U.S. Environmental Protection Agency Industrial Environmental Research     EPA-600/7-78-069
Office of Research and Development  Laboratory                  .. ^ft^o
                 Research Triangle Park. North Carolina 27711 Apfll 1978
      MINIPLANT STUDIES
      OF  PRESSURIZED
      FLUIDIZED-BED
      COAL COMBUSTION:
      Third Annual  Report
      Interagency
      Energy-Environment
      Research and Development
      Program Report

-------
                 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 INTERAGENCY ENERGY-ENVIRONMENT
RESEARCH AND DEVELOPMENT series. Reports in this series result from the
effort funded under the 17-agency Federal Energy/Environment Research and
Development  Program. These studies relate to EPA's mission to protect the public
health and welfare from adverse effects of pollutants associated with energy sys-
tems. The goal of the Program is to assure ihe rapid development  of domestic
energy supplies in an environmentally-compatible manner by providing the nec-
essary environmental data and control technology. Investigations include analy-
ses of the transport of energy-related pollutants and their health and ecological
effects;  assessments of, and development of,  control  technologies for energy
systems; and  integrated assessments of a wide range of energy-related environ-
mental issues.
                        EPA REVIEW NOTICE
This report has been reviewed by the participating Federal Agencies, and approved
for publication. Approval does not signify that the contents necessarily reflect
the views and policies of the Government, 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 Informa-
tion Service, Springfield, Virginia 22161.

-------
                                            EPA-600/7-78-069
                                                   April 1978
MINIPLANT STUDIES OF PRESSURIZED
 FLUIDIZED-BED COAL COMBUSTION:
               Third Annual Report
                             by

                  R.C. Hoke, R.R. Bertrand, M.S. Nutkis,
                  L.A. Ruth, M.W. Gregory, E.M. Magee,
            M.D. Loughnane, R.J. Madon, A.R. Garabrant, and M. Ernst

                 Exxon Research and Engineering Company
                          P.O. Box 8
                      Linden, New Jersey 07036
                      Contract No. 68-02-1312
                    Program Element No. EHE623A
                  EPA Project Officer: D. Bruce Henschel

                Industrial Environmental Research Laboratory
                  Office of Energy, Minerals, and Industry
                   Research Triangle Park, N.C. 27711
                          Prepared for

               U.S. ENVIRONMENTAL PROTECTION AGENCY
                  Office of Research and Development
                      Washington, D.C. 20460

-------
                                  ABSTRACT
     The pressurized fluidized bed combustion of coal and regeneration of
spent sorbent were studied in the continuous 480 Ib coal/hr (220 kg/hr)
"miniplant" FBC unit.  The effect of coal sulfur content and the use of pre-
calcined limestone sorbent on control of SC>2 emissions was studied.  NOX
emissions were well within the current new source performance standard for
NOX.  Particulate emissions in the flue gas after passage through cyclone
cleaners exceeded the new source performance standard.

     The continuous operation of the combustor and regenerator sections was
demonstrated in a 125 hour run.  SC>2 emissions in the combustor flue gas were
less than the new source performance standard at all times during the run.
Makeup sorbent fed to the combustor was about 25% of the rate which would
have been required in once through operation to maintain the S02 emissions
at the same overall average level.

     A granular bed filter was installed in the miniplant flue gas stream
and initial shake down was completed in a 24 hour continuous run.  The
minimum particulate concentration measured in the filter outlet gas was about
0.1 g/m^ (0.05 gr/SCF), somewhat higher than anticipated.

     A program to provide a comprehensive analysis of all potentially harmful
emissions from a pressurized FBC unit began in the miniplant.   Analysis of
materials including those present in trace concentrations was completed.

     The smaller 28 Ib coal/hr (13 kg/hr) batch unit was modified to operate
in a continuous fashion.

     This report is submitted in fulfillment of Contract Number 68-02-1312
by Exxon Research and Engineering Company under sponsorship of the
Environmental Protection Agency.  Work was completed in August, 1977.
                                     iii

-------
                            TABLE OF CONTENTS
Abstract

List of Figures                                                         v

List of Tables                                                         ix

Acknowledgements                                                        x

Sections

   I      Summary                                                       1

  II      Introduction                                                 11

 III      Combustion Studies                                           15

          Equipment, Materials, Procedures                             15
          Experimental Results and Discussion                          46

  IV      Regeneration Studies                                         65

          Equipment, Materials, Procedure                              65
          Experimental Results and Discussion                          70

   V      Granular Bed Filtration Studies                              84

          Equipment, Procedures                                        85
          Experimental Results and Discussion                         100

  VI      Modification of Batch Unit                                  115

          Combustor Section                                           115
          Regenerator Facilities                                      121

 VII      Comprehensive Analysis                                      126

VIII      Analysis of Desulfurization Data                            127

  IX      Continuing Studies                                          150

   X      References                                                  153

  XI      List of Publications                                        155

 XII      Appendices                                                  157
                                 iv

-------
                            LIST OF FIGURES






                                                                    Page




 II-l    Pressurized Fluidized Bed Coal Combustion System            12




 II-2    Exxon Fluidized Bed Combustion Miniplant                    14




III-l    Exxon Fluidized Bed Combustion Miniplant (Schematic)        16




III-2    Coal and Limestone Feed System                              17




III-3    Combustor Vessel                                            19




III-4    Cooling Coils After 600 Hour Exposure                       20




III-5    Liquid Fuel System                                          21




III-6    Liquid Fuel Injector Nozzle                                 23




III-7    Combustor Fluidizing Grid                                   24




III-8    Combustor Corrosion Probe Location                          25




III-9    Combustor Lower Corrosion Probe Section                     26




111-10   Combustor Lower Corrosion Probe and Bed Sampling Probe      27




III-ll   Combustor Upper Corrosion Probe Section                     29




111-12   Combustor Corrosion Probe                                   30




111-13   Westinghouse Erosion Probe                                  32




111-14   Combustor Bed Sampling Probe                                33




111-15   Sorbent Utilization Vs. Time                                34




111-16   Miniplant Particulate Sampling System                       36




111-17   Combustor Escentric Acurex Probe Section                    38




111-18   Acurex Probe                                                39




111-19   Coal Particle Size Distribution                             41




111-20   Sorbent Particle Size Distribution                          44




111-21   S02 Retentionvwith Dolomite Sorbent                         47




111-22   S02 Retention with Precalcined Lir.estone                    49

-------
                      LIST OF FIGURES (Continued)
                                                                    Page
111-23   Effect of Low Temperature Operation on S02 Retention        50

111-24   Correlation of NO  Emissions                                53
                          X

111-25   NO  Vs CO in Flue Gas                                       55
           X
111-26   Calculated Combustion Efficiency Vs Observed                62
         Combustion Efficiency

111-27   Secondary Cyclone Efficiency Vs. Particle Size              64

 IV-1    Miniplant Solids Transfer System (Schematic)                68

 IV-2    Miniplant Solids Transfer System                            69

 IV-3    Combustor and Regenerator Bed Heights During                75
         Demonstration Run

 IV-4    Total Inventory of Sorbent in the Combustor and             76
         Regenerator during Demonstration Run

 IV-5    Combustor S0« Emissions During Demonstration Run            78

 IV-6    S02 Emission Vs. Ca/S Ratio                                 79

  V-l    Single Bed Test Rig                                         86

  V-2    Single Bed Test Rig Schematic                               87

  V-3    Schematic of Single Ducon Filter Bed                        88

  V-4    Granular Bed Filter Operation                               89

  V-5    Ducon Filter Element and Shroud                             90

  V-6    Original Exxon Filter Element and Shroud                    91

  V-7    Ducon Sonic Filter Schematic                                93

  V-8    Schematic of a Single Exxon Filter Bed                      94

  V-9    Modified-Exxon Filter Element                               95

  V-10   Modified Filter Bed                                         96

  V-ll   Granular Bed Filter Pressure Vessel and Structure           97
                                 vi

-------
                       LIST  OF FIGURES  (Continued)


                                                                     Page

   V-12    Granular  Bed Filter  Pressure  Vessel Interior                99

   V-13    Test  Rig  Results - AP Vs.  Time                              101

   V-14    Ducon Filter with  Plugged  Screens                          103

   V-15    Effect of Blow Back  on Filter Pressure Drop                109
          Run 55

   V-16    Modified  Exxon Filter with                                 111
          Plugged Inlet Retaining Screens

   V-17    Flue Gas  Temperature Profile                                113

   V-18    Schematic of the Ejector Blow Back System                  114

  VI-1    Schematic of Modified Batch Unit                           116

  VI-2    Bench Unit Cyclone Assembly                                118

  VI-3    Bench Unit Coal Injector Vessel                            120

  VI-4    Bench Unit Regenerator Fluidizing Grid                     122

  VI-5    Bench Unit Regenerator Cyclone                              125

VIII-1    Ln (1-SR) Vs. Gas  Phase Residence Time                     129

VIII-2    Sulfur Removal Efficiency Vs. Ca/S                         130

VIII-3    Sulfur Removal Efficiency @ 2 Sec. Vs. Ca/S                131

VIII-4    Sulfur Removal Efficiency @ 2 Sec. Vs. Temperature         133

VIII-5    Effect of Temperature on Desulfurization Rate Constant     134

VIII-6    Comparison of Set  Vs. Calculated Ca/S Effects              135

VIII-7    Sulfation Rate Constant Vs. Sorbent Utilization            141

VIII-8    SO  Retention @ 2  Sec. Vs. Ca/S Set                        142

VIII-9    Time for a Particle  to Reach  Calcium Utilization           143
          Level in Atmosphere  of 2500 ppm SO  at 900°C
                                  vii

-------
                      LIST OF FIGURES (Continued)
                                                                     Page
VIII-10   Contribution of Particles in Bed for Various Time          145
          Periods to Steady State Value of Given Parameters

VIII-11   SO  Retention @ 2 Sec, 900°C Vs. Ca/S Molar Ratio          147

VIII-12   Time in Combustor Vs. Ca/S Ratio                           148
                                 viii

-------
                             LIST OF TABLES


                                                                      Page

 III-l    Coal Composition                                             42

 III-2    Sorbent Composition                                          45

 III-3    Sorbent Requirements                                         52

 III-4    Particulate Sampling Summary I (Before Modification          56
          to Sampling System)

 III-5    Particulate Sampling Summary II (After First                 57
          Modification to Sampling System

 III-6    Comparison of Particulate Loadings Measured by               58
          Total Filter (Exxon) and SASS Train (Battelle)

 III-7    Sizing of Particles Obtained on Balston Filters              58
          Before Modification

 III-8    Individual Filter Results (Runs 50 to 59)                    59

 III-9    Sizing of Particles Obtained on Balston Filters              60
          After Modification

 IH-10   Comparison of Particle Size Distributions                    60
          Measured by Three Different Methods

 III-ll   Second Stage Cyclone Lock Hopper Particle                    63
          Size Distributions

  IV-1    Operating Conditions During Demonstration Run                71

  IV-2    Feed Rates of Makeup Limestone to Combustor During           74
          Combustor-Regenerator Demonstration Run

  IV-3    Analyses of Bed Material Discharged from Combustor           81
          and Regenerator After Demonstration Run

  IV-4    Sulfur Balance Combustion-Regeneration                       82
          Demonstration Run

   V-l    Granular Bed Filter Run Summary for Modified                106
          Filter Elements

VIII-1    Comparison of Rate Constants                                146
                                   ix

-------
                              ACKNOWLEDGEMENTS
     Many individuals at Exxon Research and Engineering Company played major
roles in the conduct of this program.  The authors wish to acknowledge the
contributions made by them and to express their gratitude.  We particularly
wish to acknowledge the efforts of H. R. Silakowski, the miniplant operations
supervisor.  His contributions played a large part in the successful opera-
tion of the miniplant.  We also wish to acknowledge the efforts of the
operating and mechanical crews, A. Altobelli, J.  Bond, R. Burakiewicz, D.
Duffy, J. Fowlks, T. Gaydos, E. Hellwege, F.  Huber, M. Moroski, T. Morrison,
S. Pampinto, J. Sansone, R. Schroeder, W. Spond,  T. Sutowski,  L. Tucker and
G. Walsh.  We also wish to thank our math clerk,  S. Walther, W. Dravis and
N. Bissoni of the Mechnical Division G. Milliman of the Analytical Division,
V. S. Engleman, and G. A.  Gagliardo for their support.  A special acknowledge-
ment goes to N. Malinowsky who typed this report.

     The personnel of the Industrial Environmental Research Laboratory of the
EPA have been most helpful and deserve special thanks.  We wish to express
our gratitude for the help of D. B. Henschel, the EPA Project  Officer,
P. P. Turner and R. P. Hangebrauck.

-------
                                  SECTION I

                                   SUMMARY
     The pressurized fluidized bed combustion of coal  (PFBC) and regeneration
of sulfated SC>2 sorbent were studied in the continuous miniplant unit.  In
the combustion program, the effect of operating conditions on S02, NOX, and
particulate emissions from the miniplant was studied.  In this latest series
of combustion tests, the effect of burning a higher sulfur Illinois No. 6
coal was determined, the effect of using precalcined limestone sorbent and
operation at very low combustor "turndown" temperatures were studied.  Sorbent
levels required to meet S02 emission standards were determined.  Combustion
efficiency and cyclone collection efficiency were also measured.

     The regenerator and combustor sections of the miniplant were operated
continuously for 125 hrs and the operability of the system was demonstrated.
The drop in activity of the regenerated sorbent with time, and the SC>2 con-
tent in the regenerator off gas were measured.

     A granular bed filter was installed on the miniplant flue gas and initial
shakedown was completed.  A comprehensive analysis of all potentially hazardous
emissions from the miniplant was completed.  The smaller, semi-batch unit was
not operated during this period, but was modified to operate in a continuous
fashion.

COMBUSTION STUDIES

     The miniplant combustor consists of a refractory lined vessel 10 m (33
ft) high with an inside diameter of 32 cm (12.5 in).  A number of vertical
water-cooled tubes are mounted in the combustor to remove the heat of com-
bustion.  Two new sections of the combustor were installed to hold air-cooled
tubes used for materials testing purposes.  One of the sections holds tubes
immersed in the fluidized bed, the other holds tubes mounted in the freeboard
above the fluidized bed.

     Premixed coal and sorbent are injected into the combustor a single point
28 cm (11 in) above the fluidized bed support grid.  The combustor is capable
of operating at pressures up to 1000 kPa (10 atm), at temperatures up to the
ash agglomeration temperature of the coal (usually less than 980°C), at
superficial velocities of up to 3 m/s (10 ft/sec) and with expanded beds of
up to 6.1 m (20 ft).  The maximum design coal feed rate is 200 kg/hr (480
Ib/hr).  Flue gas leaving the combustor passes through two cyclones in series
to remove most of the particulate matter.  A granular bed filter was recently
installed and piped into the flue gas system such that the flue gas leaving
the second stage cyclone can either be sent to or by pass the filter.
Particulates captured in the first cyclone are recycled to the combustor to
improve combustion efficiency.  Particulates captured in the second stage
cyclone and the granular bed filter are rejected through lock hoppers.
Spent sorbent is also rejected from the combustor through a lock hopper
system to maintain a constant bed level in the combustor.

-------
     Runs were made with an Eastern bituminous Pittsburgh seam coal  (Champion)
containing 2% sulfur screened to a particle size distribution of 200 to
2400 microns, and an Illinois No. 6 coal containing 4.2% sulfur also screened
to a particle size distribution of 200 to 2400 microns.  Two sorbents were
used:  a Virginia limestone (Grove No. 1359) and an Ohio dolomite (Pfizer No.
1337).  Both were screened to a size range of 840 to 2400 microns.

     Operational performance of the combustor continued to be good.  As of
July 1977, over 1500 hrs of coal combustion time was accumulated.  Additional
modifications were made to the combustor which further improved performance.
The baffled vertical cooling coils performed very well, showing no sign
of erosion after about 600 hrs of running time.  A new fluidization grid was
fabricated with better distribution of the cooling water flow through the
grid.  The refractory lining in the second stage cyclone began to fail and
was replaced with a more erosion resistant material.  A sampling system was
also developed which permits the extraction of solids from the combustor
during the run.

     S02 retention results were obtained using 4.2% sulfur Illinois No. 6 coal
and Pfizer No. 1337 dolomite.  S02 retention data plotted against the Ca/S
molar ratio fit reasonably well with the data obtained with 2% sulfur Eastern
coal.  This suggests that the sulfation reaction is first order in 862 con-
centration and is consistent with results published by other laboratories.
S02 retention was also measured using precalcined Grove No. 1359 limestone
sorbent.  Calcination was carried out in the combustor burning natural gas at
930 kPa  (9 atm) and 870°C (1600°F).  The calcined sorbent was then removed
and fed with the coal during a subsequent series of runs.  The precalcined
limestone was found to be as active, at an equivalent Ca/S ratio, as dolomite
and was much more active than "raw" limestone.  The activity was also main-
tained at temperatures as low as 865°C, where limestone would be expected
to be present as the carbonates.  The precalcined limestone was also active
at a very low "turndown" temperature (760°C) where previously, "raw" limestone
was shown to be completely inactive.  At 760°C, precalcined limestone is
slightly less active than at temperatures in excess of 865°C.  It is, however,
still as active as dolomite, at an equivalent Ca/S ratio, even though the
low temperatures would strongly favor the formation of CaCOo rather than
CaO.  The high activity is believed due to the formation of very large
pores during precalcination.  The pores are apparently large enough that
carbonation of the stone does not prevent diffusion of S02 into the interior
of the particles.  Sorbent requirements were then estimated for precalcined
limestone and compared to dolomite and limestone requirements.  Precalcined
limestone is more effective, on a weight basis, than either dolomite or
limestone.  The weight requirements, expressed on an uncalcined limestone
basis,  are half the dolomite requirements and as little as 40% of the require-
ment for "raw" limestone.

     NOX emissions were found to follow the same trend line developed in
earlier studies.   The NOX emissions vary from 50 to 200 ppm or 0.04 to 0.17 g
(as N02)/MJ (0.1 to 0.4 Ib/M BTU).  The primary variable affecting NOX emis-
sion is excess air (or 02 content in the flue gas).  Temperature over the
range of 670 to 940°C (1250 to 1750°F) had a secondary effect.  The emissions
are well below the EPA new source performance standard of 0.3 g/MJ (0.7 Ib/M
BTU)  and have an average value of 0.09 g/MJ (0.2 Ib/M BTU) at 15% excess air,

-------
the level most likely to be used in commercial practice.   Indications  that
NOX could be reduced even further by two stage combustion  were made  during
an upset period of one run.  In this case, oxygen content  in  the  flue  gas
dropped to very low levels, CO content increased to 1800 ppm  and  as  it did,
the NOX content dropped to 10 to 20 ppm.

     Particulate emissions in the flue gas after two stages of cyclone cleanup
averaged about 2.3 g/m^ (1 gr/SCF) with a mass mean particle  size of 3 to 6
microns.  The second stage cyclone was found to have an overall efficiency of
85% with a 50% cut point between 3 and 3.5 microns.

     Carbon combustion efficiency results were satisfactorily correlated
using multiple regression.  Temperature and excess air were shown to be the
most significant variables.  Two other variables which had a  significant
effect were the cube of the residence time and the cross product  of  tempera-
ture and excess air.

REGENERATION STUDIES

     The regenerator consists of a refractory lined vessel with an inside
diameter of 22 cm  (8.5 in) and an overall height of 6.7 m  (22 ft).  Gaseous
fuel is burned in  a plenum below the fluidized bed to achieve the reaction
temperature.  Additional fuel is injected directly into the fluidized  bed
just above the fluidizing grid to create a reducing zone in which the  CaSO^
reduction reaction occurs.  Supplementary air is injected  directly into the
bed at a higher elevation to create an oxidizing zone.  The oxidizing  environ-
ment at the top of the bed assures high selectivity to CaO, the desired pro-
duct of the regeneration reaction, by minimizing the formation of CaS, an
undesired by product.

     The successful operation of the coupled regeneration/combustor system
was demonstrated in a 125 hr uninterrupted run.  The miniplant was operated
with limestone sorbent continuously recirculating between  the combustor and
regenerator until  the run was voluntarily terminated.  The major  purpose
of the run was to  demonstrate that the system could operate continuously for
100 hrs.  Operating conditions were deliberately chosen to be conservative
in order to maximize the chance of reaching this goal.  All conditions were
held constant except the feed rate of fresh limestone into the combustor.
This was adjusted  to maintain constant bed levels in the combustor and
regenerator.  Used sorbent was not planned to be removed from the system
during the run unless the S02 content in the combustor flue gas exceeded the
EPA new source performance standard of 1.2 Ib/M BTU.  At that point, used
sorbent was planned to be removed and the fresh sorbent feed  rate would be
increased.  Pressure in the two vessels was controlled at  a somewhat lower
level  (about 760 kPa (7.5 atm)) during most of the run in  order to increase
fluidization velocity in the regenerator to a more realistic  level and to
avoid agglomeration problems.  Regeneration temperature was also  held  con-
stant at a lower level, 1010°C (1850°F) to minimize the chances of bed
agglomeration.  Combustion temperature was 900°C (1650°F), Eastern coal and
Grove No. 1359 limestone were used.

     For the first 24 hours of the run, the regenerator was operated in
oxidizing conditions in order to establish base line operating conditions.
Sorbent recirculated between the combustor and regenerator during this

-------
 period.   Subsequently,  reducing  conditions were  established by  increasing
 the flow of supplementary fuel to  the  regenerator.  Operation of  the  system
 continued,  uninterrupted,  for the  next 100 hours.  Operation of the com-
 bustor-regenerator during the extended run was exceptionally smooth.  A hot
 spot did develop in the regenerator  just  above the grid  in which  the  tem-
 perature was 200 to 250°C (450°F)  higher  than the rest of the regenerator
 The temperature profile was  improved somewhat by decreasing the pressure to
 760 kPa, thereby increasing  the  fluidization velocity.   Some signs of
 agglomeration were noted in  the  bed  after completion of  the run.  This could
 have caused the hot spot to  form.

      The limestone addition  rate was initially set at a  Ca/S ratio of 0.74.
 However, solids inventory in the combustor and regenerator began  dropping
 and the  feed rate was  increased  to a Ca/S ratio  of 1.3.  Part of  the  drop  in
 bed levels  was due to  the inadvertent  removal of some combustor solids early
 in the run.  After the  bed levels  increased, the Ca/S was dropped to  1.06,
 then 0.46 and finally was decreased  to 0  for the last 40 hours  of the run.

      S02 emissions from the  combustor  leveled out at 550 ppm (about 1.2 Ib/M
 BTU) during the first  24 hours of  the  run while  the regenerator was opera-
 ting under  oxidizing conditions.   Emissions fell rapidly to less  than 200 ppm
 when reducing (regenerating) conditions were established in the regenerator.
 S02 emissions from the  combustor increased as the run progressed  and  as the
 sorbent  deactivated. At the end of  the run, after 100 hours of regeneration,
 the emissions were about 550 ppm.  If  the run had gone longer,  removal of
 used sorbent and an increase in  the  fresh Ca/S ratio would have been  required
 to maintain the S02 emissions at this  level.  During the 100 hours of regen-
 eration, the sorbent underwent about 15 cycles of combustion and  regeneration.
 The average Ca/S ratio  during the  run  was 0.55,  although as mentioned above,
 it was varied from 0 to 1.3.  This represents about a four fold reduction in
 the limestone feed rate which would  have  been required in a once  through
 system to control the S02  emissions  to the average level measured in  this run.

      The concentration  of  S02 in the regenerator off gas was nearly steady
 throughout  the run and  averaged  0.53 mole percent (dry basis).  This  is very
 close to the concentration predicted by a sulfur mass balance based on the
 feed rate and sulfur content of  the  coal  entering the combustor.  The cal-
 culated  equilibrium concentration  at the  operating conditions of  the  regen-
 erator was  2.9 percent;  hence, higher  S02 levels would probably have  been
 achieved by burning in  the combustor more coal of a high sulfur content.
 The  sulfur  balance of the  run was  104%, an acceptable level.

 GRANULAR BED FILTRATION STUDIES

      The  objective of the  granular bed filtration program was to  determine
 if such  a filter  could  reduce particulate concentrations in the flue  gas, at
high  temperature  and high  pressure,  to a  level which would prevent damage to
 the  gas  turbine as well  as satisfy environmental requirements.  Tentative
estimates of  the  particulate concentration which could be tolerated by the
 turbine  range  from 45 to 1 mg/m3 (0.02 to 0.004  gr/SCF), and also depend on
the particulate size distribution.   Current environmental standards requires
the particulate emission to be less  than  approximately 100 mg/m3  (0.05
gr/SCF).   To meet  these  requirements,  a particulate removal system consisting
of two cyclones followed by a high efficiency third stage device  such as a

-------
granular bed filter is now envisioned.  The granular bed filter test program
is intended to determine if the filter can achieve these particulate emission
targets, if the performance can be maintained, to uncover operating problems
and to measure the long term life of the filter hardware.

     A granular bed filter was purchased from the Ducon Company and evaluated.
The Ducon-type granular bed filter consists of a number of small beds packed
with suitable granular filter media such as alumina, quartz, etc.  A stack
of the filter beds form a single filter element.  A number of filter elements
can be used depending on the volume of gas to be filtered.  Dirty gas passes
through inlet screen sections down into the filter beds immediately below the
screen sections.  Clean gas from the beds is collected in a manifold in the
interior of the element and then passes to the clean gas outlet system.  As
the filtration step proceeds, the pressure drop across the element increases
and eventually the element must be cleaned by the reverse flow of clean gas.
This "blow back" occurs by flowing clean gas in reverse direction through the
outlet gas manifold, up through each filter bed and out through the screens.
The function of the screens is to retain the filter media during the blow
back step, keeping it inside the filter beds, while allowing the fine par-
ticulates removed from the filter media by the blow back gas to pass through.
The fine particulate then settles outside the filter elements and is collected
at the bottom of the vessel containing the filter elements.

     A pressure vessel 2.4 m (8 ft) in diameter by 3.4 m (11 ft) high was
installed on the miniplant to house up to four filter elements.  Each element
is contained within a shroud inside the pressure vessel.  Inlet gas is piped
to each shroud, passing through orifices which measure the flow rate to each
filter element.  Clean gas exits each element and fills the interior of the
pressure vessel.  Particulates removed from the filter elements during blow
back impinge on the inside surface of the shrouds, fall to the bottom and
are collected in lock hoppers.  The pressure vessel and filter elements are
heated prior to the start of a run with a gas-fired burner to a temperature
greater than the dew point of the flue gas.

     A number of high temperature runs were then attempted but the pressure
drops across the filter were extremely high and' all attempts at blow back
were unsuccessful.  Inspection of the filter elements after each of these
runs showed that a hard filter cake had formed on the inlet retaining
screens.  The filter medium was usually particulate free indicating very
little penetration through the screens.  The initial tests were made before
the preheat burner was installed and it was thought that the plugging was
caused by condensation of moisture during start up.  However, runs made
after the preheat burner was in operation were also terminated by screen
plugging problems.  As a result, this filter design was deemed unacceptable
for this application.

     Discussions with the Ducon Company led to the design and fabrication
of modified filter elements.  Ducon suggested removing the inlet screens
and designing the individual filter beds with more freeboard to prevent
entrainment of the filter media during blow back.  It was also suggested
that a gas distribution plate be used at the bottom of each bed to assure
good distribution of the blow back air.

-------
      Operability of  the modified  filter system was demonstrated.  The end
 of  the initial  shakedown  phase was  signified by the successful completion of
 a 24  hour demonstration run.  This  run was preceded by a number of shorter
 duration runs used to  establish suitable operating conditions for the demon-
 stration run.   These runs were successful in that filtration, ability to blow
 back, ability to maintain low pressure drops and collection of particulates
 after blow back were demonstrated.  Collection efficiencies of 90 to 95% were
 measured for the first few hours  of the runs based on outlet particulate con-
 centrations of  about 100  mg/m3 (0.05 gr/SCF).  Operation for up to 24 hours
 was also demonstrated  with no significant increase in base line pressure
 drop  across the filter.   Blow back  was usually required every 10-20 minutes
 during which time the  filter pressure drop had increased 14 kPa (2 psi)
 above its base  line  value.  A range of blow back conditions were used to
 restore the base line  pressure drop.  Blow back durations ranged between 2
 and 30 seconds  and superficial velocity between 0.15 and 0.75 m/s (0.5 and
 2.5 ft/s). The quantity  of blow  back air used ranged from 1 to 5% of the
 filtered gas rate.   Filtration velocities generally ranged between 20 and 24
 m/min (60 and 80 ft/min).  Filter media consisting of 300 to 600 micron
 quartz particles were  tested.

      The particulates  passing the filter had a mass median size of about 3
 microns with about 10% larger than  10 microns.

      A number of problems were defined during the shakedown and operation of
 the filter. Demonstrated particulate outlet concentrations are still higher
 than  the tentative gas turbine inlet requirements.  However, firm turbine
 requirements have not  been set as yet and it may be too early to reach any
 conclusions regarding  the suitability of the filter to protect a gas turbine.
 The lower outlet particulate loadings of 100 mg/m3 (0.05 gr/SCF) meet the
 current EPA emission standards.   However, in all runs, it was observed that
 the outlet loadings  increased with  time.  It has not as yet been demonstrated
 that  the EPA emission  standard can  be met for more than a few hours of
 operation.

      Another observation  has been the retention of a significant portion of
 the filtered particulates in the  filter beds.  This has amounted to as much
 as  30%  of the weight of the filter  media.  The retained particulates were
 also  found  to be  uniformly distributed through the filter beds instead of
 forming  a layer  on the top of the beds as had been expected.  Distribution of
 particulates throughout the bed was probably responsible for much of the
 particulate penetration through the bed, and hence the observed increase of
 outlet loadings with time.

      Loss of filter media during  blow back was another reoccuring problem.
 Since inlet retaining  screens were  found to be susceptible to plugging,
better control of the  blow back air rate must be established to minimize the
losses.

     Another potential problem with the current design is its vulnerability
to upsets.  If upsets  occur, such as bed plugging or loss of filter media,
the operating problems caused by  such upsets usually require shut down of
the system.  It is usually not possible to take corrective action which
restores good operation.   Another problem which may be unique to the mini-
plant  was the interaction of the  granular bed filter with the rest of the

-------
FBC system during blow back cycle.  An increase in system pressure was noted
during blow back resulting in problems with the coal feed system which is
controlled by the differential pressure between the coal feed vessel and
combustor.  Modifications to the coal feed control system were required to
minimize the effects.

COMPREHENSIVE ANALYSIS OF EMISSIONS

     A program was begun to carry out a comprehensive analysis of all poten-
tially hazardous materials in the effluents from the miniplant unit.  A series
of runs was completed in which samples of the input streams as well as the
effluents were taken and analyzed.  This work was done in cooperation with
Battelle Columbus Laboratory, the contractor coordinating all such activities
for the EPA.  The results of the test will be summarized in a report to be
issued jointly by Battelle and Exxon Research and Engineering Company.

MODIFICATION OF THE BENCH SCALE UNIT

     The bench scale unit which was originally designed to operate in a batch
or semi-batch fashion was modified to permit continuous operation.  The
modifications will permit the continuous feeding of coal and sorbent to the
combustor and the continuous withdrawal of spent sorbent.  Solids captured in
the first stage cyclone will be able to be recycled to the combustor if
desired.  The regenerator will also be capable of continuous feeding of
sulfated sorbent and withdrawal of regenerated sorbent.

ANALYSIS OF DESULFURIZATION DATA

     Additional analysis of flue gas desulfurization data measured in the
miniplant and batch units was carried out.  S02 retention results have been
correlated as function of the Ca/S molar ratio calculated from a sulfur
balance and based on measurements of sulfur contents in the flue gas and the
spent sorbent.  This was done to minimize the effect of mechanical dif-
ficulties with the coal/sorbent blending equipment and an incomplete approach
to steady state conditions.  The desulfurization data were again analyzed
using the Ca/S ratio set on the solids blending equipment but corrected for
variations in combustor temperature and gas phase residence time.  Residence
time corrections were based on the use of a first order rate expression which
had been developed previously and more recently verified.  Temperature cor-
rections were based on a calculated activation energy of 13 kcal/mole.
Correcting for residence time and temperature variations gave a good correla-
tion of S02 retention with Ca/S ratio set on the blender.  The correlation
agreed reasonably well with that based on the Ca/S ratio calculated from
analysis of gas and solids.

     S02 retention results measured in the batch unit and the continuous
miniplant were also compared after correcting for the effect of solids
residence time distribution in the miniplant.  Reaction rate constants were
calculated from the batch unit results and were used to calculate the S02
retention as a function of Ca/S ratio for a continuous unit.  It was assumed
that the gas moved in a plug flow fashion through the miniplant combustor,
that the solids were well mixed and the residence time distribution of the
solids could be described by an exponential distribution function.  The

-------
 corrected batch results  were  then  compared  to results measured  in  the mini-
 plant.   The corrected  batch data did not  exactly correspond  to  the miniplant
 data,  but were still quite representative of desulfurization performance with
 dolomite sorbents.

      Residence time distribution of solids  in the combustor  can also be used
 to gain some insight into the factors which control desulfurization perfor-
 mance.   It was estimated that 15%  of the  sorbent particles in the  combustor
 at typical operating conditions have a residence time between 0 and 1 hour but
 because of the higher  activity of  fresh sorbent, these particles account
 for 55% of the SC>2  retention  achieved in  the combustor.  Also,  85% of the S02
 removal occurs on particles which  have been in  the combustor for 3 hours or
 less,  even though these  particles  represent only 31% of the  combustor
 inventory.  It is this high level  of effectiveness of fresh  sorbent particles
 which results in a  fairly rapid approach  to a steady state SC>2  concentration
 after a change in the  Ca/S feed ratio.

 CONTINUOUS STUDIES

      The particulate removal  program will continue with the  objective of
 optimizing the performance of the  granular  bed  filter.  Particulate removal
 efficiency must be  improved,  loss  of filter media during blow back decreased
 and the blow back made more effective.  The use of transparent  models to
 observe the mechanics  of the  filtration and blow back steps  will begin in an
 attempt to learn more  of what is occuring in the filter.  Long  term testing
 with the filter will begin in conjunction with  the Department of Energy (DOE)
 sponsored program to evaluate gas  turbine materials.  The DOE hot  corrosion
 testing on the miniplant will be conducted  under a cooperative  agreement
 between EPA and DOE.

      Following the  granular bed filter tests, one or possibly two  alternate
 particulate control devices will be evaluated.  The choice of the  devices
 will be made after  consultation with the  EPA.   Currently, devices  such as a
 high temperature electrostatic precipitator or  a high temperature  bag filter
 are  being considered.

     The  comprehensive analysis program will be extended to  include operations
 with the  regenerator.  More quantitative  tests  (Level 2 and  Level  3 tests)
 will be carried  out based on  the results  of the initial screening  tests
 (Level  1  tests).  Two high temperature, high pressure particulate  sampling
 systems capable  of  sample particulates at temperatures up to 870°C will be
 fabricated  and  installed.

     The  regeneration program will consist  in a series of runs  of  two to five
 days duration aimed at generating sufficient data to evaluate the  economic
 feasibility  of regeneration.  The tests will determine the make up sorbent
 requirements and will indicate the level  of S02 in the regenerator off gas
which can be obtained.

-------
     Future work in the bench scale combustor unit will consist in evaluating
various coals and sorbents.  NOX control studies will also be made.  Regenera-
tion studies will determine the activity maintenance of sorbent recycled
between the combustor and regenerator while operating under various regenera-
tion conditions.  The use of coal as fuel for the regenerator will be
included.

-------
10

-------
                                 SECTION II

                                INTRODUCTION
     The pressurized fluidized bed combustion of coal is a new combustion
technique which can reduce the emission of S02 and NOX from the burning of
sulfur-containing coals to levels meeting EPA emission standards.  This is
done by using a suitable S02 sorbent such as limestone or dolomite as the
fluidized bed material.  In addition to emissions control, this technique
has other potential advantages over conventional coal combustion systems
which could result in a more efficient and less costly method of electric
power generation.  By immersing steam generating surfaces in the fluidized
bed, the bed temperature can be maintained at low and uniform temperatures in
the vicinity of 800 to 950°C.  The lower temperatures allow the use of lower
grade coals (since these temperatures are lower than ash slagging tempera-
tures) , and also decrease NOX emissions.  Operation at elevated pressures, in
the range of 600 to 1000 kPa, offers further advantages.  The hot flue gas
from a pressurized system can be expanded through a gas turbine, thereby
increasing the power generating efficiency.  Operation at the higher pressure
also results in a further decrease in NOX emissions.

     In the fluidized bed boiler, limestone or dolomite is calcined and
reacts with S02 and oxygen in the flue gas to form CaSO^ as shown in
reaction (1).

              CaO + S02 + 1/202  •*•  CaS04                              (1)

     Fresh limestone or dolomite sorbent feed rates to the boiler can be
reduced by regeneration of the sulfated sorbent to CaO and recycle of the
regenerated sorbent back to the combustor.  One regeneration system, studied
by Exxon Research and Engineering Company in the past, is the so-called one
step regeneration process in which sulfated sorbent is reduced to CaO in a
separate vessel at a temperature of about 1100°C according to equation (2).
The goal is to produce S02 in the regenerator off gas at a sufficiently high
concentration to be recovered in a by-product sulfur plant.

                      CO                 C02
              CaS04 + H2  ->  CaO + S02 + H20                           (2)

     A diagram of the pressurized fluidized bed combustion and regeneration
process is shown in Figure II-l.

     Exxon Research and Engineering Company, under contract to the EPA, has
built two pressurized fluidized bed combustion units to study the combustion
and regeneration processes.  The smaller of the two units, the batch unit,
was built under contract CPA 70-19 and was described in previous reports
(1,8,9).  Those reports also described regeneration and combustion studies
carried on in the batch unit.  The batch unit is being converted to a
continuous bench scale unit and the modifications required for the conversion
are described in this report.  No experimental work has been done in the
batch/bench unit during this current reporting period.
                                      11

-------
                                 FIGURE ll-l

                        PRESSURIZED FLUIDIZED BED
                         COAL COMBUSTION SYSTEM
        Gas Turbine
                                             Air Compressor
                     High Efficiency
                     Separator
              Cyclone
             Separators
                                           Solids
                                          Transfer
                                           System
   Coal and
Sorbent Makeup
                                                               To Sulfur
                                                               —*" Recovery
                                                              Separator
                           Combustor
Regenerator
                                                 Fuel

-------
     The larger unit, called the miniplant, was designed under EPA Contract
CPA 70-19 and built under Contract 68-02-0617.  Figure II-2 shows a photo-
graph of the miniplant.  The shakedown and operation of the unit was funded
under Contract 68-02-1312.  Previous reports  (1,7,9) described design,
shakedown and operation of the unit.  This report includes additional
results from the operation of the combustion  section of the miniplant.  The
effect of operating conditions on 862, NOX and particulate emissions, and com-
bustion efficiency was measured.  Various coals and sorbents were tested,
including precalcined limestone sorbent.  The first test in a series of tests
aimed at developing a comprehensive analysis  of all emissions especially trace
emissions, from the miniplant pressurized FBC unit was completed.  The design
and shakedown of a granular bed filter intended to remove particulates from
the flue gas to very low levels is also described in this report.

     This report also describes the operation of the regenerator section of
the miniplant.  The previous report  (1) described the regenerator and the
shakedown tests.  This report describes a 125 hour run in which the combustor
and regenerator were both operated continuously and the operability of the
system was demonstrated.

     The period of performance discussed in this report is August 1, 1976 to
August 12, 1977.
                                       13

-------
               FIGURE I1-2




EXXON FLUIDIZED BED COMBUSTION MINIPLANT
                   14

-------
                                 SECTION III

                             COMBUSTION STUDIES


     Combustion studies have been carried out  in  the EPA/Exxon pressurized
fluidized bed combustor referred to as the miniplant.  The miniplant has
provisions for continuous addition of coal and sorbent and continuous
withdrawal of sulfated sorbent.  The miniplant is shown schematically in
Figure III-l.  As of July 1977, the combustor  has been operated for a total
of approximately 1500 hours in a series of individual runs up to 240 hours
duration.  This section of the report describes the combustor equipment,
operating procedures, combustor performance and combustion results.  A
discussion of the regeneration work is given in Section IV.

EQUIPMENT, MATERIALS, PROCEDURES

     This section will focus on the major system  components which include:
1) solids feeding system, 2) combustor with internal subcomponents, 3) com-
bustor cyclones, 4) pressure control and flue  gas discharge system, 5) flue
gas sampling and analytical system, 6) process monitoring and data genera-
tion system, 7) combustor safety and alarm system, 8) coal and sorbent
properties, 9) operating procedures, and 10) analytical procedures.  A
detailed description of each of these systems  can be found in an earlier
report and only a brief discussion will be included here  (1,9).

Solids Feeding System

     Figure III-2 displays a schematic of the  miniplant coal and sorbent
feeding system.  Crushed and sized coal and limestone or dolomite are held
in separate storage bins  (20 tonnes for coal and  3 tonnes for sorbent) under
atmospheric conditions.  On demand, the solids from the bins are proportioned
to a specific coal/sorbent ratio.  Inverters control the motor speeds of
separate coal and sorbent screw feeders and volumetrically control the coal/
sorbent ratio.  A blending screw transports mixture into a solids feed ves-
sel.  The coal/sorbent mixture is held in this vessel until refill of the
injector vessel is required.

     The solids feeding system provides for continuous solids delivery
(coal and sorbent) from the injector vessel to the pressurized pressurized
combustor, while allowing intermittent refilling  of the injector vessel
(193 kg operating capacity).  Load cells located  under the injector vessel
monitor the solids feed rate and actuate control  signals for the refill
cycle.  Prior to initiation of a refilling operation, the injector vessel,
feed vessel, and the pair of solids storage bins  remain isolated from
each other.  When the load cell under the injector vessel detects a solids
loading of less than 102 kg, 91 kg of solids are  automatically transferred
pneumatically from the feed vessel to the pressurized injector vessel without
interrupting feed to the combustor.  Refilling is usually completed in about
5 minutes.  After refilling, the feed vessel is again isolated from the injector
vessel, vented, and filled with solids from the storage bins.  The feed vessel
is then isolated and repressurized to await another cycle.
                                      15

-------
                                             FIGURE Ill-l

                           EXXON FLUIDIZED BED  COMBUSTION MINIPLANT
                                                                          Cooling
                                                                           Water
                                                                                          To Scrubber
                                                                        Air (Pressure Control)
   Granular Bed Filter
Cooling Water
  Out   In
   I    f
                                                                                              To
                                                                                         Scrubber
                                                                                 Cyclone
                                                                                 Separator
                                                         .

                                                           Isolids
                                                           f Discharge
          Solids
          Reject
          Vessels
      Feed
     Water
  Reservoir

  Coal and
  Limestone
  Feed Supply
                                                                                                Natural Gas
                                                                                                Compressor
Auxiliary
Air
Compressor
                                                                       Main Air
                                                                       Compressor
    Liquid Fuel Storage

-------
                                               FIGURE 111-2
                                    COAL & LIMESTONE FEED SYSTEM
Limestone Bin
                                                                                      1/2  S.S. Pipe

-------
     Solids in the injector vessel are continuously aerated by the pressuriz-
ing air stream, which is automatically controlled at a pressure level slightly
above that in the combustor.  The coal/sorbent mixture is discharged from
the injector vessel through a 1.3 cm diameter orifice and pneumatically
conveyed by a stream of dried transport air through an s-shaped 1/2-inch
stainless steel pipe to the combustor.  A short section of high pressure
rubber hose is used to connect the discharge orifice to the injection line
to avoid interference with the load cell operation.

     Final entry of solids into the combustor is through a 1.3 cm I.D. nozzle
located 28 cm above the fluidizing grid and horizontally extending about
2.5 cm beyond the reactor wall.  The tip of the probe includes ten 0.79 mm
diameter holes which surround the solids feed opening.  They are used to
continuously inject an annular stream of sonic-velocity air to assist penetra-
tion of the solids feed into the fluidized bed and to protect the feed nozzle
from blockage with bed solids.  The flow of solids into the combustor is
controlled to maintain constant temperature in the combustor.

Combustor

     The combustor consists of a 61 cm I.D. steel shell refractory lined to
an inside diameter of 33 cm.  The 9.75 m high unit is fabricated in flanged
sections to allow insertion and removal of the cooling coils.  Various ports
are strategically located to allow for material entry and discharge.  Numerous
taps are also provided for monitoring both pressure and temperature.  A
schematic of the combustor is shown in Figure III-3.

     Heat removal from the combustor is provided by cooling coils located in
discrete vertical zones above the grid.  Each coil has a total surface area
of 0.55 m^ and consists of vertically-oriented loops constructed of 1/2-inch
Schedule 40 316 stainless steel pipe.  The number of coils can vary from one
to ten depending on the combustor operating conditions and the amount of
cooling required.  A high pressure pump is used to pump the cooling water
through a closed-loop arrangement consisting of a demineralized feed water
reservoir, cooling coils, and a heat exchanger.  The flow rate and exit
temperature from each coil can be separately controlled and monitored.

     The baffled cooling coils were inspected after Run 65 and found to be
in excellent condition.  These coils were installed prior to Run 29 and
logged more than 600 hours of running time.  There was no sign of the erosion
seen on earlier unbaffled coils.  Even the baffles were not eroded as seen
in Figure III-4.

     The combustion air to the unit is provided by a main air compressor
having a capacity of 40 Sm3/min at 1030 kPa (1400 SCFM at 150 psig).  Preheat
of the combustor during start-up is made possible by a natural gas burner
which is housed in the bottom plenum section of the combustor.  Once the
fluidized bed temperature reaches approximately 430°C a liquid fuel system
is used to heat the bed further to the coal ignition temperature.

     The liquid fuel used to heat up the combustor is "Varsol" or kerosene.
The fuel is pumped, under pressure slightly higher than combustor pressure,
from either of two 218 1 (55 gal) drums through a rotameter to an injector'
nozzle located at port #5, 15.2 cm above the combustor grid (Figure III-5).

                                       18

-------
  FIGURE  III-3
COMBUSTOR VESSEL

-------
            FIGURE II1-4




COOLING COILS AFTER 600 HOUR EXPOSURE
                  20

-------
                                                         FIGURE  III-5



                                                     LIQUID  FUEL  SYSTEM
  AtfX. A/ff J-VP

XT ffTJfOCAfg
                                                                                                                        COMBUSTOR  FUJL PffOBE



                                                                                                                              ~™" FOR DIFAIL RfFFR To DRAW If*.
                                                                                                                 I                .."-«--D


                                                                                                                 1 ^- SHfo fitsr MIU
                             JLJL             *—i'sasi smut
                                                    BAii
                                                                                                                              LtQUlO FUfL
                                                                                                                              race M/Nt PLANT

-------
The injector nozzle atomizes the fuel to insure complete combustion (Figure
III-6).  The fuel flow rate must be manually set at the rotameter, however,
a solenoid valve is used for on/off control from the control panel.  The
liquid fuel system is also activated during coal feed problems to maintain
combustor bed temperature above the coal ignition temperature.

     Coal and sorbent are injected pneumatically into the combustor through
a single side-entering port 28 cm above the fluidizing grid.  The maximum
design solids feed rate is approximately 227 kg/hr.  The combustor is capable
of operating at pressures up to 1000 kPa and temperatures to 1100°C, super-
ficial velocities up to 3 m/s with expanded beds to 6.1 m.  The expanded bed
height can be controlled at any level above 2.3 m by the continuous with-
drawal of bed solids through a port located 2.3 m above the fluidizing grid.
Solids flow by gravity through a refractory lined pipe into a "pulse pot"
from where they are pneumatically transported by controlled nitrogen pulses
to a pressurized lock hopper.

     The combustor fluidizing grid consists of 332 3/32 inch holes and four
independent cooling water loops consisting of five channels each (see Figure
III-7).  The previous grid, which was installed through Run No. 49, had
only one cooling water loop with twenty channels in the grid.  This grid
suffered from uneven grid temperature distributions due to preferential water
flow.  Some erosion was seen on the old grid, however, it did not fail.  The
current grid cooling scheme results in even grid temperatures and should
result in longer grid life.

     Combustion gases exit the combustor and go to a two stage cyclone system.
The primary intent of the first cyclone is to recirculate larger unburned
carbon particles back to the combustor to improve combustion efficiency.
Particulates collected in the second stage cyclone are dropped into a lock
hopper and disposed of on a batch basis.  Due to refractory deterioration
and severe pitting, the second stage cyclone was recast with a more resistant
refractory (Resco RS-17-E).  The barrel was also made smaller to correspond
to the smaller inlet and outlet pipes that were installed last year.  The
particulate collection efficiency was not changed significantly due to this
modification.

     Flue gas is sampled at a point about 7 m downstream of the second stage
cyclone.  The system is designed to produce a solids-free, dry stream of flue
gas at approximately ambient temperature and atmospheric pressure whose com-
position, except for moisture, is essentially unaltered from that of the
original flue gas.  The system was described in the previous report (1).
Particulates are also sampled near this point.  The particulate sampling sys-
tem is discussed in detail on page 35.

New Combustor Sections

     Two new sections fabricated for a fireside corrosion program sponsored
by the Department of Energy (DOE) and conducted in cooperation with Westinghouse
Research Laboratory have been installed in the combustor.  The DDE-funded
fireside corrosion work will be conducted on the miniplant under a coopera-
tive agreement between EPA and DOE.  The orientation of these sections is
shown schematically in Figure III-8.  The lower corrosion test section
(Figures III-9 and 111-10) measures 0.91 m in height and is located 0.91 m
                                     22

-------
        FIGURE III-6
LIQUID  FUEL INJECTOR NOZZLE


                                                                F6CC MINI - P
                                                                     1647-2-Z-D

-------
      FIGURE III-7
COMBUSTOR FLUIDIZING GRID
                                      1U-
                                                         gflcc MUSI-PLANT

-------
                        FIGURE 111-8

           COMBUSTOR CORROSION  PROBE LOCATION
 2m
1.5m
0.9m
0.9m
0.9m
0.9m
0.9m
0.9m
  I

1



1

1
m
1
m
1
m
1
m


Upper
Corrosion
Probe
Section
(Detail 'A')








r
Lower Corrosion
n

1
Tl
Probe Section
(Detail 'B')



I





1

I

1

1

1



I

                        Combustor
                        Ports
                                         Detail 'A1
  Combustor
  Ports
                 Detail  'B1
Fluidizing Grid
                            25

-------
                                                         FIGURE III-9

                                           COMBUSTOR LOWER CORROSION PROBE  SECTION
N>
ON
                                                                                                                ggiVuir mbSSF.

-------
                     FIGURE 111-10




COMBUSTOR LOWER CORROSION PROBE AND BED SAMPLING PROBE

-------
above the fluidizing grid.  The upper corrosion test section (Figure III-ll)
measures 1.5 m in height and is located 5.5 m above the fluidizing grid.  Both
sections are refractory lined to an internal diameter of 0.32 m to conform
to the dimensions of the combustor.

     These two sections will provide a test site and environment for exposure
of 21 specimens of heat exchanger materials.  Twelve of the specimen probes
will be located in the lower section within the expanded bed while the
remaining nine will be located in the upper section in the freeboard region.
The probes (Figure 111-12) will be air cooled and temperature controlled to
requirements specified by Westinghouse.

Temperature Control

     Rate of solids feed is automatically controlled in order to maintain a
specific operating temperature within the combustor.  This is accomplished
through a series of controls involving the combustor temperature, pressure
differential between the primary injector and combustor, the injector pres-
sure, and the transport air flow rate (Figure 1II-2).

     A thermocouple in the lower zone of the combustor, 46 cm above the
fluidizing grid, is used as the sensor for the control of the combustor
temperature.  The coal feed rate to the combustor is regulated by the pressure
differential between the injector vessel and the combustor - the greater
the pressure difference, the greater the feed rate.  Temperature control
is accomplished by a cascade type control loop using two controllers, one for
temperature control and another for AP control.  A deviation of the desired
temperature from the actual combustor temperature causes a signal to be trans-
mitted by the temperature controller to the AP coal feed rate controller.
This error signal actually resets the set point of the AP controller so that
a different AP will be established between the coal vessel and the combustor.
This change in the pressure difference between the coal vessel and combustor
causes a change in the coal feed rate which will tend to return the bed tem-
perature to the desired value.  Proper tuning of the controllers is necessary
for optimum reaction to system perturbations and anticipation of changes in
bed conditions.  This control system has performed very satisfactorily and
provides excellent temperature control and response.

Pressure Control

     The FBC Miniplant Combustor has the capability of operating at pressure
levels of up to 10.5 atmospheres (140 psig).  Pressure control is achieved
by restricting the discharge flow of gas from the combustor, so as to achieve
an increase in back pressure.  This is done by use of a converging nozzle
inserted in the discharge line.  Adjustment of the combustor pressure is
accomplished by metering high pressure air into the discharge line just
upstream of the flow nozzle.  A 2 inch-ball valve equipped with a pneumatic
positioner and actuator regulates the amount of air added in response to a
signal from the pressure controller.

     The converging nozzle is machined from a silicon carbide section which
is inserted into an 8 inch carbon steel blind flange.  Four nozzle inserts
are currently on hand, with throat diameters of 2.41, 2.67, 2.79 and 2.91

                                     28

-------
             FIGURE III-ll
COMBUSTOR UPPER CORROSION PROBE SECTION

-------
      FIGURE 111-12




COMBUSTOR CORROSION PROBE
            30

-------
centimeters.  By proper nozzle preselection, a wide  range of combustor flow
conditions can be handled.  Control of  combustor pressure has always been
very good, with variations from  the set pressure usually less than 0.1
atmospheres.

Turbine Specimen Erosion  Section

     A turbine specimen/erosion  section was  fabricated  for EPA by Westinghouse
Research Laboratory and inserted in the flue gas line downstream of the
secondary cyclone.  This  includes a nozzle which enables high velocities to
be reached at the specimen location with  relatively  low combustor air flow
rates.  The  section measures  45.7 cm  long and has  an entrance diameter of
8.9 cm which converges to 2.9 cm at the specimen position.  Some of the fac-
tors affecting erosion were discussed by  Westinghouse in a report to the EPA
(3).

     During  Miniplant Run 52  an  erosion test rod supplied by Westinghouse
was exposed  to superficial velocities of  259 m/sec (850 ft/sec).  The base
material of  the rod was X-45  and was  coated  with Co-Cr-Al-Y.  The temperature
at the section was  760°C  and  the grain  loading was measured to be 2.4 gm/m3
 (1.1 gr/SCF).  Inspection of  the rod  following 32  hours of exposure showed
considerable erosion  (Figure  111-13).  The rod has been returned to Westing-
house for analysis.

Sampling and Analytical  Systems

Bed Sampling Probe—
     A solids sampling probe  has been designed, constructed and installed to
enable acquisition  of  samples from the  combustor bed while a run is in
progress  (Figure  111-10). The need for such a sampling probe was evident
during the  regenerator demonstration  run. In that run, which lasted 125
hours, bed  solids did not have to be  rejected to maintain bed height, and
accordingly no information could be obtained on the  composition of the com-
bustor bed.  A cross-section  of  the probe is shown in Figure 111-14.  The
probe consists of a closed-end tube which contains a hole on the sidewall
near the  top of  the tube.  In normal  operation the top  end of the probe is
recessed  into the refractory  wall of  the  combustor vessel.  In  this position,
the bed particles form a  seal around  the  top of the  probe preventing par-
ticles from reaching  the  hole in the  sidewall.  When it is desired to take a
sample, a pneumatic piston pushes the probe  forward  into  the combustor bed,
exposing  the hole on  the  upper sidewall.   Bed particles rush into this hole
in  the sidewall  and fill  the  probe cavity.   The probe  is  pneumatically pulled
back into  the recess  in  the  combustor refractory wall  and  the captured sample
of bed particles  is removed  from the  lower end of  the  probe.

     In Run No.  51,  bed  samples  were  taken using  the sampling probe at one-
hour intervals.   The  utilization of  the bed  solids determined from analysis
of  the particles  obtained from the probe  samples  is  shown in Figure 111-15.
The  sorbent utilization  decreased with  time. This is  consistent with an
approach  toward  the expected  steady  state value  of 50  percent sorbent
utilization expected  for  the  calcium to sulfur molar feed ratio of 1.67 for
this run.   The  sorbent utilization determined  from a bed solid  reject sample
taken at  5.8 hours  into  the  run is included  for  comparison in  the  Figure
111-15 and  shows  good agreement with the  results  obtained using the  sampling

                                      31

-------
       FIGURE 111-13




WESTINGHOUSE EROSION PROBE
          32

-------
                                                      FIGURE  111-14

                                              COMBUSTOR  BED SAMPLING PROBE
CO
U)

-------
                                          FIGURE 111-15
                                  SORBENT UTILIZATION VS. TIME
                                          RUN No. 51
            0.80
1
                                                  T
         T
                        BED SAMPLING PROBE SAMPLES
                        BED REJECT LINE  SAMPLE
            0.70
u>
            0.60
            0.50
            0.40
I
I
I
                                         345
                                      TIME INTO RUN (HOURS)
                             6
                             7

-------
probe.  One of the values determined  for  the  sorbent utilization during this
period was well outside the range  of  the  other values.  This points out the
need for taking multiple samples to avoid erroneous results due to non-
representative samples.

Gas Sampling System—
     The flue gas sampling system  is  installed at  a point downstream of the
secondary cyclone.   It was described  in the previous report (1).  The system
is designed to produce a solids-free,  dry stream of flue gas at ambient
temperature and low  pressure whose composition except  for moisture, is
essentially unaltered  from that of the original  flue gas.  This system pro-
vides a gas sample which is analyzed  by the continuous, on-line, gas analyzers.
Another flue gas sample can be extracted  which has been filtered, cooled and
depressured, but not dried.  This  system  is used to obtain batch samples of
flue gas for analysis by wet chemistry methods.

Particulate Sampling System—
     The particulate sampling system  initially installed on the miniplant
is shown in Figure 111-16.  A 1.09 cm probe was  used,  and as this system
was initially intended to be used  at  about 275°C,  a three stage water cooled
heat exchanger was provided.  Total particulate  sampling was done with
Balston Type 30/25 and 30/12 filters  containing  high temperature resistant
(~530°C) BH cartridges.  The gas passing  the  filter was cooled to 200°C and
passed through another filter which removed condensed  matter.  Size distribu-
tions were to be made with a Brink impactor at 175°C.

     This system was intended to obtain samples  at temperatures high enough
to prevent condensation of moisture.   Some operating problems with this system
became apparent and  modifications  were made.  The  section containing the
impactor was removed after run 50  due to  problems  of gas leakage from the
impactor.  It was also decided to  increase the operating temperature of the
filters to the maximum to avoid condensation  of  substances other than moisture,
which could affect the particulate loading, particle size distribution and
composition.  The modifications made  to the sampling system for the above
and other reasons are  described below.

     •  Before Run 50, a second 30/25 (high temperature) filter was
        placed immediately after the  first filter. The second filter
        was used as  a back-up filter  in case  the cartridge material
        of the first filter failed or particle by-pass occurred due
        to excessive pressure drop across the first filter.

     •  In order to  increase the operating temperature of the filter
        for Run 52,  the three stage water cooled heat  exchanger was
        replaced with a five foot  length  of 1.27 cm OD ss pipe.

     •  At the end of Run 54, the  system  was  again modified.
        The changes  corresponded to  (a) the use  of a smaller probe
        diameter (0.77 cm) to decrease the pressure drop across the
        filters by reducing the flow  rate at  isokinetic velocities,
        and (b) the  ability to place  the  first and second filters
        next to the  high temperature  valve, if required, so that
        the filters  may operate at a  high temperature  (maximum^
        allowable temperature for  the filter  cartridges is 530°C).
                                      35

-------
          FLUE GAS FROM
        SECONDARY CYCLONE
u>
1
e —

HIGH
TEMPERAT
VALVE
	 . [N^-l— -
-i ^ r

— X 4
V V
I !
PURGE
NITROGE
PRESSURE PRESSURE
GAUGE GAUGE
JRE Tc9 9 TC ^
~i r 1 i 1 , _ -, r 1
LlLji U 	 p,L_^ 	 U
i L i i ' n n\A/
H
V
xj
HEAT COLLECTION STEAM pi^Rm VAI VF
1EXCHANGER FILTER MEAT CONTROL VALVE
EXCHANGER
N
                                                                              ROTAMETER
                                          FIGURE  III - 16

                             MINIPLANT PARTICULATE SAMPLING SYSTEM

-------
     Future work on particulate  sampling will  entail  the  current equipment
discussed above, as modified after Run  54,  and two  other  more  sophisticated
systems.  One of these systems will be  placed  before  the  granular bed filter
and will contain Southern Research Institute's specially  designed cyclone
set; the other system will be placed  after  the granular bed  filter and will
contain an impactor and/or a total filter.   Both  these systems have been
designed to perform at 870°C and 909  kPa, and  automatic flow and temperature
control devices have been incorporated.

     Samples of flue gas will be extracted  under  isokinetic  or constant flow
conditions and the temperature of the sample will be  maintained by line
heaters.  The sampling device  (cyclone  or impactor) will  be  enclosed in a
pressurized furnace in which the temperature is held  at the  flue gas tempera-
ture (up to 870°C) .  The pressure inside the furnace  will be the same as the
pressure inside the sampling device so  the  device will not have to be built
to withstand a high pressure.  This will minimize structural, sealing and
leakage problems.  Each system will be  self contained and can be moved to
different locations on the miniplant.   Although the systems  will be built to
operate up to 870°C, the effect  of sampling temperature on particulate
loading, size and  composition will be measured and  the most  suitable operating
temperature will be determined.

     In this report, results on  particulate sampling  will be given for all
runs up to Run 59, in which sampling  experiments  were carried out.

Acurex  Sampling Section—
     Several miniplant discharge line sections were modified in December 1976
to allow insertion of a new high temperature-high pressure (HTHP) particulate
sampling probe developed by the  Aerotherm Division of the Acurex Company.
The probe was developed under EPA contract  and is described  in a recent
report  (17).

     Exxon was requested to build a sampling section  to allow a test of the
sampling system under high temperature, high pressure conditions.  The probe
sampling section fabricated by 'ER&E has a 10 inch stainless  steel liner
inside of a 16-inch refractory lined  pipe section (Figure 111-17).  Eccentric
reducers at each end of the sampling  section were necessary  to adapt to the
existing discharge line piping and also avoid  extensive modification of the
miniplant structure.

     The initial miniplant field test of the Acurex HTHP  sampling probe was
made during the four day Comprehensive  Analysis test  series  (run #50) March
29 to April 1, 1977.  Additional tests  with the Acurex HTHP  probe were con-
ducted during the  week of May 23, 1977.  A  photograph of  the Acurex HTHP
sampling probe installation at the miniplant is shown in  Figure 111-18.
This photograph also shows the location of  the Westinghouse  turbine test
section.

Process Monitoring and Data Generation  System

     Data characterizing the miniplant  operating  conditions  are recorded on
5 multipoint recorders.  Three recorders  (24-channel  Honeywell Electronik
112) monitor output from various measuring  instruments -  thermocouples, dif-
ferential pressure transmitters, etc.   Two  recorders  continuously monitor the
                                      37

-------
                                                                         FIGURE  111-17


                                                     COMBUSTOR  - ECCENTRIC  ACUREX PROBE SECTION
U)
CO
                                                                                                                                           sn g*f#- gtt-tst-c
                                                                                                                            JS'TOB"EccfNTKic /fcovcfR .(^)
                                                                                                                                  5TANDARD  WALL 57L.
                                                                                                                                   fO TO V " tCCEHTRIC fttDVCfR (?)
                                                                                                                                   30V3.S.   SCC 0FT4/LS BtiOW
                                                                                                                                    ft)  a'f/fC J-50 LB. FiANGCS
                                                                                                                  73 MATCH UP WITH
                                                                                                                    SCHP.J'  10"
                                                                                                                    S.5. PIPE
                                                                                                                   JO MAKH t/P H/ITU
                                                                                                                     SCUD J  V
                                                                                                                     s. s. pipe
I	I
                                                                                                                                               MINI-PLANT
                                                                                                                                      £CCfHTKIC ACURfX PROBf
                                                                                                                                           SfCTldN
                                                                                                                                       EXXON HOCMCH *MO CHCIMCCmiW

                                                                                                                                            MECHANICAL OIVIWM

                                                                                                                                               I.INOKN, M J.
                                                                                                                                    "Thrums

-------
                                                  FIGURE 111-18




                                                  ACUREX PROBE
to

-------
output signals from the gas analyzers.  In addition, at one minute intervals,
the same output is recorded by a data logger system consisting of a Digitrend
210 data logger with printer and a Kennedy 1701 magnetic tape recorder.
Approximately 100 pieces of data are logged with three-quarters involving
temperature measurement while the rest deal with pressure and material flow
rate.  The points logged are given in Appendix F.

     Signals from the data logger are scanned every minute and appear as
digital output on printed paper tape and are also stored on magnetic tape.
The magnetic tape, containing about 6000 items of data per hour of run time,
is fed to a computer which converts the logger output to flow rates, pres-
sures, etc. with the proper dimensions.  The data are then averaged and
standard deviations calculated over preselected time intervals (usually 10
min.).  Other quantities are also calculated.  This includes average bed
temperature, based on four thermocouple readings covering the 15-114 cm
interval above the fluidizing grid, superficial gas velocity, excess air, as
well as the important gas concentrations.

Combustor Safety and Alarm System

     A process alarm system was designed to warn of impending operational
problems.  Two general alarm categories exist.  The first, dealing with less
critical situations, alerts the operator of the problem so that appropriate
corrective action can be taken.  The second class of more critical alarms
results in the immediate or time delayed shutdown of the complete system or
specific subsystems.  An alarm condition is brought to the attention of the
operators by a flashing light above the control panel accompanied by a high
pitch sound.  The sensitivity of the individual alarms is controlled by
potentiometers located beneath the control panel.

     Appendix G gives a brief description of the alarm triggering
condition, the mechanical sensoring device, and the follow-up actions taken
by the system.

Coal and Sorbent Properties

Coal—
     Coals used in the miniplant variables study were a high volatile
bituminous coal from the Consolidation Coal Company's "Champion" preparation
plant in Pennsylvania and Illinois No. 6 seam coal obtained from Carter Oil
Company's Monterey No. 1 mine.  The Champion coal was partly classified to
remove fines smaller than 40 U.S. Mesh.  The Illinois coal was screened to
6 X 25 U.S. Mesh to prevent plugging of the primary injector feed vessel.
Particulate size distribution and composition data for both coals are shown
in Figure 111-19 and Table III-l respectively.  During and after Run 50, the
Champion coal sulfur content changed from the average of 2+0.1% to between
1 and 1.6% due to dilution with a 0.6% sulfur Kentucky coal at the preparation
plant.   This was not detected until after the runs had begun.

Sorbent

Sorbent—
     Grove limestone (BCR No. 1359) and Pfizer dolomite (BCR No. 1337) were
the primary sorbents used in the miniplant variables study.  The composition
                                      40

-------
CO

UJ

O
h-
o:

o_

s
<
CO
co
:r
CD
LU
UJ
>
 O
lOOh


 90


 80


 70


 60


 50


 40


 30


 20


 10
                                   FIGURE 111-19

                          COAL PARTICLE SIZE DISTRIBUTION
       Illinois Coal

       Champion Coal
                         I
       0  200  400 600 800     1200

       Mill   I  I    I   I    I
                                    1600
1800

  I
2400
2800
3200 Microns
        200  50 40  30    20 18  16  14
                                       12
 10
 8
          6  Mesh
                                    PARTICLE SIZE

-------
                                       TABLE III-l.  COAL COMPOSITION
Run No.
41.1
41.2
43.1-43.5
45.1-48
50.1A-50.3
50.4-52
54-59
Coal Type
Champion
Arkwrlght
Illinois
Champion
Champion
Champion
Champion
Ultimate Analysis
Moisture
2.83
0.88
3.01
2.83
1.67
1.32
2.19
Ash
8.03
7.42
9.95
8.03
13.10
11.90
12.50
Total
Carbon
70.38
77.11
67.71
70.38
74.00
73.40
71.35
H
5.01
5.09
4.77
5.01
5.00
5.00
4.74
S
1.96
2.47
4.20
1.96
1.85
1.66
1.40
N
1.36
1.07
1.17
1.36
1.50
1.40
1.54
Cl
0.09
0.14
0.05
0.09
—
—
0.07
°2
10.43
5.96
9.19
10.43
2.40
4.20
6.31
Heating
Value
BTU/lb
13,346
13,699
12,254
13,346
12,973
13,268
12,514
(1)   Diluted with 0.6% sulfur Kentucky coal.

-------
of these stones is given in Table III-2.  Most of  the runs were made with  the
stone screened to give the distribution of limestone and dolomite shown  in
Figure 111-20.

Operating Procedures

     Prior to initiating a run, a detailed checkout procedure is followed  to
insure that the system is ready for operation.  This includes various equip-
ment checks, alarm system checks, calibration of flue gas analyzers, activa-
tion of process monitoring and control systems, and the turning on of all
cooling water systems.  All runs are begun with an initial bed of sorbent
in the combustor.  This consists of either a fresh charge of uncalcined
limestone or the bed from the previous run.

     The first operation of start-up involves heating the gently fluidized
sorbent bed by burning natural gas in the burner plenum followed by injection
of kerosene into the bed.  Prior to ignition of natural gas, an air flow
of about 9.9 Sm^/min  (350 SCFM) or about half that used at normal opera-
ting conditions is fed through the burner while combustor pressure is raised
to 280 kPa gauge.  Once ignition begins, this procedure maximizes incoming
gas temperature under conditions which allow good  natural gas combustion and
adequate bed fluidization.  Water flow rates through the combustor cooling
coils are kept low to reduce heat loss to the coils.  Ignition begins by
simultaneously feeding 0.57 Sm3/min (20 SCFM) of natural gas through the
burner while activating an ignition electrode.

     Because of the limited capacity of the gas compressor, natural gas burned
in the plenum is used only to heat the bed to a temperature of about 430°C,
sufficient to insure self-ignition of kerosene.  This generally requires 20-
30 minutes.  At this point, kerosene is injected into the lower portion of
the bed.  When rising temperatures indicate ignition of liquid fuel, natural
gas feed is discontinued to insure sufficient air  for complete combustion  of
kerosene.  Approximately 10-15 minutes are required to raise the bed tempera-
ture to 650°C, which is sufficient to achieve self-ignition of coal.

     Coal, usually mixed with limestone, is then fed to the combustor from
the primary injector.  A steady stream of 1.7 Sm^/min (60 SCFM) of
transport air is used to convey coal into the combustor.  Actual rate of
coal injection is determined by the pressure differential between the injector
and combustor.  The rate is initially set at an appropriate value based on
past experience under similar operating conditions.  Once ignition of coal is
verified by rapidly rising temperatures, kerosene  flow is stopped.  At this
time, the main combustion air feed line to the plenum is opened allowing
most of the air to bypass the burner, and both combustion air flow rate and
combustor pressure are rapidly increased to their  designated operating
values.  Flow of water to each cooling coil is adjusted to maintain steam/
water exiting temperatures of 138-150°C.  Once the desired bed temperature
has been reached, it is held constant by the automatic coal feed rate
control system.
                                      43

-------
  ISI
 - —

  "
  LU
O
LJ <

> Q-
O
cn
c/i
LJ
100



 90




 80




 70



 60




 50



 40




 30



 20




 10



  0
          I
                                               FIGURE 111-20



                                    SORBEKlT PARTICLE SIZE DISTRIBUTION
                    O Grove Limestone


                    Q Tymochtee Dolomite
             Mesh
                         3^ 30  :
                            25   20
18


 I
             Microns
                         500
                                      1000


                                        PARTICLE SIZE
                 1500
2000
2400

-------
TABLE III-2.  SORBENT COMPOSITION
Run
Number
41.1-41.2
43.1-43.5
45.1-47
48-51
52
56-59
Quarry
Grove
Pfizer
Grove
Pfizer
Pfizer
Pfizer
Sorbent
Type
Limestone
Dolomite
Limestone
Dolomite
Dolomite
Dolomite
Weight Percent
CaO
97.0
54.0
97.0
54.0
57.8
57.8
MgO
1.2
44.0
1.2
44.0
40.7
40.7
Si02
1.1
0.9
1.1
0.9
0.25
0.25
A1203
0.3
0.2
0.3
0.2
1.9
1.9
Fe2°3
0.2
0.3
0.2
0.3
0.133
0.133

-------
Analytical Procedures
Particulate Sampling

     The equipment for particulate sampling has been described on page
35.    The key to sampling consists of accurately weighing the filter ele-
ments and cartridges before and after the experiment and sustaining iso-
kinetic velocity throughout the experiment.  Before starting, air is
admitted to the probe to prevent it from plugging.  This air is shut off
after the filters are in place, and the high temperature valve is opened.
Pressure and temperature readings are taken at regular intervals, and the
rotameter is constantly and carefully checked and/or adjusted for isokinetic
sampling.  At the end of the sampling period, the high temperature valve is
shut off, and air is admitted into the sampling system and into the probe.

     After the total weight of the particulates is obtained, the solids are
wet sieved to obtain the + 45 ym fractions.  The -45 jam fraction, which usually
consists of 80%+ of the total, is then analyzed by a Coulter Counter to
obtain the particle distribution.

EXPERIMENTAL RESULTS AND DISCUSSION

     The combustion experimental program was directed primarily toward under-
standing the factors which control SC>2 emissions from the pressurized combus-
tor.  However, information on NOX emissions, particulate emissions, and carbon
combustion efficiency were also obtained.  In addition, the first test in a
series of tests was completed designed to make a comprehensive analysis of
all emissions, especially trace emissions of potentially harmful materials.
Other information such as temperature profiles in the combustor and cyclone
efficiency data was also generated in the test program.

     This section reports and discusses the results of the combustion program.

S02 Retention

     The objectives of the S02 retention program were to study the effect of
coal type and the use of precalcined limestone sorbent on -SOo retention.
The effect of other variables was studied and reported previously (1).

Effect of Coal Type—
     The effect of coal type and sulfur content was determined in a series of
runs made with an Illinois No. 6 coal containing 4.2% sulfur and Pfizer No.
1337 dolomite sorbent.  Coal and sorbent properties are given in Tables III-l,
III-2.  Details of the runs (43.2 to 43.5) are given in Appendix H-l.  The Ca/S
molar ratio, i.e., moles of calcium fed in the dolomite to moles of sulfur
fed in the coal, was varied from 0.7 to 1.6.  S02 retention results using the
Illinois coal are shown in Figure 111-21 as a function of the Ca/S ratio
and are compared to data reported previously using a Pittsburgh Seam coal
(Champion) containing 2% sulfur (1).  As seen in Figure 111-21, the reten-
tions measured with the Illinois coal fit reasonably well with the data and
the correlating line established for the Pittsburgh coal.  This suggests
that the dolomite sulfation reaction is first order in S02 concentrations
and is consistent with results reported previously by others (2,3,4).
                                     46

-------
                        FIGURE  111-21
           S02  RETENTION WITH  DOLOMITE SORBENT
  100
 CM
o
CO
                 •         •••
                            •  Pittsburgh  Seam Coal

                            A  Illinois Coal
                        Sorbent: Pfizer (No. 1377) Dolomite
                      1               2


                      Ca/S (MOLE/MOLE)
                            47

-------
Effect of Limestone Precalcination—
     Previous pressurized FBC studies with limestone sorbent indicated  that
the limestone desulfurization activity was determined in part, by  the extent
of calcination occurring in the combustor (1).  Higher combustor tempera-
tures promoted calcination and increased limestone activity.  However,  even
at the higher combustor temperatures, limestone was less active than dolomite
at an equivalent Ca/S molar ratio.  It was also found that limestone desul-
furization activity approached zero under low temperature "turndown" operating
conditions due to the inability of the limestone to calcine at those condi-
tions.   Since the level of calcination figured so strongly in determining the
activity of  the limestone, a series of runs were made in which the limestone
was precalcined and fed to the combustor in the calcined form along with the
coal.  The runs were made with Champion coal and Grove No. 1359 limestone.
Coal and limestone properties are given in Tables III-l and III-2.  The
limestone was calcined in the miniplant combustor at a temperature of about
870°C at 940 kPa pressure.  The natural gas fired preheat burner was used
during calcination to maintain the temperature at 870°C.  A single cooling
coil of  reduced size was used to remove excess heat and control the tempera-
ture at  the  proper level.  The CC>2 partial pressure in the combustor during
calcination was about 109 kPa.  The calcined limestone was removed from the
combustor, stored in steel drums and then added, by hand, to the coal/limestone
blender  located in front of the coal/stone injector vessel.  The detailed
run conditions and results (runs 46.1-46.4) are in Appendix H-l.  The runs
were made under variable temperature conditions with temperatures ranging
from 760 to  920°C.  Ca/S molar ratios varied from 0.9 to 2.5.

     The results of the runs using precalcined limestone at 865 to 920°C are
given in Figure 111-22.  Figure 111-22 also shows, for comparison, the  correla-
ting lines for dolomite and raw limestone developed earlier and reported in
the previous report (1).  As seen in Figure 111-22, the precalcined limestone
is as active, at an equivalent Ca/S molar ratio, as dolomite and is much
more active  than raw limestone.  Also, the activity is maintained at the
lower temperature (865°C) where limestone would not be expected to calcine
extensively  in the combustor.

     A run was also made with precalcined limestone at a very low "turndown"
temperature  (760°C) where previously, raw limestone was shown to be com-
pletely  inactive (1).  The result is given in Figure 111-23 where it is
compared to earlier results obtained with dolomite.  The solid line is  the
correlating line developed for dolomite and precalcined limestone.  The
individual data points are those obtained with dolomite and precalcined
limestone at the very low temperature "turndown" conditions (690-760°C).
The dashed line correlates the low temperature results obtained with dolomite
and shows that the dolomite is slightly less active at the lower temperatures.
Again, precalcined limestone is as active, at an equivalent Ca/S molar  ratio,
as dolomite, even though the low temperatures would strongly favor the  forma-
tion of  CaC03 rather than CaO in the combustor.

     The effect of precalcining limestone was studied previously in TGA
equipment by Westinghouse Research Laboratory (5).  In that study it was
found that precalcination under a high C02 partial pressure resulted in a very
active sorbent.  The results obtained in this study support the Westinghouse
results.   In other studies carried out at Argonne National Laboratory (6)
it was found that the activity of a sorbent could be related to the surface
                                     48

-------
   100
ce:
 CM
O
                        FIGURE 111-22

         S02 RETENTION WITH PRECALCINED LIMESTONE
                             DOLOMITE (840-950°C
                             Combustor Temp.)
    90
    80
    70
    60
    50
    40
    30
     20
             LIMESTONE
             (Not Precalcined)
             (925-950°C
             Combustor Temp.)
            LIMESTONE
            (Not Precalcined)
           (825-900°C
           Combustor Temp.)
       Precalcined Limestone
          865°C Combustor Temp.

          920°C Combustor Temp.
     10
      0
       0
   2        3

Ca/S MOLAR RATIO

     49

-------
                      FIGURE 111-23
 EFFECT OF LOW TEMPERATURE OPERATION ON S02 RETENTION
  100
   90
   80
LU
QL

 CM
O
CO
   70
   60
    50
   40
   30
   20
   10
    0
              I
         DOLOMITE & _
         PRECALCINED
         LIMESTONE
         (840-950°C)
I
I
                         DOLOMITE & PRECALCINED
                       ' LIMESTONE
                        (690-760°C)
                       Dolomite (690-760°C)
                       Precalcined  Limestone (760°C)
                    Ca/S MOLAR RATIO
                           50

-------
area of the sorbent occurring in large diameter  pores.   Precalcining under
high C02 partial pressure conditions apparently  produces such  a  favorable
pore structure.  Also, the pores appear  to be  large  enough  that  carbonation
of the limestone, which occurs under very low  temperature conditions, does
not reduce the pore diameter enough to prevent diffusion of SC>2  into the
interior of the limestone particles.

     Further evidence of the high activity of  the  precalcined  limestone can
be seen from the level of calcium sulfation  in the used  sorbent.  The sulfa-
tion levels of limestone varied from 20  to 39% in  previous  runs  made in the
miniplant.  In the series of runs made with  precalcined  limestone,  the sulfa-
tion level ranged from 55 to 67%.  This  is comparable  to sulfation  levels
normally measured with dolomite sorbent.

     As a result of the above and previous studies,  the  sorbent  requirements
needed to satisfy the current EPA new source performance standards  for SC>2
emissions from a coal fired boiler  (1.2  Ib S02/M BTU coal fired) can be esti-
mated.  The estimate is shown in Table III-3.  The estimate was  based on a
gas phase residence time of 2 s and a boiler temperature of 930°C.  As seen
in Table III-3, dolomite and precalcined limestone are more effective than
limestone on a molar basis.  However, on a weight  basis,  limestone  is slightly
more effective than dolomite with a coal containing  2% sulfur.   Limestone and
dolomite are equivalent for a 3% sulfur  coal.  For coals containing more than
3% sulfur, dolomite is more effective than limestone even on a weight basis.
However, precalcined limestone is more effective than dolomite for all sulfur
levels.

     The advantage of precalcined limestone  as seen  in Table III-3  is signi-
ficant.  The weight requirements, which  are  expressed on an uncalcined stone
basis, are half the dolomite requirements and  as little  as  40% of the lime-
stone requirement without precalcination.

NOX Emissions

     NOX emissions measured in all miniplant runs, including those reported
previously, are plotted in Figure 111-24 as  a  function of percent excess air.
Data obtained in the runs made in this reporting period  follow the same trend
line and are within the data scatter of  the  earlier  runs.   The NOX emissions
were found to vary from 50 to 200 ppm or 0.04  to 0.17 g  (as N02)/MJ (0.1 to
0.4 Ib/M BTU).  Though the operating conditions  varied greatly,  the only
statistically significant variables were the excess  air  (or the  flue gas
oxygen concentration) and bed temperature.   The  NOX  emissions increased 4
fold, from 0.04 to 0.17 g/MJ over the 5  to 110%  range of excess  air.  The
temperature effect in the 670 to 940°C  (1250 to  1750°F)  range was secondary
and caused only a 25% increase in the emission level.  The  emissions are well
below the EPA new source performance standard  of 0.3 g (as  N02)/MJ  (0.7 Ib/M
BTU) and have an average value of only 0.09  g/MJ (0.2 Ib/M  BTU)  at  15% excess
air, the level most likely to be used in a commercial size  boiler.

     Previous studies carried out by Argonne (4) and Exxon  (7) indicated
that NOX emissions from an atmospheric FBC unit  could be reduced by a two
stage combustion process.  In this process,  a  substoichiometric  quantity of
air would be used to fluidize the bed and partially  burn the coal.  The
balance of the combustion air would be added at  a  slightly  higher elevation
                                      51

-------
                                         TABLE  II1-3.   SORBENT REQUIREMENTS
Ul
Ca/S Wt. Uncalcined Sorbent/100 Wt . Coal
S (%) Retention (%)
2
3
4
5
Residence Time
Temperature
59
73
79
84
2 s
>930°C
Precalcined
Limestone Limestone Dolomite Limestone
1.3 0.8 0.8 8.2
2.1 1.0 1.0 20
2.8 1.2 1.2 34
3.2 1.3 1.3 51

Precalcined
Limestone
5.0
9.4
15
20

Dolomite
10
20
29
40

       S0?  Emissions    1.2 lb/M BTU

-------
                                FIGURE 111-24
0.8
                        CORRELATION OF  NO  EMISSIONS
                                         X
   0.7 -




   0.6




5   0.5


D

E   0.4
"•s.



^x  0.3




   0.2




   0.1




     0
                                           i
   o
               20
40
60        80


EXCESS AIR, %
100
120
140

-------
in the combustor.  In the zone between the coal injection point and the point
where the secondary air was added, a reducing environment would be established
which would promote NO destruction reactions such as reduction with CO, H2 or
carbon.  This concept was tested by Argonne and Exxon and was found to reduce
NOX emissions significantly.  No extensive, planned experimental work has been
carried out to investigate the use of staged combustion in a pressurized FBC
system.  However, some unplanned results were obtained in a miniplant run
which suggests that staged combustion may have merits in a pressurized system
as well.  In one run (run 47), air compressor problems resulted in poor com-
bustion of coal and the formation of CO levels in the flue gas up to 1800 ppm.
As the CO concentration increased, the NOX concentration dropped to 10 to 20
ppm.  This is shown in Figure 111-25.  These very low NOX levels measured
under reducing conditions indicate that two stage combustion may be a pos-
sibility for pressurized FBC and could reduce NOX emissions to almost insigni-
ficant levels.

Other Gaseous Emissions
     803 emissions in the flue gas were found to vary widely, usually over a
range of 0 to 30 ppm with some measurements as high as 50 to 150 ppm.  No
correlation was found with operating conditions and it is postulated that
the 803 was most likely formed in the flue gas sampling system.

     CO emissions were very low, generally in the range of 50 to 200 ppm at
bed temperatures above 825°C except in those cases, as described above, where
upsets occurred which produced poor coal combustion conditions.  As the tem-
perature was reduced below 825°C, CO emissions increased sharply to 300 to
800 ppm at 700 to 750°C.

Particulate Emissions

     Particulate emissions in the flue gas downstream of the secondary
cyclone were measured using the system described on page 35.

     Tables III-4 and III-5 summarize particulate emission results.  The
first table gives results before the sampling system was modified by place-
ment of a second Balston filter immediately after the first filter.  The
second table gives the particulate test results after the modification.  Data
are also shown after further modification (see p. 35) and after the granular
bed filter was operated.  In Table III-4 some deviation in grain loading can
be observed.  When the loading values were >4 g/m3, it was due to the plug-
ging and ineffectiveness of the secondary cyclone during those particular
runs.  Run 43 showed consistently low loadings, and it may be related to
the fact that Illinois No. 6 coal was used in that run.  In most other runs
the values fluctuate between about 1 to 2.3 g/m3 (0.4 to 1.0 gr/SCF) giving
a mean value of 1.5 g/m^ (0.65 gr/SCF) with a standard deviation of 52%.
After system modification, however, and before the granular bed filter was
operated, particulate loading value (Table III-5) were quite constant with
a mean value of 2.3 g/m3 (1.0 gr/SCF) and a standard deviation of 13%.
These results seem to indicate that when low loadings were observed before
system modification, they may have been due to particulates by-passing the
first filter either due to a leak developing in the cartridge or displacement
in the cartridge-housing assembly.  Such problems may occur due to cartridge


                                      54

-------
                            FIGURE 111-25
                      N0v VS. CO IN FLUE GAS
                        X
Q.
n.


LO

<
o

LJ
Z)
   110
   100
    90
   80
    70
L_


\   60
o

<   50
u
a
^
o
o
  40



x 30



  20



  10



    0
         O
                _L
                                       O
_L
_L
_L
J_
_L
      0   200  400  600  800 1000      1400     1800

                 CO CONCENTRATION IN FLUE GAS (ppm)
                                                        2000
                                 55

-------
                                     TABLE  III-4.   PARTICULATE SAMPLING SUMMARY I
                                       (BEFORE  MODIFICATION TO SAMPLING SYSTEM)
Ln
            Run No.
26
28
28
28
29
30.1
31
32.2
32
33
34
36
37
39
39
43
43
                .3
                .1
                .1
                .2
                .2
                .3
             43.4
             43.5
             45
             47
             48
Sampling Time
     (h)
    13.58
     1
     1
     1
     2,
     2.
     2.
     2,
     2,
     5,
     1,
     2,
  70
  30
  25
  58
  58
  00
  75
  50
  47
                             4.42
       58
       00
       5
     3.5
  ,50
  .00
2.50
1.00
               2.08
Total Solids
  Collected
     (g)
    224.5
     16.9
      4.0
      7.0
     19.0
     14.0
     45.0
     25.0
     27.0
     95.0
     85.0
     77.0
     86.0
    175.5
    260.4
     14.5
      9.9
     11.0
      4.5
                       38
                                                              Deviation from
                                                           Isokinetic Sampling
+42
0
0
0
-16
-22
-7
+4
-12
+11
+11
+10
-48
+3
+3
-4
+10
+12
+12


0.80
0.59
0.37
0.30
0.53
0.43
2.88
0.92
1.65
1.42
1.58
2.20
2.36
4.81
4.81
0.73
0.37
0.32
0.29
1.15
5.03
                                       -3
Particulate Loading
in Flue Gas
(K/m3) (Rr/SCF)
0.80
0.59
0.37
0.30
0.53
0.43
2.88
0.92
1.65
1.42
1.58
2.20
2.36
4.81
4.81
0.73
0.37
0.32
0.29
1.15
5.03
1.53
0.35
0.26
0.16
0.13
0.23
0.19
1.26
0.40
0.72
0.62
0.69
0.96
1.01
2.10
2.10
0.32
0.16
0.14
0.13
0.4
2.2
0.67

-------
TABLE III-5.  PARTICULATE SAMPLING SUMMARY II
(AFTER FIRST MODIFICATION TO SAMPLING SYSTEM)
       Total Solids
        Collected
Deviation from
Particulate Loading
     in Flue Gas
Run No.
50.12
50.2
50.4
50.5
51 (Catch 1)
51 (Catch 2)
52

57
59 (Catch 1)
59 (Catch 2)
59 (Catch 3)
(h)
5.28
1.15
1.48
0.58
0.63
0.60
0.50

1.08
1.17
13.33
2.00
(8)
112.63
35.09
33.69
13.31
14.78
13.64
14.73
RUNS WITH GRANULAR
(AFTER SECOND MODIFICATION




(%)
-5.8
+1.8
-7.4
-7.4
-4.3
-5.8
-0.4
BED FILTER
TO SAMPLING SYSTEM)
-0.9
+1.4
-3.0
-0.4
(g/m3)
2.60
2.88
2.31
2.33
1.95
1.92
2.65





(gr/SCF)
1.05
1.26
1.01
1.02
0.85
0.84
1.07

0.08
0.08
0.28
0.54

-------
Balston Filter
2.60
2.88
2.31
2.33
SASS Train
1.81
1.92
2.86
2.79
overloading and/or a high pressure drop across the filter.  Though a slight
fluctuation in grain loading may be expected for different runs, an average
value of 2.3 g/m3 (1 gr/SCF) may be used at present.

     Operation with the granular bed filter resulted in lower particulate
levels.  This is discussed further in Section V.

     A comparison was made between the loadings measured using the Balston
total filter system and a SASS train (3 cyclones plus a total filter) used by
Battelle during runs 50.1 to 50.5 (18).  As seen in Table III-6, the loadings
compare well, with an average deviation of 30%.

              TABLE III-6.   COMPARISON OF PARTICULATE LOADINGS
         MEASURED BY TOTAL FILTER (EXXON) AND SASS TRAIN (BATTELLE)

                                                          3
                              	Particulate Loading (g/m )
             Run No.
               50.1
               50.2

               50.4
               50.5

     Individual filter loadings and temperatures are given in Table III-8
for Runs 50 to 59.  The required high temperatures were realized in Run 52.
However, when the granular bed filter was operated, the temperatures were
lower due to heat loss in the granular bed filter.  Future runs will aim
towards measurement near the cartridge allowable temperature of 530°C for
the first two filters and 206°C for the third low temperature filter.  A
significant amount of material was collected in the second filter at times.
Much of the material collected in filter 3 (low temperature) was often
greenish-black in color and was due to corrosion of the piping and the heat
exchanger tubes.

     Finally, Tables III-7 and III-9 show particulate distributions.  The
values in the two tables, corresponding to runs before and after system
modification, are slightly different.  The later values, Table III-9, show
a mass median particle size of 3 to 6 ym compared to earlier values of 6
to 8 jam. Runs made with the granular bed filter (Runs 57 and 59) show a
smaller median particle size of 2 to 5 ym as expected.

                 TABLE III-7.  SIZING OF PARTICLES OBTAINED
                   ON BALSTON FILTERS BEFORE MODIFICATION
  Run No.
    31
    32.2
    32.3
    33
    36.2
    37
Particle Size (]jm)
10%
Less Than
1.8
2.0
2.3
3.3
2.5
2
25%
Less Than
3.1
3.1
3.3
4.8
4.1
3.3
50%
Less Than
8.0
5.8
5.9
7.6
7.5
6.5
75%
Less Than
23.5
12
13
13.5
14.3
14.5
90%
Less Than
~50
24
32.5
25.5
30
30
                                     58

-------
                           TABLE II1-8.   INDIVIDUAL FILTER RESULTS (RUNS 50 TO 59)
                            Filter  1
Filter 2
Filter 3
Run
50.1
50.2
50.4
50.5
51 (Catch 1)
51 (Catch 2)
52
57(2)
59 (Catch 1)
59 (Catch 2)
59 (Catch 3)
Op. Temp.,(l) °C Loading, g
55 to 330 91.4
90 to 270 35.0
50 to 360 15.2
80 to 290 12.7
125 to 390 12.79
170 to 390 12.94
370 to 500 12.63
230(3)
360
380
390
Op. Temp., °C Loading, g
50 to 290 19.4
70 to 220 0.3
40 to 310 17.9
70 to 240 0.01
80 to 310 0.3
95 to 310 0.32
255 to 440 2.21
170(3>
240
270
270
Op. Temp., °C Loading, g
70 to 170 1.9
80 to 130 0.04
70 to 150 0.5
75 to 140 0.6
100 to 180 1.69
100 to 180 0.38
200 to 210 0.09




(1)   Operating temperature at the beginning and end of sampling time.
(2)   Runs 57 and 59 made with granular bed filter
(3)   At end of sampling time.

-------
                 TABLE III-9.  SIZING OF PARTICLES OBTAINED
                    ON BALSTON FILTERS AFTER MODIFICATION


                                     Particle Size (ym)
         Run
    50,
    50,
    50,
    50,
    51
    51
    52
    57
    59
    59
    59
1
2
4
5
(Catch 1)
(Catch 2)
(Catch 1)
(Catch 2)
(Catch 3)
10%
Less Than
1.0
0.9
1.0
1.0
1.1
0.8
1.1
1.8
1.8
1.0
1.0
30%
Less Than
2.3
1.6
1.6
2.2
3.7
1.9
2.7
3.0
2.1
1.5
1.7
50%
Less Than
4.5
3.0
2.7
3.8
7.8
3.0
5.6
5.4
2.7
2.0
2.6
70%
Less Than
8.8
7.8
5.2
6.4
22.5
6.0
13.0
10.0
3.9
2.7
5.6
     A comparison was also made with the particle size distributions measured
by  Battelle using the SASS train (18) and by Acurex using a cascade impactor
for run 50.5  (17).  The result, shown in Table 111-10, show good agreement
with those measured by Exxon by using the Coulter Counter method on particles
captured by the Balston filter.

                 TABLE 111-10.  COMPARISON OF PARTICLE SIZE
              DISTRIBUTIONS MEASURED BY THREE DIFFERENT METHODS
                                              Particle Size (yim)
Laboratory

 Exxon

 Battelle

 Acurex
           Method
      Coulter Counter
      of Filter Sample

      SASS Train

      Cascade Impactor
   10%
Less Than

   1.0

   1.2

   1.0
   0.8
   30%
Less Than

   2.2

   2.6

   2.8
   2.0
   50%
Less Than

   3.8

   5.0

   5.8
   3.7
   70%
Less Than

   6.4

  10
  11
   7.4
     Attempts made to date to determine if the particulate loading, size
and composition are affected by condensation have yielded inconclusive
results.  The sampling temperatures have been too low.  The new, high tem-
perature system should help determine if condensation occurs at temperatures
less than 870°C and if this does influence the particulate measurements.

Combustion Efficiency

     Carbon combustion efficiency results published in the previous report
(1) indicated that a regression analysis determined that temperature was
the only variable influencing combustion efficiency.  However, a plot of
combustion efficiency against temperature showed three distinct curves with
a fair degree of data scatter.  The reasons for the three curves and the
                                     60

-------
scatter were not explained at  that  time.   Combustion efficiency data were
again studied using multiple regression analysis.   Data from 55 miniplant
runs were correlated using 29  possible  variables including interactions
between variables.

     The best fit correlation  indicated that  the combustion efficiency is
a strong function of temperature  (T)  and also the excess air level  (EA), a
result which is consistent with data  reported by other  investigators  (10,11)
and also found in previous batch  unit results by Exxon  (1).

     The other variables  that  showed  a  significant effect on combustion
efficiency were the cross product of  temperature and excess  air level and
the cube of the residence time (t).   The equation correlating these variables
to combustion efficiency  was determined to be:


     C.E.(%) - 161.3 = 1.62 X  10'1  (T)  + 1.0  X 10~4 (T2)  +

               7.51 X 10~5  (T)  (EA) - 4.62 X  10~4  (EA)2 + 1.73  X 10~2 (t)3

     where T is in °C, EA in % and  t  in s.

     The correlation coefficient  corresponding to  this  fit  is 82%.  The
standard error of the correlation is  0.6 combustion efficiency  units.

     A plot of the calculated  combustion efficiencies using  this equation
versus the observed combustion efficiency is  presented  in Figure 111-26.  One
explanation for the apparent data scatter in  the graph  is the difficulty in
correlating a large number of  observations with such a  small span of the
dependent variable.  It should be noted that  the miniplant uses a particulate
recycle system to increase combustion efficiency.   This may  have the effect
of reducing to insignificance  variables which may  be highly  significant in
a once through system without  particulate recycle.   Additional  work will be
required to compare the efficiencies  measured in a once through system with
those measured in a recycle system.   A  preliminary estimate  of  the effect of
recycle was obtained when a single  run  (run 51)  was made  with no recycle from
the first stage cyclone to the combustor.   The cyclone  was deactivated for
this run by slipping a blind between  the flange at the  bottom of the cyclone
and the top of the return line.   The  combustion efficiency in this run made
at 875°C at an excess air level of  45%  was 98.2%.   Runs made under similar
conditions but with recycle, gave combustion  efficiencies 1.0 to 1.5 per-
centage points higher.

Cyclone Efficiency

     The collection efficiency of the second  stage miniplant cyclone was
calculated using collection and particle size data obtained  in  recent runs
(Runs 48, 50.3 and 51).   The overall  secondary  cyclone  efficiency was as
follows:
                    Run           Overall  Efficiency, %
                                             84
                                             84
                                             86

                                      61

-------
o
LU
CO
ID
ca

o
o

Q
LJ
O
O
                         FIGURE 111-26

         COMBUSTION EFFICIENCY CALCULATED VS. MEASURED
   100
    99
97
    96
     95
               T
                              T
                        1
                              1
          96        97        98        99


           MEASURED COMBUSTION EFFICIENCY (%)
100
                             62

-------
The values given above indicate the cyclone  efficiency  to be  the same within
experimental errors.  It should be noted  that  during Run 51 the primary
cyclone was not used, and the secondary cyclone was used alone.  Yet the
overall cyclone efficiency value  from Run 51 corresponds closely to the
results from those runs in which  the  primary cyclone was also operating.
The efficiency curves for all 3 runs  are  given in  Figure 111-27.  With the
amount of data available at present,  it is difficult to state whether differ-
ences in the efficiency curves are due to the  operation of the second cyclone
by itself or not.  However, these curves  are comparable to those obtained
from earlier runs.  The 50% cut point lies between 3 to 3.5 pm, indicating
a relatively high efficiency for  a conventional cyclone.

     Particle size distribution of the material collected in  the second stage
cyclone is typified by results for Runs 50.1,  50.2 and  50.3 shown in Table
III-ll. These distributions are similar to those reported previously (1).
Additional size distribution data for other  runs are given in Appendix H-2.

                      TABLE  III-ll.  SECOND STAGE CYCLONE
                   LOCK HOPPER PARTICLE SIZE DISTRIBUTIONS


             	Particle  Size  (ym)	
              10 Wt.  %     25 Wt.  %      50 Wt.  %      75  Wt. %     90 Wt. %
 Run No.      Less  Than    Less  Than     Less  Than    Less Than   Less Than

   50.1           7            13            22           52           98

   50.2           6             9            17           38           66

   50.3           6            10            17           37           69
                                       63

-------
                 FIGURE  111-27



SECONDARY CYCLONE EFFICIENCY VS. PARTICLE SIZE



o1^-
-
o
UJ
o
u_
1 1
u_
UJ
2;
o
1-
o
UJ
	 1
o
0





100
90

80

70

60

50

40
30

20
10
0
i i i i i i i i i i ii
^OJ=s=.g==— °^°^ "
0^8^^° * ^\S"
X^A / V
/A xD
^^ S
/ /v
Of
- A s —
/ /
//D
I/
8 RUN No.
/A 048
A D 51

^y A5Q.3
a>/O
nn^X^^ | I 1 1 1 1 1 1 1 1 II

2 3 4 5 6 7 8 10 20 30 40 5
              PARTICLE SIZE,

-------
                                  SECTION  IV

                             REGENERATION  STUDIES
     Work is continuing to develop a process  for  continuously  regenerating
sulfated limestone or dolomite sorbent  (CaS04)  in a  fluidized  bed by reaction
with a reducing gas at about 1100°C and  elevated  pressure.   In the previous
annual report (1), experiments were described in  which batches of sulfated
sorbent were regenerated in the miniplant  regeneration vessel.   In those
preliminary tests, up to 3.7 mole percent  (dry basis) of  S02 was produced,
reduction of CaSO^ to CaO was nearly complete,  and agglomeration of the
fluidized bed was avoided by careful control  of temperature.   The regenerator
was coupled to the miniplant combustor so  that sorbent could be continuously
recirculated between the two vessels and a 24 hour shakedown run was com-
pleted.  Since the last report, the operability of this system was demonstrated
in a five day run in which both combustor  and regenerator operated without
interruption.  In this report, the equipment  and  procedures used to make
the run, and the experimental results are  described.

EQUIPMENT, MATERIALS, PROCEDURES

Equipment

     The equipment and materials were described in detail in the previous
report  (1).  In most cases, no major changes  have occurred since and only
summaries are reported.  When changes have occurred, for  example in the fluid-
izing grid for the regenerator, a detailed description is given here.  Also
described here are the design and operation of the system used  to transfer
sorbent between the combustor and regenerator.

Air System

     The two separate air systems are  (1)  burner  air and  (2) supplementary air.
All air is supplied by the main air compressor.   Automatic control systems,
consisting of control valves, flow measuring  orifices, and electronic con-
trollers, are used to regulate air flow.   Burner  air is supplied to the bur-
ner, located beneath the fluidizing grid,  in  sufficient quantity to completely
burn the fuel (natural gas).  Supplementary air is added  about  halfway up
the bed in order to create an oxidizing  zone  in the upper portion of the bed.

Fuel System

     The two fuel systems are (1) burner fuel and (2) supplementary fuel.
Automatic control systems, similar to those used  for air  flows, are used to
regulate the flow of natural gas.  Burner  fuel  is supplied to  the burner
where it is burned with an approximately stoichiometric amount  of air.
Supplementary fuel is added directly to  the regenerator column  just above the
fluidizing grid in order to produce reducing  gases (CO, H£) .

Off Gas Handling

     Hot pressurized gases leaving the regenerator are cooled  in a single
pass double pipe heat exchanger and expanded  to nearly atmospheric pressure
across a control valve.  Dust is removed from the gas upstream  of the cooler
                                      65

-------
by a cyclone and upstream of the pressure control valve by a sintered stain-
less steel bayonet-type filter.  The gas is filtered in order to reduce
wear on the pressure reducing valve.

     Off-gases from the regenerator are sent to a Research-Cottrell scrubber
for cleanup before venting.  An alkaline solution (Na2C03) is used in the
scrubber to absorb S(>2.

Gas-Sampling System

     A slipstream is taken downstream of the pressure reducing valve.  The gas
is filtered  (Balston Model 33 filter) and dried (Perma-Pure Model PD-1000-24S
self-regenerative membrane-type dryer) before entering the analyzers.

Fluidizing Grid

     The fluidizing grid had failed due to overheating at the conclusion of
the 24-hour shakedown run of the combustor-regenerator.  This grid had 88
holes of 3.6 mm  (9/64 in) diameter for passage of the fluidizing gases (from
burner located beneath grid) and 14 water cooling channels of 4.8 (3/16 in)
diameter.  Cooling water was supplied through two 7.7 mm (0.305 in) I.D. tubes
to a header and  flowed in parallel through the cooling channels.  Examination
of the grid showed that the metal surrounding the outer channels had been
overheated.  It  was concluded that the flow of water through the channels was
non-uniform, with the inner channels receiving most of the water.

     In order to achieve more uniform flow of water throughout all the cooling
channels, the grid was modified so that water flow could be controlled inde-
pendently through six separate cooling zones.  Hence, the 14 channels were
separated into groups of 3, 2, 2, 2, 2, and 3, each group or zone having its
own water supply.  The new design apparently solved the problem:  after the
125 hour demonstration run, the grid was examined and found to be in
excellent condition, with no signs of overheating.

Burner

     No changes were made to the regenerator burner since the last report.
This unit is identical to that used in the miniplant combustor and is des-
cribed in the previous annual report (1).

Sorbent Transfer System

     Combined operation of the combustor and regenerator required development
of a transfer system to circulate sorbent between the two vessels.  A number
of alternative approaches were considered and tested (1).  Basically, these
approaches could be divided into two types:  high and low density solids
flow.  The approach thought to be most suitable for the miniplant utilized
high bulk density (stick-slip) flow of sorbent in transfer lines.  This
technique had the important advantages that  (1) large amounts of carrier
gas would not be needed, as would be the case if pneumatic conveying were
used, and (2) short, straight transfer lines could be used.  The small distance
between the combustor and regenerator would have made it difficult to accom-
modate the U-shaped transfer lines that would have been required in a pneum-
atic system.
                                     66

-------
     Figure IV-1 shows schematically  the  sorbent  transfer  system.  Pressure
in the regenerator was maintained  slightly  higher than  that  in  the combustor.
Solids in the regenerator-to-combustor  transfer line  moved into the combustor
when a pulse of nitrogen was applied  to the lower end of the transfer line.
The flow rate of solids was controlled  by adjusting the frequency, duration,
and intensity of the pulse.  Two slide  valves  were used in the  combustor-to-
regenerator transfer line  in order to prevent  back flow of gas  from the
regenerator up the line.   These automatic valves  trapped solids in the piping
between them.  Solids were discharged into  the regenerator when the bottom
valve was opened.  The two solids'  take-off plugs shown in Figure IV-1 were
inserted into the ports during start-up to  prevent solids  from  entering the
lines.  Plugging could occur if the solids  became wet due  to water condensa-
tion during start-up.  The manual  slide valve  in  the  regenerator-to-combustor
line was also closed during start-up  and  also  during  upsets.  A photograph
may be seen in Figure IV-2.

     The components of the sorbent transfer system (valves,  expansion joints,
etc.) are described in detail in the  previous  annual  report.  The transfer
lines themselves were fabricated from 6 inch Schedule 40 carbon steel pipe
and refractory lined to an inside  diameter  of  7.6 cm.   The sloping portions
of the lines were sleeved  with 2-1/2  inch Schedule 10 316  stainless steel
pipe, which had an inside  diameter of 6.7 cm.

Miscellaneous Modifications to Equipment
Made Prior to Demonstration Run	

     In addition to redesigning the regenerator fluidizing grid, a number of
other changes were made to the combustor-regenerator  prior to the demonstra-
tion run in order to improve reliability.   The more noteworthy  changes were:

      (a)  Spare supplementary fuel and  air  inlet  ports  were  added.
          If a plug occurred at one inlet port, the second could
          be used to continue the  run.

      (b)  The double pipe  off-gas  cooler  was shortened  in  order
          to raise the temperature of gas exiting from  the cooler
          and reduce water condensation.

      (c)  A second pressure control valve was  added in  parallel
          with the original valve.  One valve  was used  at  a  time
          but the second could be  put in  service  without inter-
          rupting the run  if a problem  occurred.

      (d)  A new thermocouple port  was added just  above  the
          fluidizing grid  in order to detect high temperatures
          in this critical region.

      (e)  The system to measure pressure  drop  across  the fluidizing
          grid and across  different points  in  the bed was  modified
          to be more reliable and  provide more accurate data.
                                     67

-------
REGENERATOR
COMBUSTOR
              SOLIDS TAKE
              OFF PLUG
                           AUTO
                           SLIDE
                           VALVE
                          MANUAL
                           SLIDE
                           VALVE
                 NITROGEN
                   PULSE
                FIGURE IV-I
MINIPLANT SOLIDS TRANSFER SYSTEM (SCHEMATIC)
                    68

-------
            FIGURE IV-2




MINIPLANT SOLIDS TRANSFER SYSTEM
                 69

-------
Materials

     No changes in materials of construction occurred during the past year.

     The regenerator is lined with General Refractories Litecast 75-28, a
castable refractory with a service temperature limit of 1540°C (2800°F) and
a thermal conductivity (at 540°C) of 0.55 W/m2°C (3.8 BTU/hr ft2 °F/in).

     The thermocouples used to measure fluidized bed temperature are Type K
(chromel-alumel) protected by a silicon carbide sheath.  Silicon carbide
provides excellent resistance to high temperatures and chemical attack.
Because this material has very little strength in tension, an Inconel tube
is used inside the sheath to provide reinforcement.

     The alloy used for the gas-contacted sections of the regenerator off gas
cooler and filter is type 316 stainless steel.  This material offers good
resistance to dry and wet S02 although H2SO^ can cause serious corrosion even
at low concentrations.  This has been a problem with the filter element which,
because of its high surface area, is more susceptible to corrosion.  Corrosion
has been minimized by thoroughly washing and drying the filter element after
each run.

EXPERIMENTAL RESULTS AND DISCUSSIONS

Combustor-Regenerator Demonstration Run

     A successful combustor-regenerator demonstration run (Run 45) was com-
pleted in October, 1976.  The miniplant was operated with limestone sulfur
acceptor recirculating between the combustor and regenerator until the run
was voluntarily terminated after 125 hours of continuous operation.  No inter-
ruptions in coal feed to the combustor or fuel feed to the regenerator
occurred during the extended run.

     Two unsuccessful attempts preceeded the successful run.  The first
was shut down after only four hours when a temperature runaway in the com-
bustor resulted in agglomerated bed.  This problem was caused by a malfunction
in the load cell amplifier for the coal feed system.  Electrical problems
also plagued the second attempt and the run was terminated when a malfunction
of the AP transmitter in the control loop for regenerator fluidizing air
caused the air flow to rapidly increase, thereby blowing a portion of the bed
out of the regenerator.

Operating Plan and Conditions

     The major purpose of the run was to demonstrate that the combustor-
regenerator could be operated continuously for 100 hours.  Thus, operating
conditions were deliberately conservative in order to maximize the chance of
reaching this goal.  It was intended to keep all operating conditions steady
throughout the demonstration run, with the exception of the feed rate of
fresh (makeup) limestone into the combustor (Ca/S ratio).  Operating condi-
tions are summarized in Table IV-1.
                                     70

-------
                   TABLE IV-1.  OPERATING CONDITIONS
                       DURING DEMONSTRATION RUN
Pressure, kPa
Bed Temperature, Average, °C
Bed Height, Expanded, Avg., m
Superficial Gas Velocity, m/sec
Solids Recirculation Rate, kg/hr
Makeup Acceptor Addition Rate,
  Equiv. Ca/S, Average
  Range
Combustor Coal Feed Rate, kg/hr
Coal Type
Stone Type
Combustor
  760
  900
    3.4
    1.5
   79
                45

               0.55
              0-1.3
Regenerator
    770
   1010
      2.3
      0.6
Champion (Pittsburgh Seam),  2.0%  S
Grove Limestone No. 1359
                                  71

-------
     One of the most pressing questions the operator of a fluidized bed com-
bustion system could ask is what savings in sorbent would be realized if a
regenerator were added to his system.  In order to make a first attempt at
answering this question, it was decided to feed the smallest quantity of makeup
sorbent into the combustor that would allow operating the system with (1) SC^
emissions from the combustor below the EPA emission standard, and (2) constant
fluidized bed levels in both the combustor and regenerator.  In other words,
the amount of fresh limestone required would be determined either by the
activity or the attrition rate of the sorbent.  As it turned out, the limiting
factor during this run was maintaining bed levels.  Over the 100 hour period
of regeneration, S02 emissions from the combustor were always below the EPA
standard of 1.2 Ibs S02/106 BTU, but makeup limestone had to be added because
bed levels declined during the first half of the run.  However, it is quite
likely that had the run been continued, makeup limestone would have been
needed to increase sorbent activity and control S02 emissions.

     There was an alternative manner in which the system could have been
operated.  Since it was not known in advance exactly what attrition rates
would prevail, a makeup rate could have been chosen which was sure to be
higher than the maximum attrition rate expected, say 2.0/1 or 3.0/1 Ca/S.
The limestone makeup rate would then have been constant throughout the run
and bed levels could have been maintained by varying the rate at which stone
was rejected from the combustor.  However, this would have been a very
unrealistic way of operating a real plant, since far more fresh sorbent would
be used than was needed.  Furthermore, operating the system in this manner
would not have provided any information on the degree of reduction in sorbent
requirements possible in a regenerative system.

     For the first 24 hours of the run, the regenerator was operated under
oxidizing conditions in order to establish baseline operating conditions.
Sorbent was recirculated between the combustor and regenerator during this
period.  Subsequently, reducing conditions were established by increasing
the flow of supplementary fuel to the regenerator.  Operation of the system
continued, uninterrupted, for the next 100 hours.

Operating Performance

     Operation of the combustor-regenerator during the extended run was
exceptionally smooth and no interruptions in coal feed to the combustor, fuel
feed to the regenerator, or in the recirculation of solids occurred during
the 100 hour period when the regenerator was in reducing conditions.  The
only potentially serious problem was a hot spot just above the regenerator
fluidizing grid.  The hot spot developed during startup as the regenerator
bed (fresh limestone) was being heated under oxidizing conditions.  When the
bed temperature reached about 1000°C a sudden shift in the temperature pro-
file occurred and the temperatures at the bottom of the bed quickly climbed
to about 1200°C.  Temperature was reduced by decreasing the air and fuel
inputs.  The regenerator was maintained at constant oxidizing conditions for
24 hours in order to establish baseline operation in the combustor.  Transfer
of solids occurred continuously throughout this period.
                                      72

-------
     When reducing conditions were established  in  the  regenerator, temperature
of the hot spot increased to about 1260°C.   Fuel and air  flows were decreased
and pressure was reduced from the initial pressure of  920 to  the ultimate
operating pressure 760 kPa.  The reduction  in pressure caused an increase in
superficial gas velocity, which improved mixing of solids and flattened the
temperature profile somewhat.  Superficial  gas  velocity was 0.6 m/s, which
is just slightly above the minimum velocity to  achieve fluidization.

     There were no changes in operating conditions in  the regenerator for
the duration of the extended run.  However, the hot spot, which was initially
located at a position about 13 cm above the fluidizing grid,  increased in
size so that towards the end of the  run, high temperatures were present as high
as 43 cm. above the grid.  By comparison, the reducing zone of the regenerator
extended to 74 cm above the grid and the average total expanded bed height was
230 cm.

     After the run, the plenum of the regenerator  was  dismantled and the
fluidizing grid and bed examined.  The grid was in excellent  condition and no
signs of overheating were present.   There was a "crust" of bed material which
covered the grid to a depth of about 8 cm,  this material  was  loosely packed
and porous and would not have interfered with flow of  air through the grid.
Above the crust was about 30 cm of hard, fused  bed, much  of which must have
formed during shutdown.  Upon shutdown, air was blown  through the bed and
temperatures at the bottom of the bed increased over 1000°C.  The temperature
excursion was probably caused by oxidation  of calcium  sulfide which was pre-
sent in the reducing zone of the bed. Normally, nitrogen is  used during
shutdown to prevent this type of occurrence.  However,  in this case, the
compressed nitrogen supply was inadequate and compressed  air  was used instead.

Makeup Limestone Addition Rate.
Fluidized Bed Levels
     Makeup limestone was  added  to  the combustor  in order  to maintain reason-
 ably constant  fluidized  bed  levels  in the combustor and  regenerator.  The feed
 rates of makeup  limestone  are  summarized in Table IV-2.  The makeup rate
 from the start of  the run  was  3.6 kg/hr, which is equivalent to a Ca/S molar
 ratio (Ca  in stone to S  in coal) of 0.74.  This makeup rate was chosen because
 bed levels during  the 24-hour  shakedown run made  in July 1976 were kept con-
 stant using this rate.   Bed  height  in both combustor and regenerator is
 plotted against  hours into the run  in Figure IV-3.   It can be seen that bed
 levels in  the  combustor  began  to drop after about 40 hours.  When levels
 dropped to the height of the takeoff port to the  combustor-to-regenerator
 solids transfer  line, the  feed rate of makeup stone was  increased to 1.30
 Ca/S.  Failure to  increase the makeup rate would  have resulted in the level
 of solids  in the combustor falling  below the takeoff port. This would have
 caused the transfer of solids  to stop.

     The total mass of solids  in the combustor and regenerator is plotted vs.
 time into  the  run  in Figure  IV-4.   As can be seen,  inventory was dropping
 fairly sharply after about 40  hours and feed rates of makeup limestone had
 to be increased.
                                      73

-------
             TABLE IV-2.  FEED RATES OF MAKEUP LIMESTONE TO
        COMBUSTOR DURING COMBUSTOR-REGENERATOR DEMONSTRATION RUN
  Hours           Feed Rate of Makeup              Feed Rate of Makeup
Into Run            Limestone, kg/hr           Limestone,  Equivalent Ca/S

 0-5                    0                               0

 5-24                   3.6                             0.74

   24                  Reducing Conditions Established in  Regenerator

24 - 56                   3.6                             0.74

56 - 65                   6.4                             1.30

65 - 80                   5.2                             1.06

80 - 86                   2.3                             0.46

86 - 124                  0                               0
                                    74

-------
   500
   400
   300
   200
   100
                                    FIGURE IV-3



                       COMBUSTOR AND REGENERATOR BED HEIGHTS
E
o
LU

CO
     0
0.74 Ca/S
             i    i   i
                              <-1.30-»f— 1.06-^*0.46
0
                                         RUN 45


                                         o COMBUSTOR

                                         o REGENERATOR
H
      25 30  35  40  45 50  55  60  65 70  75 80  85  90  95 100105110115120125
                                   HOURS INTO RUM

-------
                                 FIGURE IV-4
          TOTAL INVENTORY OF SORBENT IN THE COMBUSTOR AND REGENERATOR
340
                   Ca/S INCREASED
                         I
  25  30 35  40 45  50 55  60 65  70  75  80 85  90 95100105110115120125
                               HOURS INTO RUN

-------
     It was intended not to reject  any  sorbent from the combustor during the
run.  However, after the run,  it was  discovered that 180 kg  of  sorbent had
been inadvertently removed from the combustor early in the run.  This dis-
covery explains why high limestone  addition rates were required early in the
run*

Combustor S02 Emissions

     S02 emissions from the combustor are given in Figure IV-5.  For the first
24 hours the regenerator was  operated under oxidizing conditions with sorbent
recirculating between  the combustor and regenerator.  Emissions gradually
increased to about 550 ppm  (1.1 Ibs S02/10& BTU).  Since the S02 emissions
would have been about  1330 ppm at zero  retention, 550 ppm corresponds to 59
percent retention.  The makeup rate of  fresh limestone feed  into the combustor
during this period was 3.6 kg/hr  (7.9 Ibs/hr), which is equivalent to a Ca/S
feed ratio of 0.74  (see Table IV-2).

     It is interesting to compare the S02 emissions from the combustor during
the first day (oxidizing conditions in  regenerator) to the emissions that
would have been expected without  the  regenerator.  Figure IV-6  gives S02
emissions from once-through operation of the combustor as a  function of Ca/S
feed ratio using limestone at carbonating and calcining conditions.  The com-
bustor was operated at 900°C  (1650°F) (carbonating conditions)  so at a Ca/S
of 0.74, the expected  S02 emission  level would have been about  1100 ppm.
However, the measured  S02 concentrated  started to level out  at  about 550 ppm.
The lower than expected emission  level  can be explained because even though no
regeneration was occurring in the regenerator, calcination of the limestone
was occurring.  The calcining conditions in the regenerator  were the same as
those used in the limestone precalcination tests  (p. 48).  The  S02 retention
(59%) measured during  this period was also comparable to that measured in run
46.4 made with precalcined limestone  at similar operating conditions.  There-
fore, the regenerator  was operating as  a precalciner and the S02 emissions
from the combustor were indicative  of those expected from the use of a pre-
calcined limestone, rather than from  uncalcined limestone used  under carbona-
ting combustor conditions.

     Within about four hours  after  reducing conditions were  established in the
regenerator by increasing the flow  of supplementary fuel, S02 emissions from
the combustor fell from 550 ppm to  below 200 ppm.  After about  50 hours into
the run (25 hours under reducing conditions) S02  levels gradually increased and
reached about 550 ppm  when the run  was  terminated after 100  hours of regenera-
tion.  The increase in emissions  is,  for the present, assumed to have been
caused by a gradual decline in the  activity of the regenerated  sorbent.  About
fifteen cycles of sulfur sorption and regeneration should have  occurred during
the 100 hour period during which  the  regenerator  was in reducing conditions.

     The average feed  rate of makeup  sorbent was  equivalent  to  a Ca/S ratio
of 0.55; however, the  makeup  rate was varied over the Ca/S range of 0-1.3 in
order to maintain constant bed levels.   It should be emphasized that this
variation in makeup sorbent rate was  very small compared to  the rate at which
                                     77

-------
oo
                      600
                      500
                      400

                 a.
                 a.
                 csj
300
                      200
                      100
                                                    FIGURE IV-5

                                       COMBUSTOR  S02  EMISSIONS
                  REDUCING CONDITIONS ESTABLISHED IN REGENERATOR
                                                           I	I
                                                                        _L    I    I    I
                           0   10  20  30   '10  50   60  70  80  90100110120130140
                                              HOURS  INTO  RUN

-------
                     FIGURE IV-6

             S02 EMISSIONS VS. Ca/S  RATIO

LIMESTONE  No. 1359  IN ONCE-THROUGH OPERATION (REF. 1)
 1600
 1400
   400

    310


   200
                        CARBONATING CONDITIONS
     0
 CALCINING
CONDITIONS
      0   0.55
                       Ca/S RATIO
          (Moles Ca In Feed Sorbent/Moles S In Coal)
                           79

-------
recirculated sorbent entered the combustor, which was equivalent  to a Ca/S
ratio of about 17.  There was no apparent immediate effect of varying the
makeup rate on the emissions of SO^ from the combustor.

Levels of S02 in Regenerator Off-Gas

     The concentration of SC^ in the regenerator off-gas was nearly steady
throughout the run and averaged 0.53 mole percent (dry basis).  This is very
close to the concentration predicted by a sulfur mass balance based on the
feed rate and sulfur content of the coal entering the combustor.  The cal-
culated equilibrium concentration at the operating conditions of  the regen-
erator was 2.9 percent; hence, higher S02 levels would probably have been
achieved by burning more coal, of a higher sulfur content in the  combustor.
A commercial-sized plant would not be expected to be mass-balance limited and
this would be expected to yield higher S02 concentrations in the  off gas.

Attrition Rate

     During the 125 hour demonstration run, 336 kg (740 Ibs) of fresh lime-
stone was added to the combustor to maintain constant bed levels; however, as
noted earlier, 180 kg (396 Ibs) of sorbent was inadvertently dumped from the
combustor early in the run.  The approximate amount of sorbent lost due to
attrition and entrainment was 336-180 = 156 kg (344 Ibs).  The inventory of
solids in the system was about 280 kg (620 Ibs).  Hence, the average entrain-
ment rate was 156/125 (280) = 0.0045 kg/hr kg inventory or 156/336 =0.46
when expressed as a fraction of the limestone makeup rate.  This rate is
higher than expected based on once through (non regenerative) entrainment
measurements.  More work is needed to explain this anomoly.

Analysis of Bed Solids

     After the run, samples of bed were taken from the combustor and regen-
erator for analysis.  Three samples of regenerator bed were analyzed because
of the large variability among samples.   Only one complete analysis of com-
bustor bed was made because sulfate analysis of three samples showed very
close agreement.  The large variability in samples from the regenerator was
probably a result of sampling problems caused by the regenerator bed being
partly agglomerated.  Agglomeration occurred during shutdown when air was
blown through the hot bed,  causing an exothermic oxidation of CaS and a tem-
perature rise to over 1100°C (2000°F).  Table IV-3 gives the analytical
results.   Note that no CaS  was found.
                                      80

-------
                                  Table IV-3

                  Analyses  of  Bed Material Discharged from
              Combustor  and Regenerator After Demonstration Run

    Regenerator^
                    _ _ _ Mole Percent ___
    Component       Sample  #1       Sample #2      Sample  #T     Average

      Ca°              87.0           73.2           81.2         80.5
      CaCO               2.7            1.4            2.7          2 3
      CaS04            10-3           25.4           16.1         17;3

    Combustor

      CaO              35.8
      CaCO             18.5
                       45.7
 Sulfur Mass Balance

     Table IV-4 is a  sulfur mass  balance  for  the  demonstration run
 (entire 125 hour period).  Recovery  of  sulfur was 103.5 percent.

     The sulfur accumulated section  of  Table  IV-4 comes about because
 the system was charged with fresh limestone prior to  the run, but sulfated
 limestone remained at the  conclusion.   Also,  the  sulfur balance is very
 sensitive to the sulfur  content of the  coal.  A sulfur level of 2.0 percent
 was the most recent analysis of Champion  coal available.  The actual sulfur
 level would have had  to  be only 2.07 percent  to get a calculated sulfur
 recovery of 100 percent.

 Conclusions

     Operation of the combustor-regenerator during the demonstration run was
 encouraging.  In particular, the  sorbent  transfer system operated for 125
 hours without a single problem.   The chances  are  excellent that this system
 can be used successfully in future runs.

     Several areas leave room for improvement.  The superficial gas velocity
 in the regenerator was only 0.6 m/sec,  which  is only  slightly above the
 minimum fluidizing velocity.  As  a result, mixing of  the regenerator bed was
 slow.   This factor surely  contributed to  the  poor temperature profile and
 localized agglomeration.   The velocity  will be increased slightly in sub-
 sequent runs by using higher bed  temperatures; however, it may be necessary
 to operate the regenerator at lower  pressures or  even reduce its size in order
 to obtain suitable gas velocities.

     Another area for improvement  is the  concentration of S02 in the regen-
 erator off-gas.   During  the demonstration run, the average concentration was
 only 0.5 mole percent.  This level was  determined by  a sulfur mass balance
 for the combustor-regenerator system and  not by thermodynamics, which would
have permitted 2.9 percent S02.  Hence, the S02 level could have been higher

                                     81

-------
                                            TABLE  IV-4.   SULFUR BALANCE
                                      COMBUSTION-REGENERATION DEMONSTRATION RUN
                                                            kg           % of Sulfur Entering
        •  Sulfur Entering System
           a.  Coal                                       194.8                  100
                                                 Total    194.8                  100
        *  Sulfur Leaving System
           a.  Regenerator off gas                         91.8                   47.1
           b.  Combustor flue gas                          39.5                   20.3
oo          c.  Combustor bed reject                        17.5                    9.0
N>
           d.  Combustor overhead solids (flyash)          22.6                   11.6
           e.  Regenerator overhead solids                  2.0                    1.0
                                                 Total    173.4                   89.0
        •  Sulfur Accumulated (A Inventory)
           a.  Regenerator bed                              3.1                    1.6
           b.  Combustor Bed                               25.1                   12.9
                                                 Total     28.2                   14.5

        •  % S Recovery                                                          103.5

-------
if coal of a higher sulfur content was  burned,  and  if  the  coal  feed rate into
the combustor was higher.  Runs  planned for the future will  use operating
conditions which should result in higher regenerator S02 levels.

     Continuous operation of  the miniplant  provided a  realistic way of mea-
suring the potential benefits of a regenerative system compared to a once-
through system.  An important question  is what  reduction in  sorbent require-
ments can be realized  in a regenerative system.  Since only  one continuous
run has been made thus far, and  at non-optimum  conditions, this question
cannot be answered with a high degree of certainty. However, the average
SC>2 emissions from the combustor were 310 ppm (0.63 Ibs S02/106 BTU) at an
average Ca/S ratio of  0.55.   Figure  IV-6, based on  data from a  once-through
combustion system, shows that at least  four times this makeup rate would have
been required in a once-through  system, had the combustor  been  under calcin-
ing conditions.  Actually, carbonating  conditions prevailed  in  the combustor
and Figure IV-6 shows  that the observed emission level could not have been
reached in a once-through system at  any Ca/S makeup rate.  Another way to
show the effect of regeneration  on sorbent  requirements is illustrated in
Figure IV-6.  At the average  Ca/S ratio of  0.55 used in this run, the
expected S02 emission  from the combustor in a once-through system would have
been about 1150 ppm compared  to  310  ppm measured.

     Additional combined combustor-regenerator  runs will soon be underway
to study the effects on performance  of  the  important variables, including
sorbent recirculation  rate, regenerator temperature, and type of sorbent.
It is hoped to determine which combustor and regenerator operating conditions
provide, simultaneously, low  emissions  of SQX and other pollutants from the
combustor, high concentrations of  SC^ in the off-gas from  the regenerator, low
makeup rates of fresh  sorbent, and moderate recirculation  rates between com-
bustor and regenerator.

     In the future, the use of coal  as  a fuel for the  regenerator will be
studied.  The batch combustor and  regenerator are being converted to con-
tinuous units to study this and  other areas. One problem  is recovering
sulfur from the fairly low levels  of S02 present in the regenerator off-gas.
Other problems are sorbent deactivation and attrition. These problems need
to be understood and solved before regeneration can become a commercial
endeavor.
                                      83

-------
                                   SECTION V

                       GRANULAR BED  FILTRATION STUDIES
INTRODUCTION
     The successful development  of  the  pressurized  fluidized bed coal com-
bustion process is dependent on  the ability of  particulate control devices
to remove particulates from the  hot combustor flue  gas  to very low levels.
This must be done to assure that the expansion  of the flue gas through the
gas turbine does not cause damage to the turbine by erosion, corrosion, or
deposition of solids on  the turbine blades.  At the present time, the turbine
requirements are not well defined.   Current estimates of the allowable parti-
culate concentration in  the flue gas entering the turbine range from 45 to 1
mg/m3 (0.02 to 0.0004 grains/SCF).   In  addition to  the  gas turbine inlet
requirements, the U.S. Environmental Protection Agency  has imposed limits on
the emission of particulates from coal  fired installations of 0.043 g/MJ
(0.1 Ib/M BTU).  For a typical coal, this standard  translates to a parti-
culate concentration in  the flue gas of approximately 100 mg/m^ (0.05 gr/SCF).
Therefore, at the present time,  removal efficiencies are dictated by the
turbine requirements.  To meet these estimated  requirements, the flue gas
leaving a pressurized combustor  must first be precleaned in a two stage
cyclone system and then  be sent  to  a third stage cleanup device.  To meet
the current environmental standard  the  third stage  device will be required
to have an efficiency of approximately  67%.  To meet the estimated turbine
requirements, efficiencies must  be  in the range of  95 to 99.7%.

     The objective of the flue gas  particulate  removal  program is to
evaluate two removal devices which  have the potential for reducing parti-
culate loadings to the required  levels.  The devices will be installed on the
miniplant which has a maximum flue  gas  flow rate of about 20 Sm3/min (700
SCFM).  The first device is a granular  bed filter of a  design developed by the
Ducon Company.  The results of the  initial testing  to evaluate this device
will be covered in this  section  of  the  report.  Further testing of the
granular bed filter is planned,  and will be reported in a subsequent report.
The choice of the second device  has not been made at this point.

     The objectives of the granular bed filter  test program were to measure
the outlet loading from  the filter, determine if the removal efficiency was
maintained with use, measure operational stability  of the filter (e.g., can
a low pressure drop across the filter be maintained, does the filter plug,
is the amount of blow back gas needed to maintain steady operation within
reason, etc.), and finally, to measure  the long term life of the filter
hardware.  The primary operating parameters  were:   the  filtered gas flow
rate, usually measured as the gas velocity entering each filter bed, the
reverse flow clean gas ("blow back") velocity,  the  duration and frequency
of the blow back step, the type  of  filter media used (i.e., particle size,
shape, density), and media bed depth.
                                      84

-------
EQUIPMENT, PROCEDURES

Equipment

Single Bed Test Rig—
     To provide support information for the evaluation of the granular bed
filter, a single bed test rig was designed and tested.  A photograph of the
test rig is shown in Figure V-l.  A schematic of the test rig is shown in
Figure V-2.  The test rig was designed to filter an isokinetic sample which
was extracted through a 1.6 cm (0.5 in) probe which was located in the
miniplant ducting downstream of the second stage cyclone.  The test rig
consisted of one of the Exxon designed filter beds described in a subsequent
section enclosed in a refractory lined vessel to permit filtration at system
temperature and pressure.  The gas exiting the test rig was filtered by a
glass microfiber filter to allow a measurement of particle loading in the gas
leaving the rig.  (The particle loading of the gas entering the rig was
estimated based upon past measurements on the miniplant without the granular
bed filter.)  The pressure drop across the filter bed was measured by a AP
cell.  When blow back was required, as determined by an unacceptable pressure
drop, the sample flow was stopped and the filter vessel was vented to the
atmosphere.  The blow back cycle was then initiated using air rates up to
0.14 Sm3/m  (5 SCFM).

Original Ducon Filters—
     Three filter elements were purchased from the Ducon Company for evalua-
tion.  The Ducon filter element consists of a number of small beds packed
with a suitable filter media such as alumina, quartz, etc. and supported on
a bottom retaining screen.  A stack of the filter beds form a single filter
element.  A number of filter elements can be used depending on the volume
of gas to be filtered.  A schematic of a single filter bed is shown in Figure
V-3.  The operation of the filter is shown schematically in Figure V-4.
During the filtration cycle, dirty gas passes through the 50 X 50 mesh inlet
screens down into the filter beds.  Clean gas from the beds is collected in
a manifold in the interior of the element and then passes to the clean gas
outlet system.  As the filtration step proceeds, the pressure drop across
the element increases and eventually the element must be cleaned by the
reverse flow of clean gas.  This "blow back" occurs by flowing clean gas in
reverse direction through the outlet gas manifold, up through each filter
bed and out through the screens.  The function of the screens is to retain
the filter media during the blow back step, keeping it inside the filter
bed, while allowing the fine particulates removed from the filter media by
the blow back gas to pass through.  The fine particulate then settles outside
the filter elements and is collected at the bottom of the shroud containing
the filter elements.

     The nominal flow capacity of each of the three elements purchased
from the Ducon Company was 8.5 Sm^/min (300 SCFM).  Each element was 20 cm
(8 in) in diameter by 1.8 m (6 ft) long and contained twelve beds having a
total filtration area of 0.24 m2 (2.6 ft^).  A photograph of one of the
filter elements and the shroud in which it is contained when placed in the
pressure vessel is shown in Figure V-5.  One of the Ducon elements was
designed to be blown back by short sonic pulses of high pressure air as
shown in Figure V-6.  The pulse duration was approximately 0.5 sec.  This
blow back method was not tested.  The other two Ducon elements were blown

                                      85

-------
    FIGURE V-l




SINGLE BED TEST RIG
          86

-------
                                                 FIGURE V-2

                                      SINGLE BED TEST RIG  ASSEMBLY
             Off-Gas from
            Secondary Cycle
00
                            Cooled Ball
                              Valve
Filter
                                                                 Vent
                                                                                Pressure
                                                                                Gauge
                                                                         (/\Pressure Gauge

                                                                          HT      NTyiFlow Control Valve
                                                                                         v Rotameter
                                         •low
                                       Control
                                         Valve
    Balston
    Filter
                                                                                                 Pressure
                                                                                                 Gauge

                                                                                              "| Pressure
                                                                                              J Regulator
                                                                                      Air From
                                                                                      Auxiliary
                                                                                     Compressor

-------
INLET
RETAINING
SCREEN
CO
CO
 BOTTOM
 RETAINING
 SCREEN
                                          FIGURE V-3

                            SCHEMATIC OF A SINGLE DUCON FILTER BED
20.3 cm
2.7 cm -
H-*	 3.8 cm
                                 FILTER
                                 MEDIUM
                                    CLEAN GAS
                                        o

                                        CNJ

                                        LO
                                                                              4.4 cm

-------
                        FIGURE V-4

                DUCON  FILTER SCHEMATIC
       Filtration Cycle
                        Blowback Cycle
Filter
Media"
          ZZ2
          Z22
          22Z
           (Lu.

           222
\
               LLLL
                     Clean
                     Gas  Exit
               LLLL
                    Dirty Gas
      -Retaining Screen
                                           7
                                          Fly ash
1
Lock
Hopper

i
Lock
Hopper
                                      Dirty Gas
                                        Fluidized
                                        Filter Media
                             89

-------
          FIGURE V-5




DUCON FILTER ELEMENT AND SHROUD
                90

-------
                       FIGURE  V-6

               EXXON  FILTER  SCHEMATIC
       Filtration Cycle
Dirty Gas
Filter
Media
                     Retaining
                     Screens
Blowback Cycle
    te
                                        /t>
                                               Clean  Air
                             91

-------
back with a larger volume  of  sub-sonic air for longer durations.   The  intent
was to fluidized  the  filter beds rather than shock them with a short pulse
of high pressure  air.   This method is described in Figure V-4.

Original Exxon Filter—
     A fourth filter  element  designed by Exxon was also fabricated and
tested.  The Exxon filter  element consisted of ten beds (instead  of twelve)
and was designed  to permit easy disassembly and removal of the 50 X 50 mesh
retaining screens.  A schematic of the filter element and the arrangement
within its shroud is  shown in Figure V-7.   Figure V-8 is a sketch of an
individual bed.   The  blow  back method was also different.  The Exxon filter
element used a "positive blow back" technique.  It was equipped with shut
off valves to allow it to  be  completely isolated from the feed and product
streams during the blow back.  The element was then depressurized and blown
back with a larger volume  of  low pressure air.

Current System—
     Discussion with  the Ducon Company were held after operating  problems
developed with the original filter design.  The problems are discussed in
a  subsequent section.  These  discussions led to the design and fabrication
of a third filter system.   The current system consists of two filter ele-
ments, each having five filter beds as shown in Figure V-9.  The  filter
beds incorporate  the  modifications recommended by Ducon which included
removing the inlet retaining  screens to prevent plugging and increasing
the freeboard of  each bed  to  prevent entrainment of the filter media during
blow back.  A schematic of one of the modified filter beds is shown in
Figure V-10.  An  annular fluidizing grid containing 56, 0.5 cm diameter holes
was also installed at the  bottom of each bed beneath the bottom retaining
screen, as shown  in a photograph of the clean gas side of the grid (Appendix
J-2) .  The intent of  the grid was to assure good distribution of  the blow
back air.  Dirty  gas  enters the bed through the opening below the top
flange and passes downward through the filter bed and out into the clean
gas outlet tube in the center of the element.  During blow back,  the blow
back air passes up through the fluidizing grids supporting each bed, fluidizes
the beds and blows the fine particulates out through the inlet slot.  A 18 cm
 (7 in) freeboard  above the filter beds acts as a disengaging section for the
filter media and  prevents  its entrainment through the outlet slot.

Pressure Vessel,  Internal  Piping—
     The granular bed filter  (GBF) was installed on the miniplant between the
second stage cyclone  and pressure control nozzle.  The piping from the
existing miniplant off gas line to the GBF is 30.5 cm (12 in) carbon steel
pipe.  It is refractory lined with Grefco 75-28 to an internal diameter of
10.2 cm (4 in) and is lined inside with a 4 inch Schedule 5 stainless steel
pipe.  The return line from the GBF outlet is 20.3 cm (8 in) diameter carbon
steel pipe also refractory and steel lined to a 10.2 cm (4 inch)  diameter.
The pressure vessel housing the filter elements consists of a refractory
lined vessel approximately 2.4 m (8 ft) in diameter by 3.4 m  (11 ft) high.
Figure V-ll is a  photograph of the filter vessel and shows the size and
structural relationship of the GBF to the rest of the miniplant.   The pressure
vessel is supported on a structure which is 5.6 m (18.5 ft) by 4.6 m (15 ft)
and 5.9 m (19.25  ft)  high. Appendix J-3 and J-4 are drawings showing  the
vessel dimensions and the  locations of the various access ports.   Appendix
J-5 is a closeup  photograph of the pressure vessel.  The piping arrangement
is shown schematically in  Appendix J-6.
                                      92

-------
                       FIGURE V-7

             DUCOIM SONIC FILTER SCHEMATIC
      Filtration Cycle
Filter
Media
            X
                    Clean
A
(/
j(
•24.
\^M
• i _j
M
T2lt.
'/'}_

rWV


LLlj.
\Z22j.
I
\
/
\
k
OLJ.
LlM.
LLL^
LtM.
"ZLU
f>i~i
u&
///7

7/77
\77j A
bas txi

^ Dirty Gas
Retai
•"""" Scree
Blowback Cycle
                                                    Expanded
                                                    Filter Media
                             93

-------
                                         FIGURE V-8

                         SCHEMATIC OF A SINGLE EXXON FILTER  BED
                                  	  21.3cm
                                  •*	6. 6cm	H-*
Dirty

Gas
Top Retaining
Screen
                                    Filter
                                    Medium
                                     Clean Gas
             Bottom Retaining
             Screen

-------
         FIGURE V-9




MODIFIED EXXON FILTER ELEMENT
              95

-------
Dirty
Gas
Inlet
        I   I

                             FIGURE V-10

                         MODIFIED FILTER  BED
                                  Clean Gas
                                  Outlet
I   I   1,6cm
                                                                    14.3cm
   Fluidizing Grid
   (56-0.36cm dia.
I   |         holes)
                                  6.4cm
                                 21.3cm
                                      96

-------
                              FIGURE V-ll

                  GBF PRESSURE VESSEL  AND STRUCTURE
Blowback Plungers
     ntrol Valves
                                  97

-------
     Access to the  interior can be made through a 70 cm (27 in)  manhole   The
vessel can hold up  to  four filter elements installed through four flanges at
the top of the vessel  as  shown in Figure V-ll.  Each filter element  is con-
tained within a shroud in the inside of the pressure vessel.  Inlet  gas  is
piped to each shroud,  passing through a measuring orifice to determine the
flow rate to each filter  element.  This is shown in Figure V-12.  Clean
gas exits from each shroud through openings at the top (Figure V-12)  and
fills the interior  of  the pressure shell.  Blow back air enters  each  filter
element through the top flanges of the pressure vessel (Figure V-ll)  and
flows in reverse direction through each filter element.  Particulates removed
from the filter element during blow back impinge on the inside surface of the
shroud, fall to the bottom and are collected in lock hoppers. The blow  back
gas leaving a filter element flows in reverse direction through  the  inlet gas
system into the other  filter elements which are in the filtration mode.  Each
element is blown back separately.

Startup Burner, Natural Gas Injection —
     A natural gas  burner was installed to preheat the interior  of the pres-
sure vessel to a  temperature above the dew point of the combustor flue gas
before starting a filtration test.  The burner is a self-contained,  direct-
fired air heater  capable of burning natural gas to give heat input rates of
up  to 146,500 W  (500,000 BTU/hr) .  A schematic of the burner flow systems is
shown in Appendix J-7.  Burner operations are automatically controlled as
are all the required startup sequencing and emergency shutdown procedures.
The burner fires  into the vessel through a side port for an 8 to 12  hour
period prior  to  the start of a run.  The filter vessel is at atmospheric
pressure during  this period.

     Natural  gas  injection into  the flue gas ducting between the miniplant
second stage  cyclone and the granular bed filter was found necessary  to
maintain  the  flue  gas temperature above 843°C (1550°F) at the filter  inlet.
Gas was injected  at four points  through injection probes located 2.3  m
 (7.5 ft), 6.7 m (22 ft), 9.8 m (32 ft) and 14.6 m (48 ft) downstream of  the
second stage  cyclone.   Gas flows of approximately 0.03 s m3/min  (1 SCFM)
 through each  probe  was found effective in maintaining the gas temperature
at  the desired  level.

Operating Procedures

     To prevent  condensation during startup, the pressure vessel housing the
 filter elements  vessel was preheated for a period of 8 to 12 hours.   After
 the preheat period, startup activities included calibration of all AP trans-
mitters,  turning on purge air flows for the pressure taps, and turning on
 the blow back air compressor.  The proper blow back air pressure and flow
were  then  set.   A schematic of the GBF blow back air supply is shown in
 Figure J-9.   At this point, the  miniplant combustor was started  up and flue
 gas was sent  to  the granular bed filter.  During the filtration  cycle, the
 pressure  drop across the filter  vessel and the flow to each element  were
 continuously  monitored.  When the pressure drop increased to an unacceptable
 level  (generally,  14 kPa above the baseline pressure drop), the  blow back
 cycle was  initiated.  Blow back  was accomplished by stopping the flow to
one filter element by engaging a blow back nozzle and seal P^te and blowing
back with  air at a pressure slightly above the ^ration P'^re'    e
other elements  would then pick up the additional fxltration load.
                                      98

-------
         FIGURE V-12




GBF PRESSURE VESSEL INTERIOR
             99

-------
filter elements were blown back  consecutively and  all  operations were con-
trolled from the miniplant panel board.   After a number  of filtration/blow
back cycles, the particulate  collection  lock hoppers were emptied.

EXPERIMENTAL RESULTS AND DISCUSSIONS

Test Rig Preliminary Results

     The granular bed  filter  test rig was successfully operated during mini-
plant runs 46.1 and 46.2.  The primary intent of these tests were to estab-
lish operating procedures and to determine,  at least qualitatively, how readily
the fly ash could be cleaned  from the filter bed.   A 6.4 m (2.5 in) deep
bed of granular quartz was used  during the test with a particle size ranging
between 300 and 600 microns with 50 weight percent being finer than 400
microns.  This particle size  distribution was selected as a compromise
between higher collection efficiency at  a finer particle size and support
screen plugging problems associated with too fine  a particle size.  The
filtration velocity during these tests was approximately 6.1 m/min (20
f t/min) .  The blow back duration ranged  between 3  and  10 min in these pre-
liminary tests and the blow back air superficial velocity ranged between
0.18 and 0.34 m/sec  (0.6 and  1.1 ft/sec).  Some of the more important con-
clusions of the test are as  follows:

     1.  The granular  bed was capable of filtering fly ash.

     2.  As shown in Figure V-13, the bed exhibited a  linear increase
         in AP with time.  An analysis of the data indicates that the
         AP across the bed increased at  a rate of  0.4  kPa/min
          (1.6 in H20/min).

     3.  Although blow back  conditions were more favorable than can
         be afforded in a practical cycle, the AP  across the bed was
         always able to be brought back  close to its baseline value
         of about 1 kPa.  The differences in the filter  pressure
         drop after each of  the  blow backs was caused  by differences
         in blow back  conditions.

     4.  Extrapolation of the test results indicated that to maintain
         the AP across the bed below the design value  of 14 kPa (2 psi),
         a cycle time  of approximately 20-35 minutes was required
         between blow  backs.

     5.  Although the  filter  collection  efficiency was not measured,
         an inspection of the backup filter indicated  that most of
         the fly ash was being trapped by the granular filter.

     6.  Inspection of the filter bed showed no signs  of interraction,
         sticking or agglomeration of the fly ash  and  the filter media.
         The fly ash was easily  removed  from the quartz  particles.
     Operation of  the filter was also attempted during ^J1^ ™ *6'3
 and 46?4 but was not  successful due to condensation of moisture in the
 filter vessel.  The initial pressure drop across the bed was well above


                                      100

-------
                                   FIGURE V-13
   7.5
   5.0
ro
Q-
   2.5
     0
                         TEST RIG  RESULTS - AP VS. TIME
                                             SAND - GRANULAR QUARTZ
                                             PARTICLE SIZE RANGE - -30+50 MESH
                                             BED DEPTH  -6.4 cm
           i      i     i   »   i      i
                                                         '      '

     J	L
                                                     '

                                            /
                                                                ^
0    10    20
                                                              i
                       30    40    50    60    70   80    90    100   110   120
                                    TIME (MIN)

-------
the design value of 14 kPa  (2  psi)  and the increase with  time was much
greater than during the  initial  tests.  Inspection showed plugging of screens
with caked carryover material, possibly precipitated by presence of Listure
Based on these observations, every  precaution must be taken  to assure that
the filter is not operated  below the dew point of  the flue gas.

Testing of Original Ducon and  Exxon Filters

     After installation  of  the filters, shakedown  began with ambient tem-
perature testing of the  system.

     The objectives of these preliminary tests were to (1) check combustor
pressure control with the filter on line, (2) pressure test  the system, (3)
check the alignment of the  blow  back nozzles, (4)  check out  the operation
of the blow back flow system,  and (5) measure the  distribution of flow to
each one of the filter elements. A number of mechanical  problems were dis-
covered (leaks, misalignments, etc.) and had to be corrected before further
testing could be resumed.

     Two high temperature runs were then attempted but the pressure drops
across the filter became extremely  high and all attempts  at  blow back were
unsuccessful.  Inspection of the filter elements after each  of these runs
showed that a hard filter cake had  formed on the inlet retaining screens.
This is shown in Figure  V-14.  Note in Figure V-14 that the  particles are
adhering to all surfaces, not  only  to the inlet screens,  although the cake
is thicker over the screens.   The filter medium was usually  particulate
free indicating very little penetration through the screens.  Since these
initial runs were made before  the preheat burner was installed it was orig-
inally thought that the  plugging occurred during startup  when moisture could
condense before the filter  vessel had come up to temperature.  The preheat
burner was later installed  and a run (58) was made to re-evaluate the Ducon
filter.  The same screen plugging problems occurred before even one blow
back could be successfully  completed, and the original Ducon filter was
deemed to be unacceptable for  our application.

     Runs using the Exxon designed  filter also proved unsuccessful.  Some
screen plugging was in evidence  but it was the inability  to  seal the blow
back nozzles which caused the  most  problems.   Proper engagement and sealing
of the blow back nozzles were  necessary in order to isolate  the filter from
the system so that depressurization and blow back  could be initiated.
Evaluation of the Exxon  filter was  discontinued at this point.

Modifications

     At this point, a meeting  was held with personnel from Ducon to discuss
the problem of screen plugging.   The discussions led to the  design and
fabrication of a third filter  system.  Ducon indicated that  they had
encountered the same screen plugging problem with  fly ash and prevented it
by removing the screens  and designing the individual beds with more free-
board to prevent entrapment of  the filter media during blow back.  It was
also recommended that a  fluidizing  grid be used at the bottom of the beds
to assure good distribution of the  blow back air.   The use of an ejector to
replace the sometimes troublesome plunger type blow back  nozzles was also
suggested.  A detailed description  of the modified filter was given in an
                                      102

-------
           FIGURE V-14




DUCON FILTER WITH PLUGGED SCREENS
               103

-------
earlier section   Filter  elements incorporating these suggestions were fab-
ricated and shakedown  continued using two of these elements  each of which
contained five filter  beds.   An ejector system for blow back was also designed,
but was not tested during the period covered by this  report.

Testing of Modified  System

     Operability of  the modified filter system (not including  the ejector)
was demonstrated.  The end of the initial shakedown phase was  signified by
the completion of a  24 hour run (Run 59) .  The ability to filter and blow
back, the ability to maintain low pressure drops,  and the ability to collect
particulates after blow back were demonstrated.  Collection  efficiencies of
over 90% were calculated  during the initial portion of the run based on
measured outlet particulate concentrations of about 70 rag/m3 (0.03 gr/SCF).
The particulate loading in the outlet from the secondary cyclone had been
measured a number of times prior to the granular bed  filter  runs and had
averaged about 2300  mg/m3 (1 gr/SCF).  It was assumed that this was the inlet
loading to the filter  during these tests.  In the future, provisions will be
added to sample particulates at the inlet and outlet  of the  filter simulta-
neously.  The particulate escaping the granular bed filter had a weight
median particle size of 4 microns.  A complete size distribution is shown
in Table III-8.   Stable operation for up to 24 hours  was also  demonstrated
with no significant  increase in baseline pressure drop across  the filter.
Blow back was usually  required every 10-20 minutes during which time the
filter pressure drop had  increased by 14 kPa (2 psi)  above its baseline
value.  A range of blow back conditions were used to  restore the baseline
pressure drop.  Blow back durations ranged between 2  and 30  seconds and
superficial velocity between 0.15 and 0.75 m/s (0.5 and 2.5  f t/s) .  Filtra-
tion velocities generally ranged between 18.3 and 24.4 m/min (50 and 80
f t/min) .  Filter  media consisting of 300 to 600 micron quartz  particles were
tested.  The quantity of  blow back air used ranged from 1 to 5% of the
filtered gas rate.

     A number of  problem  areas were defined during the shakedown portion of
the program.  Demonstrated particulate outlet concentrations were still
higher  than the tentative turbine inlet requirements, although the -lowest
levels measured to date are only slightly above the upper limit of the ten-
tative  target range.  Firm turbine requirements have not been set and it is
too early to judge  the suitability of the filter to protect  gas turbines.
However, at times, the filtration efficiency was very poor and the outlet
particulate concentrations were as high as 700 to 1200 mg/m3 (0.3 to 0.5 gr/
SCF) .   It was also observed that the efficiency decreased with time in some
of the  longer runs,  including the 24-hour run, dropping from 90% initially
to about 50% later in the run.  Loss of filter media during  blow back was
another reoccurring  problem during shakedown.  Since inlet retaining screens
were found  to be  susceptible to plugging, screens probably cannot be used
and better  control of the blow back air supply must be established to minimize
these losses.  A  significant buildup of particulates in the  filter beds was
also observed amounting to about 30% of the weight of the filter media.  A
 noted  during any of the shakedown runs.
                                      104

-------
     It was also observed that the particulates were not only building up
in the beds, but were uniformly mixed with the filter media.  It is possible
that the buildup and mixing of particulates in the bed could be responsible
for the increase in the particulate concentration in the outlet gas with
time.  The cleaning of the filter media by blow back must be improved.

     Another potential problem with the current design is its vulnerability
to upsets.  If upsets occur, such as bed plugging or loss of filter media,
the operating problems caused by such upsets usually require shutdown of the
system.  It is usually not possible to take corrective action which restores
good operation.  Another problem which may be unique to the Miniplant was
the interaction of the granular bed filter with the rest of the FBC system
during the blow back cycle.  An increase in system pressure was noted during
blow back resulting in problems with the coal feed system which is controlled
by the differential pressure between the coal feed vessel and combustor.
This required modifications to the coal feed control system to minimize the
effects.

Test Details—
     A detailed description of the runs made to date follows.  Table V-l
summarizes the filter operating conditions for the five high temperature runs
made to evaluate the performance of the modified filter system.  Prior to
these runs, a series of ambient temperature runs was made using fly ash and
talc particulates injected into a compressed air stream.  Some indication
of the effect of blow back conditions on restoring the bed pressure drop was
obtained.  An 8 s blow back at a velocity of 0.5 m/s seemed adequate.
However, loss of filter media during blow back and outlet screen plugging
occurred during the runs and the significance of the results was uncertain.

     Miniplant run 54 was the first high temperature run using the modified
filter elements.  The combustor was operated at a temperature of 910°C
(1670°F) and at a pressure of 595 kPa (86 psia).  The combustor air flow
rate was set to give a filter face velocity of 25 m/min (83 ft/min).  Coal
was burned for a total of five hours and the pressure drops across the two
filter elements were approximately 14 kPa (2 psi) just prior to the start
of coal combustion and were allowed to build up to as high as 41 kPa (6 psi)
during the first hour of coal combustion.  The blow back cycle was relatively
successful and was able to reduce the pressure drops to approximately 21 kPa
(3 psi).  The filter elements were blown back at superficial velocities of
0.34-0.49 m/sec (1.1-1.6 ft/sec) for durations ranging between 8 sec and
30 sec.  However, in the latter part of the run, the pressure drops across
the filter elements just after blow back began to increase indicating a
decline in blow back efficiency.  When the run was ended, the pressure
drops across the filter elements were approximately 69 and 110 kPa (10
and 16 psi).

     During the latter part of the run a flue gas particulate sample was
taken downstream of the GBF.  A grain loading of 1570 mg/m3 (0.69 gr/scf) was
measured, giving an overall filtration efficiency of approximately 20-30%.
Moisture condensed in the filter during heat up, since the preheat burner
had not been installed before this test was run, causing some of the par-
ticulates to agglomerate in the filter.  This probably influenced the
effectiveness of the blow back and the filtration efficiency.


                                    105

-------
                               TABLE V-l.   GRANULAR BED FILTER RUN SUMMARY
                                       FOR  MODIFIED FILTER ELEMENTS
Filter Description
 Run 54 (6/29/77)
Run 55  (7/14/77)
Run 56  (7/26/77)
Number of Elements
Number of Beds/Elements
Inlet Retaining  Screens
Filter Medium
Bed Depth  (cm)
Filter Medium Part.  Size  (ym)

Operating  Conditions
Filter Vessel Preheat  Temperature (°C)
Filter Vessel Inlet  Temperature (°C)
Filter Vessel Outlet Temperature (°C)
Pressure (kPa)
Filtration Velocity  (m/min)
Baseline AP (kPa)
AP Before Blow  Back  (kPa)
Run Length (hrs)

 Blow Back Conditions
 Superficial Velocity  (m/s)
 Duration  (s)
 Interval Between Blow Backs (min)

 Particulate Emissions
 Measured Part.  Concentration (mg/nH)

 Estimated Removal Efficiency (%)
         2
         5
       None
      Quartz
        3.8
      250-600
    No Preheat
        815
        540
        590
         25
14 (70-100 at end)
         40
        5.5
     0.34-0.49
       8-30
       5-10
       1570
         30
        2
        5
      None
     Quartz
       3.8
     250-600
       450
       740
       510
       570
        21
20 (35-40 at end)
        50
       1.5
    0.55-0.73
      15-30
        5
  Not Measured
        2
        5
      None
     Quartz
       3.8
     250-600
       450
       760
       510
       800
        18
        20
        50
       2.5
     0.76
       8
     5-10
 Not Measured

-------
                          TABLE V-l  (Continued).  GRANULAR BED FILTER RUN SUMMARY
                                       FOR MODIFIED  FILTER ELEMENTS
Filter Description
Run 57 (8/2/77)
 Run 58 (8/5/77)
Run 59 (8/11/77)
Number of Elements
Number of Beds/Element
Inlet Retaining  Screens
Filter Medium
Bed Depth (cm)
Filter Media Part.  Size  (ym)

Operating Conditions
Filter Vessel Preheat Temperature  (°C)
Filter Vessel Inlet Temperature  (°C)
Filter Vessel Outlet Temperature (°C)
Pressure  (kPa)
Filtration Velocity  (m/min)
Baseline AP (kPa)
AP Before Blow Back  (kPa)
Run Length (hrs)

Blow Back Conditions

Superficial Velocity (m/s)
Duration  (s)
Interval Between Blow Backs  (min)

Particulate Emissions

Measured Part. Concentration (mg/m3)

Estimated Removal Efficiency (%)
       2
       5
      None
     Quartz
      6.4
    250-600
      650
      860
      700
      815
       18
       28
       55
       6
      0.46
       8
       10
     40-180
     98-92
        2
       12
   50 X 50 mesh
      Quartz
       3.2
     250-600
Original Ducon
system used.  Run
unsuccessful because
of screen plugging.
        2
        5
  50 X 50 mesh
     Quartz
       6.4
     250-600
       650
       860
       740
       790
        26
        28
        55
        23
                             0.46
                               8
                           10, 60-90
                         180,640,1230
                          92, 72, 47

-------
     Miniplant run  55 was  also made at a low pressure because of blow
back air supply limitations.   The natural gas preheat burner was used for
the first time to preheat  the granular bed filter  pressure vessel above the

§aS h^H n^r'rftSO^rrr W°rked WeU ^ the ^"t-e within the vessel
reached 454 C  (850  F) before  particulate laden gas was  passed through the
filter elements.  Baseline pressure drops of 21 kPa (3  psi) were measured
across the filter elements during the combustor preheat.  During the combustor
preheat period, in  which kerosene was burned in the combustor, the pressure
drop across the filters increased 24-28 kPa (3.5 - 4 psi) every 30 minutes.
Blow backs at  0.73  m/sec (2.4 ft/sec) superficial  velocity were successful in
restoring the  filter AP to the original 21 kPa (3  psi).  The collected par-
ticulate had a blackish coloration indicative of collection of oil soot and
attrited sorbent.

     During a  1-1/2 hour period burning coal, pressure  build up was more
rapid necessitating blow back approximately every  5 minutes  (Figure V-15).
The filtration velocity during this period was 27.6 m/min  (90.6 ft/min).
Equal flow distribution was not achieved during the coal burning phase, one
of the elements constantly passing more gas than the other.  The collected
particulates had  the characteristic red-brown coloration of a mixture of
coal, fly ash  and attrited sorbent.  The quantity  of particulates collected
in the lock hoppers amounted to about 50% of the anticipated particulate
loading to the GBF.

     The filter elements were inspected after the  run and considerable
build-up of particulates in the individual beds was evident.  The particulate
sampling system was not in operation during the run and the filtration
efficiency was not  measured.   However, the clean gas outlet tubes and the
interior of the GBF pressure vessel showed little  evidence of particulate
pass through.  The  outlet  retaining screens were quite  clean and showed no
evidence of plugging.

     Miniplant run  number  56 was an attempt to operate  the granular bed
filter system  at  the full  operating pressure of 928 kPa (135 psi).  This  run
was attempted  after modifications were made on the blow back system to
increase the capacity of the system in terms of pressure and flow.  Coal was
burned for almost 2-1/2 hours and the pressure drop across the filter was
able to be controlled between 21 and 48 kPa (3 and 7 psi) with 8 second
duration blow  backs at  a superficial velocity of 0.76 m/s  (2.5 ft/s) every
5-10 minutes.  The  filter  operations appeared to be successful and the run
was only terminated when a fire developed in the coal  injection vessel.
However, upon  inspection after the run, it was found that most of the filter-
ing media had  been  lost from the filter beds during the blow back cycle
probably because of a excessively high blow back air flow.  Again, no parti-
culate measurements were made and the filtration efficiency was not deter-
mined .

     Miniplant run  number  57 was a repeat of the previous  run with emphasis
on maintaing control of the blow back velocity since bed  losses during blow
back were appreciable  during run 56.  Each filter bed was  6.4 cm  (2-1/2 in)
deep and consisted  of  250-600 micron diameter quartz particles.  An approach
velocity of the gas to  the filter elements of 18.3 m/min  (60 ft/mm) was
measured.


                                      108

-------
                                              FIGURE V-15
                              EFFECT OF BLOWBACK ON FILTER PRESSURE DROP

                                                RUN  55
g
    co
    0.
                                                                          Element 1

                                                                          Element 2
                                               TIME (MIN.)

-------
     The blow backs were made  at  10 minute  intervals using a pulse duration
of 8 seconds at a superficial  velocity  of 0 46 m/Q  M  •? f?/«,  ?   n  •    !u
filtration period between blow backs/the p^ssu^  £  2 ^iheTiSr^
beds increased from the baseline  level  of 28 to 55  kPa  (4 to 8 pal).  The
profile of the bed pressure  drop  was  very reproducible.  During the 6 hours
of the run, collection of the  particulates  captured by  the filters was
achieved in the lock hoppers.   Although filter medium  loss did occur, it was
not as appreciable as in run 56.   Retention of a significant amount of
captured particulates in the filter beds was also apparent.  A particulate
sample of the off gas from the GBF was  obtained during  this run.  The parti-
culate loading was measured  to be 187 mg/m3 (0.08 gr/SCF), which represents
a collection efficiency between 90 and  95%.

     Miniplant run 59 was made after  50 X 50 mesh inlet retaining screens were
installed on the modified filter  elements to prevent loss of filter media.
The objectives of the run were to:  (1)  demonstrated GBF operations over a
24 hour period,  (2) observe  any changes in  baseline AP  over a number of blow
back cycles,  (3) determine again  if inlet retaining screens could be used
to prevent loss of the filter  medium  during blow back without plugging, and
(4) measure temperature profiles  from the combustor bed through the filter
vessel.

     The GBF was continuously  operated  for  a total  of  28-1/2 hours during
which coal was burned for 23 hours.   The baseline pressure drop across  the
filter elements was measured to be 28 kPa  (4 psi) and was allowed to build up
an additional 14 kPa  (2 psi) before blowing back.   Filtration was done  at
an approach velocity of 25.6 m/min (84  ft/min) .  Blow backs at a superficial
velocity of 0.46 m/sec  (1.5  ft/sec) for an  8 second duration were successful
in restoring the AP to the baseline value.

     The run appeared to have  two distinct  segments as  far as particulate
removal was concerned.  During the first 8  hours of operation on coal,  the
time between blow backs was  8-10  minutes.   During a one hour particulate
sampling period, a particulate concentration of the gas leaving the GBF of
180 mg/m3  (0.077 gr/SCF) was measured.   This loading was comparable with that
measured during run 57.  At  this  point  in the run,  the  coal feed was inter-
rupted for a 4 hour period during which kerosene was burned to maintain the
combustor bed temperature.   No blow backs were made during this period.
During the last 14 hours on  coal, the time  between  blow backs increased to
between 1 and 1-1/2 hours.   In an 9 hour particulate sampling period, a
loading of 653 mg/m3  (0.28 gr/SCF) was  measured.  During the last 2 hours of
the run, another particulate sample gave a  loading  of  1260 mg/mJ (0.54  gr/SCF),

     On inspection of the filter  elements after the run, it was seen that the
inlet retaining screens were partially  plugged.  Figure V-16 is a photo-
graph of one of the filter beds after the run which clearly illustrates the
nature of the problem.  It appeared that the gas was entering the bed through
a narrow slit between the screen  and  the flange of  the  next bed   A consider-
able amount of particulate was retained in  the filter beds in the form of a
hard filter cake.  The fact  that  rat  holing had occurred was apparent.   The
filter cake probably formed  during the  period when  kerosene was burned and
the subsequent low measured  collection  efficiencies resulted because of the
rat holing.


                                      110

-------
                        FIGURE V-16




MODIFIED EXXON FILTER WITH PLUGGED INLET RETAINING SCREENS
                             111

-------
     Pressure control  in  the  entire  FBC  system was  susceptible to upsets
during blow back pulses.   The differential pressure between the coal injec-
tion vessel and the  combustor was  easily upset.   Three  fires occurred in the
injection vessel as  a  result  of pressure upsets  which caused hot bed solids
to flow back into  the  injection vessel.   Alternate  blow back procedures
or modifications to  the coal  feed  system must be developed in order to
minimize these upsets.

Natural Gas Injection  Tests—
     Natural gas injection into the  flue gas line between the second stage
cyclone and the GBF  was investigated as  a means  of  maintaining the flue gas
temperature above  925°C  (1700°F) before  entering the GBF.  During an initial
test, natural gas  was  injected at  a  single point 2.3 m  (7.5 ft) downstream
of the second stage  cyclone and the  effect on the GBF inlet line temperature
profile was observed.  As seen in  Figure V-17, the  GBF  inlet was able to
be increased from  749°C  (1380°F) to  829°C (1525°F)  but  a  sharp temperature
rise near  the injection nozzle caused by the instantaneous combustion of
natural gas was also observed.  To minimize this temperature rise, natural
gas was later injected at three additional points but keeping the total
natural gas flow  the same as in the  single injection point test.  This
technique  was successful  in maintaining  the GBF inlet at  925°C  (1700°F)
without any sharp  temperature rises.  A  temperature profile is shown in
Figure V-17.

Blow  Back  Gas Ejector Design—
      A blow back  gas ejector was designed to replace the  blow back gas  inlet
plunger and  seal  plate assembly.  Figure V-18 is a  sketch of  the ejector
 system.  The motive gas is air compressed to 2550 kPa.   The ejector system
was not  tested  during this reporting period but is  scheduled  to be tested in
 the  future.
                                       112

-------
o
o

L±J
Cd
LJ
O.
   910 -
   880 -
   850 r
   820 -
                                 FIGURE V-17

                        FLUE GAS TEMPERATURE PROFILE
            INJECTION
              PT. 1
                                                                       700
                                                                    - 1650
                                                                    - 1600
                                         - 1550
                                                                             m
                                                 m
                                                                     -  1500  -j
                                                                     - 1450  -n
                                                                     - 1400
                                                                     - 1350
                                                                       1300
INJECTION
  PT.  2
INJECTION
  PT.  3
                           DISTANCE DOWNSTREAM OF
                           SECOND STAGE CYCLONE (m)
INJECTION
  PT. 4 GBF
       INLET

-------
                       FIGURE V-18

     SCHEMATIC OF THE  EJECTOR  BLOW BACK SYSTEM
                               High  Pressure
                               Ejector Motive Gas
Clean Gas
  Outlet
                Ejector    I
            Secondary Gas ^*- ,
Ejector
                                                   Filter
                                                   Shroud
                               114

-------
                                  SECTION VI

                       MODIFICATION OF THE BATCH UNIT


     The combustor and regenerator sections of the batch unit were modified
to permit continuous operation.   Prior to these modifications they could only
be operated in a batch or  semi-batch manner.  Operation of the unit in the
continuous mode will increase its flexibility and enable the unit to provide
the type of data which can only be obtained from continuous units such as
the miniplant.

     The modifications will  permit the continuous feeding of both coal and sor-
bent, and the continuous removal  of solids from the combustor.  In the past the
facilities permitted only  the continuous charging of coal to the combustor.
Sorbent was changed at the beginning of the run, with no provision for with-
drawal or makeup.  The regenerator can now be operated continuously.  Pre-
viously, the  regenerator was operated only in a batchwise manner.  In addition
there were no provisions for recycling the primary cyclone's solids to the
combustor.  This has an adverse effect upon combustion efficiency, as the
first cyclone's solids contain an appreciable amount of unburned coal "fines."
The modified  system now has  provisions both to recycle or not and will permit
the measurement of combustion efficiency with either method of operation.

     In the discussion that  follows the modified facilities will be described.
Facilities retained from the batch unit were discussed in previous reports
(1,9) and will not be described in detail.

COMBUSTOR SECTION

Combustor Vessel

     A schematic flow diagram of  the modified combustor and the off gas and
solids flow are shown in Figure VI-1.

     The combustor was constructed from four sections of 25 cm  (10 inch) dia-
meter standard wall carbon steel  pipe, lined with Grefco 75-28 refractory  to
an inside diameter of 11.4 cm  (4.5 inches).  The height of the vessel above
the fluidizing grid is about 4.9  m  (16 ft).  Below the grid is a 61 cm  (24
inch) burner  section, lined  with  Grefco Bubblite Refractory.  A 4 inch  ID
hand hole is  provided directly above the grid to facilitate the complete
removal of solids after a  run.  The preheat burner was described previously
(1).

     A water  cooled fluidizing grid is inserted between the burner section
and the first combustor section.  The grid supports  the static bed and  pro-
vides a uniform flow of fluidizing air to the bed during operation.  The
fluidizing air enters through 80-0.16 cm holes  in the grid.
                                      115

-------
                                        FIGURE Vl-l

                            SCHEMATIC  OF MODIFIED BATCH UNIT
             Combustor
             Shell
Sorbent
Feed
Hopper
\y
              Refrac-
              tory  —
              Lining
         MOVs
                             1st
                            M
                      Solids
                      Overflows
                               .^ Grid
                                  Cooling
                                  Water


                               Burner
 ^, Off Gas  Precooler
*--
- Cooling Water  In
*•

^^  Off Gas Cooler

        Back Pressure
        Regulator

 off          X,
 Gas
 Filter
                                                   X
                                                                                    To Scrubber
                       To Analytical
                       Train
                                             Solids
                                             Overflow
                                             Hopper

-------
     Three sets of vertical  cooling coils are located within  the combustor
which control combustor  temperature.   The coils are made of 6.4 mm 316 SS
tubing with a surface area of  0.06 m2 per coil.  The cooling  coils are supplied
with dimineralized water, and  the flow rate is controlled to  produce a steam/
water outlet mixture.  High  flow rates can be used to prevent steam formation
when heat transfer coefficients are measured.

     Two flanged  inlets  are  welded into the side walls of the first section
above the grid at an angle of  60° to the horizontal (well above the angle
of repose for sorbents and fly ash).   These serve as a sorbent charging line
and as a return line for solids from the primary cyclone. The sorbent
charging line is  1-1/2 inch  IPS Sched. 40 and the primary cyclone return line
is 4 inch IPS, Sched. 40, CS refractory lined to a 5 cm (2 inch) ID.  Both
are projected through the refractory and terminate about 0.013 cm  (1/2 inch)
above the grid.   Introduction  of the solids immediately above the grid
improves mixing of the solids.

     The second and  third sections above the grid have outlets 0.46 cm  (18
inches) long and  inclined at 60° to the horizontal welded into three side
walls.  The outlets  are  4 inch IPS Sched. 40 CS, refractory lined to a 5 cm
 (2 inch) ID.  Both are projected through the refractory to the combustion
zone.  These serve as  solids overflow lines.  The projection  of the lower
outlet terminates 1.09 m (43 inches) above the grid, and provides a bed
volume of 0.0111  m3.  The projection of the upper outlet terminates 1.85 m
 (73  inches) above the  grid,  and provides a bed volume  of 0.0189 m3.  The
overflow lines discharge to  a solids overflow hopper.

Combustor Off Gas System

     The off gas  system may  be classified into three major components;  they
 are:

     •  Primary  and  secondary cyclones

     •  Off gas  coolers
     •  Off gas  filter  and  back pressure regulator

 Cyclones—
     Both cyclones have been rebuilt in order  to increase their  efficiency.
 They were cast with  Resco RS17-E refractory with the aid of wood  mandrels,
which were  "burned  out"  with a torch after the refractory had hardened.  The
 refractory  is highly abrasion resistant.  The  gas discharge lines,  at  the  top
 of  the cyclones,  are made of 316 SS in order to resist corrosion.   Details
may  be found  in  Figure VI-2.

      The  primary cyclone will recover unburned coal  and  some  fly ash and
sorbent  from the off gas and return it to  the  combustor  or to  a receiving
hopper.   The solids  return line  is  fabricated with  a  length of 2 inch flexible
hose  that  may  be  connected either  to  the  combustor  or to  a receiving hopper.
                                      117

-------
          FIGURE VI-2

BENCH UNIT CYCLONE  ASSEMBLY
                                                                                   UIQ1T
                                                                 EXXON RESEARCH AND ENGINEERING COMPANY
                                                                      MECHANICAL DIVISION
                                                                         LINDEN. N. J.
                                                                            1633-I&-C

-------
     The secondary cyclone will  remove  the remaining fly  ash  and sorbent from
the off gas in order to minimize the  load on the off gas  filter.  The small
volume of removed fly ash will be collected in a standpipe  and discarded
periodically.

Off Gas Coolers—
     The off gas coolers are  of  the single-pass, water cooled, counter-current
flow double pipe type and are mounted vertically.  The inner  tube  (gas flow)
is 1 inch IPS Sched. 80, 316  SS, the  outer shell (water flow) is 2  inches
IPS Sched. 40, CS.

     The off gas first  flows  through  a  precooler intended to  reduce the off
gas temperature from about  760°C (1400°F) to approximately  370°C (700°F).  A
second heat exchanger  (the  off  gas cooler) is used to reduce  further the off
gas temperature from 370°C  (700°F) to 150°C (300°F).  The precooler has a
fixed heat exchange area  (gas side) of  0.18 m2 (1.9 ft2).  The off  gas cooler
has provision for drawing off the cooling water at nine locations along the
shell.  This results in a variable area (gas side) ranging  from 0.023 m2
 (0.225 ft2) to 0.268 m2 (2.88 ft2).  The variable area cooler will  provide pre-
cise control of off gas temperature at  all combustor operating conditions.
This will insure that  the temperature of the off gas does not fall  below
its "dew point" regardless  of the system's operating conditions.

Coal Feeder

     The coal feeder vessel previously  used has been replaced with  one of
larger capacity.  The  new coal feeder vessel has a capacity of 0.184 m3  (6.5
ft3).  This will permit the storage of  109 kg (240 Ibs) of  coal and 52.7 kg
 (116 Ibs) of sorbent.   Solids feed rate will be about 13.6  kg/hr  (30 Ibs/hr)
and will permit eight  to  twelve hour runs without recharging the vessel.

     The control system previously employed will be retained for the new coal
 feeder.  As  in  the  past,  coal will be injected into the combustor  directly
 above  the grid.  In order  to insure adequate mixing of the coal with the
 contents of  the combustor bed,  the incoming coal-air stream will be surrounded
by a stream  of high pressure boost air  at sonic velocity as done in the past.
 Details  of the  coal injector may be found in Figure VI-3.

 Sorbent  Charging and Removal Systems

     The charging of a premixed coal and sorbent mixture to the combustor
would  be the optimum way  in which to charge sorbent to the combustor.
 However, this may not  be  feasible because the coal feeder orifice  may^be
 too  small to permit the passage of sorbent.  Any major increase  in orifice
 size would have an  adverse  effect upon the control of coal feed rates.  This
 possibility will be studied immediately after the "shakedown" of  the com-
bustor has been made.   If  simultaneous  coal-sorbent feeding is  not feasible,
 sorbent will be fed using a pressurized lock hopper in short and  intermit-
 tent intervals.  See Figure VI-1.
                                      119

-------
                                                                                                                                                                              S3
                                                                                                                                                                              O
                                                                                                                                                                              PC
                                                                                                                                                                               p
                                                                                                                                                                               a     <
                                                                                                                                                                               g
                                                                                                                                                                               I
1. l\\_\_ 1-AR.T
i. V
                              \mi,t>w.i
i	i
                                                                                                                                                     COAL  IMJE.CTOR  V£-V5>E-L-.
                                                                                                                                                   	BATCH   UKJlT
    EXXON RESEARCH AND ENGINEERING COUPANV

            MECHANICAL DIVISION

                LINDEN, tt.j

-------
     Intermittent charging  of  fresh  sorbent  to  the combustor bed can produce
substantial changes in  the  bed composition unless the charging is carefuMy
controlled   Frequent addition of  small  amounts of sorbent will limit the
changes in bed Composition  to  acceptable values.  An upper acceptable limit
of 5/. changes  in the bed  composition during  the addition of fresh sorbent
has been set.  Two motor  operated  valves (MOV)  spaced 30 cm apart on the
discharge line will meter the  incoming sorbent; the valves are operated
by a cycle timer.  For  a  bed height  of 109 cm (small bed) each sorbent charge
will introduce 5/, of new  material  to the bed;  for a 185 cm bed height (large
bed) each sorbent change  will  introduce  3% of new material to the bed.

     Valve cycle times  are  based upon the  maximum coal firing rate (14 kg/hr),
a 5% sulfur coal, and dolomite as  the sorbent with a Ca/S ratio of 3.0; these
are expected to be the  most severe operating conditions that will be encoun-
tered.  Under  these conditions 39  cycles per hour will be required.  Milder
operating conditions, such  as  a coal firing  rate of 9 kg/hr, 2% sulfur coal
and a Ca/S ratio of 1.0,  would require only  4 cycles per hour.  1-1/2 inch-
carbon steel-full bore  ball valves will  be used to meter solids flow.  Valve
ball and seats are stellite faced.

     Bed inventory  (or  volume) will  be controlled by overflow from the top
of the bed to  a receiving hopper of  the  same volume as the charging hopper.
Bed heights of 109 cm  (43 inches)  and 185  cm (73 inches) will be employed.
These heights  correspond  to volume of 0.11 m^ and 0.19 m3 respectively and,
in conjunction with variable gas flow rates, should provide adequate varia-
tion in residence  times.

     The capacities of  the  charging  and  overflow hoppers are 103 kg  (227 Ib)
or 0.078 m3  (2.75  ft^) .  These are more  than adequate  for 8 hour runs under the
most severe conditions.

REGENERATOR FACILITIES

     Many of  the  control  and supply  facilities for  the air, nitrogen and gas
systems are common to  the combustor  and  regenerator and were discussed pre-
viously.  Only those  portion of these systems that  are unique to the regen-
erator will be discussed  in this section.

Regenerator Vessel

     The regenerator  has  an effective height (grid  to  off  gas discharge) of
4.57 m  (15 ft) and is  constructed  of 12  inch IPS,  pipe,  and  lined with Grefco
 75-28 refractory  to an  inside diameter of 9.52 cm (3.75  inch).  The  plenum
chamber below the  grid  is 0.69 m  (2.25 ft) high and is lined with Grefco
Bubblelite refractory.

     A fluidizing  grid  is inserted between the plenum chamber and  the  first
section of the regenerator.  The grid serves to support  the  static bed, and
 to provide a  uniform  distribution of fluidizing air when the  regenerator  is
in operation.  The previously used grid will be replaced with_a new  grid
 (Figure VI-4). The new grid, having four cooling water  circuits, will
operate at much cooler  and  more uniform temperatures  than did  the  old  grid,
with only one  cooling water circuit.

                                       121

-------
N)

NJ

                                                                                                                                                             a
                                                                                                                                                             o
                                                                                                                                                                               .-L
                                                                                                                                                                                                      23

                                                                                                                                                                                                      O

                                                                                                                                                                                                      Pi
§





Hi     g
                                                                                                                                                                                                      PJ     Ml
                                                                                                                                                                                                      53
                                                                                                                                                                                                      w

                                                                                                                                                                                                      5     a
                                                                                                                                                                                                       g
                                                                                                                                                                                                       s
                                                                                                                                                                                                       s
                                                                                                                                                                                                       g
                                                                                                                                                                                                       o

                                                                                                                                                                                                       13

-------
•   h TP?  ?TraAnr " *qUlpped With  two bed overflow lines.  These are 4
xnch IPS  Sched  40, carbon  steel pipe section  that are refractory lined to
3 5 CmK I "? }   ; n ™,ioWe$ Plpe Pr°vides a  bed depth of 60 cm (24 inch)
and a bed volume of 0.00428  m3  (261 in3).  The  discharge line may be blanked
off and a refractory plug  inserted when it is desired to use a greater bed
height.  The high overflow line provides a bed  depth of 121 cm (48 inch)
and a bed volume of 0.00863  m3  (527 in3).

     The overflow lines are  connected  to an overflow lock hopper of 0.98 m3
(3.5 ftj) with the aid of  5.1 cm ID flexible stainless steel hose.  The use
of flexible hose eliminates  the rigid  pipe connections between the overflow
lines and the hopper.  This  eliminates a possible source of leakage and
alignment problems due to  thermal distortion.   The lock hopper volume is
more than adequate for eight hours of  operation under the highest anticipated
sorbent discharge rates.

     Sorbent is charged to the regenerator from a lock hopper (of 0.98 m3
capacity also) via rigid piping and a  1-1/2 inch ID flexible hose.  The sor-
bent is introduced at the  bottom of the bed to  insure complete mixing; the
charge point is 9.5 cm  (3.75 inches) above the  grid.  The sorbent charging
line is equipped with two  MOV's to meter in fresh sorbent in the same manner
as was discussed for the combustor.

     Supplementary air and methane, over and above that supplied to the
burner, are introduced at  about the middle of the bed and just above the
fluidizing grid respectively.   Supplementary air and fuel are used to produce
oxidizing and reducing zones which are needed to carry out the regeneration
reactions.

     The regeneration reactions are highly endothermic and unlike the com-
bustor, no bed cooling coils are used  in the regenerator.  Instead, a 7.5 kw
air preheater has been installed in the regenerator air supply line to provide
a  auxiliary heat source, if  needed.

     Provisions have been  made  for feeding coal to the regenerator, either as
a  primary or supplementary fuel.  These facilities have not been installed
as yet, but injector ports have been allocated  and capped off.

     Thermocouples have been installed to provide bed temperature readings
every  15 cm.  Above the bed, gas temperatures will be read at 30-38 cm
intervals.

     The plenum chamber is mounted directly below the grid and contains the
burner.  The functions and mode of operation of the burner are essentially
the same as for the combustor preheat  burner.   Safety and cooling system
functions in the same manner for both  units.

     Pressure differentials  are measured across the grid and, at partial and
total bed heights, in order  that bed heights, bed densities and pressure
drops across the grid may  be obtained.
                                      123

-------
Regenerator Off Gas System

     The regenerator off gas is passed to a single cyclone  (Figure VI-5)
whose solids are returned to a pressurized standpipe.  The clean off gas is
passed to an off gas precooler (0.194 m^ area) and an off gas cooler (0.197
m-3 area).  These are of the single-pass, water cooled counter-current flow,
tube and shell type, and are mounted vertically.  The precooler is intended
to reduce the off gas temperature from 982°C (1800°F) to 370°C (700°F).  The
off gas then passes to the cooler where the temperature is further reduced
to 150°C (300°F).

     The off gas facilities for the regenerator that are downstream of the
cooler are identical to those for the combustor.  These have been discussed
previously.
                                     124

-------
          FIGURE VI-5



BENCH UNIT REGENERATOR CYCLONE
! i!
i>fe
                                                          ,
                                                          ] 
                                                  I"
            125

-------
                                 SECTION VII

                     COMPREHENSIVE ANALYSIS OF EMISSIONS
     A program is now underway  sponsored by  the EPA to assess the environ-
mental effect of fluidized  bed  coal  combustion.  The program consists in
setting emission goals  for  all  potentially harmful emissions and measuring
the concentration of these  materials in all  the effluents from fluidized bed
combustion units.   The  contractor currently  coordinating the work for the
EPA is Battelle Columbus  Laboratories.  Exxon Research and Engineering
Company has been requested  by the EPA to participate in the program by
carrying out a series  of  comprehensive analysis tests in the miniplant.
In these tests, specified emissions  ranging  from  S02, NOX, etc which are
routinely measured  to  trace inorganic and  organic materials present in the
solid  and gaseous  effluents from the miniplant are  to be measured.  The
first  series of  these  tests (Run 50) was  completed  in April 1977 in coopera-
tion with Battelle.  Samples were obtained and analyzed by both laboratories.
The results will  be published in a summary report prepared jointly by
Battelle and Exxon early in 1978.
                                       126

-------
                                 SECTION VIII

                       ANALYSIS OF DESULFURIZATION DATA


     In this section an additional analysis of desulfurization data is
discussed.  Previous analysis of the data was based on the use of Ca/S ratio
calculated from the analysis of the spent sorbent and the flue gas.  The
desulfurization data were again analyzed using the Ca/S ratio set on the
solids feeding system but corrected for variations in combustor temperature
and gas phase residence time.  A comparison of S02 retention results obtained
in batch and continuous FBC units was also made.  In this comparison,
corrections were applied to batch data to account for the residence time
distribution effects occurring in continuous units.  An examination of the
approach to steady state in a combustor with continuous sorbent feed and
used sorbent withdrawal was also made.

ANALYSIS OF DESULFURIZATION RESULTS BASED
ON Ca/S RATIO SET ON SOLIDS FEEDER

     In the previous report  (1) and in an earlier section of this report,
desulfurization data using Pfizer dolomite sorbent were reported.  Correla-
tion of the desulfurization results was based on the use of a Ca/S molar
ratio calculated from a sulfur balance using the expression

                  % SO  Retention (SR)

           Ca/S
                  % Ca Utilization  (Xj
                                     U

This approach was used to minimize  the possible effects caused by an incom-
plete approach to steady state operation and mechanical problems with the
coal/sorbent feed blender.  As pointed out in the earlier discussions, this
approach correlated the data reasonably well with an acceptably small amount
of data scatter whereas the use of  the Ca/S ratio set on the solids feeder
resulted in more scatter.  However, the correlation of S02 retention data with
a Ca/S ratio calculated from the same S02 retention results would be expected
to smooth the data and could give a misleading correlation.  The desulfuriza-
tion data were again analyzed using the Ca/S ratios actually set on the feed
solids blender but corrected for residence time and temperature variations.
This was done to determine if an acceptable correlation could be developed
without referring to a calculated Ca/S ratio.

Test of the First-Order Reaction Rate

     Residence time corrections were to be made using the first order rate
expression discussed in the previous report (1).  In Run No. 51, operating
conditions were adjusted to permit a specific test of the validity of a
first order kinetic expression.  This established a firmer basis for the
subsequent residence time corrections.  This test was accomplished by
quickly reducing the bed height (by rejecting bed solids) while maintaining
the superficial velocity constant.  As the change in bed height occurred
over a very short interval of time, the average sorbent utilization in the
bed would remain nearly constant.  If the rate constant for desulfurization
has a first order dependence, a linear relationship between the In (1-SR) ,
where SR is the fraction S02 retention, and the residence time would be

                                      127

-------
expected, with the slope of the line equal to the negative of the rate
constant.  In Figure VIII-1, the values of In (1-SR) for twelve determina-
tions in Run No. 51 are shown plotted against the gas phase residence time.
The first seven determinations represent data taken as the bed height was
increasing during the course of the run and the last five points are data
taken during the rapid withdrawal of bed solids.  During this rapid with-
drawal of bed solids, the change in bed height caused the gas phase resi-
dence time to drop 25 percent.

     The data in Figure VIII-1 follow the expected relationship for a first
order rate constant.  The calculated rate constant is 1.28 s~l (based on
the volume of settled particulate bed).  The sorbent utilization in Run No.
51 measured at the time of the test was 47 percent.

Correction of S02 Retention
for Gas Phase Residence Time

     The measured SC>2 retention for the runs using Pfizer dolomite are shown
in Figure VIII-2 versus the Ca/S molar feed ratio determined from the coal
and sorbent feed rates set on the solids feeder.  These runs were made with
coals containing 2 and 4 percent sulfur.  While uhe SC-2 retention does not
appear to be dependent on the coal sulfur content, the scatter of the
experimental data, as measured, is considerable.  The data shown covers a
range of gas phase residence times between 1.5 and 3.8 s and average com-
bustor bed temperatures between 684 and 945°C.  These wide variations would
be expected to cause data scatter beyond that due to uncertainty in the Ca/S
ratio.

     The measured SC>2 retention (SR) values shown in Figure VIII-2, were
corrected to a constant value of 2 s for the gas phase residence time in the
expanded bed of use by the following first order rate expression.

                               2 In (1-SR    )
             on       -i        /          meas N
             SR2s  =  1 - exp  (	)
                                     meas


Where t     is the measured gas phase residence time in the expanded bed.
       meas                                                   r

     The SC>2 retention values for a 2 s gas phase residence time in the
expanded bed are shown in Figure VIII-3.  While the absolute values of the
SC>2 retention have been changed (best observed by looking at the change in
values of the open circles in Figures VIII-2 and VIII-3), the scatter in
the experimental values still exists.

Temperature Dependence of the Desulfurization Performance

     Four of the S02 retention values corrected to a 2 s gas phase residence
time shown in Figure VIII-3 which deviate markedly from the trend  (labelled
L.T.) correspond to data obtained at combustor bed temperatures of 690,
684, 762 and 829°C.  As pointed out in  the previous section, desulfurization
was found to decrease at temperatures under 800°C.
                                      128

-------
                                     FIGURE Vlll-l

                       Ln (I-SR) VS. GAS  PHASE RESIDENCE TIME
                               T
                        T
    -1.10-
2   -1.20
LU
o;

 CM
O
CO
    -1.40
                            12
    -1.50
              DATA: Run No.  51 VELOCITY =  1.52 m/S
              Points 1-7 Bed  Height  Increasing
              Points 7-12 Bed Height Decreasing
                   1.1
1.2        1.3        1.4

   RESIDENCE TIME (S)
1.5
1.6

-------
   100
    90
                           FIGURE  VIII-2


               SULFUR REMOVAL EFFICIENCY VS. Ca/S

                            o
Q
UJ
o:
Z3
CO
<
LU
^

oo
o
^
UJ
UJ
>
O
C£

C£
CO
     80
     70
<=>   60
    50
    40
    30
S

O
    r
    o
                                PF\ZER DOLOMITE

                                  •  2% S Coal

                                  O  4% S Coal
                        1
              0.5      1.0      1.5     2.0

                        Ca/S MOLAR PATIO
                                                2.5
                                         3.0
                                130

-------
                      FIGURE VIII-3
      SULFUR REMOVAL EFFICIENCY @ 2 SEC. VS. Ca/S
100
                               PFIZER DOLOMITE
                                    2% S Coal

                                    4% S Coal
                                 L.T. - LOW
                                       TEMPERATURE

                                       POINTS
                      Ca/S MOLAR RATIO
                             131

-------
     The temperature dependence of the S02 retention corrected to 2 s gas
phase residence time in the expanded bed is shown in Figure VIII-4 for  two
Ca/S feed ratios.  The use of a constant gas phase residence time for this
analysis removes much of the variation in the data which would normally
mask the effect of temperature.  Clearly a temperature dependence is evident.
The experimentally determined desulfurization rate constant for runs made
at a Ca/S feed ratio of 1.5 was then evaluated for the temperature depen-
dence using an Arrhenius plot, as shown in Figure VIII-5.  An activation
energy of 13.3 kcal/gm mole was determined which is in good agreement with
the value reported by Borgwart of 10-20 kcal/gm mole for various limestones
(14).

Correction of the SC>2
Retention for Temperature

     The measured S02 retention values were corrected to a constant 2 s gas
phase residence time in the expanded bed and to a constant reactor bed tem-
perature of 900°C using the activation energy of 13.3 kcal/g mole.  The S02
retention at a 2 s gas phase residence time and 900°C is shown in Figure
VIII-6.  It is evident that most of the experimental scatter has now been
removed.  The trend in SC>2 retention determined by this analysis is in good
agreement with that shown in Figure 111-21 although the Ca/S feed ratio used
in this analysis is based on calcium and sulfur feed rates while the data
from Figure 111-21 is based on calculated Ca/S feed ratios using the measured
values of SC>2 removal and sorbent utilization.

COMPARISON OF S02 RETENTION RESULTS
OBTAINED IN BATCH AND CONTINUOUS UNITS

     In the previous report (1) it was shown that the values of S02 reten-
tion at a given calcium to sulfur feed ratio obtained in the continuous
pressurized fluidized bed combustors from different laboratories were in
agreement when appropriate corrections were made to reduce the data to a com-
mon gas phase residence time.  Agreement of the results obtained in batch
fluidized bed combustors with that observed in continuous units, i.e., the
miniplant has been less satisfactory (1).  This lack of agreement between
the results obtained in the batch units and in the continuous units is not
surprising when one considers the difference in the operating environment
between these units.  In the batch units, a charge of fresh sorbent particles
initially removes essentially all the S02 released during combustion and in
the process becomes partially sulfated, with presumably each particle reach-
ing the same level of utilization at the same time.  As the particle utiliza-
tion increases, the ability to remove S02 decreases and the S02 retention
drops.  In the continuous units, fresh sorbent is continuously being added
to the reactor and an average mixture of sorbent particles is being removed
from the bed, such that a steady state value of both sulfur removal
efficiency and sorbent utilization is established.

     With the increased availability of continuous reactors and batch reactors
for fluidized bed combustion studies and the increasing use of thermo-
gravimetric analyies (TGA) to study sorbent utilization, it would be highly
desirable to be able to relate the desulfurization performance obtained in
the various units.  If such a relationship can be obtained, it would enable


                                      132

-------
UJ
co

CNJ
o
o
§
o
cc
ZD
co
   100
    90
                             FIGURE VIII-4

         SULFUR REMOVAL EFFICIENCY @ 2 SEC. VS. TEMPERATURE


                  I	1	


    •  Ca/S =1.5

    A  Ca/S = 0.75, 2% S Coal

    A  Ca/S = 0.75, 4% S Coal             *
     80
     70
     60
50
                                                       A
     401—
      600
700             800
                                                 900
1000
                          TEMPERATURE - DEC. CELSIUS.

-------
                        FIGURE VIII-5

 EFFECT OF TEMPERATURE ON DESULFURIZATION RATE CONSTANT
+0.5
-0.5
-1.0
              O
         8.5
                     O
9.0
                               T
                      T
                      T
         PFIZER DOLOMITE

          O  Ca/S =1.5
                                  Calculated
                                  Activation Energy
                                  13,300  Cal/gm Mole
                                                  O
                               _L
                      _L
9.5
10.0
10.5
                   1/T x 10+4 (DEC. KELVIN)
                             134

-------
                          FIGURE VIII-6


         COMPARISON OF SET  VS. CALCULATED Ca/S  EFFECTS

   100,	,	,	
    90
o
o
O
O
o
o
LU
CM

®
OH

 CM
O
CO
    80
    70
    60
    50
    40
     30
0
           / O
           /
          /
               O/
                           8
0.5
                       1.0
                           PFIZER DOLOMITE

                             •   2%  S Coal

                             O   4%  S Coal
                           -.-.   Ca/S Calculated from
                                 Analyses (Fig.111-21)
                                         1
                                           1
                         1.5
                          2.0
                                                         3.0
                         Ca/S MOLAR RATIO
                                135

-------
an expanded use of the TGA and batch units to screen new sorbents which might
be considered for use in the larger continuous units.  Such a relationship
would also increase the confidence that the performance data obtained  in
process development scale continuous units, i.e., the miniplant, can be used
for predicting the performance of commercial size fluidized bed combustors.

     This subsection describes an attempt to relate the performance observed
in these various units.

Residence Time Averaging for the Effective
Rate Constant in the Miniplant Combustor

     The miniplant combustor, unlike the smaller batch combustor, will be
in a dynamic state as far as solids entering and leaving the reactor.
                    Fl
                                     W
Where F^  =  feed rate of uncalcined stone to the reactor -  moles/hr

      F?  =  removal rate of sulfated product -  moles/hr

      W  =  bed hold up - moles

     The feed and removal rates, F^ and F2, are adjusted to maintain a uniform
quantity of material, W, in the fluidized bed.  The backmixing of solids which
occurs in the fluid bed results in an age distribution for the solids, i.e.,
the  solid particles being removed from the bed in the overflow stream F2
have been in the bed for various lengths of time.  This variation in the
residence time of the solids is of course reflected in their degree of CaO
utilization.

     The kinetic data obtained from the batch combustor gives the reaction
rate constant as a function of particle utilization, Xg.  In order to use
this rate information it is necessary to know how the utilization of the
particle changes with the time the particles are in the bed, (the particle
age  or particle residence time 8).

     Assume an increment of fresh particles of mass, w, added to this
fluidized bed at time 6 = 0.  The rate of SC>2 uptake by these particles is:


                  d"°so2

                    de    ~  k(xB)'w'cso


where  N     =  mass SO^

          6  =  time particle is in bed (s)

       k(Xfi) =  rate constant (s~ )

                                      136

-------
        CSO   =  average concentration of  SO  seen by particles
           2     (mass S02/mass particles)


The rate constant is written k(XB)  to indicate  that it is a function of the
sorbent utilization  (or sulfation level),  X  .
                                         '  B

     At steady state, CS02  is  constant during the residence time of the
particles at some average value between  C*   in and C*   out.
                                           2           2
     Equation  (1) can be expressed  in terms of  a concentration Cg09 given in
units of mass  S02/volume of settled bed, by


                  dNso2     k(xB)

                   de    -  —  -W'cso2


     where  p   = bulk density of  the settled bed.
            S

     In this analysis,  all  bed volumes and gas  phase residence times are
based on  the settled rather than the expanded bed.

     The  sorbent utilization XB  is  the  S02 uptake compared to the maximum
 S02  capacity of  the  particles.

                            wt  S00 uptake      S00
                   x    _ 	£	 _  	£.                      (3)
                   B     capacity for SO    wot

 where a = mass capacity of  S02 per unit  mass of stone.

     Differentiating equation (3)

                        dNso
                dx    =   	2                                           (4)


      Substituting  equation (4) into equation (2) and separating  variables,
                      _
                  k(XB)
                                          de                            (5)
 A logarithmic mean concentration for Cgo  is the appropriate averaged
 concentration.

      On integration between XB = 0 and XB at time 6, we obtain after
 rearrangement :
                                       137

-------
X
          dX            SO
              R
              "                     C
             0

     The integral in equation (6)  is solved numerically to various values of
X  and the particle residence time to obtain that Xg given by:


                       8  -  ^                                    (7)
where A is the area under the X_ vs.  k(XD)  curve from X  = 0 to X .
                               D        D              O         a

     In order to predict the limiting performance of the continuous  unit
it is necessary to know the age distribution of the reacting solids.  It has
been found that the following particle age  distribution function, 1(6), is
quite accurate (13).

                       K0)  =  i   exp (-e/e)                            (8)
                                o
Where 6 is the average particle residence time  defined as

                            6   =  ~
                                   Fl

     The analysis of the reaction rate constant showed that the rate constant,
k(XB), depended on the degree of CaO utilization, XB.  The degree of CaO
utilization, Xg, for a particle injected into the fluid bed reactor  was shown
above to depend on the length of time the particle was in the reactor, 6.
These considerations suggest that the reaction  constant, k(Xg), for  a given
sorbent particle will depend on the time that this particle has been in the
reactor.  The expression for this dependence of the rate constant on the
particle residence time, k(6), can readily  be obtained from the expressions
previously determined for the dependence of the rate constant on the degree
of CaO utilization, k(XB), and the dependence of the degree of CaO utiliza-
tion on the particle residence time Xg(6).

     The desulfurization performance for the miniplant combustor will be made
up of the contribution of all the particles in  the fluid bed, each with its
own particular residence time.  The average rate constant for the miniplant
combustor, k", is obtained by integrating the product of the expressions for
the variation of rate constant with particle residence time, k(6), and the
age distribution for the particles, 1(0) over all values of particle
residence time, 6.
                                      138

-------
                     0

     This expression is solved graphically.
  _   The fractional S0? retention, SR, at  steady state can now be calculated
using the rate expression for the plug flow reactor model, by specifying
the operating parameters


                  SR  =  1 - exp  (-k-|5.)                                (11)

where Hs is the settled bed height, and U  is the superficial velocity in the
miniplant combustor.

     The overflow stream for the fluid bed, F2, contains, as mentioned
earlier, a wide distribution of particle ages, each with a degree of CaO
utilization dependent on its history in the bed.  An analysis similar to
that used above leads to an expression for the average CaO utilization, XB,
in the commercial reactor.
                           XB(6)-I(0) d6                               (12)
     The degree of CaO utilization can also be obtained at a given value of
the calcium to sulfur molar feed ratio, Ca/S, by using the value of the
S02 retention, SR, calculated from equation (11), and the following expres-
sion which reflects a steady state sulfur balance around the reactor

              x   -
              X   ~
               B      SR

     The expression for the degree of CaO utilization from equation (12)  is
designated Xg (Distribution Function) to signify its direct calculation from
the assumed particle distribution function.  The expression from equation (13)
is designated Xg (Performance) to signify its calculation from the calculated
sulfur removal efficiency at a given molar feed ratio of sorbent calcium to
coal sulfur.  The adequacy of the assumed particle distribution function can
be tested by comparing the values of Xg calculated in these two ways.

Calculation of the Miniplant Desulfurization Performance

     The calculation procedure described above has been used to calculate
the desulfurization performance expected in the miniplant combustor with
Pfizer dolomite as the sorbent based on the performance data measured in the
batch combustor.  The batch unit performance data for 1500 micron Tymochtee
dolomite (tests No. 71 to 82) were used to calculate the necessary values
of the reaction rate constant, k.  As the batch unit data was limited to
sorbent utilization levels between 0.20 and 0.46, while the calculation
                                      139

-------
procedures described above require averaging over all values of particle
residence time (equivalent to all values of XB between 0 and 1), the necessary
integration of 1/k in equation (6) was performed after first fitting a
linear equation to values of 1/k measured in the batch unit.  The values of
the reaction rate, k, measured in the batch unit and the curve fitted expres-
sion for k as a function of Xg which was used in these calculations is shown
in Figure VIII-7.

     The bed hold-up, W, required by the expression for the average particle
residence time in equation (8) is determined by specifying the settled bed
height for the case to be calculated and the area of the miniplant combustor.
The S02 inlet concentration required to calculate the average SC^ concentration
seen by the particles, equation (1) , is determined by specifying the coal
feed rate, the superficial velocity and the temperature for the case to be
calculated.  The sorbent feed rate, Flf is determined by specifying the Ca/S
ratio for the case to be calculated.

     The numerical procedure used to calculate the desulfurization performance
of the miniplant combustor first estimates a value of the desulfurization
performance to enable the calculation of the average SC^ concentration seen by
a particle in equation (1).  The entire calculation procedure is performed
and the desulfurization performance calculated by equation (11).  The cal-
culated value is compared with the estimated value, a new value for the
desulfurization performance estimated, and the calculation procedure repeated
until the two values agree.

     An example of the prediction of the miniplant reactor performance from
the batch combustor data is given in Figure VIII-8 for the operating condi-
tions of 900°C and a 2 second gas phase residence time (based on the time for
the gas to transit the expanded bed) and the predicted values are compared
with the experimental values measured in the miniplant.  While the predicted
performance does not exactly correspond to the experimental data measured in
the miniplant, the predicted values are non-the-less quite representative
of the desulfurization performance with dolomitic sorbents.  The model pre-
dicts the same desulfurization performance, irrespective of the sulfur
content of the coal feed, which agrees with the experimental determinations.

Comparison with the Performance
Measured in the TGA	

     The time that a given sorbent particle requires to achieve a given level
of utilization can be calculated from equation (7).  A comparison of the time
to a given utilization level predicted from the data obtained in the batch
combustor with the time experimentally measured in the thermogravimetric
analyzer (TGA) for Pfizer dolomite is shown in Figure VIII-9.  In these
TGA experiments the temperature was maintained at 900°C and the concentration
of SC>2 passing the sorbent particle was maintained constant at 2500 ppm.
The agreement is sufficient to distinguish clearly the predicted values as
that for a dolomite sorbent, as contrasted to the value expected for other
types of sorbents, i.e., a high calcium content limestone.
                                     140

-------
CO
    8
    7-
<   5

co
^
o
    0
                             FIGURE VIII-7


           SULFATION RATE CONSTANT VS. SORBENT UTILIZATION
1 1 1 1 1
! PFIZER
1 1 1 1
DOLOMITE 1
                                  Dp =  1500 microns
                w
     0    0.1   0.2   0.3   0.4   0.5   0.6   0.7   0.8   0.9   1.0
                     SORBENT UTILIZATION (FRACTION)

-------
                       FIGURE VIII-8
  100
   90
   80
o
o
O
O
o
 ,  70
o
LU
CO

C\J
©  60
UJ
h-
   50
 CM
o
CO
  40
  30
     0
            S02 RETENTION @ 2  SEC. VS. Ca/S SET


               EXXON MINIPLANT PERFORMANCE
                     PFIZER DOLOMITE
                         8
     O
                     _L
                   PREDICTED PERFORMANCE!
                   FROM BATCH UNIT
                   DATA & MODEL
                               EXPERIMENTAL DATA

                                  • 2% S COAL

                                  O 4% S COAL
                         I
0.5     1.0     1.5     2.0

      Ca/S MOLAR RATIO (SET)
2.5
3.0
                            142

-------
o
H
O
a:
u.
2
g
i-
<
O
O
                      FIGURE VIII-9

          TIME FOR A PARTICLE TO  REACH CALCIUM
           UTILIZATION LEVEL  IN ATMOSPHERE OF
                 2500 PPM  502 AT 900°C	
                     PFIZER DOLOMITE
                           CALCULATED FROM
                           BATCH UNIT DATA
              50     100     150    200     250
                 TIME IN REACTOR (MINUTES)
                                                 300
                            143

-------
Implications of the Model

     The model provides an insight into some of the factors that control the
desulfurization performance in a continuously fed combustor like the miniplant.
Perhaps the most striking of these factors is the surprising effectiveness
of fresh sorbent particles, e.g., those particles most recently fed to the
reactor, in determining the desulfurization performance.  In Figure VIII-10
are shown three histograms which represent the contribution of particles
which have been in residence in the reactor for different hourly time
increments to the total number of particles, to the sulfur removal
efficiency and to the sorbent utilization attained at steady state condi-
tions for a sorbent feed rate corresponding to a calcium to sulfur molar
feed ratio of 1.5.  The contributions shown are for Pfizer dolomite as the
sorbent at a gas phase residence time in the expanded bed of 2 seconds.

     It is seen from the histograms that the 15 percent of the total par-
ticles which have been in the reactor for a time interval between 0 and 1
hours, account for nearly 55% of the sulfur removal achieved.   The particles
in the bed for a time duration up to three hours, while representing only
31% of the particles in the reactor, account for 85% of the sulfur removal.
It is this marked effectiveness of fresh sorbent particles for sulfur removal
that results in a psuedo steady state SC>2 concentration being  obtained within
a few hours following a change in the Ca/S ratio.

     While the fresh particles are very effective in determining the degree
of sulfur removal achieved, Figure VIII-10 does indicate that  they contribute
little to the steady state value of sorbent utilization.  Those same 31 per-
cent of the particles that accounted for 85% of the sulfur removed, account
for only 15 percent of the sorbent utilization.  The reason for this is that
a small, but significant, fraction of the sorbent particles are in the reactor
for very long periods of time.  In the example shown in Figure VIII-10, the
average particle residence time is 6.8 hours, up to which time 63% of the
total number of particles are accounted for.  The remaining sorbent particles
have longer residence times and correspondingly higher utilization levels.

     This model provides an explanation for the experimental observation
that the degree of SC>2 removal appears to approach its steady  state value
within only a few hours following a change in the Ca/S feed ratio.  The
analysis also points out the care that must be exercised in calculating
the Ca/S ratio from the experimental measurements of sorbent utilization
and SCsj retention and the expression for the sulfur and calcium mass balance
around the combustor at steady state conditions.  Clearly, one must
allow sufficient time following any change in the feed ratios  to insure that
the sorbent utilization has indeed approached its steady state value for the
calculated Ca/S feed ratio to be valid.  The danger in using this method to
calculate the Ca/S ratio can better be appreciated using an example based on
the data shown in Table VIII-1.  In this table the steady state values of
S02 retention and sorbent utilization for various Ca/S feed ratios are given
for Pfizer dolomite sorbent at a gas phase residence time of 2 seconds in
the expanded bed.
                                      144

-------
                     FIGURE VIII-10

CONTRIBUTION OF PARTICLES IN BED FOR VARIOUS TIME PERIODS
       TO  STEADY STATE VALUE OF GIVEN PARAMETERS

           Ca/S = 1.5, DOLOMITE, 2 SEC GAS RESIDENCE TIME
60-
50-
LU
fj 40-
LJ
h-
i* 20-
co
> 10-

t— 0-
01
1°

No. OF PARTICLES



|p
:/x
X^Xx
i^
:^x
xxx^

XXXx
^^ ^ H^ ^^ ^^ kXx^ rx/^i
•xxx x/x^ KXxx X/x^ rxxxi l^xx'^'l rxxxl |xXxS1 fx'x'/l f"y ^ ^ t
<" 012 3456789 10 11 12
£ 60-
Q
o 50-
cc.
LU
Q. 40-
UJ
^-—
P 30-
20
CO
y 10-
_i j. \,/
o
£ oJ
% c
%
^
x^
XxX
^X
Xx
XX
///,
^
!xxx

S02 RETENTION



1
w
yyy

x^ ^xJ
X/XX rxXxi f/Xxxl I///XI I/xVxi
11234 5 6 789 10 11 12
Q.
Ll_
0 50-
o
j= 40-
—)
CD
EE 30-
i—
o 20-
o
10-
n.
%

SORBENT UTILIZATION





r77!
///
^J^J:^ ^^F^^^f^1,^^1, — r-
    'O   i  2   3  4   5   6   7  8   9  10  11  12

        TIME PARTICLE  HAS BEEN IN REACTOR (HOURS)


                           145

-------
                  Table VIII-1.   Comparison of  Rate Constants
          Ca/S     _SR_     XB ^rformance)     kave
                                                (s"1)     (s'1)

          1.0      0.67           0.67          1.09      0.82
          1.5      0.78           0.52          1.48      1.02
          2.0      0.84           0.42          1.79      1.20
          2.5      0.89           0.36          2.11      1.42
          3.0      0.92           0.31          2.40      1.74

     As an example, assume that steady state conditions were established at
a Ca/S ratio of 2.0.  The sorbent utilization at this condition would be 42
percent.  If the Ca/S ratio were changed to 1.0, within a few hours the
measured SC>2 level above the combustor bed would begin to approach a value
corresponding to the steady state value of sulfur removal efficiency of 67%.
It has been shown that the average sorbent utilization in the reactor bed
does not respond nearly as rapidly and will within the same time interval
more closely correspond to that of the previous established steady state
value of 42 percent for a Ca/S ratio of 2.0.  It would clearly be wrong to
calculate the Ca/S ratio from the values of the sulfur removal efficiency
and the sorbent utilization measured within a few hours of this change in
Ca/S ratio.

     The marked effect of the fresh particles in determining the sulfur
removal, results in the average rate constant observed at a given steady
state value of the sorbent utilization in the continuously fed combustor, i.e.,
the miniplant, being larger than the corresponding value of the rate constant
at the same value of sorbent utilization in a batch combustor.  This can be
seen by comparing the values of kave and kbatch in Table VIII-1.  Stated
another way, if the rate constant obtained in the batch combustor is used
directly in equation (12) the S02 retention calculated at a given gas phase
residence time will be less than that obtained using the above described
model.  This is shown in Figure VIII-11 for Tymochtee dolomite at a 2 second
gas phase residence time.

     The model can also be used to explain the maximum degree of sorbent
utilization that can be achieved in a given continuously fed combustor.  To a
first approximation, the degree of sorbent utilization that can be achieved, is
determined by comparing the time that a sorbent particle spends on the
average in the combustor and the time that it takes an individual sorbent
particle to reach a given level of utilization.  As an example in Figure
VIII-12, the average particle residence time in the miniplant combustor and
the time for a particle to reach different levels of utilization is shown for
dolomite sorbent at a 2 second gas phase residence time in the expanded bed
for different Ca/S feed ratios.  It can be seen, that at a Ca/S of 1.5,  (the
level necessary to meet the EPA new source performance standards for SC>2
emissions when burning a 4 percent sulfur coal) , the achievable sorbent
utilization is only 60 percent, as at this Ca/S feed ratio, the average particle
residence time in the miniplant combustor of 10 hours does not provide suf-
ficient time for the sorbent particle to achieve a higher level of utiliza-
tion.   This is in good agreement with values measured during the course of
this program.  The figure also enables an estimate of the effect on sorbent

                                      146

-------
                       FIGURE Vlll-ll


     S02 RETENTION @ 2 SEC, 900°C VS. Ca/S MOLAR RATIO


          PREDICTED  EXXON MINIPLANT PERFORMANCE


                      PFIZER DOLOMITE


   100
o
o
O
O
o
LU
CO

CM
g

\-
•z.
Ul
I-
LU
 CN

 O
 CO
    90
    80
    70
    60
50
     40
     30
          BATCH  DATA & MODEL
                                 BATCH  DATA

                            USING EQUATION (12)-
        0
                 1.0    1.5     2.0

                 Ca/S MOLAR  RATIO
                              147

-------
   100
co
en
ID
o
o
i-
co
ID
DO
2
o
o
10
                           FIGURE VIII-12


                 TIME IN COMBUSTOR VS. Ca/S RATIO


             EXXON MINIPLANT PERFORMANCE CALCULATION


                          PFIZER DOLOMITE
                AVERAGE PARTICLE

                RESIDENCE TIME IN
        \ /COMBUSTOR
         \/
CALCIUM

UTILIZATION

ACHIEVED
                                   0.8
                                           0.7
      0.5
                                              0.4
          SUPERFICIAL VELOCITY = 6 FT/SEC

          GAS PHASE RESIDENCE TIME =  2  SEC

                  I      i     I      .      I
       0
                            Ca/S RATIO
                               148

-------
utilization if the SC>2 emission standard were to be tightened.  By way of
example, if a 3 to 1 ratio of Ca/S feed were to be required to meet the new
standard, the expected sorbent utilization which could be achieved in the
miniplant would be less  than 40 percent.
                                         149

-------
                                 SECTION IX

                             CONTINUING STUDIES


     The future program in the miniplant will be concentrated on three major
tasks; high temperature, high pressure particulate removal, a comprehensive
analysis of all emissions from the unit and regeneration studies.  After
conversion of the batch unit to a continuous bench scale unit, experimental
programs will begin centering on the study of NOX emissions, evaluation of
various coal and sorbent types and regeneration tests.  These continuing
studies are described in more detail in this section.

HIGH TEMPERATURE, HIGH PRESSURE PARTICULATE REMOVAL

     The particulate removal program is now concentrated on optimizing
the performance of a granular bed filter.  Although operability for over a
24 hour period has been demonstrated, particulate removal efficiency must be
improved, loss of filter media during blow back decreased and certain opera-
tional problems corrected.  Buildup of particulates within the filter beds
indicate the blow back has not been effective in cleaning the beds and this
will be studied.  The use of various filter media of differing size and
density will be studied, filtration and blow back conditions and methods will
also be varied and the effects measured.  It is also planned to use low tem-
perature transparent models to observe the action of the particulates and the
filter media during filtration and blow back.  This will be done in an attempt
to learn more of what is occurring in the filter at high temperatures.

     The filter will then be readied for extended testing.  These tests are
part of the DOE sponsored program to study erosion, corrosion and solids
deposition effects for a series of gas turbine blade materials exposed to
typical PFBC flue gas conditions.  The program, which is a part of the
cooperative EPA/DOE FBC effort, will use gas turbine blade materials and
a test passage supplied by General Electric Company.  In addition to the
gas turbine materials test, boiler tubing materials will also be tested in
samples mounted in the combustor vessel within and above the expanded bed.
These samples will be supplied by Westinghouse Research Laboratory.

     The extended tests will consist of a 100 hour shakedown run to test
the compatibility of the combustor, filter and materials test sections.
If the systems are compatible and the filter performs satisfactorily, long
term testing will begin.  If the systems are not compatible or if the filter
does not perform satisfactorily, a decision will be made whether to proceed
with the long term tests and with what gas cleanup configuration.  The long
term tests will consist of a series of exposure tests totalling 1000 hours
exposure time.  Gas turbine and boiler materials will be periodically
removed and inspected by General Electric and Westinghouse respectively.

     Following the completion of the filter tests, one or possibly two
alternate particulate control devices will be fabricated, installed and
tested on the miniplant in place of the granular bed filter.  The choice
of the alternate devices will be made in consultation with the EPA.  Cur-
rently, devices such as a high temperature electrostatic precipitator or a
high temperature bag filter are being considered.

                                      150

-------
COMPREHENSIVE ANALYSIS

     At the present time, a Level 1 analysis has been conducted on the
effluents from the miniplant combustor.  The program calls for the completion
of another Level 1 analysis with the combustor coupled to the regenerator.
In addition, Level 2 and Level 3 analyses are also planned with the com-
bustor and regenerator.

     In support of the comprehensive analysis and particulate removal pro-
grams, two high temperature particulate sampling systems have been designed
and are now under construction.  These will be used to sample particulates
entering and leaving the granular bed filter to determine particulate con-
centration, size distribution and composition.  Provisions will be included
to maintain temperature of the particulates at 870°C or higher to prevent a
change in composition due to condensation of volatile materials on the
particulates.

REGENERATION

     A test program is being planned to characterize and develop sufficient
information to determine the feasibility of the combustion-regeneration
system.  The primary independent variables are makeup Ca/S rate and solids
recirculation rate.  Another independent variable is sorbent type (limestone
and dolomite).  Dependent variables include S02 emissions from the combustor
and regenerator.  Parameters which are expected to be kept fixed during the
program are combustor and regenerator temperatures, regenerator air/fuel
ratio, combustor excess air level, bed depths, and coal type (Illinois coal
will be used).  The operating pressures in the combustor and regenerator
will probably be fixed at 5 to 6 atm to provide sufficient fluidization
velocity in the regenerator to permit good mixing of solids.  This pressure
level is a compromise.  Otherwise,  significant modifications must be made
to permit operation of the system at 9 atm.

     Runs of two to five days duration are planned.  However, because the
activity of the sorbent is expected to decline gradually over many cycles
of sulfation and regeneration, it may not be practicable to reach a true
steady state.  By removing samples of bed periodically, it may be possible
to follow the activity of the sorbent as a function of time and number of
cycles and extrapolate these results to longer times.  A thermogravimetric
analyzer (TGA) may be used to run a standard "activity" test by sulfating
samples of regenerator bed.  The rate of sulfation would provide a measure
of activity of the recycled stone.

BENCH SCALE UNIT

     The future work in the bench scale combustor unit will evaluate various
coal and sorbent types.  Coals of various rank, ash and sulfur content will
be tested as well as lignites, chars, etc.  The effect of recycle of parti-
culates from the first stage cyclone to the combustor on combustion efficiency
will also be measured.  Staged combustion to reduce NOX levels will be studied.
                                     151

-------
     Regeneration studies will also be carried out.   Activity maintenance of
the recycled sorbent will be measured as a function of regeneration condi-
tions including the use of coal as the regenerator fuel.
                                       152

-------
                                  SECTION X

                                 REFERENCES


 1.  Hoke, R.  C.  et al,  "Studies of the Pressurized Fluidized-Bed Coal
    Combustion Process," EPA-600/7-77-107, September,  1977.

 2.  "Reduction of Atmospheric Pollution," Vol.  1,  p.  119, National Coal
    Board  (U.K.), September, 1971.

 3.  Keairns,  D.  L. et al, "Fluidized Bed Combustion Process  Evaluation -
    Phase  II  - Pressurized Fluidized-Bed Coal Combustion Development,"
    EPA/650-2-75-027C,  September, 1975.

 4.  Vogel,  G. J. et al, "Bench Scale Development of Combustion and
    Additive  Regeneration in Fluidized Beds," Proceedings of the Third
    International Conference on Fluidized-Bed Combustion, EPA-650/2-73-053
    p.  1-1-1, December, 1973.

 5.  O'Neill,  E.  P. et al, "A Thermogravimetric Study  of  Limestone and
    Dolomite  - The Effect of Calcination Conditions," Thermochemica Acta
    1£,  209 (1976).

 6.  Jonke,  A. A. et al, "Sulfated Limestone Regeneration and General FBC
    Support Studies," Proceedings of the Fluidized Bed Combustion Technology
    Workshop, Vol. II,  p. 343, CONF-770447-P-2, April 13-15, 1977.

 7.  Skopp,  A et al, "Studies of the Fluidized Lime-Bed Coal  Combustion
    Desulfurization System," December 31, 1971.

 8.  Hoke,  R.  C.  et al,  "A Regenerative Limestone Process for Fluidized-Bed
    Coal Combustion and Desulfurization," EPA-650/2-74-001,  January, 1974.

 9.  Hoke,  R.  C.  et al,  "Studies of the Pressurized Fluidized-Bed Coal
    Combustion Process," EPA-600/7-76-011, September, 1976.

10.  "Pressurized Fluidised Bed Combustion," R&D Report No.  85,  Interim
    No.  1,  National Coal Board (U.K.), September,  1973.

11.  Vogel,  G. J. et al, "Recent ANL Bench-Scale, Pressurized Fluidized
    Bed Studies," Proceedings of the Fourth International Conference on
    Fluidized-Bed Combustion, p. 21, December 9-11, 1975.

12.  Hoy, H. R.,  Roberts, A. G., "Further Experiments  on  a Pilot-Scale
    Pressurized Fluidized Bed Combustor at Leatherhead,  England,"
    Proceedings of the Fluidized Bed Combustion Technology  Workshop
    Vol. II,  p.  43, April 13-15, 1977, CONF-770447-P-2.

13.  Nauman, E. G., Collinge, C. N., Chem. Eng. Sci.,  23., 1317 (1968).
                                     153

-------
14.  Borgwardt, "Kinetics of Reaction of S02 with Calcined Limestone,"
     Env. Sci.  & Tech.,  4_, 1, January, 1970, pp.  59-63.

15.  Vogel, G.  J. et al, "Reduction of Atmospheric Pollution by the
     Application of Fluidized-Bed Combustion and  Regeneration of Sulfur-
     Containing Additives," EPA-650/2-74-104,  September,  1974.

16.  Grumpier,  T. B.,  Yoe, J. H., Chemical Computations and Errors, John
     Wiley NY,  1949.

17.  Cooper, L., "Measurement of High-Temperature, High-Pressure Processes,"
     EPA-600/7-78-011, January,  1978.

18.  Murthy, K. S., et al, "Comprehensive Analysis of  Emissions from
     Fluidized  Bed Combustion Processes," Process Measurements for Environ-
     mental Assessment Symposium, Atlanta,  GA,  February,  1978.
                                     154

-------
                                 SECTION XI

                            LIST OF PUBLICATIONS


1.  Hodges, J. L.,  Hoke,  R.  C., Bertrand,  R.  R.t  "Prediction of Temperature
    Profiles  in  Fluid  Bed Boilers," ASME/AIChE Heat  Transfer Conference,
    St. Louis, MO,  August 9-11, 1976.

2.  Hoke, R.  C.,  "Particulate Control  in Pressurized FBC-Granular Bed
    Filter Applications," High Temperature and Pressure  Particulate Control
    Symposium.   Washington,  DC, November 9,  1976.

3.  Hoke, R.  C.,  Nutkis,  M.  S., Kinzler, D.  D., "Pressurized FBC Studies I.
    Combustion," Proceedings of the Fluidized Bed Combustion Technology
    Exchange  Workshop  Vol. II, p.  157, CONF-770447-P-2,  Reston, VA,
    April 13-15,  1977.

4.  Nutkis, M.  S.,  Loughnane, M. D., "A Program for  Hot  Corrosion/Erosion
    Materials Testing  for Application  to Fluidized Bed Coal Combustion,"
    Proceedings  of  the Fluidized Bed Combustion Technology Exchange
    Workshop, Vol.  II, p.217, CONF-770447-P-2, Reston, VA, April 13-15, 1977.

 5-  Ruth, L.  A., "Regenerable Sorbents for Fluidized Bed Combustion,"
    Proceedings  of  the Fluidized Bed Combustion Technology Exchange
    Workshop  Vol. II,  p.  301, CONF-770447-P-2, Reston, VA, April 13-15, 1977.

 6.  Ruth, L.  A., Gregory, M. W. , Bertrand, R. R., "Pressurized FBC Studies
    II.   Sorbent Regeneration and Particulate Control,"  Proceedings of the
    Fluidized Bed Combustion Technology Exchange Workshop Vol. II, p. 329,
    CONF-770447-P-2, Reston, VA, April 13-15, 1977.

 7.  Nutkis, M.S., "Hot Corrosion/Erosion Materials Program and Experience
    in the Pressurized Fluidized Bed Coal  Combustion Miniplant," Engineering
    Foundation/ASME Conference on Ash  Deposits and Corrosion Due to
    Impurities in Combustion Gases, Henniker, "NH, June,  1977.

 8.  Bertrand, R. R., "Temperature Control  in the Exxon Fluidized Bed
    Combustion Miniplant," 1977 ISA Symposium on Instrumentation and
    Control for  Fossil Demonstration Plants, Chicago, IL, July 13, 1977.

 9.  Hoke, R.  C., "Pressurized Fluidized Bed Combustion," Proceedings of
    the  12th  Intersociety Energy Conversion Engineering  Conference, Vol. I
    p.  737, Washington, DC, August 28  - September 2, 1977.

10.  Ruth, L.  A., "Sorbent Regeneration in  Fluidized  Bed  Combustion,"
    Proceedings  of  the 12th Intersociety Energy Conversion Engineering
    Conference,  Vol. II,  p. 758, Washington, DC,  August  28 -  September
    2, 1977.
                                     155

-------
11.  Ruth, L. A., Varga, G.  M.,  "Developing Regenerable S02 Sorbents for
     Fluidized Bed Coal Combustion Using Thermogravimentric Analysis,"
     paper presented at NATAS 7th Conference, St. Louis, MO, September, 1977,

12.  Nutkis, M. S., "Hot Corrosion Erosion Testing of Materials for
     Application to Advanced Power Conversion Systems Using Coal-Derived
     Fuels Task II - Fluidized Bed Combustion," First International
     Conference on Materials for Coal Conversion and Utilization,
     CONF-771025 UC-90h, V-l, Gaithersburg, MD, October 11-13,  1977.
                             OTHER PRESENTATIONS
Bertrand, R. R., Hoke,  R.  C.,  Seminar on Fluidized Bed Combustion,
Massachusetts Institute of Technology, January 1977.

Hoke, R. C., Seminar on Fluidized Bed Combustion,  Polytechnic  Institute
of New York, October, 1977.
                                    156

-------
                                  SECTION XII

                                  APPENDICES

                                                                         Page

A          FBC Solid Effluent Rates, Balances, Composition               159

B          Second Order Correlation of S02 Retention Data                164

C          Error Analysis                                                182

D          Data Management Systems                                       183

E          Analytical Techniques                                         196

F          Data Logger Channel Identification                            197

G          FBC Miniplant Alarm/Shutdown System                           199

H          Tables

  H-l      Miniplant Fluidized Bed Coal Combustion Run Summary           203

  H-2      Particle Size Distribution Spent Pfizer 1337                  211
           Dolomite Sorbent  (Except As Noted)

  H-3      Particle Size Distribution Primary Cyclone Capture            211

  H-4      Particle Size Distribution Secondary Cyclone Capture          212

  H-5      Particle Size Distribution GBF Capture                        213

  H-6      Particle Size Distribution Flue Gas                           214
           Particulates No Filter

  H-7      Particle Size Distribution Flue                               214
           Particulates After Filter

  H-8      Miniplant Solids Analysis                                     215

  H-9      Miniplant Solids Composition                                  221

  H-10     Miniplant Sample Shipments                                    224

J          Figures

  J-l      Modified Filter Bed                                           227

  J-2      Modified Filter Element                                       228
           Fluidizing Grid

  J-3      GBF Pressure Vessel                                           229


                                       157

-------
                          APPENDICES (Continued)




                                                                       Page




J-4      GBF Pressure Vessel Lining                                    230




J-5      Filter Pressure Vessel (Side View)                            231




J-6      GBF Piping Arrangement                                        232




J-7      Preheat Burner System Schematic                               233




J-8      Granular Bed Filter Blow Back System Flow Schematic           234
                                    158

-------
                                 APPENDIX A

               FBC SOLID EFFLUENT RATES, BALANCES, COMPOSITION


     Solid rates, concentrations and compositions were summarized for two
combustion runs made in the Exxon miniplant using Illinois No. 6 coal and
Pfizer No. 1337 dolomite.  Conditions and emission results for the two runs,
Nos. 43.4 and 43.5, are given in Appendix H-l.  Both runs were made at a
pressure of 940 kPa, an average bed temperature of 950°C, a superficial
velocity of 1.7 m/s and excess air of 25%.  Run 43.4 was made at a Ca/S
ratio of 1.8 and 43.5 at 0.7.  Run 43.4 achieved a 96% retention of S02,
run 43.5 a 44% retention.

SOLIDS RATES AND BALANCES

     A calcium balance was made over the entire time period during which
runs 43.4 and 43.5 were selected.  The balance was more meaningful over the
entire time period because the bed height and composition changed during
the time period.  The results of the calcium balance are shown in Table
A-l.  The imbalance of only 7% is very good.

     Individual calcium balances were then calculated for the two runs con-
sidering only the steady state periods and the results are shown in Table
A-2.  It is evident that the "loss" of calcium from runs 43.4 and
43.5 is due to bed build up.

     If the bed had been maintained at a constant height, the fraction of
calcium in the bed removal line would have been 0.65 using the overall
results and 0.59 using the combined results at steady state.  The overall
results, when normalized to 100%, give the ratio of bed-removal-calcium
to overhead-calcium of 58:42 or, roughly 60:40.

     Since the amount of calcium removed from the bed is less in run 43.4,
this suggests that the bed build up occurred primarily, if not completely,
in run 43.4.  This was verified by bed height measurements which indicated
a 30% build up in bed height in run 43.4 and a 3% build up in run 43.5.

     The ratios of solids added and removed from the combustor to the coal
feed are shown in Table A-3.  Since a build up occurred in 43.4, the solids
retained in the bed to coal feed ratio was estimated and is also shown in
the table.  The total solids removed and retained in 43.4 equal 0.27 wt/wt
coal which is about 80% of the expected weight ratio.  In run 43.5, the
total solids removed equal 0.23 wt/wt coal which is about 115% of the
expected weight ratio.  Therefore, even with the correction for an increase
in bed height, the disposition of the output solids was not properly split
between the two runs.  However, by combining the two runs and calculating
average input and output rates, an acceptable ratio of measured to expected
solids output of 95% was determined.
                                     159

-------
                     TABLE A-l.  CALCIUM  BALANCE
                  Runs 43.4-43.5 Entire Time Period

                        Percent of Ca Into Bed

                  Fly Ash Line(1)                41
                  Bed Removal Line               38
                  Particulates                    1
                  Increase of Ca in Bed          27
                                     TOTAL      107
                  (1)  Secondary cyclone
                        discharge
               TABLE A-2.  INDIVIDUAL CALCIUM BALANCES
               Runs 43.4 and 43.5 Steady State Periods
                       Percent of Ca Into Bed
Run No.
43.4
43.5
Combined Runs
Bed Removal
Line
27
44
32
Fly Ash
Line
43
40
42
Particulates
0.4
0.7
0.6
Total
70
85
75
Combined Runs Corrected        ,Q           ,„         n  ,         ino
 ,.   Tl 1 T, • I 1 TT               ->•'           ^         U.D         1U2
 for Bed Build Up
                                  160

-------
TABLE A-3.  SOLIDS/COAL WEIGHT RATIOS
Run No .
43.4
43.5
Combined
Coal
Fed
1.00
1.00
Sorbent Bed Removal
Fed Line
0.32 0.05
0.13 0.10
1.00 0.22 0.08
Fly Ash
Line(l)
0.13
0.13
0.13
Flue Gas Retained Z Solids Out
Particulate in Bed Expectation
0.004 -0.09 -0.80
0.004 0 1.15
0.004 "0.04 -0.95
(1) Secondary Cyclone Discharge

Run No. Bed

43.4
43.5

Removal

0.09
0.10
TABLE
Ash Removed /Ash Fed
Line Fly Ash Line

0.61
0.81
A-4. ASH BALANCE
Particulates

0.03
0.03
Retained Corrected
Total in Bed Balance
(Ash Balance)
0.73 -0.27 1.00
0.94 0 0.94

-------
     Ash balances were also calculated for the two runs.  The results are
shown in Table A-4.   Again, if a correction is applied to 43.4 for the build
up in the bed, good  balances are calculated.  As shown in Table A-4, 65 to
85% of the ash is removed overhead,  the balance is removed with the used
bed.

     A sulfur balance was also made  during the entire run period from start
up to shut down.  The results are shown in Table A-5.  The balance on sulfur
was 87%, while this  was not as good  as the calcium balance, it is nevertheless
reasonable.

SOLIDS COMPOSITION

     The composition of the solids removed from the bed and the fly ash is
given in Table A-6.   The percentage  of the solids present as unburned carbon
and used sorbent was calculated from chemical analysis of the solids.
The percentage of ash in the solids  was then obtained by difference.   The
composition of the calcium portion of the bed removal solids and fly ash
was also calculated  from the chemical composition and is given in Table A-6.

     Solids from Run 43.5 were also  analyzed for sodium and potassium by
atomic absorption.  The results are  given in Table A-7.
                                     162

-------
Run No.
                  TABLE A-5.  SULFUR BALANCE
                        Runs 43.4-43.5

                     Percent of S into Bed
               Fly Ash Line                  25%
               Bed Removal Line              18%
               Stack Gas                     27%
               Increase of S in Bed          17%
                                  TOTAL      87%


              TABLE A-6.  SOLIDS COMPOSITION (WT %)
Carbon
Sorbent
Ash
   Sorbent Calcium
     Composition
CaO    CaC03    CaSO/
43.4
Bed Removal Line
Fly Ash Line
43.5
Bed Removal Line
Fly Ash Line
Particulate

1
1

1
1
3

82
54

89
35
35

17
45

10
64
62

37
16

29
24
—

10
6

7
3
—

53
78

64
73
—
         TABLE A-7.  ALKALI CONTENT OF  SOLIDS - RUN 43.5
         Solid
     Dolomite
     Bed Material
       (After 43.4,
       Before 43.5)
     Fly Ash
     Particulates
            Na (Wt %)
               0.04

               0.05

               0.37
               0.80
                     K (Wt %)
                       0.008

                       0.02

                       0.63
                       1.34
                             163

-------
                                 APPENDIX B

               SECOND ORDER CORRELATION OF S02 RETENTION DATA


     In the previous report (1), sulfation data from Exxon and other
laboratories, using dolomite sorbent, were correlated by using a first order
expression for the rate constant (Eq. 1).  The rate constant was then
related to the degree of sulfation of the dolomite.


                 kx = -7 In(l-x)                                        

                 k^ = first order constant

                 t  = residence time  (bed height/superficial velocity)

                 x  = fraction SO,, retention  (SR)

This procedure gives a good correlation but has some draw backs when
used for design purposes.  For example the Ca/S molar ratio does not enter
explicitly into the value of x that would be  calculated for a given set of
conditions.  Also, the sulfation level is not known a priori (moles SO^
in bed/moles Ca in bed).

     Another correlation has been developed that not only allows the S02
retention to be calculated but also takes into account all the major design
parameters.  This technique is derived by assuming that the rate of removal
of S02 and CaO is second order  (first order in CaO and first order in 802).
It is assumed that sulfur in the coal is instantly converted to 862 and that
the dolomite is instantly calcined.   It is also assumed that the gas/solid
reduction between S02 and the sorbent can be  treated in a manner similar
to that used for homogeneous reactions.

     The second order rate expression is

                 &L =  k7(a'x')(b'-x')                                (2)
                 at       ^

                                                      3
     where a'  =  initial concentration of CaO, mols/m
                                                      3
           b1  =  initial concentration of S09, mols/m
                                              L            3
           x1  =  concentration of product  (CaSO^), mols/m

           t   =  time,  s
                                                3
           k   =  second order rate constant, m /mol-s

     Integration of  this expression gives

                                      _  ,
                                      -  V

     On dividing the numerator and denominator  of  the logarithmic  operand
by (b')2, one obtains


                                     164

-------
                             x'
              a'-b'    £l n-^li
                       u t ^•L -u ' '
                       D     D

     By definition, a'/b' = a = Ca/S molar ratio in the feed.  Similarly,
x'/b' = x = fraction of S02 retained at steady state.

One then arrives at:

                    -rW In  *:x .  =  k9t                             (5)
                    a -b     a(l-x)      2

If b1 is factored out of the prelogarithmic term, there results

                     1        a—x
                  7-7-7—TV In —T~I	r  =  k0t                             (6)
                  b'(a-l)    a(l-x)      2

     The quantity b', the initial concentration of S0_, is given by


                                  (R)
                         (11.52 x 10b)(V)(A)


     where S  =  Percent sulfur in coal,
           R  =  Coal feed rate, kg/hr,
           V  =  Superficial velocity of gas through the bed, m/sec,
           A  =  Reactor cross sectional area, m2^
                         f\                                               ^
           and 11.52 x 10  is a conversion factor so that b' is in mols/m .

The reaction time is given by

                           t  =  H/V                                    (8)

where H is the expanded bed height in meters.

Upon substitution of (7) and (8) into (6), one arrives at an equation whose
parameters are those commonly used in FBC technology:

                  11.52 x 106(V)(A)     a-x    _ ,H _ f
                              l)    ±n a(l-x)  ~ *2V ~ *
     Another method of arriving at equation (9) is given in a subsequent
section.  That method assumes SO. moves in plug flow but that the solids are
well mixed.

     The value of k2 is found by plotting f vs t, where t = H/V.  The
plot, using data from Exxon miniplant dolomite runs 19 through 51, is shown
in Figure B-l.  The several points shown for run 51 were obtained in an
experiment in which the bed height was altered while other parameters
remained constant.

                                     165

-------
                                  FIGURE B-l
                SECOND ORDER PLOT USING ACTUAL CONCENTRATIONS
5 -
4 -
 (/}

1  3
  v
r
o
r-H
 x  2
0
            O  ERE Champion Coal
            D  ERE Illinois Coal
            A  ERE Run 51
                            A
                          O
                                                           O
                                                   O
                                              O
                                                                         o
                                                                o   o
                                         D

-------
                                                                         3
     The least square fit of the data gives a value for k2 of 9.512 x 10
m /mole-sec.

     To determine the expected retention of SC»2 for any set of parameters,
equation (9) is inverted to give
                       x  =
Where c  =
                              ae -1

                  (k2) (a-1) (S) (R) (H)
                 - - - s-
                 11.52 x 10b(A)(V)
     The reliability of equation (10) may be seen by calculating the expected
S02 retention and plotting the calculated values vs the measured values.
This has been done for the ER&E data and the results are shown in Figure
B-2.  Included in this figure are runs where x > a.  These data could not
be used in equation (9) to estimate k2.  (It should be noted that in equa-
tions (9) and (10), when a = 1, indeterminate values are found.  In cal-
culations, it is only necessary to set a = 0.99999 or 1.000001 and then
perform the calculation.)

     To test the validity of equation 10 even further, data were used from
Argonne National Laboratories (ANL)  (15)  and from the National Coal Board
(NCB)(10,12).  The retentions calculated from equation (10) using the second
order rate constant determined from the Exxon data were plotted against the
measured retentions and the results are shown in Figure B-3.  Again, a
good fit is found.

     The scope of Equation 10 can be appreciated when it is realized that
the expected retention can be calculated with only one adjustable constant,
k£, which is obtained from actual experimental data.  The range of variables
in Figures B-2 and B-3 are shown in Table B-l .  It can be seen that equa-
tion (10) has been shown to be valid over a wide -range of variables.

     Equation (10) can be used to show the effect on sulfur retention of
changing one operating variable while the other variables are held constant.
Figures B-4 and B-5 show the effect of the variables in this fashion.

     The top section of Figure B-4 shows the S02 retention vs the feed
Ca/S ratio for coals of different sulfur content.  For a given Ca/S ratio,
the retention increases as coal sulfur content increases.  This effect dif-
fers from that predicted by the first order model in which S02 retention is
independent of the sulfur content of the coal.  Also shown in the figure is
the locus of Ca/S requirements to give 0.52 g S02/MJ (1.2 Ib S02/M BTU) for
coals of 2.8 MJ/kg (12,000 BTU/lb) and 1.4 MJ/kg (6000 BTU/lb) .  It is
interesting to note that for the coal with the lower heating value, a
larger Ca/S ratio is required for the 1% coal than the 6% coal.

     The superficial velocity (V) has a dramatic effect on the retention
when all other factors are constant.  This effect is shown in the central
section of Figure B-4 for two levels of Ca/S.  The large effect is not
surprising since V enters Equation 10 to the second power.
                                     167

-------
                      FIGURE B-2



       PREDICTED S02 RETENTION VS. MEASURED S09

            RETENTION, EXXON DOLOMITE RUNS
o

i-
z
UJ
I-
LlJ
 C\J
O
CO

Q
UJ
h-
o

Q
UJ
100


 90


 80


 70



 60


 50


 40


 30


 20


 10


  0
                                       I
I
      0   10  20  30 40  50 60  70  80  90 100
               MEASURED S02 RETENTION, %
                        168

-------
                     FIGURE B-3
     PREDICTED RETENTION VS.  MEASURED RETENTION
LU
I-
LJ
C£

a,
LU
i-
o
o
LJ
100


 90


 80


 70


 60


 50


 40


 30


 20


 10


  0
             _L
                               O ANL VAR Series
                               • NCB (12)
                               D NCB (10)
_L
      0   10 20  30  40  50  60  70 80  90 100


                MEASURED RETENTION, %
                      169

-------
                           TABLE  B-l.   RANGE  OF  VARIABLES  FOR  WHICH  EQUATION
                                      10  HAS  BEEN  SHOWN  TO BE  VALID
Variable              ^	    Range for Data Source

A, tn
S, %
Ca/S
V, m/s
H, m
R, kg/hr
ER&E
0.0794-0.0873
1.96-4.2
0.5-2.5
1.52-2.96
2.4-7
78-149
ANL
0.0183
2.82
1-3.2
0.64-1.49
0.915
7.75-23.43
NCB
0.558
2.97
1-2.18
0.7-0.762
1.34-2.44
144-273
OVERALL (NOMINAL)
0.018-0.56
2-4.2
0.5-3.2
0.64-3.0
0.92-7
7.8-270

-------
LJ
0"
o
I-
2
UJ
I-
UJ
 CM
o
C/)
100

 80

 60

 40

 20

  0




100

 80

 60

 40

 20
       0
                            FIGURE B-4

               PREDICTIONS USING  SECOND-ORDER  MODEL

                 EFFECT OF Ca/S  RATIO  ON RETENTION
                                             O 1.96% S Coal
                                             • 0,96% S Coal
                                             D 4% S Coal
                                             • 6% S Coal
                                Ca/S

            EFFECT OF SUPERFICIAL VELOCITY ON  RETENTION
                                         O Ca/S =1.0
                                         • Ca/S =1.5
                                         Bed Height = 4m
                                         Coal Feed Rate =
                                           100  kg/hr.
                                                1.96% S Coal
       0123456

                     SUPERFICIAL VELOCITY m/sec.

               EFFECT OF BED HEIGHT ON S02 RETENTION
                              Ca/S =1.0
                              Superficial  Velocity =  2 m/sec.
                              Coal Feed  Rate =100 kg/hr.
                              1.96% S Coal
                      468

                           BED HEIGHT, m
                                         10
12
                               171

-------
LU
h-
LU
 CM
O
100

 80

 60

 40

 20

  0
       0
                              FIGURE B-5

                PREDICTIONS USING SECOND-ORDER MODEL

           EFFECT OF  COAL SULFUR CONTENT ON S02 RETENTION
Ca/S = 1.5
Superficial Velocity = 2 m/sec.
Bed Height = 4m
Coal Feed Rate =100 kg/hr.
                              _L
                           3      4

                       % SULFUR IN COAL
    lOOi

I    80
LU
I-
LU
 CM
O
 60
     40
     20
      0
       0
             EFFECT OF COAL FEED  RATE ON S02 RETENTION
                          Ca/S = 1.0
                          Superficial Velocity = 2 m/sec.
                          Bed Height = 4 m
                          S Content o,f Coal = 1.96%
                  50              100
                         COAL FEED RATE, kg/hr.
                       150
                                 172

-------
     The lower portion of Figure B-4 illustrates the effect of bed height
on the retention.  The effect is significant, especially with shallow beds.

     The effects of coal sulfur content (S) and coal feed rate (R) are
portrayed in Figure B-5.  Both effects arise because of the higher initial
sulfur contentration with increasing S or R.

     It should be noted, on reviewing Table B-l, that Equation 10 has been
shown to hold over most of the range of Ca/S in the upper portion of Figure
B-4 (Ca/S = 0.5-3.2).  The equation holds for a good portion of the central
part of Figure B-4 (V = 0.64-3.0) and for the lower section (H = 0.92-7).
Experimental data confirm Equation 10 for a large section of the upper
portion of Figure B-5 (S = 2-4.2) and for all of the lower portion of B-5
(R = 7.8-270).

     Figure B-6 shows the percent S02 retention vs the Ca/S ratio for
different residence times, t, when other parameters are held constant.  The
actual parameter varied is the superficial velocity for the solid lines and
change of the bed height at t = 1 s for the dashed line.  As seen in Equation
(10), velocity enters as the square while H enters as the first power.
If this model is valid, there is no unique value for sulfur retention at a
given value of residence time.

     Since equation (10) can be used so effectively for design purposes, it
would be very desirable to delineate the extent for which the equation is
valid.  The region of low retention would be very valuable to investigate.

     Figure B-7 shows how this was done for one experiment during run 51.
The bed height was varied while all other variables were constant.  Figure
B-7 shows that the calculated retention fits the experimental data well over
a narrow range of bed heights.  It would be useful, however, to expand the
experimental range from, for example, a bed height of 1m to 5m.

     This second order kinetic model is an alternate way to analyze, cor-
relate data and predict results.  It has certain advantages over this first
order model described in the previous report (1), and appears to correlate
the data well.  Both methods will be evaluated further as more data becomes
available.

ALTERNATE DERIVATION OF SECOND
ORDER RATE EXPRESSION
                    Ca.
                                              Ca CaS
                                                o   o
                                     173

-------
LJ
\-
LJ
 CM
O
LO
100-


 90


 80


 70


 60


 50


 40


 30


 20


 10



   °0
                          FIGURE B-6

            PREDICTIONS USING SECOND-ORDER MODEL

                   EFFECT OF Ca/S RATIO FOR
                  DIFFERENT RESIDENCE TIMES
                                 3 SEC.
                                             2 SEC.
                                                  1 SEC
                                             1 SEC,
                                Superficial velocity varied

                                Bed Height varied
                              _J	
                                2

                               Ca/S
                              174

-------
UJ
h-
LU
or
                     FIGURE B-7


        COMPARISON OF EXPERIMENTAL EFFECT OF
             BED HEIGHT WITH THEORETICAL
100

 90

 80

 70

 60

 50

 40

 30

 20

 10
      0
—  THEORETICAL
 O  RUN 51
    (5/6/77)
                                8
              10
                    BED HEIGHT, m
                          175

-------
     The above diagram is a representation of a fluid bed combustor (FBC) .
The sulfur enters in the coal and  the CaC03 enters as a solid.  Assume that
the sulfur in the coal as instantly  converted to S02 and the sorbent is
instantly calcined.

     S.^  =  rate of sulfur input,  mols/s
         =  rate of S0_ input

    Ca.^  =  rate of Ca input, mols/s
         =  rate of CaO input, mols/s

      x  =  fraction of sulfur retained

     SQ  =  rate of sulfur out,  mols/s
         =  rate of S02 out, mols/s

    Ca   =  rate of CaO out, mols/s

   CaS   =  rate of CaSO, out, mols/s

The S0?  leaves  as  a gas and  the  CaO  and  CaSO^ leave  in various solid
streams .

     The assumption is made  that the solids are completely mixed and the
S02 moves  through  the  reactor  in plug flow.

     Now by definition,

                 CaS                                                    , _
                 -^ =  *                                             (11)
                   bi

     A material balance  and  (11) give

                 Ca.-Ca   =   CaS   = x S.                               (12)
                    10         o       i

     On  rearranging  (12),  one obtains

                 Ca    =  Ca.-xS.
                   o        11
      Factoring S.  out of (13) gives
                 Ca   —  S. Oj   ~ x;
                   o         l

     Now by definition, Cai/Si  =  Ca/S molar ratio in the feed.  On
denoting this by a,

                 Ca   =  S. (a-x)
                   o      i
                                      176
                                                                        (15)

-------
     The initial concentration of S02 is S^APR, where AFR is  the air flow
rate in m^/sec.  The concentration of CaO,  [Ca] is obtained by dividing (15)
by the air flow rate, AFR.  (This may be visualized by assuming all CaO is
taken overhead in the gas stream.)  Equation  (15)  then becomes

                                      S.
                 [Ca]  =  Ca /AFR  =  -^(a-x)                        (16)
                            O         Ar K.

     Now S-^/AFR is the initial concentration of SOo in the gas stream.   If
this concentration, in mols/m^, is denoted by b, Equation  6 becomes

                 [Ca]  =  b(a-x)                                      (17)

     If the reasonable assumption is made that the rate of disappearance of
S02 is proportional to the sulfur dioxide and calcium oxide concentrations,
one obtains


                 -^p-  =  -k[S][Ca]                                   (18)

     When  (17) is substituted into  (18), there results


                       =  -k[S]b(a-x)                                 (19)
                  U I

     Now [S] at the time of exit, T, is given by

                           S
                        _   o
                           AFR
     The value of S  is
                   o
                         S.-CaSo                                      (21)
     From  (10),  (21) becomes

                  SQ  =  S..-S..X  =  S^l-x)                           (22)

     When  (22) is substituted into  (20), there results
                  [S]T  =     (1~X)  =  b(l-x)                        (23)
     On substitution of [S] into (19), one obtains
                      T     bd(l-x)      .,2,,  .,
                         =           =  ~kb d-xXa-*)                (24)
                                     177

-------
     or, on dividing by b,

                  d(l-x)
                  —fr.    =   -kb (1-x) (a-x)                             (25)

     If one lets 1-x = y>  (25) becomes

                  dy
                  -ft  =  ~kb(y)(a-l+y)                                 (26)

     This can be integrated directly  to give
                  c  '
     Substitution of (29) into (28) and rearranging gives
                  b(a-l) *" a (1-x)
     Now b, the initial concentration of SC^, is given by

                  b =
                      11.52 x 106 (V)(A)
Where (S)  =  % sulfur in the coal
        R  =  coal feed rate, kg/hr
        V  =  superficial velocity of gas, m/s
        A  =  combustor cross section, m^
      11.52 x Id**  =  conversion factor
                                                                      (27)
     Substituting 1-x = y into  (27) gives

                  KihTlnfS  =  kT + c                              <28>
     at T = o, x = o and
                         In   *"*,  = kT                               (30)
                                    178

-------
     When (31) is substituted into  (30), one obtains


                  f  -  11.52 x 106  (V)(A)  ,_ a-x     _     H           ,,„,
                  f	(S) (R) (a-1)	  lnl(1^0   -   kV           (32)

Where H  =  bed height (H/V = T)


     The units of k are m /mol-s.  The parameters in  (32)  are  those commonly
used in FBC work.
                                    179

-------
CO
o
             Coal Feed

 Run        Rate (kg/hr)

               135
               139
               133
               122
27.6           123
27.7           132
27.8           134
27.9           142
27.10          135
27.11          149
27.12          133
27.13          134
27.14          137
27.15          136
27.17          136
27.18          143
27.19          137
35              80
37             110
38.1            90
38.2            90
43.2            88
43.3            93
43.4            96
43.5            95
48              90.9
50.1           74.5
51             78.5
                                       TABLE B-2.  CALCULATED VS MEASURED S02
                                        RETENTION USING SECOND ORDER EQUATION

                                                    Exxon Data

                                                  Coal % S = 1.96
                                   Superficial
                                   Velocity(m/s)
2.01
2.15
1.98
1.72
1.72
1.72
1.83
2.09
1.87
2.23
1.94
2.02
2.04
1.88
1.97
2.09
2.08
1.54
2.96
2.08
2.09
1.81
1.77
1.69
1.71
1.46
1.19
0.5
0.5
0.5
2.5
0.8
1.5
1.5
1.5
0.75
0.75
0.75
0.35
0.75
1.0
0.72
1.0
1.5
0.75
0.75
0.75
0.75
0.75
0.75
1.25
0.5
1.4
1.25
                                       1.52
1.4
Expanded Bed
^Height (m)_

    4
    4
    3
    6
    5
    6
    7
    5
    5
    7
    7
    6
    6
    5
    6
    7
    7
    4.4
    4
    2.9
    3.6
    5.7
    3.1
    3.6
    3.7
    3.4
    3.0
    2.4
% S02
Retention
Calc.
42
43
40
100
72
99
99
95
67
68
70
34
68
82
66
85
98
63
20
31
41
73
68
96
49
91
% S02
Retention
Meas.
55
57


80
97.2
98
98.5
72
70
71
46
64
71
72
84
90
62
46
40
45
60
63
96
44
58
75
75

-------
I-1
OO
                                 TABLE B-2  (Continued).  CALCULATED VS MEASURED  S02
                                        RETENTION USING  SECOND ORDER EQUATION


                                                       ANL Data


Run
1
2
3
4
5
6
6R
62R
7
8
9

Superficial
Velocity
(m/s)
0.64
1.46
0.64
0.701
1.037
1.098
1.098
1.067
1.280
1.098
1.494


Coal Feed
Rate (kg/hr)
8.42
23.43
7.63
7.75
13.79
13.75
13.21
13.66
19.07
12.73
18.12
A = 0.01825 m
Coal % S = 2.82

Expanded Bed
Height (m)
0.915
0.915
0.915
0.915
0.915
0.915
0.915
0.915
0.915
0.915
0.915
o

-------
                                 APPENDIX C

                               ERROR ANALYSIS


     An analysis was made of the expected error in the value of the calculated
Ca/S molar feed ratio used to correlate desulfurization results.  The Ca/S
ratio is calculated from two measured quantities, the S02 retention and the
degree of calcium sulfation  (or utilization) in the used sorbent.  The error
in each of the measured quantities was first estimated and used to determine
the error in the Ca/S ratio.

     The S02 retention is calculated from the following expression

                               Air rate x (% SO- + SO  in flue gas)
         SO,, Retention  =  1
                                        - - -— — - -
                                     coal rate x  (% S in coal)
The calcium sulfation level is given by
moles S04,          moles
         ;
                                 (         ,
                                          ;       us
            ra c_.    _       oles Ca  FA         oles Ca
            Ca
Where:  FA  =  Fly ash generation rate

        US  =  Used sorbent generation rate
and the subscripts denote  the  S04/Ca ratio in the fly ash and used sorbent.


     The errors in the S02 retention and Ca sulfation were estimated using
standard propagation of  error  calculation methods (16) .  The errors were
estimated based on the following estimated uncertainities in each of the
individual measured quantities used to calculate the S02 retention and Ca
sulfation.

                    Measured Quantity           Uncertainty

                  Air rate                          1   %
                  S02 emission                   10   ppm
                  Coal rate                         2   %
                  303 emission                   10   ppm
                  S in coal                         5   %
                  Fly ash  rate                     8.5 %
                  Used sorbent rate                 8   %
                  SO^" in  fly  ash                   6   %
                  Ca in  fly ash                     6   %
                  S04= in  used sorbent              6   %
                  Ca in  used sorbent                6   %

     Applying these uncertainties to run 50.2 gave  an estimated error in the
S02 retention of 3.3% at a retention level of 88%.  The error in the calcium
sulfation was estimated  as 13.4% at a sulfation level of 52%.  These uncer-
tainties generated an error in the calculated Ca/S  ratio of 14% at a Ca/S
ratio of 1.63.

                                     182

-------
                                  APPENDIX D

                            DATA MANAGEMENT SYSTEMS


      A run in the miniplant combustor usually spans a time period of about
ten hours.  During this period the conditions in the combustor approach
steady state values resulting in three to four hour data generating periods
characteristic of the chosen set of input variables.  In runs were the regen-
erator is included in the system, the approach to steady state conditions is
much slower, due not only to the presence of the regenerator vessel, but also
the deactivation of the recycled sorbent.  The large quantity of data gen-
erated during these extended runs requires a systematic method of data
management.  It is the purpose of this section to describe systems now being
used to collect, compute, and summarize the data generated in the miniplant.

DATA GENERATION

      Run 51 will be used as an example of the data that are generated during
a typical run in the miniplant.  In this run, only the miniplant combustor
was operated.  The regenerator and the granular bed filter systems were not
used.  The primary objective of this run, which spanned a time period of
7-1/2 hours, was to measure combustion efficiency and particulate loading in
the flue gas when operating without recycle of material from the first stage
cyclone back into the combustor.

      During the run, operational parameters and some of the desired output
results were continuously recorded by the data logging system at one minute
intervals.  Of the 100 channels of information available on the data logger,
70 channels were in use.  The remaining channels were connected to instru-
ments used only during a regeneration run.  As a back up, in case of a failure
in the data logger system, most of these parameters and output data were con-
tinuously recorded on multi-point strip-chart recorders.  The data logger
output, 70 channels of information for the 7-1/2 hours duration of the run,
resulted in 315,000 pieces of experimental data-from this source alone.

      Eight discharges from the second cyclone lock hopper, and six discharges
from the bed solids lock hoppers were made during the run.  These 12
discharges provided a measure of the rates of fly ash generation and bed
solids which are needed in determining a mass balance around the unit.  Four
of the second cyclone discharges and two of the bed solids samples were later
selected as representative of the steady state conditions achieved and sub-
mitted for detailed chemical and particle size analysis.  During the run,
seven samples of the combustor bed composition were taken using the bed
sampling probe and submitted for chemical and particle size analyses.  Two
measurements of particulate concentration in the off gas from the second
stage cyclone were made during the run and the particulate catch submitted
for chemical and particle size analyses.  The initial and final combustor bed
was also submitted for chemical analysis.  These 15 solid samples, selected as
representative of the steady state operating conditions, were each analyzed
for six chemical components and the particle size distribution in 12 size
                                      183

-------
fractions was determined.  The particulate distribution in the off gas from
the second cyclone was determined in an additional 12 smaller size fractions.
Two wet chemistry measurements for S02 and SO^ were also made to check the
analytical instruments.

      A total of 106 individual analysis, 204 particulate size fraction deter-
minations, and 14 solid generation rates were obtained for Run No. 51, a
short duration run with only the combustor operational.  These determinations
do not include any replicate chemical analysis which frequently are required
to resolve uncertainties in the analysis.

DATA NEEDS

      This quantity of generated data is used in several ways to meet the data
needs of the program.  The first need is a rapid (soon after the run) method
to select that specific time interval within the run in which steady state
conditions were achieved.  This selection needs to consider all of the data
collected by the data logger system to determine when operating conditions
(bed temperatures, coal feed rates, etc) and output variables (S02 and NOX
emissions, excess oxygen levels, etc.) have reached steady values.  Once this
time period has been selected, a determination can be made as to which solid
samples need to be submitted for chemical and particle size analysis.

      The second need is experienced when an operational problem which forces
an abrupt shutdown of the unit occurs during the run or when the value of
the output variables is not that expected from past experience.  In these
instances the problem is to access and digest the necessary data over the
time interval desired when the information is part of a set of several hundred
thousand data points on the data logger tape.

      A third need is to enable a comprehensive data analysis of a run.  Of
the 70 points recorded by the data logger each minute during a run using
only the combustor, about 40 points determine the value of output variables
necessary for an analysis of the run.  The other 30 points are operational
variables used to determine what went wrong when problems develop or to
provide internal checks on the functioning of the data recording system.  The
average value of these 40 output variables over the steady state period
represent one set of inputs to the data analysis.  Of course, in determining
the average value of each variable, checks on the functioning of the record-
ing system must be made to determine if the value of the variable at a given
time should, or should not be included in the average.  The specification
of the chemical composition of the coal and sorbent used in the run and the
average generation rates, chemical analysis and particulate size distribu-
tions of the various output solid streams complete the data input needed for
the data analysis.  Based on this information, some 30 output parameters
need to be calculated to provide a complete description of the output vari-
ables for the input variables selected for the run.  A mass balance for
the major chemical elements (C, S, Ca and Mg) and for the solid inorganics
around the various input and output streams is also desirable.

      The fourth need is to search selectively all or several runs in order
to provide run summaries or to determine the relationships between output
and input variables.   In this context it is desirable to be able to place
the recovered data into files which can be subsequently operated on by
statistical or graphical routines.
                                     184

-------
DATA MANAGEMENT SYSTEMS

      In order to handle the data generated and to satisfy the data needs of
the program, several computer oriented data handling programs have been
developed to supplement the manual recording and storage of the data.  Certain
of these programs provide a capability to input data directly from the data
logging system while also permitting an operator later to enter analytical
and particulate size distribution data into the computer storage files.
Other programs provide the capability to access the information as needed.

      Three major computer systems have been developed to meet the needs of
this program.  They are described in the sections that follow under their
call names of Hoke, NCSS and I/O.

Hoke System

      The data logging system consists of a Digitrend data logger electron-
ically interfaced with a Kennedy magnetic tape recorder.  The Hoke computer
program consists of a subprogram to convert the instrument outputs recorded
by the data logger to physical  quantities such as temperatures, flow rates,
pressures, concentrations, etc.  These values are subsequently used in
another subprogram to calculate derived quantities such as superficial
velocity, expanded bed height,  excess air, etc.  The data are then averaged
over preselected time intervals (usually 10 minutes) tabulated, and printed.
If necessary, a complete data printout of an operation can be obtained within
two hours following the completion of the run.

      Figure D-l is an example  of the tabular form of data printout received
from a run.  This printout is then used in conjunction with the continuous
recorder charts to determine the steady state period.  When a steady state
period has been determined, the magnetic tape is rerun using the Hoke steady
state program (HOKSS).  This program averages selected values over the indi-
cated steady state period.  Another feature of the HOKSS program is that
it will delete any value which  does not fall into a fixed number of standard
deviations of the average values.  Faulty values may result from failure
of a measuring device, recording device, or a momentary upset in the system.
Consequently this value is not  representative of steady state and should
not be included in the average.  Finally, the average values from the HOKSS
program are inserted into the Input/Output System for record keeping as
well as further calculations.   The Input/Output system will be discussed
later in the text.

NCSS System

      The NCSS System consists  of programs to convert the instrument outputs
recorded by the data logger to  physical quantities and programs to search
the physical quantities selectively for a given run or series of runs.
Further data reduction to calculate derived quantities from the measured
parameters, e.g., superficial velocity calculated from flow rate, temperature
and pressure, is also provided.  In addition, the system includes auxiliary
informational and diagnostic programs.  This section discusses the NCSS
System.
                                      185

-------
                                                               FIGURE D-l
                       M1NIPLANT  ANALYSIS PROGRAM
                                                                                                            PAGE  1
00
01
 1     TIME
 2     SUPERFICIAL GAS VELOCITY - FTPS
 3     COMBUSTOR AIR FLOW - SCFM
 4     CA/S MOLAR FEED RATIO
 5     COAL FEED RATE - LB/HR
 6     TEMPERATURE GRADIENT- DEC F/FT.
 7     DELTA P (PT. 15)
 8     EXPANDED BED HEIGHT - IN.
 9     AVE. COMB. TEMP - DEG F
10     LOWER COMB. TEMP - DEG f
11     PERCENT EXCESS AIR FROM FLUE GAS COMP.
12     PERCENT EXCESS AIR FROM AIR/FUEL
13     COMBUSTOR PRESSURE-ATM.
           RUN
                                                                                              10
                                                                                                  11
12
13
64.0
64.0
64.0
64.0
64.0
64.0
64.0
64.0
64.0
64.0
64.0
10
10
10
10
11
11
11
11
11
11
12
20
30
40
50
0
10
20
30
40
50
0
4.6
4.6
4.6
4.6
4.7
4.7
4.7
4.6
4.7
4.8
4.8
528.9
526.6
531.8
536.6
538.1
538.5
538.4
532.9
534.4
538.3
536.7
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
197.2
212.5
213.3
212.5
199.2
199.2
209.6
204.4
195.5
202.2
205.1
6.8
7.1
5.6
5.0
5.0
5.0
5.8
6.4
5.1
4.9
4.7
102.2
90.7
91.7
103.9
114.4
126.5
137.6
202.2
212.8
263.8
275.0
158.0
142.7
141.5
156.1
162.0
180.1
189.2
279.5
305.1
368.5
367.0
1712.7
1711.8
1721.6
1713.6
1712.1
1709.3
1710.4
1704.1
1708.2
1704.3
1708.8
1719.1
1719.7
1727.0
1718.3
1716.9
1714.9
1717.4
1713.0
1714.6
1710.0
1713.3
30.3
30.6
26.3
31.1
29.5
28.4
26.9
27.5
23.1
27.3
25.6
6.8
1.3
0.6
0.5
7.5
7.6
2.3
3.8
8.8
6.0
4.2
8.4
8.4
8.4
8.5
8.4
8.4
8.4
8.4
8.3
8.2
8.1

-------
Data Conversion Programs—
      The information on the data logger tape is converted to a form that is
readable by the IBM 370 System and this tape (the "1130 tape") is processed
onto the master tape by a program package called PR1130.  The PR1130 program
package mounts the "1130 tape," calls a short Fortran program to mount the
master tape, specifies the auxiliary files to be used, species the Fortran
programs to be loaded, converts the instrument readings to physical
quantities, calculates derived quantities, and stores the quantities on
the master tape.  There are also a number of auxiliary files used in
creating the master tape, including:

      •  library of runs on master tape

      •  names of variables recorded on master tape

      •  cross-reference between variable names and data logger position
      •  library of calibration factors and thermocouples used in
          temperature averages

      The main program in the PR1130 package calls subroutines to create
the master tape and to establish the cross-reference file between the
variables and the data logger positions.  The user can ask to see the
existing cross-reference file, use an old file, create a new file or
edit an old file.  After the cross-reference file is established, the user
can specify any changes in the calibration factors or thermocouples used in
the temperature averages.  These factors are used by subroutines which reduce
the data from the data logger tape into physical quantities, such as, tem-
peratures, flow rates, pressures, etc., and also calculate derived quantities,
such as, superficial velocity and residence time.  These quantities are then
written onto the master tape.

Data Searching Programs—
      The primary search tool is a program package which allows the user to:

      •  examine the list of available runs

      •  select a run to be search
      •  select the number of minutes to be averaged
      •  choose whether standard deviations are to be calculated

      •  select the portion of the run to be search
      •  check the millivolt standard for each minute and discard data
          for minutes during which the millivolt standard is outside
          the allowable range

      •  examine the list of variable choices

      •  select the variables to be searched

One of the variables selected can be a "key variable."  When an upper bound
and lower bound are set for a key variable and that variable is outside the
given range, not only will the value for that variable be discarded but all
the values for all variables during that minute of data will be discarded.
Even though the user may select only one key variable,  in effect he has


                                      187

-------
three key variables at his disposal since both the time and the standard
millivolt checks also function as key variables.  A preselected series of
variables, including limits, can be requested by using a single "calling
number" through the use of macro files.  When one of these calling numbers
is used, a subroutine inserts the preselected list into the search specifica-
tions.  More than one macro at a time can be requested if desired.

        After the search specifications are established, the program:  finds
the correct position on the tape to start the search, locates the variables
according to the search specifications, calculates the averages and standard
deviations, and prints the data.  The data can be printed at the console, at
a local printer, or at a remote printer, as requested by the user.  As an
illustration of the type of output available from the data searching programs,
a selection of eight recorded and calculated quantities from a 15 minute
average of data from Run No. 37 is shown, with standard deviations, in Figure
D-2.  The data can also be filed for statistical analysis.  If it is, the
program writes the data into a data file specified by the user and tells
the user how many observations are filed for how many variables and also
which variables are filed.

Input-Output Programs

ARG1 Series of Programs—
     The ARG1 Series of programs are intended to create and support a data
bank established for the storage, calculation, and retrieval of selected
miniplant operating data.  The program are written in a conversational
mode.

     The program ARG1 "drives" the subroutines which file and edit the input
information, and which calculate and print selected input and calculated
parameters.

     The subroutine ARG1C creates the data bank.  This subroutine reads the
file which contains the properties of the coal and sorbent used in the run
and reads the data logger tape data for the run generated by the program
HOKSS, described earlier.  A singular case of the ARG1C subroutine is the
creation of a list of matrix addresses (known as "the road map") for the
variables.  The road map is used to edit existing data banks as additional
information becomes available.

     The data bank consists of an unformatted file (ARGF1) of 52 records with
a capacity of 1000 words in each record.  The data from a miniplant run is
assigned to an individual record in the file.  Each record contains 228
alphameric variables and 900 numeric variables.  Provisions are available in
the data bank for including data related to operation of the regenerator
and the granular bed filter, in addition to combustor operation data.

     The subroutine ARG1E edits an existing data bank.  Additional alphameric
and numerics may be added to the data bank as they become available.  Typical
examples are chemical and physical properties of the combustor and regenerator
beds, and additional coal and sorbent data.  Erroneous or obsolete informa-
tion may be replaced with new information using this subroutine.  A desk
console is used to add this information to the data file.
                                    188

-------
                Figure D-2.,  CSS — MINIPLANT ANALYSIS PROGRAMS  — FBCPRT

        RUN  NO.  37,  PFIZER DOLOMITE, CHAMPION COAL, HIGH VELOCITY
        4/7/76


 1      CONSTANT  MILLIVOLT STANDARD SOURCE (40 MV)
 2      SOLIDS  FLOW RATE  (KG/HR)
 3      RATIO LIMESTONE/COAL (KG/KG)
 4      COMBUSTOR COAL FEED RATE (KG/HR)
 5      COMBUSTOR CA/S MOLAR FEED RATIO
 6      CALCULATED SULFATION LEVEL
 7      S02  EMISSION INDEX (L8/MBTU)
 8      COMBUSTOR-N02 EMISSION INDEX (LB/MBTU)


 TIME          1         2345678


18   7     40.045   141.490    0.097  128.972    0.804    0.657    1.483    0.303
        (   0.104)1  8.961)1  O.OOOH  8.170X  O.OOOX  0.069)1   0.175)(   0.019)


18  22     40.049   128.349    0.097  116.993    0.804    0.708    1.354    0.332
        (   0.102X  9.521)1  O.OOOX  8.682X  0.000)1  0.061X   0.154X   0.025)


18  37     40.047   131.485    0.097  119.849    0.805    0.663    1.467    0.318
        (   0.103X  6.704X  0.000)1  6.111X  O.OOOX  0.046X   0.116)1   0.015)


18  52     40.048   134.397    0.097  122.505    0.804    0.811    1.093    0.312
        (   0.102X  8.059X  O.OOOX  7.344X  O.OOOX  0.054X   0.137)1   0.018)


19   7     40.045   126.333    0.097  115.158    0.804    0.761    1.220    0.324
        (   0.104)1  5.791)1  O.OOOX  5.280)1  0.000)1  0.074)1   0.187X   0.016)


19  22     40.040   132.157    0.097  120.469    0.804    0.710    1.594    0.298
        (   0.106)1  5.027X  O.OOOX  4.580X  O.OOOX  0.374X   0.078)1   0.012)


19  37     40.034   133.277    0.097  121.483    0.804    0.734    1.288    0.295
        (   0.025)1  4.595)1  0.000)1  4.186X  O.OOOX  0.038X   0.095X   0.009)


19  52     40.037   127.229    0.097  115.969    0.805    0.650    1.501    0.309
        (   0.100X  5.939X  O.OOOX  5.415X  0.000)1  0.064)1   0.163)1   0.014)
                                        189

-------
     The subroutine ARG1D is used to compute calculated values of operating
parameters.  Calculations are performed for some 30 output variables
required to characterize a given run, i.e., superficial velocity, excess air,
Ca/S feed ratio, coal feed rate, etc.  A calculation of a mass accountability
for carbon, sulfur, calcium, magnesium and solid inorganics over each of the
input and output streams and a mass balance for these components is also
performed.

     The subroutine ARG1P prints the contents of the data bank.  The user
has many print options available to him depending upon his needs, and the
amount of information available.  The printing format has been selected such
that missing data is printed as a series of asterisks.  An "XXXXXX" rotation
is used for those parameters (such as the sulfur content of the combustion
air) that are not obtained.  These conventions distinguish between those
parameters which are necessary and those which are never obtained.  Data may
be printed on a DS-40 terminal or on the IBM-1403 terminal.

     As an example of the use of the ARG1 series of programs,  a segment of
the computer print out obtained using the ARG1P subroutine is  shown in
Figures D-3 to D-5.  The first three figures are examples of experimental
data inputs to the file and the last two figures are calculated outputs.
Figure D-3 shows the average values of the operational parameters for Run
No. 50 and is illustrative of the data logger input to the data file provided
through the HOKSS program.  The values of the output variables required to
characterize Run No. 50, as calculated by subroutine ARG1D, is shown in
Figure D-4, while the mass accountability and mass balance for the same
run is shown in Figure D-5.

Ancillary Programs—
     The program ARGZ is used to retrieve data selected by the user from the
data file for subsequent statistical analysis or for out-filing the data in
an unformatted data file for other uses.

     The program ARG4 is used to summarize mass balances for selected mini-
plant runs that have been previously filed in the data bank.  The program's
output is presented in tabular form suitable for inclusion in  finished
reports.  An option to summarize the mass balances for all miniplant runs to
date is also provided.  This option is becoming more valuable  as the number
of miniplant runs on file in the data bank increases.

     The program ARG5 is used to summarize preselected parameters for
selected, or all, miniplant runs filed to date in a tabular form suitable
for inclusion in finished reports.  The preselected parameters cover opera-
ting conditions, flue gas emissions and overall performance.  An illustrative
example is shown in Figure D-6.
                                      190

-------
                                      FIGURE D-3
                                          EXXON MINIPLANT RUN NUMBER    O0.220
DESCRIPTION:
      COMPREHENSIVE ANALYSIS RUN V/BATT^LLE
RUN DATE      MONTH  3.      DAY 30.      YEAR 77.
START TIME     611. HOURS        DURATION    11.60 HOURS
PERIOD FOR THIS SUMMARY    1231. TO 1751.    DAY   30.
NUMBER 0? COOLIE COILS IN COV/!BUSTOR= ***
COAL:    FILE NO. *****    TYPE EASTERN         MINS CHAMPION
PARTICLE SIZE:    PLUS ***** *IESH, MINUS ***** MESH
SORBENT:    TILE NO. *****    TYPE DOLOMITE        SOURCE PFIZSR
PARTICLE SIZE:    PLUS ***** MESH,MINUS ***** MESH
                                       £CP  MO.1337
  COAL ANALYSIS
      MOISTURE
      ASH
      TOTAL CAREON
      HYDROG-iN
      SULFUR
      NITROGEN
      CHLORINE
      OXYGFN
      KEATING VALUT
      CALCIUM
      MASN3SIUM
      POTASSIUM
      SODIUM
    1.67 WPCT
   13.10 WPCT
   74.00 WPCT
    5 . ?, 0 WPCT
    1.85 WPCT
    1.53 WPCT
******** tfPcr
    2.40 WPCT
12973.    ETU/LE
         PFM
         PPM
******** PPM
£*£#***# pp^l
  SORBENT ANALYSIS (CALCINED^

      CAO
      MSO
      SI02
      AL303
      FE203
      POTASSIUM
      SODIUM
   S4 .00 WPCT
   44.00 'VPCT
    0.90 WFCT
     .20 WPCT
     0.30 WPCT
******** PPM
                                             191

-------
                                         FIGURE D-4
                                                        MASS BALANCES fWPCT)
RUN NUMBER
                   48.02!
                             50.11
                                       5 .18
                                                 50.20
50.42
51.(
TOTAL MASS         99.93    100.32     99.90     99.81     99.14     99.55     99.73     99.71
SULFUR             94.75     52.63    131.71    106.24     58.42     81.15     88.10     73.37
CARBON            104.59    109.45    104.62     87.90     B9.G0     93.83     93.32    137.51
CALCIUM           108.78     15.42    149.63    104.£5     57.73     31.66     93.51     63.99
REACTIVE OXlfGFN   100.71    102.21    102.96     92.10     97.65     98.27     99.35    105.12

MAGNESIUM          90.19     10.63     9 .66     84.31     54.36     70.94     79.69     50.63
BB3               ******    ******    ******    ******    ******    ******    **-.):#*<(    **#«**
SOLID INORGANICS  101.66    163.93     96.54    101.96     72.52     97.32    104.30     94.23
                                               192

-------
                                                       FIGURE D-5
                                         MINIPLANT  FLUIDIZED  BED  COAL  COMBUSTION  RUN
Lo
    RUN  NUMBER
OPERATING CONDITIONS:

 RUN LENGTti(HRS)
 PRESSURE(KPA)
 LOWER BED TEMP(CELSIu'S )
 AVERAGE BED TEMP(CELSIUS)
 SUPERFICIAL V?LOCITY(M/SEC;
 SETTLED BED HEIGHT(M)

  INITIAL
  FINAL
 EXPANDED BSD HSIGHT(M)
 COAL FEED RATEUG/HR)
 CA/S MOLAR FEED RATE-SET
 CA/S .MOLAR FEED RATE-CALC
 EXCESS AIR(VPCI)
 SORBENT
 COAL
    FLUE GAS EMISSIONS
                                      43.00    50.11    5 .12    50.20    50.30    50.40    50.5?
     SOZ(PPM)
     NOX(PPM)
     CO(PPM)
     C02(VPCT)
     02(VPCT)
    RESULTS:

     RBTiiNTION(WFCT)
     SULFATION(WPCT)
     S02(LBS/METU)
     NOX(LESXMBTU)
     COMBUSTION EFFICIENGT
8.25
932.
934.
874.
1 .4?
1.^?
2.15
3.37
90. 96
1.3?
0.91
39.1?
DOLO
iAST
510.
92.
114.
13.29
5.99
55.78
61.45
1.13
0.10
93.25
9.50
916.
901.
891.
1.19
•1.58
******
3.2)2
74.30
1.34
******
34.59
DOLO
iAST
373.
97.
95.
13.84
5.43
68.10
******
0.91
0.12
99.33
12.00
919.
395 .
688.
1 .21
******
******
3.93
74.54
1.34
1.29
35.91
DOLO
iAST
355.
104.
93.
13.75
5.54
70.23
54.93
0.32
.12
99.15
11.50
901 .
814.
903.
1 .22
******
******
3.19
70 .86
1 .34
1 .63
55.05
DOLD
SASP
151.
138.
131.
3 . 92
7.55
85.53
52 .58
0 .41
0.19
93.93
15.20
901 .
95"0.
885.
1.32
******
******
3.22
32.94
1.34
1.48
40.97
DOLO
SAST
62.
129.
90.
12.06
6.18
93.46
63.03
0. 14
0.15
99.89
12.50
900.
902.
399.
1.33
******
if if * If. * If
3.31
83.43
1.43
1.65
39.01
DOLO
EAST
41.
127.
55.
12.50
5.93
95.81
57.83
0.03
0.14
99.42
18.03
902.
Q?C _
89S.
1.34
******
2.16
3.29
85.20
'1.49
1.45
34. 90
DOLO
E*5I
29.
119.
53.
13.10
5.51
97.02
66.89
0. 06
0.13
99.36
3.25
930.
375.
376.
1.65
1.93
******
2.32
78.78
1.69
1.31
44.76
DOLO
FAST
253.
143.
125.
13.30
6.53
59.42
52.91
.74
0.21
98.19

-------
                                        FIGURE  D-6
                            SUMMARY  OF  SELECTED  PARAMETERS.
 RUN NUMBER                          48-00        52.11        50.12        50.20
 PARAMETERS
 COME  EFF  (PCT)          694        99.25       99.33       99.15        98.93       99.89
 NOX  (PPM)               440        107.65       113.50       115.40       130-09      141.35
 CO  (PPM)                441        133.39       111.50       108.58       171.66       93.60
 S02  ;LB/MBTU)           444         1.13        0.91         0.32         0.41        0.14
 NOX  (LB/MBTU)           445         0.10        0.13         0.12         0.19        0.15
 SULF  REM  EFF  (PCT)      447        55.78       68-10       72.23        85.63       93.46
 COAL  FEED  (KG/HR)       464        90.36       74.30       74.54        70.86       82.94
 SORBENT  FEEO  (K3/HR)    695        15.35       11.38       11.42        10.86       12.71
 RUN  NUMBER                          50.40       50.50        51.
^PARAMETERS
 COMB  FFF  (PCT^          694        99.42       99.86       93.19
 NOX  (PPM,               440       141.15      126.27       210.25
 CO  (PPM)                441        60.72       55.82       183.80
 S02  (LB/MBTU)           444         0.09        0.06         2.74
 NOX  (L£/MBTU)           445         2.14        0.13         0.21
 SUIF  REM  EFF  (PCT)      447        95.Bl       97.0a       69.42
 COAL  FEED  (rCS/HR)       464        B3.4B       86.20       78.78
 SORBENT FEED  (K&/HR)    695        12.83       13.21       13.71
                                            194

-------
     The program ARG6 is used to summarize selected parameters  from selected,
or all, miniplant runs whose data has been filed in the data bank.   The
results are presented in tabular form, suitable for inclusion in finished
reports.  Unlike the two preceding programs, the user may select the para-
meters that he wishes to review.  The program also permits the identification
of these parameters  (by the user) with a 20 character description.
                                      195

-------
                                 APPENDIX E

                            ANALYTICAL TECHNIQUES
Analysis of Solids
     Solids from combustion runs were analyzed for SO/"2, C0.,~2, Ca
Na+, carbon and total sulfur.  --     -   -   -
described below.
                                                                   +2
                                                                Mg+2,
                        The analytical techniques that were used are
     SO
     CO,
        -2
        -2
        The sample was  treated with acidic BaCl2 solution.
        The BaSO^ precipitate was weighed.

        HC1 was added to an acidified sample.  The solution was
        stripped with N2 and the gas passed through drierite,
              and ascarite.
                                       "  was determined from the
     Ca
       +2
     Mg
     Na
+2
'+
     Total  -
     Sulfur
weight gain of the ascarite.

The sample was digested by heating vigorously in a
medium of perchloric acid/nitric acid.  The determination
of Ca, Mg and Na was made by atomic absorption.

(Dietert Sulfur Method) - The sample is combusted in an
oxygen atmosphere at 1250°C.  The 502-803 products in
the effluent gas were analyzed by an automatic Leco
titrator .
     Total   -   (Carbon on Catalyst Method)  -  The  sample  is combusted
     Carbon     in  an oxygen atmosphere  at  1200°C.  The C02 evolved was
                determined from the weight  gain  of ascarite.

Analysis of  Flue Gas
by Wet Chemical Methods

     SO-     -   The amount absorbed by an 80%  isopropanol solution was
                determined titrimetrically  using 0.01N barium per-
                chlorate as the titrant  and thorin as the indicator.

     SO      -   The amount absorbed by a 3% hydrogen peroxide solution
                was determined  titrimetrically using 0.01N sodium
                hydroxide as the titrant and methyl orange as the
                indicator.
                                      196

-------
            APPENDIX F
DATA LOGGER CHANNEL IDENTIFICATION
Ft.
No.
01
02
03
04
05
06
07
08
09
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
Loop
No.
1-0
21
12
13
14
15
16
18
1-9
1-10
10
11
36
34
35
40
44
32
22
23
24
25
26
29
29
30
33
20
1-29
1-30
1-31
1-32
1-33
1-34
1-35
1-36
1-37
1-38
1-39
1-40
1-41
1-42
1-43
1-44
1-45
1-46
1-47
1-48
Identification
Constant Millivolt Std. Signal (40 MV)
Coal, Weight in Primary Injector
Combustor, Water Flow Rate - Coil #1A
Combustor, Water Flow Rate - Coil #1B
Combustor, Water Flow Rate - Coil #2A
Combustor, Water Flow Rate - Coil #2B
Combustor, Water Flow Rate - Coil #3A
Combustor, Water Flow Rate - Coil #3B
Batch Unit, Analyzer, NOX (10,000 ppm)
Batch Unit, Analyzer, S02 (10,000 ppm)
Combustor, Flow Rate - Main Air
Combustor, Pressure
Combustor, AP - Coal Feed Vessel
Combustor, AP - Fluidizing Grid
Combustor, AP - Bed (Port 4 to 31)
Combustor, AP - Bed (Port 4 to 11)
Combustor, Water Flow Rate - Heat Transfer Loop
Spare
Regenerator, Flow Rate - Burner Air
Regenerator, Flow Rate - Burner Fuel
Regenerator, Flow Rate - Supplemental Air
Regenerator, Flow Rate - Supplemental Fuel
Regenerator, Pressure (or AP to Combustor)
Regenerator, AP - Fluidizing Grid
Regenerator, AP - Bed (Port 29 to 34)
Regenerator, AP - Bed (Port 29 to 31)
Limestone to Coal Weight Ratio
Main Air Pressure at Measuring Orifice
Run Identification Number
Combustor, Temperature - Burner Grid Metal
Combustor, Temperature - Burner Grid Cooling Water
Combustor, Temperature - Fluidizing Grid Metal
Combustor, Temperature - Fluidizing Grid Cooling Water
Combustor, Temperature - Port #3 (Burner Zone)
Combustor, Temperature - Port #5 (6")
Combustor, Temperature - Port #7 (18")
Combustor, Temperature - Port #8 (27")
Combustor, Temperature - Port #9 (48")
Combustor, Temperature - Port #12 (64")
Combustor, Temperature - Port #13 (85")
Combustor, Temperature - Port #14 (94")
Combustor, Temperature - Port #16 (103")
Combustor, Temperature - Port #20 (140")
Combustor, Temperature - Port #22 (169")
Combustor, Temperature - Port #26 (207")
Combustor, Temperature - Port #28 (252")
Combustor, Temperature - Port #32 (354")
Combustor, Temperature - Coal Injector Line
                197

-------
Pt.
No.

49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
                               Identification
1-49
1-50
1-51
1-52
1-53
1-54
1-55
1-56
1-57
1-58
1-59
1-60
1-61
1-62
1-63
1-64
1-65
1-66
1-67
1-68
1-69
1-70
1-71
1-72
1-73
1-74
1-75
1-76
1-77
1-78
1-79
1-80
1-81
1-82
1-83
1-84
1-85
1-86
1-87
1-88
1-89
1-90
1-91
1-92
1-93
1-94
1-95
1-96
1-97
1-98
           Temperature
           Temperature
           Temperature
           Temperature
           Temperature
           Temperature
           Temperature
                         Cooling Water to Coils
                         Cooling Water from #1A
                         Cooling Water from #1B
                         Cooling Water from #2A
                         Cooling Water from #2B
                         Cooling Water from #3A
                         Cooling Water from #3B
           Analyzer Range Switch - 862
Combustor,
Combustor,
Combustor,
Combustor,
Combustor,
Combustor,
Combustor,
Combustor,
Combustor, Analyzer Range Switch - NOX
Combustor, Analyzer Range Switch - CO
Combustor, Analyzer Range Switch - Q£
Combustor, Temperature - 1st Cyclone Gas Discharge
                         2nd Cyclone Gas Discharge
                         Off-Gas Upstream of Nozzle
                         Off-Gas Downstream of Nozzle
                         1st Cyclone Dip Leg
                         2nd Cyclone Dip Leg
                         Solids Reject Line (Before Pulse Pot)
                         Solids Reject Lock Hopper
                         Fly Ash Lock Hopper
                         Surface - Lower Deck
                         Sullair to Measuring Orifice
Combustor,
Combustor,
Combustor,
Combustor,
Combustor,
Combustor, Temperature -
Combustor, Temperature -
Combustor, Temperature -
Combustor, Temperature -
           Temperature -
           Temperature -
           Temperature -
           Temperature -
           Temperature -
Combustor, Temperature -
Regenerator, Temperature
Regenerator, Temperature
Regenerator, Temperature
Regenerator, Temperature
Regenerator, Temperature
Regenerator, Temperature
Regenerator, Temperature
Regenerator, Temperature
Regenerator, Temperature
Regenerator, Temperature
Regenerator, Temperature
Regenerator, Temperature
Regenerator, Temperature
Regenerator, Temperature
Regenerator, Temperature
Regenerator, Temperature
Regenerator, Temperature
Regenerator, Temperature
Regenerator, Temperature
Combustor, Analyzer - SO
Combustor, Analyzer
Combustor, Analyzer
Combustor, Analyzer
                           Port #2
                           Port #3
                           Port #4
                           Port #5
                           Port #6
                           Port #7
                           Port #9
                           Port #10
                           Port #11
                           Port #12
                           Port #13
                           Port #14
                           Port #16
                           Port #18
                           Port #21
                           Port #24
                           Port #26
                           Cyclone Gas Discharge
                           Off-Gas from Cooler
                        2 (3,000 ppm)
                      NOX (10,000 ppm)
                      02 (25%)
                      C02 (25%)
                      CO (6250 ppm)
Combustor, Analyzer
Regenerator, Analyzer - S02 (15%)
Batch Unit, Analyzer - 02 (25%)
Batch Unit, Analyzer - C02 (25%)
Batch Unit, Analyzer - CO (5000 ppm)
                                     198

-------
13

14

15
                                                 APPENDIX G
                                      FBC MINIPLANT ALARM/SHUTDOWN SYSTEM
Alarm
1
2
3
4
5
6
i
i
8
9
10
11
12

Unit
Comb
Comb
Comb
Comb
Comb
Comb
Comb
Comb

Comb
Comb
System
Description
Fluidizing Air
Fluidizing Air
Cooling Coil
Water Loop
Cooling Water to
Each Coil (5)
Condenser Cooling Water
Coal Injection
1st Stage Cyclone
Dipleg
2nd Stage Cyclone
Dipleg
Cooling Tower Water
Reactor Vessel
Fluidizing Grid
Alarm Condition
Low Flow
High Flow
Low Total Flow
Low Flow
Low Flow
Low Flow
Low Flow
Low Flow
Low Pressure
High Pressure
High AP
Comb   Fluidized Bed
Comb
Coal Injection
Vessel
Demineralized Cooling
Water Reservoir
High AP or
Low AP
Low AP
                                 Low Level
                                                                                                Action
Flow Switch

Orifice AP Cells
Flow Switch
Ultrasonic Microphone

Thermocouple

Thermocouple

Pressure Cell
AP Cell
AP Cell-
*TD Before Alarm
*Close Coal Block Valve
and Valve AV10
Thermowatch
Capacitance Switch
                                                                                                4 min
                                                                                                   X
                                                                                                          X
                                                                                          /*
                                                                                         0*

-------
                         FBC MINIPLANT ALARM/SHUTDOWN SYSTEM  (Continued)
                                                                                         Action





NJ
O
O
Alarm
#
16
17
18
19
20
21

Unit
Comb
Comb
Comb
Comb
Comb
Comb
System
Description
Natural Gas
Flame
Burner Grid
Fluidizing Grid
Bed Temperature
at 3 Points
Bed Temperature
Upper Bed
Temperature
r*/-*n1 -I i-irr TJo-f-QV

Alarm Condition
Flame Out
High Temperature
High Temperature
High Temperature
Low Temperature
High, High Temp.
Sensor /Function
UV Sensor
*Fuel Shutdown
Thermocouple
Thermocouple
Thermocouple
Thermocouple
Thermocouple
Alarm Time Shut-
Only Delay Down
X*
^
/
/
/

22

23

24

25
26
27
28
29
30
Comb
Comb
From Each Coil
Coal Injection
Probe
Comb   Coal Injection Line
Comb
Comb
Comb
Comb
Comb
Coal Injection Ready
Off Gas from Cooler
Combustor to Regenerator
Transfer Line
Fly Ash Lock Hopper
Reactor Shell Surface
Comb   Coal  Injector Vessel
High Temperature

High Temperature

High Temperature

Low Combustor Temp,
High Temperature
Low Temperature
High Temperature
High Temperature
High AP
Thermocouple

Thermocouple
Thermocouple
*Close Coal Block Valve
Thermocouple
*Close Valve AV10
Thermocouple
Thermocouple
Thermocouple
Thermocouple
AP Cell
*Close Valve AV10
                                                                                         4 min
                                                                                    0*

                                                                                    0*
                                                                                            0*

-------
                                FBC MINIPLANT ALARM/SHUTDOWN SYSTEM (Continued)
                                                                                                Action
Alarm
#
33
34
35
37
38
39
40
41
43
44
45
46

Unit



Reg.
Reg.
Reg.
Reg.
Reg.

Reg.
Reg.
Reg.
System
Description
Main Air Compressor
Aux. Air Compressor
Site Air
Burner Air
Burner Fuel
Supplemental Air
Regenerator/
Reg . to Comb .
Burner Air
Nitrogen Compressor
Reactor
Supplemental Fuel
Fluidizing Grid
Alarm Condition
Low Pressure
Low Pressure
Low Pressure
Low Flow
High or Low Flow
High or Low Flow
High Pressure/
High Pressure Drop
High Flow
Low Pressure
High Pressure
High or Low Flow
High AP
Sensor /Function
Pressure Switch
Pressure Switch
Pressure Switch
Orifice AP Cell
Orifice AP Cell
Orifice AP Cell
Pressure Cell/
AP Cell
Orifice AP Cell
Pressure Switch
Pressure Cell

AP Cell
Alarm Time Shut-
Only Delay Down
/
/
^
0 X
S
S
'
s
S
o /

/
47
Reg.   Fluidized Bed
                                 High AP or Low AP
                      AP Cell
                      *T.D.  Before Alarm
                                                                                          /*
49
Reg.   Natural Gas Flame
50     Reg.   Burner Grid
Flame Out
                                 High Temperature
                                                       U.V.  Sensor

                                                       Thermocouple
10 sec
  or
 5 min

-------
FBC MINIPLANT ALARM/SHUTDOWN SYSTEM (Continued)
                                                                Action
Alarm
#
51
52
53
54
55
K3
O
10 57
58

Unit
Reg.
Reg.
Reg.
Reg.
Reg.
Reg.
Reg.
System
Description
Fluidizing Grid
Bed Temperature
At 2 Points
Bed Temperature
At 2 Points
Bed Temperature
Upper 2 Points
Off Gas from
Cooler
Supplemental Fuel
Injection Ready
Reactor Shell

Alarm Condition
High Temperature
High Temperature
High Temperature
High, High Temp.
High Temperature
Low Temperature
High Temperature
Alarm Time Shut-
Sensor/Function Only Delay Down
Thermocouple /
Thermocouple /
Thermocouple /
Thermocouple /
Thermocouple /
Thermocouple /
Thermocouple V

-------
                        APPENDIX H-l.  MINIPLANT FLUIDIZED BED COAL COMBUSTION RUN SUMMARY
to
o
(jO
   Operating Conditions;
Run Length, hrs.
Pressure, kPa
Air Flow Rate, m3/min
Avg. Bed Temperature, °C
Sup. Vel., m/s
Settled Bed Height, m
  Initial
  Final
Expanded Bed Height, m
Coal Feed Rate, kg/hr
Ca/S Molar Feed Ratio-Set
Ca/S Molar Feed Ratio-Calc
Excess Air, %
Sorbent
Coal

Flue Gas Emission:
S02, ppm
NOX, ppm
CO,
C02,
         ppm
    o
     2,
    Results;

    Retention, %
    Sulfation, %
    Lb S02/M BTU
    Lb NOX/M BTU
                                   41.1(7/26-27/76)*
                                          20.5
                                          912
                                          19.6
                                          889
                                          1.6
                                          2.3
 86
0.84

 44
 GL
 CH
                                      350
                                      111
                                      116
                                       14
                                        6
                                       68

                                      0.90
                                      0.20
               41.2(7/27/76)*
                      5
                    912
                    19.4
                    875
                    1.5
1.9-2.2

  69
 0.84
 3.92
  37
  GL
  AR
                      0
                    127
                    133
                     12
                      6
                    100
                     26
                    0.00
                    0.30
               43.1(8/24/76)
                    10
                    840
                    18.4
                    940
                    1.7
1.4

3.9
 98
1.25

 24
 PD
 IL
                     60
                     80
                    180
                     14
                      4
                     98

                    0.14
                    0.13
             43.2(8/24-25/76)
                     7
                   880
                   22.9
                   845
                   1.8
                                                                                                2.1
                                                                                                5.7
                                                                                                 88
                                                                                                0.75
                                                                                                1.23
                                                                                                 42
                                                                                                 PD
                                                                                                 IL
                   720
                   120
                   250
                   9-12
                     6
                    60
                    49
                   2.58
                   0.31
    *  Combined Combustor/Regenerator Runs; Combustor Results Only.

    CH = Champion Coal           PD = Pfizer Dolomite (BCR No. 1337)
    AR = Arkwright Coal          GL = Grove Limestone (BCR No. 1359)
                                                                 Run Discussed in Reference 1.

                                                                          IL = Illinois Coal No.  6

-------
             APPENDIX H-l (CONTINUED).   MINIPLANT FLUIDIZED  BED  COAL COMBUSTION RUN SUMMARY


   Operating Conditions;       43.3(8/25/76)     43.4(9/2/76)     43.5(9/2/76)     45-1(10/11-12/76)*

Run Length, hrs.                      6              6.3              4.3                  24
Pressure, kPa                       885              940              940                 780
Air Flow Rate, m3/min               20.8             21.1             21.2                14.2
Avg. Bed Temperature,  °C            940              940              945                 880
Sup. Vel., m/s                      1.8              1.7              1.7                 1.4
Settled Bed Height, m
   Initial                           1.6              1.6               —                 2.3
   Final                             1.9               —              2.2
Expanded Bed Height, m             3.1              3.6              3.7                 3.2
Coal Feed Rate, kg/hr                93               96               95                  82
Ca/S Molar Feed Ratio-Set           0.75             1.25             0.50                0.74
Ca/S Molar Feed Ratio-Calc          1.44             1.80             0.68
Excess Air, %                        25               25               25                  37
Sorbent                              PD               PD               PD                  GL
Coal                                 IL               IL               IL                  CH

Flue Gas Emissions;
S02, ppm                            870              114              1328                493
NOX, ppm                             90               79               75
CO,  ppm                            240              201              201
COo, %                               12               14               14
02,  %                                444                   6

Results:
Retention, %                         63               96               44                  65
Sulfation, %                         45               53               65
Lb S02/M BTU                        2.39             0.26             3.62                1.04
Lb NOX/M BTU                        0.18             0.13             0.15
*  Combined Combustor/Regenerator Runs; Combustor Results Only.
PD  =  Pfizer Dolomite (BCR No. 1337)              CH  =  Champion Coal
GL  =  Grove Limestone (BCR No. 1359)              IL  =  Illinois Coal No.

-------
           APPENDIX H-l  (CONTINUED).  MINIPLANT FLUIDIZED BED COAL COMBUSTION RUN  SUMMARY
N>
O
Cn
   Operating Conditions;

Run Length, hrs.
Pressure, kPa
Air Flow Rate, m^/min
Avg. Bed Temperature, °C
Sup. Vel., m/s
Settled Bed Height, m
  Initial
  Final
Expanded Bed Height, m
Coal Feed Rate, kg/hr
Ca/S Molar Feed Ratio-Set
Ca/S Molar Feed Ratio-Gale
Excess Air,  %
Sorbent
Coal

Flue  Gas Emission;

SO,
                             45-2(10/12-13/76)*   45-3(10/13-14/76)*   45-4(10/14/76)*   45-5(10/14-15/76)*
NO
         ppm
      'x> PPm
    CO,  ppm
    C02, %
    02,  %

    Results;

    Retention, %
    Sulfation, %
    Lb S02/M BTU
    Lb NOX/M BTU
                                     30
                                     780
                                     14.2
                                     885
                                     1.4
                                     2.8
                                      99
                                     0.74

                                      40
                                      GL
                                      CH
                                     107
                                      92

                                     0.23
                                                   14
                                                  780
                                                  14.2
                                                  890
                                                  1.4
                                                  2.9
                                                   81
                                                  1.30

                                                   33
                                                   GL
                                                   CH
                                                  217
                                                   84

                                                  0.46
 12
780
14.2
870
1.4
3.4
 83
1.06

 30
 GL
 CH
249
 82

0.52
  28
 780
 14.2
 875
 1.4
3.6
 85
0.00

 28
 GL
 CH
635
 56

1.29
*  Combined Combustor/Regenerator Runs; Combustor Results Only.
GL  =  Grove Limestone (BCR No. 1359)
CH
Champion Coal

-------
                 APPENDIX H-l  (CONTINUED).  MINIPLANT BED COAL COMBUSTION RUN SUMMARY
K>
O
O\
   Operating Conditions;

Run Length, hrs.
Pressure, kPa
Air Flow Rate, m3/min
Avg. Bed Temperature,  °C
Sup. Vel., m/s
Settled Bed Height, m
  Initial
  Final
Expanded Bed Height, m
Coal Feed Rate, kg/hr
Ca/S Molar Feed Ratio-Set
Ca/S Molar Feed Ratio-Calc
Excess Air, %
Sorbent
Coal

Flue Gas Emission;

S02, ppm
NOX, ppm
CO,  ppm
C02, %
02,  %

Results;

Retention, %
Sulfation, %
Lb S02/M BTU
Lb NOX/M BTU
                                     45-6(10/14-15/76)*    46.1(11/18/76)
                                            12
                                            780
                                            14.2
                                            875
                                            1.4
 2.2
 3.6
 85
0.00
1.69
 27
 GL
 CH
                                           555
                                            62
                                            37
                                           1.13
3.67
940
15.0
920
1.3

1.8

 3.0
 81
2.50

 37
GL(P)
 CH
                    130
                     80
                    100
                     13
                      6
                     90

                    0.29
                    0.13
                                  46.2(11/18/76)

                                       4.5
                                       940
                                       15.0
                                       865
                                       1.2
>2.3
 3.3
 78
2.50
1.49
 47
GL(P)
 CH
                   170
                   120
                   150
                    11
                     7
                    86
                    58
                   0.40
                   0.20
              46.3(11/22/76)
                   7.67
                   940
                   16.7
                   760
                   1.0
>2.3

 2.7
 65
2.50
1.01
 40
GL(P)
 CH
                   416
                   120
                   180
                    10
                     6
                    55
                    55
                   1.31
                   0.27
     *  Combined Combustor/Regenerator Runs; Combustor Results Only.
     GL = Grove Limestone  (BCR No. 1359)
     GL(P) « Grove Limestone (Precalcined)
     CE = Champion Coal

-------
                     APPENDIX H-l  (CONTINUED).  MINIPLANT FLUIDIZED BED COAL COMBUSTION  SUMMARY
Ni
O
   Operating Conditions;

Run Length, hrs.
Pressure, kPa
Air Flow Rate, m3/min
Avg. Bed Temperature, °C
Sup. Vel., m/s
Settled Bed Height, m
  Initial
  Final
Expanded Bed Height, m
Coal Feed Rate, kg/hr
Ca/S Molar Feed Ratio-Set
Ca/S Molar Feed Ratio-Calc
Excess Air, %
Sorbent
Coal

Flue Gas Emission:

S02, ppm
         NOX,
         CO,
         co2,
         o2,
      ppm
      ppm
         Results:
         Retention, %
         Sulfation, %
         Lb S02/M BTU
         Lb NOX/M BTU
                                        46.4(11/22/76)
                                             3.5
                                             940
                                             16.7
                                             865
                                             1.4
2.0
3.1
 81
1.50
1.14
 35
GL(P)
 CH
425
150
100
 11
  5
                                     64
                                     56
                                    1.07
                                    0.27
47(12/9/76)

    3.5
    1075
    30.0
    885
    2.2

    2.1
     2.3
    3.8
    179
    1.50
    0.06
     18
     GL
     CH
    1400
    100
    175
     14
      3
                  2.7
                   49
                  2.86
                  0.15
48(2/24/77)
      8
    930
    17.9
    875
    1.5

    1.5
    2.2
    3.4
     91
    1.40
    1.32
     39
     PD
     CH
    510
     92
    114
     13
     6
                     58
                     44
                    1.16
                    0.16
50.1A(3/29/77)
     9.5
     910
     13.9
     890
     1.2

     1.6

     3.0
      75
     1.25

      41
      PD
      CH/K
    373
     97
     94
     14
      5
                     69

                    0.87
                    0.16
         GL = Grove Limestone  (BCR No. 1359)
         GL(P) = Grove Limestone (Precalcined)
         PD = Pfizer Dolomite  (BCR No. 1337)
         CH = Champion Coal
         CH/K = Mixture of Champion Coal (2% Sulfur) and Kentucky Coal (0.6% Sulfur)

-------
                   APPENDIX  H-l  (CONTINUED).  MINIPLANT FLUIDIZED BED COAL COMBUSTION RUN SUMMARY
NJ
O
oo
   Operating Conditions;

Run Length, hrs.
Pressure, kPa
Air Flow Rate, m^/min
Avg. Bed Temperature, °C
Sup. Vel., m/s
Settled Bed Height, m
  Initial
  Final
Expanded Bed Height, m
Coal Feed Rate, kg/hr
Ca/S Molar Feed Ratio-Set
Ca/S Molar Feed Ratio-Calc
Excess Air, %
Sorbent
Coal
                                    50.1B(3/29-30/77)    50.2(3/30/77)    50.3(3/31/77)    50.4(3/31/77)
                                            12
                                           915
                                           14.3
                                           890
                                           1.2
3.9
 75
1.25
1.52
 45
 PD
 CH/K
                   11.6
                   900
                   15.3
                   805
                   1.2
3.2
 71
1.25
1.95
 63
 PD
 CH/K
                  15
                 900
                 15.3
                 890
                 1.3
3.2
 83
1.25
2.35
 40
 PD
 CH/K
                 12.5
                 900
                 15.4
                 890
                 1.3
3.3
 84
1.25
2.44
 40
 PD
 CH/K
      Flue Gas Emission;
NOX,
CO ,
C02,
02,
           ppm
           ppm
           ppm
      Results;

      Retention, %
      Sulfation, %
      Lb S02/M BTU
      Lb NOX/M BTU
355
102
100
 14
  6
                                      70
                                      46
                                     0.85
                                     0.18
151
136
130
 10
  8
                    86
                    44
                   0.41
                   0.26
 62
128
 90
 12
  6
                  95
                  41
                 0.14
                 0.21
 41
124
 56
 12
  6
                  96
                  40
                 0.09
                 0.20
      PD = Pfizer Dolomite  (BCR No.  1337)
      CH/K = Mixture, of  Champion  Coal  (2%  Sulfur)  and Kentucky Coal  (0.6%  Sulfur)

-------
           APPENDIX H-l  (CONTINUED).  MINIPLANT FLUIDIZED BED COAL COMBUSTION RUN  SUMMARY
   Operating Conditions;

Run Length, hrs.
Pressure, kPa
Air Flow Rate,
Avg. Bed Temperature, °C
Sup. Vel., m/s
Settled Bed Height, m
  Initial
  Final
Expanded Bed Height, m
Coal Feed Rate, kg/hr
Ca/S Molar Feed Ratio-Set
Ca/S Molar Feed Ratio-Calc
Excess Air, %
Sorbent
Coal

Flue Gas Emission;
S02,
NOX,
CO,
co2,
o2,
ppm
ppm
ppm
 Results;
 Retention,  %
 Sulfation,  %
 Lb S02/M BTU
 Lb NOX/M BTU
50.5(3/31-4/1/77)
        18
       900
       15.5
       895
       1.4
       2.2
       3.3
        86
       1.25
       2.00
        36
        PD
        CH/K
                                        29
                                       119
                                        53
                                        13
                                         6
        97
        49
       0.06
       0.19
51(5/6/77)
     8
   932
   18.3
   871
   1.5

   1.9

   2.3
    79
   1.40
   1.60
    46
    PD
    CH/K
                          253
                          143
                          125
                           14
                            7
    71
    45
   0.72
   0.29
52(5/24/77)

     23
    930
    17.8
    890
    1.5

    1.7
    1.6
  2.7-3.5
   75-102
    1.25

   26-50
     PD
     CH/K
                  115-710
                   90-125
                   80-125
                     14
                      6
     53
54(6/29/77)
    5.5
    595
    14.6
    910
    1.9
    2.7
     81
    0.00

     29
    None
     CH/K
                     660
    31
    52
   1.41
 PD = Pfizer Dolomite  (BCR No.  1337)
 CH/K = Mixture  of Champion Coal  (2% Sulfur) and Kentucky Coal (0.6% Sulfur)

-------
            APPENDIX H-l  (CONTINUED).  MINIPLANT FLUIDIZED BED COAL COMBUSTION RUN SUMMARY
   Operating Conditions;

Run Length, hrs.
Pressure, kPa
Air Flow Rate, m3/min
Avg. Bed Temperature, °C
Sup. Vel., m/s
Settled Bed Height, m
  Initial
  Final
Expanded Bed Height, m
Coal Feed Rate, kg/hr
Ca/S Molar Feed Ratio-Set
Ca/S Molar Feed Ratio-Calc
Excess Air, %
Sorbent
Coal

Flue Gas Emission:
                         55(7/14/77)    56(7/26/77)     57(8/2/77)     59.1(8/11/77)     59.2(8/12/77)
S02,
NOX,
CO,
CO-,,
ppm
ppm
ppm
02,  %
Results;

Retention, %
Sulfation, %
Lb S02/M BTU
Lb NOX/M BTU
                             3.5
                             570
                             12.5
                             930
                             2.1
                             2.5
                              83
                             0.00
                             0.37
                              13
                             None
                              CH/K
792
110
2'13
 15
  5
                              30
                              82
                             1.57
                             0.16
                 4
               800
               14.8
               815
               1.4

               1.8
 62
104
 11
  8
  2
                93
                89
               0.16
               0.20
                6
              810
              13.7
              935
              1.4
—
2.6
73
0.69
1.05
34
PD
CH/K
1.6
3.0
91
0.75
1.07
14*
PD
CH/K
150
 35
145
 13
  3
               87
               81
              0.29
              0.05
                10.5
                890
                17.7
                956
                1.6
                                             3.7
                                             105
                                             0.75

                                              33
                                              PD
                                              CH/K
120
 78
156
                 88

                0.58
                0.27
                  12
                 873
                 17.7
                 949
                 1.7
                                               1.3
                                               7.2
                                               109
                                               0.00
                                               0.80
                                                26
                                                PD
                                                CH/K
360
 55
163
 19
  4
                  66
                  83
                 0.79
                 0.08
PD = Pfizer Dolomite (BCR No. 1337)
CH/K - Mixture of Champion Coal  (2% Sulfur) and Kentucky Coal  (0.6% Sulfur)
*  Excess Air  Percent derived from flue gas composition.

-------
                      APPENDIX H-2.   PARTICLE SIZE DISTRIBUTION SPENT PFIZER 1337
                                   DOLOMITE SORBENT (EXCEPT AS NOTED)
Run No,
   ;(D
 45
 48
 50. IB
 50,
 50.
 50.
 51
.2
.3
.5
Particle Size (ym)
5%
Material
Less
Than
10%
Less Than
25%
Less Than
50%
Less
Than
75%
Less
Than
90%
Less
Than
95%
Less Than
  56
Final Bed            29
Final Bed           295
Final Bed           200
Rejected Solids
Rejected Solids     160
Rejected Solids     130
Final Bed           560
Rejected Solids     180
Rejected Solids     320
Rejected Solids     300
Final Bed           220
150
350
260
120
235
215
870
280
440
450
270
 320
 440
 680
 420
 480
 440
1260
 470
 660
 560
 340
 490

1520
 960
1040
1080

 750
1060
 920
 540
                                              1500
                                              1720
1240
1540
1360
 920
                                                                                 1400
                                                                    1700
 (1)   Grove  1359  Limestone-Regenerated
                    APPENDIX H-3.  PARTICLE SIZE DISTRIBUTION PRIMARY CYCLONE CAPTURE
      Run  No.

        48
        56
        57
        59
               5%
            Less Than

               130
               102
   10%
Less Than

   140
   130
                                     Particle Size (ym)
                                 25%         50%         75%
                              Less Than   Less Than   Less Than
                                 190
                                 235

                                 140
              300
              710
              135
              190
                 880
                1150
                 240
                 365
                90%
             Less Than


                1550
                 800
                 850
                 95%
              Less Than


                 1850
                 1175
                 1250

-------
                  APPENDIX H-4.   PARTICLE SIZE DISTRIBUTION
                          SECONDARY CYCLONE CAPTURE
                                 Particle Size (pm)
            5%       10%       25%       50%       75%       90%       95%
Run No. Less Than Less Than Less Than Less Than Less Than Less Than Less Than
41(1)
43<1m
43.2)^
( / )
43.3^'
43.4U;
43. 4,
I / I
43.5U;
45(D
48
50. 1A
50.2
50.3
51 Bi
ci VJ/
J-L / T \
( J )
c-i V-'/
JJ- /ON
( _J )
c -I \~/ /
52
52
52
55
56
57
57
59
59
59
5.6
9.2
7.4

3.2
5.6
12

5.4
—
4.4
6
4.3
4.5
6.3
5.4

6.0

4.2
4.1
4.1
4.2
5.2
5.4
4.8
4.7
3.8
3.9
4.0
8
12
11

8.0
9.8
18

8.8
3.1
5.6
8.2
5.8
6.1
9.6
7.4

8.0

5.5
5.6
5.7
5.7
7.3
8.0
6.8
6.4
5.2
5.2
5.8
15
21
20

16
20
49

18
8.6
8.9
14
10
10
15
13

14

9.4
11
10
11
15
15
12
12
9.8
9.6
12
36
39
57

35
45
72

39
23
16
26
18
19
53
23

27

20
19
19
20
45
31
23
23
19
18
26
58
59
79

68
70
80

77
60
29
72
39
44
180
75

94

67
36
34
38
140
54
47
46
36
35
48
80
92
115

92
95
94

110
120
68
115
72
86
320
140

210

160
54
59
58
275
76
78
69
70
70
80
98
110
145

110
112
110

145
230
90
185
100
115
375
215

320

250
72
78
76
335
94
100
84
105
105
115
Pfizer 1337 Dolomite, Champion Coal Except as Noted

(1)  Grove 1359 Limestone-Regenerated

(2)  Pfizer Dolomite, Illinois Coal

(3)  First Cyclone Deactivated
                                     212

-------
                               APPENDIX H-5.  PARTICLE SIZE DISTRIBUTION GBF CAPTURE
NO
Run No .
54
55
55
55
55
55
55
55
59
59
59
59
Filter
Element
No.
—
2
2
2
2
3
3
F(1)
2
2
3
3
5%
Less Than
2.5
2.0
2.1
1.8
1.6
2.3
1.8
1.4
1.2
1.4
1.6
1.5
10%
Less Than
3.3
2.6
2.7
2.2
1.9
2.9
2.2
2.2
1.6
1.8
2.2
1.9
Particle Size
25% 50%
Less Than Less Than
5.5
4.5
4.7
4.0
3.0
4.5
3.7
4.1
2.6
2.7
3.9
3.0
10
9.5
10
13
6.5
8.0
8.8
7.3
5.2
4.9
7.1
5.4
(ym)
75%
Less Than
18
22
24
25
18
21
22
13
9.8
9.0
11
9.6
90%
Less Than
30
36
38
33
35
38
35
20
25
15
18
18
95%
Less Than
37
—
—
—
—
—
—
—
—
21
—
30
       (1)   Fines  Collected  on  Exterior of Filter Element

-------
                                     APPENDIX H-6.  PARTICLE SIZE DISTRIBUTION FLUE
                                               GAS PARTICULATES NO FILTER
Particle Size (vim)

Run No.
48
50.1
50.2
50.4
50.5
51
51
5%
Less Than
__
—
—
—
—
—
—
10%
Less Than
__
—
—
—
1
—
—
25%
Less Than
1.5
1.8
1.4
1.5
1.9
2.0
1.5
50%
Less Than
3.1
4.7
2.4
2.5
3.3
4.6
2.8
75%
Less Than
6.5
7.2
4.7
5.0
6.5
9.2
6.9
90%
Less Than
11
13
8.8
9
11
15
11
95%
Less Than
15
16
12
12
14
__
14
N5
                                      APPENDIX H-7.   PARTICLE  SIZE  DISTRIBUTION  FLUE
                                               PARTICULATES  AFTER FILTER
Particle Size (ym)

Run No.
54
57
59
59
59
5%
Less Than
1.3
2.0
1.3
1.0
1.0
10%
Less Than
1.8
2.3
1.7
1.3
1.3
25%
Less Than
2.7
3.4
2.4
1.7
1.9
50%
Less Than
5.0
6.7
3.4
2.5
3.3
75%
Less Than
17
15
2.7
4.0
8.5
90%
Less Than
—
25
13
9.0
20
95%
Less Than
__
31
20
16
30

-------
                               APPENDIX H-8.  MINIPLANT SOLIDS ANALYSIS
 Run No.

41.1


41.2

43.1



43.2
                                                     Weight Percent
 43.3
 43.4
	Source	

Second Cyclone #2
Second Cyclone #3
Second Cyclone #4

Final Bed
Second Cyclone #5

Initial Bed
Second Cyclone #8
Second Cyclone #9
Second Cyclone #10

Second Cyclone #13
Second Cyclone #14
Second Cyclone #15
Flue Gas Particulates
Intermediate  Bed

Second Cyclone #25
Second Cyclone #26
Second Cyclone #27
Flue Gas Particulates
Final Bed

Initial Bed
Rejected Solids #2
Rejected Solids #3
Rejected Solids #4
Second Cyclone #5
Second Cyclone #6
Second Cyclone #7
Flue Gas Particulates
 Ca

10.8
 8.8
 6.6

36.4
16.8

34.5
15.8
14.8
15.6
10,
11,
10.3
 6.1
25.3
 9.1
 9.9
10.1
 7.3
26.8
26.8
23.3
26,
30.
15.3
12.4
14.0
 7.2
                                             .6
                                             .3
	 s_
4.0
2.3
1.9
7.2
3.8
3.0
7.8
7.4
7.6
5.7
5.5
5.1
9.4
9.0
5.9
5.8
5.9
9.8
8.7
8.7
11.2
11.6
11.9
8.2
8.5
7.5
11.5
SQ4
12.4
9.6
8.1
22.3
12.1
14.0
27.1
26.1
26.6
19.6
18.5
26.0
27.6
29.8
18.3
20.4
18.2
25.7
28.6
28.6
28.0
31.5
33.0
25.9
26.2
26.1
30.4
C03
0.6
0.6
0.5
5.8
1.7
37.6
1.7
1.7
1.7
1.6
1.8
1.6
0.0
18.6
0.7
0.8
0.9
0.0
1.3
1.3
2.9
4.2
4.6
1.5
1.3
0.8
0.1
Total C

   6.8
   4.4
   3.4

   1.1
   8.7

   8.6
   0.5
   0.6
   0.2

   1.6
   1.9
   2.2
   0.5
   2.6

   1.2
   1.0
   3.8
   4.1
   0,6

   0.6
   1.2
   1.2
   1.3
   0.7
   0.6
   0.4
   1.4
           Other
  0.5
  8.4
  8.3
  9.0

  5.6
  5.7
  5.5
  3.1
  8.9

  4.9
  5.5
  5.5
  3.2
11.5

11.5
 8.3
 5.8
 8.9
 8.7
 6.8
 7.7
 2.7

-------
APPENDIX H-8 (CONTINUED).   MINIPLANT SOLIDS ANALYSIS
                           Weight Percent
Run No.
43.5



45











46.1-.2

46.3



46.4



47





Source
Rejected Solids #8
Second Cyclone #11
Flue Gas Particulates
Final Bed
Second Cyclone #6
Second Cyclone #9
Second Cyclone #17
Second Cyclone #22
Second Cyclone #27
Second Cyclone #30
Second Cyclone #33
Second Cyclone #36
Second Cyclone #37
Second Cyclone #39
Flue Gas Particulates
Final Bed
Final Bed
GBF
Second Cyclone #7
Second Cyclone #8
Rejected Solids #4
Rejected Solids #5
Second Cyclone #10
Rejected Solids #6
Primary Cyclone
Final Bed
Second Cyclone #1
Second Cyclone #2
Second Cyclone #5
Flue Gas Particulates
Primary Cyclone
Final Bed
Ca
26.0
9.1
5.7
27.5
5.8
6.5
8.2
8.1
5.6
4.3
3.4
3.8
3.0
3.3
5.9
32.6
26.3
8.3
5.4
4.5
26.2
27.4
5.3
26.1
5.6
24.4
5.2
7.4
7.8
4.6
13.5
29.7
S
10.4
5.1
9.3
12.4
1.9
2.7
2.7
2.5
2.0
1.5
1.1
1.2
1.0
0.9
—
10.9
12.4
2.8
2.9
4.6
11.4
10.5
2.5
11.8
—
13.1
—
—
—
2.7
5.2
12.6
S04
40.4
16.0
28.4
34.7
7.0
10.7
8.1
7.4
5.3
4.1
2.9
2.9
2.6
2.7
17.9
28.7
36.7
9.7
7.5
7.1
34.5
33.5
7.5
35.1
10.6
37.6
12.8
8.6
7.2
7.1
12.9
35.1
CQ^
2.6
0.4
—
2.8
0.3
0.5
0.5
1.0
0.6
0.1
0.1
0.1
0.1
0.1
0.1
2.3
7.0
7.5
1.5
3.8
5.3
7.2
1.5
7.8
1.1
10.5
2.8
1.1
1.7
0.9
1.5
13.6
Total C Mg Other
1.4 10.2 Na = .04
1.1 4.9 Na=0.4
2.6 2.2 Na=0.8
0.5 11.3
13.9
12.8
8.4
13.6
11.6
12.4
13.5
12.4
15.5
11.7
1.7
0.8
1.5
7.1
16.3
17.3
1.9
2.5
8.5
2.2
1.8
2.4
36.6
12.4
8.3
36.7
3.0
4.4

-------
                         APPENDIX H-8 (CONTINUED).   MINIPLANT SOLIDS ANALYSIS
 Run No.
50.3
50.4
                                                    Weight Percent
 50.5
 51
Source
Second Cyclone #43
Second Cyclone #46
Rejected Solids #25
Rejected Solids #28
Second Cyclone #49
Second Cyclone Mix #1
Second Cyclone Mix #2
Rejected Solids #31
Rejected Solids Mix #1
Rejected Solids Mix #2
Flue Gas Particulates
Second Cyclone #61
Second Cyclone #63
Second Cyclone #65
Rejected Solids #44
Rejected Solids #46
Rejected Solids #47
Flue Gas Particulates
Second Cyclone #4
Second Cyclone #5
Second Cyclone #6
Second Cyclone #7
Rejected Solids #1
Rejected Solids #2
Flue Gas Particulates
Flue Gas Particulates
Initial Bed
Bed Probe #1 (> 100 Mesh)
Bed Probe #1 ( < 100 Mesh)
Bed Probe #2
Bed Probe #3
Ca
8.8
7.7
23.5
22.5
7.6
8.7
7.5
22.4
20.0
19.9
4.0
7.7
8.9
6.5
17.3
21.8
21.8
4.1
5.0
4.5
4.8
3.8
21.9
9.6
4.0
3.6
22.2
18.4
19.7
19.2
20.6
S
5.0
4.0
6.8
7.2
4.3
4.9
4.5
6.0
7.5
7.5
3.1
5.2
5.0
4.8
6.7
7.8
8.5
1.2
2.1
2.4
4.1
2.0
7.6
7.8
4.3
4.7
9.9
11.3
9.4
8.9
7.0
SQ4
14.7
13.2
21.9
22.8
12.9
14.0
12.8
19.1
18.0
22.0
10.0
14.6
14.6
14.0
21.7
24.2
24.9
6.4
6.7
6.1
6.6
5.5
26.3
24.9
15.7
13.7
35.1
32.3
29.3
28.0
22.4
C03
2.1
1.9
22.3
21.6
1.6
2.1
2.0
23.7
14.8
14.5
0.0
2.7
2.0
1.5
20.1
19.7
17.9
0.0
1.9
1.8
2.0
1.4
21.2
23.8
0.3
0.1
14.7
7.7
17.5
17.3
23.0
Total C
3.7
3.8
1.4
1.3
4.6
3.3
3.6
1.5
3.0
2.9
2.0
3.1
2.6
2.4
1.6
1.4
1.0
1.7
6.6
9.0
12.5
9.8
3.8
4.7
3.6
3.1
3.8
1.6
2.8
3.0
3.2
Mg
4.4
4.3
13.5
13.0
4.0
4.9
4.1
13.1
16.9
12.1
1.6
4.4
4.8
3.5
9.6
13.1
13.0
2.0
2.6
2.3
2.4
2.1
12.9
5.6
1.7
1.6
12.9
10.6
12.1
12.3
12.3
Other





















Na = 0.6%
VT f\ f\&/
Na = 0.0%
TVT f\ O Of
Na = 0-3%
Na = 0.5%






-------
                                 APPENDIX H-8 (CONTINUED).  MINIPLANT SOLIDS ANALYSIS
         Run No.

        48
                                                            Weight Percent
NJ
i-1
00
50.1-50.5






50.1A



50. IB
       50.2
Source
Second Cyclone #4
Second Cyclone #5
Second Cyclone #6
Second Cyclone #7
Flue Gas Particulates
Flue Gas Particulates
Primary Cyclone
Initial Bed
Final Bed
Initial Bed
Final Bed
Primary Cyclone
Flue Gas Particulates
Second Cyclone Mix #1
Second Cyclone Mix #2
Second Cyclone #12
Second Cyclone #15
Second Cyclone #17
Rejected Solids #1
Rejected Solids #2
Rejected Solids #3
Flue Gas Particulates
Second Cyclone #25
Second Cyclone #27
Second Cyclone #28
Rejected Solids #8
Rejected Solids #10
Flue Gas Particulates
Ca
5.2
4.7
4.7
4.3
3.6
3.7
11.0
17.2
22.1
22.9
16.4
14.9
3.8
7.3
8.3
6.5
5.7
5.2
23.4
23.4
22.3
4.4
7.7
6.0
5.8
20.3
22.2
3.5
_S 	
3.1
2.9
2.6
2.8
5.1
6.8
5.4
13.6
7.1
7.4
8.4
7.3
5.3
4.6
3.8
4.3
3.2
3.1
9.4
7.9
7.9
6.7
4.0
3.3
3.3
7.6
7.5
4.6
S04
9.8
9.2
8.9
9.1
19.1
17.8
16.2
42.2
23.5
21.7
25.7
23.4
12.3
13.4
13.0
11.7
10.6
9.4
28.2
23.9
24.5
21.8
11.9
9.1
9.5
23.5
21.3
14.3
C03
0.7
0.5
0.3
0.4
0.0
0.0
1.2
2.5
22.9
22.9
20.6
2.3
0.9
5.3
4.7
1.3
0.9
0.7
11.2
20.4
21.7
0.2
3.5
2.8
3.0
16.1
22.7
0.1
Total C
8.9
6.5
6.2
6.8
2.9
1.8
2.2
0.8
4.1
3.9
2.3
1.4
1.9
5.7
7.1
8.8
8.2
7.2
2.3
3.7
1.6
3.1
9.3
9.4
9.8
2.0
2.3
3.5
Mg
2.3
2.3
2.4
2.4
1.5
1.5
6.1
9.4
13.2
13.7
9.8
9.5
1.9
3.4
3.9
3.1
2.7
2.4
6.5
9.9
12.7
1.5
4.3
3.2
2.8
11.8
13.0
1.3
Other
Fe = 10.1
Fe = 10.5
Fe = 10.4
Fe = 10.1
Fe = 7.4
Fe = 7.0
Fe = 11.7
Fe = 2.1
Fe = 0.3




















-------
                               APPENDIX H-8  (CONTINUED).  MINIPLANT  SOLIDS ANALYSIS
                                                          Weight Percent
       Run No.
       51  (Cont)
       52
KJ
       54
       55
Source
Bed Probe #4
Bed Probe #5
Bed Probe #6
Bed Probe #7
Final Bed
Second Cyclone #7
Second Cyclone #9
Second Cyclone #11
Rejected Solids #2
Rejected Solids #4
Rejected Solids #5
Flue Gas Particulates
Initial Bed
Bed Probe #3
Bed Probe #7
Bed Probe #10
Final Bed
Second Cyclone #4
Second Cyclone #5
Bed Probe #4
Bed Probe #5
Second Cyclone #1
Second Cyclone #2
GBF
Primary Cyclone
Final Bed
Ca
20.3
22.9
20.0
21.1
21.3
6.8
5.6
7.0
16.9
21.9
17.1
4.0
23.1
22.2
20.3
20.6
21.7
7.9
5.0
32.6
33.0
12.8
5.4
4.1
15.2
23.5
S
8.2
5.8
8.4
9.0
8.1
3.9
3.9
3.9
7.5
5.6
7.6
3.8
6.4
7.0
8.7
7.7
8.9
3.0
2.8
14.0
14.3
5.5
2.4
3.5
7.2
14.6
S04
29.4
18.0
26.8
24.9
20.8
11.5
10.7
11.9
23.4
20.4
26.6
8.6
26.7
17.4
24.4
22.6
29.2
8.2
5.7
43.4
40.8
15.1
7.8
9.9
21.8
46.3
C03
16.4
25.7
18.7
17.4
22.3
0.7
1.0
1.2
15.2
24.5
21.3
0.4
11.6
24.7
19.3
20.8
16.1
0.6
0.2
3.9
4.7
0.3
0.2
0.4
0.6
0.6
Total C
2.4
3.1
2.5
4.2
3.1
3.9
3.5
2.6
1.8
1.8
1.7
2.8
3.5
3.6
2.8
2.2
4.0
12.4
15.5
0.6
0.7
14.3
17.1
1.3
0.6
1.4
Mg Other
12.8
13.1
12.0
13.4
12.8
3.9
3.3
4.3
10.8
13.2
9.9
1.6
12.8
14.3
12.8
13.4
13.3
1.6
1.3
0.4
0.4
4.9
1.6
1.2
5.7
7.3

-------
                                APPENDIX H-8  (CONTINUED).  MINIPLANT SOLIDS ANALYSIS
                                                           Weight Percent
        Run No.

       56
       57
M
O
       59
Source
Second Cyclone #1
Initial Bed
Bed Probe #1
Bed Probe #2
Final Bed
Primary Cyclone
Second Cyclone #4
Second Cyclone #5
Initial Bed
Bed Probe #4
Primary Cyclone
Second Cyclone #8
Second Cyclone
Second Cyclone
Flue Gas Particulates
Flue Gas Particulates
Primary Cyclone
Final Bed
Ca
13.8
24.7
22.6
21.8
20.8
19.6
9.3
8.3
24.2
19.7
16.9
4.4
4.5
3.3
3.3
1.9
11.6
15.6
S
5.6
10.4
8.5
11.2
14.3
10.4
4.6
4.1
11.7
12.5
9.1
1.6
1.7
1.5
4.1
3.1
4.9
10.3
SO&
17.6
34.2
25.5
34.8
44.3
29.3
13.0
12.8
40.6
38.5
27.0
4.5
5.0
2.9
14.4
10.2
13.1
31.0
CQ-j
9.5
6.2
18.6
12.5
0.6
4.4
0.3
0.3
8.6
3.1
0.5
—
—
—
—
—
0.4
0.1
Total C
10.3
0.9
1.7
1.5
0.5
1.1
4.0
6.2
1.9
0.6
0.8
3.1
5.2
2.8
1.9
1.2
0.4
1.0
Mg Other
7.8
10.0
12.5
11.4
6.2
11.6
5.0
4.5
5.3
7.1
10.5
2.4
2.4
1.8
0.9
0.6
7.7
9.1

-------
                 APPENDIX H-9.  MINIPLANT SOLIDS COMPOSITION
  Run
Number
 41.1
 41.2

 43.1
 43.2


 43.3


 43.4


 43.5


 45
Composition (wt.
    Source
Sec.  Cyclone

Final Bed
Sec.  Cyclone

Sec.  Cyclone

Final Bed
Sec.  Cyclone
Flue Gas Part.

Final Bed
Sec.  Cyclone
Flue Gas Part.

Reject Solids
Sec. Cyclone
Flue Gas Part.

Final Bed
Reject Solids
Sec. Cyclone

Final Bed
Sec. Cyclone
Flue Gas Part,
46.2

46.3
46.4
Final Bed
GBF
Sec . Cyclone
Reject Solids
Final Bed
          Reject Solids
          Sec.  Cyclone
          Prim. Cyclone

  47      Final Bed
          Sec.  Cyclone
          Flue  Gas Part.
          Prim. Cyclone

  48      Final Bed
          Sec.  Cyclone
          Flue  Gas Part.
          Prim. Cyclone
c
(Combustible)
4.7
0.0
8.3
0.1
0.0
1.6
0.5
0.3
1.8
4.1
0.5
0.3
1.4
0.0
0.9
1.0
0.4
12.5
1.7
0.1
5.6
16.2
0.8
0.3
0.6
8.2
1.6
1.7
18.7
36.5
2.7
0.0
7.2
2.7
0.6
Ash

74
26
57
41
12
55
63
18
59
59
20
45
59
12
10
64
28
75
75
27
69
69
30
26
28
77
81
18
62
50
67
11
74
74
60
Sorbent

48
74
35
59
89
43
37
81
39
37
79
55
40
88
89
35
71
12
23
72
25
15
69
73
72
15
17
81
19
13
31
89
19
23
40
                                                  Sorbent Portion
                                                Composition (Mole %)
CaO
48
64
63
21
2
7
(1)
52
13
34
16
41
28
24
59
50
24
-9
-2
31
7
24
22
8
20
23
53
-13
12
CaC03
4
11
7
7
49
10
—
3
5
10
6
7
7
3
5
4
18
(2) 60
41
14
29
20
19
13
31
18
8
69
7
CaS04
48
26
30
72
49
83
—
45
82
56
78
53
65
73
37
46
58
49
61
55
64
56
59
80
49
58
40
44
81
                                                 31
                                           61
                                     221

-------
           APPENDIX H-9 (CONTINUED).  MINIPLANT SOLIDS COMPOSITION
  Run
Number


 50

 50.1A

 50. IB
 50.2



 50.3


 50.4



 50.5




 51
                     Composition  (wt,
     Source
 52
 54

 55



 56
Final Bed

Sec. Cyclone

Sec. Cyclone
Reject Solids
Flue Gas Part.

Sec. Cyclone
Reject Solids
Flue Gas Part.

Sec. Cyclone
Reject Solids

Sec. Cyclone
Reject Solids
Flue Gas Part.

Sec. Cyclone
Reject Solids
Flue Gas Part.
Prim. Cyclone

Final Bed
Sec. Cyclone
Reject Solids
Flue Gas Part.
Bed Probe

Final Bed
Sec. Cyclone
Reject Solids
Flue Gas Part.
Bed Probe

Sec. Cyclone
Bed Probe

Final Bed
Sec. Cyclone
Prim. Cyclone
Flue Gas Part.

Final Bed
Sec. Cyclone
Prim. Cyclone
Bed Probe
c
(Combustible)
0.0
5.4
7.9
0.0
3.1
8.9
0.0
3.5
3.3
0.0
3.4
0.0
2.0
2.3
0.0
1.7
0.0
0.0
9.1
0.0
3.3
0.0
0.8
3.1
0.0
2.7
0.0
13.8
0.0
1.2
17.1
0.5
1.2
0.4
8.4
0.2
0.0
Ash

26
63
70
18
70
66
18
78
65
14
66
19
81
66
21
84
41
16
74
25
76
17
10
71
23
82
41
69
15
15
66
50
83
23
38
25
14
Sorbent

76
32
22
83
27
26
83
19
32
89
31
82
17
31
82
14
58
85
17
75
20
84
89
26
79
16
61
17
85
84
17
49
16
77
54
74
88
  Sorbent Portion
Composition (Mole %)
CaO
-49
-14
13
2
2
-5
13
-4
14
4
4
-12
24
-11
16
-13
-13
-6
17
-25
-11
50
38
16
37
38
9
1
23
-3
CaC03
84
43
11
51
32
61
16
64
16
57
18
63
10
70
26
71
59
50
10
73
69
4
9
2
3
2
2
46
15
47
CaS04
65
71
76
46
66
44
71
41
70
40
78
49
65
41
58
42
54
56
73
53
43
45
54
82
60
60
89
53
62
57
                                     222

-------
           APPENDIX H-9  (CONTINUED).  MINIPLANT  SOLIDS COMPOSITION
  Run
Number
Source
Composition (wt. %)	
 C         Ash   Sorbent
                          (Combustible)
         Sorbent  Portion
       Composition (Mole  %)
       CaO   CaCO^   CaSOA
 57      Sec. Cyclone
         Prim. Cyclone
         Bed Probe

 59      Final Bed
         Sec. Cyclone
         Flue Gas Part.
         Prim. Cyclone
                  5.1
                  0.7
                  0.0

                  1.0
                  3.6
                  1.5
                  0.3
           64
           35
           26

           36
           83
           83
           59
31
64
74

63
13
15
40
37
32
 8

17
50

51
 2
 2
10

 0
 8
61
67
81

83
42

47
  (1)   Flue gas particulates sulfate during sampling at  lower  temperatures
        giving erroneous results.

  (2)   Negative values caused by calculation procedure and possible error
        in analyses.
                                      223

-------
                                  APPENDIX H-10.  MINIPLANT SAMPLE  SHIPMENTS
         Requestor
Brookhaven National Labs
Uptin, NY

Curtiss-Wright
Wood-Ridge, NJ

Dravo Corp.
Pittsburgh, PA

EPA
Research Triangle Park, NC

ERE
Florham Park, NJ
Baton Rouge, LA
Baytown, Texas
        Sample Description
Foster Wheeler Co.
Livingston, NJ

General Electric Co.
Schenectady, NY

MIT
Cambridge, MA
Fly Ash
Primary Cyclone

Fly Ash
Bed
Bed
Flue Gas Particulates
Limestone
Fly Ash
Illinois Coal Fines
Limestone
Limestone Fines
Limestone
Dolomite
Aragonite
Limestone
Dolomite

Primary Cyclone
Fly Ash
Fly Ash

Dolomite
Illinois Coal Fines
Illinois Goal No. 6
 Sorbent

Dolomite
Dolomite

Dolomite
 Run No.

43.5
43.5

37
Dolomite    23
Alumina     28
Limestone   45
                                                                  Limestone   47
Dolomite    37
Dolomite    51
  Amount

  5 Ibs.
  5 Ibe.

110 Ibs.
                                                                                                       Date
 3/14/77
 3/14/77

10/29/76
            5 gallons    3/4/76
            2 gallons    4/1/76
            13.5 grams   1/7/77
—
Limestone
Dolomite
—
—
—
—
—
—
—
—
45
43
—
—
—
—
—
—
—
2 gallons
360 Ibs.
360 Ibs.
10 Ibs.
900 Ibs.
1 Ib.
1 Ib.
1 Ib.
5 Ibs.
5 Ibs.
9/21/76
11/2/76
11/2/76
1/7/77
5/14/76
5/23/77
5/23/77
5/23/77
8/2/77
8/2/77
                          5 Ibs.
            240 Ibs.
            168 Ibs.

             25 Ibs.
            100 Ibs.
            100 Ibs,
                         1/7/77
              3/2/77
              6/6/77

              7/13/76
              2/3/77
              2/14/77

-------
                            APPENDIX H-10  (CONTINUED).   MINIPLANT SAMPLE SHIPMENTS
	Requestor	

National Gypsum Co.
Buffalo, NY

O.R.N.L.
Oak Ridge, TN

Pratt & Whitney  Co.
Middletown,  CT

 Ralph Stone  & Co.,  Inc.
 Los Angeles, CA
        Sample Description
Bed
 Research Cottrell
 Bound Brook, NJ
 T.V.A.
 Muscle Shoals, AL
Fly Ash
Fly Ash
Fly Ash

Bed
Fly Ash
Bed
Fly Ash
Bed
Fly Ash

Flue Gas Particulates
*Flue Gas Particulates
*(under non-isokinetic conditions)

Bed
Bed
Bed
Bed
Bed
Bed
Sorbent
Limestone
Dolomite
Limestone
Dolomite
Limestone
Dolomite
Dolomite
Dolomite
Dolomite
Dolomite
Dolomite
Dolomite
Limestone
Limestone
Limestone
Dolomite
Limestone
Dolomite
Run No.
19.6
38.1
45
43.4
19.7
37
43.5
43.5
37
37
50.1
50.1-50.5
19.3
30.1
30.2
32.3
19.5
43.5
Amount
1 gallon
2 Ibs.
20 Ibs.
5 Ibs.
200 Ibs.
200 Ibs.
100 Ibs.
150 Ibs.
100 Ibs.
100 Ibs.
25 grams
200 grams
500 grams
500 grams
500 grams
500 grams
7 Ibs.
7 Ibs.
Date
1/19/77
4/23/76
2/9/77
4/8/77
4/27/76
4/27/76
10/22/76
10/22/76
12/15/76
12/15/76
5/17/77
5/17/77
3/4/76
3/4/76
3/4/76
3/4/76
12/21/76
12/21/76

-------
                               APPENDIX H-10  (CONTINUED).  MINIPLANT  SAMPLE  SHIPMENTS
            Requestor
Sample Description
Westinghouse Research Labs. Bed
Pittsburgh, PA Fly Ash
Bed
Bed
Fly Ash
Fly Ash
Flue Gas Particulates
Fly Ash
Bed
Bed
Fly Ash
Regenerator Bed
Dolomite
Dolomite
Limestone
Limestone
Limestone
Limestone
Dolomite
Dolomite
Dolomite
Limestone
Limestone
Limestone
27
27
19.6
30.2
19.6
26
34
43.3
43.3
45
45
45
50 Ibs.
50 Ibs.
50 Ibs.
50 Ibs.
2 Ibs.
2 Ibs.
100 grams
300 Ibs.
275 Ibs.
20 Ibs.
35 Ibs.
50 Ibs.
3/8/76
3/8/76
3/8/76
3/8/76
3/8/76
3/8/76
4/12/76
11/18/76
11/18/76
1/7/77
1/7/77
1/7/77
N>

-------
   APPENDIX J-l




MODIFIED FILTER BED
        227

-------
     APPENDIX J-2

MODIFIED FILTER ELEMENT
    FLUIDIZING GRID
           228

-------
   APPENDIX J-3




GBF PRESSURE VESSEL

-------
                                                         APPENDIX J-4


                                                  GBF PRESSURE VESSEL LINING
K>
CO
o

-------
     APPENDIX J-5

FILTER PRESSURE VESSEL
      (SIDE VIEW)
         231

-------
     APPENDIX J-6




GBF PIPING ARRANGEMENT


-------
                                         APPENDIX J-7

                              PREHEAT BURNER SYSTEM SCHEMATIC
                                                              Pressure Gauge
                                                High  Limit Gas
                                                                        Low Limit
10
OJ

Pressure
Vessel
Lim
Gas
Flame
Scanner
N
i
— n f 	 r—>
it
iting Ori
Valve
\
IP
r •3-K'K
/ V"
Burner J?L
v<
G;
fi
jriable
js Reg
:e

>


V*

1

•H\
Ignition
Transformer
Pressure Switch W ' pressu,.e
^'° / A V Switch
ulator / i J3 Y
X^l 1 IS^l ^ rXl
r
S--, Bilking V"?' "
w , Reset
Valve Shutoff Valve
JX^--- \s&
—--Pressure Gauges -""i
t
Motorized
Air Valve
p


XI 	 1
i/lain Gas
Shutoff \
., L
9"^s
Natural Gas
(2 psig)
telve
-ow Limit
Ur Pressure
>witch

                                                                               Blower

-------
                   APPENDIX J-8




GRANULAR BED FILTER-BLOW. BACK SYSTEM FLOW SCHEMATIC

-------
                                TECHNICAL REPORT DATA
                         (Please read Instructions on the reverse before completing)
   EPORT NO
 EPA-600/7-78-069
                                                     !. RECIPIENT'S ACCESSIOf+NO.
4. TITLE AND SUBTITLE MiniplantStudies Qf PrgSSUHZed

Fluidized-bed Coal Combustion: Third Annual Report
                                                     5. REPORT DATE
                                                      April 1978
                                                     6. PERFORMING ORGANIZATION CODE
         R.C.Hoke, R.R.Bertrand, M.S.Nutkis, L.A
Ruth,  M. W. Gregory, E. M. Magee  M. D. Loughnane.
R. J.Madon  A.R. Garabrant, and M. Ernst	
                                                     8. PERFORMING ORGANIZATION REPORT NO.
9. PERFORMING ORGANIZATION NAME AND ADDRESS
Exxon Research and Engineering Company
P.O. Box 8
Linden, New Jersey 07036
                                                     10. PROGRAM ELEMENT NO.
                                                     EHE623A
                                                     11. CONTRACT/GRANT NO.

                                                     68-02-1312
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
                                                     Annual: 8/76 - 8/77
                                                     14. SPONSORING AGENCY CODE
                                                      EPA/600/13
is. SUPPLEMENTARY NOTES jERL-RTP project officer is D. Bruce Henschel,  Mail Drop 61,
919/541-2825.  EPA-600/7-76-Oil and EPA-600/7-77-107 are earlier annual reports.
  . ABSTRACT
              repOrt presents further results of studies of the environmental aspects
 of the pressurized fluidized-bed coal combustion process , using the 218 kg coal/hr
 'miniplant' continuous -combustion/sorbent-regeneration system (0. 63 MW equiva-
 lent), and a 13 kg coal/hr bench-scale system.  Previous combustion studies on the
 miniplant combustor were extended to investigate emissions of SO2, SO3,  NOx, and
 particulates during combustion of a high-sulfur coal, and with the use of precalcined
 limestone as the SO2 sorbent. Percentage SO2 removals obtained with the  high-sulfur
 coal were similar to earlier ones with intermediate-sulfur coal. Precalcined lime-
 stone proved to be as effective as dolomite (on a Ca/S molar basis) in removing SO2.
 The performance of the miniplant regenerator was demonstrated in an uninterrupted
 125-hr run with continuous circulation of sorbent between the combustor and the
 regenerator. A granular bed filter for high temperature/pressure flue gas particu-
 late removal was installed on the miniplant: initial shakedown was completed with a
 24-hr continuous run. Sampling was completed on the miniplant combustor for com-
 prehensive analysis of emissions.
17.
                             KEY WORDS AND DOCUMENT ANALYSIS
                DESCRIPTORS
                                         b.IDENTIFIERS/OPEN ENDED TERMS
                                                                   COSATI Field/Group
 Air Pollution
 Coal
 Combustion
 Fluidized Bed
  Processing
 Pressurizing
                      Limestone
                      Sorption
                      Regeneration
                      Sulfur Oxides
                      Nitrogen Oxides
                      Dust
Air Pollution Control
Stationary Sources
Particulate
13 B
21D
21B

13H,07A
08G
07D

07B

11G
13. DISTRIBUTION STATEMENT

 Unlimited
                                          Unclassified
                            234
                                         20. SECURITY CLASS (This page)
                                         Unclassified
                                                                 22. PRICE
EPA Form 2220-1 (9-73)

-------