamic transition temperature., As the liquid
raises in temperature to point A, thermodynamic transition occurs
and continues to a predetermined point B. At point B, the fluid
is in a liquid/gaseous state, the percentage of each phase being
a design variable which will be discussed later. At point B,
a staged pump raises the pressure of the heated fluid to the
final exit pressure desired. Because of the increased pres-
sure, the two-phase fluid again increases in liquid content and
enters thermodynamic transition at C. Thermodynamic transition
continues until the heated fluid is all vapor, at E. At point
E, all of the fluid has been converted to a vapor state and the
temperature again begins to rise as superheated vapor is produced,
The heated fluid exits the heat exchanger at temperature T,,-..
982
-------
The dotted line A-C in figure 2 represents the temperature
profile for the heating of a fluid which undergoes thermodynamic
transition carried out in a conventional heat exchanger, i.e.,
without multiple evaporators and a staged pump. In order to ap-
preciate the advantages of the pressure-staged heat exchanger,
one need only compare the differences between necks A/AO, C/CO
and T^. The necks are orders of magnitude larger than the single
neck of the prior art and, thus, the heat transfer achieved is
much greater than in conventional heat exchangers.
Operation of a Two-Evaporator Pressure-Staged Heat Exchanger
Among the variables which can be used to determine the overall
characteristics of the pressure staged heat exchanger
is the fluid quality, hereafter referred to as Z,, which is the
percentage of liquid which has been converted to a vapor in the
first fluid evaporator before the staged pump acts to increase
the pressure of the heated fluid. When Z, is between approxi-
mately 0 and 10% or within the range of approximately 85 to 100%,
a single pump can adequately be used to pressurize the heated
fluid. However, when Z, is in the range of approximately 10 to
85%, it has been found that the liquid-vapor mixture is difficult
to compress using a single stage pump. The pump generally suf-
fers from "cavitation" which is a phenomenon which occurs when
the bubbles of vapor within the liquid collapse. If the pressure
ratio is not particularly high, a standard positive displacement
pump can be used. However, the cavitation problem can also be
greatly reduced by using a liquid-vapor separator followed by
separate pumps to compress different fractions of the liquid-
vapor mixture. Once the separate fractions are compressed, they
arc remixed before being added to the next evaporator.
Such a configuration is shown in figure 4 wherein heated fluid S
enters primary pump 40 and travels through coils 45. Heated
fluid S travels through the preheat section and then enters evap-
orator 1 at which time the fluid enters into a state of thermo-
dynamic transition. Once the fluid has partially vaporized, it
enters liquid-vapor separator 41 at which time the liquid is
pumped separately through stage pump 42 while the vapor is drawn
off and pumped through stage pump 43. Once each component has
been compressed to the desired pressure, the fluids are remixed
in mixer 44 and passed into the second evaporator wherein a
second thermodynamic transition occurs. Upon exit from the se-
cond evaporator, the heated fluid is now entirely vaporous and
is superheated in the superheat section and exits from the heat
exchanger at T.
The diagram in figure 5 depicts a further advantage of the pres-
sure staged heat exchanger. Naturally, one would seek to maxi-
mize the final exit temperature of the heated fluid and thus
would strive to achieve a maximum T,... Once the temperature of
the heating fluid T31-T32 is set, a temperature profile of the
983
-------
heated fluid cannot rise above the heating fluid temperature and
thus the temperature T^ is limited under conventional heat ex-
changer designs. The dotted line in figure 5 shows that under
conventional techniques, if one were to start with a fluid temp-
erature T.,. and end at a temperature T.,.,, an impossible situation
would occur in which the "neck" temperature T" would be negative
(i.e., the temperature of the heated fluid wou?d be greater than
that of the heating fluid). This violates the second law of
thermodynamics which prevents heat flowing from a low temperature
source to a high temperature source spontaneously. This is pre-
vented by the use of the design of this new heat exchanger which
uses a multi-evaporator system separated by a staged pump, the
profile which is shown by solid lines D1 -A1 -B' -C1 -E1.
Application of Multiple Evaporators
Another variable is the use of multiple e'vaporators. For example,
figure 6 shows a temperature profile employing three evaporators
and two stage pumps. Under conventional systems, the heated fluid
would follow the temperature profile shown by the dotted line
which results in a "neck" of T'?t. However,by employing a triple
evaporator system, the heated fluid would preheat in sections
D"-A", enter transition between A"-B", be compressed at B"-C",
enter second phase transition at C"-E", be recompressed by a
second stage pump at E"-F", enter a third phase transition at
F"-G" and exit the exchanger at T43 A number of temperature
"necks" are formed at A"-A°", C"-C°"and F"-F ". One can see
by this figure that the "necks" are greatly increased over T ''',
the "neck" of a conventional system. Thus, the log-mean tempera-
ture difference is increased and the heat transition rate is im-
proved.
Use of Superheated Steam to Drive the Stage Pumps
Figure 7 shows a further modification of the present system.
Schematically, heated fluid enters primary pump 71 and passes
through heating coils 73 in the preheater section. The temper-
ature of the fluid increases .until evaporator 1 is reached, at
which time thermodynamic transition occurs and the fluid partially
vaporizes. Instead of simply increasing the pressure of the
heated fluid and causing the fluid to immediately enter the
second evaporator, the fluid is separated into its liquid and
vapor states in order to minimize pump work. As stated previous-
ly, this is particularly advantageous when the fluid has been
converted into a vapor state such that the fluid contains be-
tween approximately 10 and 85% vapor. Thus, the liquid phase
is fed into stage pump 76 while the vapor phase is pumped through
stage pump 75. Both phases are then mixed in mixer 78 and fed
into the second evaporator section. The pressure within the
second evaporator can be controlled by means of throttle valve
79 in order to gain further flexibility within the system.
984
-------
Upon exit from the second evaporator, the fluid, now entirely
in a vapor state, is superheated in the final section of the
exchanger. At this point, the majority of the superheated
vapor is drawn off at Y although a quantity of such vapor can
be bled by means of throttle valve 80 and fed into turbine 77
which can drive stage pumps 75,76. In this way, much of the pumping
work can be performed by the latent heat of condensation of the
heated fluid. Once the heat of condensation is exhausted within
turbine 77, the liquid can be drained and fed into preregenerator
70 together with make-up fluid 72. This has the further advantage
of preheating the entering fluid.
EXAMPLE OF THE APPLICATION OF A PRESSURE-STAGED HEAT EXCHANGER
A waste heat boiler is employed where hot gases consisting
mostly of air and petroleum combustion products at one at-
mosphere pressure were employed to heat water from an arbi-
trary starting temperature of 59 F to superheated steam at
high pressure. For the purposes of these calculations, the
heating gases were assumed to have a flow rate of 100 Ib/sec
and a specific heat at constant pressure of 0.25 Btu/°F/lb
on the average during the entire heat exchanging process.
Water, being the fluid to be heated, is assumed to have a
specific heat of 1 Btu/lb/ F. It is assumed that the average 2
heat transfer coefficient within the boiler is 20 Btu/ F/hr/ft
which is a realistic value governed by the gas coefficient of
the air-petroleum gas mixture.
The water at 59°F enters the heat exchanger precompressed to
a certain pressure below the final exit pressure. After the
water is boiled to a quality Z,, the mixture of vapor and liquid
is compressed again to a final pressure and quality Z~- The
ratio of the final pressure to the precompressed pressure, R,
together with the first thermodynamic transition temperature,
specific heat ratio A and Z, are design variables.
The steam's final temperature is chosen as a required condi-
tion as.the temperature is important for steam turbine opera-
tion and various chemical processes. The amount of steam that
can be generated is calculated as a function of the "neck"
temperature T . The steam flow rate M_ is then a direct measure
of the amountnof heat being recovered. In this example, the
hot gas temperature at the heat exchanger inlet is 950 F,
and the steam is assumed to be 900 F at a pressure of 400psia.
Reference 1 describes the analytical equations used to compute
heat exchanger performance. The calculation process is a
standard one and is not repeated here.
Assuming a 50°F temperature differential at the neck (T^ in
985
-------
figure 1), a conventional heat exchanger is required to have a
surface area of 18,443 ft2 and will generate steam at a rate of
10.88 Ib/sec. The average heat flux is 3040 Btu/ft2/hr.
The same problem is calculated parametrically for a two evapor-
ator pressure-staged heat exchanger using stage pressure ratios,
R, of 2, 4 and 8 and varying the quality at the staging point, Z ,
from 0 to 1. The results are shown in figure 8, and the advant-
ages of the system can be summarized as follows:
1. The use of the staged evaporative heat transfer system
results in a significant reduction in heat transfer surface area
because the constraint of the apparent "neck" temperature is
removed.
2. High values of Zi and high compression ratios R give max-
imum heat flux values; that is, greater reductions in heat
transfer area or equipment costs.
3. At higher values of R, ther are regions of Z-^ where the
heat exchanger cannot operate because of a negative "neck" and
sometimes the mixture cannot reach the boiling temperature at
final pressure. This region is labeled "forbidden zone" on
figure 8.
One can see that using a "neck" of 50°F with an M_ of 10.88 Ib/sec,
a conventional heat exchanger-having a Z, equal to 0 would have
a heat flux of 3040 Btu/hr/ft . By using the pressure-staged
heat exchanger in which Z-, could be selected, to .95, the heat
flux would be in the vicinity of 4000 Btu/hr/ft . Thus, by use
of a staged pressure heat exchanger, the heat transfer area
can be reduced by 25% as compared to a conventional heat exchanger
while yielding the same energy transfer.
Similar results are shown in figure 9 for a neck temperature of
-20°F. This, of course, is a fictitious condition for a con-
ventional heat exchanger, but has meaning for a properly designed
pressure-staged heat exchanger. Results from figure 8 are re-
produced on figure 9 to show the effect of reducing the neck
temperature. The heat flux is reduced because of the smaller
temperature differences. However, the neck occurs at a lower
absolute temperature which allows an increase in the steam
rate from 10.88 Ib/sec to 12.56 lb/ sec - an increase in energy
recovery of 15%.
Figure 10 was generated in a similar manner except "*fc&at the
steam pressure was dropped to 100 psia. This condition corres-
ponds to the typical operation of a heating plant. One can see
by comparing figure 10 to figure 9 that the graphs are quite
similar except that the "forbidden zone" of figure 10 is some-
986
-------
what narrower. Also, the effects of the compression ratios are
not as large for a large "neck" as it is when the "neck" is
small or negative.
COST EFFECTIVENESS OF THE PRESSURE-STAGED HEAT EXCHANGER
The advantages of the staged counterflow heat exchanger are
four-fold. First, its use results in cost reductions by re-
ducing the surface area of the heat exchanger. Second, one can
achieve the highest possible temperature in the heat receiving
fluid so that the equipment associated with the system can
be designed more efficiently. Third, energy requirements are
reduced which, in turn, saves operating costs. Fourth, the
weight can be furthur reduced by using thinner walls in the
pre-heater and low temperature evaporator sections within the
bounds of the ASME Boiler Code.
In order to dramatize the actual savings, a "Figure of Merit",
Y[f has been devised. This Figure of Merit can best be appreci-
ated by citing actual estimated cost savings. Generally,
boilers cost in the range of $5.00 to $10.00 per square foot
of surface area. The pump, compressor and accessories are
estimated to cost between $10.00 and $30.00 per pumping horse-
power depending on the value of Z, . The Figure of Merit is
defined as the surface Area AO without staging minus the surface
area A with staging times C, , the cost/ft of the heat exchanger,
minus the pump costs expressed in horsepower, M->W, times the
cost per horsepower, C2 - The difference is divided by the
surface area times cost Cl without staging.
A)CI - M2WC2
A0Ci
Thus, r^ is really a fraction which is achieved by subtracting
the pump cost from the cost difference between a heat exchanger
without and with staging divided by the cost of a heat exchanger
without staging. Thus, the greater this fraction, the greater
are the economies of using a pressure-staged heat exchanger.
In order to present the fairest comparison, figures were chosen
which would least point out the advantages of this heat exchanger,
For example, C^ was chosen at $5.00/ft2 and C2 at $30.00/hp. The
Figure of Merit in terms of capital cost for compression ratios
of 2 and 4 are shown in figure 11. The greatest advantage occurs
when Zi is between 0.2 and 0.4.
To optimize a pressure-staged heat exchanger, operation in the
"negative neck" region is preferred. Although a mathematical
comparison between the pressure staged heat exchanger and one
of conventional design can be made, in actuality a conventiona
heat exchanger cannot operate in a negative neck area. If a
negative neck temperature of -20°F is chosen, Figures of Merit
987
-------
for pressure ratios of 2 and 4 are shown in figure 12. For
a compression ratio of 4, the synthesized Figure of Merit has
a peak at Z± between 0.5 and 0.7. At a compression ratio of
2, the Figure of Merit increases with Z]_. Thus, design para-
meter selection indicates that complete evaporation should be
employed at low compression ratios.
CONCLUDING REMARKS
The pressure-staged heat exchanger is a novel method to improve
the performance of heat exchangers where the cool fluid being
heated undergoes a phase change. The advantages can be realized
either as reduced surface area for the heat exchanger or as an
increase in the mass flow of the fluid being heated.
The pressure staging of the evaporative process requires an
increase in mechanical complexity with the addition of pumps,
separators and mixers. In addition, the work input to the
pumps must be evaluated. However, even with this complexity,
the pressure staged heat exchanger appears to be cost effective
for many thermodynamic conditions in light of potential perfor-
mance gains and weight reductions.
REFERENCES
1. Anon, Pressure Staged Heat Exchanger, U.S. Patent No.4,072,182,
7 February 1978.
988
-------
w
rt
p
u
£X
W
E-"
LENGTH
Figure 1 Temperature Profile for a Conventional Heat Exchanger
989
-------
in
30^
n
39
H
SUPERHEATER
EVAPORATOR 2
EVAPORATOR I
PREHEATER
Figure 3 Mechanical Concept of a Pressure-Staged Heat Exchanaer
having 2 Evaporators
W
PH
§
H
LENGTH
Figure 2 Temperature Profile for a Pressure-Staged Heat Exchanger
having 2 Evaporators
990
-------
,R'
42
4
rL~i'-
T
43-fe
rV
i. ^
.j ^
41
» w
;
C
}
)
C
^
c
")
^)
c
^)
c
}
c
SUPER}
EVAPOR
EVAPOR
PREI
•IEATER
ATOR 2
ATOR I
1EATER
5
Figure 4 Concept for Separating and Pumping Liquid and Vapoi
in a Pressure-Staged Heat Exchanger
991
-------
LENGTH
Figure 5 Avoiding the Limiting Conditon of the Neck through
the use of a Pressure-Staged Heat Exchanger
992
-------
w
a
H
Oi
I
EH
LENGTH
Figure 6 Multiple Evaporators in a Pressure-Staged Heat Exchanger
993
-------
78
76
77
•s\
75)
fl
C
C
C
73
7
D
3
3
3
3
Y
SUPERHEATER
EVAPORATOR 2
EVAPORATOR
PREHEATER
Figure 7 Use of Superheated Steam to Drive the Stage Pumps
in a Pressure-Staged Heat Exchanger
994
-------
Ol
6.OOO
5,000
Q/A
4,000
B.TU.
HRrFT.2
3,000
2,000
1,000
ZONE
- R=2
----- R=4
___ D - Q
r\ -o
GAS SIDE
950°F, AThot=50°F
STEAM PRESSURE.400 PSIA
FORBIDDEN
ZONE
0
4 6 fl
z, STEAM QUALITY"
I R
1.0
Figure 8 Increase in the Heat Flux Available with a Pressure-Staged Heat Exchanaer
Steam Pressure = 400 psia anye-t
-------
6,000
5,000
Q/A
4,000
B.T.U.
HR-FT.2
3,000
2,000
1,000
ZONE
R=2
GAS SIDE
= 950°F, AThot=50°F
STEAM PRESSURE,400 PSIA
FORBIDDEN
ZONE
= 10.88
M2= 12.56
FORBIDDEN
ZONE
0
.4 .6 .8
Z| STEAM QUALITY
1.0
Figure 9 Increase in Steam Mass Flow Available with a Pressure-Staged Heat Exchanger
Steam Pressure = 400 psia
-------
6.0OO
5,000
Q/A
B.T.U.
HR.-FT.'
4,000
3,000
2,000
FOR BID DEN ZONE
GAS SIDE
°
T0 =950F,
STEAM PRESSURE, 100 PSIA
= 50°F
FORBIDDEN
ZONE
1,0005—
FORBIDDEN
ZONE
= 13.58
LB.
.4 .6 .8
Z. STEAM QUALITY
Figure 10 Increase in Heat Flus and/or Steam Mass Flow Available with a Pressure-Staqed
Heat Exchanger - Steam Pressure = 100 psia
-------
.Surface Area. ,CostA _ .Pump . .Cost/.
Reduction ' Area Power Power
.Conventional Heat, .Cost/.
Exchanger Area Area
x 100
CO
.2
GAS 950°F, ATHOT=50°F
STEAM 400 PSIA,
ATN=50°F
AQ=I8,443 FT/
C; =S5/FT2
C =$30/H.P
.4 .6
STEAM QUALITY
1.0
Figure 11 Cost Effectiveness of the Pressure-Staged Heat Exchanger
-------
.Surface Area, ,Cost/. _ ,Pun\p . .Cost/.
Reduction ' \Area ^Power Power
,Conventional Heat, ,CostA
Exchanger Area Area
x 100
20
15
10
o/
/o
GAS 950°F, ATHOT=50°F
STEAM 400 PSIA,
LB.
FORBIDDEN
ZONE
LB.
A0= 34,716 FT/
C|= S5/FT.2
.4 .6 8
STEAM QUALITY
1.0
Figure 12 Cost Effectiveness of the Pressure-Staged Heat Exchanger
-------
HEAT RECOVERY FROM WASTE FUEL
Y. H. Kiang
Trane Thermal Company
Conshohocken, Pa. 19428, U.S.A.
ABSTRACT
The attention of industry has been focused on fuel shortages and the high
costs of available fuels. Recovery of available energy from sources once
considered as only waste is in practice in many plants and processes
today. Industrial wastes which have fuel value can be in any form -
solid, liquid or gaseous. Many presentations and discussions have
centered the utilization of heat available from solid waste materials.
In this paper, the possibility of recovering heat from waste liquid and
gaseous materials will be discussed. This paper will present the problems
of handling these various wastes, combustion equipment, and the effect of
waste properties on combustion and heat transfer. Case histories of
installations where systems have been applied in industry to recover waste
fuel value will also be presented.
1000
-------
HEAT RECOVERY FROM WASTE FUEL
by Yen-Hsiung Kiang, Trane Thermal Company, Pa. 19428, U.S.A.
Industrial waste materials which have fuel values are defined as the waste
fuel. This paper will discuss the heat recovery from liquid and gaseous
waste fuels.
LIQUID WASTE FUEL
In many types of process plants, whether they be chemical, petrochemical,
metallurgical, automotive, paper, food, pharmaceutical, etc., there are
liquid wastes generated that contain heating values. The ultimate
solution to these waste problems is combustion.
The following are some of the problems involved in the combustion of
liquid waste fuels:
(1) Low heat of combustion: High water content or high
ash and halogen content makes a waste liquid less
liable to sustain combustion in conventional burners.
(2) High viscosity or mixture of solid particles: These
factors adversely affect the atomizing of the liquid.
A proper selection of injector is required to ensure
trouble free operation.
(3) Polymerization or decomposition: In some cases, the
waste undergoes polymerization in the pipe line or
in the nozzle before atomization. Some times,
thermal decomposition takes place and corrosive
substances are formed. This can be corrected by
properly designing the piping and injectors.
(4) Contaminated combustion product gases: The
contaminants in the fuel will become contaminants
in the combustion product gases. Properties of
the product gases will determine the selection
of heat recovery equipment. A pollution control
system is also required to ensure clean exhaust
gases.
In order to ensure an optimal combustion-heat recovery system, certain data
has to be generated on the waste fuels. They are:
1. Chemical Composition
2. Heat of Combustion
1001
-------
3. Viscosity
4. Corrosive problems to be considered for pumps,
piping, valves and injectors.
5. Chemical reaction with other compounds (e.g.
steam to waste reaction in the injector).
6. Polymerization
7. Solid cortent (solids tend to plug valves,
orifices, etc. in the piping system).
8. Ash reaction with refractories at high
temperatures.
9- Slag formation (its reaction with plugging
of tubes).
10. Analysis of combustion gases and their
effect on heat exchange surfaces.
11. Nitrogen composition (NOX formation)
These are critical data to be reviewed. Typical application of some waste
fuels is shown in Table 1.
Liquid Waste Fuel Injector - In liquid waste fuel combustion, the atomizers
used to inject waste into the combustion zone are critical equipment. For
relatively clean waste fuel, a conventional burner injector can be used^'^)
For high viscosity or highly undissolved solid content liquids, specially
designed injectors are required. The TEAT atomized tip, Figure la, as
developed at Trane Thermal Company has been used successfully for this
type of application. This design operates at low pressures, thereby
avoiding the problems of high pressure pumping. These nozzles have been
used on materials with viscosities as high as 4500 ssu at 300°F. For
two non-compatible waste fuels, a dual liquid TEAT atomizer can be used,
as illustrated in Figure Ib. The TEAT design generates a solid cone.
Combustion rates slow down because of poor fuel air mizing in the center
fuel mass. Increased residence time or a high intensity burner is
usually required for a TEAT nozzle application.
The heat atomizer is another externally atomized tip developed by the
Trane Thermal Company. Figure Ic illustrates the schematics of the
heat tip. The spray generated by heat nozzle is hollow cone, improving
fuel air mixing.
Combustion Equipment - Liquid waste fuels, in general, do not combust
efficiently. A special burner is needed to increase the combustion
1002
-------
efficiency. The requirement of the special burner is high heat release.
The Vortex burner developed by Trane Thermal Company, Figure 2, has been
used effectively in waste liquid combustion.
In the vortex design, waste fuel is introduced through a nozzle at the
centerline of the burner. Combustion air is brought in tangentially and
passes through swirl vanes which impart rotational energy. A twisting,
high velocity vortex action results in complete mixing with the fuel spray
at its point of injection, and at the same time creates a low pressure
zone immediately downstream from this point. As the highly turbulent
air-fuel mixture eaters the flame zone, the low pressure area causes; a
recycling of the hot gases of combustion back into the mixture. Thus,
the mixture is preheated, vaporized and raised to ignition temperature
almost instantaneously. The flame rotates tangentially within the
combustion chamber. This high intensity combustion allows the combus-
tion chamber to be considered as a reaction chamber.
Flame length is short, about one to one-and-a-half times the chamber
length, with heat release rates upwards of a million BTU per hour per
cubic foot in the standard unit. This vortex action provides most
efficient oxidation reaction for waste disposal.
The Vortex burner can be used to burn waste fuels with heating values
4500 Btu/lb. and up. For waste fuels having heating values lower than
4500 Btu/lb., two-stage combustion is necessary. The two-stage combus-
tion equipment can be either a modified Vortex burner or a standard two-
stage combustion equipment. Details are illustrated in Reference 2.
Secondary Chamber - It is important in any conversion to a waste fuel
fired burner that a proper review of the burner location be made. Its
relationship to tube surface is most important. One must be sure that
the waste is first completely oxidized to its final products and that
there is no chance of unburned materials getting into the stack and
exhausting to atmosphere. With some slower burning waste fuels, increased
residence time is necessary. A secondary combustion chamber is usually
required prior to entry into the heat exchange device. The secondary
chamber will ensure the complete oxidation. This will prevent the deposit
of unburned hydrocarbons which could condense and attack the heat ex-
changer surfaces.
Combustion and Heat Transfer - In order to illustrate and analyze the
effect of w?.ste fuels, two waste fuels - as listed in Table 2 - are used
for study.(3) NO. 2 oil is used as the reference.
The data presented in Table 3 are the fuel composition, stoichiometric
products and heat transfer coefficients for these three fuels. The
emissivity values of the various fuel .products of combustion are related
to the water vapor, carbon dioxide values and the gas temperature. The
radiation heat transfer coefficients are determined for the fuels tabulated.
1003
-------
The mass flow of products of combustion per million Btu of heat release
generally increases in value as the heating value drops off.
The decrease in combustion temperature as the excess air rate increases
is illustrated in Figure 3. The gas emissivities and radiation heat
transfer coefficients as a function of gas temperature (thus, excess air)
are illustrated in Figures 4 and 5. At the same temperature, methanol
has the highest radiation heat transfer coefficients and No. 2 oil the
lowest. The convective heat transfer coefficients are shown in Figure 6.
The convective heat transfer coefficients are lowest for methanol and
highest for No. 2 oil.
The data used in this section are limited to the special geometrical
configurations (identical for all fuels), temperature, fuels, and other
parameters of the cases examined. This can only be used as a general
guide line for waste fuel application. It is advisable to study the
theoretical combustion and heat transfer analysis before the designing
of a system.
Discussion
Before any waste material is burned in a heat recovery unit, it is
recommended that a test run be made to determine the composition of the
products of combustion and if a particulate problem exists. This infor-
mation is necessary so that the designer of the heat exchanger will be
able to determine the effectiveness of the surface, and also if any
problems will exist in fouling of the surface. This will also indicate
whether any clean-up equipment is necessary prior to discharge to
atmosphere.
The physical and chemical properties are also necessary for the designer.
The physical properties are necessary to design waste fuel handling
systems. The chemical properties are the key to a successful combustion-
heat transfer system. Besides heat transfer, the selection of equipment
is determined by waste chemistry. ^>->»") One example is the chlorinated
hydrocarbons. A gas to gas heat exchanger cannot be used and special
characteristics must be built into the boiler design. References 4 to 6
give illustrations of the selection of heat transfer equipment as the
waste chemistry changes.
Another problem often encountered in the waste fuel application is the
changing of composition, heating values, etc. It is necessary to provide
a day tank to mix the wastes so that the change in composition is
gradual and the control scheme can compensate for the gradual changes.
GASEOUS WASTE FUEL
In many areas of the process industries, gases from the process must be
vented in order to:
1004
-------
1. Prevent pressure build-up in the system
2. Purge undesirable constituents in the reaction
3. Purge a vessel of residual products after
emptying the vessel.
If the gases contain combustibles, combustion may be used to recover waste
heat.
Combustion Equipment - The Vortex burner, described before, has been used
successfully to burn waste gases with a heating value as low as 100 Btu/CF.
Combustion & Heat Transfer - The gases used for comparision are listed in
Table 4. The properties of the stoichiometric combustion products are
shown in Table 5. The combustion temperature, emissivity, radiation and
convective heat transfer coefficients are illustrated in Figures 7, 8, 9,
and 10.
Case History('»") - A case in point is a waste gas with the following
composition by volume:
C02 - 0.9 percent
02 - 0.2 percent
H2 - 26.2 percent
CO - 5.3 percent
CH4 - 0.4 percent
N£ - 67 percent
This waste gas is defined as WAG in Tables 5 and 6.
The average heating value varies between 88 and 100 Btu/cu.ft. This value
is 10 percent of that for natural gas. The available heat from this vent
gas is on the order of 75 MM Btu/hr. In this particular process, 40 MM
Btu/hr. was needed for gas preheating. A test burner was set up to
determine whether there would be problems in burning this waste gas in a
standard combustion chamber configuration. One of these problems was the
high mass flow of combustion products that would result from the combustion
reaction at a specific heat release.
Since the waste gas contained 67 percent nitrogen, this acts as a diluent
and at the same time increases the total nitrogen in the combustion
products. The fuel-air ratio, in this case, is approximately 1 part of
air to 1 part of fuel. Normally, when burning natural gas, 10 parts of
air is required for 1 part of natural gas. The exact requirements for
this waste gas is covered in Table 5. The waste gas has almost 150
percent as much product resulting from the combustion reaction as compared
with natural gas. This increase in mass flow would have to be reviewed
for both heat transfer and pressure drop in an existing heater design.
1005
-------
A heat transfer analysis, in this case, indicated that the waste gas could
be burned in an existing heater design without causing any problems from a
pressure drop or overall heat transfer design. In fact, it provided an
additional margin of safety from a temperature standpoint. With natural
gas, the maximum flame temperature that could be reached at theoretical
conditions would be 3450°F. In this case, the maximum flame temperature
possible was 2412°F. This was a benefit since in this particular appli-
cation, a heat sensitive material was flowing inside the tubes of the heat
exchanger and tube metal temperature was critical. The lower temperature
level of the combustion products added an additional margin of safety.
Burner tests were run with a burner 25 percent c2 the size necessary in
the full scale unit. These tests were necessary to determine optimum gun
size for the waste gas, optimum gun position and optimum combustion chamber
dimensions. It was found that the gun position was critical to prevent
flashback and also to insure uniform mixing of the waste fuel with the
combustion air. It was also determined that if this waste were injected
directly into the burner throat section without thorough mixing with
combustion air, a rumble or vibrating resulted. This was due to fuel being
injected into a zone of combustion products deficient in oxygen. Proper
mixing of both the combustion gas and oxygen is necessary to provide smooth
burning without vibrations and instability.
Due to the lower flame temperature, a decrease in the combustion chamber
diameter could be made without deterioration of the burner efficiency.
The smaller volume aided in having the reaction temperature very close to
the theoretical flame temperature. At the same time, this provides more
complete combustion in the reduced combustion chamber volume.
These modifications were made to the full size burner (LV-24) and installed
in three units. These units have been operating in this application since
1964. These units operate continuously, 24 hours a day, 7 days a week, and
are shut down once per year for maintenance turnaround. By using this waste
gas, approximately 75,000 standard cu.ft./hr. of natural gas is saved and
over a ten year period has resulted in a saving of $3.6 Million (based on
average cost of 0.60/1000 cu.ft.). The only natural gas that is used in
this particular heater is the operation of a constant pilot which requires
approximately 250 SCFH of natural gas. This is used for safety reasons in
the event that there is an interruption in the flow of waste gas or instan-
taneous drop of the hydrogen content in the waste gas. The pilot will
permit re-ignition if this occurs, prior to shutdown by the flame-out
controls.
The waste gas is burned in three separate fired heaters. The units are
constructed as shown in Figure 11. The burner fires directly into a
secondary chamber where the gases are tempered to 1600°F. This is
necessary due to the critical tube metal temperature limitations and
maximum heat flux permitted on the material being heated within the tubes;
however, in another application this has been fired directly at a flame
1006
-------
temperature of 2200°F. The short flame characteristics of the vortex
combustor permits a short mixing chamber to be used to dilute these
products from flame temperature to design temperature. Normally, a
chamber five to six times this size would be required with a conventional
burner system to enclose the flame which might be as long as 20 ft. at a
firing rate of 24 MM Btu/hr. In this case, the flame was approximately
5 ft. long downstream of the burner exit. Complete combustion in the
burner also permitted the unit to be installed adjacent to the heat
transfer surface without fear of high radiant losses causing incomplete
combustion of the waste gas. This also saved in the total installed cost
of the burner system necessary for these waste gases.
Discussion - If a fume contains at least 16 percent oxygen, it may be
used as combustion air for the main burner. A savings of 745 SCFH of
natural gas per thousand SCFM of fume may be realized over the use of
outside combustion air. This should be reviewed carefully to determine
what problems may result. Some fumes contain condensible materials which
could deposit on blower wheels, control valves, and burner internals.
Figure 12 shows schematics of typical fume systems.
One problem associated with gaseous waste fuel is the cyclic properties
of both flow rates and composition. A well designed control system is
required to ensure trouble free operation. The control system usually
keeps the oxidation temperature and stack oxygen content relatively
constant by adjusting the flow of air and auxiliary fuel. An alternate
approach is to base load the system with auxiliary fuel. The latter
approach usually gives a more stable system.
REFERENCES
1. Santoleri, J.J. , "Spray Nozzle Selection", CEP, ^,9 ,p.84,1974.
2. Kiang, Y.H., "Total Hazardous Waste Disposal Through Combustion",
Industrial Heating, December 1977.
3. Ashburn, L., "Techniques of Energy Recovery from the Combustion
of Low Heating Value Fuels and Industrial Fluid Wastes",
presented to Ass. of Iron and Steel Engineer 1977 Convention,
Cleveland, 1977.
4. Hung, W., "Results of a Five Tube Test Boiler in Flue Gas with
Hydrogen Chloride and Fly Ash", ASME WAM, Houston, Nov. 1975-
5. Kiang, Y.H., "Prevent Shell Corrosion for Chlorinated Hydrocarbon
Incineration", Presented to Seminar on Corrosion Problems in
Air Pollution Control Equipment, sponsored by AQCA, IGCZ, and
NACE, Atlanta, Jan. 1978.
6. Kiang, Y.H., "Technology for the Utilization of Waste Energy",
presented to IEC 23rd Annual Meeting, L.A., April 1977.
7. Santoleri, J.J., "Energy Recovery from Low Heating Value
Industrial Waste", presented to ASME Industrial Power Conference,
Pittsburgh, May 1975.
8. Santoleri, J.J., "Waste Energy - The New Source of Plant Profits",
presented to AIChE 85th Annual Meeting, Philadelphia, June 1978.
1007
-------
ACKNOWLEDGMENT
The author wants to acknowledge his gratitude to Mr. J. J. Santoleri and
Len Ashburn of Trane Thermal Company for providing reference materials
and assistance in preparing this manuscript.
TUU 1
UOBU wm nti. 4fn.ic.vn™
Vma f»l
1. MMkylMt trlaolaa Tar
2. KmUatva' oceanic aaaaracor Pack.****1 lollar
Cfm* aMCl«ar caaccor
PacrolaM Uaacav
J. Tan. *««. Kt>e.
9. blwM • rtetkfl Scjmn
ftrial lUea, ieu/hi
1 a IB*
1 E 10*
« » 10*
7 « JO*
H M 10*
7.10*
ugau yjjTi ratu
•a. 2 «il
Haita»ol
LOT lu Vaau liquid
l»/lk.
EOT
li.KW
10.100
3.050
un
H.4W
1,078
7,035
Alr/fua
Wt.
14.4
4.52
•nVlk. c
EOj U.UO (1.0
KT 10,140 17.5
Ut, I.OU M.I
uoou mm
jaltioj. Ut. I
HI
11.4
12.1
4.1
1 0, .j
0.5 0 0.1
0 U 0
ttoiehiaoarriB fcoJueta
A »ol. 1 ._H If
F Mj B3! "t° K>UCtt *F
T|i Claao tt^AT,
0 0 14.4 :,.. 11.1 14.4 m
0 0 4.12 41.• 10.4 II.) 719
14.1 0.1 l.M 45.1 11.0 21.4 Ml
I.I.IIY;
•C hr ti
HS< O.OB1 11.) 14.1
WAI 0.1)4 14.4 14.4
1)11 0.11 11.1 11.4
iv* h««t tr«uf«r co«tflcl«ot
TABU 4
cinon van nti
IrWlCT Ita/lk.
H
000
LK
VAC
•n
Vftcvral Caa
C«ka Ov.o Caa
lav liu Co
Vaaca Cu
•laac Funiaca CA*
•n
1.3M
347
'. T6
104
11
HHV
11.970
22,070
2.30^
1.171
1,02)
LKV
:o . :«o
19.650
: ,»;•
1.4)4
1.011
/Ltr/Fml
Ut.
14.56
14.40
l.iT
1.07
O.i»
TtllC 5
c*s««r» HASTE nitL->Taiaiiiintriiic
rvu
K
uc
uc
H.1C
we
HHV
It./ Ik.
11.910
11,010
1.104
l.aJt
1.0 Jl
Co,
CO
0
4.1
24.1
5.1
11.4
0
5M
11.0
,14.1
1.4
91.]
10.)
!.)
0.4
O.I
otnar
4.9
1.1
0.)
0.1
0
"1 a.
1.1 0
1.1 l.t
II. 1 4.1
41.0 0.9
SI.) 11. 4
Si.>lcMuavtrlc fro<
i *«'• « ______
' •) CO] HjO
14.4 11. | 9.1 It.l
14.4 ».0 1.9 U.I
1.17 li.l 14. J 10.)
1.07 11.4 4.0 14.4
0.19 )(.) :4.) 1.1.
luCI.
Hi 11- 'f
I'O 1141
n> MII
91) H1J
1101 1400
lit* !»U
«C
0.1M1
0.091
0.104
0.119
0.141
kr
11.)
11.4
11.1
11.1
9.1
he
11.9
11.4
14.9
.4.1
11.)
1008
-------
Fluid
Fu.1
Fluid
II
fu.1 f 2
fittl
Atonlzlng
Fluid
• 3000
• Jto°
I 1*00
1 1100
uoo
I U«tth: 4 ft.
0.0« 0.10 0.12 0.11 0.16 0.11 O.:i
Cw tal»lvtey
•( UquM
Ixunul Atoaized Tips
MOO
.. am
'.-3000
g 1MO
J UOO
' awo
UOO
1*00
10 i] 20 :s »
Hot Tr«i:ff«r l-.wt I liM.-nt
•r LI..U co^i»cii»
1009
-------
•» tno
I".
e am
JTltOO
J two
2200
2000
ino
1100
cl: MOO lk/Ht-f('
I t I i t
U i» » i3 -0 - 24
feeueffl^e Meat Tran-i(rr Cavffletcnt
C«meU»»t
• I ftM CMbuelCtOH
figure 11
Direct fired heater
BJHC hnt racovcrv schemer with hot exhaust qasta
from incirMramr u«ed to prvneat the incominq ?rocra
funtcs. Temperatures up to 1200 F are achieved.
3-
Hot 4*h«u«t from incinn jtcw pjv%«i ih'Oixt
feCupVJtOf to pfehfJl pdrtTi-M .v.iMt: t.*m.-t jrto 'trt-
throteM|n wcond hc.it exchjit^-f 10 tccuv«f jiiihi.'v
he«t «*h*oh mwy be rvtutnod to p*o< ru or uthif 51 j
Figure 12
Schematic! of fune eyste*
1010
-------
SURFACE-SKIN TEMPERATURE GRADIENTS IN COOLING LAKES
S. S. Lee, S. Sengupta, C. R. Lee
University of Miami
Coral Gables, Florida 33124
The thermal structure at the air-water interface indicates the
direction of heat transfer. Detailed understanding of the tem-
perature profile at the interface is imperative for formulating
boundary conditions and determining the relationship of radio-
metric measurements of surface skin temperature and the bulk
temperature in the mixed region immediately below. When heat
flow is from water to air the skin temperature is cooler than the
bulk temperature. Roll (1965) provides a summary of measurements
of skin temperature anomalies.
THEORETICAL BACKGROUND
The surface heat flux Q is a sum of sensible heat flux from water
to air, flux of latent heat due to evaporation and net flux of
long-wave radiation from water to air. Fig.l taken from Hasse
(1971) shows the individual heat transfer components.
In general environmental flows are turbulent. However, a thin
layer just below the air-water interface may be assumed to be
laminar. This implies that the transport processes in this layer
are molecular. Saunders (196?) through dimensional arguments
derived that
i-VC^)11 (1)
Where 6 is the thickness of the molecular layer, T' is the viscous
stress. v is the kinematic viscosity of water and p is the den-
sity. Saunder's, also assumed that the major part of the differ-
ence between the surface temperature and bulk temperature occurred
in the molecular layer, such that
n K AT (2)
Q ~ — g-
Where K is the thermal conductivity. A combination of (1) and (2)
yields
AT= ,
K(r/pw)
Where T is the surface wind stress and A is a numerical coeffi-
cient.
1011
-------
Hasse (1971) developed a similar relationship from data analysis
and theoretical considerations. For negligible solar radiation
be derived
Am_ o Q(ly/min)
U10(m/sec)
Where C is a parameter varying slowly with 6 and UIQ is winu
speed 10 meters above the surface.
If the wind stress is represented as pC^U n2, where Cn is the
drag coefficient.
K P2CD2 10
Using coefficients provided by Hasse we find that A=8.
Thus the relationships of Hasse and Saunders are quite similar.
Paulson and Parker (1972) discusses these two relations.
It has been conjectured that X is independent of Q and u. The
implication of such an independence is profound since it would
allow computation of Q using only wind speed and AT. Further,
usefullness of this method would be to compute water loss through
evaporation. Solar radiative fluxes are measured directly. The
purpose of this paper is to utilize field measurements taken in
Lake Belews North Carolina to investigate the constancy of A.
This will bridge the gap between data available for laboratory
scale experiments and ocean data.
PREVIOUS STUDIES
Several investigations have been made in the past few years using
field and laboratory data. Table I shows a summary of some of
these studies. Hasse used oceanographic field data for his in-
vestigations. From his data the values of A calculated is about
8. Wind speed varied from 1.45 m/sec to 11.35 m/sec. Saunders
also used oceanographic data and found A to lie between 5-10.
Laboratory experiments by McAllister and Mcleish (1969) indicate
a value of A=4.5 which is considerably smaller than the values
obtained by field measurements. (The values of L in the chart
Indicate the characteristic horizontal length of experimental
basin). Paulson and Parker (1972) conducted laboratory experi-
ments on a relatively small basin with no wave generation. They
varied wind speed from 1.39 m/sec to 3-64 cm/sec. They calculated
A=15 ±1. This is significantly larger than those obtained from
field observations. Hill (1972) conducted a series of experiments
in laboratory scale with and without surface waves. He obtained
for the no wave case a value of A equal to 4 with wave he obtained
A equal to 11. Paulson et al suggests that the variation in A
1012
-------
is primarily caused by the wave state, the effect of surface
waves on X being more pronounced in the laboratory than in
field conditions. This conjecture was emphasized by noting that
capillary waves are more dominant in laboratory water-tables than
in the field. Witting (1972) has conducted an analytical investi-
gation of wave effects on surface temperatures. He concluded
that waves particularly steep capillary waves decrease AT and
thereby A. Waves also reduce the shear stress T thereby -increa-
sing A. However, the first effect dominates especially in la-
boratory experiments since capillary waves are more prevalent in
the laboratory tanks than in the open ocean conditions. Saunders
(1973) provides a review of the existing literature on this topic.
FIELD MEASUREMENTS
The field observations were made at Lake Belews North Carolina
during May 18 and 19, 1976. This is a cooling lake for a ther-
mal power plant of the Duke Power Co. It has two basins connected
by a connecting canal. The smaller basin is well mixed and re-
ceives the heated effluent. The larger basin is seasonally strati-
fied and the intake to the plant is located there. The measure-
ments of skin temperature gradients were made in the smaller pond.
A Barnes model PRT-5 radiometer was used to measure surface tem-
perature. A thermistor attached to a float measured temperatures
approximately 3 centimeters below the surface. Since the thick-
ness of the surface skin layer is of the order of 1 mm, the ther-
mistor reading is the bulk temperature. These readings were made
from a boat. Wind speed, solar insolation, humidity and air-temper-
ature were recorded at a meteorological tower in the larger lake.
Fig.2 shows the configuration of Lake Belews. Details of the
field experiment are provided by Lee et al (1976).
CALCULATION PROCEDURE AND RESULTS
The objective of the present investigation is to calculate Q
using existing formulae and measured meteorological parameters
and obtain values for A and C . The motivation for this is to
establish what effect length scale of the domain has on these
constants. Since studies on oceanographic scales and laboratory
scales have been made, this study on an intermediate length scale
domain is important. This study is also directed to determine
What degree of variation of A and C is observed for a single
basin.
General Description of Heat Flux Across the Air-Water Interface
The net heat transfer through a water surface is composed of
radiation penetrating the water surface from above, radiation
out of the water surface, evaporation, and conduction transfer.
1013
-------
These are indicated schematically in the Pig,3.
Qnet=
=^sn + Qan- (Qbr 4 «e ± Qc>
Where Q = net absorbed solar radiation
s n
Q0~ = net absorbed atmospheric radiation
CLi 1
Each of these terms is discussed below.
1. Short-Wave Solar Radiation
The total incident solar radiation impinging on the water
surface QSJ_ was recorded by pyriheliometer at the weather
island at Lake Belews, operated by Duke Power Company.
Dake and Harleman (1969) estimate that about 40$ of the
solar radiation arriving at the air-water interface is ab-
sorbed in a thin layer of water at the surface, and the
reflected solar radiation is typically 6% of incident solar
radiation. Hence the net solar radiation absorbed by the
water surface is
Qsn = QS - Qsr = °'9^ * 0.4 x Qs±
2. Long-Wave Atmospheric Radiation
Clear sky incident atmospheric radiation, Qacj may be
expressed as,
"13 *
Qac = 1.2 x 10 (Ta )6 (B/ft2 day}
Where Ta* = absolute air temperature ( R)
The presence of clouds tends to increase the average radia-
tion received at the ground from atmosphere. Harleman et al
(1975) recommend an equation of the form
Qa = Qac(l + 0.17C2)
Where Qa = heat flux at the surface. (B/ft2day)
C = fraction of the sky covered by clouds.
A figure of 3% is usually accepted as reflectance of a water
surface to long-wave radiation. Thus the net atmospheric
radiation absorbed by the surface is
Qan = Qa - Qar = °-97 Qa
and, therefore, we have
1014
-------
B
Q = 1.16 x 10" (T., )6 (1 + 0,17C2)
call a.
or
Qan = 1.16 x 3.14 x 10~19(Ta*)6,. (1 + 0.17C2){-^
3. Long-Wave Back Radiation
Harleman et al (1975) note that the emmissivity of a water
surface is independent of,temperature and salt or colloidal
concentrations, and gives a value of 0.97- Thus we obtain
V = °
Where T.-, = water surface temperature ( K)
_i 2 _
a = 1.354 x 10 (cal/cm2 sec K")
4. Evaporation Heat Loss
Following the recommendation of Harleman et al (1975) the
"Lake Hefner" formula is used in this study to estimate
heat loss due to evaporation. The formula is
Qe = f(w)(es - ez)
Where f(w) = wind speed function
e = saturated vapor pressure of air at T=l (mbar)
s
e = vapor pressure at height 2 m above the surface
z (mbar)
c* m
w = wind speed ( /sec)
The wind speed function can be written as
f(w) = 0.9 x 10" W {cal/cm2 sec(^f-) mbar}
To faciliate calculations, the saturated vapor pressure of
air is computed as
e
s
= 0.0435 T2 - 0.0917 T + 7-80
gl
Where T n = water surface temperature ( C)
si
and the unsaturated vapor pressure is
e = e '
z r s
1015
-------
Where cj> = relative humidity
e , = saturated vapor pressure of air at the tempera-
ture existing 2 m above the surface.
5- Conduction Heat Loss
Bowen's ratio is used to estimate the heat flux across the
air-water interface by conduction. The equation is
T . - T .
Qc = 0.639 (e _ a—) Qe {Cal/cm2 sec}
s z
Where T = air temperature at a height of 2 m ( C)
a
Heat Transfer Calculations
The data measured in Lake Belews, North Carolina for this study
are presented in Table II.
According to the formulus described in previous sections, each
of the heat flux components calculated are shown in Table III.
Where Q is negative, which means the heat flux is directed
upward §way from the water surface. This is expected for the
smaller mixing lake since it receives the heated effluent from
the power plant.
The Values of X and GI
The temperature difference AT between the surface and a lower
well mixed region of nearly constant temperature is
Xvp \ Q
AT = r—r^^ according to Saunders (1967)
Kp "2CD^ W
K CD"2 p 2 WAT
or A= £— —Q
v PW^ net
Where X = dimensionless constant
K = thermal conductivity of water
v = kinematic viscosity of water
p = density of air
PW - density of water
AT = Tsl - Ts2
1016
-------
3
with CD = 1.21 x 10" , we have
,-" W'AT
\ = 1.73 x 10
Qnet
Hasse (1971) examined AT as a function of mean wind speed at a
height of 10 meters, and Q , he finds
II GU
AT -_ c Uy
/sec
or C = 1 W (°m/sec) x AT (°C) = 1
1 6000 Q (cal/cm2sec) 1.038
ne u
Using Qnet obtained from above calculations, C. and X are cal-
culated. They are shown in Table IV.
DISCUSSION AND CONCLUSIONS
From Table IV it is seen that the value of A lies between 10.4
and 5-6. The value 10.4 is significantly larger than its near-
est value 8.4 and may be an isolated data point involving error
in observation. Without neglecting this point the average value
for X is 7.1. This value lies between the range calculated from
Saunders data ie 5-10. The value of A calculated in this study
is in agreement with Basse's value of 8. It is considerably
smaller than the Paulson and Parker's calculated value of near
constant 15. The value of 7-1 is also quite different from
Hill's with wave result of A = 4. The value of C, shown in
Table IV is between 5.4 and 10.4. This compares with Basse's
value of 9.4.
The following conclusion can be made from the calculated values
for A.
a). From the calculations of Saunders, Hasse and the pre-
sent study a value of A between 7 and 8 is a reasonable
estimate for field situations.
b). There is no apparent relationship between A and length
scale in field conditions, since oceanographic and
lake data yielded approximately the same range for A.
c). Calculations using laboratory measurements yield much
larger values of A for no wave condition eg. 15. The
effect of waves reduce this value to 4-4.5-
1017
-------
d). The effect of waves is to reduce the value of X with
maximum effect on laboratory scale conditions owing
to predominance of capillary waves.
The direction of investigation presented provides encouraging
results since a near constant value of A implies the net heat
flux may be estimated by measuring AT and wind speed only.
Thus the empirical relations for obtaining Q may be avoided.
It is imperative however to conduct parametric studies both
in laboratory and field length scales to understand further
the variations in X that have been observed to date.
ACKNOWLEDGEMENTS
This work was conducted under funding from National Aeronautic
and Space Administration, Kennedy Space Center.
1018
-------
REFERENCES
1. Dake, J.M.K., and Harleman, D.R.F., "Thermal Stratification
in Lakes: Analytical and Laboratory Studies," Water Resour.
Res., 5(2), 484-495, 1969,
2. Harleman, D.R.F., and Stolzenback, K.D., "Engineering and
Environmental Aspects of Heat Disposal from Power Generation,"
Dept. of Civil Engineering, M.I.T., Jan., 1975.
3. Hasse, L., "The Sea Surface Temperature Deviation and the
Heat Flow at the Sea-Air Interface," Boundary-Layer Meteorol.,
1, 368-379, 1971.
4. Hill, R.H., "Laboratory Measurement of Heat Transfer and Ther-
mal Structure Near an Air-Water Interface," J. Physical
Oceanography, Vol.2, pp.190, 1972.
5. Lee, S.S., Sengupta, S., and Mathavan, S.K., "Three Dimensional
Numerical Model for Lake Belews," Final Report NASA Contract
NAS 10-9005, June, 1977-
6. McAlister, E.D., and McLeish, W,, "Heat Transfer in the Top
Millimeter of the Ocean," J. Geophys. Res., 74, 34o8-34l4,
1969-
7. Paulson, C.A., and Parker, T.W., "Cooling of a Water Surface
by Evaporation, Radiation, and Heat Transfer," J. Geophys.
Res., Vol.77, No.3, pp.491, Jan. 1972.
8. Roll, H.U., "Physics of the Marine Atmosphere," pp.227-247,
Academic, New York, 1965.
9. Saunders, P.M., "The Temperature at the Ocean-Air Interface,"
J. Atmos. Sci., 24, 269-273, 1967-
10. Saunders, P.M., "The Skin Temperature of the Ocean," Contri-
bution No.3148 from the Woods Hole Oceanographic Institution,
1973-
11. Witting, J., "Temperature Fluctuations at an Air-Water Inter-
face Caused by Surface Waves," J. G'eophys. Res., Vol.77, No. 1.8,
pp.3265, June 1972.
1019
-------
TABLE I
CHART SHOWING X FOR DIFFERENT INVESTIGATORS
HASSE
SAUNDERS
PAULSON
PARKER
HILL
MCALISTER
MCLEISH
A
8
5-10
7
15
4
11
4.5
CONDITION
Wind speed 1,45-11.35
m/sec ,v is used at
temperature=15 C
neglect solar radiation
Wind speed > 2 m/sec ,
neglect the divergence
of solar radiation
In middle latitude
winter AT reaches to
1 °C.
Neglect wave generation,
v is used at tempera-
ature=25°C
With wave
. Without wave
Wind speed 4 . 5 /sec
With wave
LENGTH SCALE
Field measurement
Field measurement
Laboratory measure-
ment
L=13 cm
Laboratory measure-
ment
L=90 cm
Laboratory measure-
ment
L=220 cm
1020
-------
TABLE II
LAKE BELEWS DATA
No.
1
2
3
4
5
6
7
8
9
10
11
Station
-. A
A
A
H
G
B
C
D
A
F
F
Date
5/18/76
it
11
it
it
11
ti
11
5/19/76
it
11
Time
8:45
8:52
9:19
10:00
10:40
11:10
11:35
12:30
13:50
14 :00
15:05
T
a
20.6
20.6
21.1
21.1
21.7
21.7
21.7
22.2
19-0
19.1
20.0
•T
a
69
69
70
70
71
71
71
72
66
66
68
T-l
28.9
29.0
29.0
29.0
29.2
28.9
28.8
29.0
27.2
26.1
27.0
os2
29.8
29.8
29.8
30.2
30.2
29.8
29.7
29.8
28.0
26.8
27.9
AT
_ rp rn
si s2
-0.9
-0.8
-0.8
-1.2
-1.0
-0.9
-0.9
-0.8
-0.8
-0.7
-0.9
w
sec
232.4
223-5
245-9
290.6
335-3
393-4
420.2
447-0
662.0
635.7
657-0
1 of \
( /" )
67
67
67
61
61
60
60
59
62
61
61
C
0.7
0.7
0.6
0.4
0>.6
0.6
0.7
0.9
0.3
0.3
0.3
(cal/cm2sec )
7-7xlO~3
8.0
9-0
13-0
12.7
11.0
8.3
8.0
16.8
16.8
15-8
a
tj
H
-------
TABLE III
HEAT FLUX COMPONENTS
No.
1
2
3
4
5
6
7
8
9
10
11
Q xlO3
sn
2.9
3.0
3-4
4.9
4.8
4.1
3.1
3-0
6.3
6.3
6.0
Q xlO3
an
8.6
8.6
8.6
8.4
8.6
8.6
8.8
9.4
8.1
8.1
8.3
Qu. xl°3
^br
10,9
10.9
10,9
10.9
11.0
10.9
10.9
10.9
10.6
10.5
10.6
Q xlO3
5-3
5.1
5.7
6.9
7-9
9.1
9.6
10.3
14.3
12.4
13-5
Q xlO3
c
1.1
1.1
1.1
1.3
1.4
1.6
1.7
1.7
3-1
2.6
2.7
Q xlO3 cal .
net (cm2sec)
-5.8
-5-5
-5.7
-5-8
-6.9
-8.9
-10.3
-10.5
-13-6
-11.1
-12.5
1022
-------
VALUES OF
TABLE IV
AND X CALCULATED FOR LAKE BELEWS
No.
1
2
3
4
5
6
7
8
9
10
11
Qnct x 103(cai }
IltT U A _L U V 2 )
cm sec
-5.8
-5-5
-5.7
-5.8
-6.9
-8.9
-10.3
-10.5
-13-6
-11.1
-12.5
w(cm }
yy i i
^sec
232.4
223.5
245-9
290.6
335.3
393.4
420.2
447.0
662.0
635-7
657.0
AT
-0.9
-0.8
-0.8
-1.2
-1.0
-0.9
-0.9
-0.8
-0.8
-0.7
-0.9
A
6.2
5.6
6.0
10.4
8.4
6.9
6.4
5-9
6.7
6.9
8.2
Cl
6.0
5-4
5.8
10.0
8.1
6.6
6.2
5.7
6.5
6.6
7-9
1023
-------
Solar
Radiation
P
0-
0.001
0.01
: 1.0
L
3 10
100
I0001-
Latent heat transport
Sensible heat transport
Effective back radiation
Molecular transport
v.X
5=6(U)
Increasing influence ot
turbulent transport
Turbulent transport
K
t
Pig.l Schematic Diagram of Heat Flow at
the Sea-Air Interface.
1024
-------
Ash Basin
Stations
(b)
Pig.2 Map of Lake Belews Showing (a) Location of
Meteorological Towers (b) Mixing Pond Station
Locations on May 18, 19 of 1976.
1025
-------
Q =Short-wave Solar Radiation
s
Q =Long-wave Atmospheric Radiation
a
Q, =Long-wave Back Radiation
=Evaporation Heat Loss
Q ^Conduction Heat Loss
O
^Reflected Solar Radiation
sr
ar
=Reflected Atmospheric Radiation
T ^Surface Temp.
i S J_
'T ?=Temp. at 1" Below
Surface
Fig.3 Net Rate at Which Heat Crosses
Water Surface
1026
-------
FOUR THERMAL PLUME MONITORING TECHNIQUES: A COMPARATIVE ASSESSMENT*
ROBERT S. GROVE, RONALD W. PITMAN, AND JACK E. ROBERTSON1
ABSTRACT
Four different methods of monitoring thermal plumes were compared: two from a
vessel and two from an airplane. The study area was the Pacific Ocean offshore of
the 450 MW San Onofre Nuclear Generating Station in southern California. The
ocean provides the once-through cooling water which is discharged through a sub-
merged, single port, twelve foot diameter conduit. Water temperature data were
taken along with other oceanographic and meteorological data on four separate
days, and three of the four different plume monitoring techniques were conducted
simultaneously.
The plume monitoring methods consisted of: 1) an in-hull solid state thermistor
recording surface temperature while the survey vessel traversed the area of the
thermal plume for approximately one hour with vessel position recorded continu-
ously using an electronic range positioning system, 2) an airplane traversing the
thermal field for approximately one hour at an altitude of 1000 feet using a narrow
beam infrared radiometer calibrated by ground truth temperature measurements,
3) an airplane traversing the thermal field for Approximately 15 minutes at an
altitude of 1000 feet using an infrared thermal scanner that photographically
recorded the configuration of the thermal plume, and 4) vertical temperature pro-
files taken from a vessel at pre-determined positions in the area of the thermal
plume over a three to four hour period.
Assessment of the study methods revealed that each had certain advantages de-
pending on what plume characteristics were being determined. Comparison of plume
configuration indicates good general agreement among methods. The infrared scanner
provided the best picture of the surface plume but the least degree of absolute
isotherm definition, while the vertical temperature profiling method provided accurate
absolute temperatures but produced comparatively distorted plume configurations
due to the duration required for monitoring.
*This paper was not presented.
Respectively: Research Engineer, Southern California Edison Company, Rosemead,
California; Project Oceanographer, Brown and CaldweD, Pasadena, California;
and Project Manager, Brown and Caldwell, Pasadena, California.
1027
-------
EXPERIMENTAL RESULTS OF DESTRATIFICATION
BY BUOYANT PLUMES
D. S. Graham
Dept. of Civil Engineering
University of Florida
Gainesville, Florida U.S.A.
ABSTRACT
The effects of ambient stratification upon buoyant plumes have been
studied in detail, but the converse case has received little attention.
A literature review of destratification experiments in the laboratory
and field tends to show a rapid decrease in mixing efficiency of plumes
associated with an apparent change from overall mixing to interfacial
formation and descent (ascent). A rigorous dimensionless scheme for
interpretation of the results of such experiments is given, and an
analogy to the Fourier equation for one of the mixing regimes is out-
lined. Finally, sample results of experiments are presented which show
that two distinctive mixing regimes termed "diffusive" and "interfacial"
can be identified. The former is associated with high Richardson numbers
and the latter with low. The latter is especially pronounced near the
orifice. The point of change from one to the other can be predicted
from dimensionless criteria for the particular experimental geometry used.
INTRODUCTION
Ejection of waste heat to the environment by means of outfall diffusers
(line or source) into lakes, reservoirs and coastal seas can be expected
to continue to increase. While use of evaporative cooling towers is
currently being encouraged by the EPA, once-thru cooling systems usually
have substantially lower cost and cause minimal disturbances to the
atmosphere in warm humid locations like Florida. Usual diffuser locations
are either near the water surface for rapid radiation of heat to the
atmosphere, or at depth for efficient mixing of the plume with the ambient
wa ter.
Many bodies of water into which the outfalls discharge are density-
stratified by temperature, salinity, or both. Stratification may be
temporally intermittent (eg., diurnally, seasonally, or over a portion
of a tidal cycle) or persistent. Two types of stratification commonly
occur-linear and interfacial. The former is characterized by an approx-
imately linear density gradient and has received more intensive study
because several closed-form solutions are possible (see (1) and (24),
for a partial review). The latter type of stratification has a readily
identifiable interface between almost homogeneous masses of water of
different density. The interface location and sharpness are functions
of both environmental conditions and mixing induced by the usually
1028
-------
less dense thermal plume. While plume rise and entrainment properties
have been well studied, interaction of a buoyant plume with ambient
interfacial-type stratification has not received the same thorough
experimental investigation.
The orientation of this study should not be confused with the many excel-
lent studies of the effects of ambient stratification upon plume or jet
behavior (see, for instance (1) , (2), and (24)). For small receiving
bodies, and locally, the plume and ambient stratification are linked to
one another and it is the effect of the plume upon the ambient strati-
fication that is investigated here.
PREVIOUS STUDIES
Studies by Rouse and Dodu (3), Turner (4), and others, which were reviewed
by Turner (5) and Long (6), showed destratification due to interfacial
entrainment without shear to be proportional to a Richardson number based
on overall length scales and a Peclet number based on molecular dif-
fusivity despite the fact that Reynolds numbers were high away from the
interface. The entrainment velocity could be expressed as a power of
the Richardson number (about-1.5 to -1) which was dependent upon the
Pecle£ number for destratification without shear (i.e., all destratifi-
cation accrued from the energy flux divergence term of the turbulent
kinetic energy equation) and a constant (-1) for shear-induced entrain-
ment. Kantha (14) notes that the range of Peclet influence appears to
be dependent upon the Reynolds number. Subsequent experimentation has
been ongoing to better define these processes, but the relevant dimen-
sionless parameters have been identified. Prior to these experiments
most mixing studies by chemical engineers had tried (incorrectly) to
relate mixing time to a Reynolds criterion (see Uhl and Gray (7)).
For the case of a vjet or plume aimed at an interface, the literature
may be divided into several categories - 1) small scale laboratory in-
vestigations, 2) chemical engineering studies using intermediate-sized
containers, and 3) large scale destratification experiments in lakes and
reservoirs. The orientation and utility of each of these groups differs
greatly. The chemical engineer or reservoir manager often wants to know
time until complete mixing, while the fluid mechanics scientist is more
interested in defining entrainment velocities.
Baines (8) describes experiments in which a dense salt plume is allowed
to fall to an interface, but not penetrate it. He found
-3/2
Entrainment flux = const. *(Jet Richardson No.) (1)
Sullivan (9) described similar experiments with the exception that
1) finite quantities of liquid were used, and 2) some cases were forced
plumes. These are reviewed by Brooks (1). Linden (10) shot vortices
of freshwater at a salt-fresh interface and found the depth of maximum
penetration to be inversely proportional to a Froude-type number while
1029
-------
the entrainment rate was proportional to the cube of a Froude number
[or to the -1.5 power of a Richardson number, again].
A series of experiments by Brush, et al., (11), Brush (12) and Neilson
(13) attempted to relate mixing of different scales thru a dimension-
less format. In the more sophisticated 1970 experiment Brush (12) varied
both density difference and jet discharge. His results have some com-
putational (and likely typographical) errors as reported, but after recal-
culation they are presented as Figure 1. It can be seen that dimension-
less entrainment velocity is a function of Richardson, Peclet, and perhaps
Reynolds, effects. At lower Richardson (higher Reynolds and Peclet)
numbers, the molecular effects disappear and the data follow a -1 slope
as energy considerations alone would imply. This is consistent with
many other results (Kantha (14)) . These results are based upon jetting
one layer into the other, and measuring the difference in density to
compute the entrainment coefficient.
Neilson (13) repeated Brush's (1970) experiments with an air plume. He
found the air plume destratified the system, that a Peclet influence
was again evident, and that the interface approached the nozzle almost
asymptotically with time making computation of the entrainment velocity
by density measurements very difficult. Again, the shape of the density
profile during destratification was not measured.
An interesting result of this set of experiments, which was not discussed
in detail by the authors, was the apparent nonlinearity of the destra-
tif ication process. As shown in Figure 2, two apparent mixing regimes
were found by Neilson. In one the mixing time decreased rapidly with
increasing Richardson number (air flow rate), while it decreased only
very slowly for the second. Similarly Brush, et al., (11) state (p.49):
"The mixing time [for liquid jets or plumes] decreases with increasing
distance from the interface and for a reason not apparent, the mixing
time is less when the outlet is placed in the lighter fluid."1 In the 1970
experiments (12) with air plumes, Brush found a dimensionless mixing time
t v*
a _ (*-m jair.
m Depth '
decreased much more rapidly with increasing ^jair (and hence inverse
Richardson number) when depth (and hence volume) and distance to the inter-
face from the nozzle were greater. As the latter two parameters were kept
as a constant ratio with only vjair varying, differing effects of each
were not isolated.
sic, the comma may be misplaced here.
*V. . = velocity of the air jet based on orifice discharge divided by area.
jair
1030
-------
In none of the experiments reviewed thus far has the change in the density-
depth function been related to the mixing process. The mixing process was
described either by measurement of density in one, or both, layers, or as
a time to complete mixing. Few observations have been made of the actual
destratification process, but a comparison of those available sheds light
on several properties including the apparent change of plume-mixing effi-
ciency. Crapper and Linden (15) measured changes in density profiles for
salt and heat from mechanical mixing (grids). In particular they found
that 1) the interface thickness decreased with increasing agitation (i.e.,
lower Richardson number) and 2) the destratification process appears, at
times, to be "diffusive" [that is, the interface does not descend dis-
tinctly as most mechanical mixing models implied, but the density-related
scalar appears to propagate across a plane of constant density at the
position of the initial interface in a manner analogous to an error
function solution of the Fourier equation] at "high" Richardson numbers.
An illustration of the mixing process from their article is provided as
Figure 2. Because they assumed the turbulence to be homogeneous, Crapper
and Linden's analysis is suspect however.
Finally, several lake or reservoir destratification experiments using air
plumes have been reported in the literature. Many of these are of no
utility at all since either incremental volume was not calculated or only
the time to complete mixing irrespective of initial stratification was
measured. A few papers report results of interest however. Knoppert,
Rook and Oskam (16) destratified a lake of 8.02 H^ volume and 30 m depth
with an air plume. The progress of destratification is shown in Figure 4.
Note that an initial linear density profile sharpened to an interface as
the nozzle was approached. Furthermore the efficiency of mixing dropped
quickly as the interface sharpened. After measuring data from their figure,
the depth of the lower layer (hypolimnion) was found (Graham (17)) to be
described well by the empirical equation
h2 = 63 - 41.7 (Zqalr*E-4)-°696 (2)
where ha - distance from nozzle to thermocline in feet
Eq . - cumulative air discharge, in ft3/s 1
313T
Graham (17) also calculated the mixing efficiency, E^, defined as volume
of water raised per unit volume of air released, from Knoppert, et al.'s
data to increase with ha:
E =3.316 exp (+0.204 h2) (3)
m
empirically, based on a best-fit criterion.
1 U.S. customary units were used in the original paper.
1031
-------
From a now-classic set of experiments by Koberg and Ford (]8) it is possible
to use their data to show that the change lake stability (in Kg - mE6)
decreased rapidly as compressor operation time increased. A good fit is
provided by
-1.083
Stability = 50.1 t
op
(4)
where top is duration of compressor operation, in hours.
correspond with these of Knoppert, et al.
These results
A very recent paper by Moretti and McLaughlin (19) shows even more clearly
some aspects of the destratification process previously described. Their
figures 11 and 13 (*) show a "diffusive" type of destratification with
some evidence of interface sharpening in the prototype (an Oklahoma lake),
and a very clear interfacial re-sharpening as the thermocline approached
the jet (liquid, not air) in the model. A plot of stability index vs.
time (or cumulative discharge) is almost identical to those of Neilson,
Knoppert et al., and Koberg and Ford in form.
While there are many papers on the subject in the literature (20, 21 for
instance), this brief review has been adequate to define the problem. It
appears that plumes, jets, and forced plumes affect, as well as are af-
fected by, the ambient stratification. While numerous investigators have
studied plume behavior in stratified environments (particularly linear ones),
and interfacial entrainment velocities under laboratory conditions, very
little attention has been paid to the actual shape of the vertical density
profile as a plume or jet acts upon it. It appears that two distinct
"regimes" characterize the mixing process - a more efficient "diffusive"
regime occurring far from the orifice and in "High Richardson No." cases
and paradoxically, interfacial formation close to the orifice or with
"high" degrees of agitation. A series of simple experiments was devised
to test this hypothesis in a dimensionally rigorous format.
DIMENSIONAL ANALYSIS
From geometrical and physical reasoning it may be postulated that the fol-
lowing function defines the mixing process.
$1 [t^, Pi, pz, Pa. Ki2, Ki3> K23» hi, h2, do, RO> g>
r»
0
(5)
where t^ - time until the fluid is H% mixed locally
Pi - density of the lower (denser) fluid
P2 - density of the upper (lighter) fluid
These cannot be reproduced for copyright reasons.
readily available however.
The journal is
1032
-------
P3 - density of the plume
KH ~ mo:1-ecular diffusion between fluids i and j at different con-
centrations of heat where 1 - lower fluid; 2 - upper fluid;
3 - plume
hi - depth of upper layer (see Figure 5)
h2 - depth of lower layer (see Figure 5)
do - orifice diameter (see Figure 5)
RO - vessel radius (see Figure 5)
g - gravitational acceleration constant =9.81 ms~2 (6)
VljL ~ molecular viscosity of fluid i
ae - bubble radius
r - radial coordinate (see Figure 5)
z - vertical coordinate (see Figure 5)
<|> - azimuthal coordinate (see Figure 5)
The geometrical parameters are illustrated in Figure 5.
The presence of the third reference fluid of the plume (or jet) makes solu-
tion of equation 5 nearly intractible. A simplification can be made if an
air plume be used since 1) almost no mass is introduced into the system
(since Pair << Pwater) » 2) the density and viscosity of air are much less
than water, and 3) density and viscosity differences between the air and
water are much greater than between those of the water layers themselves.
If an air plume is used as an agitator, equation 5 can be reduced to
$2 UH, Po, Ap, K12, hi, h2j do, RO, g, P, ae, r, z, ] = 0 (7)
where po - reference (Boussinesq) density for Pi and p2
y - reference mol. viscosity, i.e., yi » y2 (8)
Ap = p2 - pi
Let Ki2 = K
It is assumed that Pair> §» ^ may ^e considered constant. Pair i-s deleted
as a parameter if the bubbles are large and not concentrated since it can
be neglected in the momentum terms because Pair << Po-
If ae = $3[do, Po, qair, Pair, D, g, y] (9)
where D = hi + h2 (10)
3air ™ air discharge rate
as found from the literature, then in the experiments
only since all other effects are fixed. Since, over Dmax = 38.735 cm,
differences in D account for a volume change of about 3% at most, then
approximately.
1033
-------
Using equations (10) and (12), and holding d0, and R0, and K constant
in the experiments (see Figure 5) equation 7 reduces to:
$6 [% Po, Ap, qair, hlf D, r, , z] = 0 (13)
Due to the measuring technique, to be described presently, variation of
density with r and <$> is small. While RQ is not varied in the experi-
ments, it is retained since common sense and a literature review by
Graham (17) indicates CH is inversely proportional to volume and not
depth cubed. d0 is also retained, although not varied, in order to
scale the plume. Nine terms remain which may be formed into six dimen-
sionless groups:
Ap hi Rp_ £ D ,
D R02 ' Po ' D ' D ' D ' d0
These are the simplest forms of the parameters - the experiments attem-
pted to relate all but Ro/D which had to be left fixed, unfortunately.
It may be enlightening to show that more classical dimensionless para-
meters may be defined if g, K, and y be reintroduced so that 12 - 3 = 9
dimensionless parameters exist -
Ap gD5 ^H D z hj. f^H Rfl. , = 0 ,-,5)
2' ' ' ' ' J U ^ '
D R02 ' po » q2air' p0R0' d0' d0' d0 ' Ro2
Now . . _ . f =
D po ^iir S ^ po (^air/do2)2
ytH Ro2 D do y do
(17)
(
V }
= Rep (19)
_
Kt.. Ro2 D d0 Kd0 P
rl
D = f
^ ;
and
1034
-------
where Ri , Re , Pe and Sc are plume Richardson, Reynolds, Peclet and
Schmidt fiumbefs respectively. Equation 15 may thus be rewritten as
, Hlp> Rep, Pep, Scp, £ . ^ i • £l J - 0 (22)
Since K, y, RQ, do were not experimentally varied, then effects of Pe ,
Sc , and Ro/do were not analysed; and analysis of Re is redundant ifP
R:L be examined since the parameters which were variid - Po» q . and D -
are common to both. If D and qair be kept constant, then the relation
between tfl lair D"1 Rg~2 (= $tH henceforth) and Ri is the same. as that
with Ap/pQ (hence the validity of equation 14) provided fcH is not sensitive
to Rep which is usually the case in turbulent conditions. Note in comparing
equations 15, 16 and 22 that the Richardson function should be derivable by
varying either Ap/pQ or ^air for constant geometry.
A. DIFFUSION ANALOGY
Harleman, et al. , (23) suggested in 1962 that the decoupling and mixing
action at an interface could be described by a turbulent diffusion
equation. Their arguments were based upon the equation of continuity of
mass which implied a convective - diffusive equation would describe the
process. Experimental results of Crapper and Linden (Figure 3) and
Graham (Figure 7) indicate that a boundary condition of constant density
over time near or at the initial interface location, and the associated
symmetry of the density transects about this level, would be amenable to
description with an error solution of the heat equation.
The fundamental conservation equation may be stated as
Dp_ = 0 = I (pu) + _d_ (pu) + JL _9_ (pv) + _§_ (pw) (23)
Dt r 3r r 8ef> 3z
where (u,v,w) are instantaneous velocities in the coordinate directions
(z,r,<(>), (see figure 5). While space is not available to include the
derivation here, Graham (17) averaged fluctuations of u(z ,r,4> ,t)} v(z,r,,
t), w(z,r,,t) and p(z,r,,t) over r,, and t, and. used mass continuity,,
to derive
. "
(24)
where the tilde is a time average, and a prime indicates a time fluctuation.
The y-variable is just y = z-\\2 which centers the coordinates at the
interface at t=0.
Equation 24 results when variations in r and are removed by turning off
the plume and allowing the fluid system to equilibrate. What equation 24
describes then is the effect of the plume agitation, not the plume dynamics.
1035
-------
Traditional phenomenological arguments based on the Prandtl mixing-
length concept and Reynolds' analogy allow equation 24 to be altered to
7 *'' ' (25)
Since Ey(y) because h.2/do is not scaled to hi/b_2 in the experiments, Ey
should not be placed inside a V operator. Solution for Ey has been
done numerically from experimental data. Finally, equation (25) has the
same form if the parameters are nondimensionalized. Further discussion
of this concept may be found in reference (l7).
SOME EXPERIMENTAL RESULTS
The experimental apparatus is illustrated in Figure 6. It consisted of
a plexiglas cylinder (to provide radial symmetry) 38.74 cm high and
12.17 cm in diameter. The orifice was 0.476 cm in diameter and centered.
A plate above the bottom and 2 side outlets allowed sharp interfaces to
be made. Only saline solutions were used so that no conservation prob-
lems would arise, and K = constant. (Temperature losses or gains to the
atmosphere during an experiment, would aliase the results). Air discharge
rate (^air), total depth (D), buoyancy difference (Ap/po), and initial
interfacial height (h.2) were systematically varied. The experimental
values are given in Table 1.
In general, the rather small volume of the cylinder resulted in the
liquid being very agitated at high air flow rates. Because Ro could not
be varied, the results cannot be generalized to other geometric configur-
ations since volume-dependency has not been removed from the coefficients.
On the other hand, a very clear picture of the sequential .destratification
process adumbrated in the various references could be discerned. A
detailed description of the experimental results is not possible within
the length constraints of this article, but some selected results will be
given in the hope that it might encourage some prototype-scale experiments
along these lines.
Figures 7 and 8 are actual reduced reproductions of conductivity-depth
X-Y recordings for experiments (6-2) and (8-4) respectively. After the
air plume had been passed through the system for a period of time it was
turned off and the liquid allowed to come to rest. A transect was then
made with a very sensitive salinity recorder and a depth-conductivity
signal was fed into an X-Y recorder. These figures have not been adjusted
for density calibration so that the final "fully-mixed" trace does not
lie at an appropriate proportional distance from the initial upper and
lox^er density tracings. The patterns .on the figures selected are quite
clear nevertheless however.
All features of the full "classic" destratification process can be seen
in Figure 7. First, note that the density at the interface did not change
during the first few transects. This is the so-called "diffusive" regime.
1036
-------
Note also that the rate of change of density at locations remote from the
interface occurred rapidly in this regime. After transect 4 (45 sec.)
the stratification changed to one being progressively more interfacial
and which approached the nozzle only very slowly. The upper layer became
rapidly homogeneous while material from the upper layer did not seem to
mix into the lower so easily. A zone just below the initial interface
location actually became more saline.
If the traces represented some density-associated water-quality parameter,
such as DO, it is easy to see that DO could reach locations below the
thermocline much more readily and efficiently if a "diffusive" regime
prevailed. In light of the dimensional arguments given previously,
particularly in the discussion regarding Rip (equation 17) it was found
that interfacial destratif ication occurred sooner (dimensionless time
$tn) at a lower Richardson number, that is, at higher air discharge rates
or lower initial buoyancy differences. The slower (in terms of nondim-
ensional time $t ) mixing after a change from "diffusive" to "interfacial"
regimes results in graphs similar to Figure 2, and Figure 12 of refer-
ence (19).
It was possible to demarcate the point of transition from one regime to
the other in all the experiments. This dimensionless time was termed
$t and an empirical equation describing its occurrence was found, by
best fit, to be
$t = constant * J\p . 1.07 * ,hov 0.65 * , .hn, ,,,,
D (~) (^ «10 (-1.) (26)
where hj, b.2 represent the initial values of these parameters (see Figure
5). Note that in equation (26) t^ ^air ~ , and the exponent for the
buoyancy difference is also close to 1, while the relations with respect
to hi/do and h^/h£ are nonlinear, as hypothesized in equations(2 ,3,4) .
Additional analysis and the form of "f^Qcan be found in Graham (17).
Finally, the dominance of the 'interfacial' regime at small b.2/do (near
the orifice) can be seen in Figure 8. This appears to be a local scaling
phenomenon since it occurred for all D/dg, and near the end of all experi-
ments. Paradoxically the interface is most distinct where the local
agitation is greatest (measured by jet velocity), but the jet width is
least. Entrainment at this location is very weak and destratif ication
occurs very slowly.
SUMMARY AND CONCLUSIONS
As mentioned, many studies have been made of the effects of local stratif-
ication upon plume behavior (7,2,24) but very few of the opposite case.
Thermal plumes and jets obviously affect locaJ stratification, particularly
in smaller lakes and reservoirs. It has been shown that similar distinct
1037
-------
mixing sequences seem to occur in both the field and in laboratory
studies. A more 'efficient' overall type of mixing (diffusive) is
characteristic of high initial stability and low jet agitation; while
classic interfacial descent occurs where there is low initial stability
and/or high jet agitation, and especially when the interface is near
the diffuser. While greater analysis of the results appears in (17),-
experiments at a prototype scale are needed to extrapolate these results
to different geometries (25). Such experiments should also follow a
rigorous dimensional format such as the one presented. Additional
laboratory and field experiments are necessary to l) determine the
validity of the Fourier analogy, and 2) properly understand the cause
of the change from one mixing phase to the other.
If this plume mixing process is more clearly understood, then thermocline
location and sharpness (and other associated water quality parameters)
can be better modeled and predicted.
ACKNOWLEDGEMENTS
This study was conducted as private research. Laboratory space and
equipment were provided by the Johns Hopkins University, Baltimore,
Assistance with drafting was provided by Carol Dillard, engineering
student at UF. Thanks are also due to Irene Urfer of the Department of
Geography, Brandon University for typing the manuscript.
REFERENCES
1. Brooks, Norman H. (1972) Dispersion in Hydrologic and Coastal
Environments. CIT Keck Lab. Rpt. KH-12-22. 203 pp.
2. Harleman, Donald R.F. (1972) "Thermal stratification due to heated
discharges", Proc. Int'l. Symp. on Stratified Flows, ASCE,
Novosibirsk, 35-68.
3. Rouse, H. and J. Dodu (1955) "Turbulent diffusion across a density
interface". La Houille Blanche. H), 405-410.
4. Turner, J. S. (1968) "The influence of molecular diffusivity on
turbulent entrainment across a density interface". JFM. 33. 639-656.
5. Turner, J.S. (1973) Buoyancy Effects in Fluids. Cambridge Univ. Press.
6. Long, Robert R., (Oct. 1974) Lectures on Turbulence and Mixing Processes
in Stratified Fluids. Tech. Rpt. No. 6 (Series C). Dept, of
Earth and Planetary Sciences, the Johns Hopkins Univ., Baltimore, MD.
7. Uhl, Vincent and Joseph B. Gray (1966), Mixing, Theory and Practise
Vol. 1, Academic Press, N.Y.
8. Baines, W.D. (1975) "Entrainment by a jet in plume at a density
interface". JFM, 68., (2), 307-320.
1038
-------
9. Sullivan, Paul J. (1972). The penetration of a density interface
by heavy vortex rings. Air. Water and Soil Pollution. I. (3),
322-336.
10. Linden, P.P. (1973). "The interaction of a weak vortex ring with
a sharp density interface: a model for turbulent entrainment1'
JFM. 60, 467-480.
11. Brush, Lucien M. Jr., Francis, C. McMichael and Chen Y. Kuo (1968).
Artificial Mixing of Density-Stratified Fluids: A Laboratory
Investigation - Princeton Univ., Moody Hydrodynamics Laboratory
Ii.pt. MH-R-2. 80 pp.
12. Brush, Lucien M. Jr., (1970) Artificial Mixing of Stratified Fluids
Formed by Salt and Heat in a Laboratory Reperyojr. N.J. Nat.
Res. Instit., Rutgers Univ. 33 pp.
13. Neilsen, Bruce J. (1972) Mechanism of Oxygen Transport and Transfer
by Bubbles. Ph.D. Diss. , the Johns Hopkins Univ. , Baltimore,
MD. 131 pp.
14. Kantha, Lakshmi H. (August, 1975) Turbulent Entrainment at the
Density Interface of a Two-Layer Stability Stratified System..
Publ. of Dept. Earth and Planetary Science, The Johns Hopkins
Univ., Baltimore, MD. 161 pp.
15. Crapper and Linden (1974), "The structure of turbulent density
interfaces." JFM. 6.5, (1), 45-63.
16. Knoppert, P.L. , J.J. Rook and G. Oskam (1970). "Destratification
experiments at Rotterdam", Jrl. AWWA, 62., 448-454.
17. Graham, Donald S. (Sept. 1976) An Experimental Study of the Mixing
of 2-layer Density-Stratified Liquids., by.. an__ Air Plume_._in_a_
Small Cylindrical Container. Submitted to The Johns Hopkins
Univ. in partial fulfillment of the requirements for the Ph.D.
degree, Oct. 6, 1976. 872 pp. Draft.
18. Koberg, Gordon E. and Maurice E. Ford, Jr. (1965) Elimination of
Thermal Stratification in Reservoirs and Resulting Benefits.
U.S.G.S. Water Supply Paper 1807-M. 28 pp.
19. Moretti, Peter M. and Dennis K. McLaughlin (Apr. 1977). "Hydraulic
modeling of mixing in stratified lakes". Proc. ASCE, 103, HY4,
367-380.
20. Henderson - Sellers, Brian (June 1976). "Role of eddy diffusivity
in thermocline formation". £roc.- ASCE. 102, (EE6), 517-531. With
British references.
1039
-------
21. Ito, Takeshi (1972), "Mixing method of stratified water layer in
reservoirs" (sic). In International Svmp. on Strat. Fluids,
Novosibirsk, USSR. ASCE Publ. 567-577.
22. Gebhart, Glen E. and Robert C. Summerfelt (Dec. 1976).' "Effects
of destratification on depth distribution of fish". Proc.
ASCE, 102. (EE12), 1215 - 1228.
23. Harleman, D.R.F., J.A. Hoopes, D. McDougall and D. A. Goulis (1962)
Salinity Effects on Velocity Distributions in an Idealized
Estuary. MIT Parsons Lab. Tech. Rpt. No. 50. 45 pp.
24. Wright, Steven Jay (May, 1977). Effects of Ambient Crossflows and
Density Stratification on the Characteristic Behavior of Round
Turbulent Buoyant Jets. CIT Keck Lab. Rept. KH-R-36. 254 pp.
25. Graham, Donald Steven (June, 1978). Disc, of "Aeration of hydro
releases at Ft. Patrick Henry Dam". Proc. ASCE, 104,, (HY6),
943-945.
10 30
-------
TABLE I
VALUES OF EXPERIMENTAL PARAMETERS
Experiment
Number Abbr.
10.
10.
10.
10.
10.
10.
10.
10.
10.
10.
10.
10.
10.
10.
10.
10.
10.
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
.1.-
.1.-
.1.-
.1.-
-1.-
.1.-
.1.-
.1.-
.1.-
.1.-
.1.-
.1.-
.1.-
.1.-
.1.-
.1.-
.1.-
(5-1)
(5-2)
(5-3)
(5-4)
(6-1)
(6-2)
(6-3)
(6-4)
(6-5)
(7-1)
(7-2)
(7-3)
(8-1)
(8-2)
(8-3)
(8-4)
(8-5)
5-1
5-2
5-3
5-4
5-1
6-2
6-3
6-4
6-5
7-1
7-2
7-3
8-1
8-2
8-3
8-4
8-5
qair
cm3
35
80
120
210
49
49
49
49
49
49
49
49
49
49
49
49
49
D
cm
38.735
do.
do.
do.
do.
do .
do.
do.
do.
38.735
do.
do.
38.735
29.05
19.37
19.37
10.80
Ap
• •
(Initial)
0.0130
do .
do.
do.
.0124
.0124
.0062
.0034
.0014
.0141
do.
do.
.0080
do.
do .
do.
do.
*
1.
1.
1.
1.
1.
1.
1.
1.
1.
1.
1.
1.
1.
1.
1.
1.
1.
PO
•
0154
0154
0154
0154
0130
0130
0161
0175
0185
0114
0078
0164
0103
0103
0103
0103
0103
hi/D
• •
(Initial)
1/2
do.
do.
do.
do.
do.
do.
do.
do .
1/2
3/4
0.148
1/2
1/2
1/2
1/2
1/2
1041
-------
Figure 1
too
10
i
O
X
•I .->
1
O.I
1
A
&J A
' ^^
^ + O
V CH-A A SF-A * *
O
O CH-B + SF-B +
: o
; 0 CH-C * SF-D + °
O
C— Heat-Induced Stratification
S — Salt - Induced Stratification ^
i l ii
O'6 ICT5 IO'4 IO"3 IO'2
0 JJ — \] .
J«t Entrainment. Velocity v. Richardson Number
Source' Brush (12), modified by Graham (17)
1042
-------
Figure 2
50
40
•; 30
c
'i
«
_§
~ 20
10
3 4 5 6
* n (number of orifices)
10
Mixing Time as a Function of Air Flow Rate According to Neilson
Source: Neilsen (13)
1043
-------
Figure 3
1.000 1.001 1.002 1.003 1.004 1.005
Density
A Series of Depth - Density Transects From Cropper and Linden
Source : Modified from Cropper and Linden (15)
Figure 4
Air: 88 cfm
(170 Holes,
I mm diameter}
4-8-67
10-8-67
I7-8-S7
23-8-67
28-8-67
7-9-67
Progress of DestraMfication of Lake Maarsseveen
Source^ Knoppert, et ol. . (16).
1044
-------
o
£>
Ul
Figure 5
orifice
Experimental Configuration and Symbol Definitions
Figure 6
-12.7
Side
Constructed
of plexigtas
0.953 _
valve
\63S
11
outlet-*
outlet-"
i
K6.35-1
0.6350.0.
orifice*1 ** .982
S>n JJ
i -
is
± -
19
37
37
2.22_i_
?rr
All dimensions in cm.
Orifice O.D. 1/4 inch
I.D. 3/ !6 inch
or .4V6 cm
outlets are 1/4
(.635 cm.)
Scale-' h4
Plan
brine connection
to air source
Diagram of Experimental Mixing Apparatus
Source '• Graham (I?)
-------
Figure 7
orifice leve
Conductivity
X-Y Graph of Density - Depth (or Experiment (6-2)
Source' Graham (17)
Figure 8
waler
orifice level
Conductivity
D= 19.37 cm.
D/d.= 40.67
h/DsO.5
9nir = 49 «"i' "tin"'
run time (s)
i T
I
10
40
91
X-V Graph of Density - Oeplh for Experiment (8-4)
Source •• Graham (17)
-------
THREE-DIMENSIONAL FIELD SURVEYS OF THERMAL PLUMES FROM BACKWASHING
OPERATIONS AT A COASTAL POWER PLANT SITE IN MASSACHUSETTS
A.D. Hartwell, Normandeau Associates, Inc., Bedford, NH 03102 and
F.J. Mogolesko, Boston Edison Company, Boston, MA 02199, U.S.A.
ABSTRACT
Using specially designed temperature profiling equipment, two surveys
were conducted during thermal backwashing operations at Pilgrim Nuclear
Power Station to determine the spatial and temporal extent of temperature
rises above ambient, Backwashing formed a thermal plume about 5 to 6-ft
thick (1.5 to 1.8 m) in front of the intake screenwall. Maximum observed
surface temperatures were 101.0 F (38.3 C), representing a AT of 43.4 F
(24.1 C) above ambient. The frontal zone of the plume spread gradually
seaward at about 0.2 kn. Its outer edge became thinner and rapidly
cooled, presumably by advection and turbulent diffusion associated with
currents from the reverse pumping and local changes from dissipation to
the atmosphere. Along the intake shoreline, the plume was often less
than 1 ft (0.3m) thick. Most of the hot water was dissipated within
several hundred feet of the intake with AT's of about 10.0 to 15.0 F (5.6
to 8.3 C) above ambient. Under the influence of strong southwesterly
winds during the second survey, some warmed water was apparently carried
beyond the outer breakwaters into Cape Cod Bay. These surveys provided
real-time data indicating that the backwashing operation caused a rela-
tively thin thermal plume, which spread rapidly from the intake out
across the study area and along the seaward breakwater. Within a few
hours these backwash thermal plumes were completely dissipated.
INTRODUCTION
Although thermal backwashing is a commonly used technique for control of
biofouling in condenser tubes and intake structures of operating power
plants, only limited published information is available on the receiving
water temperature structure caused by such operations. Boston Edison
Company, Boston, Massachusetts, conducted two thermal surveys of actual
mid-summer backwashing operations under varying tidal conditions at
Pilgrim Nuclear Power Station during 1977 to establish a synoptic picture
of the plume's three-dimensional structure fl].
The Pilgrim Nuclear Power Station, located on the shore of Cape Cod Bay
in Plymouth, Massachusetts, is a 655 MW light-water moderated, boiling
water nuclear reactor with a once-through condenser cooling water system.
Water used for cooling the condenser is removed from Cape Cod Bay through
a shoreline intake (Fig. 1). It enters the intake between two break-
waters via a dredged channel which is about 18 to 24 ft deep (5.5 to 7.3 m)
at mean low water (MLW).
1(147
-------
Under normal operating conditions, the water is drawn into the intake by
two pumps (designated herein as east and west), circulated through the
condenser system and discharged via a surface canal at a rate of about
510 million gallons/day and a AT (difference between the discharge and
intake temperatures) averaging 30.0 F (16.7 C). Condenser tubes are
cleaned by backwashing on a 1 to 2-week interval, depending upon bio-
fouling severity. Generally 45 to 60 min are required to treat each of
the two circulating water pumps, with elevated temperatures averaging
around 100.0 F (37.8 C). Occasionally the temperatures peak at from
110.0 F (43.3 C) to 120.0 F (48.9 C), depending upon the amount of heat
treating necessary. Because plant load must be reduced during backwash-
ing, the operation is generally conducted at night during off-peak hours.
METHODS
This study conducted by Normandeau Associates, Inc. (NAI), of Bedford,
New Hampshire, consisted of overnight three-dimensional temperature and
current surveys, supplemented by continuous thermal monitoring. For the
first survey on July 9 and 10, 1977, backwashing began at low water and
continued into early flood tide. During the second survey on July 16 and
17, 1977, backwashing began at high water and continued into early ebb
tide. Both surveys concentrated on the time history of plume build-up
and dissipation.
Temperature and depth data were collected at selected stations (Fig. 1)
and plotted on board the survey boat using a Naico Model 3100-TD Profiling
System (Fig. 2). Current velocity profiles were acquired using Bendix
Model Q-15 current meters and Model 270 recorders. Precise location was
continuously recorded using a Motorola MiniRanger III System with two
shore based transponders.
Two Naico Model 200 Digital Field Temperature Recorders were utilized to
periodically measure temperature profiles from water surface to bottom at
two stations in the intake channel. The arrays were assembled so they
could be moved quickly within the survey area to check thermal anomalies.
In addition, two Naico Model 1001-T Temperature Recorders were installed
to monitor water temperatures 1) inside the intake screenwall and the
discharge canal, and 2) in ambient receiving waters of adjacent Cape Cod
Bay.
Observed temperatures were transformed to true temperatures using regres-
sion equations based on calibration data for each respective field
instrument. From measurements of ambient near-bottom waters mid channel
between the two intake breakwaters (Fig. 1), a AT or approximate temp-
erature rise above ambient was calculated for each temperature observa-
tion.
1048
-------
FIELD SURVEYS
Low-Water Backwash Survey
The July 9 and 10 low-water backwash survey consisted of five sampling
runs keyed to actual plant operations. For this survey, NAI's ambient
temperature measurements along the bottom of the intake channel started
around 49.0 to 50.0 F (9.4 to 10.0 C) and then gradually rose to about
58.0 F (14.4 C) by the time of low water. Throughout the rest of the
night, ambient temperatures continued to rise slowly, reaching about 60.0
F (15.6 C) by the end of the survey. This rise may represent some recir-
culation of the discharge plume toward the intake area because of local
winds and coastal currents.
As backwashing was initiated, plant load was gradually brought down.
NAI's readings of discharge canal temperatures showed a drop from 87.0 F
(30.6 C) to 74.6 F (23.6 C; Fig. 3). Next, the west pump was backwashed
from about 0030 to 0119 EST. The in situ temperature monitors recorded a
sudden rise in discharge temperature to about 83.0 F (28.3 C), followed
by a sharp drop to about 65.3 F (18.5 C). Simultaneously water box
temperatures rose quickly to about 104.0 F (40.0 C) and remained at this
level for much of the backwashing period (Fig. 3) . As backwashing of the
first pump neared completion, discharge temperatures rose again to 83.2 F
(28.4 C) and water box temperatures dropped back down to below 70.0 F
(21.1 C). From about 0150 to 0227-EST the east pump was backwashed in
the same way with similar backwash temperatures observed for both pumps.
During this backwashing period, discharge temperatures dropped to about
70.6 F (21.4 C), then rose to 87.0 F (30.6 C) for a short time, dropped
back down to about 75.0 F (23.9 C), and finally rose back toward normal
operational levels (Fig. 3).
A prebackwash survey conducted during late-ebb showed surface temperature
rises (AT) ranging from 9.1 F (4.1 C) near the offshore discharge to 4.9
F (2.7 C) near the plant intake.
As backwashing started, the first visible evidence was a sudden rush of
hot, turbulent water marked by foam and a steamy vapor right in front of
the intake. With continuing backwashing, the hot water formed a surface
layer about 5-ft (1.5 m) thick, which reached temperatures as high as
100.0 F (37.8 C) in front of the intake screenwall. A distinct frontal
zone moved slowly northward (or seaward) away from the intake, bulging in
the middle and slightly restrained along shore due to frictional effects.
The water temperatures in the near-surface thermal plume gradually
decreased with both distance away from the intake and time, presumably
due to evaporative heat loss and dilution (mixing with ambient waters).
1049
-------
At the surface, AT's of 42.1 F (23.4 C) in front of the west pump and
24.8 F (13.8 C) in front of the east pump were observed (Fig. 4).
Within less than 100 ft (30.5 m), the AT from the western pump was
28.0 F (18.6 C) or less. High AT water hugged the outer breakwater,
apparently because of momentum effects and southwesterly winds during
the night. Surface AT's of 10.0 F (5.6 C) and higher were confined to
the western third of the intake area between the breakwaters (Fig. 4).
The remainder of the area experienced AT's equal to or colder than
observed prior to backwashing.
At the 3.3 ft (1.0 m) depth level, observed AT's were 23.4 to 24.3 F
(13.0 to 13.5 C) in front of the intake. Within less than 200 ft (61.0 m),
AT's were down to 18.2 F (10.1 C). Beyond that distance they dropped
from 14.8 to 6.7 F (8.2 to 3.7 C). Near the outer end of the break-
waters, AT's were only 2.4 to 3.3 F (1.3 to 1.8 C).
At the 9.8 ft (3.0 m) level, AT's were 4.6 F (2.6 C) or less in front of
the intake and 1.2 to 2.1 F (0.7 to 1.2 C) along the dredged channel.
Along the bottom all of the AT's were negative, or colder than conditions
at the outer end of the breakwaters. At Station 6 minimum values were
-4.4 F or -2.4 C (Fig. 5).
The detailed profiles at Station 6 showed that the backwashing from the
western pump formed a distinct slug or pulse of hot water along the
surface, which eventually extended-down to about 7 ft (2.1 m). The
heated effluent apparently took about 15 rain to reach and about 75 min to
pass the anchored boat in its seaward progression (Fig. 5). Maximum
observed AT at the surface was 22.6 (12.5 C), which represented an actual
temperature of 79.0 F (26.1 C). Near-bottom temperatures were 53.4 to
56.2 F (11.9 to 13.4 C) which represented negative AT's of up to -4.7 F
(-2.6 C). By about 0119 EST backwashing of the west pump was complete.
At about 0150 EST backwashing of the east circulating water pump started.
As before, there was a sudden surge of hot, turbulent and steamy water at
the surface. Within minutes a thin thermal plume and a distinct seaward-
moving frontal zone was observed. At the surface, AT's were essentially
the same as during backwashing of the west pump, averaging 20.0 F (11.1 C)
and more across the western third of the study area, 10.0 to 20.0 F (5.6
to 11.1 C) in the middle, and 5.0 to 10.0 F (2.8 to 5.6 C) across the
eastern third. As before, the highest temperatures were along the outer
breakwater. At 3.3 ft (1.0 m) AT's were 15.5 to 23.4 F (8.6 to 13.0 C)
next to the intake and gradually decreased seaward. Below this level
there was no evidence of the backwash plume, whereas along the bottom
AT's remained negative.
At Station 6 the second backwash manifested itself as another pulse of
hot water, which was warmer than before (up to 81.1 F or 27.3 C) but
slightly thinner and shorter-lived (Fig. 5). This plume had surface AT's
of up to 23.5 F (13.1 C). Apparently it took about 10 to 15 min for this
1050
-------
second plume to reach the anchored boat, but its effects were only evi-
dent for about 60 min. By the time the plume had passed, it was only
about 1 to 2 ft (0.3 to 0.6 m) thick. Near-bottom temperatures showed
little change, ranging from 54.1 to 56.2 F (12.3 to 13.4 C) and repre-
senting negative AT's (down to -3.4 F or -1.9 C). By about 0227 EST
backwashing of the east pump was complete and the plant began to return
to normal operation.
Subsequent surveys for the rest of the night showed that the elevated
surface temperatures from the backwashing operation persisted for only
about 2 to 2.5 hrs in the western portion of the study area and even less
in the eastern portion, before being completely dissipated.
High-Water Backwash Survey
One week later on July 16 and 17, a second survey was conducted under
high-water tidal conditions. Throughout this survey ambient temperature
measurements along the bottom of the intake channel showed very little
variation, ranging from 52.0 to 55.0 F (11.1 to 12.8 C). Backwash temp-
eratures were about the same for both pumps (peak of 107.0 F or 41.7 C);
however, this series of backwashes lasted 20 to 25 min longer than
respective ones the week before because of increased fouling of the
condenser tubes.
At about 2354 EST on July 16, backwashing started on the west pump. This
time, in sharp contrast to the low-water backwashing, the surface appear-
ance of the backwash waters was much less dramatic. The thermal plume
was somewhat turbulent and steamy, but the thermal front along the inter-
face with Cape Cod Bay waters was much less distinct than it had been the
week before. Apparently this was because more dilution or "receiving"
water was available at high tide.
The observed surface AT's were 28.2 F (15.7 C) in front of the west pump
.and 17.1 F (9.5 C) in front of the east pump (Fig. 6). Warmest tempera-
tures were along the west side of the study area with AT's from 28.0 F
down to about 14.8 F (15.6 to 8.2 C). Across the middle portion of the
study area, AT's ranged from 15.0 to 10.0 F (18.3 to 15.6 C), with most
of the warmed water apparently being blown against the outer breakwater
by the strong southwesterly winds which persisted throughout the survey.
Much lower AT's were seen along the shore in front of the power plant
(6.1 to 9.1 F or 3.4 to 5.1 C). In the eastern portion of the study
area, some warm water was observed along the outer breakwater (8.7 to
11.8 F or 4.8 to 6.6 C) ,- but, close to shore temperatures remained
unchanged. At the discharge the temperature rise was 14.8 F (8.2 C). At
the 3.3 ft (1.0 m) level, AT's were lower than at the surface, but the
general distribution of the backwash plume was about the same. At 9.8 ft
(3.0 m) AT's were small, while near-bottom AT's were negative apparently
due to cold water being drawn into the intake area.
1051
-------
Temperature measurements from the boat anchored at Station 6 showed that
the west pump's backwash plume arrived within 5 to 10 min of the start of
backwashing (Fig. 7). The AT's rose sharply to 14.4 F (8.0 C) or an
actual temperature of 69.1 F (20.6 C). The resulting thermal plume
seemed to be about 2 to 3 ft (0.6 to 0.9 m) thick and persisted for
almost 90 min. Actual backwashing of the west pump was completed around
0113 EST.
At about 0159 EST backwashing of the east pump started. Surface AT's
were 43.2 F (24.0 C) in front of the east pump and 25.2 F (14.0 C) in
front of the west pump. Elsewhere AT's were generally higher than during
the previous sampling run. Temperature rises of 20.0 F (11.1 C) and more
were found across the channel to the outer breakwater. As before the
elevated AT's were observed along the outer breakwater (AT's of 15.0 to
20.0 F or 8.3 to 11.1 C) , possibly due to continuing wind influence.
Slightly deeper at 3.3 ft (1.0 m), the temperature distribution was about
the same as at the surface; but deeper down and along the bottom, temp-
eratures were much warmer than earlier in the evening.
At Station 6 the passage of the east pump thermal plume was very evident
(Fig. 7). It took less than 10 min for the backwash water to arrive and,
as before, it persisted for about 90 min. The temperatures were slightly
higher this time, with the greatest rise occurring after backwashing was
complete. At about 0307 EST backwashing of the east pump was completed
and the plant started to return to 'normal operation.
Subsequent surveys during the rest of the night showed that the elevated
surface temperatures and thermal backwashing plumes persisted for almost
4 hrs'in the western portion of the study area and somewhat less in the
eastern portion, before dissipating. Backwashing momentum effects, as
well as local winds, seemed to play a role in forcing the warmed water
along the outer breakwater and keeping it away from the shore in front of
Unit 1 (Fig. 6).
DISCUSSION
Each backwashing was first evidenced by a pulse of warmed water at depth
from the intake (Fig. 8). As the pumping continued, the hot buoyant
water rose to the surface and within a few minutes formed a warm thermal
plume averaging 3 to 5 ft (0.9 to 1.5 m) thick. Below the plume was a
steep gradient to the colder near-ambient waters along the bottom of the
intake channel. During the first weekend survey, the thermal plume
formed a distinct frontal zone of foam and turbulent, steaming water
which could be easily tracked by eye. Under the influence of the reverse
intake flows, the initial jet momentum, the plume buoyancy effect and the
localized hydrostatic head in front of the screenwall, the frontal zone
moved slowly across the study area. Along shore and in shallow water,
1052
-------
frictional effects slowed the frontal zone, causing the plume to bulge in
the center. The hot water propagated toward the western portion of the
study area and the outer breakwater; but relatively little hot water
contacted the shoreline area in front of Unit 1 during both of the sur-
veys (Figs. 4 and 6). During the second survey the frontal zone behaved
in a similar manner; but was much less distinct, probably because of the
increased volume of receiving water (high-water condition).
Because of the relative thinness of the thermal plume and the pronounced
stratification it created, it appeared to be highly susceptible to wind-
shear effects. During both weekend surveys, momentum effects and south-
westerly winds apparently forced much of the plume against the outer
breakwater, leaving the shoreline area much less affected. During the
second weekend some warmed water was apparently forced out into Cape Cod
Bay beyond the outer breakwater by transient wind effects (estimated to
be only a small percentage of the surface backwash thermal plume). In
general, the eastern portion of the study area remained relatively
unaffected by the hot water during both studies. Where the thermal plume
impinged the shoreline, such as along the breakwaters, it was generally
less than 2 ft (0.6 m) thick.
SUMMARY AND CONCLUSIONS
These surveys showed that backwashing operations at Pilgrim Station form
a relatively thin thermal plume averaging 3 to 5 ft (0.9 to 1.5 m) thick.
Higher temperatures were observed during the low-water backwashing than
during the high-water backwashing, presumably due to lesser amounts of
available entrainment water. During the first survey the thermal plume
persisted for about 2 to 2.5 hrs before being completely dissipated. The
second weekend more heat treatment was required due to accumulated bio-
fouling and the thermal plume persisted for almost 4 hrs. Initial momen-
tum effects of the backwashing flows apparently tend to cacry the thermal
plume northward and along the outer breakwater, with little tendency for
warmed water to impinge the shoreline in front of Unit 1. During both
surveys local winds also appeared to play a role in pushing the thermal
plume seaward. Finally, observed near-bottom ambient temperature vari-
ations suggest that some water from the plant discharge can recirculate
into the intake area.
REFERENCE
Normandeau Associates, Inc. 1977. Thermal surveys of backwashing opera-
tions at Pilgrim Station during July 1977. Conducted for Boston
Edison Company, Boston, Massachusetts. 73 pp.
1053
ADH
-------
•r> PROFILING STATION • IH^smj TEMPERATURE MONITOR
+ ANCHORED BOATS T TIDE STAFF
PLANT TEMPERATURE 0 MINI-RANGER TRANSPONDERS
.30
PLANT
CAPE COD BAY
Fig. 1 Location map showing approximate sampling stations
and in situ instrumentation for the July 1977
Pilgrim Station backwashing studies.
SHORE BASED
TRANSPONDER E
I P^ECISiON _
I
I V
£W_ | ^^.
SURVEY
VESSEL
110 v AC
GENERATOR
DATA
LO G G E ^~~~-~?*C~~S*\
BP
X-Y RECORDER
DATA ACQUISITION SYSTEM
Fig. 2 Instrumentation set up for field surveys.
1054
-------
u, 80-
cz:
5= 70-
6CH
5O-
o
en
01
-40
-35
PILGRIM STATION THERMAL SURVEY
DATE: 7410-?' TIME: 0039-0203
DEPTH: SlIRrACf TIDE; CflBU FLOOD
S 5'i i 5
JULY 9 TIME, EST JULY 1
Fig. 3 Temperature monitor data from the west
pump waterbox, the discharge canal and
the ambient in situ unit at Station 28
during backwashing operations on July
9 and 10, 1977.
BACKWASH )WES^ £*24T
21 22 23 24 1 Z 3 4
JULY 9 TIME, EST JULY 10
Fig. 5 Temperature data from an anchored survey
boat at Station 6 on July 9 and 10, 1977
showing actual temperatures and corres-
ponding AT's above ambient in degrees F.
PILGRIM STATION THERMAL SURVEY
DATE: 7-10-77 TIME: 0039-0203
OEPTH:3.3'(M)TIDE: EARLY FLOOB
Fig. 4 Contour maps of observed temperature rises
(AT) in degrees F above ambient during early
flood (backwash west pump) at surface and
3.3 ft (1.0m) on July 10, 1977.
-------
o
un
CTi
PILGRIM STATION THERMAL SURVEY
DATE: 7-17-77 TIME:0003-0139
DEPTH: SURFACE TIDE: EARLY EBB
DISCHARGES Hr28-2 • PILGRIM
INT4KF H STATI°N
INTAKE ^^^B
PILGRIM STATION THERMAL SURVEY
DATE: 7-17-77 TIME: 0003-0139
DEPTH;3.3'tlm)TIDE: EARLY EBB
II I/ 1
Fig. 6 Contour maps of observed temperature
rises (AT) in degrees F above ambient
during early ebb (backwash west pump)
at surface and 3.3 ft. (1.0m) on July
17, 1977.
4 6
TIME.EST
Fig. 7 Temperature data from an anchored survey
boat at Station 6 on July 16 and 17, 1977
showing actual temperatures and correspon-
ding AT's above ambient in degrees F.
50
5-
10-
15-
20-
60
I
TEMPERATURE (F)
7O
ao
_ i
90
I
-1
-2
Fig. 8 Temperature profiles from Station 1
during the start of backwashing of the
east circulating water pump on July 17, 1977.
-------
SHORT-TERM DYE DIFFUSION STUDIES IN NEARSHORE WATERS
D.E. Frye* and S.M. Zivi**
ABSTRACT
Short-term dye diffusion studies were conducted in the nearshore waters of
Lake Michigan and Massachusetts Bay. Measurements of horizontal and
vertical diffusion of continuous and batch dye releases were made using a
combination of fluorometric and aerial photographic techniques. Results
of experiments performed in summer of 1973 on Lake Michigan and fall of
1974 on Massachusetts Bay showed a similar range of values for diffusion
coefficients. Horizontal diffusivities ranged over about 2 orders of
magnitude with most values falling between 500 and 5000 cm^/sec.
An efficient and practical method of using aerial photography for quantita-
tive diffusion studies was developed following work of Ichiye and Plutchak
(1966).LlJ Comparison between the photographic method and standard
boat-based fluorometry indicates good agreement between the methods.
1, INTRODUCTION
Studies of diffusion in nearshore waters, conducted in Lake Michigan and in
Massachusetts Bay, were motivated by a need to forecast the behavior of
thermal plumes from power plants. The Lake Michigan experiments were
conducted in the summer of 1973 at three sites: the Point Beach Nuclear
Generating Station, Two Rivers, Wisconsin; the J.H. Campbell Generating
Station, Holland, Michigan; and the D.C. Cook Nuclear Generating Station,
Bridgman, Michigan. These sites are located on long, straight shorelines
with gently sloping, sandy bottoms, typical of much of the east and west
shores of Lake Michigan. Currents at the sites typically flow shore-
parallel and are primarily wind-driven. The lake is strongly stratified in
summer, and periods of upwelling (particularly on the western shore) and
downwelling (particularly on the eastern shore) are common occurrences.
Diffusion measurements were made in water 7 to 10 meters deep between 0.5
and 1.0 km offshore.
In the fall of 1974, a similar series of diffusion studies were conducted
near the Pilgrim Nuclear Generating Station, Plymouth, Massachusetts. The
study site is located just south of Plymouth Harbor on the western shore of
the Bay at a point separating Massachusetts Bay in its entirety from Cape
*EG&G, Environmental Consultants, Waltham, Massachusetts, U.S.A.
**Argonne National Laboratory, Argonne, Illinois, U.S.A.
1057
-------
Cod Bay. The waters are semi-enclosed; open to the ocean to the north-
east. Currents are primarily shore-parallel and wind-driven. Semidiurnal
tidal currents of the order of 5 to 10 cm/sec result from tides whose
range is about 3 meters (EG&G, 1976).[2J The bottom gently slopes from
shore, and waters are vertically well-mixed during the fall.
At both sites, similar data collection and analysis techniques were used
to obtain estimates of horizontal and vertical eddy diffusion coeffi-
cients. The Lake Michigan experiments were performed using sodium
fluorescein as a tracer (due to an EPA ban on Rhodamine at that time).
Rhodamine WT was used in the Massachusetts Bay studies.
Dye dispersion studies in natural water bodies have been performed by. 3
number of investigators including Csanady (1973),[3] Murthy (1972),14]
Huang (1971),[5] Ichiye and Plutchak (1966), [1J Eliason, et al. (1971)16]
and others. However, due to the complex nature of the diffusion phenome-
non, neither an adequate theoretical model nor a well-founded engineering
approximation exists to describe the range of turbulent diffusion in the
ocean or large lakes. In defining an eddy diffusivity, it is assumed that
turbulent diffusion is analogous to molecular diffusion with the coeffi-
cient of molecular diffusivity replaced by an equivalent, but much larger,
coefficient of eddy diffusivity. Studies conducted over a variety of
conditions indicate a broad range of diffusion coefficients governed in
part by the complex interaction of current shears, thermal stratifica-
tions, wave action, wind effects, and topography. The data presented here
add to the base of information on diffusion processes in nearshore waters
in both large lakes and in semi-enclosed oceanic waters, and suggest
diffusion rates similar to those found in previous oceanic studies. The
methods described represent techniques for measuring the spread of
fluorescent dye tracers in nearshore waters in a more efficient and
comprehensive manner than is generally used.
2. METHODS
Several experimental techniques were employed in these investigations of
nearshore diffusion. These techniques included the use of continuous and
instantaneous dye releases measured by continuous-pumped-fluorometry,
discrete water sampling, and quantitative aerial photographic dye measure-
ments. A brief discussion of each of the measurement techniques is
presented below.
For those measurements where a continuous point-source of dye was used to
produce a continuous plume, the dye (as a 5 or 10% solution) was pumped
from a 15-gallon drum located on a moored raft. A continuous flow rate of
5 cm-^/sec was maintained by a peristaltic metering pump injecting the
dye 1.5 meters below the surface. Density of the dye solution was
adjusted, using ethyl alcohol, to match the density of the water.
1058
-------
Batch dye releases consisted of either 1 gallon of 5% dye solution, adjus-
ted for density, pumped into the water at a depth of 1.5 meters at a
single point, or "T" shaped patches deployed from a fast-moving boat.
These "T's" were about 300 meters on a side, with an initial width of
about 3 meters (see Figure 1). Diffusion of each leg of the "T'"s"
provided information on the smaller scales of turbulent diffusion (3 to
100 meters), while gross distortion of the "T's" indicated the presence of
larger scale motions.
In those experiments where the absolute dye concentration was measured, it
was sampled using a small boat equipped with a pumping system coupled to a
Turner Model 111 fluorometer having a flow-through door. Water was drawn
from a single depth through the fluorometer while the boat traversed the
dye at a constant speed. The boat position was obtained using a microwave
navigation system. Dye concentration and position information (and
temperature, when appropriate) were recorded on a strip chart recorder or
on an automatic digital data acquisition system. Vertical profiles of dye
concentration were obtained by either taking bottle samples from several
depths or by lowering the water intake hose and recording the fluorometer
output for a series of intake depths.
Relative dye concentration was measured using an aerial photographic
technique developed by Ichiye and Plutchak (1966).d] With this tech-
nique, aerial photographs of the diffusing dye were taken at frequent
intervals using a standard 9 inch format aerial camera with either black
and white or color film (both were used). In the Lake Michigan experi-
ments, Kodak Tri-X Aerographic film No. 2403 was used with a Wratten No.
61 filter to enhance the contrast between the fluorescein dye and the lake
water. Flight altitudes of 4000 to 5500 feet were used. In the
Massachusetts Bay experiments, Kodak Aerocolor No. 2445 negative color
film was used at altitudes of 2000 and 4000 feet.
Relative dye concentration is proportional to the intensity of the light
in the wave band emitted by the fluorescent dye, assuming low background
levels at that wavelength and uniform vertical structure of the dye. For
dye released near the surface and initially uniformly mixed to some depth,
these assumptions are reasonably accurate. The optical density recorded
on the film negative is related to the intensity of the fluorescent
emission (or the relative dye concentration) through the characteristic
curve for the film. From this curve, the relationship between optical
film density and relative dye concentration was determined. Film density
was then measured densitometrically using precision microdensitometers (a
color microdensitometer was used for the color photography).
The microdensitometers automatically scanned across the film negative and
recorded a signal proportional to the film density. Aperture size of the
densitometer was chosen such that a rectangle approximately 0.5 meter x 2
meters (on the water surface) was viewed at any one time. This rectangle
was oriented such that the larger dimension was along the axis of the dye
patch, resulting in a smoothed densitometer trace. Between 10 and 25
1059
-------
scans across each segment of dye were averaged together to obtain a single
representative measure of dye spread.
In addition to measuring dye concentration, ambient conditions at each of
the study sites were recorded. The results of the ambient measurements
are summarized in Table 1.
Data Analysis
Distribution of a diffusing substance, assuming a uniform flow field with
a constant diffusivity, is described by Csanady (1973)1-3]
where ^ is the concentration of the diffusing substance, A is a constant,
x is the distance from the origin, K is the eddy diffusivity coefficient,
and t is time. Equation 1 describes a Gaussian distribution with its mean
at the origin of the diffusing substance (assumed to be a point source),
and its standard deviation, a, equal to s/2Kt. It is thus possible to cal-
culate a coefficient of eddy diffusivity from knowledge of the concentra-
tion distribution of the diffusing substance and Equation 1.
Calculation of diffusion coefficients from the data collected by the
aerial photograph and boat-based fluorometric techniques is outlined
below (following Tokar, et al., 1975). L^J Densitometric reduction of
photographic data and digitization of boat fluorometric data provided dye
concentration as a function of position. Background fluorescence was
removed from the records. The center of gravity MI (first moment) of
each transect across the dye patch was calculated from:
Mi •
f x. 4> (x,)
Those transects taken at about the same time and on the same leg of the
"T" were averaged together to produce an average concentration distribu-
tion across the leg. From this average concentration distribution, the
variance, M2 (second moment), was calculated from:
f (xrM,)2 tUJ
1060
-------
Theoretically, if enough dye patch transects were synoptical ly recorded,
the average concentration curve would approach a Gaussian distribution.
In practice, this was approached in the aerial photographic results,
unless strong current shears distorted the dye motion. Fluorometric
measurement of dye patches, however, was limited by the time necessary to
make the measurements, and only a limited number of transects could be
made in a short period of time; therefore, the Gaussian assumption was
less reliable. After the second moment was calculated, the eddy diffusion
coefficient, K, was calculated as
where a^2 = ^ (t]_) is the variance of the distribution at time, ti, and
Og2 is the variance of the distribution at time tg.
Results and Conclusions
Results of horizontal measurements of nearshore dye diffusion spanning
time periods up to 6 hours and space scales up to several hundred meters
are shown in Table 2. Observed values for the eddy diffusion coefficients
calculated from Equation 4 ranged from about 100 cm^/sec to about 5000
cm2/sec for both the lateral and longitudinal directions. No strong
evidence for a significant difference between these two directions was
observed in those experiments which yielded data on both directions (those
studies using the "T" shaped dye patches).
Both the Lake Michigan studies and the Massachusetts Bay studies resulted
in a similar range of values for eddy diffusion coefficients. No signifi-
cant difference between the data sets is distinguishable, although lateral
diffusivities seen on the lake were slightly lower than those observed at
the oceanic site. Comparison between diffusing dye released in the far-
field thermal plume and dye released in nearby ambient waters showed no
coherent difference in calculated eddy diffusivities.
The effects of existing oceanic and meteorological conditions on the
observed diffusivities are not apparent in the data. In general, the
range of wind speeds, wave conditions, and current speeds seen during the
studies do not correlate in any obvious way with the measured diffusion
coefficients. On October 27, 1974, in the Massachusetts Bay studies,
conditions of strong winds and high waves (~2 meters) forced a halt to the
boat measurements, and aerial measurements were cut short due to rapid
disappearance of the dye. This was apparently attributable to increased
vertical mixing due to wave action; even though horizontal diffusivities
were at the high end of the measured range, they were not large enough to
account for the rapid dispersal of the dye. Similar horizontal diffusivi-
ties were observed on October 29 and 30, 1974, when winds were light and
wave heights were minimal.
1061
-------
Figure 2 contains diffusion diagrams (after Okubo, 1971)[8] showing eddy
diffusivity plotted against a length scale, L, defined as 4g. The line
labeled Okubo on these figures shows the results of a large number of
oceanic dye diffusion studies.
In general, the eddy diffusivity measurements made in Lake Michigan and
Massachusetts Bay produced results similar to these oceanic measurements
with respect to the rate of increase of the diffusion coefficient as a
function of patch size. At a particular patch size, however, the near-
shore measurements show higher diffusion rates than the oceanic data
indicates. The results of Murthy (1970)L8] and Huang (1971)L5J taken
in the Great Lakes indicate diffusivities as a function of size very
similar to the nearshore results shown here.
Vertical dye measurements made in Lake Michigan indicate values of verti-
cal diffusivity ranging from 0.3 to 2.7 cm2/sec. A single set of profiles
obtained in Massachusetts Bay indicates little or no vertical mixing
following an initial mixing to a depth of several meters and most of the
Lake Michigan data are amenable to this interpretation also. Thus,
our conclusions are that vertical diffusion under conditions described
here does not exceed 3 cm^/sec after an initial mixing period and may
actually be less than this value. This result has been observed pre-
viously on the Great Lakes (Csanady, 1973).L^J
Table 3 summarizes the comparison between fluorometric and photographic
measurement techniques employed in the Massachusetts Bay experiments.
Fluorometric and aerial photographic determinations of the standard devia-
tion of the dye distribution correspond well for most of the measurements.
Inconsistencies in the measured values such as at 1048 on October 28,
1974, were probably the result of distortion of the "T," resulting in
fluorometric measurements at inappropriate locations. The aerial data
collection method has much to recommend it, including ease of data collec-
tion and comprehensive spatial results, though it does lack the
sensitivity of the fluorometric technique.
Summary
Results of" short-term, nearshore dye studies in Lake Michigan and
Massachusetts Bay indicate a range of horizontal diffusivities between
about 100 cm2/sec and 5000 cm^/sec over time scales of 0.1 to 6 hours.
These results were obtained under calm to moderate conditions, about 1 km
offshore in waters about 10 meters deep. They agree well with previous
data taken on the Great Lakes, but indicate slightly higher diffusivities
at a particular scale size than are generally seen at oceanic sites. The
increase in eddy diffusion coefficients as a function of scale size is
similar to that observed at oceanic sites. While the short time period of
these observations limits their usefulness, the measured values indicate a
range of diffusion coefficients applicable to nearshore waters under calm
to moderate conditions.
1062
-------
Vertical diffusivity of less than 3 cm2/sec was observed on several
days; but a meaningful numerical result was not obtained, since most of
the data are also amenable to an interpretation invoking rapid vertical
mixing throughout a well-mixed layer of some depth followed by an
extremely low rate of vertical mixing. This well-mixed depth was of the
order of 2 meters, but is probably a function of the wind and wave
conditions present during the study.
The use of aerial photography to obtain quantitative results for diffusion
processes appears to be a valuable technique. Limited time periods and
sensitivity may reduce its usefulness, but in some nearshore applications
it can result in significantly better and more easily obtained results
than boat-based fluorometric techniques. One of the common problems in
making fluorometric measurements from small boats is the lack of a visual-
ization of the gross nature of the dye motion. This often results
in poorly run experiments and inaccurate results, which the aerial
technique can help eliminate. In practice, a combination of the two
techniques results in the most accurate and convincing measure of dye
mixing.
Acknowledgments
The authors are pleased to acknowledge the members of the Argonne National
Laboratory Great Lakes Project for support and assistance in the Lake
Michigan measurements, and members of the EG&G, Environmental Consultants
staff on the Massachusetts Bay program. The Lake Michigan portion of the
work was sponsored by the U.S. Energy Research and Development Agency
(ERDA). The Massachusetts Bay portion was sponsored by ERDA, Public
Service Electric and Gas Company, Electric Power Research Institute,
Boston Edison Company, New England Power Company, and the Commonwealth of
Massachusetts Division of Water Pollution Control.
REFERENCES
1. Ichiye, T. and Plutchak, N.B., "Photodensitometric Measurement of Dye
Concentration in the Ocean," Limnology and Oceanography, 11(3): 364,
July 1966.
2. EG&G, Environmental Consultants, "Phase II Final Report, Forecasting
Power Plant Effects on the Coastal Zone." Report B-4441, Waltham,
Mass., 1976.
3. Csanady, G.T., "Turbulent Diffusion in the Environment," D. Reidel
Publishing Company, Boston, 1973.
4. Murthy, C.R., "Complex Diffusion Processes in Coastal Currents of a
Lake," Journal of Physical Oceanography, 2:80, 1972.
5. Huang, J.C.K., "Eddy Diffusivity in Lake Michigan," Journal of Geo-
physical Research, 76(33): 8147, November 1971.
1063
-------
6. Eliason, J.R., Daniels, D.G., and Foote, H.P., "Remote Sensing
Acquisition of Tracer Dye and Infrared Imagery Information and Inter-
pretation for Industrial Discharge Management," Pacific Northwest
Laboratories of Battelle Memorial Institute, March 1971.
7. Tokar, J., et al., "Measurements of Physical Phenomena Related to
Power Plant Waste Heat Discharges: Lake Michigan, 1973 and 1974,"
Argonne National Laboratory, ANL/WR-75-1, 1975.
8. Okubo, A., "Oceanic Diffusion Diagrams," Deep Sea Research 18:789-802,
1971.
9. Murthy, C.R., "An Experimental Study of Horizontal Diffusion in Lake
Ontario," Thirteenth Conference on Great Lakes Research, Buffalo, New
York, March 31 - April 3, 1970.
TABLE 1. AMBIENT CONDITIONS DURING DIFFUSION MEASUREMENTS.
Site
Point Beach 1
Point Beach 2
Point Beach 3
Campbell 1
Cook 1
Pilgrim 1
Pilgrim 2
Pilgrim 3
Pi Igr im 4
Pi Igr im 5
Date
8-9-73
8-23-73
8-24-73
9-12-73
10-23-73
10-25-74
10-27-74
10-28-74
10-29-74
10-30-74
Wind Velocity
Time Tide Stage (m/s)
1700-2000 5 at 225'
1430-1700 5 at 175'
1000-1200 3 at 120'
1356-1444 2 at 195*
1345-1520
1330-1430 Ebb 8 at 210*
0935-1309 High-Ebb 6 at 300'
1000-1603 Ebb-Low Variable
0915-1045 High 4 at 215'
0940-1040 High 2 at 200'
Current Velocity
Wave Height (Near-surface)
(m) (CB/S)
5
13
0.5 8
0.2 13
5
0.3 7
1.5 3
7
8
11
at
at
at
at
at
at
at
at
at
at
240'
045'
015'
200'
200*
280*
255'
150'
300'
310'
Water Temp.
(Near-surface)
CC)
10
18
18
11
14
10
11
10
10
10
Air Temp.
CC)
24
16
20
14
18
16
13
8
IS
17
-------
TABLE 2. RESULTS OF DIFFUSIVITY MEASUREMENTS.
Site
Point Beach 1
Point Beach 2
Point Reach 3
Campbell 1
Cook 1
Pilgrim 1
Pilgrim 2
Pilgrim 3
Pilgrim 4
Pilgrim 5
Diffusion
Time
(s)
6000
5000
2500
4600
3100
3500
4300
4900
5200
6000
1260
2280
2880
1500
4500
5700
1200
1500
1680
3600
1560
4860
6180
11640
12840
840
1800
3600
5700
6240
7200
7560
9000-
10200
10620
10800
11880
18300
19980
21780
2880
3420
5400
960
1740
2940
3660
KX
(an2/s)
_ ._ _
....
290
480
1130
600
2160
5000
....
302
487
3983
3524
481
770
1167
1096
1326
841
3052
1108
2579
5864
3734
5153
5323
2763
3778
137
2002
2055
3142
KY
(cm2/s)
390
75
3500
2500
12 ,000
130
65
470
1260
2600
473
1798
3740
....
2286
5026
747
1782
284
691
....
624
343
872
1520
— i.
14
110
280
1621
2487
•"— •" —
KZ
(cm2/s)
0.5
2.7
1.4
0.4
—
1.4
0.3
—
0.3
_._
—
—
...
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
--
--
—
—
—
—
—
—
—
...
Technique
f luorometry-
continuous injection
f luorometry-
continuous injection
fluorometry-
batch injection
photography?T-shaped
injection
photography/T-shaped
injection
photography/T-shaped
injection
photography and
f 1 uorometry /T-shaped
injection
photography and
f 1 uorometry/T-shaped
injection
photography of T--
shaped injection
Comments
thermal plume,
ambient waters
thermal plume,
ambient waters
thermal plume,
ambient waters
after an initial
mixing period,
no changes in
vertical dye
distribution
were observed.
1065
-------
TABLE 2. RESULTS OF DIFFUSIVITY MEASUREMENTS.
Site
Point Beach 1
Point Beach 2
Point Beach 3
Campbell 1
Cook 1
Pilgrim 1
,
Pilgrim 2
Pilgrim 3
Pilgrim 4
Pilgrim 5
Diffusion
Time
(s)
6000
5000
2500
4600
3100
3500
4300
4900
5200
6000
1260
2280
2880
1500
4500
5700
1200
1500
1680
3600
1560
4860
6180
11640
12840
840
1800
3600
5700
6240
7200
7560
9000
10200
10620
10800
11880
18300
19980
21780
2880
3420
5400
960
1740
2940
3660
KX
( cm2/s)
290
480
1130
600
2160
5000
302
487
3983
3524
481
770
1167
1096
1326
841
3052
1108
2579
5864
3734
5153
5323
2763
3778
137
2002
2055
3142
(cm2/s)
390
75
3500
2500
12 ,000
130
65
470
1260
2600
473
1798
3740
____
2286
5026
747
1782
284
691
624
343
872
1520
....
14
110
280
....
1621
2487
....
(cm2/s)
0.5
2.7
1.4
0.4
—
1.4
...
0.3
—
0.3
...
—
—
—
—
—
...
—
—
—
—
—
—
...
—
—
—
—
—
—
—
—
—
—
—
—
—
—
...
—
—
...
—
—
...
Technique
f luorometry-
continuous injection
fluorometry-
continuous injection
fluorometry-
batch injection
photography/T-shaped
injection
photography /T-shaped
injection
photography/T-shaped
injection
photography ancl
f 1 uorometry /T-shaped
injection
photography and
fluorometry/T-shaped
injection
photography of T-
shaped injection
Comments
thermal plume,
ambient waters
thermal plume,
ambient waters
thermal plume,
ambient waters
after an initii
mixing period,
no changes in
vertical dye
distribution
were observed.
1066
-------
TABLE 3. FLUOROMETRIC METHOD VERSUS AERIAL PHOTOGRAPHIC METHOD.
Date Time x-Aerial x-Boat y-Aerial y-Boat
10-27-74 1001
1015 17.9
1030
10-28-74 1027
1030 16.9
36.8
14.9
25.8
10.0
28.1
1033
1048
1100
1135 36.6
1144
1200 44.4
1206
1250
1257
1300
19.7
22.1
43.0
41.6 53.0
36.4
34.8
34.7
73.5
59.4
49.2
Figure 1. Results of Photography of a T-shaped Dye Patch. Motion,
Diffusion, and Distortion of the "T" as a Function of Time is
Shown in This Figure. Graphs Show Relative Dye Concentration
(Averaged Over 10 Densitometer Scans) For Each Leg of the
"T."
1067
-------
o
01
<
^v
CM
E
o
* IO4
>-
l_
>
^
u.
UL
O
O Irt3
Ijj IU
_J
CC
111
H
I02
IO1
I SYM DATE SITE
r -f 10/25/74 PILGRIM 1
• A 10/27/74 PILGRIM 2
' • 10/28/74 PILGRIM 3
. • 10/29/74 PILGRIM 4
* 10/30/74 PILGRIM 5
0 8/9/73 POINT BEACH 1
: V 8/24/73 POINT BEACH 3
- 0 9/12/73 CAMPBELL 1
; D 10/23/73 COOK 1
_
•[
- A j
c£. /
** •^r
r 0 -/^-OKUBO
" ^ X
r *
7
! "
i i i 1 1 1 1 1 1 i i i 1 1 1 ni i i i 1 1 1 1 1
IO2
I06
IO3
IO4
I05
L(cm)
I05
E
o
t IO4
>
in
Q
O
I03
z
a
o
o
I02
10'
IO2
D
D
OKUBO
i i 1111
IO3
IO4
L(cm)
IO5
Figure 2. Dye Diffusion Coefficient as a Function of Dye Patch Size "L" = 4a.
Okubo Shows Results of Numerous Oceanic Studies.
Line Labeled
-------
EFFECTS OF BOTTOM SLOPE, FROUDE NUMBER, AND REYNOLDS
NUMBER VARIATION ON VIRTUAL ORIGINS OF SURFACE JETS:
A NUMERICAL INVESTIGATION.
by
J. Venkata, S, Sengupta, S. S. Lee
University of Miami
Coral Gables, Florida U,S,A.
(ABSTRACT)
A i diiru-.iioional numerical model which was developed to predict
thr behaviour of thermal discharges is used to investigate the
ef->cts of variation of bottom slope, Reynolds, number and Froude
number on the virtual origins of surface jets. Two types of
virtual origins, one based on the jet width and the other based
on the centerplane velocity and temperature decay are considered
Th-1 results Indicate that jet width is independent of bottom
Slope. Increasing Reynolds number moves the virtual origins
upstream of the discharge point.
1069
-------
INTRODUCTION
Soirje of the three dimensional numrrrlCu-i models which have been
developed recently for velocity a.nd temperature predictions are
those of Brady & Geyer (1972), Till, J. (1973), Waldrop & Farmer
•(1974), Sengupta & Lick (.197*0 , Paul & Lick (197*0, and Markham
(1975). An excellent, review of numerical models is given by
Policastro et al (1975), The model that was used to obtain re-
sults in this paper is the modified version of the model developed
by Sengupta & Lick (197^0, The modified version of the model was
successfully applied and verified by the authors (Venkata, J &
Sengupta? 1977f Matha,van & Lee, 1977 and Sengupta & Lee 1976).
This paper is concerned with the numerical investigation of the
effects of bottom slope, Froude number and Reynolds number vari-
ation on the virtual origins, therefore, so the details of the
verification of the model will not be discussed here. For the
verification of the model details the reader is advised to look
into the references (7,8 and 12),
Two types of virtual origins have been defined for free jets.
One is based on the widening of the jet and is known as geometric
virtual origin, and the second, based on the centerplane velocity
decay, is known as kinematic virtual origin (Flora & Goldschmidt,
1969). It was found experimentally that the two origins of si-
milarity do not coincide and do not depend on the discharge chan-
nel aspect ratio /N but, found to depend on the turbulence in-
tensity (Flora & aSldschmidt, 1969; Jenkins & Holdschmidt, 1973
and Kostovinos, 1975). In the present investigation it was at-
tempted numerically to see the possible influence of bottom slope,
Froude number and Reynolds number variation ori the geometric and
kinematic virtual origins for incompressible surface jets.
THEORY
The flow of an incompressible surface jet entering into a quies-
cent atmosphere of the same fluid Is considered. Fig.(l) shows the
discharge canal a.nd receiving basin geometry considered along with
the boundary conditions which are discussed elsewhere in this paper.
The equations that describe the motion of the fluid and^heat trans-
fer for incompressible fluids are the three Navier-Stokes equations
of momentum, conservation of mass, conservation of energy and equ-
ation of state, which couples energy equation to momentum equations.
These equations are stretched in the vertical direction using the
relation j=Z/h(x,y). The details of the stretching are discussed
in reference (12). The advantage of this stretching is that it
allows the same number of grid points at shallow and deeper parts
of the basin without variable grid spacing. The other approxima-
tions that are made before the final set of equations obtained are:
(1) The vertical equation is replaced by the hydrostatic equation,
1070
-------
(',:) A rigid-lid is plaaed on the top of the surface which allows
horizontal velocities but not vertical velocities. In order to
determine the pressure on the surface, a Polsson equation is deri-
ved from the two horizontal momentum equations. (3) The fluid
is treated as incompressible; the coupling between momentum
-and energy exists through the equation of state. (4) The effect
of turbulence is modelled using eddy transport coefficients. The
final set of non-dimensional, stretched equations which are si-
milar to Sengupta (1974) are given below.
Continuity:
• M™""* U / -\ \
3y (1)
u-Momentum:
3(hu) + 3(huu) +3(huv) + .aj^u) _ h_
3t 3a 3g 3y Rg
3P i -,
= _ h S - h B + — — (h—) + — —(h—)
3a X Re 3av 3a; Re 3BV 36'
2~ T
v-Momentum:
Hydrostatic Equation:
1071
-------
F I? (-Ayi - Ay2 4 S - V
h L dot 3a 33 33 3t v
where
h
Ax-j = / {— (uu) + |- (uv) + |y (uw) * dz
o
h
Axo = HD / udZ
c Ko
0
h
1 5 * an a * an 1
r = — r r— f A — } + — f A °u- \ + _L
x Re j 1 3x l MH 3x ; 3y { MH 3y ; _2
/ l ax ( PdZ) ' dZ
o
ENERGY EQUATION:
3_(hT) a(huT) + 3(hvT) + ,
' 9t 3a 33
_ _ . J _ 1 3_ /R*
Pi" 3a W Pe ae 2 F~3Y l v
EQUATION OF STATE:
P = 1 . 02943] - . 00002QT- . OOOOO^T2 ( 6 )
The Poisson equation for pressure is of the form
? 2
3 P 3 P -
1072
-------
and
/{^ M
0
- ivv) + (vw) } dZ
Ay =
1
udZ
Cy Re
L_l_
e2 3Z
<\ if
Vp = E" } {
WHERE
dZ
dZ
—
u = D — ; v =
Uref
; W = n]
ref b uref
ref
x =
L '
£-•• z = fr
P =
pref Uref
'"'rpf P'
• T re' . p =
> i —f > p —
'ref
pref
* A
H = •
A
v =
ref
B* _ BH . Bv _ BV
> v -
ref
B
ref
B
ref
1073
-------
This set of equations are to be solved with appropriate initial
and boundary conditions'. The initial conditions used on the
velocity are, at time t = 0, all velocities are zero. The temper-
ature at t=0 is equal to the reference temperature. The boundary
conditions are schematically presented in Figure (1). The condi-
tions on solid walls and bottom are no slip and no normal velocity
for all time, except w is not equal to zero due to the hydrostatic
approximation. The temperature boundary condition at solid walls
is handled by assuming the walls and bottom as adiabatic i.e.
8 T —
=. Where n is in the direction normal to the wall or bottom.
°n
The boudary condition at the open boundaries used in this investi-
gation is ^Yr =0 where v is velocity in the direction normal 'to
3n
the boundary. The boundary conditions in summary for the verti-
cally stretched co-ordinate system are
Boundary Conditions:
On solid lateral wall :
u = 0
v = 0
n f 0
_ =
ay 93 " h 93 9y
At the bottom of the basin (Y=!)
n = 0
u = 0
v = 0
Along free boundaries:
At (a = 0, g, Y)
UI=1, K =UI=2, K
v = 0
1074
-------
w i- 0
TI=1, K=TI=2, K
PS = constant
At (a = aL, 6, K)
UI=IN, K = UI=IN-1, K
v = 0
w i- 0
TI=IN, K= TI=IN-1, K
P = constant
At (a, 6 " B, Y)
UJN, K = UJN-1, K
v = 0
w i- 0
P = constant
At the air water interface
n = 0 (Rigid lid)
gu , hH ,
hHKc
s } (T - T
U '
1075
-------
The equations 1 to 7 are solved with the above boundary conditions
using finite difference approximations on a UNIVAC HOC Computer.
Computer Simulations
The list of cases run is given in Table (1). First a constant
density jet entering a constant depth .basin is studied for a
Reynolds number equal to 100. Then for the same Reynolds number
the bol''Mti is changed from constant depth to smoothly sloping
bottom i, Van 6 = 0.00*1) and is studied for a constant density jet.
The slope is then doubled (Tan 6-0.008) and the above case is
repeated for a constant density jet at Re=100. All the above
cases are run until steady state is reached. It took approxi-
mately 65 minutes to reach steady state. The jet width (b/D)
and centerline velocity in the form /Uo.2 are plotted against
centerline distance (^) for the above £hree cases and are shown
in Figures (2 to 7). The geometric and kinematic virtual origins
are obtained in the following manner. A straight line is fitted
in the near region of the jet and the straight line is extended
to cut the x-axis. The intercept gives the geometric virtual
origin for the jet width diagram and kinematic virtual origin
for the centerplane velocity decay diagram.
What is interesting from these figures is that the geometric vir-
tual origin, and, hence, jet width,.do not seem to depend on the
bottom slope. Where as kinematic virtual origin is increasing
(moving upstream of the discharge point) indicating that the surface
centerline velocity decreases more rapidly with increase in slope.
As the bottom slope increases there is more bottom entrainment
causing the jet velocity to decay at a more rapid rate.
The next step was to consider how the jet behaves when density
effects are included. The above three cases are repeated inclu-
ding the effects of density (i.e. Froude number is changed) and
keeping the Reynolds number the sa.me. The results of these
cases are shown in Figures (8 to 13)- Again it can be seen that
for the same Reynolds number the geometric virtual origin and
hence the jet width is independent of bottom slope. The kinema-
tic virtual origin increased as before indicating that surface
centerplane velocity decreased more ra.pidly with increased bottom
slope. But, an important effect of including density can be
found by comparing the cases with variable density/to that of
cases with constant density. The geometric virtual origin increased
(i.e. moved towards the discharge point) for variable density cases
This is because the jet is spreading in the lateral direction more
rapidly because of density differences between the discharged fluid
and ambient fluid and consequent spreading in a thinner layer at
the surface. The kinematic virtual origin decreased (i.e. moved
upstream) for cases with variable density indicating that the
surface center plane velocity decay is slower than that of the
1076
-------
case with constant density. This is because the fluid is rising
due to buoyancy in the cases where density effects are included,
causing flow through a smaller effective cross-section.
In all the above six cases, the Reynolds number (Re) is kept
constant equal to 100. Its effects are studied by increasing
Re from 100 to 285 by decreasing the reference eddy viscosity.
The results are plotted and ar.e shown in. Figures (14 to 19).
It can be again observed here that the geometric virtual origin
is independent of bottom slope, and kinematic virtual origin is
dependent on bottom slope. The important effect of' Reynolds
number on virtual origins, as can be seen from Table (1) is,
geometric and kinematic virtual origins moved upsream with
the increase in Reynolds^number. The results of all the above
nine cases are summarized* In Figures (20'to 23).
CONCLUSIONS
From the different cases studied, the following conclusions can
be drawn:
(1) From the constant depth and two bottom slope cases, it is
concluded that for increasing bottom slope, the decay of
centerplane velocity and temperature is faster due to in-
creased entrainment, where as j-et width would be Indepen-
dent of bottom slope. Also, it is noticed that geometric
virtual origin is independent of bottom slope, and kine-
matic virtual origin decreases with increase of slope.
(2) Comparison between non-buoyant and buoyant jets indicate
that geometric virtual origin for non-buoyant jets is more
upstream than for buoyant jets. The kinematic virtual
origin moves further upstream when density effects are
included.
(3) It Is found that increasing the Reynolds number moves the
geometric and kinematic virtual origins further upstream
of the discharge point,
ACKNOWLEDGEMENTS
This work was conducted under funding from National Aeronautic
and Space Administration, Kennedy Space Center.
1077
-------
REFERENCES
1. Abramovich, G.N., "The Theory of Turbulent Jets", The
M.I.T. Press, Cambridge, Mass., 1963.
2, Brady, D.,-and Geyer, J., "Development of General Com-
puter Model for Simulating Thermal Discharges in Three
Dimensions", Report No.7, Dept. of Geography and
Environmental Eng., Johns Hopkins University, Baltimore,
Md., (1972).
3. Dunn, W.E,, Policastro, A.J., and Paddock, R.A., "Sur-
face Thermal plumes: Evaluation of Mathematical Models
for the Near and Complete Field", Water Resources Re-
search Program ,• Energy and Environmental Systems Divi-
sion, Argonne National Laboratory, Argonne, Illinois
(Part one and two), 1975-
4. Flora, J., and Goldschmidt, V., "Virtual Origins of
a Free Plane Turbulent Jet", A.I. A. A., Journal 7, PP
23^-2346.
5. Jenkins, P.E., and Goldschmidt, V., "Mean Temperature
and Velocity in a Plane Turbulent Jet", A.S.M.E.,
Journal of Fluids Engineering, 95, PP 581-584.
,6. Katsovinos, N.E., "A Mote on the Spreading Rate and
Virtual Origin of a Plane Turbulent Jet:, Journal of
Fluid Mechanics, 1967, Vol.77, Part 2, pp 305-311.
7. Lee, S.S., and Sengupta, S., "Proceedings of the
Conference on Waste Heat Manatement and Utilization",
Miami Beach, Fla., 9-11 May 1977-
8. Mathavan, S.K.M., "Experimental and Numerical Study
of Current and Temperature Fields in Lake Belews an
Artificial Cooling Lake", Ph.D. Thesis Submitted to
the Department of Mechanical Engineering, University
of Miami, Coral Gables, Florida, August 1977-
9. Paul, J., and Lick, W.J., "A Numerical Model for a
Three Dimensional Variable Density Jet", /FTAS/TR
73-92, Case Western Reserve University (1974).
10, Sengupta, S,, 'and Lick, W.J., "A Numerical Model for
Wind Driven Circulation and Temperature Fields in
Lakes and Ponds", FTAS/TR-7^-79, Case Western Reserve
University (1974).
11. Sengupta, S., Lee, S.S., Venkata, J., and Carter, C.,
1078
-------
"A Three Dimensional Rigid Lid Model for Thermal
Predict.ionG11 ? presented at the Waste Heat Management
and Utilization, Miami Beach, Florida, May 9-11, 1977-
12. Venkata, J., "A Numerical Investigation of Thermal
Plumes", Ph,D, Thesis Submitted to the Department of
Mechanical Engineering, University of Miami, Coral
Gables, Florida, August 1977-
13, Waldrop, VJ.K,, and Farmer, R. , "Three Dimensional Com-
putation of Buoyant Plumes", Journal of Geophysical
Research, Vol,74, No,9, (March, 1974).
1079
-------
P=Constant
Y,B,J
v=o
Z,Y,K
I
w
C'3
OH
•
(TOP VIEW)
i
4J H
C II
CtJ *~3
-p >
W II
c ?^
O M
O II
'II 1-3
0, >
I
1 !
1
! 1
r-H
1
g
M
II
l—J
EH
II
2
M
0 II
II I-J
P=Constant
ui=
V=0
UI=IN~UI=IN-1
[ "Y,3,J
i w
s CD T =T
T ** TANGENTIAL FLOW ONLY
l-'H , F
Z Y K o L »•
^ > I >^ rn • h
(SIDE VIEW)
Q
X,a,I
r: f\i
aj II
X II
O r-
O II O II |
II H II H|
NO SLIP
li* TANGENTIAL FLOW ONLY
HH
E-1 !
" i (VIEW ALONG SECTION J-J)
H
II
rH
I
&
M
I O H H
I O II O II
I II hH II H
NO SLIP
Fig.l Boundary Conditions for
the Region of Computation
1080
-------
LIST OF'CASES STUDIED AND VIRTUAL ORIGIM RES'JLTS OBTAINED
2ase No
:>r Run
1 No
I
2
3
4
5
6
7
8
9
Slope
Constant Depth
(Slope=0.0)
Low Slooe
(Tan0=G~. 004)
High Slope
(Tan0= 0.008)
Constant Depth
(Slope 0.0)
Low Slope
(Tan0 = 0~. 004)
High Slope
(Tan0=0. 008)
Constant Depth
(Slope 0.0)
Low Slope
(TanG=0.004) "
High Slope
(Tan0=0.008)
Froude Number
(Fr) = Uo
/Ao o-ho
/ gr.o
P
0". 053
(Const Density
0.058
(Const Density
0.058
(Const Density
0.0208
(Variable
Density)
0.0208
(Variable
Density)
0.0203
(Variable
Density)
0.058
(Const Density
0.058
(Const Density
0.058
(Const Density
i
Reynolds Number
(Re)
TT T
= ref
Aref
100
100
100
100
100
100
285
285
285
Geometric Virtual
Origin
(C )
1.8
1.8
i:8
0.8
0.8
0.78
11.0
11.0
11.0
Kinematic Virtual
Origin
c2)
3-0
2.8
2.2
4.0
3-5
3-0
14.4
5-5
3-0
-------
-5 -4 -3 -2
9
8
7 -
6 -
5
4
3
2
Conntant Depth i1.2ra
Discharge Ve'l »20cnv/sec
Donoity «Constant
A—* 2
Ra
t
i10,000cm /Bee
tlOO
«4hra I5min
total
K2- 0.271, C2- -0.124 x 25
I
I
4 5
X/D
10
Fig.2 Kinematic Virtual Origin (Constant Depth)
Constant Depth :1.2rn
Discharge Vel :20cm/sec
Density :Constant
fl _ _ _ f ^ /^ /i /-i i»\ *
"ref
Re
^otal
K - 0.343,
:10,000cm /sec
: 100
:4hrs ISmin
-0.072X 25
8
7
6
5
4
X/D
I
Fig.3 Geometric Virtual Origin (Constant Depth)
1082
-------
1?
1),
1C
e.
!_._. 1
' Discharge Width: 25ra
' Discharga Vel. : 20 cm/sec
/ Density : Constant
/
K2 " 0.3028
-0.116 x 25
-5 -4-3-2-10 12 34 5 67 8 9 10 11
Fig. 4 Kinem
Tan 0
ll
10
9
b 8
D
7
6
5
4
3
2
1
atic Virtual Origin (For Sloping Botton
=0.004)
^-^
-
Discharge Width:. 25m
Discharge Vel. : 20 cm/sec
Density : Constant
KX " 0.28. CJL -= -O.Q72 x 25
-
-
-
^^£-+
^^^^^^ ^
i i i i i i i it l .
-5-4-3-21-1- 0 1 2 3 4 5 6 7 8 9 10
hH g
Fig.5 Geometric Virtual Origin (For Sloping Bottom
Tan 0=0.004)
1083
-------
Discharge Width i 25m
Discharge Vol. : 20 cm/sec
i Constant
C « 0. -0.033 x 25
5-4 -3-2-10 1 2 3 4 5 67 8 9 10
Fig.6 Kinematic Virtual Origin (For Sloping Bottom
Tan 0 =0.008)
b
D
11
10
9
e
7
6
1
^
Discharge Width : 25m
Discharge Vel. : 20 cm/sec
Density « Constant
t
KX = 0.28. CJL" -0.072 x 25
t i i i i
' < '
L
5-4-3-2-10 1 2 3 4 5 6 7 8 9 10 11 12 13
3t
O
Fig.7 Geometric Virtual Origin (For Sloping Bottom
Tan 0=0.008)
1084
-------
Constant Depth
Discharge Vcl:20cm/s«e
Discharge Temp:35.9°C
"ref
Re
t
total :65min
Kinematic Origin :-4
-0.16x 25
-------
Slope Case
Discharge Velocity
Discharge Tamparature
Re
: 20cm/
-> r n^
sec
: 10,000cm /sec
:100
total :65min
Kinematic.Origin s-3.4
K-- 0.294, C-- -0.136x 25
-4 -3 -2-1 01
23456789
X/D
Fig.10 Kinematic Virtual Origin (For Sloping Bottom
ian 0—0.004)
Low Slope Case
Discharge Velocity :20cm/sec
'Discharge Temparature :35.9°c
Aref :10,OOOcm2/Eec
Re :100
total :65min
Geometric Origin :-0.8
0.714, C,= -0.032X 25
234567S9
-4 -3 -
Fig. 11
Origin (For Sloping Bottom
1086
-------
High Slope Caoe
Discharge Velocity
Discharge Temparatura :35.9°C
-4
"ref
Re
^otal
Kinematic Origin
tlO,000cm />ec
jlOO
t 6Smin
:-3
K2=0.33,
-0.12 x 25
Fig.12, Kinematic Virtual Origin (For Sloping Bottom
Tan 0=0.008)
High Slope Case
Discharge Velocity
Discharge Temparatuo
Aref
Re
t
: 20cTr/sec
:35.9°C
j10,000cm /sec
jlOO
total :65min
Geometric Origin :-C.73
0.714, C= -0.032X 25
li i i XI I I I I I I i a U
-4 -3 -2 -1 0 1 2 3 4 5 6 7 B 5^
X/D
Fig.13 Geometric Virtual Origin (For Sloping Bottom
Tan 0=0.008)
1087
-------
CONSTANT DEPTH
Discharge Vel
Discharge Width
max
«20 en/sec
s25 m
i 500 m
i425 m
:3,500 en /sec
j235
total :65min
K2 - 0.069, c_- -0.576X 25
P.£f£olds No (Re)
t
-16 -14 -12 -10 -8-6-4-2 0 2 4 6 8 10 12 14 16 13 20
X/D
Fig.14 Kinematic Virtual Origin (Constant Depth)
CONSTANT DEPTH
Discharge Vel
Discharge Width
Rt-.ioldg No (Re)
t
total
K1=0.09, c," -0.44 x 25
=1.2 m
i 20 cm/sec
: 25 m
: 500 m
i 425 m ,
: 3,500 cm /sec
: 285
« 65 min
-11 -10 -9 -8 -7 -6-5-4-3-2-1 01 2 3 4 5
X/D
Fig.15 Geometric Virtual Origin (Constant Depth)
1088
-------
LOW SLOPE
Discharge Vel
Discharge width
L
.max
ref
Reynolds No (Re)
ttotal
i 20 en/sec
i 25 meters
i SOOmeters
: 425 meters
,3500 cn2/sec
,285
: 65 minutes
-0.22 x 25
-6 r5 -4 -3 -2 -1 0.0 12
345
X/D
10 11
Fig.16 Kinematic Virtual Origin (For Sloping Bottom
Tan 0=0.004)
13.0
12.0
11.0
10.0
I
9.0
(b/0) 8.0
7.0
6.0
5.0
4.0
3.0
LOW SLOPE
Discharge Vel:
Discharge width
,
ref
Reynolds No (Re)
fc
: 20 cm/cec
: 25 meters
: SCO rceLers
: 425 naters
2
: 3 , 500 czi /sec
: 285
"total
^=0.09,
: 65 ninutea
-0.44x 25
I I
-11
-10 -9 -a
-7 -6
Kn
-5 -4
-3 -2 -1 0.
D 1
2
345
X/D
C
7
8
9 10
Fig.17 Geometric Virtual Origin (For Sloping Bottom
Tan 0=0.004)
1089
-------
Severe Slope
Discharge Velocity: 20 cm/sec
Discharge Width, D: 25 m ,
i 3,500 cm V»«c
No i 285
: 65 minutes
total
K2-T).3, Cy -0.12 x 25
J L
4.0
3.5
3.0
2.5
2.0
1.5
1.0
X
>.5
_L
J i I I
i il l
-6 -5 -4 -3 -2 -1
4 5
(X/D)
7 8
Fig.18 Kinematic Virtual Origin (For Sloping Bottom
Tan 0=0.008)
Severe Slope
Discharge Velocity : 20 cm/sec
Discharge Width : 25 m
: 3,500 cn\2/sec
: 2B5
Reynolds NO (Re)
= 0.037,
: 65 minutes
-0.44 x 25
-12 -11 -10 -9 -8-7-6-5-4-3-2-10 1 2 3 4 5 6 7
Fig.19 Geometric Virtual Origin (For Sloping Bottom
Tan 0=0.008)
1090
-------
o
<£>
15
Re=285
Re=100
0.0
Fig.20
0.002 0.004 0.006 0.008
~Bottom Slope in radians
0.01
Figure Showing the Relation
Between Bottom,Slope and
Kinematic Virtual Origin for
Two Reynolds Numbers (Constant
Density Jet)
1
1. Constant Depth
2. LOW Slope
3. High Slope
1
I
I
0.0 0.002 0.004 0.006 O.OOJ 0.01
Slope in radians
Fig.21 Figure Showing the Relation
Between Bottom Slope and
Kinematic Virtual Origin for
Re=100 (Variable Density Jet)
-------
APPENDIX
AH horizontal kinematic eddy viscosity
Ay vertical kinematic eddy viscosity
A^ vertical eddy viscosity
iLt
A reference kinematic eddy viscosity
AV VAref
BH horizontal diffusivity
By vertical diffusivity
Bref reference diffusivity
BV VBref
B vertical conductivity pC EL,
z p v
C specific heat at constant pressure
Eu Euler number
f Coriolis parameter
Fr Froude number
g acceleration due to gravity
h depth at any location in the basin
H reference depth
I grid index in x-direction or a direction
J grid index in y-direction or 3 direction
K grid index in z-direction or y direction
k thermal conductivity
K surface heat transfer coefficient
s
L horizontal length scale
P pressure
1092
-------
P surface pressure
5 A
Pr turbulent Prandtl number (5——-)
Bref
Pe Peclet number
Q* heat sources or sinks
Re Reynolds number (.turbulent)
T temperature
T . air temperature
air b
T „ reference temperature
T,-, equilibrium temperature
ilj
t time
t „ reference time
ref
u velocity in x-direction
v velocity in y-direction
w velocity in z-direction
x horizontal coordinate
y horizontal doordinate
z vertical coordinate
Greek Letters
a horizontal coordinate in stretched system
6 horizontal coordinate in stretched system
Y vertical coordinate in stretched system
U absolute viscosity
p density
$ dissipation terms in energy equation
T surface shear stress in x-direction
1093
-------
T surface shear stress in .y-dlrection
yz ^
Superscripts
( ) dimensional quantity
'( %) dimensional mean quantity
( '} dimensional fluctuating quantity
( ) dimensional quantity
ref reference quantity
The relation between K-,, K2, jet width and centerplane velocity
decay are given by the following relation
I- vt-v
2 - Vt - V
where
K, = rate of widening of the jet
Kp = s.lope of centerline velocity deca,y
C =• location of the geometric virtual origin from the dis.cha.rge
canal made dimenslonles.3 by discharge canal width
Cp = location of the kinematic virtual origin from the dis.cha.rge
canal made dimensionless by discharge canal width
b = jet width
d = discharge canal width
Urn = velocity at the axis of the jet
Uo - discharge velocity
x = distance along the axis measured from the mouth of th.e jet
1094
-------
METEOROLOGICAL EFFECTS FROM LARGE COOLING LAKES
F. A. Huff and J. L. Vogel
Illinois State Water Survey
Urbana, Illinois U.S.A.
ABSTRACT
A 30-month field program to evaluate atmospheric effects from waste heat
dissipation by large cooling lakes was recently completed in Illinois.
Extensive meteorological instrumentation was employed along with radar and
satellite data in the evaluation. Results indicated that meteorological
effects perpetrated by single power plants are usually insignificant with
respect to initiation or enhancement of clouds, precipitation, and fog
under the climatic and topographic conditions prevalent in Illinois.
Although fog initiation and enhancement are not infrequent in the cold
season, the induced visibility restrictions are seldom severe and the
downwind extent of the lake effect is usually less than 0.8 km.
INTRODUCTION
As the demand for electrical energy increases, many more power plants
will use auxiliary cooling methods, such as cooling lakes and cooling
towers, for the disposal of waste heat. The effect these auxiliary cooling
methods have upon the atmosphere are largely unknown [l]. To measure the
atmospheric effects associated with waste heat disposal from large cooling
lakes, an extensive field program was conducted by the Illinois State
Water Survey under contract with the Electric Power Research Institute
(EPRI). The program was carried out at Baldwin Lake in southwestern
Illinois where an 1800 MWe power plant is operated by the Illinois Power
Company. The investigation centered on the 2200-acre cooling lake and the
surrounding region to determine possible effects on the initiation and
enhancement of steam fog, cloudiness, and rainfall. Some results from
this recently completed 30-month program are presented here.
Baldwin is situated 72 km (45 miles) SSE of St. Louis, Mo., in a temperate
climate characterized by frequent intrusions of cold air, especially in
winter. Most of the instruments were installed by the summer of 1976
and the field program ran until March 1978.
Within the instrument network (Fig. 1), temperature and humidity were
measured at ground level at 21 locations within an area of approximately
50 km^. At five sites, wind was measured at a single level. At five
instrumented towers, three levels of temperature and humidity and two
levels of wind were recorded. Water temperatures were measured at six
sites. Two net radiometers, a recording evaporimeter, recording raingage,
1095
-------
microbarograph, and transmissometer were operated to provide a complete
array of meteorological measurements. A non-recording raingage network
extending within and beyond the basic instrument network was operated also.
Routine visibility measurements and photographs of weather conditions were
made by the project observer. Satellite data from the summer of 1975 were
used to help assess cloud conditions.
STEAM FOG
A major atmospheric effect and one of the most visible effects associated
with cooling lakes is the initiation and/or the enhancement of steam fog
[1, 2, 3, 4]. During the 20-month period from September 1976 to March
1978, 185 steam fog days were recorded by the Baldwin observer (Table 1).
The frequency of these events by season and visibility (intensity) were
further divided into initiation and enhancement days. Enhancement days
were defined as those when natural fog and steam fog occurred simultaneous-
ly and the steam fog significantly reduced the visibility. Initiation
days were those having steam fog with no natural fog present. The maximum
frequency of all steam fog events occurred in winter with a secondary
maximum in fall. The frequency of steam fog was at a minimum in both
spring and summer.
For this study, dense fog was considered to have a visibility of 0.4 km
(0.25 mile) or less. It was felt that such visibilities would be intense
enough to impair normal driving. The Transportation Research Board of the
National Research Council [5] indicates that the performance of a driver
is not affected seriously-until the visibility drops below 0.2 km (600
feet). Thus, the dense fog definition for Baldwin provides a conservative
estimate of the number of times this event could impair normal road traf-
fic, if the steam fog moved from the lake across a road surface.
Dense fog maximized over Baldwin during winter with twice as many
occurrences than any other season. During the winter of 1976-1977 all
dense fogs but one were due to the initiation of steam fog over the
cooling lake. During the winter of 1976-1977 little natural fog formed,
although it tends to maximize during this season [2]. However, during the
winter of 1977-1978 natural fog formed more frequently. Winter enhance-
ment was observed on 10 days, and seven of these were associated with
dense fog. Dense fog over the cooling lake in the other seasons was
associated usually with fog enhancement, rather than initiation.
The frequency and intensities of steam fog initiation maximized during
fall and winter and decreased markedly during spring. Only five incidents
of steam fog initiation were noted during the summer, all with visibili-
ties of 1.6 km (1 mile) or greater. Although the enhancement of natural
fog by steam fog occurred in all seasons, it maximized (unexpectedly)
during the summer of 1977. Normally, the enhancement effect will maximize
during fall and winter, when the climatic maximum of fog days occurs over
1096
-------
Illinois [2]. However, it will have temporal variance since it is
strongly related to the frequency and intensity of natural fog events.
Steam fog over the cooling lake will have only minor impacts upon the
movement of vehicles if it is confined to the boundaries of the lake, and
no roads are built over or immediately adjacent to the lake. However, if
steam fog moves off the lake it can reduce significantly the visibility
across roads and cause problems for motor-vehicle traffic. Only 38% of
all the steam fog events (71 of 185) were observed to travel beyond the
boundaries of the lake, and 78% (55 of 71) did not extend more than 0.2 km
(-Table 2). On days when steam fog was initiated over the cooling lake
only 25% of these fogs moved beyond the boundary of the lake, while nearly
50% (23 of 48) of the enhancement days experienced some horizontal movement
from the lake.
In general, the more intense the steam fog the farther it moved beyond
the lake. All fogs which moved 1.6 km (1 mile) or more were associated
with steam fog that initiated over the lake and the lowest visibility
associated with these steam fog events was 0.4 km (0.25 mile) or less.
On enhancement days no visibility reductions due to steam fog were noted
beyond 0.8 km (0.5 mile). However, it is possible that steam fog could
reduce the visibility farther downwind under certain conditions, especially
if the natural fog formed with nearly calm wind conditions and the steam
fog traveled along natural low-lying areas adjacent to the cooling lake.
Such a situation was not observed with natural fog present, but it was
observed to occur on two days when steam fog was initiated over the cooling
lake. The fog on these days drifted from the lake and traveled by gravity
flow along dry creek beds up to distances of 6.5 km (4 miles) from the
lake, with much reduced visibilities.
The relation between the intensity of steam fog and the horizontal extent
is quite strong. Of the 24 steam fog events which initiated over the
cooling lake with visibility of 0.4 km (0.25 mile) or less, 22 experienced
some horizontal movement. Similarly, 13 of the 21 dense fog cases with
enhancement experienced some horizontal movement from the lake when
visibility was 0.4 km or less.
The initiation of steam fog has been linked to the difference between the
water and air temperatures and the saturation deficit of the ambient air
[2], The water-air temperature difference for initiation days was greater
than on days when the natural fog was present for the same fog intensity.
This is due to the greater saturation deficit of the air on initiation
days. More water vapor has to be evaporated before -condensation in the
air is reached on initiation than on enhancement days. The more intense
steam fogs that initiated during winter formed i\dth ambient saturation
deficits of 1 gm/kgm or less, and the water-air temperature difference
was generally 19-4°C (35°F) or greater.
Synoptically, many of the steam fogs formed when a cold air mass was over
the cooling lake. The formation was not often associated with frontal
1097
-------
activity. Typically, it occurred with cold, stable air when the water
temperature was much warmer than the air temperature. Comparisons with
steam fog observations over the Dresden cooling lake in the colder winter
climate of northern Illinois showed similar steam fog distributions [6].
ENHANCEMENT OF CLOUDINESS
Satellite photographs taken during the summer of 1975 were used to
investigate cooling lake effects on the time-space distribution of
cloudiness. Analyses were made for two cooling lakes in southern Illinois
(Baldwin and Coffeen) and for a much larger, control lake (Carlyle). No
evidence was found that the cooling lakes or control lake had any signifi-
cant effect upon the summer cloud frequencies, and, consequently, upon
precipitation. There was some evidence, however, that local terrain
features in the study region, ridges and river valleys, do influence the
spatial distribution of clouds, primarily cumulus and cumulonimbus.
The potential cooling lake effect was investigated further through use
of available radar data for summer during 1971 to 1975. Results supported
the satellite findings with respect to the two cooling lakes. However,
the larger control lake (Carlyle) appeared to have some influence on the
initiation of convective precipitation when atmospheric motions were
parallel to the major axis of this elongated lake.
Thus, .it was concluded from the satellite and radar evidence that cooling
lakes the size of Baldwin and Coffeen have little or no effect upon the
initiation of convective cloudiness or precipitation. However, much
larger cooling lakes, as indicated by the Carlyle findings, could enhance
convective activity when the low-level air and clouds have a relatively
long travel time over the lake. Otherwise, most cooling lakes have a
minimal impact upon the initiation and enhancement of convective cloudiness
and should produce no environmental problems of significance in this
direction.
RAINFALL
A dense raingage network was operated in the Baldwin area during
July-November 1976 and March-November 1977. The objective was to
investigate potential effects upon the regional rainfall pattern resulting
from waste heat discharges into the cooling lake associated with the
Baldwin power plant. Analyses were performed to determine the seasonal
distributions of total rainfall, frequency of rainfall events, effect of
storm movements on network rainfall patterns, and the relation between
rainfall and synoptic weather types.
Results of the 2-year study were inconclusive. There was a persistent
high in the Baldwin network located 10-15 km E-ENE of the center of the
lake when rainfall for the two years was combined, and this apparent
1098
-------
anomaly was especially prominent with storms moving from the SW quadrant,
which place the lake directly upwind of the observed maximum. However,
the rainfall maxima within the network were in agreement with the natural
rainfall distribution for southern Illinois during the sampling periods,
as revealed by the National Weather Service climatic network data.
The most positive evidence of a localized anomaly was its persistence in
location during the sampling period. If this is a localized anomaly, it
could also be related to topographic features to the west (upwind) of the
network where ridges and bluffs apparently stimulate the development of
cumulus and cumulonimbus, as indicated in the previous discussion on
enhancement of convective clouds. Since no evidence was found of con-
vective cloud stimulation downwind of the lake, it appears unlikely that
.the relatively high rainfall in the eastern part of the Baldwin Network in
1976-1977 can be attributed to a cooling lake effect on the environment.
On the basis of presently available information, it is concluded that
cooling lakes of the size of Baldwin (2200 acres) will not significantly
modify the precipitation regime in the surrounding area.
CONCLUSIONS
Meteorological effects from cooling lakes associated with single power
plants are usually insignificant in Illinois and other areas of similar
climate and topography. There was no evidence in the Baldwin study of
significant lake effects upon clouds and precipitation. Most cases of
fog initiation or enhancement occurred in the cold season, and the down-
wind extent of lake-influenced fog was usually less than 0.8 km (0.5 mi).
Dense fog (visibility 50.4 km) occurred in less than 25% of the fog
events.
REFERENCES
1. Ackermann, William C. Research Needs on Waste Heat Transfer from
Large Sources in the Environment. Urbana, 111.: Report to National
Science Foundation, Grant GI-30971, Illinois State Water Survey, 1971.
2. Huff, F. A., and J. L. Vogel. Atmospheric Effects from Waste Heat
Transfer Associated with Cooling Lakes. Urbana, 111.: Report to
National Science Foundation, Grant GI-35841, 1973.
3. Vogel, J. L. , and F. A. Huff. "Fog Effects from Power Plant Cooling
Lakes." J. Appl. Meteor., Vol. 14, 1975, 868-872.
4. Murray and Trettel, Inc. Report on Meteorological Aspects of Operating
the Man-Made Cooling Lake and Sprays at Dresden Nuclear Power Station,
Chicago, 111.: Prepared for Commonweath Edison Company, 1973.
1099
-------
5. Heiss, W. H. Highway Fog Visibility Measures and Guidance Systems.
Washington, D.C.: Transportation Research Board, National Research
Council, 1976. National Cooperative Highway Research Program Report
171.
6. Vogel, J. L., and F. A. Huff. "Steam Fog Occurrences over Cooling
Lakes." Boston, Mass.: Preprints Sixth Conference on Planned and
Inadvertent Weather Modification. American Meteorological Society,
1977, 69-72.
1100
-------
Table 1
FREQUENCY OF BALDWIN COOLING LAKE STEAM FOGS AND VISIBILITIES
Season
Fall 76
Winter 76-77
Spring 77
Summer 77
Fall 77
Winter 77-78
March 78
Total
, initiations
'Visibilities (km)
Enhancements
50.4
2
10
0
0
2
9
1
>0.4-
51.6
5
5
1
0
1
2
0
>1.6-
58.0
7
4
6
4
4
8
1
>8
8
10
9
1
16
20
1
Total
22
29
16
5
23
39
3
24
14
34
65
137
<0.4
3
1
0
4
4
7
_!
21
Visibilities
>0.4-
51.6
3
0
1
6
5
2
2
>1.6-
58.0
1
0
0
4
0
1
2
(km)
>8
0
0
0
0
0
0
0
19
Total
7
1
1
14
9
10
_6
48
-------
o
K)
Table 2
FREQUENCY OF BALDWIN COOLING LAKE STEAM'FOGS WITH HORIZONTAL EXTENT
Total
Initiation Days
Horizontal Extent (km)
Enhancement Days
Horizontal Extent (km)
Visibility
50.4
>0.4-5l.6
>1.6 58.0
>8.0
50.2
11
4
8
12
>0.2-
50.8
6
2
0
0
>0.8-
51.6
2
0
0
0
>1.6
3
0
0
0
Total
22
6
8
12
50.2
10
8
1
0
>0.2-
fO.8
3
1
0
0
>0.8-
fl.6
0
0
0
0
>1.6
0
0
0
0
Total
13
9
1
0
35
48
19
23
-------
2
o
W
V-V/ELEVATION, 430 ft:'
E X P L A
A TOWER
• SHELTER
0 SHELTER
NA
AND
WATER AND AIR
TEMPERATURE
D RAINGAGE, EVAPORIMETER
TRANSMISSOMETER
6
•
Figure 1. Baldwin instrument network
H03-
-------
COMPUTER SIMULATION OF MESO-SCALE METEOROLOGICAL EFFECTS OF
ALTERNATIVE WASTE-HEAT DISPOSAL METHODS
J.P. Pandolfo and C.A. Jacobs
The Center for the Environment and Man, Inc.
Hartford, Connecticut U.S.A.
ABSTRACT
The preliminary use of a physically complete land-sea-air boundary layer mo-
del is described in the simulation of the meteorological effects of artifi-
cial heat inputs. The model provides solutions obtained by temporal inte-
gration of the Eulerian conservation equations. Taken into account are
stability-(Richardson number)-dependent mixing, complex topography, spatially
varying interface properties, and cloud-dependent radiative heating (cooling).
Clouds may be externally specified, or internally generated in the model by
exercising a model input option. In the example described, simulation of the
effects of hypothetically arranged dry cooling towers in the Rhine valley of
Switzerland was carried out (in cooperation with the Institute of Reactor
Research of Switzerland). Simulated effects on the mesoscale temperature
structure of the atmosphere's lowest kilometer, as well as on the slope-
valley circulation as resolved on a 3-km horizontal grid, are presented for
a clear summer day.
INTRODUCTION
A physically complete land-sea-air boundary layer model has been used in the
simulation of the meteorological effects of artificial heat inputs. The mo-
del provides solutions obtained by temporal integration of the Eulerian con-
servation equations on a relatively fine (1- to 10-km horizontal, 1- to 100-m
vertical) spatial mesh, with complex momentum, heat, and moisture sources.
Taken into account are stability-(Richardson number)-dependent mixing, com-
plex topography, spatially varying interface properties, and cloud-dependent
radiative heating (cooling). Clouds may be externally specified, or intern-
ally generated in the model by exercising a model input option. The model's
physical equations, and a previous application in studying inadvertent wea-
ther modification, are described by Atwater fl].
In the study described in this paper, the overall objective was defined by
the Institute for Reactor Research (EIR) of Switzerland. It is
"to identify and quantify the impact of man's industrialization on
the climate of the Rhine River Valley in the region about Basel.
In particular, the climatic effects resulting from alternative
scenarios involving the size and locations of new electric power
generating facilities will be explored."
1104
-------
The reason for this objective becomes apparent when it is pointed out that
with present and projected power generating facilities, the man-made heat in-
put to the atmosphere over the region will amount to 50 percent or more of
the solar heat input to the region in the winter-time.
The initial phases of this study involved extensive data gathering efforts in
the region, including the construction of a pilot-test cooling tower, and the
derivation of a general numerical model of the cooling tower plume [2]. This
work was begun, and continues, at EIR (Switzerland).
Later in the project, the preparation and use of a valley-scale meteorologi-
cal model to be used in conjunction with the research products of these ac-
tivities was begun in a joint CEM-EIR project. A series of numerical experi-
ments was defined to assess the feasibility of using such a general meteoro-
logical model in this study. These began with one-dimensional (horizontal
variation terms of the Eulerian equations prespecified from observations)
simulations of the diumally varying vertical structure of the atmospheric
boundary layer [3]. They continued with two-dimensional (along-valley vari-
ations specified from observations) simulations of the diurnally varying atmo-
spheric structure in a cross-valley section. We have now completed the first
of our three-dimensional simulation experiments, which is described here.
In this experiment, we wished to determine whether practically obtainable in-
itial three-dimensional data sets (derived from scattered observations inte-
grated into one- and two-dimensional model simulations) could be used in a
usefully detailed three-dimensional spatial grid, and integrated over useful
periods of time (a few days per simulation), without generating computational
errors so large as to make the three-dimensional simulation results uselessly
unrealistic. Furthermore, this was to be carried out within practical limi-
tations on computer size and availability.
THE FIRST THREE-DIMENSIONAL SIMULATION EXPERIMENT
The scenario used for this feasibility experiment introduced 2000 MW of waste
sensible heat at each of three ("dry tower") locations spaced approximately
at equal intervals along the main valley floor (grid-square centers marked
with the symbol m on Fig- la). Figure la also shows the smoothed topography
on the basic 3kmx3km horizontal grid. The main Rhine valley is generally
oriented E-W, but turns sharply to the north at the west end of the region
(in grid columns 1-4). A side valley branches generally north at grid column
12. Three other side valleys branch generally south at grid columns 4, 8,
and 15. Features apparently associated with these secondary valleys are evi-
dent in the solution temperature and flow patterns shown in later figures.
The boundary ridge elevations are highly asymmetric. The most pronounced
ridge lies on the eastern half of the northern boundary, with a much lower,
interrupted, ridge along the southern and southwestern boundary. The most
intense slope is oriented N-S, east of grid column 12, and north of the main
valley. Features of the solution temperature and flow fields apparently re-
lated to this topographic feature are also evident in later solutions.
1105
-------
The large-scale weather situation specified for this experiment was one of
clear, early-summer (near-solstice) conditions with weak synoptic-scale flow.
The integration was carried out on a three-dimensional spatial mesh contain-
ing 7(north-south) x 20(east-west) x 27(vertical) points. It was found ex-
perimentally that 36-second time steps were required to ensure numerical sta-
bility. As a consequence, it was found that almost exactly one hour of CRAY-1
CPU time (or about two hours of CDC-7600 time) was required for each simula-
tion day of real time.*
SOLUTION FIELDS FOR THE FIRST EXPERIMENT
Solution features are shown for two times of day—viz., 1707 sun time (ST)
June 23rd, and 0507 ST June 24th. There are presented for each of the two
times of day the basic ("CONTROL") unperturbed temperature and horizontal
flow fields at 8-m and 300-m elevation (above terrain). There are also shown
difference fields in which the temperature and horizontal flow vector differ-
ences ("DRY TOWER MINUS CONTROL") are plotted at 8-m (above terrain) elevation.
Only a cursory discussion of these results is justified at this time. We
point out that generally reasonable "CONTROL RUN" results have been obtained.
Results at 1900 and 3100 time steps of integration are shown. These results
are reasonable in that they show:
11 a general, relatively deep, upslope and upvalley daytime flow
(Fig. 2a,b);
2) a generally shallow downslope, downvalley flow at night (Fig. 4a,b);
3) relatively weak, deep, daytime temperature maxima over the valley in
terrain-parallel surfaces (not shown) -, and
4) relatively intense, shallow night-time temperature minima over the
valley in terrain-parallel surfaces (Fig. 3a,b).
In addition, the general correspondence between the scattered wind measure-
ments (Fig. Ib) and the solution wind fields (Fig. 4a) is to be noted.
Details in these general fields require more investigation. For example, the
ridge-parallel jet-like detail at 300 m found to be:
1) N-S in the daytime flow pattern along grid columns 2, 3, 4, and W-E
at grid columns 17-20, rows 3-5 (Fig. 2b);
2) N-S in the night flow pattern along grid columns 4-9 (Fig. 4b);
is similar in general intensity, orientation, and vertical structure to that
found in the much more highly idealized models of Mason and Sykes [4]. A se-
quence of stepped idealizations from our model to their model will serve to
investigate the underlying physics of the apparent similarity. Though this
sequence of experiments is easy to formulate, it will require a few hours of
computer time to carry out.
The temperature perturbations by the intense waste-heat sources are qualita-
tively reasonable. Relatively weak (against a strong solar radiation back-
ground) daytime maximum differences are evident (Fig. 5a). More intense
night-time maxima are evident in Fig. 6a. The wind disturbance is less
Acknowledgement is made to the NCAR, which is sponsored by the NSF,
the computing time used in this research.
1106
-------
systematic. The daytime flow at 8 m is generally countered, and locally re-
versed, with wide-spread vector differences as large as and opposing the con-
trol run winds (Figs. 2a, 5b). The night-time flow is perturbed strongly
only locally in the S-W portion of the region (Figs. 4a, 6b).
SUMMARY AND PLANS FOR FURTHER INVESTIGATION
The first experimental results exhibit significant disturbance of the control
run daytime meso-scale slope-valley circulation. They also show wide-spread
temperature increases of about 1°C in clear, summer, daytime in-valley tem-
perature and 3°C increases in clear, summer, night-time temperatures.
The large-scale weather situation dealt with in this experiment constitutes a
"near-worst" case in terms of the computational requirements, a "far-from-
worst" case in terms of the relative magnitude of waste-heat to natural solar
heat input (about 4% over the total area, and for the day of the year, and
the clear conditions considered) and perhaps a "semi-worst" case in terms of
natural flow disruption because of the weak large-scale flow component and
the strong solar input.
These characterizations remain to be more precisely defined by carrying out
other experiments in other weather (particularly cloudy-winter) conditions.
There also remains to be more precisely assessed the level of "computational
noise" still present in the solutions. A two-pronged approach to this assess-
ment is planned: one branch obtaining more detailed observations scheduled
and placed in accordance with previously obtained model solutions; the other,
more theoretical, branch studying stepwise physical idealizations, and ap-
plied mathematical questions (e.g., the influence of the finite-difference
schemes and the lateral boundary conditions chosen).
Finally, and to some extent, concurrently, alternative waste-heat disposal
methods must be simulated, including variation of type (wet cooling towers)
and location (e.g., ridge or on-slope rather than in-valley location).
REFERENCES
[1] M.A. Atwater, "Urbanization and Pollutant Effects on the Thermal Struc-
ture in Four Climatic Regimes," J. Appl. Meteoy., 16, 1977 (Sep).
[2] F. Gassman, D. Burki, D. Haschke, R. Morel, Flugmessungen in de? atmo-
spharisahen Grenzsohidht, EIR-Bericht Nr. 334, Eidg. Inst. fiir
Reaktorforschung Wvirenlingen, Schweiz (Switzerland), 1978.
[3] D. Haschke, F. Gassman, F. Rudin, Eindinrensionale3 zeitabhangige Simu-
lation dez> planetarischen Grenzschicht weber eine 48-Stundsn Periods,
EIR-Bericht Nr. 337, Eidg. Institut fur Reaktorforschung Wurenlingen,
Schweiz (Switzerland), 1978:
[4] P.J. Mason and R.I. Sykes, "On the interaction of topography and Ekman
boundary layer pumping in a stratified atmosphere," Quart. J* Roy.
Meteor. Soc., 104, 475-490, 1978.
1107
-------
FIGURE la. Basic topography with dry cooling tower locations. Elevation
is in meters.
468.
i ' i
—r
T 1 1 r
FL
KE't
I
OB
I
I
I
1
PR
I
I
I
I
J
I
I
I
I
FIGURE Ib. Observed hourly mean wind at 10 meters, 0500-0600 sun time,
23 June 1976. The magnitude of the maximum wind vector (cm/s)
in the field is shown in the upper right-hand corner.
1108
-------
640,
FIGURE 2a. Control run horizontal wind at 8 meter elevation at 1707 sun
time. The magnitude of the maximum wind vector (cm/s) in the
field is shown in the upper right-hand corner.
917,
\
FIGURE 2b. Control run horizontal wind at 300 meter elevation at 1707 sun
time. The magnitude of the maximum wind vector (cm/s) in the
field is shown in the upper right-hand corner.
1109
-------
FIGURE 3a. Control run temperature at 8 meter elevation at 0507 sun time.
i i i i\ \ i
FIGURE 3b. Control run temperature at 300 meter elevation at 0507 sun time.
1110
-------
468.
\
FIGURE 4a. Control run horizontal wind at 8 meter elevation at 0507 sun
time. The magnitude of the maximum wind vector (cm/s) in the
field is shown in the upper right-hand corner.
933,
FIGURE 4b. Control run horizontal wind at 300,m elevation at 0507 sun
time. The magnitude of the maximum wind vector (cm/s) in the
field is shown in the upper right-hand corner.
1111
-------
i—i—i—i—i v i } \\—i i
FIGURE 5a. Temperature difference dry cooling tower minus control run at
8 meters at 1707 sun time.
655
FIGURE 5b. Horizontal wind vector difference dry cooling tower minus con-
trol run at 8 meters at 1707 sun time. The magnitude of the
maximum wind vector (cm/s) in the field is shown in the upper
right-hand corner.
1112
-------
FIGUBE 6a. Temperature difference dry cooling tower minus control run at
8 meters at 0507 sun time.
623
FIGURE 6b. Horizontal wind vector difference dry cooling tower minus con-
trol run at 0507 sun time. The magnitude of the maximum wind
vector (cm/s) in the field is shown in the upper right-hand
corner.
1113
-------
A NUMERICAL SIMULATION OF WASTE HEAT EFFECTS ON SEVERE STORMS
H. D. Orville and P. A. Eckhoff
Institute of Atmospheric Sciences
South Dakota School of Mines and Technology
Rapid City, South Dakota U.S.A.
ABSTRACT
A two-dimensional, time-dependent model has been developed which gives
realistic simulations of many severe storm processes — such as heavy
rains, hail, and strong winds. The model is a set of partial differential
equations describing time changes of momentum, energy, and mass (air and
various water substances such as water vapor, cloud liquid, cloud ice,
rainwater, and hail). In addition, appropriate boundary and initial con-
ditions (taken from weather sounding data) are imposed on a domain
approximately 20 km high by 20 km wide with 200 m grid intervals to
complete the model.
Cases have been run which depict realistic severe storm situations. One
atmospheric sounding has a strong middle-level inversion which tend to
inhibit the first convective clouds but give rise later to a severe storm
with hail and heavy rains. One other sounding is taken from a day in
which a severe storm occurred in the Miami area.
The results indicate that a power park emitting 80% latent heat and 20%
sensible heat has little effect on the simulated storm. A case with 100%
sensible heat emission leads to a much different solution, with the
simulated storm reduced in severity and the rain and hail redistributed.
INTRODUCTION
A two-dimensional, time-dependent cloud model has been modified to
simulate the addition of heat and vapor from a hypothetical power park.
The cloud model has been under development for many years and successfully
applied to several convective situations. The most recent application
was a simulation of a hailstorm reported by Orville and Kopp [1].
For this study, the model was run using two types of severe storm
atmospheric soundings. The first type can be classified as Type I using
the classification system established by Fawbush and Miller [2], This
type of sounding generally produces a family of tornadoes. The atmos-
pheric sounding from the well documented Fleming Storm [3] was used as
a Type I- sounding. This was a dangerous hailstorm which eventually
produced a tornado in its twelve plus hours of existence.
1114
HDO
-------
The second sounding used can be classified as a Type 2 atmospheric
sounding [2], The sounding used was taken three hours prior to a
tornado touching down in downtown Miami, Florida [4]. This storm is
typical of a Type 2 which produces a single tornadic event.
For each sounding, the total effluent from the cooling towers in the
power park was calculated and inserted into the model in a cross sectional
area of the park's heating and moistening volume (see Fig. 1).
The model was run until all the precipitation had fallen or until the
simulation had progressed where valid comparisons could be made. Then
the model was run again using the same initial sounding except that the
effluent (vapor and heat) from the power park was excluded. Several
other effluent variations were also simulated. For the Fleming stor.m
cases, three other runs were made. One involved doubling the power park
concentration of effluent which, in effect, halved the area of the power
park. Another involved using an effluent that was made up of 100%
sensible heat which is designated to simulate a park made up of dry
natural draft cooling towers [5], The last case in this series involved
placing the power park on the other side of the ridge. This was done to
see the effects location had on storm development.
The Miami storm cases were done in a similar manner with fewer park
variations. In the end, there were seven cases that could be analyzed.
RESULTS
Flem_lng_ jatorm
The cross sections for 66 min. and 102 min. show the general development
of the storm in the 5 Fleming storm cases.
The first four cases (Figs. 2a-d) show the main cloud being fed by air
from both the right and left. The strength of the main updraft in Figs.
2a-e draws in air from the lower left-hand corner into the main cloud.
In the first four cases of Fig. 2, the closed circulation pattern just
to the right of the main updraft is a main feature. Each pattern is
shaped differently, and the contours indicate that the flow of air in the
main updraft is weakest in the natural case (Fig. 2a), followed by the
100% sensible heat case (Fig. 2e). The standard park and double flux
cases (Figs. 2b and d) are strong but of about equal strength. The main
updraft is the strongest in the left park case (Fig. 2c). Also notice
the shape of the zero contour below the main cloud in Figs. 2a-e, and
that the zero contour is in a different position in each case. The 100%
sensible heat case has formed a strong secondary circulation over the
right side of the grid, causing a second cloud to form.
HDO 1115
-------
The sequences at 102 min. (Fig. 3) show significant differences in most
of the cases. The standard park case is most like the natural case. The
storm in the left park case has moved further to the right in the domain
and is weakening. The double flux case shows slightly less rain and hail,
with most of the precipitation distributed below 5 km. The 100% sensible
heat case exhibits the greatest differences. The major convection has
ceased and precipitation has nearly all fallen to the ground.
The dynamics of each storm is slightly, to almost completely, different
from that of the natural case. This difference in dynamics is evident in
the accumulated rain and hailfall and the time at which the storms end.
Figure 4 compares the natural case rainfall with that of the standard
park, the left park cases, the 100% sensible heat, and the double flux
cases.
Notice the similarity in rainwater distribution between the natural case
and the standard park and double flux cases. However, the latter two
cases exhibit a small distribution shift to the right. The left park
case does not show the two-peak distribution of these three cases. The
100% sensible heat shows greatly reduced rainfall. The total accumulated
rain on the ground for the Fleming storm cases shows the natural case
with 178.2 kT km"1. This is followed closely by the standard park and
double park cases with 174.6 and 174.9 kT km"*, a decrease of about 2%
for both cases when compared to the natural case. The left park case
shows a rainfall accumulation of 151.6 kT kin , a decrease of rainfall
when compared to the natural case. The smallest rainfall amount was
produced in the 100% sensible heat case. This case produced 65.13 kT
km", which is a 64% drop from the natural case.
Each case shows a maxima of hail at 10 km on the horizontal axis (the
ridge line); however, the park and 100% sensible heat cases show a second
maxima to the right. The total accumulated hail for the natural case is
47.4 kT km"*. The standard park shows the next highest accumulation with
41.3 kT km"1, or a 13% drop in hail. Next highest is the left park case
with 31.5 kT km , followed very closely by the double flux case with
31.3 kT km. Both cases show a drop of about 34% when compared to the
natural case. The case with the smallest hail accumulation is the 100%
sensible heat case with 3.8 kT km"1, or a decrease of 92% when compared
to the natural case.
Miami Storm
The Miami storm results are shown in Figs. 4b-5a-b. The natural case at
141 min. (Fig. 5a) shox^s a vigorous, active convective storm, with con-
vergent inflow (flow from both left and right in the lower levels). The
power park case storm is nearly as big (Fig. 5c), but not as broad as the
natural case. In addition, the power park case is being fed by low-level
flow primarily from the right side. Figure 5b shows the natural case
storm still active, with copious amounts of rain and precipitating ice.
However, 5d shows that the power park case storm has nearly dissipated,
1116
-------
mostly anvil cloud remaining. Figure 4b shows the accumulated rainfall;
much more has fallen in the natural case. There were reports of over
6 inches of rain in some south Florida areas on this day.
DISCUSSIONS AND CONCLUSIONS
The seven runs have shown some of the influence of power parks on severe
storm development. Storm development was different and was affected to
varying degrees by the effluents of the power park. The power parks
create their own dynamics which interact with the flow of the developing
storm to produce storms of less, to greatly less, precipitation output.
One of the really signficant changes comes about after 66 min. of
real-time simulation in the Fleming storm case. This is a time when the
heat and/or moisture from the various parks had enough time to develop
and interact with the natural dynamics to produce readily noticeable
changes. The addition of heat and moisture from the wet cooling towers
have supplied enough moisture to sustaiTn the growth of the original cloud.
In the dry cooling tower or 100% sensible heat case, there was enough
heat affecting the dynamics to create a more vigorous cloud growth to the
right of the original cloud development. The vigorous cloud developed a
downdraft that interacted with the downdraft from the cloud system to the
left. The result was a cessation of low-level moisture into both cloud
systems and the premature death of both systems.
The cloud in the natural case was very weak at 66 min., and the new
development to the rear saved the original cloud from dying slowly. The
new development took over with good growth characteristics and rejuvenated
the natural case. However, the wet cooling tower cases grew faster, and
by 102 min. their gust fronts were more developed. This can be attributed
to the effects of the power parks.
One of the more noticeable changes is the quantity and distribution of the
rain and hail. All the power park cases produced less rain and hail, with
the 100% sensible heat case showing around a 75% decrease in both rain and
hail maximums. The wet cooling tower cases show a small decrease in
precipitation with a shift in the location of rainfall. The standard
power park case shows less differences than any of the other cases in its
rain and hail distribution for the Fleming storm series of cases. The
double park case showed slightly more of a change with a little less rain
and hail than in the standard park case. However, the distributions of
rain and hail were very similar to the standard case. The left park
case showed a total rainfall slightly less than the natural case in the
Fleming storm series, but the distribution shows a large single peak
instead of the double peaks as in the other wet tower cases. The 100%
sensible heat has rain and hail peak amounts that are 30% and 13% of the
natural case. This can be directly attributed to the rapid cloud develop-
ment in front of the storm, which saps the energy of the storm leading
to early dissipation of the storm system.
1117
-------
One point brought out in the left park case is the earlier cloud
formation if the power park is under the area where initial cloud develop-.
raent would normally take place.
In the two Miami runs, the power park effluents interact significantly
with a cloud developing overhead. The clouds develop more rapidly in the
power park case, but never become as organized a system as in the natural
case. The flow that develops is not "complementary" to the flow in the
natural case. This results in a 50% decrease in the rain maximum and
a 66% decrease in accumulated rain at 174 min. in the power park case.
Hail develops in the storms, but hail accumulation on the ground is
insignificant.
Results of this study indicate that the incorporation of the added heat
and moisture in a developing storm is a very nonlinear process and does
not necessarily yield a more severe storm.
Other studies, one by Orville et^ al^ [6], showed that slight increases
in rain could occur if all of the added moisture were stored in the region
and released to the storm at one time, such as might occur in a very
stagnant flow condition.
The ultimate effects of power park effluents on severe storms are not
readily determined by simple additive calculations. Complex interactions
occur which can only be tested through realistic numerical simulations.
Careful observations of the long term climatological changes near large
power plants should be maintained for long periods of time to determine
the actual effects of the plants on the weather.
ACKNOWLEDGMENTS
This research was sponsored by the U.S. Nuclear Regulatory Commission
under Contract No. NRC-04-76-350. Acknowledgment is given to the National
Center for Atmospheric Research, which is sponsored by the National
Science Foundation, for granting us use of their computing facilities.
REFERENCES
1. Orville, H. D., and F. J. Kopp, 1977: Numerical simulation of the
life history of a hailstorm. J»_ Atmos. Sci., 34, 1596-1618.
2. Fawbush, E. J., and R. C. Miller, 1954: The types of air masses in
which North American tornadoes form. Bull. Amer^. Meteor. Soc.,
35_, 154-165.
1118
-------
3. Browning, K. A., and G. B. Foote, 1975: Airflow and hail growth in
supercell storms and some implications for hail suppression.
National Hail Research Experiment Technical Report No. 75/1,
May 1975, 75 pp.
A. Hiser, H. W., 1967: Radar and synoptic analysis of the Miami tornado
of 17 June 1959. Preprints jth Conf. Sey^er^e Local Scorns,
St. Louis, Missouri, Amer. Meteor. Soc., 260-269.
5. Lee, J. L. , 1978: Potential weather modification caused by waste
heat release from large dry cooling towers. Proposed for
presentation at the 2nd AIAA/ASME Thermpphysics and Heat
Transfer Conf^, May 24-26, 1978, Palo Alto, California.
6. Orville, H. D., F. J. Kopp, and P. A. Eckhoff, 1977: The application
of a numerical model to determine the effects of waste heat on
severe weather. Preprints 10th Conf. Severe_LocalStorms,
Omaha, Nebraska, Amer. Meteor, Soc., 271-276.
1119
-------
TOP VIEW
Mountain _^^
Ridge ^N
X-Z PLANE N
\
-
IW
MW
° « °
OOOOO*OG
00oo°o0o0
0 0 0°0°
oo o0ooo
o oo o
o o °° o o
22.3km
5.2 km
19.2
km
SIDE VIEW
Mountain Ridge
Heating + Moistening
Volume "\.
2OOm
210m
19.2 km
Fig. 1: The top view shows the standard park configuration to the
right of the ridge. The side view shows the volume into which the
moisture and heat is added.
1120
-------
(a)
STREAM FUNCTION tuf iff W)
NATURAL CASE
-37OO , STANDARD PARK CASE T-66 MNS
(b)
10
01 STANCE-(km)
Fig. 2: The stream function field
of the Fleming storm cases at 66
minutes. The clouds are the
shaded areas.
1121
Fig. 3: Same as Fig, 2 but for
a contour interval of 5000 kg m"1
sec"*, Rain and hail over 1 gm
kg~* are depicted as dots and
asterisks, respectively.
-------
(a)
3.5
3-0
£2.5
I
<2.0
o
LU
13
-------
(a)
STREAM FUNCTION OflrfW) NATURAL CASE T • 141 MNS
(b)
STREAM FUNCTION (kfl m W) NATURAL CASE T •
SSSSS^SSSSJSSSSSSSSSSSSSSSSS
f>.'5555l55S}§S5§3555-555S5Ss5S555l|S"JI
§Si§l
i "
(c)
STREAM RJNCTON OtgirfW) PCWER PARK CASE T* I4> MWS
(d)
STREAM FUNCTION (kgrtlW) POWER PARK CASE T-
llllllHlnimilttl"|iMH>lll|'Ml|
Fig, 5: The stream function for the natural (a & b) and power park
(c ?* d) cases of the Miami storm at 141 and 168 minutes^ Contouring
is 10000 kg m"1 sec"1 for a, b, & d and 5000 kg nT1 sec"1 for c.
1123
-------
ON THE PREDICTION OF LOCAL EFFECTS OF PROPOSED COOLING PONDS
B. B. Hicks
Radiological and Environmental Research Division
Argonne National Laboratory, Argonne, Illinois U.S.A.
ABSTRACT
A Fog Excess Water (FEW) Index has been shown to provide a good
measure of the likelihood for steam fog to occur at specific
cooling pond installations. The FEW Index is derived from the
assumption that the surface boundary layer over a cooling pond
will be strongly convective, and that highly efficient vertical
transport mechanisms will result in a thorough mixing of air
saturated at surface temperature with ambient air aloft.
Available data support this assumption. An extension of this
approach can be used to derive a simple indicator for use in
predicting the formation of rime ice in the immediate downwind
environs of a cooling pond. In this case, it is supposed that
rime ice will be deposited whenever steam fog and sub-freezing
surface temperatures are predicted. This provides a convenient
method for interpreting pre-existing meteorological information
in order to assess possible icing effects while in the early
design stages of the planning process. However, it remains
necessary to derive accurate predictions of the cooling pond water
surface temperature. Once a suitable and proven procedure for
this purpose has been demonstrated, it is then a simple matter
to employ the FEW Index in evaluations of the relative merits
of alternative cooling pond designs, with the purpose of minimizing
overall environmental impact.
INTRODUCTION
Industrial cooling ponds often give rise to localized environ-
mental effects, particularly in winter when steam fog and rime
ice can become problems downwind of the hottest areas. Fog
generation above artificially-heated water surfaces has been the
subject of a number of studies1'2'3, but similar studies of
rime ice have not been found. A preliminary study of the matter
demonstrated the practical difficulties likely to confront experi-
mental investigations of riming1* 5 ^. This study, performed at the
Commonwealth Edison Dresden plant (near Morris, Illinois) during
the winter of 1976/7, provides a four-month record of the occur-
rence and intensity of fog and rime associated with the operation
of a fairly typical industrial cooling lake.
Earlier studies at Dresden succeeded in obtaining direct measure-
ments of turbulent fluxes of sensible and latent heat from the
heated waterG. The resulting improved formulations of these
convective and evaporative heat: losses can be used in much
±124
-------
the same way as the familiar wind speed functions that are used
in most contemporary cooling pond design studies. In this regard
the earlier Dresden experiments, which were conducted over the
three-year period 1973-1976, addressed the question of how to
predict the water temperature characteristics of cooling pond
installations. Subsequent studies have refined these techniques
by parameterizing the subsurface thermal boundary layer?, which
effectively limits heat exchange between deep water and the air.
The premise of the present study is, therefore, that we can predict
the temperature characteristics of a proposed cooling pond, but
need to assess the potential environmental impact.
STEAM FOG
2
Hicks introduced a Fog Excess Water Index, e , based on the
supposition that air saturated at surface temperature rises and
mixes with equal quantities of ambient, background air. The excess
vapor pressure e of the mixture can be written as
j\. o
e = (e (T ) + e )/2 - e ((T + T )/2)
xs s s a s s a
where e (T) is the saturated vapor pressure at temperature T,
T is tne ambient air temperature and e is the air vapor pressure.
Wnen tested against the data of Currier et al.l, the FEW Index
was found to provide a good indication of the occurrence of steam
fog, as well as some measure of its intensity. The FEW Index
was further verified by use of observations of fog generated by
cooling-pond simulators at Argonne National Laboratory and by
data from Dresden.
Figure 1 is a further test of the FEW Index, again largely based
on observations made at 'Dresden but supplemented by a series of
measurements made at the Cal-Sag shipping canal, a major inland
waterway which passes conveniently near Argonne. Canal, water
temperatures in winter are typically more than 20°C higher than
in nearby lakes and streams, due to heavy industrial usage. The
data illustrated in the diagram give further support for the
validity of the FEW Index method.
RIME ICE DEPOSITION
A few obvious (and perhaps trivial) considerations should be set
down at the outset. Firstly, it is clear that rime ice deposition
is a cold-weather phenomenon which is constrained, by definition,
* to occasions when the surface temperature is below freezing.
This constraint does not apply to the generation of steam fog,
and hence rime deposition might well be considered as a sub-set
of steam fog cases. Secondly, it follows that riming will be
mainly a wintertime phenomenon, most often at night. ^ In the
nocturnal case, it seems likely that accurate prediction of riming
will prove extremely demanding, since nocturnal surface temperatures
1125
-------
are highly variable both in space and in time and thus great care
must be taken in selecting an appropriate data base.
Figure 2(a) illustrates the first point; the Dresden 1976/7 winter
data do indeed show riming to be a subset of the fog occurrences.
Observations were made on a total of 84 mornings. On no occasion
was the observation of overnight rime deposition not accompanied
by steam fog from the pond. Furthermore, the amount of rime
deposited is well correlated with a measure of steam fog intensity.
To show this, rime deposits have been quantified according to
the visual observations; none = 0, slight rime = 1, moderate
rime = 2, heavy rime = 3. The fog intensity is conveniently
quantified by the reported depth of the fog layer over the hottest
part of the cooling pond, estimated from a comparison with the
known heights of surrounding obstacles. Figure 3 demonstrates
the correlation. Thus, it appears reasonable to expect the FEW
Index to be an appropriate measure of the intensity of rime
deposit, since it has already been shown to be an indicator of
steam fog intensity. The present limited set of data do not allow
direct investigation of the interrelation between rime intensity
and evo, since reliable nocturnal evaluations of evc; at the
A. o n ^s. o
Dresden site are not availableo
.Figure 2(b) shows the frequency of occurrence of fog and rime
that would have been expected on the basis of the arguments
presented above. It is assumed that steam fog will occur when
e > 0, based on the observed Dresden water temperatures and
overnight air temperatures and humidities measured some 40 km
away at Argonne National Laboratory. Rime is then predicted on
each of those occasions for which sub-freezing overnight tempera-
tures were reported. Comparison between Figures 2(a) and 2(b)
shows fairly good agreement: the rime curves are drawn to be
identical.
DISCUSSION AND CONCLUSIONS
Although it is clear that the depth of steam fog and the amount
of rime deposited are well correlated, there is no strong
dependence of riming upon meteorological quantities,such as wind
speed, nocturnal net radiation, etc. To a considerable extent,
this is as must be expected as a consequence of the lack of
correlation between e and wind speed (see Figure 1). The
1976/7 results are no£Ssuitable for investigating this matter
with confidence. Nor is it clear that the physics involved will
permit a clear-cut conclusion to be obtained. Nevertheless, it
is intended to proceed with investigations of the thermal and
moisture plumes arising from heated water surfaces, in part to
derive better methods for predicting the frequency of events in
the design stage but also to investigate the role of steam fog
as an interference with the natural infrared radiation regime of
a water surface.
1126
-------
ACKNOWLEDGEMENTS
The work performed at Dresden was made possible by the complete
cooperation of the Commonwealth Edison Company. The Dresden
data reported here were obtained during a field program directed
by Dr. J. D. Shannon. Dr. P. Frenzen obtained the canal data.
This study was supported by the U. S. Department of Energy, as
part of an investigation of the Meteorological Effects of Thermal
Energy Release.
REFERENCES
1. Currier, E. L., J. B. Knox, and T. V. Crawford, Cooling pond
steam fog,, J. Air. Poll. Cont. Assoc., 24, 860-864, 1974.
2. Hicks, B. B., The prediction of fog over cooling ponds, J.
Air Poll. Cont. Assoc., 27, 140-142, 1977.
3. Leahey, D. M., M. J. E. Davies, and L. A. Panek, A study of
cooling pond fog generation, Paper #78-40.2 presented at
the 71st. Annual Meeting of the Air Pollution Control
Association, Houston, Texas, June 25-30, 1978.
4. Everett, R. G., and G. A. Zerbe, Winter field- program at the
Dresden cooling ponds, Argonne National Laboratory
Radiological and Environmental Research Division Annual
Report, January-December 1976, ANL-76-88 Part IV, 108-
113, 1976.
5. Shannon, J. D. and R. G. Everett, Effect of a severe winter
upon a cooling pond fog study, Bull. Amer. Meteorol. Soc.,
59, 60-61, 1978.
6. Hicks, B. B., M. L. Wesely, and C. M. Sheih, A study of heat
transfer processes above a cooling pond, Water Resources
Res., 13, 901-908, 1977.
7. Wesely, M. L., Behavior of the thermal skin of cooling pond
waters subjected to moderate wind speeds, Proceedings,
Second Conference on Waste Heat Management and Utilization,
Miami Beach, FL, XI-A-40 )1-8), December 4-6, 1978.
8. Hicks, B. B., The generation of steam fog over cooling ponds,
Environmental Effects of Atmospheric Heat/Moisture Releases,
Proceedings of theSecond AIAA/ASME Thermophysics and Heat
Transfer Conference, Palo Alto, California, 24-26 May 1978
(Library of Congress Catalog Card Number 78-52527).
1127
-------
X-B-122
12
o
UJ
LU
CL
o
z 4
$
CO
00 O o
o „ _
o o oo o°,
nr* O On n°° O '
o
o
O
A
0 1 ° o ° ° o
°oo °
o A
A A A
00
I
e
• ©••
A **
-8-404
FOG EXCESS WATER INDEX (mb)
Fig. 1. Observations of the Fog Excess Water Index
made at the Dresden cooling lake (circles),
over cooling pond simulators at Argonne^
(triangles), and above a shipping canal near
Argonne (circles and crosses). Except in
the last case, solid symbols indicate that
fog was observed; fog was always observed
over the canal.
1128
-------
100
80
o 60
z
LJ
o 40
UJ
tr
u.
20
(a) Observed
I
1
_L
DEC JAN FEB MAR APR MAY
100
- 80
o 60
o 40
LJ
20
0
(b) Predicted
DEC JAN
FEB MAR APR MAY
Fig. 2. Observed (a) and predicted (b) frequencies of
occurrence of overnight steam fog (open circles)
and local rime ice deposition (solid circles)
at the Dresden cooling lake.
1129
-------
30
J:
I
I-
o.
UJ
o
UJ
O
o
20
10
0
I
I
I
I
0123
RIME INTENSITY CLASS
Fig. 3.
The relationship between the intensity of
overnight rime deposition and the reported
depth of the fog layer at Dresden.
1130
-------
MEASUREMENT AND EVALUATION OF THERMAL
EFFECTS IN THE INTERMIXING ZONE AT
LOW POWER NUCLEAR STATION OUTFALL
P. R.KAMATH, R. P. GURG, I. S. BHAT and P. V. VYAS
ENVIRONMENTAL STUDIES SECTION, H. P. DIVISION
BHABHA ATOMIC RESEARCH CENTRE, BOMBAY 400085
ABSTRACT
The paper reports observations and evaluation of thermal effects in
the Rana Pratap Sagar Lake in Rajasthan, India where one unit (200
MWe) of the Rajasthan Atomic Power Station is in operation. The
coolant waters are drawn 8-10 m below the lake surface through a
conduit and discharged through an open discharge channel with a
temperature rise of lO^C. There was a small increase in lake water
temperature in the vicinity of the outfall. Temperature profiles and
spread were mapped using insitu monitors.
These studies showedevidence of thermal stratification in the period
following winter and the existence of a well established thermocline.
Thermal stratification brought out specific advantages for thermal
abatement when the hypolimnion waters were well below the temper-
ature of surface waters. Parasitism and eutrophication were observed.
The thermal effects were accentuated by photosynthetic effects.
Proposal to utilise waste heat for algal culture in the Kalpakkam
nuclear site in South and mariculture (Lobsters, Prawns) in the
heated effluents canal at the Tarapur Atomic Power Station near
Bombay are discussed.
INTRODUCTION
The fresh water nuclear site in operation in India is in Rajasthan
on the R. Chambal drawing its coolant water from the man made
lake Rana Pratap Sagar(RPS). Only one reactor unit of 200 MWe
is commissioned. Of the two coastal sites, the Tarapur Atomic
Power Station has two reactors 200 MWe each commissioned in
1970. Kalpakkam Atomic Project is under construction.
Monitoring of heat distribution in foreshore waters, assessment of
thermal effects and investigations for waste heat utilisation are at
present carried out by the Environmental Survey Laboratories
installed on each site. •*•
RAJASTHAN ATOMIC POWER SITE (RAPS)
The RPS lake is about 3 km wide at the reactor site and extends
1131
-------
to 5 km downstrean upto the Darn. It receives about 8000 cfs of
water as tailings from a Dam located 32 km upstream. The lake is
in the main fed by R Chambal and its tributaries. The cooling
waters from the lake are drawn through a conduit 8-10 m below the
surface and 300 m off shore. The warm condenser effluents are
discharged through a canal which is open and virtually discharges
to the lake surface.
The water movement in the RPS is dependent on wind speed and its
direction. When the wind speed is less than 8-10 km/hr, the lake
waters are stagnant. The discharges remain close to the bank at
that time and spread along gradually. At wind speeds 15-17 km/hr
there is a conspicuous movement of water on surface in the
direction of wind. At speeds greater than 20 km/hr there is good
turbulence and mixing. ^
RAPS is an inland site located in the central part of the country
subjected to large differences (10-15°C) between day and night air
temperatures and about 25°C between peak day temperatures
between summer and winter. Temperature stratification in the lake
depends on the severity of winter which is considered acute when
the air temperatures reach 15-18°C at noon.
Stagnancy of water movement in the lake and design of the condenser
circulation system in RAPS held out interesting possibilities of heat
build up. Torrid summers when air temperatures touched 39°C and
lake water, 31.9°C at surface, showed the chance of thermal
pollution effects being observed attributable directly to the power
station reject heat.
Materials and Methods :
The condenser discharge was identified as it moved along or spread
over the lake surface by spiking the effluent stream with Rhodamine
B dye for visual marking.
Electronic Temperature Meter
An insitu temperature meter for taking vertical temperature
profiles was developed on the temperature dependent characteristics
of a semiconductor diode. The system used matched pair of
diodes as temperature sensors(°C). The response time is 80 sees
for a difference of 25°C. The sensitivity is 0. 1°C. Cable reach
20 m.
Dissolved Oxygen (DO) Meter "*
Field Lab Oxygen analyser supplied by Beckman Instruments, USA
1152
-------
was used for DO measurement in the laboratory study. The In situ.
monitor employed in the field was also obtained from USA - YSI
DO Meter model. The dual probe has a built in temperature sensor
for in situ measurement. DO concentration is read in ppm or as
percentage of saturation at the set temperature.
Observations -Results
Temperatures at surface observed during set hours of the day in
the different months following winter are presented in Table 1. The
temperature readings have to be read bearing in mind that the
intake is obtained from the hypolimnion (Elevation 335 m) and the
condenser discharge is made to the surface of the lake. Table 2 is
important because it gives the basic water quality change and
indicates what happens when water from the hypolimnion goes through
a churning motion in the condenser tubing. Values of DO(as percent
saturation), COD and BOD are given in the intake and discharge
streams. ^ Fig 2 gives the horizontal spread of heated effluents
giving the different Thermal zones. Fig. 3 gives the vertical pro-
files of temperatures in the different seasons to illustrate the
formation of thermocline and its gradual disappearance. "
DISCUSSION-EVALUATION
Data presented in Table 1 has brought out the important features
of the environment which go a long way to effect thermal abatement.
Even as the lake surface temperature reached 23. 5°C in February,
the hypolimnion waters were nearly 5. 6°C lower than surface. The
intake waters were cooler than if the design was a surface intake.
The difference between the intake water and surface temperatures
(5. 3°C) was greater than that between condenser intake and
discharge (4. 3°C), resulting a station output of heated effluents at
one degree Celsius less than the lake surface temperature. This
situation continued till peak summer temperature was reached- in
air (39°C). In the last week of May there were stray showers
which brought in the welcome change in air (air temperature
dropped by 4°C 39 to 35). There was a sudden change also in the
lake water temperature profile. After the rains came down the
picture changed entirely because the lake received plenty of water
supply from hinterland and tributaries (Fig 3 for profiles).
Fig 2 gives the different thermal zones around the outfall. There is
a 1°C rise in the close vicinity of outfall and then an intermixing
or a well spread out mixing zone where at the peripheri the temper-
ature was hardly above the ambient. The stretch of the mixing
zone was about 1. 3 km along the bank and 0. 3 km off shore. As
1133
-------
the Fig 3 recorded 8 km/hr wind speed, the condition was a
stagnancy in the lake. The lake stretched to about 3 km in the off
shore direction - and the thermal impact was therefore felt only
upto a tenth of available width.
Comments
Thermal Water Quality standards are generally set round the
*7 Q
following criteria. '»
1. Mixing Zone - an area where water quality standards are not
applicable. This depends on the limited spread of affected
region as a small fraction of the width at outfall.
2. Temperature standard - In cold climates a temperature
maximum of 32-32.2°C is recommended ; at any time however
the increase in temperature should not be greater than 2. 5°C.
in any part of the river system. In summer such increases
shall be less than 1.1°C.
In conditions available in tropics namely, wide differences in
diurnal and seasonal temperatures, and summer water temperatures
at surface exceeding 35-38°C, the above criteria are not relevant.
Except where the fishes get trapped or sedentary organisms are
present, the impact of heated effluents is not likely to be felt on
fish life directly at a low power nuclear station sites.
3. Table 2 suggests that under the prevailing circumstances of
design, the intake waters drawn from the hypolimnion and dis-
charged to surface, there was an enhancement in its DO content
in the process of circulation through the condenser. This effect
is demonstrated in the last four columns of Table 2. High C. O. D
and B. O. D of intake waters can be caused because of pollution at
depth (away from sunlight) and because the lake contained rotting
wood.
4. There is a matted growth of Vellesneria grass in the outfall
region which was found to be spreading and needed removal.
This may not be a direct impact of thermal discharge.
INVESTIGATIONS CARRIED OUT AT E. S. LABORATORY(ESL)
Gas supersaturation
Long hours of day light in tropical and subtropical regions can
cause algal growth in stagnant reservoirs. Photosynthesis can
lead to increased oxygen output in waters. If heated effluents
are also discharged, oxygen supersaturation can result because
of elevated temperature. These effects were studied in the ESL
experimental tank as follows :
20, 000 litres of raw water were transferred to a concrete tank of
size 9.6 m x 5.2 m x 0.5 m. The waters were inoculated with
1134
-------
culture of Sconedesmus and dosed with urea and other nutrients.
Measurements of DO, pH and temperatures were made through
out the duration of experiment. DO rose from 8.2 to 15 ppm
and pH 8.9 to 11.1. After 12 days exposure in sun and when
the water appeared as a pea soup from algal growth, fingerlings
of Indian carp (C. mrigala-8; L. rohita-4) were introduced in the
tank. In a week's time DO concentration exceeded 100% over the
saturation limit and the fishes progressively died. An examination
of dying fish showed that the fishes were breathing with difficulty
and the fish died from excess of oxygen. Laceration of tissue in
the gill region was seen in the dead fish.
The outfall region cannot be treated as stagnant because of
turbulence but these observations are likely to be met with in the
intermixing zone (Gas supersaturation) under tropical conditions.
Gas bubbles were seen to escape from the tank waters during the
day (14. 00 hrs).
Parasitism
Two 'happas1 (floating cages) were fabricated from nylon netting
built around a wooden frame 180 cm x 80 cm x 60 cm. The
happas were tied loosely to fixed pegs on the bank and released ,
one, into the discharge canal and the other (control) in the lake
upstream. The nylon cages were weighted to submerge partially
so that the introduced fishes always remained under heated
effluents in the 'test' cage, as the waters flowed through the net.
Each happa was charged with 20 numbers each of C. carpio,
L. rohita and C. mrigala. The experiment had to be given up as
the large fishes chewed away chunks of nylon. A set of impro-
vised cages was prepared with steel wires and placed as before,
in the discharge canal and upstream as Test and Control
respectively. 50 fingerlings of L. rohita each were placed in
each cage. The cages were provided with slit opening in the
top cover for addition of fish feed and to conduct periodical
examination. All the fishes were found dead in a month -
severely mauled in belly and mouth.
The experiment was repeated with fingerlings of L. rohita and C.
mrigal - weighing them before placing them in the cage :
Species No Discharge Canal Control
Av. Wt-g Length-cm Av. Wt-g Length-cm
L. rohita 13 238 26 (av ) 215 25
C. mrigal 11 210 26 234 27
JJ35
-------
5 fingerlings from the Test and Control were taken out for
examination. L. rohita had suffered very severely in the Test
cage losing nearly 70g weight. The controls were steady.
C. mrigala were also similarly affected but not to the same
extent. On closer scrutiny, the fishes were found to be infested
with an Ectoparasite identified as Alitropus typus. The parasite
Alitropus Typus is a blood sucking type and it attacks the soft
parts of the fish. '
Twin Aquaria Assemblies-Synergesis
The study of parasite proliferation, and synergesis caused by
waste heat as prima'ry pollutant, are being conducted in ESL.
attached to nuclear sites. For this purpose two sets of aquaria
experimental tanks were electronically connected in such a way
that the water temperature in one is 2. 5°C higher than in the
other which represents unaffected water of lake, upstream of
outfall. The temperature difference represents the peak temper-
ature increase in the intermixing zone. The experiments under
way are of two types, namely, where
1- Both the control and heated one are charged with 5 finger-
lings and 5 parasites (parasite behaviour and prolifration), and
2- In addition to heat other pollutants are added to the control
and the Test aquaria e.g., l^S, Clg, Hg, Chromates etc. to
study synergistic effects.
TARAPUR ATOMIC POWER STATION
The Tarapur Atomic Power Station is 100 krn north of Bombay
on the West Coast. The station output of 400 MWe is generated
by two BWR units. The reactors are located on a promontory
jutting out about 200 m into the sea. The intake is an open
channel drawing waters from upto one fathom depth, the channel
sloping into a stilling pool after a silt trap. The intake waters
are about 1°C less than the ambient sea water temperature.
The condenser discharge which has a temperature of 10°C above
the intake water, across the condenser ends, flows out to the
sea through two discharge canals - one north of and the other
south of the intake. 3 Although originally intended to prevent
recirculation of heated effluents by directing the discharge to
follow the tidal flow-at present the discharge flows out of both
the canals. The discharge canals are 14 m wide, 4 m deep
and nearly one kilometer long and do not contribute to thermal
abatement by themselves, except in high tides because of
-------
dilution and intermixing with on rushing waters. During other
periods the wind cooling takes place only to the extent of
lowering by 1°C as the water reaches the end of the canal.
Two important natural factors that help control thermal
pollution are i) monsoon rains lasting for 3 months and strong
breeze, and ii) turbulence caused by semidiurnal tides which
may rise upto 5-6 m, giving effective mixing and dilution. ^
There is no evidence as yet to demonstrate the negative
effects of thermal discharges at Tarapur. Even in the many
sedentary species present along the coastline and creeks, no
accelerated growth of vegetation or radioactivity uptake in fish
have been observed. The temperatures of heated effluents
drop suddenly by about 3°C, even under neap conditions, when
the effluent stream meets the sea at the discharge inal end in
the first abatement step. Under low tide conditions about 400 m
of the shelf are virtually bare and the effluents flow over the
exposed rocks. Thermal monitoring did not show any increase
in temperature beyond 1.5 km from the discharge canal end.
WASTE HEAT UTILISATION
Studies on Waste Heat Utilisation have been initiated for some-
time in ESLs and it may take sometime before effective
techniques for waste heat utilisation are developed for commercial
exploitation.
At the Kalpakkam E S Laboratory, where the Madras Atomic
Power Station is located it is intended to use waste heat for
large scale production of algal cultures. An algal pond of
size 12 m x 9 rn is in operation from last 3 years using solar
heat and domestic waste nutrients. When the power station goes
into operation, waste heat will come up as an additional source
of lowly rated heat flow. Mariculture is on the cards- parti-
cularly growth of shrimps, prawns and lobsters.
Prawns form a major exchange earning industry in the
country. 10 Among the different species Penaeus indicus and
Penaeus mo nod on are widely employed for developing creek and
estuarine fisher-as. Experiments are being "initiated in co-
operation with Tamilnadu mariculture teams for setting up
1137
-------
experimental assemblies in the discharge canals at Tarapu'r
for production of Prawns and Lobsters. Laboratory study
is also being planned at the Rajasthan -E S Laboratory for
production of fresh water prawns-type M. rosenbergii.
Acknowledgements : The authors desire to acknowledge
assistance from several colleagues and particularly of
Mr K V K Nair (ESL-Kalpakkam) Mr B Dube (ESL-
Rajasthan) and Shri S. Chandramouli (ESL-Tarapur). The
authors gratefully acknowledge support received from
Dr. A. K. Ganguly, Director, Chemical Group and
Mr.S.D. Soman, Head H. P. Division, BARC.
1138
-------
REFERENCES
1. Kamath, P. R. t 'Environmental Surveillance at Nuclear
sites in India1, NUCLEAR INDIA April-May 1978
(Publ: DAE, Bombay 400 001)
2. Kamath, P. R., Bhat, I.S., Gurg, R. P. , Adiga, B. B. , and
S. Chandramouli. 'Seasonal Features of Thermal Abatement
of Shoreline Discharges at Nuclear sites' presented at the
IAEA Symposium on Environmental Effects of Cooling
Systems at Nuclear Power Plants; OSLO 26-30;Aug 1974.
3. Kamath, P. R., Bhat, I. S. , and Ganguly, A. K. , 'Environ-
mental Behaviour of discharged Radioactive Effluents At
Tarapur Atomic Power Station1 IAEA-USAEC Symposium
on Environmental Aspects of Nuclear Power Stations;
10-14 Aug 1970 N. Y., USA.
4. Gurg, R. P. , Bhat, I. S., and Kamath, P. R.-Progress report
of ESL Rajasthan Atomic Power Station - BARC report
No 1-369, 1975.
5. Gurg, R. P., Bhat, I. S., and Kamath, P. R. 'Thermal
Pollution from Nuclear Power Production under Tropical
Conditions' Presented at the symposium on Operating
Experience of Nuclear Reactors and Power Plants.
Feb 7-9, 1977 Bombay, DAE.
6. Kamath, P. R., Gurg, R. P. , Sebastian, T. A. , Vyas, P. V.,
Dube, B. , and Nair, K. V.K. , 'Impact on water quality
from Discharge of Thermal Effluents in RPS Lake'
Presented at the IAEA Research Coordination Meeting on
Thermal Pollution, Kalpakkam Dec 5-9, 1977.
7. Miller, D. C. , and Beck, A. D. 'Development and Application
of Criteria for Marine Cooling Waters' Paper IAEA
-SM-187/10 in Symposium mentioned in Ref 2 above.
8. Jeter, C 'An Approach to Thermal Water Quality Standards'
Presented at the Conference on Waste Heat Management
and Utilisation, 9-11 May 1976, Miami Beach, Florida,
University of Miami.
9. Chaudhary, R.S. , and Walker M. F. , 'Parasitic Behaviour
of Fresh Water ISOPOD ALITROPUS TYPUS in fishes of
Rana Pratap Sagar India' (In press) 1978. Preprint
communicated.
10. Fishes and Fisheries - publication of CSIR, New Delhi
1962 Supplement to 'The Wealth of India1 Vol IV.
1139
-------
Table - 1
RAJASTHAN: SEASONAL VARIATION
IN AMBIENT
AND COOLING TEMPERATURES
Date
and time
(hrs)
26.2.74
(11. 00)
9. 3.74
(11. 10)
18.3.74
(14. 30)
10.4.74
(15. 30)
25.4. 74
(13. 00)
20.5.74
(15. 30)
25.5.74
(10. 30)
1.6.74
(11.30)
5.6. 74
(17.15)
13.6. 74
(10. 30)
27.6.74
(14. 10)
Power
MW
130
155
150
170
170
175
150
180
160
175
125
Ambient
Air
23.8
32. 2
32. 5
36. 5
37. 0
39.3
35. 0
35.0
36. 0
35.5
38. 8
temp. °C
Lake
surface
23.5
26.0
24.7
29.2
28.5
31.7
30.0
30.0
30.0
31.0
29.7
Coolant
intake
°c
18.2
17.2
18. 1
19.2
19.6
19. 0
21. 5
24. 0
26. 0
23. 0
27.5
Coolant
outlet
°C
22,5
22.4
24.2
28.5
27. 8
30.0
32.0
36.0
37.0
35.0
29.7
1140
-------
Table - 2
Date
30
11
27
25
6.
3.
11
.6.
.8.
.8.
.9.
10.
11.
. 11
76
76
76
76
76
76
.76
COOLANT
Reactor
Power
MWe
165
170
0
172
160
180
175
WATER QUALITY
-Temp.
differ-
ence
dt °C
10.5
11.0
2.0
10. 5
10. 5
8. 0
11.0
AT
Dissolved
oxygen ppm
percent
saturation
Int- Disch-
ake arg e
92
91
76
83
75
85
72
.1
.0
. 9
.5
.6
.0
.8
105.
110.
81.
100.
92.
95.
101.
5
8
6
0
4
7
5
RAPS
C.O. D
Intake -
Disch-
arge
(ppm)
+ 1.2
+ 0
+ 0. 3
+ 0.7
+ 2.0
+ 1. 1
+ 0. 3
B. O. D
Intake
Disch-
arge
(ppm)
+ 1.
+ 0.
+ 1.
+ 0.
+ 1.
+ o.
+ 0.
2
5
1
5
5
9
3
1141
-------
1142
-------
01
FIG. 2 THERMAL ZONE OF RAPS DISCHARGE J18-1-77)
POWER LEVEL *7C MWe
CONDENSES INTAKE ' B°C
CONDENSER D!SCHA3G£-2SC
AMBIENT TEMP 2£C
WIND SPEED 3Krr.|Hr
W1NTJ DIRECTION 5E
3_'JM£ CS.-'
AT A 3- J M
AT 9 1-5 w
AT C -1-0 v
i? 0
-------
FIG. 3
'HERMAL PROFILE IN LAKE KPS DAM
5-8-74
352 ! -
i
i
343 (-
344 -
340 :-
336 -
332-
328 j-
1
324 i-
17-6-74
23-10-74
13-12-74
7-2-75
J7-4-75
29-5-75
29-7-75
=2
<
(.T
cc
§
^
<
320
318
Intake
Conduit
i r
Locationi 333 m
20 24 28 30 20 24 28 30 20 24 28 30 23 24 28 30 16 20 24 26
'C r *C *C *C
18 20 24 23 20 24 28 30
"C *C
20
28 30
-------
RAJASTHAN
ATOMIC
POWER SITE
4435
45'
7/50*
1145
-------
RIVER THERMAL STANDARDS EFFECTS ON COOLING-RELATED
POWER PRODUCTION COSTS
by
T.E. Croley II, A.R. Giaquinta,
M.P. Cherian, and R.A. Woodhouse
Iowa Institute of Hydraulic Research
The University of Iowa
Iowa City, Iowa USA
ABSTRACT
Power plant cooling costs and water consumption for various river tempera-
ture standards are presented for existing and proposed future power plants
located along the Upper Mississippi River. Three models previously devel-
oped at the Iowa Institute of Hydraulic Research are combined to evaluate
the cooling-related costs of river thermal standards. These costs depend
on the meteorological conditions at each power plant site, and they are sum-
med for the river reach of interest. The existing thermal standards case,
the free-discharge or no-thermal standard case (all plants employ open-
cycle-cooling) , and the extreme case of no allowable discharges are chosen
to show the dependency of power-production-related cooling costs and water
consumption on various criteria. A critical appraisal of the worth of ther-
mal standards in terms of water consumption and other costs is thereby pos-
sible, so that subjective assessments of the standards can proceed with full
knowledge of the trade-offs involved between the costs of power production
and environmental impacts.
INTRODUCTION
A joint meeting of state and federal governmental agencies on Mississippi
River temperature standards was held in St. Louis, Missouri, on March 3,
1971. Temperature standards were proposed because it was felt that heated
effluents from nuclear-and fossil-fueled power plants could raise river
temperatures enough to harm the biota. The report recommended that the
maximum "artificial" rise in water temperature not exceed a prescribed limit
above the recorded natural temperature, nor should the actual temperature ex-
ceed the maximum safe temperature, whichever constraint dominates. It was
decided at this meeting that power plants could easily comply with the stan-
dards with closed-cycle cooling being the most economically feasible means.
The existing standards now governing thermal discharges into the Mississippi
River include a specified maximum allowable water temperature for each month
of the year and a maximum allowable temperature rise of 5°F along the entire
length of the river. Future regulations will further limit thermal discharges
into the river. The U.S. Environmental Protection Agency has mandated that
1146
-------
thermal discharges into natural rivers from power plants placed into ser-
vice after 1 January 1970 (or 1974 depending on the size of the plant) will
not be permitted after 1 July 1983 [1].
The standards were aimed at environmental enhancement with little consid-
eration of resultant costs. It is extremely difficult to determine a set of
standards which adequately represents both the environmental and beneficial
use viewpoints. The difficulty stems from a lack of knowledge about the
level of environmental preservation (or beneficial use) to be maintained
by the standards and how that level should be measured. There have been
many studies of the environmental ramifications of thermal loads on rivers.
The common characteristic of them all is that the environmental impacts
either are not quantifiable or are multidimensional, or both. In any
event, it has been impossible to associate a numerical indication of environ-
mental impact with a set of river standards. However, the environmental im-
pact is real and must be addressed in any intelligent determination of river
temperature regulations. This problem of evaluating alternate standards in
terms of their environmental impacts is typical of situations requiring sub-
jective evaluations to be made.
If the economic impacts of environmental standards are understood by de-
cision makers, then alternate sets of standards can be evaluated in terms
of the "costs" required to meet those standards and the amount of environmen-
tal protection consequent to those constraints. In other words, trade-off
"costs" of providing different levels of environmental protection
(resulting from different sets of thermal standards) can be investigated.
The question can be asked for each set of standards to be evaluated: "Are
the environmental gains justified in relation to the expenditures?" This
question still involves a subjective choice, but it is much easier to answer
than the original question: "How much environmental protection should be
provided?" The trade-off question can be asked over and over for increas-
ingly stringent sets of standards until a desired balance between environ-
mental objectives and consequent economic penalties and water consumption
is established.
This study looks at the question of the "costs" (both economic and water
consumption) to the utilities (and the public) of meeting various thermal
standards for the Upper Mississippi River from the source to the southern
Iowa border. The costs of the existing thermal standards are assessed by
computing the marginal increases in monetary expenditure and water consump-
tion over the "free-discharge" or no-standards case wherein all utilities
are assumed to utilize the most economical and lowest water consumptive
system: the once-through cooling system. The additional "costs" (over
the existing standards and over the free-discharge case) of more restric-
tive thermal standards (the "zero-discharge" thermal standard) are also
assessed for a complete realization of the implications of impacts of these
standards. It is extremely important to realize that the figures given
herein are illustrative only since fixed unit costs were assumed across-
the-board for all utilities along the study reach and fixed assumptions
were made for the operation of all plants. The numbers cannot be taken
1147
-------
as indicative of true costs of any one utility but serve to indicate the
generalized total costs for the entire study region. The power plants in-
cluded in the study are those presently operating and those proposed for fu-
ture construction (through 1994) having capacities of 25 MW or greater; all
of the utilities lie in the Mid-Continent Area Power Pool (MAPP) geographical
area, and most are MAPP members. The computational scheme to assess the
worth of thermal standards requires the use of three models previously
developed at the Iowa Institute of Hydraulic Research. The first model ex-
amines the steady-state thermal regime along the study reach of the Upper
Mississippi River. The model is used to locate regions where river tempera-
tures exceed the allowable limits for any prescribed set of thermal standards,
and to assess river evaporation for heat loadings consequent with those ther-
mal standards. The second model evaluates cooling-related costs of back-
fitting existing power plants (identified as requiring backfitting with
the first model under a set of thermal standards) with mechanical draft wet
cooling towers. The third model computes cooling-related costs of out-
fitting proposed power plants (identified as requiring outfitting with the
first model for a set of standards) with once-through and closed-cycle (wet
tower) cooling systems.
COMPUTATIONAL MODELS
Iowa Thermal Regime Model (ITRM)
A predictive computational model for computing temperature distributions
along natural rivers (ITRM) was developed by Paily and Kennedy [2]. The
steady-state version presented by Paily et al. [3] is used to compute the
thermal regimes for the natural, free-discharge, existing, and no-discharge
cases.
The model is based on a numerical solution of the one-dimensional convection-
diffusion equation, and it predicts the longitudinal distribution of cross-
sectional average temperature along a river. The total river length is
divided into smaller reaches, and temperature distributions are computed for
each reach separately. The solutions for adjacent reaches are linked by
the common conditions at the junction points connecting them. Each reach
of the river can have multiple thermal inputs and tributary inflows. The
formulation allows for changes in the channel characteristics and the river
flow rate. Variations in weather data from place to place also are taken
into account. The model is one-dimensional and assumes complete mixing of
the heated effluent with the river. To compute thermal discharges of pro-
posed plants, the model assumes in-plant efficiencies of 85 and 95 percent,
an overall plant efficiency of 32 and 36 percent, and a condenser tempera-
ture rise of 18°F and 25°F for fossil-fueled and nuclear power plants, re-
spectively.
Based on these assumptions, it is clear that the steady-state thermal re-
gime model presents only an overview of the aggregate thermal profile of
a river; it does not yield a detailed assessment of the actual temperature
1148
-------
distribution. However, this model does give adequate representation of the
river temperature distribution.
The thermal regime model is used to determine the temperature profile along
the Mississippi River in the MAPP geographical area corresponding to average
flow and weather conditions. The input data used for the computations are
the following:
1. heat loads from power plants of rated capacity 25 MW or greater,
industries, and municipalities located on the main stem of the river;
2. monthly mean values of daily flow rates measured at 12 U.S. Geological
Survey gaging stations along the river;
3. monthly mean values of daily weather conditions including air tempera-
ture, wind speed, relative humidity, atmospheric pressure, cloud cover, and
solar radiation measured at 6 first-order weather stations of the National
Weather Service; and
4. channel top widths at approximately ten-mile intervals determined from
river profiles developed by the U.S. Army Corps of Engineers.
Temperature profiles also were determined for the 7-day, 10-year low flow
along thd river combined with average weather conditions for the months of
August and November to compute the extreme assimilation capacity of the
river at locations of proposed power plants. Evaporation rates along the
river were computed for the average flow thermal regimes. The river discharge,
climatological variables, and channel geometry parameters were assumed to
vary linearly between adjacent measuring stations.
Backfitting Model
This model evaluates the cost of backfitting a power plant or unit currently
operating on open-cycle cooling with a mechanical draft wet cooling tower
[4]. The major factors considered in the economic assessment of backfitting
an existing unit are:
1. the cost of installing the cooling tower, including materials, labor,
site acquisition, and preparation;
2. the plant downtime for system changeover;
3. the provision of additional generating capacity to replace the power
consumed by the cooling system;
4. the operation and maintenance costs of the cooling-system; and
5. the additional cost of power generation due to limitations imposed by
the use of the closed-cycle system.
The first three of these quantities are capital costs and the last two are
operating costs incurred over the remaining lifetime of the plant. Once
these factors have been determined, the total cost may be computed by using
the fixed-charge-rate method [5]. It is possible to design mechanical
draft wet towers of any size, but realistically the lowest-cost tower would
be built in practice. Therefore, a range- of tower sizes must be investiga-
ted at each site to determine the optimum design. The characteristics of
the power plant required for backfitting calculations are the accredited
1149
-------
capacity of the unit, the type of plant (fossil or nuclear), the thermody-
namics of the existing turbine and condenser systems, and the economic ;sit-
uation of the utility operating the unit. Simplifications and assumptions
made in the development of the model are:
1. the power plant is operating at 80 percent of the accredited capacity
throughout the year to satisfy a constant demand of the same amount;
2. the plant or unit is considered to be operating with a constant, rela-
tively low, turbine back pressure, and the corresponding heat rejection rate
is known for an existing open-cycle cooling system;
3. the existing condensers are retained without modification; and
4. the same capacity factor is used both before and after backfitting.
With these assumptions, computation of capital and operating costs of back-
fitting with a mechanical-draft wet cooling tower may be achieved by using
calculation procedures outlined by Croley et al. [4] and described briefly
in the following subsection on outfitting. Of foremost importance in the
backfitting calculations are the capacity loss, the energy loss, the excess
fuel consumption (the difference between the fuel consumption with an open-
cycle cooling system and the backfitted system) and the excess maintenance
cost. The model also may be used for the computation of water consumption
by the cooling tower.
Outfitting Model for Once-Through and Closed-Cycle Cooling Systems
The economics of power plant cooling performance is mainly dependent on the
turbine-condenser subsystem characteristics and on the size and type of
cooling system. Two basic types of turbines are considered in this study
as representative of those currently in use. The first type is a high
back-end loaded unit of contemporary design, and the second type is a low
back-end loaded unit primarily used in older plants. Operating character-
istics of these turbines are given in reference [5].
Cooling characteristics curves may be determined for any specified size and
type of cooling system using the appropriate model. The size of a once-
through cooling system is primarily determined by the condenser flow rate
and by the design heat assimilation capacity of the receiving water body.
The cooling characteristics curve is determined by the size of the system
and by the actual heat assimilation capacity of the stream. For mechani-
cal draft wet cooling towers, the size of the cooling system is specified
by the dimensions of the tower and the design meteorological conditions chosen
at the site. The cooling characteristics may then be determined from the
basic thermodynamic model described elsewhere [6]. The operation point of
the cooling tower is defined as the intersection point obtained by super-
imposing the appropriate cooling characteristics curve on the turbine charac-
teristics curve as described in reference [6]. This operation point com-
pletely describes the performance of the power plant cooling system in
terms of hot water temperature, heat rejection rate, power output, arid tur-
bine back pressure for any given set of meteorological conditions and
power loading. Condenser sizing is obtained from the operation point
1150
-------
corresponding to the design characteristics curves using design meteorologi-
cal and stream conditions. Capacity loss is obtained from the operation
point corresponding to the extreme characteristics curve evaluated using
extreme meteorological and stream conditions. Fuel consumption, make-up
water (water evaporation), energy loss, and other quantities may be obtained
from the operation point corresponding to the "prevailing or existing"
characteristics curve. If 'the cooling system is smaller or if the meteoro-
logical or stream conditions become more adverse, the operation point will
shift such that the resulting power output will decrease. The model has
the capability of computing fuel consumption, water evaporation, energy
loss, and other parameters for a given distribution of meteorological
and stream conditions.
Power plant cooling costs are composed of capital costs which include the
cost of tower structures, once-through cooling structures, condensers,
pump and pipe systems, and replacement capacity; and operating costs which
consist of the costs of fuel, make-up water, water treatment, maintenance,
and replacement energy. These costs are determined by using appropriate
unit costs and monographs, I see TABLE I and reference [7]). The unit costs
are expressed in terms of 1976 dollars and are valid only for the MAPP
region. The total cooling-related cost of power production is computed
using the fixed-charge-rate method.
PROCEDURE
Changes in thermal standards would alter existing assimilation capacities
of the river and, hence, the operation of once-through cooling systems lo-
cated along the river. Thermal violations (defined as cases where the
maximum river temperature or the increase of river temperature due to heated
effluents exceeds the allowable limit) which occur as a result of a change
in thermal standards would require these power plants to derate or to back-
fit a closed-cycle cooling system. Associated with these alternatives
are high energy losses and capital expenditures in terms of cooling tower
construction and associated changes in operating cost, and water evapora-
tion as a result of the closed-cycle operation. To relate thermal standards
to the cooling-related costs of power production and water consumption,
the following procedure was developed:
1. Determine the thermal regime of the river for the heat loads from
existing and proposed power plants at average flow conditions. Compute
net evaporation resulting from thermal discharges.
2. Choose a specified set of thermal standards.
3. Determine cases where thermal standards are exceeded, if any.
4. Backfit power plants with mechanical draft wet cooling towers wherever
thermal violations occur, and compute annual backfitting costs (including
all capital costs). Calculate the corresponding annual water consumption.
-5. Determine the annual energy loss at plants where thermal standards are
exceeded, and determine the corresponding cost of purchasing replacement
energy, without adding auxiliary cooling capacity.
6. Choose the more economical alternative of steps 4 and 5 at each affec-
ted location.
1151
-------
7. Compute the annual operating costs of existing power plants that do not
violate thermal regulations. (The capital costs of these cooling systems are
not considered.) Determine the corresponding annual water evaporation.
8. Size cooling systems of proposed power plants (the capital costs of these
cooling systems are considered). Determine total annual costs and water
evaporation.
9. Compute the total annual cooling-related costs of power production and
the total water consumption corresponding to the specified thermal standard.
10. Repeat the computations for other thermal standards.
Computations in steps 1 and 3 are carried out for the months of February,
May, August, and November. These four months are assumed to adequately re-
present the meteorological and hydrological conditions that exist through-
out the year. Backfitting is carried out at an existing plant if thermal
standards are exceeded during any one of these months. Annual operating
costs for existing and proposed power plants are computed corresponding to
meteorological and stream conditions that exist during the chosen study
months.
Proposed plants are sized corresponding to extreme meteorological conditions
and the 7-day, 10-year low flow hydrological condition. Capital costs of
proposed power plants are added to annual operating costs by using the
fixed-charge-rate method. For existing power plants a uniform remaining
life of 20 years is assumed; for proposed power plants the expected life
horizon is assumed to be 35 years.
RESULTS
The cost of meeting a specified thermal standard is defined as the differ-
ence in total cooling-related costs associated with the thermal standard
and the total cooling-related costs associated with no thermal standards.
In the latter case, all existing and proposed power plants utilizing a
closed-cycle cooling system are considered as employing once-through
cooling, as noted earlier. Marginal cost changes are computed and the worth
of the thermal standard is evaluated; also changes in water consumption that
occur as a result of variations in thermal standards are presented. The
three thermal standards considered herein are the existing regulations, the
"no-discharge" standard, and a free-discharge condition.
Water Consumption
Water consumption resulting from power plant operation is due to the increased
river temperatures caused by heated effluents and to evaporation from wet
cooling towers. Natural evaporation from the study reach (without power
plants) was obtained from the ITRM with the appropriate data set and is shown
in Fig. 1. The annual equivalent of this figure integrated over the river
and the year is 257.1 million m^. The variations in natural evaporation are
a result of the natural variations of the top width of the river. To eliminate
1152
-------
the effects of top width from this and succeeding figures, the unit evapora-
tion is calculated by dividing evaporation by the river width. Unit natural
evaporation is depicted in Fig. 2. Now, the dips and peaks in the curve are
seen to correspond to the locations of the weather stations which are labeled
at the tops of Figs. 1 and 2. It is noted that natural evaporation for the
month of February and most of November is zero because of the presence of
ice cover on the river during these months. Sublimation from ice is
neglected in this study.
The actual river evaporation corresponding to the existing and proposed
power plants with existing thermal standards was computed with the ITRM and
appropriate data sets on a unit basis, and the unit natural evaporation of
Fig. 2 was subtracted to give the unit net evaporation on the river, which
is plotted for August conditions in Fig. 3. It is important to note that
this figure pertains to unit net river evaporation only and does .not include
the other cooling-related evaporation losses from wet cooling towers since it
is not possible to present those losses on a unit basis. Sudden spikes
in the evaporation curve are a result of thermal discharges at those locations.
Certain interesting features can be observed in the unit net evaporation for
the month of February, Fig. 4. As a result of ice cover, no evaporation
occurs unless the temperature of the river water is above 0°C as a consequence
of heated discharges from power plants. Water temperatures above freezing
are not sustained over any long reach of the river because of adverse meteoro-
logical conditions; therefore, the unit net evaporation abruptly drops to zero.
It also should be noticed that the magnitude of the evaporation rates is much
greater than the corresponding rates for August. This increase is primarily
due to the fact tha,t the air temperature in February is below the water temp-
erature at these locations; other meteorological conditions are also conducive
to this phenomenon during February.
By integrating the net river evaporation along the river and over the year
and adding the total evaporation from wet cooling towers (if there are any),
the total annual evaporation can be calculated for each set of thermal
standards. This calculation was made for the free-discharge condition, the
existing thermal standard, and the no-discharge thermal standard; the
results are tabulated in TABLE II. It is seen from this table that the ex-
isting standards result in an annual increase of about 2.60 million m3 over
the free-discharge condition of 156.9 million m3 (an increase of 1.67 per-
cent). The no-discharge standard represents an annual increase of 5.68
million m3 over the free-discharge condition (an increase of 3.62 percent)
and an annual increase of 3.08 million m3 over the existing thermal standards
of 159.5 million m3 (an increase of 1.93 percent). Net evaporation for the
free-discharge condition represents total evaporation (no cooling tower
evaporation) and is smaller than the total annual evaporation for the
existing thermal standards. For the no-discharge thermal standard, net
evaporation is zero. On the other hand, evaporation from wet towers is higher
for the no-discharge condition. Thus, it is easily seen that total annual
evaporation is greater with the no-discharge standard since water consump-
tion from wet towers is higher than evaporation from the river surface for
comparable cooling duties.
1153
-------
Economic Costs
Costs for the various thermal standard conditions are computed from the back-
fitting and outfitting models for each utility identified as exceeding the
thermal standards with ITRM. Costs for existing thermal standards are pre-
sented in TABLE III. These computed results indicate that the average
cooling-related cost of power production in the region of study is of the
order of 15.20 mills/kw-hr for the present thermal standards, which repre-
sents a relative increase of about 1.58 mills/kw-hr over the free-discharge
case. The value of 1.58 mills/kw-hr may then be considered as the average
"cost" of the existing thermal standards. Details of the cost reduction
as a result of the free-discharge thermal standard are presented in TABLE IV.
It should be mentioned that the fuel consumption cost with the once-through
cooling system is higher than the corresponding cost for the same power
plant outfitted with a mechanical-draft wet cooling tower. This phenomenon
also is observed in the backfitting operation and is due to the fact that
with a wet tower, the power plant is derated at certain times as a result
of adverse meteorological conditions. Consequently, fuel consumption is
lower with the wet tower. Under these conditions, however, large amounts
of replacement energy are required which result in high replacement energy
costs. The decrease in fuel consumption of plants with cooling towers is,
of course, counteracted by an increased fuel consumption of the plants
supplying the replacement energy.
The no-discharge thermal standard involves additional costs incurred as a
result of backfitting once-through cooling systems with wet cooling towers;
the cost increases are listed in TABLE V. These costs must be added to
the costs obtained with the existing thermal standards to compute the cost
of the no-discharge standard. It is seen that the no-discharge thermal stan-
dard represents an average increase of 2.042 mills/kw-hr over the existing
average annual cost. The "cost" of the no-discharge standard is, therefore,
of the order of 3.62 mills/kw-hr as compared to the free-discharge condition.
All regional cost figures are summarized in TABLE VI.
CONCLUSION
The knowledge of costs and water consumption associated with the free-dis-
charge, existing, and no-discharge thermal standards should provide a use-
ful guide in reexamining present criteria and perhaps setting up new thermal
regulations for the Upper Mississippi River. If the "worth" of these thermal
standards in environmental terms has been established, the standards can be
assessed with regard to their "costs" and the trade-offs between the costs
of power production and environmental impacts is made clear. Undoubtedly,
it is extremely difficult to determine the level of environmental protection
required; however, subjective assessments can be made with an understanding
of the costs associated with thermal standards. The procedure and the re-
sults presented in this paper should help in enabling intelligent decision-
making with regard to the establishment of thermal standards. The costs
cannot be judged as accurate on a site-to-site basis, but serve to illus-
trate the general impact of various standards for the entire study reach.
1154
-------
ACKNOWLEDGEMENTS
This project was financed in part by a grant from the Mid-Continent Area
Power Pool (MAPP) and by a grant from the U.S. Department of the Interior,
Office of Water Research and Technology under Public Law 88-379 as amended,
and made available through the Iowa State Water Resources Research Institute.
Funds for computer time were provided by the Graduate College of The Univer-
sity of Iowa.
REFERENCES
1. U.S. Environmental Protection Agency, "Development Document for Effluent
Limitations and Guidelines and New Source Performance Standards for the
Steam Electric Power Generating Point Source Category," United States
Environmental Protection Agency, Washington, D.C., Oct. 1974.
2. Paily, P.P. and Kennedy, J.F., "A Computational Model for Predicting
the Thermal Regimes of Rivers," IIHR Report No. 169, Iowa Institute
of Hydraulic Research, The University of Iowa, Iowa City, Iowa, Nov. 1974.
3. Paily, P.P., Su, T.Y., Giaquinta, A.R., and Kennedy, J.F., "The Thermal
Regimes of the Upper Mississippi and Missouri Rivers," IIHR Report No.
182, Iowa Institute of Hydraulic Research, The University of Iowa, Iowa
City, Iowa, Oct. 1976.
4. Croley, T.E., II, Giaquinta, A.R., and Patel, V.C., "Wet Cooling Tower
Backfitting Economics," Journal of the Power Division, ASCE, Vol. 104,
No. PO2, Apr. 1978, pp. 115-130.
5. Giaquinta, A.R., et al., "Economic Assessment of Backfitting Power Plants
with Closed-Cycle Cooling Systems," Report No. EPA-600/2-76-050, U.S.
Environmental Protection Agency, Research Triangle Park, North Carolina,
' Mar. 1976.
6. Croley, T.E., II, Patel, V.C., and Cheng, M.-S., "Thermodynamic Models
of Dry-Wet Cooling Towers," Journal of the Power Division, ASCE, Vol.
102, No. PO1, Jan. 1976, pp. 1-19.
7. Croley, T.E., II, Giaquinta, A.R., Lee, R.M.-H., and Hsu, T.D., "Opti-
mum Combinations of Cooling Alternatives for Steam-Electric Power Plants,"
IIHR Report No. 212, Iowa Institute of Hydraulic Research, The University
of Iowa, Iowa City, Iowa, July 1978.
1155
-------
TABLE I
Description
UNIT COSTS
(assumed to be spatially uniform in the region of study)
Cost _
Unit cost of wet towers
Unit condenser cost
Unit replacement capacity cost
Unit land cost
Unit make-up water cost
Unit waste-water treatment cost
Unit fuel cost
Unit maintenance cost of wet towers
Unit replacement energy cost
$21/TU
$12/sq. ft of
surface area
$400,000/MW
$5,000/acre
$1.8/1000 gals
$0.15/1000 gals
$0.004/kw-hr
$300/yr/cell
$0.02/kw-hr
TABLE II
WATER CONSUMPTION COMPARISON OF DIFFERENT THERMAL STANDARDS
Thermal Standard
Net Annual
Evaporation from
River Surface, m^
Annual
Water Consumption
of Wet Tower?, m^
Total Annual
Evaporation
Free-discharge
Existing
No-discharge
156,923,000
68,851,000
0
90,670,000
162,600,000
156,923,000
159,521,000
162,600,000
1156
-------
TABLE HI
COMPUTER COSTS OF EXISTING THERMAL STANDARDS (197'i do
Plant 4 Unit Ho.
Clav Boswfcll #4
LlK River »l-3
Kiversiae tit!
High Bridge *=-£
Alma Si- 5
Aima fit
^enOd -lA-^D , J
..=msi: •-• *1-1
Mfl^o,-. Dewey
DoLijqut «2-4
K.L. Kapj. Hi-;
Molinr- *'—'
Louisa ».
Location
1167
lie-
got
^0'
391
c.51.7
r^.\
5S(
= lr
4 = 7.1
4-"";. ;
4?7. :
4U4
Capacity
150
500
1420
5fc8.fi
49. ^5
239. 3C
209. y
350
Jt2. j
:.C
= 1.75
2:7
BO
23b.S
1600
73.6-5
1 2 3 . t
65C
Cn^,-a
r*'W.
riwr-
nm.
NiCL
FiOK
FiOF
Fl'Jt-
Fl'-f
Fi
F;
FI
F^
F2
N^
F2
Fl
flWT
~,JU. .
"
Com;
3. ;:•..
--
j",,.,
ca i
..•Jl '
.';:.::
t.l C05t~ !«•
LOOI.,,0
2. 161
j .f-7:
ii.t-4t.ti
.1 lion dolla
4.74:^
o
7.5^7
151
Land
u.25
"
0. 325
rotaJ
]9.9t^
1.585
1.772
27. OL
1 f^
2.^44
0..7C
i-.Jt.13
3.990
O.Btd
J.27ft
J.J49G
0.8T60
Kuel ]
13. bi
4fl.il
124.0
51 . li*
4.<»03
22 . Of
.1 ',.'_. 9
19. 34
1^.24
3^.11
4.C13
'>. 7U7
2i.4e)
4 . 7uy
i!tt.94
7. 37J
J 1 . 98
•,.V8fc
11. 3y
J3.H2
'j7.lJ
Kei-lacenwM
0
3. SOI
13.09
4. 192
fi
0
0
0
u
0
0
u
0
0
0
0
0
c
S. 329
MaKe-U}.
--
4 . 02C.
11.17
1J.72
1.^17
::
--
!;__,7e
—
0.101E.
l'.307J
0. 3464
0,0307
::
0.1333
MaJJ-LcnaricP
'j.0:7h
0.005-i
0.0141
0.0159
0.0051
0.0030
0.0663
0.003-
0.0742
0.0634
J.0097
0.0559
0.0fj91
0.0282
U.P501
0.00-i3
C.03C3
^.0192
J.0287
u.0178
C,'. 0069
Total
13.85
52.04
145.5
IbS.b
5&.63
4.606
22.15
19.35
32.32
33.47
4.621
S.S14
23.55
21.01
7.4C1
22.03
196. (f
6.791
6.009
11.42
13.34
l'>-^
miJl^kv-hr
3. 17fc
4. as;
5.U28
5.032
4.20S
3.15S-
3.2D3
3.153
3.1T5
3.175
3.189
3.210
3. 176
3.205
3.2
3. 18
7.479
3.139
3.195
J.183
3.102
•1.900
Twtal
13. H5
54.99-
149.5
178.:
56.63
4.606
Ji.15
2S.64
19.35
32.99
33.47
4.621
3.814
23.55
21.01
7.401
22.03
196.0
f, . 790
'..O09
n.4:
14. 1C,
7 1 . «G
_\_'i--^7 ^
HillsAw-hr
1 . 7C
. 92
. 26
. 91
. 08
. 59
. 03
. 80
. 53
. 51
. 75
• 69,
. 10
. 7fe
. faO
. 05
. 8
. 79
. 39
. 93
. 83
. 10
. 76
1Z.177
-------
TABLE IV
COST REDUCTIONS OF FREE-DIfCHARGE THERMAL STANDARD (1976 dollars)
Plant & Unit No.
Clay Boswell #3
Clay Boswell #4
Sherburne County #1-2
Sherburne County #3-4
Monticello
Prairie Island
Quad-Cities
Carroll County #1
Louisa #1
Capital Costs
Total
Annual
106$
1.681
6.847
29.02
2.857
mills/kw-hr
0.480
0.611
3.765
0.627
Operating Costs
Total
Annual
106$
3.287
5.907
18.51
20.90
4.077
17.70
18.41
7.893
mills/kw-hr
3.127
1.686
1.860
1.864
1.023
2.248
1.642
1.733
Total Cost Reduction
Annual
106$
3.287
7.588
18.51
27.74
4.077
17.70
18.41
29.02
10.75
mills/kw-hr
3.127
2.166
1.860
2.475
1.023
2.248
1.642
3.765
2.360
TABLE V
ZERO-DISCHARGE COST INCREASES ABOVE COSTS OF EXISTING THERMAL STANDARDS (1976 dollars)
Plant &
Unit No.
Clay Boswell #1-2
Monticello
Elk River #1-3
Riverside #8
High Bridge #5-6
Prairie Island #1-2
Alma ftl-6
Genoa #1A-2D,2,3
Lansing #1-4
Stonemari
Nelson Dewey
Dubuque #2-4
Carroll County #1
M.L. Kapp #1-2
Moline #5-7
Riverside #3,3HS,4,5
Fair #1-2
Muscatine #5-9
Burlington
Capital Costs
Total
Annual
106$
2.060
0.7935
7.032
7.002
9.693
6.808
5.424
0.7727
6.489
1.304
6.398
6.023
1.254
3.901
1.126
4.937
0.726
mills/kw-hr
1.961
2.267
4.19?
3.600
3.008
2.335
2.435
2.131
4.079
2.326
0.830
3.604
2.430
2.490
2.473
2.515
5.873
Operating Costs
Total
Annual
106S
2.160
17.938
0.6743
4.582
4.644
5.153
8.600
6.453
4.846
0.7985
4.214
1.124
16.01
4.107
l.lSb
3.584
1.064
4.640
5. 368
mills/kw-hr
2.054
4.500
1.926
2.731
2.J87
0.655
2.668
2.232
2.176
2.202
2.649
2.004
2.U77
^.457
2.295
2.28B
2.336
2.420
3.613
Total Cost Increase
Annual
106$
4.221
17.94
1.468
11.61
11.65
5.153
18.29
13.26
10.27
1.571
10.70
2.428
22.41
10.13
2.439
7.485
2.190
9.576
14.09
mills/kw-hr
4.015
4.500
4.193
6.923
5.987
0.655
5.676
4. 583
4.611
4.333
6.728
4.330
2.907
6.061
4.725
4.778
4.809
4.995
9.486
TABLE VI
REGIONAL COST COMPARISONS OF DIFFERENT THERMAL STANDARDS
Thermal Standard
Free-Discharge
Existing
No-Discharge
Total Costs
Annual
106$
1180.3
1317.4
1494.3
mills/kw-hr
13.622
15.204
17.246
Incremental "Cost" of
Standard above Free-
Discharge
Annual
106$
118.7
295.6
mills/kw-hr
1.582
3.623
1158
-------
:' V-1
e 2 Unil Nalural Evaporation Along 'he Upper Mississippi Riv
en
<£>
Figure 3 Unit Net EvODOrQlion tor Auqusl Along Ihe Upper Mississippi '
-------
THERMAL PLUME MAPPING
J.R. Jackson and A.P. Verma
Envirosphere Company, Division of Ebasco Services Incorporated
Atlanta, Ga. and New York, N.Y., U.S.A.
ABSTRACT
An accurate description of thermal plume characteristics is fundamental to
the evaluation of plant performance as it relates to technical specifica-
tions and state imposed mixing zone criteria. This paper presents a
generalized approach for mapping thermal plumes with considerations given
to discharges in different types of receiving water bodies, variability of
ambient conditions, and other parameters which must be measured. Rivers,
lakes, estuaries and oceans all present widely varying conditions for which
several alternative methods of sampling and positioning are available.
Also, any single receiving water body type can be sampled in several ways
due to the wide variety that exists in instrumentation and data logging
equipment.
The basic elements of a thermal plume survey can be grouped in three
phases. These phases consist of (1) logistics and planning, (2) execution,
and (3) data reduction and evaluation. The criticality and interrelation-
ship between them are highlighted.
INTRODUCTION
Before the planning phase of any survey can begin, it is necessary to
carefully examine the client's, needs with respect to not only the mea;js and
extent of data acquisition but also the uses to which the data will
ultimately be applied. Quite often the client is not fully aware of his
own needs in terms of the level of effort required to insure the adequacy
of the data and the degree of scrutiny to which the results will be
subjected. There will also be cases in which the client will only supply
a general requirement, leaving the responsibility of detailed planning to
the surveyor. In any event, it is imperative, when one's client must
ultimately deal with the federal and state regulatory process, that the
mediocre or "cost effective" survey does not become adequate after the
fact.
The most critical piece of information needed but not always asked for by
the client is a precise definition of the end product of the survey effort.
For the purpose of this paper, we shall define this initially in terms of
a set of maps or map overlays which show the following data:
1160
-------
- Isotherms, in terras of temperature or temperature variation ( T)
over ambient and/or contours of other simultaneously measured
data such as dissolved oxygen or dye concentration,
- Wind and current vectors,
- Ambient (if definable), intake, discharge and air temperatures,
- Calculated area within critical isotherms (where required),
- Tide or stage level data,
- Time, date and depth(s) of the survey(s).
Certain basic items must also appear on the maps or in the title block
including:
- Shoreline and discharge structure outlines (unnecessary on an
overlay),
- Scale and north arrow,
- Grid reference points and an explanation of the coordinate
system.
Other pieces of information which might be included or discussed separately
in an accompanying report are:
- Positioning system reference station locations or control
points,
- In situ meter and vertical profile station locations,
- Plant operational .data for the time of the survey,
- Vertical profile data and/or receiving water body temperature
cross-sections,
- Plant structure outline,
- Bathymetry,
- Drogue plots,
- A discussion of field and analytical methodology,
- Any subsequent analytical results,
- An evaluation of discharge performance with respect to
mathematical models and/or thermal water quality criteria,
Discussions with the client prior to detailed planning of the survey should
include not only the appearance and content of the finished maps, but also
and of even greater importance the adequacy of the information being
presented. Considerations such as sample density and data redundancy must
be dealt with and agreed upon, primarily due to their obvious impact upon
the survey cost, before detailed planning can .begin.
OPERATIONAL ELEMENTS
The basic elements of a thermal plume survey can be grouped in three
phases. The first and most critical phase consists of logistics and
planning. The success and credibility of the survey will directly depend
upon adequate preparation in terms of equipment and supplies, timing and
coordination with respect to ambient conditions and plant operating
schedules, as well as pre-plotting of tracklines and profile locations,
arranging for accurately surveyed horizontal control and, if necessary,
site reconnaissance. The need for redundancy in data collection for
certain parameters is also an important planning consideration,
particularly when operational constraints are imposed by economic factors.
1161
-------
With a reasonable effort during the planning phase, the next phase,
execution of the actual survey, becomes reasonably straightforward with the
exception of unavoidable scheduling difficulties which often arise as a
result of weather or changes in plant operation. With the more
sophisticated equipment now available, much of the data reduction and
evaluation, the third and final phase of the survey, can be accomplished
in the field. This is not always economically feasible, however, when
surveys are of limited duration and scope.
DATA REQUIREMENTS AND PLANNING
Preliminary Background Investigation
The complete thermal survey should include the definition of effects of
natural and man-made variability of environmental conditions on plume
characteristics. Consequently, thermal mapping is not confined to the
measurement and mapping of temperature alone. Other important parameters
include ambient current, tides, water mass distribution, bathymetry, wind
conditions for heat transfer considerations and accurate locations for the
discharge and any other structure or naturally occurring object which may
affect the plume's shape. Additional considerations include horizontal and
vertical ambient temperature fluctuations, variations in plant heat output,
interference from other heat sources, extrapolation between normal and
extreme thermal conditions and definition of ambient conditions with
respect to regulatory agency requirements. Many of these parameters can
be anticipated and evaluated for their relative importance prior to the
field work. The most immediate source of site specific information is
the plant operator who may be able to make available the results of
previous or ongoing data collection programs. Available parameters might
include intake and discharge temperatures, stage or tide, and meteorologi-
cal conditions. In addition, the operator should provide horizontal
control (p"!ant or state grid) information, charts or plans of the plant and
discharge areas, any hydrological information obtained through studies
conducted during pre-operational phases, a plant operating schedule as it
effects the operation of the circulating water system and/or blowdown flow,
the heat rejection rate, thermal criteria and technical specifications to
which the effluent is subject and an understanding of any local political
sensitivities which might affect the way survey operations are to be
conducted. Other potential sources of information include but are not
necessarily restricted to:
- U.S. Geological Survey,
- Corps of Engineers,
- National Oceanic and Atmospheric Administration,
- State and local agencies including water management and
irrigation boards,
- Privately-owned reservoir managements,
- Universities and private research institutions.
1162
-------
Adequate attention given to background data collection will make possible
a proper definition of the scope and duration of the study and the
conditions leading to the definition of critical situations.
Logistics and Support Arrangements
Most problems connected with field programs can be prevented or alleviated
by applying anticipation, communication, and money. Furthermore, t
weakness in any one factor, usually the latter, will proportionately
increase a need for the others. The following is, therefore, an attempt to
anticipate at least the basic field planning steps necessary in advance of
survey execution. While some may seem obvious, they are listed for the
sake of completeness. In that any two surveys may have as many differences
as similarities, greater detail in the area of general arrangements is
beyond the scope of this paper.
- Survey Boat selection:
While the selection and layout of the boat to be used is often a
matter of institutional or personal preference (not always that
of the surveyor), a few basic points should be mentioned. The
size of the vessel should be large enough to provide adequate
protection as well as mounting and working space for the
instrumentation. In addition, space should be available for
working over the side along with, if possible, some light
lifting capability in the form of a davit for lowering and
raising line depressors or profiling instruments. The boat
should also be small enough to have a high degree of
maneuverability for work in and around the plume but not so
small as to make it overly sensitive to weather and sea
conditions. The draft should be shallow and wake small to
minimize surface water (plume) displacement and, of course,
appropriate safety and emergency equipment, including extra
protection for the instruments, should be on board at all tines.
If diving is involved with, for example, the placement of in
situ recorders or inspection of the discharge structure,
facilities should be provided according to Occupational Health
and Safety Administration Requirements as set forth in the
Federal Register, July 22, 1977. Finally, berthing, launching,
fueling and insurance should also be prearranged.
— Preparation, shipping and calibration of equipment:
A thermal survey can involve a wide variety of types of
instrumentation including in situ and onboard monitoring
devices, vertical profiling and towed sensors, electronic
positioning systems and a variety of recording equipment (strip
chart, magnetic tape, film and x-y plot). Additional support
equipment might consist of communications and navigation
equipment, diving and mooring gear plus testing and calibration
instruments. Since each individual piece of equipment may have
its own distinct preparatory requirements, only general rules
1163
-------
can be stated to cover this phase of the effort. Most
importantly, any and all setup, maintenance and calibration
procedures and logs should be pre-defined, recoverable and
meticulously documented, even in those areas where the client
has not imposed specific quality assurance guidelines.
Where possible, pre-operational checks and calibrations should
be performed after shipping and as close as possible to the
time of deployment. While these procedures are best left to
the responsible technician, it is imperative that the principal
investigator be familiar with and able to defend the selection
and preparation of the equipment should questions arise, as they
often do, concerning the credibility of his or her methods and
data. One more consideration in the preparation of equipment
is, simply, how much to use. Whenever possible backup equipment
for onboard systems should be available along with redundant
data collection by in situ instruments. This must of course,
be weighed against limitations of time, space and budget.
Finally, a common mistake in the shipping of equipment is
inadequate insurance coverage. The automatic coverage generally
provided by airlines and other carriers is minimal. V"hile
additional coverage is often expensive, the high replacement
costs of much of the equipment involved generally justifies the
expense.
- Local support and purchase arrangements:
In addition to the arrangement for plant data which is
concurrent with the many operations, several mundane but ever-
present problems must be addressed. These include lodging,
security for the boat, positioning equipment and in situ
instruments, local availability of moving materials, marine
supplies and rental equipment and land access permission, if
necessary, for the location of shore stations. Finally,
arrangements must be made for land surveyors to establish the
positioning system reference points.
POSITIONING ALTERNATIVES AND REQUIREMENTS
Several methods exist for the determination of a moving boat's position
while crossing and recording the temperature distribution of the thermal
plume. The selection of any one necessitates the weighing of the need
for accuracy and data density against cost and level of sophistication of
equipment and personnel. It is not always less expensive to resort to
the simpler visual (versus electronic) positioning methods when one
considers the often greater time requirements for setup and data reduction
and the extended use of land survey parties. These methods might include:
- Shore based surveyors (two or more with radio communication)
continuously turning angles on the boat's position as it moves
and recording fix locations which must be later calculated
individually,
- Recording horizontal angles between shore landmarks (at least
three) from the boat for each fix location,
1164
-------
- Steering along predetermined and-surveyor fixed transects by
means of buoys or aligned range stakes on the shore.
All of these methods share the disadvantages discussed above. Electronic
positioning, however, provides the capability of instantaneous position
determinations as little as one-half second apart which, when coordinated
with continuously measuring temperature or other sensors, can provide
enormous amounts of information over large areas not necessarily
constrained by visibility limitation. Unfortunately, this costly piece of
equipment cannot be utilized to its fullest potential without a reasonably
sophisticated digitizing and recording system which is capable of
assimilating all of the parameters being measured, properly sequencing
and tagging them with times, and recording them in analog or preferably
digital format which can be later recovered by a computer and in hand copy.
At this point one must also consider the use of an onboard processor which,
in addition to its ability to key, organize and feed the data to a tape
recorder, can also instantanteously process incoming position information,
converting it to a simultaneous track plot by which the boat operator can
steer. This enables the surveyor to preplot the survey tracklines and
simply over-print these lines on an x-y plotter during the actual survey.
The obvious advantages are the completeness of coverage made possible by
close, regularly spaced tracklines without overlap, repeatability and
increased ease of survey operation. Likewise, data reduction time can be
greatly reduced as a result of the system's computer compatibility. There
are several further variations and refinements to this system but all
produce the same end result.
SURVEY TIMING AND AMBIENT CONDITION VARIABILITY
The inherent differences between rivers, lakes, estuaries and oceans with
their varying levels of complexity determine the timing of a thermal
survey and the number of surveys required to sufficiently define the plume.
In general, the timing should reflect the periodic fluctuations of plume
characteristics, from seasonal to tidal, in terms of conditions surrounding
the critical case(s). Aperiodically changing conditions such as storm
effects are more difficult by far to plan around and can only be marginally
predicted on a seasonal basis.
Some of the more important variables to consider include:
- Periodic and aperiodic changes in direction of flow,
- Velocity, magnitude of flow and dispersive characteristics,
- Degree of natural stratification,
- Presence of vertical water mass boundaries,
- Potential heat accumulation or ponding areas which occur only
under certain conditions,
- Conditions of maximum thermal impact,
- Occurrences of heat input from sources other than the plant,
- Relative location of ecologically sensitive areas,
1165
-------
- Minimum levels of wave action and atmospheric heat transfer,
- Worst case basin configuration in the receiving water body,
- Plant operating conditions producing maximum temperature
elevation.
FINAL CONSIDERATIONS
Once on board the survey boat, it is generally too late to significantly
modify the program to account for oversights. However, there are certain
questions that can be raised during initial survey activities which may
expedite a successful completion.
- Are the instrument preparations and calibrations complete and
traceable?
- Are in situ instruments and profiling stations adequate in terms
of location and density to properly define the system?
- Have variations in plant output been accounted for?
- Is there sufficient definition of the wind and air temperature
conditions over the actual plume?
- Is there interference from other heat sources present and, if
so, can it be discriminated from the plume under study?
- If the discharge is subsurface, has it been accurately located
in terms of the survey positioning system?
- Is a detailed bathymetric survey available or necessary?
- Is the minimum temperature elevation to be measured within the
range of horizontal ambient temperature variability?
- Are the depth settings of the temperature sensors such that
they will skip in and out of a thin surface plume layer?
- Is the deepest sensor consistently below the far-field plume or
are vertical temperature profiles along the longitudinal plume
axis necessary?
- Can the boil area location and migration from a subsurface
discharge be accurately determined?
- Have the short term periodic ambient temperature fluctuations
been adequately defined?
- What are the thermal characteristics relative to both the normal
and extremes and can they be extrapolated with respect to
ambient conditions and plant output?
1166
-------
THERMAL SURVEYS NEW HAVEN HARBOR
SUMMER AND FALL, 1976
W. Owen
College of Marine Science, University of Delaware
Lewes, Delaware, U.S.A.
J. Monk
Normandeau Associates, Inc.
Bedford, New Hampshire U.S.A.
ABSTRACT
Thermal surveys conducted in New Haven Harbor, Connecticut during July,
August and October 1976 were designed to define the thermal plume of the
New Haven Harbor Station as required by the National Pollution Discharge
Elimination System (NPDES) Permit to Discharge. Since New Haven Harbor
has a complex temperature structure due to both natural and man-made
sources of heat, Rhodamine WT dye was used in conjunction with a three
dimensional temperature sampling program to distinguish the thermal load
introduced by the New Haven Harbor Station from other natural and man-
made thermal influences. The results of a dye and thermal study con-
ducted in October were used to interpret the data from the July and
August thermal surveys. This report includes a presentation and analy-
sis of the assumptions upon which the dye study design, calculations and
projections were made.
INTRODUCTION
Dyes and more specifically Rhodamine dye have been successful as tracers
in studies of transport, dispersal and dilution patterns of solids or
liquids subjected to the naturally occurring forces of a water system.
Since 1960 this technique has been adapted to profiling the movement of
effluent discharges in receiving waters. In the present study the dye
was used as a tracer of heat input to New Haven Harbor resulting from
operation of the New Haven Harbor Station, a 460 MW oil fired power
plant. *
Given the temperature increase across the condensers (At) and the cool-
ing water flow, dye concentration can be related to temperature and it
is possible to calculate At1 per part per billion of measured dye
concentration. Dye concentration was measured in the field using
standard, continuous sampling fluorometric techniques in conjunction
with temperature measurements. The dye distribution was converted to a
'•At is used here to describe the elevated temperature due to the cooling
water discharge from the New Haven Station only. At is a function of
position and time.
116/
-------
temperature distribution indicating the At independent of other thermal
sources. This method is based on the assumption that the temporal and
spatial distribution of dye and temperature will be the same. For this
assumption to hold, the following conditions must be met: (1) the
density of the receiving water must not be affected by the tracer mater-
ial, (2) the power plant must be operating under normal conditions to
ensure a representative density difference between the thermal discharge
and the receiving water, and (3) the dilution of the thermal discharge
water must occur rapidly enough to ensure that the effects of cooling to
the atmosphere can be neglected.
The first condition (1) was satisfied by the method of discharge of dye
The proof of the validity of conditions (2) and (3) lies in the base
temperature computations which serve as a check on the correspondence of
dye and temperature. Base temperature is defined as the temperature the
water would have been if it were thermally unaffected by the New Haven
Harbor Station but still affected by natural thermal sources and man-
made sources of heat other than the power plant in question. The base
temperatures were determined by subtracting the At computed from dye
concentration from the actual temperatures measured in the Harbor (If
the dye concentration and the temperatures were not similarly distribu-
ted, the assumptions would be invalid to the extent that the base tem-
perature was perturbed). Conditions (2) and (3) can only be checked by
the base temperature's agreement with temperature distribution expected
in the body of water in question at the time of the survey. If there
was a region of anomalously low temperatures observed in the computed
base temperature distribution, the implication is that the plant has not
run long enough to create a quasi-steady state in temperature. Thus
there was too much dye for the given temperature. The opposite effect
would be caused by interruption in the impact of dye.
Experimental Procedure
Instrumentation
Instrumentation and material used in the study included Turner Designs
Model 10-000 full flow fluorometer and an NAI Model 3100-TD temperature
profiling system (BT), and Rhodamine WT dye, 20% aqueous solution. The
dye was injected into the plant cooling water just downstream of its in-
take. Other equipment and materials used were all standard off the
shelf items commonly employed in this sort of work.
1168
-------
Measurements at Cooling Water Intake
During the October dye study, water was pumped continuously from in front
of the New Haven Harbor Station intake through the fluorometer set up in
the pump house. A Rustrak 1332 temperature probe was in the line down-
stream of the fluorometer. Fluorescence and temperature were recorded on
a 3 channel Rustrak recorder. This set up is shown in Figure 1.
Shipboard Measurements
During the thermal surveys, vertical temperature profiles were obtained
at a series of stations whose positions were established using a mini-
ranger (August) or sextant as a pelorus (July). The stations were occu-
pied at each critical phase of the tide on two successive days each month.
While the vessel was on station, the BT submarine unit was lowered from
the surface to the bottom using the hand winch; the vertical temperature
profile was recorded on the x-y plotter set in its temperature-depth mode.
Shipboard measurements during the October dye/temperature surveys were of
two types:
a. Horizontal continuous sampling of surface dye concentration and
temperature.
b. Vertical profiles of.dye concentration and temperature.
Both types of measurements were made during each of the four critical tide
phases. The equipment set up for both types of measurement is shown in
Figure 2. Water was pumped continuously through a hose to the deck where
it passed through the shipboard fluorometer and a housing containing a
Rustrak 1332 temperature probe. The hose was clamped to the submarine
unit of the BT system, which was also used during shipboard measurements.
Outputs from the fluorometer and the temperature probe were recorded on
a Rustrak recorder. BT system output was recorded on a Houston Instru-
ments x-y recorder. The purpose of the Rustrak temperature system was to
correct ^he fluorometer. The purpose of the Rosemount temperature system
was to provide an accurate temperature measurement of the water at the
hose intake for the fluorometer.
To obtain continuous horizontal data, the research vessel traversed a
series of pre-selected transects (Figure 3). Start, end and intermediate
positions on all transects were established using the mini-ranger navi-
gation system. While the vessel was underway, water was pumped through
the instruments and data were recorded as described above. The x-y
plotter was set in its time sweep mode so the BT system served as a con-
tinuous temperature monitor.
1169
-------
For the vertical profiles (figure 3) the research vessel occupied a
series of stations. The positions of these stations were established
using the mini-ranger. While the vessel was on station, water was
pumped through the instruments/ fluorescence data were read from the
fluorometer and recorded by hand and temperature was recorded on the x-y
plotter. The plotter was set in its temperature-depth mode so that
temperature was recorded on the y-axis and depth was recorded on the x-
axis.
RESULTS
October Dye and Thermal Survey
Figures 4a through lla are maps of surface At prepared from dye concen-
tration measurements made on October 13, 14 and 15, 1976. Included are
two sets of data obtained for each phase of the tide. The attendant
maps of actual measured temperature are contained in Appendix B. In
these and all other figures the caption block includes wind speed and
direction and power generated by the station. Figures 4b through lib,
which accompany the At maps, indicate the respective surface base tem-
peratures. The purpose of the base temperature determination is to
assess the validity of the assumptions made in using the method and to
look for problems which may have occurred during the survey. The com-
puted base temperature distributions agreed well with what was expected.
Generally they indicated the weak surface temperature gradients charac-
teristic of autumn with the following exception: on October 14 the dye
pump failed from 0452 to 0652 EST due to the destruction of the foam
cushion around the motor. The base temperature for the ensuing low
water slack measurements clearly showed a warm spot (insufficient dye).
The At map for the corresponding tide was not used in this report. It
was replaced by data from low water slack measurements made on October
15, 1976. This exception and the normal base temperatures determination
previously discussed all serve as verification of the assumptions made
above.
Figure 4 is the At map determined from field measurements during low
water slack on October 13, 1976. The wind was from the southwest at 13
knots. The plume, as would be expected, was in the immediate vicinity
of the discharge, but there was also evidence of warmer water on the
west side of the harbor. Results of the second low water slack survey
(October 15) are presented in Figure 5a. The wind (southwest, 14 knots)
was nearly the same as during the earlier low water slack survey, but
the plume was larger. This is due at least partially to the higher
power output of the plant on October 15 (463 MWH, 3 hr average) compared
to output on October 13 (446 MWH).
1170
-------
The At maps developed from flood tide surveys are shown in Figure 6a
(October 13) and 7a (.October 14) . On October 13 the wind was 15 knots
from the south-southwest, and, as such, it had a sizeable component in
the same direction as the flood tide. Wind and tide combined to create
the narrow, elongated plume seen streaming to the north in Figure 6a. On
October 14, the wind was from the west-northwest at 19 knots. This
opposed the flood tide and the resultant plume was smaller than it had
been on the day before even though the average power output of the plant
was higher (463 MWH on October 14 and 447 MWH on October 13). Also in
Figure 7a there is a small patch of water on the west side of the harbor
characterized by a At of 2F. This may be a remnant of the ebb tide plume
which broke away from the main plume as a result of the turning tide and
the vigorous wind.
The At distributions for high water slack are illustrated in Figure 8a
(October 13) and 9a (October 14). On October 13 the wind during high
water slack was nearly the same as it had been during the previous flood
tide, and the elongated character of the plume was still evident, al-
though it covered a smaller area. On October 14 the wind had diminished
somewhat from mid ebb to high water slack when it was 14 knots from the
west-northwest. The plume had broadened considerably compared to its
shape during flood, and, because of the wind, it extended further south
than the high water slack plume on the day before.
There was a substantial difference in the relative extents of the ebb-
tide plume on October 13 and 14. On October 13 (Figure lOa) the plume
swumg far down stream despite the continued brisk wind from the south-
southwest. In contrast, the ebb-tide plume on the next day (Figure lla)
was very small. The reason for the large difference was the sharp re-
duction in power on October 14, when the average plant output was only
288 MWH during the ebb-tide survey. Plant output during the ebb survey
on October 13 was 420 MWH. If the plant had been operating at full power
on October 14, the plume would have extended far downstream (i.e., like
Figure lOa), and it would have been narrower than the October 13 ebb
plume because of the 18 knot west wind that blew during the ebb survey
on October 14.
It is clear that the position of the plume varied greatly. This varia-
bility resulted from two effects: the wind's direct effect on the
harbor and the wind's effect on the ocean and Long Island Sound which in
turn directly affect the tide in the harbor.
Figure 12 is an example of the cross-sectional At observations in the
harbor. Higher subsurface temperatures were found only in the immediate
vicinity of the discharge (Transect F). The lower temperature water of
the plume did spread further but did not extend across the harbor. The
percentages of the cross-sectional area affected by the plume are dis-
cussed below. The At's in the cross sections were determined by the
same method as was used for the surface maps.
1171
-------
July Temperature Survey
During the July survey the power plant was slow to come on line each
morning. On July 21, it was operating at about 31% power early in the
day, and on July 22 it was operating at 60% power early in the day.
These times correspond to high water slack surveys, which were conducted
from 0540 to 0741 EST on July 21 and from 0640 to 0826 on July 22. For
the remainder of each day the plant operated at between 80% and 83%
power. Nevertheless, there was generally no readily recognizable plume
on the surface and only slight indications of a subsurface plume.
Since there was no clearly defined plume, the techniques used for compu-
ting base temperatures in Section 3.1 cannot be applied here. Instead,
base temperatures for the July survey were determined by analyzing the
plant records of intake temperature, correcting them for recirculation
effects and by examining the far-field temperatures measured while the
plant was coming on line. These analyses disclosed surface base temper-
atures which generally were in the range of 69-70F and subsurface base
temperatures which ranged between 67F and 69F. The results of the dye
study suggested that with the power plant operating at 100% the highest
maximum At was about 5F. A At of 5F would have produced a maximum
surface temperature in the harbor of 75F during the July survey. The
subsurface maximum would have been between 72F and 74F, except in the
immediate vicinity of the discharge where it would probably be higher.
August Temperature Survey
The power plant was not operating during the August survey; therefore
the temperatures measured were base temperatures. To determine the
nature of the plume that would have existed had New Haven Station been
operating, results of the October dye study were applied to the August
temperature data. The principal assumption was that if dye had been
pumped during the August survey then the distribution of dye (and At)
in the harbor would have been the same as it was in October if the winds
were the same. In reality, the winds were not the same, but the differ-
ences were taken into account in hindcasting the positions of the August
plume at the various phases. No adjustments were made in the plume
projections to allow for natural cooling of the discharge water during a
given tidal phase, so actual areal extent of the plume would likely be
smaller than that depicted.
Recirculation Effects
Recirculation is defined here as the ratio of the dye concentration
measured at the intake structure, well upstream of the dye injection
point, to the dye concentration at the discharge (expressed as a per-
centage) . Recirculation is a function of the tidal current and the wind;
tidal current is the dominant factor. Figure 13 displays recirculation
plotted against time on October 13, 14 and 15. The range of the recir-
culation was from 0 to 12.9% which is equivalent to a temperature range
1172
-------
of OF to 1.9F. Because of the upstream location of the intake, recir-
culation is usually highest neat the time of high tide. Variations in
recirculation during the same tide phases on different days are prima-
rily the result of wind effects.
DISCUSSION
Characteristic maximum discharge plume surface temperatures observed or
predicted during the three months included in the survey were as follows:
July — 75F; August — 80F; October — 64F. The maxima for the indivi-
dual tide phases generally ranged no more than 1 or 2F from the average
for each study. Table 1 lists the observed maximum surface temperature
for each of the surveys. The maximum temperatures for the July survey
are all 74-75F and were obtained by adding a At of 5F to an assumed
maximum base temperature of 69-70F. Maxima for the August survey were
determined by the superposition of the At's calculated from the October
dye study onto the August measurements; maximum predicted temperature
was 82F during low water slack on August 25. Maximum temperatures for
the October survey were determined by actual measurements; the maximum
observed temperature of 65F occurred during high water slack on October
14.
Generally, the percentage of the surface area of the inner harbor sub-
jected to a temperature rise of 1 to 4F was small. This is indicated in
Table 2. which is a compilation of the percentage of the surface area of
the inner harbor bounded by At's of 1, 2, 3 and 4F. In each case the
area bounded by a At of 4F was equal to or less than one tenth of one
percent of the total inner harbor area. The areas bounded by At's of
3F ranged from zero to 0.4% in October and from 0.1% to 0.5% in August.
Much more area was bounded by the 2F At isotherm; the percentage affec-
ted ranged from less than 0.1% to 3% in October and from 0.5% to 5.5% in
August. The percentage of the surface area bounded by a At of IF ranged
between 0.7% and 12.6% in October and between 2.7% and 11.2% in August.
In discussing the cross-sectional area affected by temperature rises of
1 to 4F, attention will be limited to Transect F. This transect corres-
ponds to the position of the New Haven Harbor Station discharge; there-
fore, it shows far more subsurface plume effects than any other. Table
3 lists the percentage of cross-sectional area of Transect F bounded by
At's of 1, 1.5, 2, 3 and 4F at the various phases of the tide. The
total cross-sectional area of the harbor at Transect F is, of course, a
function of tide height, so average values were determined for each tide
phase during both the October and August surveys. Specifically, this was
done by drawing cross sections of the harbor at Transect F for each tide
phase and determining the areas by planimetry. Predicted tides and tidal
heights obtained from National Ocean survey tide tables are listed in
Appendix D.
In October the percentage bounded by 4F was zero during four of the eight
tide phases studied, and it ranged to a maximum of 2.9% during low water
1173
-------
slack on October 13. For August the range was between zero and 3.2%.
The 3F At contour bounded up to 6.7% and 8.5% of Transect F in October
and August, respectively. The extent of the 2F At contour was highly
variable in October but somewhat less variable in August. The area
bounded by the 2F isotherm ranged from 1.9% to 22.8% in October and from
10.4% to 20.6% in August. The areal extent of the IF At was also quite
variable in October, ranging from 7.4% to 40.2%. Again, August showed
less variability; the range then was from 22.3% to 41.6%.
The percent of cross-sectional area bounded by the 1.5°F contour of At
was also included in Table 3 as Connecticut water quality regulations
use this temperature in defining a "mixing zone". The percentages of
Transect F bounded by the 1.5F At ranged from 3% to 27.6% in October
with an average of 17.4%. In August the range was from 15.8% to 30%
with an average of 22.6%.
Generally, the data reported herein have shown surface and cross-sec-
tional contours of At in which the 3F and 4F contours encompassed a re-
latively small area. Since power plant design specifies a At of about
3.75F where the discharge jet impinges on the surface (and near field is
the immediate vicinity of the discharge jet), it is reasonable to re-
present the near field as the 3F and 4F contours of At. Therefore, the
1.5F At contour is included in the far field, and it will probably not
be affected by any change in plant operation as long as the heat input
to the harbor does not change.
LITERATURE CITED
Fan, L.N. Numerical solutions of turbulent buoyant jet problems.
W.M. Keck Laboratory of Water Resources and Hydraulics Report
No. KH-R-15. California Institute of Technology. June 1967.
Pritchard, D.W. and H.H. Carter, 1965. On the prediction of the
distribution of excess temperature from a heated discharge in an
estuary. Chesapeake Bay Institute. Technical Report N
Res. 65-1. 45 p.
1174
-------
TABLE 1. MAXIMUM SURFACE TEMPERATURES. THERMAL SURVEYS NEW HAVEN
HARBOR, SUMMER AND CALL, 1976.
DATE
July 21, 1976
July 22, 1976
August 24, 1976
August 25, 1976
October 13, 1976
October 14, 1976
October 15, 1976
TIDE PHASE1
HWS
ME
LWS
MF
HWS
ME
LWS
MF
MF
HWS
ME
LWS
MF
HWS
ME
LWS
LWS
MF
HWS
ME
ME
MF
HWS
LWS
MAXIMUM
SURFACE TEMPERATURE
(F)
74-75
74-75
78
80
80
81
79
80
31
82
63
64
64
64
64
64
65
63
HWS = High water slack
ME = Mid ebb
LWS = Low water slack
MF = Mid flood
Temperatures are estimated.
See Section 4.2 for discussion.
1175
-------
TABLE 2. PERCENTAGE OF THE SURFACE AREA OF THE INNER HARBOR BOUNDED
BY at's of 4F OR LESS. THERMAL SURVEYS,NEW HAVEN HARBOR,
SUMMER AND FALL, 1976.
DATE
8/24/76
TIDE PHASE
8/25/76
10/13/76
10/14/76
10/15/76
4F
3F
2F
IF
MF
HWS
ME
LWS
MF
HWS
ME
LWS
LWS
MF
HWS
ME
ME
MF
HWS
LWS
<0.1 0.1
<0.1 0.3
<0.1 0.4
<0.1 0.3
<0.1 0.2
<0.1 0.1
0.1 0.5
<0ul 0.2
0.1 0.4
0.0 <0.1
0.0 0.0
<0.1 0.4
<0.1 <0.1
<0.1 0.1
<0.1 0.1
<0.1 0.1
0.8
0.8
5.5
0 7
1.1
0.5
2.3
0.8
0.6
0.3
<0.1
3.0
0.2
1.0
0.6
0.5
2.8
3.4
11.2
4.7
4.7
2.7
6.5
5.4
1.0
2.4
2.2
12.6
0.7
1.2
3.7
10.1
Predictions based on dye study
1176
-------
TABLE 3. PERCENTAGES OF THE CROSS-SECTIONAL AREA OF TRANSECT F BOUNDED
BY At's OF IF, 1.5F, 2F, 3F AND 4F. THERMAL SURVEYS,NEW HAVEN
HARBOR, SUMMER AND FALL, 1976.
DATE
8/24/76
8/25/76
10/13/76
TIDE
PHASE
MF
HWS
ME
LWS
MF
HWS
ME
LWS
LWS
MF
HWS
ME
ME
MF
HWS
4F
3.2
2.8
0.0
1.6
1.4
2.1
2.1
1.1
2.9
0.0
0.6
0.0
0.0
0.7
1.1
3F
8.5
8.5
2.4
4.8
5.5
5.6
5.6
3.8
6.7
0.5
2.0
1.6
0.0
3.4
5.9
2F
17.0
20.6
12.8
11.3
15.7
12.4
10.4
12.5
17.4
1.9
21.3
3.9
22.8
9.6
11.7
1.5F
25.2
30.0
22.2
22.1
23.5
19.7
15.8
22.7
20.8
3.0
26.3
5.4
27.6
16.3
14.9
IF
34.1
41.6
35.8
36.5
34.1
28.8
22.3
36.5
24.4
15.1
30.5
7.4
32.7
24.9
20.3
10/14/76
10/15/76 LWS 0.0 0.0 3.1 24.7 40.2
1177
-------
PUMPHOUSE WALL
RUSTRAK TEMPERATURE PROBE
Figure 1. Illustration of instrumentation set up in intake pump
house. Thermal Surveys, New Haven Harbor, Summer and
Fall, 1976.
1178
-------
RUSTRAK TEMPERATURE PROBE
FLUOROMETER
PUMP
TO XY RECORDER
BT SUBMARINE UNIT
Figure 2. Illustration of instrumentation set up aboard the R/V
Gale. Thermal Surveys, New Haven Harbor, Summer and
Fall, 1976.
1179
-------
NEW HAVEN HARBOR STATION
DATE-
TIDE:
tlME(EST)
PLANT OUTPUT(MWH)
WIND FROM: AT KNS
LONG
WHARF
BOULEVARD
S.T.P.
INTAKE
o
NEW HAVEN
HARBOR
STATION
EAST
SHORE
PS.T.P.
NAUTICAL MILES
.1 .2 .3 .4
KILOMETERS
FORT HALE
PARK
SANDY
POINT
Figure 3. Transect locator map. Thermal Surveys, New Haven Harbor,
Summer and Fall, 1976.
1180
-------
NEW HAVEN HARBOR STATION
NEW HAVEN HARBOR STATION
DATE: 10-13-76
TIDE: LOW WATER SLACK
TIME(EST) 08I5-IOO5
PLANT OUTPUT(MWH) 446
WIND FROM:SW AT 13 KNS
AT°F
DATE: 10-13-76
TIDE: LOW WATER SLACK
TIME(EST) 0815- IO05
PLANT OUTPUT (MWH) 446
WIND FROM :SW AT 14 KNS.
BASE T°F
BOULEVARD
S.T.R "
BOULEVARD
S.T.P. •
INTAKE
0
NEW HAVEN
EAST
SHORE
»S.T.P.
EAST
SHORE
•s.T.P
NAUTICAL MILES
.1 .2 .3 .4
FORT HALE
PARK
FORT HALE
PARK
Figure 4. Surface At, (a) and surface base temperature (b), October
13, 1976 - Low Water Slack. Thermal Surveys, New Haven
Harbor, Summer and Fall, 1976.
-------
NEW HAVEN HARBOR STATION
NEW HAVEN HARBOR STATION
DATE: IO-I5-T6
TIDE: LOW WATER SLACK
TIME(EST) 0949-1139
PLANT OUTPUT (MWH) 463
WIND FROM: SW AT 14 KN3.
h BASE T*F
DATE: 10-15-76
TIDE: LOW WATER SLACK
TIME(EST)0949-
PLANT OUTPUT(MWH) 463
WIND FROM: SW AT 14 KNS.
3 AT'F
BOULEVARD
S.T.R
BOULEVARD
S.T.R
INTAKE
O
NEW HAVEN
HARBOR
STATION
INTAKE
O
NEW HAVEN
HARBOR
STATION
NAUTICAL MILES
3 .1 .3 .4
NAUTICAL MILES
.1 .2 .3 .4
.4 'l
KILOMETERS
FORT HALE
PARK
FORT HALE
PARK
Figure 5. Surface At, (a) and surface base temperature (b), October
15, 1976 - Low Water Slack. Thermal Surveys, New Haven
Harbor, Summer and Fall, 1976.
-------
00
NEW HAVEN HARBOR STATION
DATE: 10-13-76
TIDE: MAX.FLOOD
TIME(EST) 1047-1257
PLANT OUTPUT (MWH) 447
WIND FROM: SSW AT 15 KNS
AT F
BOULEVARD
S.T.P. •
INTAKE
Q
NEW HAVEN
HARBOR
STATION
NAUTICAL MILES
.1 .Z .3
FORT HALE
PARK
NEW HAVEN HARBOR STATION
DATE: IO-I3-76
TIDE: MAX. FLOOD
TIME(EST) 1047-1257
PLANT OUTPUT(MWH) 447
WIND FROM: ESE AT 15 KNS.
BASE TrF
INTAKE
0
NEW HAVEN
HARBOR
STATION
Figure 6. Surface At, (a) and surface base temperature (b), October
13, 1976 - Mid-Flood. Thermal Surveys, New Haven Harbor,
Summer and Fall, 1976.
-------
oo
-Pi.
NEW HAVEN HARBOR STATION
NEW HAVEN HARBOR STATION
DATE: 10-14-76
TIDE: MAX. FLOOD
TIME(EST) 1144-1342
PLANT OUTPUT (MWH> 463
WIND FROM:WNW AT 19 KNS
a
DATE: IO-I4-76
TIDE: MAX.FLOOO
TIME(EST) 1144-1342
PLANT OUTPUT(MWH) 463
WIND FROM:WNWAT 19 KNS.
BASE T"F
BOULEVARD
S.T.P.
BOULEVARD
S.T.R •
INTAKE
O
NEW HAVEN
INTAKE
O
NEW HAVEN
HARBOR
STATION
NAUTICAL MILES
.1 .2 .3 .4
NAUTICAL MILES
.1 .2 .3 .4
FORT HALE
PARK
FORT HALE
PARK
Figure 7. Surface At, (a) and surface base temperature (b), October
14, 1976 - Mid-Flood. Thermal Surveys, New Haven Harbor,
Summer and Fall, 1976.
-------
CO
tn
NEW HAVEN HARBOR STATION
NEW HAVEN HARBOR STATION
DATE: 10-13-76
TIDE: HIGH WATER SLACK
TIME(EST) 1354-1545
PLANT OUTPUT (MWH) 451
WIND FROMtSSW AT 16 KNS.
3 AT'F
DATE: 10-13-76
TIDE: HIGH WATER SLACK
TIME(EST) 1354-1545
PLANT OUTPUT (MWH) 451
WIND FROM.SSW AT 16 KNS.
D BASE T°F
BOULEVARD
S.T.P. •
BOULEVARD
S.T.R
INTAKE
O
NEW HAVEN
HARBOR
STATION
INTAKE
O
NEW HAVEN
HARBOR
STATION
NAUTICAL MILES
.1 .2 .1 .4
NAUTICAL MILES
.1 .Z .3 A
FORT HALE
PARK
FORT HALE
PARK
Figure 8. Surface At, (a) and surface base temperature (b), October
13, 1976 - High-Water Slack. Thermal Surveys, New Haven
Harbor, Summer and Fall, 1976.
-------
NEW HAVEN HARBOR STATION
DATE: 10-14-76
TIDE.- HIGH WATER SLACK
TIME(EST) 1301-1633
PLANT OUTPUT (MWH) 463
WIND FROM:WNWAT 14 KNS.
a AT F
r
INTAKE
a
NEW HAVEN
r-v HARBOR
STATION
NEW HAVEN HARBOR STATION
DATE: 10-14-76
TIDE: HIGH WATER SLACK
TIME(EST) 1501-1633
PLANT OUTPUT(MWH) 463
WIND FROM :WNW AT 14 KNS.
b BASE T°F
BOULEVARD
S.T.R •
INTAKE
a
NEW HAVEN
NAUTICAL MILES
.1 .2 .3 A
FORT HALE
PARK
Figure 9. Surface At, (a) and surface base temperature (b), October
14, 1976 - High-Water Slack. Thermal Surveys, New Haven
Harbor, Summer and Fall, 1976.
-------
NEW HAVEN HARBOR STATION
DATE: IO-I3-76
TIDE: MAX.EBB
TIME(EST) 1717-1910
PLANT OUTPUT(MWH) 420
WIND FROM: SSW AT 16 KNS.
9 AT'F
DATE: IO-I3-76
TIDE: MAX.EBB
TIME(EST) 1717-1910
PLANT OUTPUT (MWH) 4ZO
WIND FROM:SSW AT 16 KNS.
h BASE T"F
INTAKE
a
NEW HAVEN
HARBOR
STATION
INTAKE
a
NEW HAVEN
HARBOR
STATION
EAST
SHORE
•S.T.R
EAST
SHORE
•S.T.P.
KILOMETERS
/''SANDY
// POINT
FORT HALC
PARK
Figure 10. Surface At, (a) and surface base temperature (b), October
13, 1976 - Mid-Ebb. Thermal Surveys, New Haven Harbor,
Summer and Fall, 1976.
-------
NEW HAVEN HARBOR STATION
DATE: 10-14-76
TIDE: MAX.EBB
TIME(EST) 0555-0743
PLANT OUTPUT(MWH) 288
WIND FROM: W AT IB KNS.
a ATT
r
00
00
INTAKI
O
NEW HAVEN
r-t_ HARBOR
STATION
NEW HAVEN HARBOR STATION
DATE: 10-14-76
TIDE: MAX.EBB
TIME(EST) 0553-0743
PLANT OUTPUT(MWH) 288
WIND FROM: W AT 18 KNS.
5 BASET'F
BOULEVARD
S.T.P.
INTAKE
Q
NEW HAVEN
HARBOR
STATION
NAUTICAL MILES
.1 .2 .3 A
.4
KILOMETERS
FORT HALE
PARK
Figure 11. Surface At, (a) and surface base temperature (b), October
14, 1976 - M1d-Ebb. Thermal Surveys, New Haven Harbor,
Summer and Fall, 1976.
-------
NEW HAVEN HARBOR STATION
DATE: 10-13-76
TIDE: LOW WATER SLACK
TIME(EST):Oai5-K>05
PLANT OUTPUT(MWH):446
WIND FROM:SW AT: 14 KNS.
AT'F
"SJx
6
it
I''
5°
ie
it
4»
,-f
n
'* i
'- i
-'"I
; *
'|5 *"
11
K
Figure 12. Cross sectional At, October 13, 1976-Low Water Slack.
Thermal Surveys, New Haven Harbor, Summer and Fall, 1976.
-------
14-
12-
lo-
0000 0400 0800 1200
HOURS (EST)
1600
2000
4-
2-
10-14-76
0000 0400 0800 1200
HOURS (EST)
1600
2000
12-
0
UJ
2-
10-15-76
0000 0400 0800 1200
HOURS (EST)
1600
2000
Figure 13. Percent recirculation as a function of time. Thermal
Surveys, New Haven Harbor, Summer and Fall, 1976.
1190
-------
BEHAVIOR OF THE THERMAL SKIN OF COOLING POND WATERS
SUBJECTED TO MODERATE WIND SPEEDS
M. L. Wesely
Radiological and Environmental Research Division
Argonne National Laboratory, Argonne, Illinois U.S.A.
ABSTRACT
The temperature difference AT across the partially laminar skin of water on the
surface of a water body is determined by the total heat transfer 0 through the
skin, the wind speed u, and the mean temperature T of the skin. Systematic
measurements of these variables were made over a wide range of conditions
at a cooling pond in northeastern Illinois. Waves were present in all cases;
the wind speeds were u = 2.5-7.0 m s"1 at a height of 1 m and water tempera-
tures were T = 18-37.5°C. The main result is the equation
where \ is the water viscosity, K is the thermal diffusivity of water, k is the
water thermal conductivity, T is the wind shearing stress, and p is the water
density. W
INTRODUCTION
The transfer of heat across the uppermost millimeter of a body of water is
limited partially by the slow rate of molecular heat diffusion in this poorly-
mixed cool skin. In the relatively warm water of industrial cooling ponds,
the magnitude of the temperature drop across the skin can approach 1°C. Use
of bulk water temperature instead of actual surface temperature can cause
significant errors both in estimating the total heat loss by use of bulk aero-
dynamic formulae , and in predicting the onset and severity of steam fog .
For oceanic waters, the temperature drop from the surface to the water beneath
the skin is usually much less than 1.0°C . Although of ten small, this
difference is significant because the gradient of air temperature above is
also usually very small. Thus, significant errors in the estimate of sensible
heat flux (and evaporation) from the sea can result if water temperature below
the skin is used instead of surface temperature.
In the present study, an attempt is made to determine the relationship of the
temperature drop across the cool skin to atmospheric conditions and water
properties at a cooling pond. Only a wavy surface subjected to moderate wind
speeds is considered. The rather wide range of water temperatures encoun-
tered allows a systematic examination of the effects of varying (molecular)
water viscosity and thermal diffusivity.
1191
-------
SIMPLE THEORETICAL DESCRIPTIONS
This section reviews some of the past work on the behavior of the cool skin.
A common formula for relating the temperature difference AT across the depth
6 of the thermal skin to the total heat flux Q through the skin is
Q = kAT/6, (1)
o
where k is the thermal conductivity of the water. Rather than a fixed value,
6 is considered the variable to be determined. According to the measurements
of Khundzhua and Andreyev , 6is the depth at which the remaining temperature
drop is about 37% of AT. Rather than a detailed examination of profile
descriptions4'7, a parameterization of the bulk properties is considered here.
To do so, we will assume initially that flow in the skin is mostly laminar, as
would be the case if the air-sea interface were a smooth, nonmobile surface ,
although opinions have been expressed that excessively turbulent flow might
exist when waves are present . If the flow is mostly laminar, the viscosity
of the water and thus its temperature strongly affects the depth of the thermal
skin.
With the assumption that the depth of ±he skin for heat is proportional to that
for momentum, dimensional arguments lead to the relationship
V(Tv/p
w>~*
where T is the viscous stress, Y is the viscosity of water, and p is the
water density. By combining (1) and (2) and assuming that r is proportional
to the shearing stress T aloft in the atmospheric surface layer, Saunders
finds the relationship
(3)
where \ is a numerical coefficient that absorbs the relationship between r
w
and T and other unknown factors. Resorting to this empirically -derived
coefficient may be one of the disadvantages that result from the assumption
that a rigid boundary exists when in fact waves are present. One limited
data set indicates thatr is considerably less than T, perhaps by about 80%,
when waves are present.
For the case of forced convection in the skin, (3) appears acceptable except
that no adjustment has been directly for the difference between the thickness
of the thermal boundary layer in the skin and the depth of the viscous boundary
layer. Since the Prandtl number Pr = Y/K, where K is the thermal diffusivity of
water, is greater than one, the thermal layer should be smaller than the
viscous layer. Approximately, this can be taken into account in accordance
with the theory of flow near a rigid boundary layer by multiplying the right-
hand side of (2) by Pn . This is equivalent to replacing \ in (3) by a new
1192
-------
_ _
coefficient A such that X = A Pr~ 3 . The resulting formulae is
Q = k(r/p /AT /(ArM). (4)
W o
One of the aims of the present experimental effort is to determine if A is better
suited than \ to relate AT to Q.
Deacon derives an equation similar to (4), but with modifications that allow
consideration of cases when AT is across depths much greater than 6are
considered. We shall neglect such elaborations here. His equation in the
present notation becomes, after rearrangement,
]~1 (5)
0 39
where (Pr) is given by his Figure 1 , about equal to 15.2 Pr * for water.
Another somewhat similar approach for describing the thermal skin is given by
Hasse , who finds that a temperature difference across the upper 35 cm of
sea water not exposed to solar radiation can be represented by
where CIQ is a constant appropriate for wind speed u _ measured at a height
of 10 m. Saunders° has found this to be roughly in agreement with (3),
provided absorbtion of solar radiation in the water layer is not significant.
Although a fairly large amount of solar radiation can be absorbed by a water
layer of 35 cm, absorption by layer of depths of 6 ~ 1 mm can usually be
ignored.
MEASUREMENTS
All measurements were taken at the cooling pond complex of Commonwealth
Edison's Dresden nuclear power generating facility near Morris, Illinois,
U.S.A. Many aspects of the cooling lake have been described in a previous
publi cation-^ . Briefly, it is a man-made lake of about 5.3 km surface
area divided into five pools connected by narrow channels. Typically, the
warmest pool is about 10°C warmer than ambient air and the coolest is within
5°C of air temperature. Thus, the heat fluxes from most of the lake are large
and can be measured by use of atmospheric bulk techniques with a relative
accuracy better than above most natural, unheated water bodies.
A wide range of temperatures and wind speeds at the Dresden cooling pond
can be found if samples are taken over an entire year. The measurements
considered here were taken on various occasions when no steam fog was
present during 1973-1978. For the 64 10-min samples taken, the temperature
T of the surface skin ranged from 18 to 37.5°C and averaged 27.5°C, where
6
1193
-------
each T was determined as the average of the surface and the bulk water
temperatures. Wind speeds at a height of 1 mm varied from 2. 5 to 7.0 m s
(which extrapolates to about 3.0-8.5 m s"1 at a height of 10 m), with an
average of 5.1 m s"1. Atmospheric conditions above the pond were unstable
during all data collection periods, with a heat flux upward through the water
skin. The temperature difference AT across the skin varied from 0.3 to 1.5°C.
6
In all cases, measurements were taken aboard a pontoon boat positioned
200-1500 m downwind from the nearest shoreline. A cup anemometer measured
wind speeds at heights of 0.5-1.5 m above the surface, usually supported on
a shaft about 2 m to the side and cross wind of the low-slung boat faced into
the wind. At a height of 0.5 m at the upwind edge of the boat, an aspirated
psychrometer provided wet- and dry-bulb air temperatures. An immersed
mercury-in-glass thermometer supplied water temperatures at a depth of 2-5 cm,
and a hand-held infrared thermometer detected the surface temperature.
The temperature difference AT across the skin was found by several techniques,
usually by vigorously stirring the water in the field of view of the hand-held
infrared device. Other techniques, somewhat less successful, were employed
also. For example, using stirred water in an insulated container as a reference,
the experimenter could obtain a fairly accurate measurement of surface temper-
ature, so that AT could be determined as the difference between the surface
and the immersed temperature.
A difficulty encountered in measuring AT is that it varies in magnitude as
different portions of the wave field are viewed, and is highly responsive to
wind speed variations . Breaking waves might cause serious aberrations, but
if breaking waves were present during sampling at the Dresden pond, the
breaking portion of the wave usually was not included in the view of the
thermometer. Typical variations were noted during one 10 min data-collection
period. With the total heat flux Q across the skin averaging 418 W m and
the mean wind velocity being u = 4.1 m s~* at a height of 0.5 m, AT
fluctuated within the range 0.55-0. 85°C as the wind speed varied from 6.7 to
2.7 m s~*. The values of AT appeared to be roughly proportional to u~l.
How should one properly average these variables? The present approach is
that common in studies of momentum, heat and mass transfer in the atmospheric
surface layer. That is, while gradients of horizontal wind speed, temperatures
and humidities can vary greatly from minute to minute (especially during
unstable conditions), valid relationships of fluxes to mean gradients can be
found by simple linear averaging of the measured variable over periods of
10-60 min. Admittedly, the results are partially empirical and do not explain
the details of the transfer mechanisms involved.
RESULTS
For each 10 min run, the friction velocity u^ = (r/p ), sensible heat flux H,
and latent heat flux LE is calculated by use of a low-level bulk aerodynamic
MLW 1194
-------
method as described by Hicks1 with minor modifications as given by Hicks
et_ah13 . The upward infrared energy flux density is calculated as
Ru = eaTs4' (7)
where e^ 0.95 is the emmisivity of the water surface, a is the Stephan-
Boltzman constant, and T is the surface temperature. The downward infrared
flux is estimated (in unit of watts per square meter) as
R = 5.31 T 610~13 - 20-0.3e aT 4c , (8)
« ci C C
which is Swinbank's15 formula as modified by Paltridge16 and Paltridge and
Platt . The numerical coefficients are empirical, Ta is the average air temper-
ature, e a* 1 is the thermal emissivity of clouds present, T is the estimated
temperaftire of the cloud lower surfaces, and c is the fraction of cloudiness.
For most of the data taken, c was zero. Combining the atmospheric estimates
of fluxes results in
Q = H + LE + R - R, . (9)
u a
This estimate of Q is independent of the procedures used to examine directly
the cool skin, as in the discussion to follow.
Upon inspection of (4) it becomes evident that Q should be unique function
of AT , wind speed (u-^ at a height of one meter), and T . A simple multiple
linear regression of the 64 data points results in the equation
Q = - 631 + 457AT + 88u, + 17.1T. (10)
o l
Figure 1 compares the results of (10) with the atmospheric estimates given by
(9). Although the correlation coefficient for the 64 samples is a highly
significant 0.92, wide scatter in the data is evident. This is most likely due
to the errors in obtaining a reliable average of AT during each 10 min run.
Because of this scatter, statistical techniques win be used to obtain values
of A , X, and CK).
First, to test the expectation that 6 ~ u^ as given by (1), Figure 2 shows
the relationship between dand u^. For this case, Sis calculated as
6=Q/(kAT6). (11)
Even though a rather extensive range of water temperatures were encountered
(18-37. 5°C), the range of u# for each small interval of water temperature was
rather large, so that no systematic change of 6 with u^ due to a correlation
of u with T should have resulted. Figure 2 seems to verify that 6~u^~^.
Since waves were present in all cases, the aerodynamically smooth case is
not considered. These results do not compare well with the results of
1195
-------
who used a wind-water tunnel. His estimates of 6 are considerably greater
for u^ <35 cm s , for which the tunnel produced aerodynamically smooth flow.
Extrapolation of the results of Figure 2 for u^ > 35 cm s"1 yield present
estimates of 6considerably greater than Hill's values. Further comparisons
can be made by examination of Figure 5 of of Kondo^, in which the present
data would follow very well the theoretical calculations derived from
Brutsaert6 for u,. = 15-30 cm s"1.
*
Figure 3 shows the values of A, X / and CIQ determined at the Dresden cooling
pond and plotted as a function of T . The variability of k, Y/ and K with
temperatures are taken into account by application of readily-available,
published estaimates. The density p is estimated for a height of about 0.5 m,
and p is assumed to be 1 g cm~3. jfaso, CIQ is computed on the basis of an
extrapolation to wind speeds at a height of 10 m. The regression lines in
Figure 3 show the dependency of A , X , and CIQ on T . It appears that C10
is negatively correlated with T , with a correlation coefficient of- 0.27 for
the 64 runs, while X and A are positively correlated with T , yielding corre-
lation coefficients of 0.22 and 0.06, respectively. Thus,x seems to suffer
from overadjustment relative to CIQ, but A is not significantly correlated with
T . Additionally, analyses indicate that (Pr) from (5) is about 12.8Pr-0-39,
irthe exponent is fixed at -0.39. The numerical coefficient 12.8 has a
behavior very similar to A ; the use of A Pr~3 or (Pr) = 12.8Pr~°-39 provide
equally good fits in a statistical sense.
The present estimate of X~ 6 is very near to the estimate of about 7 given by
Saunders (1967) for the oceanic case, and to the value of about 8 that can be
inferred from the oceanic data of Hasse^ for temperatures near 158C.
Grassl^O recomputes a value of X ~ 6 for Hasse's data by choice of a different
drag coefficient. For his own data at sea with surface temperatures near
26.58C, Grassl obtains X increasing with wind speeds, with X ~ 4 at UJQ =
3ms"1 and X » 5.5 at u10 = 8.5 m s"1 corresponding to the approximate
range of wind speeds in the present study. On the other hand, Hill1" obtains
X ~ 4 for waves present and X * 11 at lower wind speeds without waves in a
wind-water tunnel; Paulson and Parker^l discuss Hill's results more fully.
The Dresden data do not indicate a significant correlation of X (or A and C
with wind speed; the scatter in the data over the relatively small range of
wind speeds would prevent detection of this correlation if it were only slightly
significant.
For all 64 runs obtained at the Dresden cooling pond, the average value of
Cin is 5.1°C m s'Vdy min'1), which corresponds to 0.0073°C m3 W'1 s"1
1 j T
Hasse's value of about 9.2°C m s~v(ly min"1) is larger, as it should be
because the temperature difference was measured over a much greater depth
(35 cm versus the present 2-5 cm). The average value of C±Q can be used to
derive a simple expression for 6 . That is, (1) and (6) can be combined to
form the expression
1196
-------
6=kC10fc/U*'
where fc is the friction coefficient suitable for a 10 m height, inferred from
the aerodynamic calculations18 to be about 0.0349 for the unstable conditions
at the Dresden lake. With use of k = 0.637 W m"1 "K corresponding to an
overall mean water skin temperature of about 27.5°C encountered, Sis found
to be
<5 = 16.2/u^ . (13)
The curve drawn in Figure 1 shows that (13) presents an acceptable fit to the
data.
CONCLUSIONS
Data collected at the Dresden cooling pond indicate that expressions for the
transport of heat through a viscous layer appear to describe sufficiently the
temperature differences found when waves are present. While the range of
windspeeds examined is small (3-8.5 m s~* at z = 10 m), the rather wide
range of water temperatures encountered (18-37.5°C) allows determination
statistically of the temperature dependency of empirical numerical coefficients.
Values of \ appear to increase slightly with temperature from 6 to 7, and C^g
decreases from 6 to 4.5. When the ratio of the viscous to thermal Boundary
layer thicknesses is assumed to be approximately proportional to Prs , which
is appropriate for a rigid boundary, the resulting coefficient A is found to be
roughly independent of temperature. The overall result is verification of (4),
which can be rearranged to show that the temperature drop across the skin is
AT = 11.5QY¥K'[k(T/p )»] . (14)
o w
The measured thermal skin thickness is in fair agreement with some theoretical
predictions, indicating that the assumption of a rigid, smooth boundary at the
air-water interface appears valid for the wavy surface. As stated by Saunders8,
this might be fortuitous if the effect of possibly significant transfer of wind
stress to the waves by normal pressure forces is compensated by the effects
of turbulence in the water near the surface. Whether fortuitous or not, a
working relationship has been found.
Similar relationships can be found to describe the transfer of nonreactive gases
across the viscous water layer. Because the transfer is greatly impeded by the
low diffusivity of gases in water, the main problems that need to be addressed
deal with the transfer through the water rather than in the lower atmosphere,
especially if reactions in the surface water can substantially increase the
uptake rate.
1197
-------
ACKNOWLEDGE ME NTS
Data were collected at the Dresden cooling pond with the permission and
cooperation of the Commonwealth Edison Company.
REFERENCES
1. Hicks, B. B. , A procedure for the formulation of bulk transfer coefficients
over water, Boundary-Layer Meteorol., 8, 515-524, 1975.
2. Hicks, B. B. , The prediction of fog over cooling ponds, J. Air Pollut.
Contr. As see., 27, 140-142, 1977.
3. Khundzhua, G. G., and Ye. G. Andreyev, An experimental study of heat
exchange between the ocean and the atmosphere in small-scale
interaction, Izv. Acad. Sci. USSRAtmos. Oceanic Phys., Engl.
Transl., J10, 685-687, 1974.
4. Owen, P. R. , and W. R. Thomson, Heat transfer across rough surfaces,
T. Fluid Mech., 15, 321-334, 1963.
5. Yaglom, A. M., and B. A. Kader, Heat and mass transfer between a
rough wall and turbulent fluid flow at high Reynolds and Peclet
numbers, J. Fluid Mech. ,62, 601-623, 1974.
6. Brutsaert, W., A theory for local evaporation (or heat transfer) from
rough and smooth surfaces at ground level, Water Resour. Res.,
JU, 543-550, 1975.
7. Liu, W. T., andj. A. Businger, Temperature profile in the molecular
sublayer near the interface of a fluid in turbulent motion.
Geophys. Res. Letr., 2, 403-404, 1975.
8. Saunders, P.M., Space and time variability of temperature in the
upper ocean, Deep-Sea Res., 19, 467-480, 1973.
9. Saunders, P. M., The temperature at the ocean-air interface, J. Atmos.
Sci., 24' 269-273, 1967.
10. Dobson, F. W. , Measurements of atmospheric pressure on wind-
generated sea waves, J^_Jluid_Mech_., j48, 91-127, 1971.
11. Deacon, E. L., Gas transfer to and across an air-water interface,
Tellus, 29, 363-374, 1977.
1198
-------
12. Hasse, L., The sea surface temperature deviation and the heat flow
at the sea-air interface, Boundary-Layer Meteorol., J,
368-379, 1971.
13. Hicks, B. B., M. L. Wesely, and C. M. Sheih, A study of heat
transfer processes above a cooling pond, Water Res our. Res.,
13., 901-908, 1977.
14. McLeish, W., On the mechanism of wind-slick generation, Deep-Sea
Res., L5, 461-469, 1968.
15. Swinbank, W. C., Long-wave radiation from clear skies, Quart. T. Roy.
Meteorol. Soc. . 89. 339-348, 1963.
16. Paltridge, G. W., Day-time long-wave radiation from the sky, Quart.
T. Roy. Meteorol. Soc.. 96, 645-653, 1970.
17. Paltridge, G. W., and C. M. R. Platt, Radiative Processes in Meteor-
ology and Climatology, Developments in Atmospheric Science 5,
Elsevier, New York, 1976.
18. Hill, R. H., Laboratory measurement of heat transfer and thermal
structure near an air-water interface, T. Phys. Oceanogr., _2,
190-198, 1972.
19. Kondo, J. , Parameterization of turbulent transport in the top meter of
the ocean, T. Phys. Oceanogr. , 6, 712-720, 1976.
20. Grassl, H., The dependence of the measured cool skin of the ocean on
wind stress and total heat flux, Boundary-Layer Meteorol., 10,
465-474, 1976.
21. Paulson, C. A. , and T. W. Parker, Cooling of a water surface by
evaporation, radiation, and heat transfer, T. Geophys. Res.,
77, 491-495, 1972.
1199
-------
1200
200 400 600 800 1000 1200
MEASURED Q(Wm-2)
Fig. 1. Comparison of total heat transfer through the skin as computed from
the regression equation (10), to Q estimated from measurements as
given by (9).
1.5
6 1.0
0.5
8
6
o
o~4
2
15
10
X
oD
10
15
20
u_ (cm s" )
25
30
Fig. 2. Values of 6 estimated via
(11) versus u* calculated
from bulk aerodynamic
relationships. The numbers
near the points and the
standard error bars are
the numbers of 10 min
samples.
15
Fig. 3
z ri—r
5-r-o- r 1
20
25
30
35
40
Measurements of coefficients
as a function of skin tem-
perature. The numbers give
the number of 10 min samples
for each skin temperature
interval and standard error
bars are shown. The dashed
lines represent a linear
regressions.
1200
-------
Alternate Energy Conservation Applications for Industry *
Lawrence J. Schmerzler
Increasing costs of electrical,energy and fuel requires management to
evaluate alternate fntf.tLocl.-j of providing their total c.ucL^y requirement vith
a view towards saving money as well as fuel.
A number of energy conservation illustrations are presented utilizing
cogeneration, regenerators, recompression, and heat pumps. Thermoeconomic
analysis is made for a few industrial cogeneration applications from 50 to
1200 kw.
While cogeneration systems may not always be practical or economical,
it was found possible to obtain payback periods of under 4 years. In
addition to possible economy, cogeneration systems have other merits such
as reliability, uninterrupted service, and national security.
*This paper was not presented.
1201
-------
Mineral Cycling Model of the Thalassia Community as Affected by Thermal
Effluents.
By Peter B. Schroeder and Anitra Thorhaug, Department of Biological Science,
Florida International University, Tamiatni Campus, Miami, Florida 33199.
ABSTRACT
The cooling water effluents from fossil fuel and nuclear power sta-
tions often contain low levels of heavy metals from nuclear plants and
potentially contain radionuclides. When these effluents discharge into
subtropical or tropical estuaries in the Caribbean area and the Gulf of
Mexico, the biological community most directly affected is the tropical
seagrass bed dominated by the marine angiosperm, Thalassia testudinum.
These seagrass beds are very productive and form the basis of a food
chain which supports many marine organisms harvested by man. In order
to understand the process and explore the possible consequences of intro-
ducing pollutants at the base of a food chain leading to man we have pre-
pared an energy circuit diagram and mathematical model of the flow of
heavy metal and radiopollutants through a tropical seagrass community
and into higher tropic levels.
The model comprises seven compartments: water, substrate, seagrasses,
macroalgae, epiphytes, detritus and macroanimals. The model will be trans-
lated into CSSL (Continuous Systems Simulation Language) and coupled to a
productivity-biomass model which is also presented. Simulations of dif-
ferent size communities under various environmental conditions and differ-
ent levels of energy-related pollution will be made.
1202
-------
INTRODUCTION
"The importance of seagrass meadows to roastal marine ecosystems
is not fully understood and is generally uncicrc-s t ima ted . . . Oospi to the
extensive studies on seagrass productivity and on the temporal and spa-
tial variability in biological composition of seagrass communities, little-
is known of the general principles of ecosystem function and factors con-
trolling "ecological success' of the communities" (Thayer, £t a_l_. , 1975).
Because of the location of seagrass communities in estuaries and often
directly adjacent to the shoreline, they are one of the marine systems
most directly impacted by man's activities. Seagrasses are directly
subjected to increased nutrient loads, heavy metal, thermal, and radio-
active pollutants, dredge and fill; and the effects of recreational
activities, sueh as boating in the estuaries. The seagrass communities are
the major basis of a food web that leads in many locations to man.
Destruction of seagrasses leads to a great decrease in invertebrate and
fish species (Thorhaug e_t a_l_., 1974). In particular, as energy related
industry expands in the coastal regions, increasing impact occurs on marim-.
grassbeds. One of the large impacts of energy related industry is the
release of trace metals into the environment. A number of studies have
been made on the uptake and content of trace metals by specific marine
organisms. Goldberg (1965) summarized these findings. A major deficiency
in these studies is that most of the environmental parameters that might
have affected the results were unrecorded .
In chemical studies in the subtropical estuary, Biscayne Bay, Florida,
the seagrasses and macroalgae were found to contain a significant pro-
portion cf trace metals and to cycle significant fractions of the amount
each year ("egar, et a]., 1971, Gilio and Segnr, 1976). Studies by
1203
-------
Parker (1962, 1966) indicate that two compartments,'the sediment and the
seagrass Thalassia testudinum KtJnig, constitute prime reservoirs for radio-
nuclides added to the estuary and there can be a rapid flux between these
two.
To understand trace metal flux in the seagrasses, one must consider
the plants1 physiology. Research on absorption of trace metals and other
elements by aquatic plants was reviewed by Sutcliffe (1962). Early work
was done with Vallisneria, Elodea and Lemna species, all fresh water aqua-
tics not subjected normally to environments of fluctuating salinities.
Submerged marine angiosperms such as Thalassia secondarily migrated to
the sea; in readapting to an environment of fluctuating salinity, they
have developed complex osmoregulatory microstructures and are able to
maintain a steep electrochemical gradient with the surrounding water
through selective exchange of ions (Jagels, 1973; Gessner, 1971). The
result is a highly dynamic system dependent primarily on the maintenance
of osmotic balance by the seagrasses. Osmoregulation in seagrasses is not
only a function of salinity, but also of temperature and the molecular or
ionic state of the minerals in the water (Schroeder, 1975), and possibly
of light and other factors (Bachmann and Odum, 1960). Unless mineral cycling
in these systems is studied as a dynamic function of interacting environ-
mental parameters such as light, temperature, salinity;and biological fac-
tors such as phytosynthesis and plant growth and senescence, at best it
can be only partially understood.
Schroeder (1975) and Schroeder and Thorhaug (in press) found that
radioactive cation uptake by the seagrass Thalassia testudinum occurred
primarily by absorption on cation exchange sites and was largely reversible
1204
-------
by a wash with a solution of the nonradioactive isotope. A two step process
of uptake was suggested. Initially, there was a large uptake, probably on
the cation exchange sites located in the "outer space" of the plant. This
process was readily reversible. A slower uptake of lesser magnitude also
occurred, thought to represent cations which had been transported into the
cell by the cell membrane.
Elemental content of this subtropical and tropical marine angiosperm
was found to be concentration and temperature dependent; uptake may be in-
creased several times by a temperature change of less than 5°C (Thorhaug
and Schroeder, this volume). Gessner (1971) showed that Thalassia cells
overcame plasmolysis resulting from increased salinity by uptake of cations.
Bachmann and Odum (1960) strongly suggested that zinc-65 uptake by macro-
algae was light controlled. Parker (1966) found that cobalt-60 content in
Thalassia may be five times greater at night than during the day. As indi-
cated by Walsh and Grow (1973), and Pulich e£ 4!t. (19(?6)and others, sampling sur-
veys of the distribution of elements in seagrass communities collected at
only one discrete time may lead to erroneous conclusions concerning mineral
cycling because of the magnitude of seasonal or diurnal changes. We con-
clude that the cycling of minerals through seaqrass beds and the exchange of
elements between the compartments in these ecosystems are dynamic processes
and must be studied as a function of time and environmental variables.
The Model
The Thalassia seagrass community is a highly diverse system comprised of
many species of plants and animals comparable to coral reefs in diversity
(Thorhaug and Roessler, 1977 ). The composition and numbers of organisms
1205
-------
comprising the system change and shift with the seasons, ospeciallv in sub-
tropical locations, complicating the statistical analysis of the effects of
stress.
However, as a first approximation, the seagrass community can be con-
ceptualized as a limited number of separate compartments which can be treated
as components representing the entire system. Changes in the composition and
functions of the organisms comprising a compartment can be interpreted as a
quantitative change in the function of the compartment, j_.e. , the interchange
of the compartment with the other compartments in tho system.
Because it is possible to compartmentalize the Thalassla seagrass system
and quantify the flows between the compartments under different environmental
conditions, it is possible to create mathematical models of productivity and
mineral cycling in the Thalassia community. These models were created not
only to simulate the response of the system to environmental change but also
to provide a structure on which to design future research.
Two interlocking conceptual models were made. One is an energy flow
model (Figure 1) used to quantify a representative Thalassia bed, and de-
scribe annual changes in its composition. The second is a model of the
interchange of minerals (heavy metals, radiopollutants, micronutrients)
between the compartments of the Thalassia seagrass system. It will be used
as the basis for simulations of micronutrient cycling in the system and
the biological concentration in marine food chain:; of pollutants released
into nearshore waters or estuaries dominated by the Thalassia seagrass com-
munity.
Figures 1 and 2 show the biomass-productivity model and the model of
mineral cycling respectively diagrammed in the symbolic modelling language
1206
-------
developed by H.T. Odura (1971). Both models are composed of major
ments: water, substrate, seagrasses, their epiphytes, macros lp,ac>, detritus,
and the macroscopic animals. The seagrass compartment is composed of two
sub-compartments: the sunsediwent parts of the plant (roots and rhizomes)
and the above-sediment parts (vertical short shoots, living and attached
dead leaves).
The variable names used in the models are found in Table 1. The co-
efficients that pertain to each flow or exchange are shown as labels on
the arrows indicating the pathways in the diagrams. Definitions of the rate
symbols used in the biomass-productivity model are given in Table 2. De-
finitions of the rate symbols used in the mineral cycling model are given
in Table 3. Values for all rates used in the models are dependent on en-
vironmental parameters, principally temperature and salinity.
The biomass-productivity model expressed in mathematical equations is
found in Table 4. This model is relatively straightforward. A Michaelis-
Menten expression is used to describe energy flow to Thalassia, the epiphytes,
and the macroalgae. The Km symbol used in these equations follows the con-
vention of Michaelis-Menten equations and represents the level of light or
nutrients allowing one-half maximum growth. A logistic expression is used
to describe energy flow to the heterotrophic compartments. Respiration is
assumed to be a simple exponential decay function of each biomass compartment.
As a first approximation, epiphyte and Thalassia loaf conversion to detritus
is assumed to be a mutually dependent function; when leaves are heavily epi-
phytized, they die and slough off, carrying their epiphytes with them.
The mineral cycling model expressed in mathematical equations is given
in Table 5. Biomass values used in this model are generated by the biomass-
productivity model previously described. Otherwise it is a simple mass
1207
-------
balance model. The water compartment is volume dependent as would be the
case in tank microcosm studies. When used to simulate an actual or field
situation, the water compartment would be considered infinite in volume.
Losses by other compartments to the water would be considered losses from
the system, and concentration of the mineral or element under study would
remain unaffected in the water. The concentration in the water would be
determined from a table function which would simulate the change in con-
centration in the water as it passed over the seagrass bed affected only
by the source of the mineral and the flow of the water.
The mathematical models are presently being translated into Continuous
Systems Simulation Language (CSSL) in order to produce computer simulations
with the UNIVAC 1106 at the University of Miami Computer Center in Coral
Gables, Florida.
Coefficients for the mineral cycling model are being determined by a
series of microcosm studies using radiotracers. Initial values of elemen-
tal content in the compartments are being taken from existing studies (Eis-
ler et al., 1972; Gilio and Segar, 1976; Goldberg, 1965; Schroeder, 1975;
Windom, 1972).
Coefficients for the biomass productivity model are also available
for previous reports (Jones, 1968; Thorhaug and Garcia-Gomez, 1972; Thor-
haug and Kellar, 1972; Bach, 1975; Josselyn, 1965; Penhale, 1976; Edwards,
1977; Greenway, 1977; Thorhaug and Roessler, 1977; Thorhaug, 1977). Ini-
tial conditions for simulations of this model will reflect actual condi-
tions in the particular seagrass bed under study, or conditions in typical
Thalassia communities will be used. Seagrass community data from upper
subtropical (Thorhaug jit al., in preparation), subtropical (Thorhaug and
1208
-------
Roessler, 1977; Thorhaug et^ al., 1973; Thorhaug and Stearns, 1972), and
tropical (Puerto Rico - Schroeder, 1975; Jamaica - Greenway, 1977; Cuba
- Buesa, 1974) can be compared for effects of energy-related industry.
1209
-------
KIO
-------
-------
Table 1. List of Variable Names Used in the Models.
S=SUN LIGHT
N=NUTRIENTS
I=TOTAL ACTIVITY OF COMPARTMENT
B=BIOMASS OF COMPARTMENT
V=VOLUME OR WEIGHT OF COMPARTMENT (SURFACE AREA OF SEDIMENT)
C=CONCENTRATION PER UNIT (GRAM-MILLILITER-AREA) IN COMPARTMENT
B=CHANGE IN BIOMAS5 (B) WITH TIME
V=CHANGE IN VOLUME (V) WITH TIME
C=CHANGE IN CONCENTRATION (C) WITH TIME
Subscripts
L=LEAVES
R=ROOTS-RHIZOMES
E=EPIPHYTES
M=MACROALGAE
A=ANIMALS
F=FECES
D=DETRITUS
W=WATER
G= SEDIMENT
0=INITIAL CONDITIONS
I=NEW CONDITIONS
Rates
K= Rates
1212
-------
Table 2. Definitions of Rate Svmbols Used in Bionans-Productivity Model.
kl=Coefficient of above substrate Thalassia parts growth
k2=Coefficient of below substrate Thalassia parts growth
k3=Coefficient of epiphyte growth
k4=Coefficient of macroalgae growth
k5=Coefficient of conversion of above sediment Thalassia parts to detritus
k6=Coefficient of conversion of epiphytes to detritus
k7=Coefficient of macroanimal feeding on Thalassia
k8=Coefficient of macroanimal feeding on epiphytes
k9=Coefficient of macroanimal feeding on macroalgae
klO=Coefficient of macroanimal feeding on detritus
kll=Coefficient of conversion of macroanimals to detritus
k!2=Coefficient of conversion of macroalgae to detritus
k!3=Coefficient of macroalgae respiration
kI4=Coefficient of macroanimal respiration
k!5=Coefficient of respiration in detritus
k!6=Coefficient of respiration by above sediment Thalassia
k!7=Coefficient of respiration by below sediment Thalassia
k!8=Coefficient of respiration by epiphytes
1213
-------
Table 3. Definitions of Rate Symbols Used in Mineral Cycling Model
klL=Rate of uptake by leaves from water
k2L=Rate of loss to water from leaves
k3L=Rate of translocation from root-rhizomes to lcnvos=kAR
k4L=Rate of translocation from leaves to root-rhizomes=k3R
k5L=Rate of translocation from epiphytes to leavos=k4E
k6L=Rate of translocation from leaves to epiphy teH=k3F!
klR=Rate of uptake by roots from sediment
k2R=Rate of loss to sediment from roots
k3R=k4B
k4R=k3B
klE=Rate of uptake by epiphytes from water
k2E=Rate of loss to water from epiphytes
k3E=k6L
k4E=k5L
klM=Rate of uptake by rnacroalgae from water
k2M=Rate of loss to water from macroalgae
klA=Rate of uptake by animals from water
k2A=Rate of loss to water by animals
k3A=Rate of feeding on leaves
k4A=Rate of feeding on epiphytes
k5A=Rate of feeding on macroalgat
k6A=Rate of feeding on detritus
k7A=Rate of excretion=k6D
klD=Rate of uptake by detritus from water
k2D=Rate of loss to wacer from detritus
k3D=Rate of conversion of leaves to detritus
k4D=Rate of conversion of epiphytes to detritus
k5D=Rate of conversion of macroalgae to detritus
k6D=k7A
k7D=Rate of mineralization of detritus
klG=Rates of uptake by sediment from water
k2G=Rate of loss to water from sediment
1214
-------
Table 4. Biomass-Productivity Model.
i = klNL( N )(^i ) . k L _ k LE _ k LA(l-cA)
R = k2NL(- - ^
L vKmn2
£ = kNE(- N
D = k5LE
Krr.s,-f-S
Kms2+S
- kR
)( S ) _ k k LE - koEA(l-cA) - k,oE
Kms3+S 3 b o 10
= k4NM( ^
A = (k?L + ksE + k9M 4- klQD)(l-cA)A -
~ k15D
-------
Table 5. Mineral Cycling Model.
Leaves-Vertical Shoots
ILI = ILO + CL BL + CL BL
CL = K1L CW - K2L CL + K3L CR - K.4L CL + K5L CE - K6L CL
Roots-Rhizomes
IRI = IRQ + CR BR + CR BR
CR = KIR CG - K2R CR + K3R CL - K4R CR
Epiphytes
IEI = LEO + CE BE + CE BE
CE = K1E CW - K2E CE + K3E CL - K4E CE
Macroalgae
IMI = IMO + CM BM + CM BM
CM = KIM CW - K2M C>5
Macroanimals
1AI = IAO + CA BA + CA BA
CA - K1A CW - K2A CA + K3A CL + K4A CE + K5A CM + K6A CD - K7A CA
Detritus
TDI = IDO + CD BD + CD BD
CD = KID CW - K2D CD + K3D CL + K4D CE + K5D CM + K6D CA - K7D CD
Water
IWI - IWO + CW VW + CW VW
CW = K2L CL - K1L CW + K2E CE - KlE CW + K2M CM - KIM CW + K2A CA - K1A CW
+ K2D CD - KID CW + K2G CT, - KIG CW
Substrate
IGl ••= ISO + CG VG + CG VG
CG - KIG CW - K2G CG + K2R CR - KIR CG + K7D CD
1216
-------
Captions for Illustrations
Fig. 1 Ifoalassia tastudinum Blomass-produclivity model
Fig. 2 Thalassia testudinum Mineral cycling model
1217
-------
LITERATURE CITED
Bach, S.D. 1975. The Distribution and Production of Calcareous Macroalgae
in Card Sound, Florida. Ph.D. Dissertation. . University of Michigan,
Michigan.
Bachmann, R.W. and E.P. Odum. 1960. Uptake of zinc-65 and primary produc-
tivity in marine benthic algae. Limnol. Oceanogr., 5(4):349-355.
Buesa, R.J. 1974. Population and biological data on turtle grass (Thalassia
teatudinum KHnig, 1805) on the Northwestern Cuban shelf. Aqiiaculture,
4(2):207-226.
Edwards, R.E. 1977. Respiration of a Shallow-water Benthic Community^
Associated with the Seagrass Halodule wrightii. Masters Thesis.
University of Miami, Florida.
Eisler, R., G.E. Zoroogian and R.J. Hennekey. 1972. Cadmium uptake by
marine organisms. J. Fish. Res. Bd. Canada, 29:1367-1369.
Gessner, F. 1971. The water economy of the seagrass Thalassia testudinum.
Mar. Biol., 10:258-260.
Gilio, J.L. and D.A. Segar. 1976. Biogeochemistry of trace elements in
Card Sound, Florida. Inventory and annual turnover. IN: Symposium
on Biscayne Bay, University of Miami, Florida.
Goldberg, E.D. 1965. Review of Trace Element Concentrations in Marine
Organisms. 2 vol. P.R. Nuclear Center, Mayaguez, P.R.
Greenway, M. 1977. The Production and Utilization of Thalassia testudinum
in Kingston Harbor, Jamaica. Ph.D. Dissertation. University of West
Indies, Kingston, Jamaica.
Jagels, R. 1973. Studies of the marine grass Thalassia testudinum. I.
Ultrastructure of the osmoregulatory leaf cells. Amer. J. Hot., 60(10):
1003-1009.
1218
-------
Jones, J.A. 1968, Primary Prr.hu-r. ! vl rv bv :!i.- Tn>j-ii_a! _M.ir i_ii" 'i"Mrt.l_e
Crass, Thalassia tcstudinum Konip and U.s J''j1.ij'"y t *••>!. I'h.n. Disser-
tation. University of Miami, Florida.
Josselyn, M.N. 1975. The_Cn >T.h an.: Oisrrjbuti TI "• Two :-• p. -c- i ,••••. of
Lavirencia, a Red M;.KT;I.I 1 ^in , in ('_i_rd_ Somi ! . i'ioridri. Masters
Thesis. University of Miami, Florida.
Odum, H.T. 1971. Ejnv^rr'nrwn^, I'owor, and Society. John Wiley <:- Son--, N'.:'.
Parker, P.L. 19^" .nc in a Texas bay. Pub 1 . Inst ._Mar . _.S_ci_._ Hniv. _'l*:-: ..s .
8:75-79.
_ 1966. Movement of radioisotopes in a marine bay: cob.il t-60, iron- 59,
manganese-54, zinc-65, sodium-22. Piij^l. Tjist. Mar. Sr.i. Univ. Ti'xa_s_.
11:102-107.
Penhale, P. A. 1976. P r imary Prod nc.ti v rt y ._jjj^jj_lji!;-A ^Ir
and Nutrient Transport _i_n__nn Epiphjv-tc'^Kj^ Igr.iss (Zoster a narin-i) __ S vs_t e~ .
Ph.D. Dissertation. N.C. State L'nivestity, N.C. s
Pulich, W. . S. Barnes and P. Parker. 197ft. Trace metal cycles in se,u:r.-;..-
communities, in Wiley, M. (ed. ). Es t na r i no P r 01 • ess e s . _ K Academic
Press, N.Y.
Schroeder, P.B. 1975. Thermal _St_ress_j_n_ Tlialassin U-stud inum. Ph.D. Dis-
sertation. University of Miami, Florida.
_ , and A. Thorhaug. (in press) Uptake of zinc-6;J by Jhji]j3f;_sjji_^_us_n.ijl_|_i:];ir;.
Mar. Biol.
Segar, D., S. Gerchakov and T. Johnson. 1971. Chemistry, in R.C. Badc-r and
M.Roessler (eds. ) . ^cjoJ^i^L^JLii^^
Sound. University of Miami. Florida.
121£
-------
Sutcliffe, J.F. 1962. Mineral Salts Absorption in Plants. Pergawon
Press, N*.Y.
Thayer, G.W., D.A. Wolfe and R.B. Williams. 1975^ The impact of man
on seagrass systems. Amer. Scientist, 63:288-295.
Thorhaug, A. (in preparation) Primary production measured on a long term
basis of the seagrass Thalassia testudinum in two subtropical estuaries
fringing the tropics.
, and J. Garcia-Gomez. 1972. Preliminary laboratory and field growth
studies of Laurencia complex. J. Phycol., 8(S):10.
, and K.F. Kellar. 1972. Laboratory and field growth studies of four
greea calcareous algae. I. Preliminary results. J. Phycol., 8
-------
Synergistic Effects of Substances Emitted from Power Plants on Subtropical
and Tropical Populations of the Seagrass ThaJLassLa testudinum: Temperature,
Salinity and Heavy Metals.
By Anitra Thorhaug and Peter B. Schroeder, Department of Biological Science,
Florida International University, Tamiami Campus, Miami, Florida 33199.
ABSTRACT
The seagrass Thalagsla testudinum is the dominant species in much of the
Gulf of Mexico and Caribbean nearshore marine system. Dense meadows of sea-
grasses appear immediately adjacent to the shoreline where energy related in-
dustry has often been sited. Power plants have released their heated effluents
accompanied by salinity changes (dilution causing lower salinity, or waters
.evaporated in cooling ponds raising salinities) along with heavy metals on
seagrasses causing damage. Although the effect of temperature on the heated
effluents in the tropics and subtropics is of fundamental importance in mor-
tality of organisms, it has been shown that there are sublethal temperature
regimes where synergistic effects of other effluent components probably figure
importantly. Field data from two subtropical and on tropical effluent canals
have recently been compared (Thorhaug, Blake and Schroeder, 19?8) . Unfortu-
nately, measurements of all parameters are usually not frequent enough in
field situations to delineate the entire topological surface of synergy. There-
fore, detailed laboratory experiments using tropical and subtropical Thalassia
were undertaken to describe synergistic effects. Temperature versus uptake of
heavy metals appears fairly similar and predictable for most cations in the
20 to 30 C range; there are minima in all cationic uptake examined at 30° to
32 C; above this there is a strongly accelerated rate of uptake of metals.
High salinities (50%) lower the upper lethal temperature by one to two degrees
centigrade; lower salinities (20%) have a much smaller effect. A comparison
of power plant effluents from Gulf of Mexico to Central Caribbean shows summer
effluent temperatures are in the range of 31° to 35°C, which is the sublethal
to lethal area of maximum effect of synergy examined in laboratory experiments.
This emphasizes that the tropics are "on the brink of disaster."
1221
-------
I. INTRODUCTION
Throughout the Gulf of Mexico, Caribbean and southeastern Florida coast,
the seagrass Thalassia testudinum is the dominant nearshore species. It is
most dense very close to shore, decreasing in productivity and abundance as
one goes seaward. Thus, energy-related industry,often sited on estuaries or
marine shorelines, has in the past impacted this densest zone of Thalassia
in a series of sites (Thorhaug, 1974; Thorhaug e£ £l•, 1977; Thorhaug and
Schroeder, 1977). Heated effluents from cooling canals have been shown to
hav.e a lethal effect on Thalassia populations from the subtrooics and tropics
above 35°C for extended time periods such as ten days (most recently reviewed
by Thorhaug et al., 1977). However, it is in the sublethal temperature regime,
where temperature (usually the dominant factor from heated effluents from
power plants) is not lethal, that interests us. At sublethal temperatures, the
synergistic effects of other substances, such as salinity and/or heavy metals,
come into focus.
There has been little quantified field evidence for the effects of heavy
metals and/or salinity changes on such populations. The critical question is
"How long does the maximum or minimum condition impinge on the population?"
The question of lethal substances at power plants has usually been answered
by data which reports the average salinity or average heavy metal concentration
during a certain season, or gives the yearly limits, but not their duration.
Because detailed $eries of measurements of heavy metals and other pollutants
at power plants were not available during the summer months when the sublethal
effects were encountered, and also due to the many interaction factors present
in thermal effluents, we have necessarily resorted to laboratory measurements
to understand the detailed synergistic effects of high temperature, salinity
and heavy metals.
J222
-------
The effect of heavy metals on Thalassia has received little attention in
the literature. Biogeochemists have examined the cycling of radionuclides.
Parker (1962, 1963, 1966) in a Texas Bay showed zinc was important in the
mineral cycling in these shallow waters. Gilio and Segar (1976) showed a good
deal of the trace element content in <:he Card Sound estuary was accumulated in
Thalassia in an inventory of the major trace metal components (Tables 1 and 2)
and these authors extrapolated the rate of cycling through Thalassia would be
a critical factor in mineral cyclirsg in Card Sound (Table 3).
Schroeder and Thorhaug (in press) showed that Thalassia seedlings take
up zinc-65 both through roots and leaves, dependent on which organ is exposed
to the radionuclides. Cations such as zinc can be translocated from the root
to the leaf or vice versa. Seeds do not have a significant role in uptake.
Much of the original cation concentration is adsorbed to exchange sites on the
surface and can be washed off. Heavy metals can be concentrated in tissues up
to 500 times ambient sediment water concentration within 10 days in these sea-
grasses which form the base of the detrital food chain in subtropical and tro-
pical estuaries. Pumping effects of heavy metals from sediment to water or
vice versa did not appear significant. Previously McRoy et al. (1972) have
postulated a phosphate pump by another seagrass Zostera, which mechanism
was suggested by Gilio and Segar (1976) as possibly occurring with the cations
in Thalassia.
In this present discussion we examine the combined effect of temperature
and salinity on the uptake of metals in Thalassia from the tropics and from
the subtropics to begin a determination of synergistic effects of temperature
dependencies which might be found in sites impacted by energy-related industry.
1223
-------
II. METHODS
A. Radiotracer Experiments
The ability to grow Thalassia under laboratory conditions has been demon-
strated in the past (Thorhaug, 1971, 1972, 1974). Large plugs of the seagrass
Thalassia testudinum with sediment were removed from the field in snug fitting
vessels and returned to the laboratory. Glass tanks into which these plugs
fit snugly were provided with grow-lux lights (eight hours on, sixteen hours
off), calcareous sediment, and aeration.
Temperature was controlled by aquarium immersion heaters (+ 0.5°C). Fil-
tered seawater was used; salinity was adjusted by using artificial seawater
diluted to the appropriate salinity level in daily checks.
A cocktail of salts of zinc-65,cobalt-57, cobalt-60, cesium-137, manganese-
54, silver-108, and iron-59 was prepared so that all isotopes would have approxi-
mately the same radioactivity. The mixture was added to tanks of unfiltered sea-
water 35%o; pH was adjusted to normal within a short time period. All cations
except cobalt-60 (added as cobaltamine) were introduced as elemental ions.
At 4, 11, 18, 23 and 31 days, plant samples were removed, rinsed thoroughly,
fractionated into roots, rhizomes, blades and seeds, dried at 105°C and weighed.
Samples were then placed mixed with plaster of Paris in petrie dishes to pro-
vide the same geometry for all samples.
Counting procedures included counting on a shielded germanium semiconductor
detector. Multichannel pulse height analyzer separated counts into 1024 chan-
nels of which 299 were used for analysis. Energy intervals were 1.5KeV. 120
samples and 8 standard spectra were used. A computer program analyzed disinte-
grations per minute per g dry weight for each radionvclide. Two cobalt radio-
1224
-------
Isotopes provided an additional internal check for running a dummy sample
spectrum.
B. Salinity-temperature Experiments
Plants collected as seeds in the Florida and Bahamas areas by methods of
Thorhaug (1974) were held in out-of-door running-seawater tanks with six inches
of peat sediment (described also in Thorhaug, 1974) replanted in these tanks
for observation.
Two types of temperature control devices were utilized to determine upper
temperature limits: polythermostats and controlled temperature baths. The*pro-
cedure and apparati have been thoroughly described by Thorhaug (1976). Basically
Millipore filtered (Whatman #42) seawater was adjusted to appropriate salinity
with "Instant Ocean," then equilibrated in the temperature device. In the two
polythermostats, six small seedlings per cuvette were utilized; large numbers
of seedlings or mature plants could be used in the temperature baths. Speci-
mens were exposed to a steady-state temperature for a given time, at the con-
clusion of which specimens were carefully examined, tagged and planted in a
second outdoor running-seawater tank for observation of death (1 month holding
time). Two to four month seedlings and mature plants with rhizomes and roots
intact were utilized. Time of exposure was 12, 24, and 48 hours; salinities
20, 36 and 50%o, + 0.5%o; temperature 30° to 45°C, + 0.05°C.
III. RESULTS
A. Trace Metal Uptake
In experiments at ambient salinity with Puerto Kican specimens, uptake
rates of cations to whole plants (fractionated into blades and roots) showed
1225
-------
one d?p3nd«icy from 30° (or 32°C) to 37° (or 38°C) and usually around 37° to
38°C an abruptly decreasing dependency (Figures 1 and 2), which is the accu-
mulctea data of four trials, six plants per temperature per trial. In some
c-<;,
-------
between 20 and 50%o salinity, 25°C to 40°C, and 6 to 96 hours showed that upper
temperature lethal limits were more sensitive to salinities 15%o above (i.e.,
50%o) mean than 15%o below (Figure 3). The temperature-mortality curve resembled
a step-function for 20 and 36%o, while the 50%o resembled a skewed Gaussian
curve. At 36%o for 12 hours the upper survival temperature limit was 36°C,
which dropped to 35°C at 48 hours, and was the same at 20%o, while at 50%o,
48 hours, the limit was 33°C. Tests showed no difference between seedling and
^ture plant lethal limits.
T*-e two cobalt isotopes acted as an internal check on the methodology it-
self. Behavior of these two radioisotopes in leaves, roots and rhizomes was
quite similar.
IV. DISCUSSION
The effect of temperature on the survival and physiology of the seagrass
Thalassia is profound. In many subtropical and tropical estuarine areas, am-
bient summer temperatures are near a mean of 30°C with mid-afternoon excursions
in shallow water 1o 33° or 34°C. Yet the upper tolerance limits for Thalassia
over a long time period are 34° to 36°C (Thorhaug ejt £]L. , in press), very close
to non-impacted summer temperatures. High salinity (50%o) as seen in our re-
sults, affects this upper temperature limit, causing it to fall to 33°C. Lower
salinities (in the range of 20%o) do not have much effect from normal (34%o)
on upper temperature tolerance, perhaps indicating less adaptation to saline
conditions than the brackish water from which ancestors of Thalassia arose.
The sublethal temperature area from 31 C upward is particularly of interest
to those concerned with thermal effluents from power plants. The uptake of most
of the radioisotopic cations investigated is highly temperature dependent above
1227
-------
31° to 32 C. A minimum occurs at about 30° to 32° C with a maximum near 35°
to 38°C. Below 30°C, the temperature dependency varies from a low temperature
coefficient in elements such as zinc-65 and bisrauth-207 to high temperature
coefficients in cesium-137 and sodium-22. Both above and below 30°C, bismuth
showed less strong temperature dependency than the other elements. Tempera-
tures in the range above 31° to 32°C produce higher uptake of the heavy metals
than temperatures below 30°C in general. This is significant for thermal ef-
fluents of semitropical and tropical power plants, which release both heavy
metals and cause temperature regimes above 31° to 32°C for periods of three
to five months in the semitropics and longer in the tropics.
There are several levels of information to be obtained from this data in-
cluding implications for: the organismic level of Thalassia functioning; and
cooling canal operation.
The major implications for Thalassia plant physiology are several. Trace
metals accumulate rapidly from external sources in Thalassia tissue. Within
hours.concentrations of several times ambient appear and within days, accumula-
tions build to several hundred times ambient. The elements can enter through
either roots and rhizomes from sediment concentrations of metals or from the
water column via leaves. Translocation between tissues occurs rapidly within
hours within the plant (Schroeder and Thorhaug, in press). Coefficient of up-
take in roots and leaves differs but the effect of temperature on uptake in
roots and leaves is fairly similar.
The implications of these studies to cooling canals from power plants are
several. First, 15£»concentration of salts in normal seawater (32 te 35%o)
such as released from cooling poods (or other evaporating devices) appears to
have an effect in lowering Thalassia lethal temperature limits, whereas lowering
1228
-------
the salt in seawater 15%
-------
ACKNOWLEDGEMENTS
The authors appreciate the support of ERDA grant if E(40-l) 4493 for the
.support of the bulk of the reported work. Part of the Puerto Rican work was
sponsored by the Puerto Rican Nuclear Center in Mayaguez, Puerto Rico and
the laboratory graduate participation program of the Oak Ridge Associated
Universities, Inc., to Dr. Schroeder.
1230
-------
LITERATURE CITED
Gilio, J. L. and D. A. Segar. 1974. Biogeochemistry of trace elements
in Card Sound, Florida. Inventory and annual turnover, pp. 1-17.
IN: A. Thorhaug (ed.) Biscayne Bay: Past, Present and Future. Sea
Grant Sp. Report No.5, Univ. of Miami, Miami, Florida.
teRoy, C.,R. Barsdate and Nebert. 1972. Phosphorus cycling in an
eelgrass (Zostera marina L.) ecosystem. Limn. Oceanogr. 17:58-67.
Parker, P. L. 1962. Zinc in a Texas bay. Publ. Inst. Mar. Sci., Univ.
Texas. 8:75-79-
Parker, P. L., A. Gibbs and R. Lowler. 1963. Cobalt, iron and
manganese in a Texas bay. Publ. Inst. Mar. Sci. Univ. Texas.
9:28-32.
Parker, P. L. 1966. Movement of radioisotopes in marine bay: cobalt-60,
iron-59, manganese-54, zinc-65, sodium-22. Publ. Inst. Mar. Sci.
Univ. Texas. 11:102-107.
Thorhaug, A. 1971. Grasses and macroalgae. IN: R. G. Bader and M. A.
Roessler (eds.) An Ecological Study of South Biscayne Bay and Card
Sound. Progrs. Rpt. to AEC (AT(40-l)-3801-3) and FPL Co. ML 71066.
Thorhaug, A. 1972. Laboratory thermal studies. IN: R. G. Bader and M. A.
Roessler (eds.) An Ecological Study of South Biscayne Bay and Card
Sound. Prgrs. Rpt. to AEC (AT(40-l)-3801-4) and FPL Co. RSMAS 72060.
Thorhaug, A. 1974. Effect of thermal effluents on the marine biology of
southeastern Florida, pp. 518-531. IN: J. W. Gibbons and R.R.
Sharitz (eds.) Thermal Ecology. AEC Symp. Series (Conf. 730505).
Thorhaug, A. 1976. Tropical macroalgae as pollution indicator organisms.
Micrones ica 12(1):49-68.
1231
-------
Thorhaug, A.,N. Blake and P. Schroeder. 197 . The effects of heated
effluents from power plants on the seagrass Thalassia testudinum
quantitatively comparing estuaries in the su.b tropics to the
tropics. Mai. tfoll. Bull. 9(7):181-187.
Thorhaug, A. and P. Schroeder. 1977. A comparison of the biological
effects of heated effluents from two fossil fuel plants:
Biscayne Bay, Florida, in the subtropics:Guayanilla Bay, Puerto
Rico, in the tropics. Vol 3, 118:133-164. IN; S. Lee and S,
Sengupta (eds.). Waste Heat Management and Utilization. Miami,
Florida.
1232
-------
Table 1. Trace element concentrations for marine organisms of Card Sound,
Florida as determined by Gilio & Segar, 1976. Numbers in paren-
thesis are the number of samples. (from Gilio & Segar, 1976)
... Elements fu^/r; drv ueiohr) + $t .-inrl.irri Frrnr nf fhp Mp,in
Hacrophyta
Thalassla testudinum (46)
Laurencia poitel (14)
Penlcillus capitatus (34)
Halimeda Incrassata (4)
tfcltepbora mangle <7)
' L«ave» (b)
(c)
Seedlings In water (3)
Decaying stems in water
(2)
Microphyta
Phytoplankton (d)
Epiphytes on Thalassla
blades (e)
Hacrofauna
Detritlvores and Carnivores (4)
Sponges (7)
V Fe Cu Zn Cd ?b
8.5 + 1.2
96.0 + 58
4.8 + 0.72
2.4 + 0.78
.43 + .29
.52 + .22
.48 + .41
.056 + .055
0.33
96
0.77 + 0.07
2.8 + 1.5
320 + 46
420 + 75
560 + 77
230 + 75
100 + 76
71 + 20
12 + 5.6
140 + 0
730
420
41 + 8.9
530 + 150
1.6 + 0.33(a)
12 + 2.4
1.2 + 0.17
0.70 + 0.26
1.3 + .67
5.8 + 4.6
0.81 + 0.79
0.52 + 0.46
12 + 8.0
21 + 9.4
7.4 + 0.67
3.7 + 1.5
18 + 1.3
34 + 5.1
12 + 3.5
3.7 + 1.2
3.1 + .88
2.3 + .52
2.2 + .58
8.1 + 5.9
180 + 80
150 + 59
28 + 20
24 + 9.8
D.20 + 0.021
0.20 + 0.047
0.11 + 0.012
0.16 + 0.12
.044 + .028
.24 + .11
.017 + .0059
.056 + .055
.20
.20
0.19 + 0.08
.44 + .18
0.72 + 0.16
0.59 + 0.16
1.1 + 0.21
1.2 + 0.56
.39 + .11
.79 + .23
.23 + .17
.099 +_ .0072
.33
0.59
0.39 + 0.15
.36 + .15
Notes; (a) Possible error due to a high Cu blank; (b) Live leaves; (c) Dead leaves; (d) Values for V. Fe, Cd,
and Pb, 15-fold lover than Bowen's 1966 data. Cu and Zn values determined in this study as 15-fold
lower than Bowen's values; (e) Values same as Laurencia poitel.
1233
-------
Figure 2. Trace element inventory for Card Sound, Florida.
(from Gilio & Segar, 1976)
CoTupartrent
f,
Sediment (1)
Water
Bifita
uicropnyta
Thalassia testudinum
Laurencia poitei
Halir.eda group
Penicillus group
Microphyta
epiphytes
phytoplankton
Macro tauna
sponges
Detri tivores
Carnivores
Bl.'ta Total
Bin:-.;,*. tier.cn ts ir.g/n ;
drv
3.
3.
1.
6.
8.
3.
10
0.
1.
0.
3.
V.'C<
4 x
0 x
irn
105
10C
67 x 102
1 (2)
7
1
28
1 x
18
1 x
(3)
(4)
102
(5)
IQ2
8
2
1
0
0
0
0
0
2
V
.0 x 103
.6
.4
.23
.023
.015
.96
.0092 x 102
.032
.021 x 10"2
.7
Fe
6.3 x 105
5.2 x 102
53
2.6
2.2
1.7
4.2
.064
89
0.011
150
Cu
6.7 x 102
1.2 x 102
.27
0.073
0.0068
0.0037
.21
.0034
0.19
0.002
0.76
1.
2.
3.
0.
o'.
0.
1.
.
2.
0.
7.
Zn
4 x 103
6 x 102
0
21
036
037
5
050
1
075
0
Cd
23
2.1
.033
.0012
.0016
.00034
.002
.0076 x
.030
.055 x
0.068
Pb
3.4 x 102
15
.12
.0036
.J12
.0034
^•-
.0059
10~3 .0092 x 10"2
0.014
10~3 .011 x 10"'
0.16
Notes: (1)_ Calculated from concentration data of trace elements in Card Sound. Florida from Pellenbarg
(1973) . Includes total sediment depth and total element concentrations
(2) Josselyn (1975).
(3) Calculated tiom Bach (1975). Includes Ha lineda incrassaca (5.5), H. nonile (1.5). and
H. opunta (1.4).
(4) Calculated from Bach (1975). Includes Penicillus capitatus (1.7), Rhipocephalus phoenix (0.51)
and Udotea flabellun (0.48).
(5) Gilio et al. (in prep.).
1234
-------
Table 3. Trace element biological turnover potential. Derived from
the product of annual net production data of Card Sound and
trace element concentrations. (from Gilio &'Segar, 19/b)
Trace Elenent Biological Turnover Potential nig/m /yr
Annual
Coppartiaent Net Production V Fe Cu Zn Cd Pb
g/m'/yr
Macrophyta
Thalassia (blades) 609. 5.2 200. 0.97 11. 0.12 0.44
i. 11. ia 4.6 0.13 0.37 0.0022 0.0065
Pgnicillua 4.3 (1) 0.021 2.4 0.0052 0.052 0.00047 0.0047
Ralimeda 8.6 (2) 0.021 2.0 0.0060 0.032 0.00014 .010
Microphyta
epiphytes 180. 17. 76. 3.8 27. 0.036 0.11
phytoplankton 120. 0.040 28. 1.4 22. 0.024 0.040
Macrof auna
sponges
detrltivores
carnivores
21.
6.6
(3)
0.059
0.0051
11.
0.27
0.078
0.049
0.50
0.18
0.0092
0.0043
0.0076
0.0026
Total (4) 960. 24. 320. 6.4 61. 0.19 0.12
Notes: (1) Calculated from Bach (1975) as Peniclllus capitatus (56Z), Rhlpocephalus phoenix (372), and
Udotea flabellua (7Z).
(2) Calculated fron Bach (1975) as H. incrassata (94%), and H. nontle (62).
(3) Assumes ingestlon of O.lSg m d"1 since cost members of this group are Juveniles which ingest
their own body weight/day (J^rgensen, 1966) of which 10X is net production.
(4) Excludes detritivor and camlvor group as only Initial uptake by primary production from
either water or sediment is relevant to potential turnover.
4235
-------
Table 4. Correlation Coefficients of Radionuclide Uptake by Thalassia testudinum Including Mean Values
Silver-108 Cesium-137 Manganese-54
.909
.989
.886
.872
.738
.894
Leaf
Bismuth-207
Silver-108
Cesiura-137
Manganese-54
Zinc-65
Sodium-22
Cobalt-57
Cobalt-60
Root Material Silver-108 Cesium-137 Manganese-54
Bismuth-207
Silver-108
Cesium-137
Manganese-54
Zinc-65
Sodium-22
Cobalt-57
Cobalt-60
.689
.990
.662
.883
.552
.883
• gram dry
Zinc-65
.962
.873
.958
.894
Zinc-65
.944
.648
.943
.900
weight) .
Sodium-22
.999
.895
.991
.877
.958
Sodium-22
.997
.656
.991
.869
.929
Cobalt-57
.236
.242
.272
.580
.331
.221
Cobalt-5/
. 191
.222
.188
.495
.372
. 141
Cobalt-60
.687
.664
.703
.805
.721
.675
.646
Cobalt-60
.852
.625
.846
.870
.890
.834
.476
-------
MANGANESE-54
ZINC-65
COBALT-57
COBALT-60
99
T
9) 7
fX
flfi
-2
+•> 2
•H
o
<
tH_
«>7
PL,
OH
20
25 30 32 38
20
25 30 32
< 20 25 30 32 38
is
H
8,7
g.6
•^ ^
"o5
25 30 32
Temperature (degrees CJ Temperature (degrees L.j Temperature (degrees C) Temperature (degrees C)
BISMUTH-207
SILVER-108
CESIUM-137
SODIUM-22
,r
•05
o 7
ex,
24
X
291
M8
fc*
fi6-
P*
J?sJ
S4
«
^31
^2
•r-t
fr
O
g;
M
« 7
P,^
e6,
,"5j
^
^3
?2^
< 20 25 30 32 38 < 20 25 30 32 38 < 20 25 30 32 38^^
Temperature (degrees Cj Temperature (degrees C ) Temperature(degrees C)
o
20 25 30 32 38
Temperature (degrees C)
Figure 1. . Trace metal activities In disintegration per minute per g dry wt. in blades as
a function of temperature. Each point represents mean of five samples,
-------
MANGANESE-54
ZINC-65
COBALT-S7
COBALT-60
A
8-
Is
04
>H
I*
'
s
bO
0)
|9
M8
2
1
20 25 30 32 38 < 20 25 30 32 38 < 20 25 30 32 38 < 20 25 30 32
Temperature (degrees C) Temperature (degrees C) Temperature (degrees C) Temperature (degrees C)
oo
BISMUTH-207
§9
Mg
_
•a 5
10 4
o 4
r-l
X 7
w_/ O
o
<
20 25 30 32 38
Temperature (degrees
SILVER-108
CESIUM-137
SODIUM-22
o4
r-t
3g
rtQ
o4
3,
^3
x2
20 25 30 32 38 « 20 25 30 32 38
Temperature (degrees Cp Temperature (degrees C)
Temperature (degrees C)
Figure 2. , Trace metal activities in disintegration per minute per g dry wt. in
a function ot temperature. Each point represents mean oE five samples.
-------
Figure 3. Mortality of Thaiassla as a function of temperature and salinity.
o
I .
-• ,
2 '
I? HOURS
38 JO -1.
i
-------
WASTE HEAT MANAGEMENT AND UTILIZATION:
SOME REGULATORY CONSTRAINTS *
William A. Anderson, II
P.O. Box 1535
Richmond, Virginia 23212
ABSTRACT
The need for rational management and utilization of waste
heat is undeniable. Yet conflicting governmental policies
have resulted in regulatory constraints that often fore-
close rational solutions. Effluent limitations and water
quality standards under the Federal Water Pollution
Control Act restrict use of surface waters, including
cooling lakes, for waste heat management. Other regula-
tory constraints, including provisions of the Clean Air
Act Amendments of 1977, may preclude the use of evapora-
tive cooling towers under some circumstances due to salt
drift emissions. Waste heat utilization schemes involving
clusters of industrial facilities will also encounter
environmental regulatory constraints. Provisions of
both the Air Act and the Water Act will limit industrial
concentration in any given locality. Thermal aqua-
culture may be possible under governing EPA regulations
only in blowdown streams from closed-cycle systems.
Concentrated contaminants in these blowdown streams may
make the produce unmarketable.
*This paper was not presented.
1240
-------
TECHNICAL REPORT DATA
(Please read Instructions on the reverse before completing)
REPORT NO
EPA-600/9-79-031b
2.
3. RECIPIENT'S ACCESSION NO.
T.TLEAND.U.T.TLE proceedings: Second 'Conference on
Waste Heat Management and Utilization (December
1978, Miami Beach, FL), Volume 2
5. REPORT DATE
August 1979
6. PERFORMING ORGANIZATION CODE
. AUTHOR(S)
8. PERFORMING ORGANIZATION REPORT NO
S.S.Lee and Subrata Sengupta, Compilers
9. PERFORMING ORGANIZATION NAME AND ADDRESS
University of Miami
Department of Mechanical Engineering
Coral Gables, Florida 33124
10. PROGRAM ELEMENT NO.
EHE624A
11. CONTRACT/GRANT NO.
EPA Purchase Order
DA86256J
12. SPONSORING AGENCY NAME AND ADDRESS
EPA, Office of Research and Development*
Industrial Environmental Research Laboratory
Research Triangle Park, NC 27711
13. TYPE OF REPORT AND PERIOD COVERED
Proceedings: 12/78
14. SPONSORING AGENCY CODE
EPA/600/13
is.SUPPLEMENTARY NOTES ffiRL-RTP project officer is Theodore G. Brna, MD-61, 919/541-
2683. Cosponsors are: EPRI, Florida Power and Light Co. , Univ. of Miami, U.S.
DoE, U.S. EPA, and U.S. Nuclear Regulatory Commission.
16. ABSTRACT
The proceedings document most presentations made during the Second Con-
ference on Waste Heat Management and Utilization, held December 4-6, 1978, at
Miami Beach, FL.,Presentations were grouped by areas of concern: general, utili-
zation, mathematical modeling, ecological effects, cooling tower plumes, cooling
towers, cogeneration, cooling systems, cooling lakes, recovery systems, aquatic
thermal discharges, and atmospheric effects. Causes, effects, prediction, monit-
oring, utilization,, and abatement of thermal discharges were represented. Utiliza-
tion was of prime importance because of increased awareness that waste heat is a
valuable resource. Cogeneration and recovery systems were added to reflect this
emphasis.
17.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
b.lDENTIFIERS/OPEN ENDED TERMS
COSATI Field/Group
Pollution
Heat Recovery
Management
Utilization
Mathematical Models
Ecology
Cooling Towers
Plumes
Pollution Control
Stationary Sources
Cogeneration
Cooling Lakes
Thermal Discharges
Atmospheric Effects
13B 07A,13I
20M,13A 21B
05A
14B
12A
06F
18. DISTRIBUTION STATEMENT
Release to Public
19. SECURITY CLASS (ThisReport)
Unclassified
21. NO. OF PAGES
637
20. SECURITY CLASS (Thispage)
Unclassified
22. PRICE
EPA Form 2220-1 (9-73)
1241
« U.S. GOVERNMENT PRINTING IWI.CB—1979/640-013/3934
------- |