STUDY OF CHARACTERIZATION AND
CONTROL OF AIR POLLUTANTS
FROM A FLUIDIZED-BED COMBUSTION
UNIT
THE CARBON-BURNUP CELL
for
Division of Control Systems
Office of Air Programs
Environmental Protection Agency
POPE, EVANS AND ROBBINS
INCORPORATED
-------
STUDY OF CHARACTERIZATION AND
CONTROL OF AIR POLLUTANTS FROM
A FLUIDIZED-BED COMBUSTION UNIT
THE CARBON-BURNUP CELL
for
Division of Control Systems
Office of Air Programs
Environmental Protection Agency
un de r
CONTRACT CPA 70-10
E. B. Robison, R. D. Glenn, S. Ehrlich
J. W. Bishop and J. S. Gordon
. FEBRUARY, 1972
POPE, EVANS AND ROBBINS
INCORPORATED
-------
ACKNOWLEDGEMENT
, .
The authors wish to acknowledge the skillful, energetic
and of times courageous efforts of the other members of
the Alexandria staff of Pope, Evans and Robbins.
Liman
Osborn and Wayne Pugh built the apparatus, kept it
operating and logged data.
Milton Lee and Edward Payson
satisfied the insatiable appetite of the Fluidized-Bed
Column for fly ash.
Yong Park and William Lowenbach
performed the analytical work.
LeRoy DeRuyter located
the odd things needed for the program.
The figures in
this report were drawn by John Fitch, Sam Pomeroy'and
Rene Paredes.
The typing was done by Luci Santi, Cie1
Zubrod, Jean Hamilton and Barbara Cianto.
D.
Bruce Henschel of the Division of Process Control
Engineering, Office of Air Programs, has been project
officer for Pope, Evans and Robbins work since 1968.
ii
POPE. EVANS AND ROBBINS
INCORPORATED
-------
TABLE OF CONTENTS
Paqe.
1. SUMMARY. . . . . . . . . . . . . . . . . . . . . 1
1.1 General . . . . . . . . . . . . . . . . . . 1
1.2 Te s t s in the Fluidized-Bed Column . . . . . 4
1.3 Tests in the Fluidized-Bed Module . . . . . 8
1.4 Use of Data Required in Both FBC and FBM. . 11
2. CONCLUSIONS. . . . . . . . . . . . . . . . . . . 13
3. RECOMMENDATIONS. . . . . . . . . . . . . . . . . 15
4. INTRO DUCT ION . . . . . . . . . . . . . . . . . . 17
4.1 Description of the Fluidized-Bed Boiler . . 17
-;-
4.1.1 Background: The Need for a New
Technique for Burning Coal . . . . 17
4.1.2 De f in ing a Fluidized-Bed Boiler. . . 19
4.1.3 A Simpl ified Description of a
Fluidized-Bed Boiler . . . . . . . 20
4.2 Air Pollution Control Potential of
Fluidized-Bed Boiler. . . . . ... . . . . 25
4.2.1 Perspect ive. .. . . . . . . . . . . . 25
4.2.2 Discussion . . . . . . . . . . . . . 26
4.3 Pope, Evans and Robbins Prior Work. . . . . 32
4.3.1 General. . . . . . . . . . . . . . . 32
4.3.2 Once-Through Limestone Injection . . 33
4.3.3 Re ge ne rat i ve Limestone Process . . . 34
4.3.4 Pollutants Other Than SO 2.. . . . . 35
4.3.5 Recommendations Based on Prior Work. 36
4.4 Prior Work Leading to the Burnup Cell
Concept . . . . . . . . . . . . . . . . . 37
4.5 Specific Objectives of This Work. . . . . . 38
5. APPARATUS. . . . . . . . . . . . . . . . . . . . 40
5.1 Pilot Scale Combustor, FBC. . . . . . . . . 40
iii
POPE. EVANS AND ROBBINS
INCORPORATED
~
.;
-------
5.2
5.3
5.4
6.
TABLE OF CONTENTS
(Continued)
Full-Scale Boiler, FBM. .
5.2.1
5.2.2
. . .
. . . .
. . .
Boiler Module, FBM . . . . . . . . . .
The Simulated Carbon-Burnup Cell,
CB C. . . . . . . . . . . . . . . . .
Instrumentation. .
.......
. . . . . .,
Ma te ria 1 s . . .
. . . .
. . . .
. . . .
5.4.1
5.4.2
5.4.3
. . .
Fly Ash for Tests in the FBC . . . . .
Coa 1 for FBC and FBM Tests. . . . . .
Limestone for FBCand FBM Tests. . . .
TESTS WITH THE FBC.
PROCEDURES AND RESULTS:
6.1
6.2
6.3
6.4
. . .
General. .
. . . .
. . .
. . . . .
. . . . .
Statistical Design of Experiments.
6.2.1
6.2.2
6.2.3
6.2.4
6.2.5
6.2.6
. . . . .
Pe r spec t i ve. . . . . . . . . . . . . .
Identification of Variables. . . . . .
Selection of Most Important Variables.
Variables Actually Studied. . . . . .
Preliminary Modeling (A Simple
Physical Model). . . . . . . . . . .
The Design of the Experiments: .
Selection of Test Conditions. . . .
General Results: Exploring the Regime of
Stable Operation. . . . . . . . . . . . . .
Statistically Derived Model for Fly Ash
Carbon Combustion in a Fluidized-Bed
Combus tor. . . . . . . . . . . . . .
6.4.1
6.4.2
6.4.3
6.4.4
. . .
Description of Model. . . . . . . . .
Re suI t s. . . . . . . . . . . . . . . .
Test of Model in Same Apparatus. . . .
Relative Significance of Parameters
in Model Equation. . . . . . . . . .
6.4.4.1
6.4.4.2
Effect of Bed Temperature, T.
Effect of Bed Depth, H. . . .
iv
POPE, EVANS AND ROBBINS
INCORPORATED
paqe
46
46
54
60
67
67
69
70
71
71
71
71
72
77
78
81
86
87
92
92
94
98
100
100
102
-------
6.5
6.6
6.4.5
6.4.6
6.4.7
6.4.8
TABLE OF CONTENTS
(Continued)
6.4.4.3
6.4.4.4
Effect of Air Rate, A. ...
Effect of Carbon Content of
Fly Ash Fuel, c. .....
Effect of Firing Rate, C. . .
Conclusion Regarding
Significance of Parameters.
6.4.4.5
6.4.4.6
Maintaining the Bed Temperature for
High Combustion Efficiency: Effect'
of the Coal Addition to the FBC. . . .
The Effect of Fine Limestone Sorbent
on Combustion Efficiency in a
Carbon-Burnup Cell. . . . . . . . . .
The Effect of Particle Size of Input
Fly Ash. . . . . . . . . . . . . . . .
Conclusions Regarding Statistically
Derived Model for Carbon-Burnup Cell
Performance. . . . . . . . . . . . . .
Applying the Performance Model in Design. . . .
6.5.1
6.5.2
6.5.3
6.5.4
Gene ra 1. . . . . . . . . .
. . .
. . . .
Normalizing the Model for Use in Other
Apparatus. . . . . . . . . . . . . . .
The Bed Temperature Anomaly. . . . . . .
6.5.3.1
6.5.3.2
Bi-Stable Combustion. . . . . .
Effect of Energy Release
Patterns on Temperature of
Fluidized Bed. . . . . . . .
Energy Loss to Cool Surfaces
Above the Dense Phase
Fluidized-Bed. . . . . . . .
6.5.3.3
An Example: Using the Carbon-Burnup
Cell Performance Model in Design. . .
Other Results from Tests in the FBC
6.6.1
6.6.2
6.6.3
. . .
. . .
Effect of Heat Removal. . . . . . . . .
Nitric Oxide Emissions. . . . . . . . .
Sulfur Dioxide Emissions from FBC
Without Presence of Limestone
Sorbents . . . . . . . . . . . . . . .
v
POPE. EVANS AND ROBBINS
INCORPORATED
Paqe
102
104
110
110
113
119
123
130
132
132
133
136
137
142
145
149
154
154
156
158
-------
6.6.4
6.6.5
6.6.6
7.
TABLE OF CONTENTS
(Continued)
Behavior of Limestone and Sulfur in
a Carbon-Burnup Cell. .. . . . . . .
6.6.4.1
Rernova1of 802 by Lime in a
Carbon-Burnup Cell. . . . .
Release of 802 by Lime. . . .
6.6.4.2
Particulate Emissions. . . . . . . . .
Hydrocarbon and Carbon Monoxide
Em iss ions. . . . . . . . . . . . . .
DEVELOPMENT OF AN INTEGRATED FLUIDIZED BOILER AND
CARBON-BURNUP CELL: TESTS WITH THE FBM . . . . .
7.1
General. .
.........
. . .
. . . . . .
7.2
Coal Feeding to the FBM.
........
7.2.1
7.2.2
. . .
General. . . . . . . . . . . . . . . .
Results. . . . . . . . . . . . . . . .
7.3' Feeding Fly Ash to the CBC .
7.3.1
7.3.2
. . . .
. . . . .
General. . . . . . . . . . . . . . . .
Fly Ash Feeder Designs Tested. . . . .
7.4
Interchange of Bed Material. .
........
7.4.1
7.4.2
7.4.3
Gene ra 1 . . . . . . . . . . . . . . . .
Measurement of Rate of ' Interchange. . .
Effect of Interchange of Bed Material
on Carbon-Burnup Cel'1' . . . . . . . .
Paqe
161
165
167
169
176
177
177
179
179
180
185
185
187
188
188
190
191
7.5 Prevent ion of Carryover of Jetted Particles
From a High Velocity Fluidized Bed . . . . . 192
7.6 Performance of a Baffle Screen as a Heat
Exchange r . . . . . . . . . . . . . . . . . . 196
7.7 Pe rformance of Integrated FBM/CBC System . . . 197
7.8 Sulfur Capture With Fine Limestone Injection . 200
vi
POPE, EVANS AND ROBBINS
INCORPORATED
-------
TABLE OF CONTENTS
(Cont inued )
8. . ECONOMICS OF THE CARBON-BURNUP CELL.
10.
8.1
8.2
9.
. . . .
. . . .
Perspect ive .
. . . .
. . . .
. . . . . .
. . .
Analysis. . . . . .
8.2.1
8.2.2
. . .
. . . .
. . . .
. . .
Evaluation of a Fluidized-Bed Boiler
Without Fly Ash Recycle. . . . . . . .
Evaluation of a Multi-Cell Fluidized-
Bed Boiler: Optimum Duty Split
Between Primary Cells and Carbon-
Burnup Cells. . . . . . . . . . . . .
SUMMARY OF DESIGN CRITERIA FOR CARBON-BURNUP CELL. .
REFERENCES
..........
.......
. . . .
vii
POPE, EVANS AND ROBBINS
INCORPORATED
Paqe
204
204
206
209
213
218
220
-------
No.
10.
11.
12.
13.
14.
15.
16.
LIST OF FIGURES
1.
Simplified Schematic of Multicell Fluidized-
Be d Bo i Ie r. . . . . . . . . . . . . . . . .
. . .
2.
Schematic of Fluidized-Bed Boiler
. . .
. . . . . .
3.
FluIdized-Bed Column (FBC) Construction Detail -
Front View. . . . . . . . . . . . . . . . . . . .
4.
Fluidized-Bed Column' (FEC) Construction Detail -
Side View. . . . . . . . . . . . . . . . . . . .
5.
Air Distribution Grid Button.
. . .
. . .
. . . . .
6.
Section Through Fluidized-Bed Column Showing
Insulation Steel Liner and Adjustable Cooling
Sur face. . . . . . . . . . . . . . . . . . . . .
7.
Schematic of FBC Air and Exhaust Gas Ducting
Showing Sampl ing Points. . . . . . . . . .
. . .
8.
Fluidized-Bed Module (FBM) Internal Construction. .
9.
Section Through Fluidized-Bed Module (FBM) and
Simulated Carbon-Burnup Cell (CBC). . . . .
. . .
Schematic of FBM Test System Showing Various
Subsystems. . . . . . . . . . . . . . . . . . . .
Mushroom Feeder for Fluidjzed-Bed.Combustor .
. . .
Details of Wall Between FBM and CBC . . . . .
. . .
Schematic of Gas Transfer System for Continuous
Monitoring of Sulfur Dioxide, Nitric Oxide,
Carbon Dioxide and Hydrocarbons. . . . .. . . .
Schematic of the FBM Gas Sampling System. . .
. . .
Schematic of CBC Gas Sampling ~ystem. . .
. . . . .
Predicted Air-Carbon Operating Regime of the FBC. .
viii
POPE. EVANS AND ROBBINS
INCORPORATED
Paqe
3
21
41
42
43
45
47
49
51
53
57
58
62
63
65
82
-------
No.
17.
18.
19.
20.
21.
22.
23.
24.
25.
26.
27.
28.
29.
30.
31.
LIST OF FIGURES
(Cont inued )
Comparison of Observed and Calculated Combustion
Efficiency in FBC Tests. . . . . . . . . . . . .
Comparison of Observed and Calculated Residual
Oxygen Leve 1 . . . . . . . . . . . . . . . . . .
Effect of Bed Temperature on Carbon-Burnup Cell
Performance Model Pr~dictions. . . . . . . . . .
Effect of Static Bed Depth on Carbon-Burnup Cell
Performance Model Predictions.. .......
Effect o"f Excess Stoichiometric Air on Carbon-
Burnup Cell Performance Model Predictions. . . .
Predicted Effect of Carbon ~ontent of Fly Ash
Feed on Performance of Carbon-Burnup Cell. . .
Predicted Combustion Efficiency versus Carbon
Content of Fly Ash from Primary Fluidized-Bed
Boiler Cell. . . . . . . . . . . . . . . . . . .
Effect of Firing Rate on Predicted Plan Area of
Carbon-Burnup Cell. . . . . . . .. .....
Predicted Combustion Efficiency of Mu1tice11
Fluidized-Bed Boiler. . . . .. ..... .
Particle
Distribution for FBC Test C-312. . .
Size
Particle Size Distribution of Fly Ash Discharged
to Atmosphere - FBC Test C-312 . . . . . . . . .
Illustration of Bi-Stab1e Combustion Process
. . .
Effect of Energy Release Partition, a , on
Temperatures Hypothetical Fluidized-Bed
Combustor. . . . . . . . . . . . . . . .
. . . .
Effect 'of Heat Removal on Combustion Efficiency
in FBC . . . . . . . . . . . . . . . . . . . . .
Nitric Oxide Concentration in Flue Gas from FBC
Versus Temperature. . . . . . . . . . . . . . .
ix
POPE, EVANS AND ROBBINS
INCORPORATED
Paqe
96
97
101
103
105
106
109
111
117
125
128
140
143
155
157
-------
No.
32.
33.
34.
35.
36.
37.
38.
39.
40.
41.
42.
43.
LIST OF FIGURES
(Continued)
Typical Nitric Oxide Emission ResPQnses in FBC . .
Comparison of Observed and Calculated Nitric
Oxide Emission from FBC. . . . . . . . . .
. . .
Particulate Emission from FBC as a Function of
Be d Tempe ra t ure. . . . . . . . . . . . . . . . .
Ash Bed Particle Size Distribution Before and After
High Temperature Operation in FBC Test C-308 . .
FBM Coal Injector Modifications. . . . .
. . . . .
Typical FBM Residual Oxygen Maps FBM Test B-3. . .
Comparison of Observed and Calculated Combustion
Efficiency of CBC. . . . . . . . . . . . . . . .
Sulfur Dioxide Reduction With Fine Limestone
Injection into FBM . . . . . . . . . . . .
. . .
Carbon Combustion Efficiency in a Coal-Fired
Fluidized-Bed Combustor. . . . . . . . . .
. . .
Thermal Losses from Fluidized-Bed Boiler Operating
Without Recycle of Carbon-Bearing Fly Ash. ~ . .
Thermal Losses fr'om Mu1tice11 Fluidized-Bed Boi1er-
o
Exhaust Temperature 400 F. . .,. . . . . . . . .
Thermal Losses from Mu1tise11 Fluidized-Bed Boi1er-
Exhaust Temperature 250 F ... . . . . ~ . . .
x
POPE, EVANS AND ROBBINS
INCORPORATED
Paqe
159
160
171
174
182
184
201
203
208
211
215
216
-------
No.
10.
11.
12.
13.
14.
15.
16.
LIST OF TABLES
1.
Combustion Air Feed Variables. . . . .
. . . . . .
2.
Fuel Feed Variables. . . .
. . . .
. . .
. . . . .
3.
Bed Va riables.
. . .
. . .
. . . .
. . .
. . .
. .
4.
Combustion-Gas Stream. . .
. . . . .
. . . .
. . .
5.
Flue Gas Solids Variables. .
. . .
. . . .
. . . .
6.
Reactor Variables. . . . .
. . . .
. . . . .
. . .
7.
Control and Pseudo-Control Variables Selected for
Carbon-Burnup Cell Investigation. . . . .'. . .
8.
Reliability of Carbon-Burnup Cell Performance
Mode 1. . . . . . . . . . . . . . . . . . .
. . .
9.
Regression Coefficients for Use in Response
Funct ion. . . . . . . . . . . . . . . .
. . . .
Illustration of Carbon-Burnup Cell Performance
Mo de 1. . . . . . . . . . . . . . . . . . . . . .
Predicted and Observed Carbon Combustion Efficiency
for Carbon-Burnup Cell Operation with Coal in
Fe e d . .'. . . . . . . . . . . . . . . . . . . .
Adjusted Values of Combustion Efficiency When Coal
is Fed to a Carbon-Burnup Cell. . . . . . . . .
Results of FBC Test 320-1 With Fly Ash Containing
Limestone By-Products. . . . . . . . . . . . . .
Carbon Content of Input and Output Fly Ash as
Funct ion of Particle Size. . . . . . . . . . .
Illustration of Bi-Stable Combustion in
FBC. . . .
Coal Analysis for Fuel to Fluidized-Bed Boiler.
An Example of Carbon-Burnup Cell Design. '. . . .
xi
POPE, EVANS AND ROBBINS
INCORPORATED
Paqe
73
74
75
76
76
77
79
95
99
98
114
115
121
126
138.
150
-------
No.
17.
18.
19.
20.
LIST OF TABLES
(Cont inued )
Sulfur Balance Summary for FBC Test. .
. . . . . .
802 Appearing in Exhaust of FBC Firing Fly Ash
Containing 19.4% Ca, 2.42% S . . . . . . . . . .
Effect of Fly Ash Feeder Design on Temperature
Distribution in a Carbon-Burnup Cell. . . . . .
Performance of FBM, CBC and System in Current Test
Se r ie s . . . . . . . . . . . . . . . . . . . . .
xii
POPE. EVANS AND ROBBINS
INCORPORATED
Paqe
162
169
188
198
-------
APPENDICES
A.
FBC Specifications
B.
FBM Specifications
C.
CBC Specifications
D.
Selected Carbon Burnup Results
E.
Summary of FBC Data for Carbon-Burnup Cell Tests
F.
Statistical Analysis and Modeling of FBC Fly Ash
Carbon-Burnup Data, by Arthu~ E. Hoerl, University
of Delaware
G.
Gas T~mperature Gradient in a Hot Fluidized Bed
H.
A Theoretical Analysis of Sulfur Retention
Phenomena (The Pseudosulfate Theory)
I.
Adjustment Procedure -for Carbon-Burnup Cell
Performance Model with Coal in Fuel
J.
Bed Material Flow Rate between Regions
K.
FBM-CBC Test Summary Data
L.
Discussion of Fluidized-Bed Combustion Variables
M.
Why the Carbon-Burnup Cell?
xiii
POPE, EVANS AND ROBBINS
INCORPORATED
-------
1
1.
SUMMARY
1.1
General
Pope, Evans and Robbins, in a continuation of a project
sponsored by the Office of Air Programs of the Environmental
Protection Agency (OAP), has monitored air pollutant emissions
from the combustion of coal in a fluidized-bed boiler. This
report describes the results of experiments anc studies
carried out between October 1969 and October 1970. In earlier
work (~)* it was found that sulfur oxide emissions could be
markedly reduced by injecting finely divided limestone into a
coal-burning, fluidized bed operating at l5000F to l6000F and
with about 3% residual oxygen** in the flue gas. However, with
3% residual oxygen in the flue gas, an economically unacceptable
fraction of the input coal's fuel value would be lost as carbon
in the flyash that is blown out of the furnace.
A comprehensive search for methods to reduce the loss of
fuel, which ranged as high,as 15%, led to the invention. of the
Carbon-Burnup Cell (~). The Carbon-Burnup Cell is simply a
region of a fluidized bed boiler in which the fuel. is the
carbon bearing flyash carried out of the adjacent coal burning
regions of the boiler. The Carbon-Burnup Cell is characterized
by design features which permit operation at a relatively high
temperature.
*Underscored figures in parenthesis refer to items in the list
of references in Section 10 of this report.
**Values for "residual oxygen" are given in lieu of the more
common "excess air" because the latter is ambiguous in
discussing the fluidized-bed boiler. A 3% residual oxygen
level is the equivalent of 15 to 17% excess air.
POPE, EVANS AND ROBBINS
INCORPORATEO
-------
2
The term "cell", which has an interesting origin, is
defined in Section 4. In this summary section it should be
explained that a commercial fluidized-bed boiler will be
composed of a combination of primary, burnup and, probably
regenerator cells. These cells are distinct regions of the
boiler in which specialized functions are carried out. In one
design concept 75% of the fluidized-bed boiler's cross-
sectional area is devoted to the coal-burning primary cells,
about 15% to the flyash burning Carbon-Burnup Cells and 10%
to regenerator cells in which calcium sulfate is decomposed.
A simplified schematic diagram of such a system is shown in
Figure 1.
OAP selected the Carbon-Burnup Cell for extensive study
because realization of the air pollution control potential of
the fluidized-bed boiler depended on the units commercial
feasibility, which in turn depended on the efficient combustion
of the fue 1. The program descr ibed in this report cons isted,
for the most part, of an effort to produce design criteria for
an effective Carbon-Burnup Cell.'
The experimental work was conducted in two different
test rigs. The more basic work, that which led to a
statistically derived performance model, was done in a
separate column which depends on an external source for its
. fuel supply. While this permitted the rate of fuel injection
to be controlled it also meant that the flyash fuel was days
or even weeks old. The second set of tests was performed in
an actual fluidized-bed boiler, generating the flyash which
is then immediately burned in its own internal flyash fired
section, the prototype Carbon-Burnup Cell. The work in the
boiler led to engineering insights on the.design requirements
for a commercial fluidized-bed boiler which included a
POPE. EVANS AND ROBBINS
INCORPORATED
-------
CLEAN FLUE GAS
STACK
L~
" DUST COLLECTOR
~
FLUE GAS
CONTAINING
CAR80N-BEARING
FLY ASH
f \
CARBON- BEARING L"
FLY ASH
ITI
~ ~
~(J)
o »
]J 2
~ 0
~ ;U
: ffi
~
~J
PRIMARY CELLS
CARGON-
.....UD,".IUP
o 1\1...
CELLS
DUST COLLECTOR
COAL~," "
AIR \ : P~-l
""". ~: 7 d,
LIME~~~~/~ ['" " .... ",n> .. " "" ,,,. :1 ,I.', -
. ClRCULATlNG LI MESTO-;;;J
ASH TO WASTE
v
REGENERATOR
CELLS
GAS RICH IN SULFUR DIOXIDE
TO RECOVERY PROCESS
C,.tJ
FIGURE I
SIMPLlFIE"D SCHEMATIC OF MULTICELL FLUIDIZED-BED BOilER
-------
4
Carbon-Burnup Cell. The operation of the boiler with a
simulated Carbon-Burnup Cell also served as an effective
demonstration tool for the process and design concepts.
The apparatus and the work are discussed below.
1.2
Tests in the Fluidized-Bed Column
The major portion of the effort to produce the design
correlations for flyashcombustion was carried out in the
Fluidized-Bed Column (FBC). The FBC is an experimental
fluidized-bed combustor which has a cross-sectional area of
0.86 square feet.
Carbon-containing flyash was burned in this device
over a wide range of conditions, summarized as follows:
Parameter
Ranqe Studied
l7500F to 2l40oF
Bed temperature
Bed depth (static)
10" to 2 2 "
385 to 1000 lb/hr ft2
48 to 350 lb/hr ft2
Air rate
Fuel rate
Carbon content of the
flyash fuel
Heat Removal Rate
28% to 65%
15% to 40% of heat release
For any given test the specific values within these
ranges were selected on the basis of a statistically-designed
program of experiments. As discussed in Section 5.4.1, the
carbon content of the flyash fuel was not controllable. The
bed material in these tests was crushed, sintered coal-ash.
The model equations, based on an analysis of 38 FBC
experiments are given below.
The equations are presented in a normalized form.
The
POPE. EVANS AND ROBBINS
INCORPORATED
-------
5
normalization assumes that the proper scaling factor is cross-
sectional area. This has not been demonstrated and will not
be until a large Carbon-Burnup Cell is constructed.
(bed temperature, of)
(air rate, lbjhr ft2)
(static bed depth, inches)
(carbon feeq rate, lb of carbonjhr ft2)
(inert feed rate, lb of inertjhr ft2)
(carbon feed rate x inert feed rate,
lb2jhr2ft4)
The standard deviation is 2.4%.
Combustion efficiency, %=
-13.78
+ 0.05193
+ 0.03973
+ 0.3831
- 0.7514
- 0.1638
+ 0.0020
(1 )
This model has been tested using data obtained in runs
not used in deriving the model. The results of these tests
indicated that the model was not limited to simply reproducing
itself but was probably a useful prediction tool.
It was further shown that by adjusting the control
parameters, the combustion efficiency of mixtures of flyash
and bituminous coal could also be predicted. Section 1.4
discusses another use for Equation 1.
The residual oxygen content of the flue gas could also
be predicted, although somewhat less accurately than the'
combustion efficiency. This model equation is given on the
following page.
A prediction tool for the C02 content of the gas also
comes out of the statistical analysis. It was reassuring to
note that the equation C + 02 = CO2 was reproduced by the
model equations.
POPE, EVANS AND ROBBINS
INCORPORATED
-------
6
Residual Oxygen, % =
+22.91
-0.007353 (bed temperature, of)
+0.01118 (air rate, 1b/hr ft2)
-0.1390 (static bed depth, inches)
-0.1521 (carbon feed rate, 1b carbon/hr ft2)
-0.0151 (inert feed rate, 1b of inert/hr ft2)
+0.0002653 (carbon feed rate x inert feed rate,
1b2/hr2ft4)
(2 )
The standard deviation is 0.69%, based on the 38 tests
in the FBC.
The body of the report contains a sample calculation
which shows how one may use these two model equations in the
design of a Carbon-Burnup Cell.
Mixtures of carbon-bearing f1yash and partially sulfated
fine lime were also fired in the FBC. These tests were
conducted because it was possible to control S02 emissions by
injecting finely ground limestone into the primary cells of a
fluidized-bed boiler. The fine limestone, would calcine, be
partially converted to calcium sulfate and then leave the
primary cells as part of the f1yash in flue gas. Most of this
dust would be removed from the flue gas by some gas cleaning
device, either a cyclone or electrostatic precipatator but
never a wet scrubber. Since no practical method exists "for
separating the spent sorbent from the carbon particles or ash
all of the dust collected would be injected into the Carbon~
Burnup Cell.
The tests, in which the fuel to the FBC contained
calcium sulfate, were intended to determine how, or rather if,
" the calcium sulfate content affected the carbon combustion
POPE, EVANS AND ROBBINS
INCORPORATED
-------
7
process and in turn how exposure to the high temperatures in
the FBC would affect the sulfate.
In the few tests conducted it was found that when the
partially sulfated lime entered the FBC alo~g with the carbon-
bearing fly ash, its sulfur content could be released as 802
if the FBC operated at a temperature of approximately 20000F
with residual oxygen ~ 2.0"/0. The highest 802 concentrations
observed from the FBC during these tests, approximately 1/2%,
were measured when the residual oxygen level was approximately
0.1%. On the other hand, when the residual oxygen level in the
FBC was in the range 3.5 to 6.0%, the fine sulfate did not
decompose at burn up cell temperatures. In fact, in one test
at 6% residual oxygen the partially sulfated limestone appeared
o
to still be reactive at a bed temperature 0t 1980 F. Two
possible explanations of this result are provided in the body
of the report. With high residual oxygen sulfur oxide emission
from the FBC was as low as 350 ppm.
When no sorbent was present, the sulfur in the fly ash
would burn with an efficiency equal to that of the carbon.
This fact is of no special significance but it is interesting
to note that when coal is burned at low temJ?eratures in a
fluidized-bed the carbon is more refractory than the sulfur.
The data gathered on nitric oxide emissions from a
Carbon-Burnup Cell was correlated less accurately than other
parameters, though'the highest emissions were detected at the
highest bed temperatures. A mean value of 539 ppm was measured
for all fly ash combustion tests in which the bed temperature
was above 1900oF. No efforts were made to reduce 'nitric oxide
emissions which remains an area requiring research. However,
even at the level of 539 ppm from the Carbon-Burnup Cell the
(}
POPE, EVANS AND ROBBINS
INCORPORATED
-------
8
fluidized-bed boiler, as a whole, compares favorably with
conventional boilers firing coal.
Particulate emissions from the FBC were found to
decrease with increasing bed temperature, possiblY through
agglomeration of ash matter to bed particles. The correlation
produced by the statistical analysis of the' independent
variables used for Equation 1 and 2 is
2
Pounds of particulate emission per hour per ft
=
+15.57
- 0.00664 (bed temperature, of)
- 0.00135 (air rate, lb/hr ft2)
+ 0.03034 (carbon rate, lb/hr ft2)
+ 0.00461 (inert rate, lb/hr ft2)
+ 0.00034 (carbon rate x inert rate, lb2/hr2ft4)
The residual st.andard deviation is 0.75 lb/hr ft2
(3 )
Since in some tests the fuel was approximately 65% ash,
particulate emissions were remarkably low. The potential of
retaining a large fraction of the ash in the bed of a Carbon-
Burnup Cell should be explored further.
Hydrocarbons and carbon monoxide from the FBC were
essentially nil when operating in the Carbon-Burnup Cell mode,
Le., high temperature and high excess air.
1.3
Tests in the Fluidized-Bed Module
Tests were also conducted in the Fluidized-Bed Module
(FBM), an actual boiler, with a grate* of approximately 9
square feet.
*The terms air distribution grid, grid, air distributor and
grate are used interchangeably in this report.
'0
POPE, EVANS AND ROBBINS
INCORPORATED
-------
9
A small (1.1 ft2 cross-sectional area) test column was
appended to the FBM in which the carbon bearing fly ash,
produced by the FBM, could be fired. The purpose of these
tests was to confirm that the results achieved in the FBC,
under carefully controlled conditions, could be achieved,
under the more realistic conditions of an actual boiler
operation. With this addition the FBM had a grate area of
slightly over 10 ft2.
The experimental test column was designated the "CBC".
In this report the term CBC is only used when referring to
the specific item of experimental apparatus added to the FBM.
The CBC only simulated a Carbon-Burnup Cell, as this is
defined in Reference 2 and in Section 4 of this report.
While demonstration was an important part of the FBM/CBC
design and test program it was also necessary to determine the
problem areas in operating a fluidized-bed with two distinct
temperature regions. A temperature of about l5000F was to be
maintained in the FBM, the CBC was to .operate at 2000oF. Bed
material could circulate between the two sections.
It was found .that for the small CBC, an opening in the
barrier between the two sections of just 2 square inches
provided adequate bed material circulation and permitted the
desired temperature difference to be achieved. An interchange
rate between the two regions of 12,000 pounds of bed material
per hour per square foot of opening was estimated from two
transient heating tests described in the report.
The tests with the FBM were only a partial success; in
most tests the combustion efficiency was less than 99%.
However in the two final tests it was possible to demonstrate
that the burnup results achieved in the FBC could be repr~duced
POPE, EVANS AND ROBBINS
INCORPORATED
-------
10
in an actual boiler.
So it was shown that an overall
combustion efficiency of 99% could be achieved. Some
important ins ights, i.e. "know how", were obta ined but
the CBC had some significant flaws that made it difficult
to continue a test for more than about eight hours. Subsequent
to the work described in this report a second generation CBC
was built after radical changes were made to the FBM. The
new column, which retained the air distributor and fuel
injector of the old, can operate at its design point
continuously.
Some of the problems which made it
the desired parametric studies before the
outiined below.
impossible to perform
program ended are
A new coal feeder, required by the addition of the CBC
to the FBM, performed very poorly until a number of minor
alterations corrected the problems. The new feeder concept
appears to be more practical for commercial use.' Coal feeding
to a fluidized-bed boiler is an area which requires further
development effort.
Two fly ash feeders were tested in the CBC.One design,
termed a "mushroom" feeder, performed well, giving a relativelY
even fuel distribution.
Two bed particle "knockouts" (one water-cooled, the
other uncooled) were also tested in the CBC. These baffles
both consisted of cylinders arranged in a horizontal triangular
array placed just above the fluidized bed. Without a baffle
screen, large quantities of bed material would be lost from the
CBC; with either baffle screen, the quantity of bed. material
carried over was small. When the baffle screen was made up of
water-cooled tubes, an initial heat transfer coefficient of
POPE, EVANS AND ROBBINS
INCORPORATED
-------
11
approximately 35 BtU/ft2 hroF was measured, where the
referenced temperature is that of the fluidized bed. This
was approximately twice the value that would be predicted from
simple radiation and convection, indicating that an active
heat transfer region exists above the dense phase of a high
velocity (hence turbulent) fluidized bed.
When .all of these changes had been made and operational
problems overcome, the FBM achieved an overall combustion
efficiency of approximately 99%.
When finely divided (-325 mesh) No. 1359 limestone was
injected into the FBM, less sulfur was removed than in the
tests described in Reference 1 (approximately 70% at Ca/S =2.5
compared to approximately 80% at Ca/S = 2.5 measured previously).
The reasons for the less favorable performance are unknown but
may have been the result of a different injector design. This
problem was not pursued for reasons brought out later.
1.4
Use of Data Acquired in Both FBC and FBM
Data gathered in coal combustion tests during earlier
work (1) at relatively low temperatures (1500-l7500F) in both
the FBC and the larger FBM allowed an equation for fuel energy
loss, as high carbon fly ash, to be derived. Equation 4 shown
below, links carbon combustion efficiency with the most
significant control variable for combustion of a coal in a "low-
temperature" fluidized bed.
Carbon Combustion Efficiency, % =
71.2 x [Residual oxygen,]
The standard deviation for
data points is S = 4.1%.
%
0.145
(4 )
this equation based on 28
POPE, EVANS AND ROBBINS
INCORPORATED
-------
12
Since it is the carbon combustion efficiency in the
coal-burning regions (or primary c~lls) of the fluidized bed
boiler which determines the quantity and composition of fuel
fed to the Carbon-Burnup Cell, the combustion efficiency of
the system as a whole may be described with Equations 1 and 4.
A parametric study was performed to determine the optimum
split in duty between the primary cells and the Carbon-Burnup
Cells. As shown in the body of the report, the lowest fuel
costs are realized if the Carbon-BurnupCe11 burns about 10%
of the input fuel (i.e., if the carbon combustion efficiency in
the primary cell is 90%). Referring to Equation 4, this is
equivalent to saying that the primary cells should operate with
about 3% residual oxygen in the flue gas.*
*NOTE that the hydrogen and' sulfur burnup efficiencies in the
primary cell are essentially 100%. The "carbon combustion
efficiency" as stated here inc1l.;1des burnup ~o both,C
-------
13
2.
CONCLUSIONS
Based on an analysis of the experimental work and the
performance model derived under this program, the following
conclusions were drawn:
a. A coal-fired, fluidized bed boiler may be
constructed which can operate at relatively high air rates
(~ 800 lbs/hr ft2) and achieve 99¥/o combustion efficiency.
b. The desired level of combustion efficiency may be
achieved by recycling carbon bearing fines collected from the
flue gas of the coal burning regions to a region of the
fluidized-bed boiler in which the bed temperature is in the
range 1950-20500F. The requirements for bed depth, firing rate
and air rate for a Carbon-Burnup Cell may be determined using
the performance model derived in this study.
c. Fuel costs are minimized; i.e., the system is
optimized, when the boiler is designed and operated so that
approximately 90% of the fuel value, fed as coal to the
fluidized-bed boiler, is consumed in the primary cells and
10% in the Carbon-Burnup Cell(s). As will be shown in the
text optimizing fuel costs is equivalent to optimizing system
costs.
d. Calcium sulfate formed in the low temperature (i.e.,
coal-burning) regions of a fluidized-bed boiler by injection
of fine limestone will not decompose in the high-temperature
Carbon-Burnup Cell if the residual oxygen level is maintained
at above 3.5%.
The following are considered valid, significant but
(unfortunately) tentative.
e.
Decomposition of this calcium sulfate can be
POPE, EVAN$ AND ROBBINS
INCORPORATED
-------
14
achieved by operation of the Carbon-Burnup Cell at low levels
of residual oxygen. However, the 9)2 concentration achievable
with fine lime in a high velocity apparatus appears to be
relatively low, approximately 0.5%. No firm conclusions on
-
the maximum 802 level maybe drawn from this work since
attempting to locate the maximum was outside the scope of this
effort. However several patents were located for processes
indicating how higher concentrations could be achieved.
f. While fine limestone injection does not lend itself
to r~generation and reuse of the lime, (no method exists for
separating fine lime from equally fine fly ash) the coal ash
and lime may be reacted within the Carbon-Burnup Cell, to form
uniform agglomerates which may be of commercial value.
No firm conclusions on the requirements of the agglom-
eration process or market potential for the product are available
since these tasks were also outside the scope of this study.
POPE, EVANS AND ROBBINS
INCORPORATED
-------
15
3.
RECOMMENDATIONS
In order to realize the air po11ut.ion control potential
of the fluidized-bed boiler as rapidly as possible, the
following further actions are recommended:
a. A set of cost and performance goals for f1uidized-
bed boilers should be established. This would include air
pOllution control goals on SO , NO , hydrocarbons, CO,
x x
halogens, particulates and plume opacity.
b. Pope, Evans and Robbins, together with one of the
major boiler manufacturers and a public utility, should
perform an engineering design for a large (600 to 1000 MW)
coal-fired, fluidized-bed boiler which may meet these goals.
c. Based on the questions which arise about that
particular design, an experimental program should be conducted
to answer those questions.,
d. If, based on the assessment of an actual design,
OAP's goals will be met, a prototype boiler should be
constructed and operated. The prototype should be large
enough to produce from 10 to 40 MW.
above,
control
e. Pending a decision to proceed with the plan outlined
the following experimental work on the air pollution
aspects of the fluidized-bed boiler should be carried
out:
1) Tests of the S02 Acceptor Process (~) should
be conducted which will determine, for geographically matched
limestones and coals, the required sorbent circulation and
makeup rates.
2) Experiments which indicate the directions in
which reduced levels of oxides of nitrogen may be achieved
POPE. EVANS AND ROBBINS
INCORPORATED
-------
16
should be initiated.
3) A definition of the particulate and plume
opacity control requirements of a fluidized-bed boiler should
be found in cooperation with a leading manufacturer of dust
collection apparatus.
POPE, EVANS AND ROBBINS
INCORPORATED
-------
17
4. .
INTRODUCTION
4.1
Description of the Fluidized-Bed Boiler
4.1.1
Backqround:
Coal
The Need for a New Technique for Burning
Since 1962 Pope, Evans and Robbins has carried out
design and experimental studies (under the sponsorship of the
Office of Coal Research, U. S. Department of the Interior) aimed
at reducing the cost of utilizing coal as a boiler fuel.
The initial studies were concerned, primarily, with
improvements in the design of plants and boilers for industrial
steam generation; i.e., systems which would be large enough to
supply power and/or process heat to a factory, but too large to
be used by a laundry or apartment house and too small to be used
by a large electric utility.
Boilers in this size range were selected for development
for a number of reasons which can be summarized by a single
statement - a novel boiler could be commercially successful in
this size range with less development expense. Success at this
level could then lead to scaling both downward and upward.
Conventional methods of firing coal, as a fixed-bed
on a stoker grate or as a suspension flame in a pulverized fuel
burner, were not found to hold promise for major reductions in
size and cost regardless of development effort.
It was found that the cost of an oil or gas-fired boiler
was low when compared to coal-fired boilers with the same
steam capacity primarily because of differences in furnace
size. Oil and gas could be burned in a smaller furnace than
could coal. For industrial sized boilers this difference
POPE, EVANS AND ROBBINS
INCORPORATED
-------
18
meant that oil and gas-fired boilers could be assembled in a
factory and shipped to the user's site by rail, while coal-
fired boilers could not. The coal-fired boiler would have to
be assembled piece-by-piece, often out-of-doors, at the user's
site and by workers who sometimes required close supervision.
A method of firing coal was therefore needed which would reduce
the size of the furnace sO,that large coal-fired boilers could
also be factory assembled.
An evaluation of alternatives led to the selection of the
fluidized-bed boiler as the most promising method of aChieving
the goals of our sponsor.
The fluidized-bed boiler would be smaller because a
higher volumetric heat release rate and higher heat transfer
coefficients could be achieved in a fluidized-bed than in a
conventional furnace; i.e., the coal would burn more intenselY,
in less space, and with faster heat removal.
Therefore, in 1965 a development effort was begun, again
under the sponsorship of the Office of Coal Research (OCR). By
1966 a primitive experimental boiler had been built and operated
thus proving that fluidized-bed combustion could be used to
develop a commercial boiler. The results were sufficiently
promising to warrant an actual fluidized-bed boiler being built.
The construction of the FBM(See Figure 7 and Section 5.2) was
planned and the scope of the study portion of the effort
expanded to include conceptual designs for the much larger
boilers required for central power stations. (See Reference
4 for a description of this work.) A major incentive here was
the fact that the fluidized-bed,boiler could burn poorer grades
of coal than could the pulverized coal boiler. The other
advantages of the fluidized-bed boiler which would apply to
POPE, EVANS AND ROBBINS
INCORPORATED
-------
19
utility size units as well as the smaller industrial units
were the high heat transfer coefficients, leading to reduced
tubing requirements, and the high volumetric heat release
rates leading to a smaller fUrnace.
Also in 1966 dolomite was mixed with coal being fed in
an experiment with the FBC (See Figure 2 and Section 5.1).
Now another reason that this new technique for burning coal
was needed became evident-- pOllution control.
Representatives of the National Center for Air Pollution
Control (now the Office of Air Programs) observed some of these
first crude experiments and in 1967 began to support the air
pollution control aspects of the fluidized-bed boiler.
4.1.2
Defininq a Fluidized-Bed Boiler
Almost from the beginning of fluidized-bed combustion
(2, 6) an extensive patent literature has been developing on
techniques for using this new tool to generate steam from
fossil fuels. (Reference 7 is believed to be the earliest U. S.
patent of this type.)
Only the effort by Albert Godel (~), out of over twenty,
has been commercially successful and this exception is generally
known as a fluidized-bed boiler although it differs substantially
f rom the. system be ing discussed in th is report.
To avoid confusion with Godel's design Pope, Evans and
Robbins referred to its design as "The Direct Contact Heat
Trans ferr ing Fluid ized-Bed Boiler" (.2.). This ted ious nomenclat ure
has been dropped because the definition of boiler "the part of a
steam generator in which water is converted into steam" cannot
be applied to Godel's most significant contribution: the only
commercially successful fluidized-bed furnace.
POPE, EVANS AND ROBBINS
INCORPORATED
-------
20
So as to make very clear what is meant in this report
by "fluid ized-bed boiler" the fOllowing de fini tion is presented.
A fluidized-bed boiler is defined as a system which
meets each of the following criteria:
a) The system's primary function is the generation of
steam. Therefore, the materials of construction, the mode of
operation, the arrangement, auxiliary power requirements, etc.,
are consistent with existing practices and economics in the
conventional boiler field.
b) The fuel i$ added to and burned within a turbulent
bed which has been termed a fluidized bed.
c) A significant fraction of .the heat released by the
burning fuel is extracted by heat transfer surfaces in contact
with the turbulent bed.
A fluidized bed, in turn, is defined as a mass of
particulate solids held in suspension by an upward current of
fluid such that the bed has zero angle of repose and exhibits
certain other properties of a liquid. Among the liquid-like
properties of a fluidized bed which are important to the boiler
designer, are that the bed becomes well mixed, with sufficient
agitation, and the bed material can be caused to flow about or
out of the system without the aid of mechanical devices.
4.1.3
A Simplified Description of a Fluidized-Bed Boiler
A fluidized-bed boiler consists of an enclosure contain-
react with coal which will
layer, or bed, of granular solids.
is perforated so that air may enter
solids to become suspended, and to
be added to the bed. Such a system
ing both boiler tubes and a
The bottom of the enclosure
the enclosure to cause the
is shown schematically in Figure 2.
In addition to the
POPE. EVANS AND ROBBINS
INCORPORATED
-------
21
FLUE GAS TO
STACK
ICONTl\CTED"
HEAT
EXCHl'.UGE
SURFACE
STEAM
COAL SUPPLY
FLUIDIZED-BED
COMBUSTION ZONE
FEEDh'ATER
IIVIr:~~CD II
EXCHANGE
FACE
r.'~.~... ..,.,
J..L...':-1..l
SUR-
.-.
'.. ~ .
BOILEH. TUBES
IN FORr.! OF
~ ~vATER \'li\LLS
FUEL
AIR
.
IN~ l:'"''
o l
~--
,I,. ~.. .....'
. ,
. ., .
. .
. .
::-
FAN
AIR DISTRIBUTION
GRID
COMBUSTION
AIR
PLENUM
CHAMBER
FIGURE 2.
SCHE~~TIC OF FLU!OIZED-BED BOILER
POPE. EVANS AND ROBBINS
INCORPORATED
-------
22
components
exchangers
combustion
shown in the sketch, there would be convective heat
and a system to remove particles carried out of the
space with the flue gas.
When the temperature of the bed is raised by an auxiliary
means, such as a gas burner, to above about 800oF, bituminous
coal added to the bed will ignite. After ignition the
temperature will rise until the system achieves a thermal
equilibrium; i.e., the energy added to the bed by the burning
fuel precisely equals the energy extracted by the boiler tubes
contacting and viewing the bed and by the gases and dusts
leaving the bed. Because the coal burns very rapidlY the
fluidized-bed consists, almost entirely, of inert particles
with a small quantity of coal. Typically a sample removed from
the bed would be from 0.1% to 1.0% carbon by weight.
, 0
The equilibrium temperature may be as low as 1200 F or
.0,00
as hlgh as 2500 F, although a narrower range, 1450 F to 2050 F,
is of practical interest. Below about 14500F it is difficult
to consume the carbon monoxide or to prevent refractory hydro-
carbons from appearing in the flue gas. Above about 20500F the
bed particles may agglomerate into large pieces and, the bed
collapse.
Assume that the simplified fluidized-bed boiler shown
in Figure 1 is operating with a bed temperature of l500oF. To
increase the bed temperature additional coal and air might be
added but it is more effective to reduce the depth of the bed
by removing some of 'the bed material. This may not seem
obvious but it should be recalled that the temperature at
which the bed operates is that temperature at which the heat
losses (which are proportional to the bed's temperature)
precisely equal the heat input by the burning fuel. To
increase the bed temperature one may either add more heat or
POPE. EVANS AND ROBBINS
INCORPORATED
-------
23
remove less heat.
If the bed depth is reduced less of the boiler tubes
are contacted by the bed and less heat is removed by these
tubes. Since the heat input has remained constant the only
way in which the heat losses can equal the heat input is for
the bed temperature to rise. Then most of the extra heat
available can be carried off by the hotter gases and dust.
The amount of heat transfer surface in the furnace has
not been changed; but a tube which is contacted by the bed
removes more heat than a similar tube which only views the bed.
In a commercial fluidized-bed boiler the bed temperature
will not be changed and the bed depth will remain constant but
the preceding example, hopefully, illustrates how bed depth and
heat transfer surface are related and how they are used in design.
The design depth of the fluidized-bed may vary over a
wide range - from a few inches to many feet. However, practical
systems must 'operate in the range of about 12 to 47 inches.
Below about 12 inches the combustion efficiency degrades and.
will become unsatisfactory; above about 48 inches the power
required by the blowers becomes excessive since the power
consumption will be proportional to the depth of the bed.*
Having set the operating temperature and bed depth in
the fluidized-bed boiler, the energy release rate still remains
unspecified. Rather than specify an energy, or heat, release
rate, it is more appropriate to specify an easily measurable
*Fluidized-bed boilers might be incorporated into combined
cycles in which the hot flue gas, at high pressure, is passed
through a turbine which is used to compress the combustion
air as well as drive an alternator. In this system the
economics are not so sensitive to bed depth. Supercharged
boilers might logically evolve from the work described here.
POPE. EVANS AND ROBBINS
INCORPORATED
-------
24
equivalent, the mass flow rate of combustion air, in pounds
of air per hour per square foot of grate area. In conceptual
designs air rates between 100 lb/hr ft2 to over 2,000 lb/hr
ft2 have been considered. Practical systems will operate with
between 400 and 1200 lb/hr ft2.
Specifying the air rate is the equivalent of specifying.
the heat release rate because in a fluidized-bed boiler the
rate of fuel injection is almost precisely set by the air rate.
This statement assumes that a practical excess air level of
about 10 to 20% is used and that the fuel is known. For every
oil or coal there is a proper, albeit different, fuel-to-air
ratio. The point of specifying air rate as the primary factor
is brought out in the following paragraphs.
From a specification of the air rate, the bed depth and
bed temperature plus the energy release per unit air, it is
possible to compute the apparent plan and vOlumet,ric heat release
rates, the superficial gas velocity and residence time and
other parameters of interest. For example' an air rate of 1000
lb/hr ft2 is usually the equivalent of a superficial gas
velocity of 12 to 15 fps, a plan heat release rate of about
106 Btu/hr ft2 and, in a five foot high furnace, a volumetric
heat release rate of 200,000 Btu/hr f't 2.
The a ir rate se lected determines the size and dens i ty
distribution of the particles which will make up the bed.
Particles above a certain maximum size and density will sink
to the air distributor. Particles below a certain minimum size
and density will be entrained in the gases leaving the bed and
will be carried out of the system. The particles which make
up the bed may be any noncombustible, granular solid which is
sUfficiently tough to retain its shape and size in the bed
over an extended period.
POPE, EVANS AND ROBBINS
INCORPORATED
-------
25
In some coals a portion of the ash, (small stones,
shale particles, etc.) is coarse enough and strong enough to
remain in the bed and the coal can be burned in a fluidized-
bed composed of its own ash. In fact, for this type of coal
bed material must be continuously removed or the bed depth
will increase. For other coals the ash may be too fine to be
retained in a fluidized-bed, even if the coal itself is quite
coarse, unless the ash is deliberately sintered. For the
coals which cannot maintain their own bed, material must be
added with the coal to make up for particles of the starter
bed which are lost to the system through attrition. This
material, besides ~eeting the properties listed above, must
also be inexpensive. In many areas limestone may be the
material of choice because of its cost and relatively low
density* whlch permits a relatively deep bed to be used with
a moderate pressure drop through the bed. Limestone possesses
other properties of interest to the designer of a fluidized-bed
boiler and these are discussed in the following paragraphs.
4.2
Air Pollution Control Potential of Fluidized-Bed Boilers
4.2.1
Pe rspect i ve
Fluidized-bed combustion of coal is, in itself, not a
remedy for air pOllution; a fluidized-bed boiler is not
a pollution control device. Howev~r, certain properties of
fluidized-bed combustion and of a properly designed fluidized-
bed boiler can be exploited to produce a steam supply which is
*The limestone is dense but it soon becomes .lime when it is fed
into a fluidized-bed boiler. This report uses the term lime-
stone for lime, or calcined limestone, when precise definition
is not required.
POPE. EVANS AND ROBBINS
INCORPORATED
-------
26
"clean" at less cost than ava i1ab1e a1 ternat ives .
What follows is intended ,to provide an overview of
the air pOllution control potential of the fluidized-bed
boiler in the view of the authors of this report.
A report (to be published for EPA in 1972) by
Westinghouse Research Laboratories entitled "Evaluation
of the Fluidized-Bed Combustion Process" contains a summary
of all of the available data and should also be read by
anyone interested in the state-of-the-art.
4.2.2
Discussion
Air pollutants are what t0e laws define as air
pollutants. For steam generators (boilers), air pollutants
proposed and promulgated in 1971 are sulfur oxides, nitrogen
oxides, the mass of the particulates and ~he plume opacity;
i.e., the "look" of materials leaving the stack.
It is likely that before too many years have elapsed,
rules will be established limiting the amount of arsenic,
cadmium, mercury and other trace pollutants (e.g., beryllium)
which can be emitted and just as for the automobile, carbon
monoxide and hydrocarbons will be limited by law. For hydrogen
chloride and chlorine, which have been mentioned, limits may
be established some day.
A fluidized-bed boiler may be able to meet most of these
regulations because of certain inherent characteristics of the
fluidized bed combustion process.
Carbon monoxide and hydrocarbons are always produced
while coal is being burned in a limited space. The FEC
program was not undertaken primarily to control CO and hydro-
carbons, since conventional utility boilers also have low
POPE, EVANS AND ROBBINS
INCORPORATED
-------
27
emissions of these pollutants due to their large volume.
They enter the atmosphere when the burning process is not
complete. By careful design, partial combustion products
can be consumed within the fluidized-bed and in the freeboard.
And although no work appears to have. been done on methods of
reducing CO and C. H (beyond decreasing the coal:air ratio),
x y
emissions on the order of 100 ppm for each can be anticipated
for fluidized-bed boilers. Additional research may provide
methods for a further reduction in emissions. A low-cost
combustion catalyst approach may be economical
bed boiler, but is less likely to be practical
boilers.
in a fluid ized-
for conventional
By the use of a tough limestone as the bed material,
it is possible to absorb virtually all of the sulfur released
by the coal. To achieve very low levels of 802 emission
requires that a very.large excess of gctive lime be present
in the bed and freeboard at all times. Economics, in turn,
require that the limestone be kept active by continuously
stripping off much of the sulfur so as to regenerate the
limestone. The regenerator off-gas can then be processed to
recover sulfur or sulfur products or be scrubbed, using a
portion of the lime produced by the boiler. The property of
a fluidized-bed boiler which makes effective in situ sulfur
control possible is the relatively low temperature of the
medium in which the combustion occurs. The bed may be kept
cool by immersing heat exchange surfaces in the combustion bed
or, alternatively, by circulating the sorbent bed through a
separate heat exchanger. In addition, the re fluxing of sorbent
particles in the freeboard provides a degree of concurrency
to an otherwise well stirred system. Additional development to
provide countercurrency may make .it possible to limit sulfur
POPE. EVANS AND ROBBINS
INCORPORATEO
-------
28
oxides emissions to 50 ppm.
If a wet scrubber were added to
control particulate emissions, the sulfur oxide emissions
might be so low that the boiler could not longer, reasonably,
be considered a source of sulfur oxides.
Nitrogen oxides are produced in a fluidized-bed boiler
through the fixation of atmospheric nitrogen and the
oxidation of nitrogenous compounds' in the coal. * The
nitrogen in coal is converted principally to gaseous N2.
Most coal-fired boilers produce less than 1000 ppm NO .
x
Because of the relatively low temperature of the
combustion medium (about l5000F compared to 30000F in the
conventional flame), it had been anticipated from equilibrium
data on the fixation of atmospheric nitrogen that NO from a
x
fluidized-bed boiler would be very low compared to conventional
bpilers. However, this was not found to be the case, apparently
due to oxidation of nitrogenous coal comp~unds. Measured values
of NO from most fluidized-bed combustors have usually been only
x
moderately lower than emissions from conventional coal-fired
boilers.
Little work has been done on methods of reducing NO
, x
emissions from a fluidized-bed boiler, but it is hoped that
research to be carried out in the' future will lead to means
for significantly reducing NO
x
emissions.**
*Argonne National Laboratories, in an experiment
an "artificial air" where argon was substituted
nitrogen, demonstrated that nitrogen oxides can
the nitrogenous compounds in the coal (10).
utilizing
for
arise from
**Pope, Evans and Robbins ,showed, in a recent test series,
that more NO resulted from combustion of 3/4" coal than
from combustion of 1/4" coal. If only fine coal is burned,
the carbon inventory in the bed is minimized.
POPE. EVANS AND ROBBINS
INCORPORATED
-------
29
For the control of particulate
fluidized-bed boiler possesses certain
over most conventional boilers. Since
emissions, the
inherent advantages
the coal fed to the
fluidized-bed boiler is crushed and not finely ground, some
of the larger pieces of ash termed non inherent will not be
entrained in the gas at all. Further, most of the particles
which are entrained may be removed by a relatively inexpensive
mechanical dust collector. Such a collector, while inefficient
on a pulverized-fuel furnace, is efficient on the ash from a
fluidized-bed boiler.
Inherent ash, on the other hand, may be very fine even
if the coal is very coarse* and will be entrained by the gas.
Some of this fine material will pass through a mechanical dust
collector and could be removed by an electrostatic precipitator.
This final collection is potentially difficult when sulfur
oxides** in the gas are low. However, in a fluidized-bed
boiler in which a Carbon-Burnup Cell is used, the high carbon
content of the fly ash from the primary cells should act as a
natural resistivity conditioning agent. This should allow
*About 10 pounds of coal was taken from a pile. The largest
pieces were over 1". The coal was put in an ashing furnace
and heated carefully so that the ash would not melt. The
ash was then carefully screened and weighed. About 1/4 of
the ash was larger than 1/2" with one piece 1", about 15%
was between 1/20" and 1/2". The rest, 60%, was less than
1/20" and most was as fine as face powder.
**It has been shown that, although the electrical resistivity
of most of the ash matter (silica, alumina, and ferric
oxide) is high at conventional collection temperatures,
sulfur oxides which would normally be in the gas reduce the
resistivity to the point where the electrostatic precipitator
can function efficiently.
POPE. EVANS AND ROBBINS
INCORPORATED
-------
30
efficient dust removal at conventionally low temperatures
from the gas leaving the primary cells. The dust fed to the
Carbon-Burnup Cell could be caused to sinter into larger
pieces and that which is emitted could be removed by a bag
filter without a significant cost penalty (less than 1% of
the plant costs).
A second standard for particulates is concerned with
the opacity of the plumes issuing from the power plant's
stacks. "Smoke" control laws were the first promulgated *
and may be the most difficult to meet in the long run.
Standards now proposed would require that the emission be
"invis ible". While the opacity of the plume is related to
the weight of particulates, it is'not congruent with weight.
A relatively opaque plume may result, .for example, when
firing conditions result in incomplete combustion'causing
micron-sized carbon particles to form from cracked hydro-
carbons. Gas turbines produce smoke through this mechanism.
This will not occur in a properly designed fluidized~bed
boiler.
A second source of light-scattering particles (causing
the plume to be visible) is sulfur trioxide droplets formed
as the dew point is reached in the gas stream. When the bed
of the fluidized-bed boiler contains limestone, the quantity
of sulfur trioxide is vanishingly small.
A third source of small particles is the inorganic
fumes and smokes which result when mineral matter is
volatilized during combustion.
Because of the low temperature
*One of the ea,rliest smoke control laws provided for beheading
of violators. This penalty has not yet been suggested as the
solution to our current pollution problems.
POPE. EVANS AND ROBBINS
INCORPORATED
-------
31
(about 1500oF) of the combustion medium in a fluidized-bed
boiler, less mineral matter will volatilize than in a
conventional boiler flame. As much as 99% of certain
normally volatile species can be retained in a f1uidized-
bed boiler operating at low temperatures.
Mercury has been declared a hazardous pollutant.
Standards for CL2 - HCL, among others, are being contemplated.
The pollutants over which there is no current (1971)
concern but which may soon be considered are heavy metals and
radioactive isotopes. There is some evidence that arsenic
and beryllium would be retained in a bed of lime, but cadmium
and mercury would not be retained. Radioactive species
should be lower from the fluidized-bed boiler than from a
pulverized fuel unit but little thought has been given to
this. Chlorides have not been measured but a plausible set
of chemical reactions and some intuition suggests that they
may not appear as gases from a fluidized-bed of lime.
From the above it may be seen that the air pollution
control potential of the fluidized-bed boiler is quite
promising. This potential follows from certain inherent
characteristics: the low temperature of the bed, the low
temperature of the combustion, the high carbon content of
the fly ash in the main flue gas stream, the stability of
the combustion, the ability of lime to react with volatile
inorganic materials to form stable solids or liquids.
The first generation of fluidized-bed boilers may not
achieve all of the goals set in the preceding paragraphs.
However, by the careful and ingenious application of tools
POPE. EVANS AND ROBBINS
INCORPORATED
-------
32
made available by this new form of
be possible to. evolve designs that
economical power.
steam raising it may
will provide clean and
4.3
PO}:e, Evans and Robbins' Prior Work
4.3.1
General
Pope, Evans and Robbins, under contract with the Office
of Air Programs (OAP) of the Environmental Protection Agency
(formerly the National Air Pollution Control Administration,
Department of Health, Education .and Welfare), has
characterized emissions from a fluidized-bed boiler developed
for the Office of Coal Research, Department of the Interior.
Two reports have been prepared describing this work
(References 1 and 4).
The pollution control aspects of a fluidized-bed
boiler had been considered as early as 1965 when it was
discovered that if coal distribution was uniform, smokeless
combustion was achieved at very low excess air levels.
As noted before, in 1966 a test was conducted in which
dolomite was mixed with the coal and the sulfur oxide control
potential of the process was demonstrated. A literature and
patent search revealed that the use of limestone in a fluidized-
bed combustor to control sulfur oxides from a coal-fired
boiler was nove~, though a patent had been granted earlier on
the use of limestone in an analogous process.*
*The use of limestone in a coal-fired, .f1uidized-bed boiler
should not infringe this existing patent which was narrowly
drawn.
POPE, EVANS AND ROBBINS
INCORPORATED
-------
33 '
4.3.2
Once-Throuqh Limestone Injection
Based on this early work, the initial studies conducted
for OAP investig~te,d the use of relatively coarse limestone
and dolomite (crushed to -7+14 mesh) on a once-through basis.
Once-through is the term used to describe a non-cyclic or
non-regenerative process. The limestone is 'added to the bed
and allowed to blow away or overflow and is not reused.
Although in a fluidized-bed boiler the dust which blows out
of 'the primary cells is returned to the Carbon-Burnup Cell,
it would still be a once-through process if the sorbent is
not regenerated and reused.
The bed in these tests was primarily coal ash and the
air rate was chosen to yield a nominal superficial velocity
of 10 to 15 feet per second. The coarse limestone particles
fed to the bed would not e1utriate at these velocities until
they were reduced in size by attrition. It was determined
that coarse stones would not retain nonregenerative1y more
than about 30% of the sulfur released by the coal at
reasonable rates of sorbent addition.
Each test in this series was run long enough to assure
that a steady state value of the sulfur oxide emission' had
been achieved, no more than two or three hours and often much
less. Depending on the coal and stone used, the physical
composition of the bed mayor may not have been at steady state.
If a tough limestone and a well washed coal were used
the steady state bed, almost all limestone, would take over a
hundred hours to establish. With the dolomite used in these
tests, which decrepitated easily, the steady state bed, which
contained little dolomites was achieved in only a few minutes.
POPE, EVANS AND ROBBINS
INCORPORATED
-------
34
One unexpected result of these first tests was that
the capture rate decreased with increasing bed temperature.
Bench-scale studies by others in which S02 was removed from
a gas stream by a fixed bed of lime had indicated more rapid
o . 0
capture at 1800 F than at 1500 F. It had been assumed .that
the injection process into a fluidized-bed boiler was
kinetically limited, yet the capture of sulfur was far better
at l6000F than at l8000F.
In order to enhance the process kinetics, the stone was
finely ground to pass through a 325 mesh screen. Although
such a fine powder is rapidly elutriated f~om the bed at the
air rates used here, the efficiency of sulfur capture was
much higher with the finely ground sorbent. It was determined
that about an 80% reduction in 802 emissions could be achieved
by injecting a finely-ground limestone (identified by OAP as
No. 1359) at Ca/S = 2.5. Raw stone and hydrate performed
equally well, while stone calcined by the supplier performed
poorly.
4.3.3
Reqenerative Limestone Process
Towards the end of
(May 1969), we found that
No. 1359 limestone could
the pOllution-characterization effort
beds composed almost entirely of the
be made to release the sulfur that
had been captured during the first few hours of a test in just
a few minutes. This was done by increasing the coal-feed rate
so as to increase the. bed temperature and decrease the oxygen
content of the flue gas. A regenerative cycle was thereby
devised and a pa tent appl ica t ion prepa red for the "SO 2 Acceptor
Process" C}). This process was particularly attractive for
application to a fluidized-bed boiler, while the previous
POPE, EVANS AND ROBBINS
INCORPORATED
-------
35
state-of-the-art for decomposing calcium sulfate either
damaged the limestone so that it could not be reused, were
costly to apply, or exchanged pollution as carbon monoxide
and hydrocarbons for pollution as sulfur oxides. This
discovery was also felt to provide an explanation for the
anomalous temperature behavior of the lime-sulfur oxide
reactions noted earlier. In this hypothesis, sulfur retention
in a fluidized-bed combustor was seen as an easily reversible
process, sensitive to temperature, oxygen partial pressure and
the presence of reducing gases.
4.3.4
Pollutants Other than S02
Pollutants other than sulfur dioxide were character ized
--nitrogen oxides, hydrocarbons and particulates.
Oxides of nitrogen were found to increase as the amount
of oxygen remaining .in the flue gas increased. Experiments
in which some air was diverted from the base of the bed to
a port above the bed resulted in a marked reduction in NO.
This might indicate that increased reducing gases in the 'bed
were decomposing the NO. Nitrogen oxides were not materially
and reproducibly affected by the injection or fine limestone.*
No levels of less than 400 ppm were found in this first
test program; 275 ppm was typical.
*Other workers have shown a reduction in NO with limestone
injection (10). An increase in NO with limestone injection
has also been predicted on the ba sis of experiments at Esso
Research and Engineering (11). These apparent contradictions
indicate that the statement - fluidized-bed combustion of. coal
is not an adequate description of the process. Each
experimenter's reactor must behave differently. Since no
attempt has been made to quantify these differences (for
example, bubble frequency and diameter), an understanding of
how the various results relate to boiler design parameters is
not available.
POPE, EVANS AND ROBBINS
INCORPORATED'
-------
36
Hydrocarbons and, presumably, carbon monoxide were
found to be sensitive to the quality of the fuel distribution
and to the coa1:air ratio. With between 3% and 4% oxygen
remaining in the flue gas, the hydrocarbons were reduced to
below 100 ppm. Hydrocarbons were not affected by fine
limestone injection.
The .major fraction of the mineral matter in the coal
normally appears as fly ash. The larger particles of non-
inherent ash remain in the bed. The starter bed material adds
some particulate matter to flue gas, and when fine limestone
is injected, essentially all of this appears in the fly ash
stream.
At the high dus.t loadings, when fine limestone was
injected, a low-pressure-drop It!echanica1 collector removed
about 90 wt. % of the fly ash. Without limestone the
collector was 90";6 efficient.
Therefore, with or without
limestone injection, the dust emission was on the order of
1.5 pounds of particulate per million BTU of fuel input.
More efficient dust collectors are anticipated for a commercial
application.
The particles not removed by the collector were all
smaller than 20 microns. It is these particles which would
have to be removed by an electrostatic precipitator, bag
collector, or wet scrubber.
Recommendations Based on Prior Work
4.3.5
The previous phase of the program was ended with
recommendations for further study of the most promising
leads. Sixteen of these recommendations were listed in
Reference 1.
One of these sixteen recommendations was the
POPE, EVANS AND ROBBINS
INCORPORATED
-------
37
characterization of the carbon burn up cell, a region of a
fluidized-bed boiler in which carbon-bearing fly ash is the
primary fuel, as a means of promoting high combustion
efficiency. This suggestion formed the basis of the work
reported here. It was carried out during the period October
1969 through October 1970.
4.4
Prior Work Leadinq to the Burnup Cell Concept
Economics demand that commercially viable fluidized-
bed boilers will have superficial gas velocities (primary
cell) in ~he 5-10 ft/sec range; otherwise the size and cost
of the unit is excessive. Economics also demand that
commercially viable fluidized-bed boilers will burn essentially
all the coal delivered to the boiler; i.e., f.ine coal as well
as 3/4" coal. One of the economically attractive features of
the fluidized-bed boiler is low coal preparation cost, since
the coal need not be finely ground.
Bench scale studies by others have indicated that if
freeboards are excessively large and superficial gas
velocities are unrealistically low (e.g., 2 ft/sec) there is
less need for a carbon burn up cell since the primary cell
effluent carbon content is low by virtue of long residence
time in the primary cell.
Authorities are generally agreed that if only large
coal lumps (e.g., 3/4") were fed to a fluidized-bed boiler,
the need for a carbon burn up cell would be less, by virtue
of long coal residence time in the bed. Authorities are
generally agreed that there is no realistic way to feed
3/4" size coal lumps to the fluidized-bed boiler, since
of the heating value in a 3/4" top size lot of coal must
recovered, and metering screw feeders cause additional
only
all
be
attrition.
POPE, EVANS AND ROBBINS
INCORPORATED
-------
38
Particles of coal, like the other
up the fluidized bed, will appear in the
because they are small and are elutriated
because they are thrown out of the bed by
For ~urther discussion of this topic, see
particles which make
freeboard either
from the bed, or
violent agitation.
Appendix L.
Based on this discussion, it is clear that if particles
of coal, fed to the bed, have not been totally consumed before
they come to the top of the bed, some coal particles will
appear in the freeboard. It is also clear that the freeboard
is an integral part of a fluidized-bed system. It is also
clear from this discussion that unless the coal particles
return to the bed or are totally consumed in the freeboard,
unburned carbon values will be chilled by contacting cooled
surfaces and will leave the furnace.
An examination,' over a
period of several years, of possible ways to utilize these
carbon values, led to the development of the two-stage
combustion system incorporating a carbon-burnup cell.
4.5
Specific Obiectives of This Work
As noted in Section 1, SUMMARY, the primary objective
of this work was to demonstrate the Carbon-Burnup Cell
concept and to develop a design correlation which would aid
in the design of an actual Carbon-Burnup Cell.
A second objective was the derivation of a technique
for optimizing the performance of a coal-fired, fluidized-
bed boiler. Basically this requires a method to select the
division in duty between the primary coal-burning cells and
the flyash-burning Carbon-Burnup Cells so as to maximimize
the boilers' thermal efficiency.
Some specific factors to be considered included;
POPE. EVANS AND ROBBINS
INCORPORATED
-------
39
injection of limestone into a burnup cell; the effect of the
high temperature of the cellon partially sulfated limestone
present in the flyash, especially decomposition; and the
effect of fine limestone by-products on combustion efficiency.
It was not found practical to study the effect of
horizontal heat transfer tubes within the bed of the high
temperature burnup cell. The experimental apparatus was too
small to permit a realistic tube array to be installed without
absorbing an excessive fraction of the combustion energy.
Two movable vertical water-cooled tubes were used to vary bed
temperature so that some insights were achieved on the effects
of heat removal. A larger apparatus, in which wall losses
were relatively smaller, or a small apparatus with heated
walls might each be useful for studies with horizontal tube.
bundles.
Design criteria for a fluidized-bed Carbon-Burnup Cell
were to be provided. The effects of the burn up cellon the
design and economics of Pope, Evans and Robbins' Multicell
Fluidized-Bed Boiler concept were to be identified. The
study was to be conducted in two parts~ First, a statistical
experimental program on the FBC. Second, a program of
confirmation on the FBM. The scope of work prepared by OAP
outlined the variables to be considered and these, as well as
others, are discussed in some detail later in the report.
(See Section 6.2.2).
POPE. EVANS AND ROBBINS
INCORPORATED
-------
40
5.
APPARATUS AND MATERIALS
5.1
Pilot Scale Combustor, FBC
The majority of the tests were conducted in a pilot
scale combustor, designated Fluidized-Bed Column (FBC). The
FBC consists of a water-walled combustion chamber with plan
dimensions of 12" x 16". A flat air distributor forms the
bottom of the chamber and a vertical 1288 diameter duct forms
the outlet.
3 and 4.
Front and side elevations are shown in Figures
In operation, air at ambient temperature flows into
the plenum below the air distributor, up through the holes
in the distributor and into the combustion chamber where it
fluidizes the bed material and provides the oxygen for
combustion. The bed material for the present tests consisted
of sintered coal ash crushed and screened to the desired size.
Fuel, coal and/or carbon-containing fly ash, is injected
through a port at the base of the bed.
The air distributor contains a matrix of stainless
steel bubble caps (buttons) mounted in a'mild steel plate.
A typical button is shown in Figure 5.
The bed temperature is monitored with a number of
thermocouples spaced vertically in the combustor. Kaowool
seals were provided between flanges to prevent flue gas.
leakage out of the system. Specifications for the FBC are
presented in Appendix A.
The fuel feed system is capable of delivering
. .
approximately 250 pounds of coal each hour which would be
sufficient for a firing rate of 3.5 x 106 Btu/hr. The air
feed system is capable of delivering oxygen sufficient for
POPE, EVANS AND ROBBINS
INCORPORATED
-------
o
--1
-- I
- - - -A- -...
- - - - - -\::.J I ,
-' - - .- .-'" I 1-...1
- -.." I""""
- - - - , - --;0 : FEED PORT
-- I
-- I I
- I
I ---
START-UP
COAL FEED
SCREW ~
FLY ASH
FEED SCREW \ .'.
FLUID
BED
ASH RECIR-
CULATION
PORT
GRID
BUTTONS
PLENUM
CHAMBER
WELDED SEAM DUCT
41
SIGHT PORT
WATER COOLED HOOD
WATER JACKETS
KAOWOOL GASKET
THERMOCOUPLE PORTS
o
WATER WALLED COLUMN
AIF~
LIGHT-OFF
GAS BURNER
FUEL INJECTION
AIR LINE
--~--- . ~
, =-»
j ..,~ 2 AUXILIARY FEED PORT
(ONE RIGHT, ONE LEFT)
-4
COMBUSTION
AIR INLET
~,i. Ii: 4o,,!o).'I-:':" ,",' ':';',!'f:.'I"':,~'. '~~\:~, ...":.,' ","":"''':';'h,.ti.r-,; " .
IJ4 ...'; r~' "'..:. ~-':!' ,"t'..j 'p. :: . . .~.... :. .-'. '-'~ tI "",...J...- ...;-;. \ /1 ~" 'i.A I"..e, .'..~ 0' 1&1
r":t".~.~.... ",'''.''''''.'':,... u: . ,,':~.':~:".t.~~~ ""''''''.';;~'
FIGURE 3.
FLUI DIZED- BED COLUMN (FBC) CONSTRUCTION DETAIL
FRONT VI EW
POPE. EVANS AND ROBBINS
- ,"'.1<:'(2 r=:2'C: ~l;."'->~::=[,
-------
42
WATER WALLED
COlUMN
FLY ASH (FUEL)
FEED SCREW
START-UP,
COAL FEED
SCREW
..-...
, \
" . \
I ,
\ ,
" ,
... ."
AIR
INLET
"-..'.. '... .,. ... t .....' ',.,','.' ., .'....~ ,1,,,--,
,. ","~'. ...'~'. ~.. -.. ",. ...."'..1'. ,.'! -"
FIGURE 4 FLUIDIZED-BED ~OLUr.~N (faC)
CONSTRUCTION DETAIL -SIDE VIEW
POPE, EVANS AND 'ROBBINS
- ~ IE:
WATER COOLED HOOD
LIGHT-OFF BURNER PORT
THERMOCOUPLE PORTS
WATER WALL
AUXILIARY FEED
PORTS (2)
PLENUM CHAMBER
-------
J
1
0.078" dia.
(Typical for -p)
FIGURE 5.
45°
45°
.5 II
8
.1
':- ~.r;. 5 0
-.
AIR DISTRIBUTION GRID BUTTON
POPE, EVANS AND ROBBINS
I~ICORP()R.O.TED
43
TOP VI E'i'l
SIDE VIEI'l
-------
44
a heat release of approximately 2 x 106 Btu/hr. These are
both in excess of the actual operational values, since the
FBC's performance as a fluidized-bed combustor deteriorates
at firing rates much in excess of 1.2 x 106 Btu/hr. The
water walls of the FBC will remove on the order of 0.5 x 106
Btujhr from the fluidized-bed. As will be shown, it is not
possible to sustain the fly-ash combustion process with that
level of heat loss. It was therefore necessary to reduce the
energy removed by the water-walls, and the FBC was insulated
internallY, as shown in the partial cross-section of Figure 6.
The water-cooled hood was insulated in a similar manner. The
insulation consists of a sleeve, or liner, of ASTM 446, one
of the most refractory steels, backed with 1" of Kaowool, a
refractory insuiation. with the insulated sleeve, the FBC
has an internal cross section of 0.86 ft2. With the.
insulation in place, high bed temperatures, in the l8000F
to 2l000F range, could be achieved even with relatively deep
beds, up to a 22" static depth.
The temperature of the bed could be controlled by the
use of cooling surface in the form of bayonet tubes, also
shown in Figure 6. The temperature of the bed could be
decreased by lowering one or more bayonets. Four, fully
inserted, would provide 4 ft2 of heat transfer surface.
The bed material consisted of sintered coal ash,
crushed and screened to either -16 +22 or -8 +16 U.S.
Standard sieve size. The size was selected on the basis of
the air rates to be used in a particular test; the largest
particles for the highest air rates.
To begin a test, the bed is heated to coal ignition
temperatures with a premix gas burner flame directed downward
onto the bed as shown in Figure 3. The ignition procedure
POPE. EVANS AND ROBBINS
INCORPORATED
-------
COOLING
WATER
L
OUT
II II
II II
II II
'I : I
II II
II II
II II
I, '.
II 'I
II II
. (I II
,I COOLING III
I I
'I SU!1FACE II
:' TRAVEL II
I ,I
I, II
I
L! t!
45
ADJUSTABLE COOLING SURFACE
~
I' ANNU U\R SPACE
FILLED WITH KAOWOOL
WATER JACKET
LINER OF ASTM 446
NOTE:
INSULATED LI~JER Ar-!D
ADJUSTABLE COOL!i!-3 SU7(F/~CE
PEm,HT. OPERATIOr~ AT
TE~.1PERATUr\ES A[JOVE
10000 F WITH BEDS 2i' DEr;:P
SECTION THROUGH FLUIDIZED-OED COLUMN SHOWING
INSULATED STEEL LINER Ar~D ADJUSTABLE COOLING
SURFACE
FIGURE 6
POPE. EVANS AND ROBBINS
'-:::;--':;-'I;.\,,"-,:-:r::::.~1
-------
46
involves fluidizing the bed material with minimum air flow,
raising the bed temperature to 800oF, then injecting coal
until the combustion is self sustaining while raising the
air rate. About 10 minutes of heating is required for
ignition. After the bed is, operating stably with coal, the
feed of carbon containing fly ash may be initiated. Fly ash
feed rate is then increased as coal feed rate is reduced
toward zero.
The entire FBC test system is shown schematically in
Figure 7. Air is supplied to the plenum of the 'FBC by two
blowers operating in series. Combustion products from the
FBC pass through an induced draft fan (which may be by-
passed), through a dust collector and on to atmosphere. The
slanted duct between the FBC and the dust collector provides
gas cooling without causing wall surface temperatures to fall
below the dew point of the gas. A control damper may be used
to provide a variable back pressure on the system. Fluidizing
combustion air was monitored both by a pitot tube and a
venturi meter located in the long entrance duct. A gate valve
in the line was used to control airflow to ~he unit. The fuel
feed rate was controlled by variable speed drives on the fly
ash and coal feed screws.
Collected dust was discharged into
A dust recirculation system, indicated in
used in this test series; neither was the
bags and weighed.
Figure 7, was not
induced draft fan.
5.2
Full-Scale Boiler, FBM
5.2.1
Boiler Module, FBM
The full-scale boiler module, designated the FBM,
is a
POPE, EVANS AND ROBBINS
INCORPORATED
-------
~
111
2 ~~
.0
.0
I ~ (J)
o ~
J! Z
~ 0
8 ;iJ
~
~
J.D. FAN
BYPASS
" J.D.
COAL AND FLY ASH
FEED SYSTEMS
.-1. D. FAN NOT USED IN THIS TEST SERIES
FIG~RE 7
SPRING SUPPORT
FROM CEILING
12" WELDED
ROUND DUCT
DAMPERS
CONTINUOUS ANALYZER
SAMPLING POINT
PARTICULATE
SAMPLING PT.
FLUE GAS
DISCHARGE
--
4" TUBE
FORCED DRAFT FAN
.eo.
~
SCHEMATIC OF FBC AIR AND EXHAUST GAS DUCTING SHOWiNG SAMPLING
POINTS
f:-
1 ~" \
- 2
;OIPE - ! 1 VENTURI
. _! PITOT
;
---,- _/GATE VALVE
D--___--Jf
r:l--WET TEST
J A~AL YZER
SAMPLE POINT
-
-------
48
coal-fired boiler capable of generating
of steam per hour at steam pressures up
large amount of work has been conducted
past stud ies (1., .1).
about 5,000 pounds
to 300 psig. A
with this unit in
The term FBM is used when referring to the test unit
as a whole, which includes auxiliaries and appendages, and
is also used when referring to that part of the whole being
described here. Whether the whole or part is being discussed
will be clear from the context. The term "primary cell" is
also used when referring to the FBM (part) and "burnup-ce11 "
is applied to the CBC. The FBM represents one half-cell of.
one of the original Mu1tice11 Fluidized-Bed Boiler concepts
developed under contract with the Office of Coal Research.
Two modules placed end-to-end would comprise one cell. A
nUmber of cells placed side by side without intervening
insulation would make up a full-scale boiler.
A 250,000 pound per hour boiler uses about 25,000
pounds of coal per hour. Since the FBM" is capable of firing
800 pounds of coal per hour, about 30 FBM-1ike modules would
be required for the large boiler.
A cutaway sketch of the FBM is provided in Figure 8.
A simulated Carbon-Burnup Cell, designated the CBC, added at
the rear of the FBM, is also shown in Figure 8. The CBC is
discussed in Section 5.2.2.
The FBM cross section (not including the CBC), is
approximately 18 by 72 inches, roughly seven times the cross
section of the FBC without the insulating liner. The bed is
surrounded by vertical boiler tubes which extend from the
two cross headers, below the grid plate, to the steam drum.
No other tubes were placed in the bed for the tests discussed
in this report. The boiler tubes are joined together by
POPE. EVANS AND ROBBINS
INCORPORATED
-------
FBM EXHAUST
F B M GAS
BREECHING
I '
I " , VERTICAL
u JCOAL FEEDER
INLET
, :,
, .
, , ~
" ",
. ,
-"..-." ,
(/(,,,
'n' .'
! -' /i
:';' '\ 1::-
I i'/ ;, ,.1
?/ ::.frWATER TUBES
/l 1/,( /:q
. ~ t'
d :.I~.! 1.\\
:: I ~ ~ c:::::a::a r-... '
I,', .....-
I,,,,, ' i D::a
,:i' I' ~
-1. '''::',1,'",1 1:'\
'4- . .,~'I, :1",., :' ~ ~~
-,' -;", ",,:"":;~"":'" j j' ';, J, . if,' ~ ~
,~',,~,::,,:' II """'i' ' ;;
-~ \:... -'.' , .".':' ",' ~ :;.IU'" [:-]" i,' r--......
.... ",... '. , ..",,'-,.. -' ,. ," -.. , ~
... " ',,,.-;, "":>;. >~~~ :;:.I-'II! ' "
.. '::,'r,' ~'~ }J , ~ r:.
£'. '::::~:GL=-::::-:-:', j' S\D
, " :. " ')' I
',~ \ ~ ~( ~''','
'- ).....- _.-\,....,h I'
o ,-,.~':;'\~
f\""".,"~
STE AM
DRUM
~
FRONT PANEL
LI GHTOFF
BURNER
ITI
Z ~
8 z
~ UJ
o ~
:D Z
~ 0
~ ::u
~
~
ADDITIVE OR ASH
FEED PORT
(AUXILIARY
GRID PLATE
HEADER
FLUIDIZED BED
~
I,Q
FIGURE 8 FLUIDIZED-BED MODULE (FBM) INTERNAL CONSTRUCTION
-------
50
welded fins and are backed by insulation. The fins do not
extend the full height of the tubes so that flue gas passes
between the tubes at the top of the unit, around the steam
drum and into the breeching.
The combustion. space of the FBM is accessible through
a water~coo1eq panel at the front of the unit. The panel
contains a pre-mix gas burner used to fire the bed. The
burner directs a flame downward onto the front of the bed.
From the plenum at the base of the unit, combustion air is
directed upward through the grid and then through the bed.
The grid consists of a mild steel plate containing buttons
of the same spacing and deslgn used in the FBC. The bed
material. used in the FBM tests was sintered ash, -8 +20 U.S.
Standard mesh. The static bed depth may be varied from 6 to
over 30 inches although the useful range is narrower.
Thermocouples were mounted throughout the bed as shown in
Figure 8. Detailed specifications of the FBM are presented
in Append ix B.
For this test series a vertical coal feeder, shown in
Figures 8 and 9, and discussed in Section 7.2 was utiiized
although a number of other' designs have been considered (~).
Sorbent and, in some tests, fly ash were fed via four
horizontal ports that penetrated a side wca11 of the unit,
through holes cut in the fins, just above the air distributor.
(See Figure 9). The ports -- numbered 1, 2, 3 & 4 from the
front of the FBM -- were spaced along the wall opposite the
the rmocoup1es shown on Figure 8. The ports extend 1-1/2" into
the bed and end with a deflector to impart an initial downward
path to the fine powders fed.
In operation, the bed is raised to the ignition point
of coal by use of the gas burner. Combustion of the coal
POPE. EVANS AND ROBBINS
INCORPORATED
-------
fl.. EXHAUST
."
: .
'.
.'
",'
~
!. ~
, " .
. ,. .. " .. ,- -.. C:' ,-.. -
~"' "... f., ~ ~.1 .-~ J.i . -, ~-, 1>), r...,
!. . . ,:1. .,,;.. .:,). ::.:>: .,~: . ~J' .:,:) . : ~:1': . ,..:.. : /
. ... . . I . . . .
"
o
"
rTI
~~
:;z
~(J)
~1>
"'z
°0'
fl..
::tI
O'
(II
~
z
(J)
ADDITIVE' ASH poltY-n
4TYPICAU
~~W)
~
"
~
II
..
d
.....-
FIGURE
1
SCR£..EH
2 - 18 X 28
INTEROOr1fl1,.,ICATION SLOTI'
ACCESS DOCJIt
~ nEDOJ (£rl.~D
~iii oaYRViU1~ ~
1:)-
I" PLENUIiIt
"
5'
"
.
WI.
.,;.
t
4'
GRAPHIC SCALE
(FBM)
9.
SECTION THROUGH FLUIDIZED BED MODULE
S!ML~LATED C~.RBON BURN UP CELl. (cae)
AND
-------
52
begins in the vicinity of the light-off burner and propagates
rapidly throughout the bed.
The energy not absorbed by the waterwalls leaves the
FBM as hot products of combustion. Two water-cooled tube
banks and an air heater were installed in the breeching beyond
the FBM to absorb some of this energy.
A schematic drawing of the FBM test system is shown
in Figure 10. Air from an external forced-draft fan passes
through the air preheater and into the FBM plenum. Coal feed
is controlled by the speed of a screw feeder which drops the
coal into a pneumatic feed tube at the injection port.
Sorbent materials were also screw fed to another pneumatic
injection line at a rate controlled by a variable speed
screw drive. Fly ash transport is pneumatic. A wear
resistant star feeder provides the pressure seal between the
dust collector hopper, maintained under negative pressure,
and the positive pressure transport line. Although the
details can not be shown in Figure 10 primary cell fly ash
can be discharged to a storage hopper, sent to the CBC or
be returned to the FBM. Generally in this test series the
collected fly ash was fed to the CBC. Flue gas from the FBM
passes across the first gas cooler above the unit to reduce
temperature before the gas enters the air preheater. As the
flue gas passes through the air preheater, a portion of the
fly ash drops out and is collected in the hopper shown. The
bulk of the fly ash is removed by a multi-cone collector
downstream of the air heater. During recirculation, the ash
knocked down by the preheater is screw fed into the dust
collector hopper. From the collector the gas flows through a
long duct to an induced draft fan and then to atmosphere.
A damper is provided in the ducting to control pressure in
POPE, EVANS AND ROBBINS
INCORPORATED
-------
~
-
f
PARTICULATE
SAMPL I NG POINT
DRAFT BALANCE
DAMPER
DU-ST COLLECTOR
DUST COLLECTOR
HOPPER
FLEXIBLE
CONNECTION
COAL INJECTION
AIR
SECOND GAS COOLER
-
~SAMPLE GAS TO ANALYZERS
- - - +
/
-_/
I'
I'
I'
II
I'
I'
11
I I
FIRST GAS "
COOLER '.J
~ 0 0 0 0 0 0 0 0 y~
o 0 0 0 0 0 0 0 0 '//
"
00 00
00000 AIR PRE HEATER
000
o 0
0000
o /--,
00000 /' \
OO$) , "
~
--
,,-,
,
, I,
\ I
,'-'" ,
--
\
\
ASH ffi
DROP-OUT a::r~
HOPPER ct ~
~w
wO:::
~Q..
~ f2.
ADDITIVE
HOPPER
(
FBM
STUB STACK
53
INDUCED DRAFT FAN
ROOF
COAL
HOPPER
STEAM DRUM
DOOR
j
DUCT FROM
FORCED DRAFT FAN
STAR FEEDER SCALE
SIGHT PORT
LI GHTOFF BURNER
--- TO CBC
ASH REINJECTION LINE
HOT AIR II NE
GRI D PLATE
LOWER
HEADER
,J
.--/
INLET AIR FROM PREHEATER
-
-
PLENUM
FIGURE 10 SCHEMATIC OF FBM TEST SYSTEM SHOWING VARIOUS SUBSYSTEMS
POPE, EVANS AND ROBBINS
U:Jcoo pC=ATl!D
-------
54
the combustion chamber. The system is operable without the
induced draft fan but is not usually run pressurized.
5.2.2
The Simulated Carbon-Burnup Cell, CBC
A demonstration of the Carbon-Burnup Cell concept,
operating on a continuous basis, was required. Although it
would have been possible to operate the FBM and FBC in
tandem (FBM firing coal, the FBC firing the FBM's fly ash)
this might not have been considered an adequate demonstration
of the efficacy of the Carbon-Burnup Cell concept. Building
a test system with an actual Carbon-Burnup Cell, while
providing the most effective demonstration, was not possible.
It was therefore decided to modify the FBM by adding an .
appropriately sized simulated Carbon-Burnup Cell. This
device was designated the CBC.*
The CBC was to operate continously, burning the carbon
containing .fly ash produced in the FBM while the FBM was
fired with crushed coal. The CBC was to operate at a higher
bed temperature that the FBM, but -- since bed material could
circulate between the two sections -- they shared a common
bed. No precendents existed for the design of such a system,
accordingly the apparatus was modified as the program
progressed.
The CBC consisted of a chamber with a rectangular
*The FBC and CBC were both simulated Carbon-Burnup Cells
because they did not have cold walls. To use cold walls the
cell must be large. In a small unit,the chamber walls must
be insulated. When refractory walls are used the ash,
attacks the refractory and sintered slag accumulations can
grow. When insulated, oxidation-resistant metal walls are
used, as in these tests, slag does not grow but the walls
are destroyed in only a few hundred hours of operation.
POPE, EVANS AND ROBBINS
INCORPORATED
-------
55
cross-sect ion, 10-5/8" x 15-5/8". From the grid leve 1,
considered the base line for all vertical dimensions, to
the roof of the CBC measured 56" but to the ga s outlet the
height of the CBC was only 48". These dimensions are shown
in Figure 9.
Figure 9 also shows why these inadequate height
dimensions were used. The only way to make the CBC an
integral part of the FBM without cutting pressure parts was
to locate it at the back and under the steam drum.' The
maximum width was set by the down-comer spacing and the
depth, by the location of one of the building's structural
walls. Except for height, the dimensions were appropriate
for the expected firin9 rate (500,000 to 750,000 Btujhr).
The decision to avoid cutting the drum or removing
some side wall tubes* was an error as the difficulties
encountered with bed material loss (discussed below)
indicated.
The uncooled, insulated walls of the CBC were
fabricated of 27% chromium steel, designated ASTM 446. This
metal has the lowest thermal coefficient of expansion of any
stainless steel and is able to resist oxidation at 2,0000F.
The grid plate was of the same design as that in the FBM.
As shown in Figure 9 the exhaust was an 8" I.D.
horizontal duct. This duct extended horizontally under the
steam drum, through the laboratory wall and then turned
vertically to a dust collector installed on the laboratory
*In the work which followed that reported here the FBM
was essentially remade, in place, by a very small, very
skillful ASME certified welder.
POPE. EVANS AND ROBBINS
INCORPORATED
-------
56
roof. .Dust collected here could be sent to a storage hopper,
routed back to the FBM via the four point injector or
returned to the CBC.
FBM.
Fuel for the CBC was the fly ash generated by the
Coal was fed along with the fly ash when the fly ash
could not maintain the desired 19000F to 2l00oF bed
alone
temperature level. Fly ash from the CBC's own collector was
also fed back into the CBC during most tests described in this
report. All fuel entered the unit through one port.
The fuel injector initially employed in the CBC was a
square, horizontal tube jutting into the unit almost
identical to that used in the FBC.However, this design was
soon changed to a "mushroom" feeder such as that shown in
Figure 11 and positioned as in Figure 9.
Fluidizing/combustion air is supplied via the air
heater from the main FBM air supply. Although the design
of the mushroom feeder makes it capable of supplying the bulk
of the CBC's fluidizing air the plenum and air distribution
grid were always used.
. The FBM and CBC share a common wall which cons ists of
the FBM's 2 inch G.D. back-wall tubes with intervening fins.
Portions of these fins were removed to enable bed material to
circulate between the two sections.
As shown in Figure 12
five slots were cut out of the fins 18" high and 1" wide.
A 2" G.D. boiler tube was between each pair of slots. The
part of each of these tubes facing the CBC was insulated by
a semicircle of insulation held in place by a thin metal
shield. The slots provided 90 square inches of open an=a
in the wall between the two bed regions.
The open area between the two sections, provided by
POPE, EVANS AND ROBBINS
INCORPORATED
-------
57
.' .
. .. . ~
'..' .
3-
CARBON STEEL
. "MUSHROOM"
THREADED
STUD (3)
COUPLING
AIR
DISTRIBUTOR
. FUEL/AIR
. SUSPENSION
NOT TO SCALE'
FIGURE II MUSHROOM FEEDER FOR
FLUIDIZED - BED COMBUSTOR
POPE, EVANS AN:> ROE3BII\JS
-------
eBe
58
(INSULATION
BED MA TERIAL-......... (TYPIC/\L)
PATH. )1000 p
A SLOT:jI'\-
ti --
TOP VIEW
18",
INTEG~t;l FIN
WALL CONST;
-------
59
removing the fins, was found to be excessive for the volume
of bed material in the CBC. Bed material circulates freely
between the two sections and this interchange provides a
significant heat sink for the CBC. Each pound of bed
particles which enter the CBC at 15500F and leave at 20500F
removes 120 Btu from the CBC.
The 90 suare inches of area provided so great a rate
of bed material exchange that the desired temperature
difference, 5000F between the bed material in the FBM and
the bed material in the CBC, could not be achieved. Thus,
in order to reduce the rate of bed material circulation,
the open area in the common wall was reduced in three stages:
to 45 square inches, to 4 square inches, and then finally to
2 square inches. The effects of these changes are discussed
in Section 7.4. The reduction of slot area was achieved by
placing a carbon-steel baffle on the CBC side of the common
wall of boiler tubes thus blocking the open area left by
the removal of the fins. Holes cut in the baffle provided
the desired opening.
The initial test results with the CBC as part of the
.FBM were poor becau~e of the initial injector design, the
excessive open area between the 'two sections and finally
because of the ins.ufficient freeboard in the CBC chamber.
The horizontal exhaust from the CBC was simply too close
to the top of the expanded bed to allow bed particles,
"splashed" up into the freeboard, to fall back into the bed
before being swept out by the gas. Cooler bed material from
the FBM, would flow into the CBC to replace the material
being lost at a.higher temperature. To reduce this energy
loss two designs for "particle knockouts" were tested. The
basic design concept was based on information obtained in
POPE. EVANS AND ROBBINS
INCORPORATED
-------
60
April 1970 from BCURA Industrial Laboratories.
how these two knockouts were placed in the CBC.
Figure 8 shows
The first design consisted of horizontal triangular
array (4 rows) of 1" tubes connected by U-bends to form a
single continuous water circuit. Water flowed into the tube
at the bottom and out at the top. Thermometers installed
in the inlet and outlet permitted the temperature rise to be
determined. A total of 5.2 square feet of heat exchange
surface was provided by this knockout screen. The effect of
this heat exchange surface is discussed in Section 7.6.
The second design consisted of a similar array of
uncooled rods fabricated of ASTM 446. This screen, as wel~
as the first, was installed within the CBC extending.from the
32" level to the 39" level. The knockouts are discussed
further in Section 7.5.
The test procedures for both the FBC and FBM operations
involved igniting the bed and stabilizing the combustion at
the desired bed temperature until steady-state conditions
prevailed. Steady-state was assumed when the bed temperature
was constant and the gas composition analyzers indicated
constant flue gas composition.
5.3
Instrumentation
Emissions of sulfur dioxide, nitric oxide and hydro-
carbons were monitored continuously. Infrared analyzers
(Beckman 215) were used to monitor sulfur fioxide and nitric
oxide. Hydrocarbons were detected with a flame ionization
analyzer (Beckman 109A) using methane as the reference gas.
The signal output of each of these units was displayed on
strip chart recorders.
POPE. EVANS AND ROBBINS
INCORPORATED
-------
61
The gas transfer system used with these analyzers is
sketched in Figure 13. The system permitted rechecking of
calibrations on any of the three units at any time during a
test by switching from sample gas to reference and zero gases
at the rotameter valves. The sample gas was drawn from the
flue gas stream through a sintered sta inless steel filter and
conditioned to remove water. The sample gas was again filtered
before entry into the analyzers to prevent possible contamina-
tion 'of the optical cells and the hydrogen burner.
The FBC gas sample was drawn into the instrument room
through a heated 1/2" O.D. tube from the FBC flue gas
discharge duct, shown in Figure 8 which passed directly over-
head.
In sampling the FBM flue gas, special precautions were
necessary because of the possibility of infiltration of
dilution air in the duct above the unit. Also, the poor
instrument response which would result from drawing a small
sample a long distance (approximately 60 feet) from unit to
instrument room was undesirable. Thus a system was devised
to draw a large gas sample from the FBM, just above the first
gas cooler (See Figure 10), pass it through a dust collector,
and then through a loop above the instrument room. The sample
tube was a 3" pipe with sections screw-fitted and then seal
welded. The system was driven with an I.D. fan located at
the discharge to atmosphere. Where the sam~le line passed
over the instrument room, a small sample was withdrawn via
a 1/2" O. D. tube. A schema tic drawing of the sys'tem is shown
in Figure 14.
The CBC gas sampling system was similar in design to
that for the FBM. The CBC sample line was a 2" pipe located
in the exhaust duct upstream of the dust collector, as shown
POPE, EVANS AND ROBBINS
INCORPORATED
-------
SYMBOLS:
o CALIBRATION REFERENCE AND COMBUSTION GAS SOURCES
~
CONTROL VALVE
ZERO
ROTAMETERS
REFRIGERATOR
NO
MIX
502
MIX
C02
MIX
FLUE GAS
62
FILTER 9'U.m
FILTER 0.3 Um
AIR
CONDENSER
PUMP
FILTER 0.3 1.(.m
I RECORDER I I RECORDER I I RECORDER I RECORDER
FILTER
0.31-lrn
NO 502 C02 HC
ANALYZER ANALYZER ANALYZER. TO ATM. ANALYZER
RELIEF
CH4
MIX
FIGURE 13 SCHEMATIC OF GAS TRANSFER SYSTEM FOR CONTINUOUS MONITORING OF SULFUR DIOXI DE, :
NITRIC OXI DE, CARBON DIOXIDE AND HYDROCARBONS
POPE, EVANS 4ND ROBBINS
,.cO.Pe,;....!'TID
-------
~
m
~ ~
~U>
0»
:D z
~ 0
8 ::u
~
Z
U>
Filters
To'
Atmosphere
50'
3" Pipe, Welded
Connections
I
I I
1
SO
NOx
x
Analysis
(vlet Tests)
To IR (S02 & NO)
HC + C02
Analyzers
Dust. Collector,
Welded Seams
From CBC Exhaust
.-
I.D. Fan
Flue Gas
Sample Line
~~
I .
I '
'1'
ii'
I I
I J "'"""
L.!
00000000
0000000
000000
~
\
Flue Gas froM
Fluidized Bed
FIGURE 14. SCHE~~~TIC OF THE FBM GAS Sfu~PLING SYSTEM.
FBM
IV
0\
.w
-------
64
in Figure 15.
When the FBM was operated, it was not possible to
simultaneously analyze the gas from the primary cell and
the burnup cell. A pair of ball valves permitted a sample
to be drawn from the pipes above the instrument room carrying
the FBM and CBC sample streams. The time that the analyzers
spent on each stream varied but was gen~rally about 15
minutes.
Particulate emissions were monitored with the
isokinetic probe system described in Reference 1. The prob~
design permits equalization of the sampling velocity with
the stream velocity. Locations of particulate sampling
points in the FBC, FBM, and CBC test systems are indicated
in Figures 7, 14, and 15, respectively.
Carbon dioxide was continuously monitored, using a
self-referencing thermal conductivity analyzer (Beckman Model
7C). The reference and span calibration gases contain 16%
C02' The zero reference gas is purified air. Instrument
cabinet temperature is controlled at 130°F. The C02 analyzer
was checked using a conventional Orsat analyzer which also
determined 02 and CO.
Bailey oxygen analyzers (Type OC1530A) were used as
operator guides to indicate oxygen concentration in the FBM
and CBC flue gases. During a test period, the air input rate' .
was held constant and the fuel rate adjusted to maintain the
oxygen concentration at the desired value. The Bailey
instruments have been calibrated periodically with known
mixtures of 02' N2' and C02 and found to be very reliable.
The flue gas oxygen was also verified, using the standard
Orsat technique. When the FBM was operated, a separate oxygen
POPE. EVANS AND ROBBINS
INCORPORATED
-------
FBM Steam prum
Exhaust
..~Building
~--
BM
I
1
BC
Access
iGrid Level -0"
65
62"
Leve17
Gas Sample Point
(2" pipe)
Wall
Exhaust to Dust Collector
and I. D. Fan
,2" pipe r~~o"~,_~ver Instrume.nt Room
... - .. ~ Cyclone Dust separator)'
To Sample- Fan ------. ::-;'" -~
r --1/2"
j" Stainless
Steel Tube
CBC
B.
Schematic of Sample Flow
FIGURE 15.
SCHEMATIC OF CBC GAS SAMPLING SYSTEM
POPE. EVANS AND ROBBINS
Gas to Instruments
-------
66
analyzer was connected to the primary cell and the CBC.
Temperatures in the bed and at various of her points
in the system were recorded on a Honeywell Multi-point.
Recorder. A multiple switch panel was used to connect the
recorder to either the FBC or FBM systems, as required.
Locations of thermocouples in the systems are indicated
in Appendices A and B.
The infrared analyzers and the hydrocarbon analyzer
were calibrated with gas mixtures supplied by vendors. The
concentration of the active components in the calibration
gases were checked after delivery to the laboratory. The
methane mixture was analyzed by the National Bureau of
Standards - a report is shown in Reference 1. This gas,
.containing 1265 ppm CH4' was used to calibrate a second methane
mixture before the first was depleted.
The sulfur dioxide calibration gas was analyzed with a
peroxide absorption train. A gas concentration of 2650 ppm
was used in the program. Analysis of the nitric oxide
calibration gas is also given in Reference 1.
The output signal of the infrared sulfur dioxide
analyzer varies in a nonlinear manner with S02 concentration.
The calibration curve provided with the instrument was checked
by precision dilution of the known calibration gas. The curve
was found to be correct except for a slight deviation at the
low end of the range. The calibration curve and check points
are given in Reference 1.
The calibration curve was used without correction since
the deviation is not more than 1% of full scale.
The calibration curve for the nitric oxide infrared
analyzer is given in Reference 1.
The contribution of water
. POPE. EVANS AND ROBBINS
INCORPORATED
-------
67
vapor to the signal output is significant with this analyzer.
The water vapor correction determined by the supplier (180
ppm) was checked by testing a dry gas in the analyzer for
comparison with a moist sample. A correction of 220 ppm was
noted and incorporated in the data reduction. The range of
this unit is 0-1000 ppm NO.
5.4
Materials
Three materials were used in the test program: fly ash,
coal and limestone. The sources of each are discussed below.
5.4.1
Fly ash for Tests in the FBC.
Fly ash having the properties desired for this test
series was not easy to obtain. The two properties required
were a coarse size consist and a relatively high carbon content.
An analysis of the fly ash from several of the nearby
utility power stations, showed their fly ash to be unsuitable.
Interestingly a few samples had carbon contents in excess of
30%,"but the particle size distribution was considered too
fine; the utility plants all used pulverized coal.
Stoker fired steam plants serve most of the federal
facilities in the Washington area; their fly ash was of the
proper type. Unfortunately the ash required for these tests
had to be dry and free of bottom ash and most plants either
wet the fly ash or mixed it with bottom ash as it entered the
transport system to the ash silo.
The search for fly ash took us to the prisons of the
State of Maryland and to the Glidden's Titanium Dioxide plant
in Baltimore. Both had stoker fired boilers. The prison
officials thought our request for fly ash was rather odd but
POPE, EVANS AND ROBBINS
INCORPORATED
-------
68
were willing to cooperate if we could obtain the proper
clearances to bring a 6000 gallon solids tanker onto the
prison grounds. Glidden was chosen.
A tanker was leased and parked adjacent to the fly
ash hopper of the mechanical collector on one of Glidden's
boilers. Pope, Evans and Robbins personnel constructed a
system for withdrawing the fly ash from the hopper, which was
negative pressure, and over a period of seven days the tanker
was filled.
Unfortunately the material which arrived in the tanker
was low in carbon ( ~ 30%), not at all like the initial
samples ( ~45% carbon). Investigation revealed that the plant
personnel in an attempt to be helpful had increased the load
of the boiler filling the tanker, so as to fill it faster, and
had thereby improved the units combustion efficiency.
The steam plant serving the U. S. Capitol, the Supreme
Court and the Library of Congress agreed to operate their ash
disposal system in a manner which would allow Pope, Evans and
Robbins to obtain fly ash free of bottom ash. They would
store fly ash while the ash silo was emptied, then transfer
fly ash and store bottom ash. By this means a few hundred
pounds of fly ash could be obtained during each trip, enough
for a few tests. Fly ash obtained by this method constituted
about 50% of the material used in the FBC statistical test
program.
After the CBC had been added to the FBM so that the
FBM could be operated, coal was burned in the FBM and the
fly ash stored. This material was also used in the FBC.
The test data (Appendix E) do not indicate the source
of fly ash for each test since this was not considered
POPE, EVANS AND ROBBINS
INCORPORATED
-------
69
re1event by the statistical consultant. Capitol Power
Plant fly ash was used in tests C-304 through C-312. FBM
fly ash was used in tests C-313 through C-320. For tests
C-301 through C-303, (these data were not used in the model)
fly ash from PER laboratory's waste pile was used. No
attempt was made to interpret FBC data for each fly ash
source separately. The performance model provuded with all
of the data indicates that this decision was valid.
5.4.2
Coal for FBC and FBM Tests
Most of the coal used in the tests reported here
was from the same mine as that used in Reference 1. This
coal is an East Ohio Pittsburgh No.8 Seam coal designated
"Perfect 8". The high ash content indicates the coal was
unwashed. The coal had the following analysis:
Constituent
Ultimate Analysis, Wt. %
(As Received)
Carbon
Hydrogen
Oxygen
Sulfur
Ni trogen
Moisture
Ash
68.35
4.57
5.48
3.30
2.50
2.30
13.50
Total
100.00
Net heating value
11,498 Btu/1b
POPE, EVANS AND ROBBINS
INCORPORATED
-------
70
5.4.3
Limestone for FBC and FBM Tests
The limestone used in those tests in which fine 1ime-
stone injection was investigated was the .same No. 1359 stone
used in the studies described in Reference 1. .
POPE. EVANS AND ROBBINS.
INCORPORATED
-------
71
6.
PROCEDURES AND RESULTS:
TESTS WITH THE FBC
6.1
General
Tests in the insulated FBC were intended to determine
the effect of important operating variables on the combustion
of fly ash in a Carbon-Burnup Cell. The FBC results were
independent of any complicating factors resulting from the
interchange of bed material between the burnup cell and
primary cell. Tests in the FBM, in which bed material
interchange occurred, and in which the feed rate of fly ash
. to the CBC could not be controlled, were useful for
demonstration of, and insights into, the dynamics of the
process. However, the FBM could not be used for the derivation
of a performance model with the time and other resources
ava ila b1e .
6.2
Statistical Desiqn of Experiments
6.2.1
Perspective
Predictions, based on prior work in the field of
fluidized-bed combustion, were made at the outset of the
program. It was predicted that combustion efficiency,
defined below, would be increased by increases in the following
pa rame te rs :
5.
Bed temperature.
Bed depth.
Ratio of air to fuel.
Superficial gas residence time; i.e., decre~se
in superficial gas velocity.
Carbon content of fly ash; i.e., less inert matter
would be available to absorb energy.
1.
2.
3.
4.
POPE. EVANS AND ROBBINS
INCORPORATED
-------
72
6.
Mesh size of feed;
i.e. ,
decrease in feed particle
size.
On the other hand, the cost of a fluidized-bed boiler
might be favored by a low value of each of these parameters.
Thus, it might be necessary to find an optimum combination
for combustion efficiency, auxiliary power requirements and
capital costs. The goal of the experimental design was to
provide, with a minimum number of experiments in the FBC,
a mathematical relationship between the control variables
and the principal response, combustion efficiency. The
mathematical relationship would show if (and how) it might
be possible to optimize the system.
Combustion efficiency (CE) is defined as:
CE
H
= (1 0:.. H ~) x 100
1
(5 )
where
H
o
=
higher heating value of unburned fuel
components leaving the system, Btu/1b x 1bs.
H.
1
=
higher heating value of fuel entering system,
Btu/1b x 1bs.
6.2.2
Identification of Variables
As part of this program an attempt was made to identify
the variables describing the fly ash combustion process in a
fluidized-bed. The system was broken down into six major
components: (1) the air feed; (2) the fuel feed; (3) the bed,
(4) gaseous products, (5) gas-borne solids, and (6) the
reactor. Three categories of variables were enumerated for
each of the six components: independent (or control)
variables; dependent (or response) variables; and design
POPE. EVANS AND ROBBINS
INCORPORATED
-------
73
variables. The potentially important parameters were listed
under each category. These are given below in Tables I
through 6.
It should be recognized that the classification is
somewhat arbitrary and that all controls influence all
responses. A fourth category, natural material variables,
was also established to show how the fuel might be
characterized.
Some of the symbols may not be understood by all
readers. It is unimportant since they are not used elsewhere
in the report.
TABLE I.
COMBUSTION AIR FEED VARIABLES
A.
Control Variables
I.
2.
3.
4.
2
Mass flow rate, lb/hr ft
Temperature above ambient, of
Pressure, atmospheres
Composition; e.g., % oxygen, %
humidity
B.
Desiqn Variables
l.
2.
3.
Distributor design
Injection point of secondary air
Arrangement of secondary injection system
C.
Response Variables
None
POPE, EVANS AND ROBBINS
INCORPORATED
-------
74
TABLE 2.
FUEL FEED VARIABLES
A.
B.
C.
D.
1.
2 .
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
Control Variables
Mass flow rate, 1b/hr ft2
Temperature above ambient, of
Composition; e.g., % carbon, % inert diluent,
% moisture
Particle size distribution
Injection gas rate; lbs of sOlid/lb of injection gas
Use of homogenous combustion catalyst
1.
2.
3.
4.
5.
6.
Desiqn Variables
1 .
2.
3.
4.
Injector design
Injector location
Recycle rate
Injection gas selection;
gas, steam
e.g.,
air,
compressed flue
Response Variables
None
Natural Material Variables
Ignition temperature
Porosity
Ash particle size
Ash sintering temperature
Degree of graphitization of carbon
Ash/carbon matrix characteristics
Attrition chara~teristics
Swelling characteristics
Surface area
Shape factor
Densi ty
Emissivity
Diffusivity
POPE. EVANS. AND ROBBINs.
INCORPORATED
-------
75
TABLE 3.
BED VARIABLES
A.
B.
C.
9.
10.
11.
12.
Control Variables
1.
2.
Mass, pounds (weight)/square foot of grid area
Composition; e.g., particle size distribution,
particle density, bulk density, thermal properties,
hardness
Temperature history; e.g., is the steady-state
temperature approached from a higher or lower
initial temperature
3 .
Desiqn Variables
1.
Combustor shape; e.g., constant cross section or
conical
Interna1s; e.g., arrangement of heat transfer surfaces,
inclusion of a knockout; i.e., baffles in freeboard
Circulation of hot or cold particles
Heat sink; e.g., surface area, effectiveness,
temperature, location
Solids removal method 2
Cross-sectional area, ft
2.
3.
4.
5.
6.
Response Variables
1 .
2.
3.
Nominal or average bed temperature, of
Bed temperature distribution; e.g., T(x,y,z)
Effective thermal conductivity - axial and
horizontal, Btu/ft2hroF/ft , 2
Diffusivity - axial and horizontal, ft /hr
Expanded bed height, ft
Bed density distribution; e.g., p (z)
Bed circulation patterns, bubble frequency, channel
1 i fe and movement
Equilibrium particle characteristics; e.g., particle
size and shape
Residence time distribution of fine particles, t (d )
Fines capture rate through agglomeration r p
Ash particle size increase through agglomeration
Viscosity of fluidized-bed
4.
5.
6.
7.
8.
POPE, EVANS AND ROBBINS
INCORPORATED
-------
76
TABLE 4.
COMBUSTION-GAS STREAM
A.
B.
C.
Control Variables
None
Desiqn Variables
1.
2.
3.
Arrangement of above-bed heat transfer surfaces
Volume of freeboard
Laminar or turbulent flow
Response Variables
1.
Mass flow rate, Ib/hr (or suPerficial velocity,
ft/sec)
Final composition; e.g. %
Composition distribution;
S02(x,y,z)
Temperature distribution
Final or average temperature, of
Solids loading, grains/scf
Electrical properties of gas (for electrostatic
precipitation)
2.
02' N2' C02' CO, HxCy
e.g.,02(x,y,z)
3.
4.
5.
6.
7.
TABLE 5.
FLUE GAS SOLIDS VARIABLES
A.
B.
C.
Control Variables
None
Desiqn Variables
l.
2 .
3.
Freeboard (disengaging height)
Above-bed baffles
Secondary air injection
Response Variables
1.
2.
3.
4.
Mass flow rate, lb/hr
Fina 1 compos it ion, %C, o/aH2' o/eS, %N2' %Cl, %NA +K
Size and distribution and density aistribution
Resistivity (for electrostatic precipitation)
POPE, EVANS AND ROBBINS
INCORPORATED
-------
77
TABLE 6.
REACTOR VARIABLES
A.
Control Variables
B.
C.
1.
position of cooling surfaces; e.g., in-bed, above
bed
Coolant temperatures
Note: There were also included in Table 3 as design
variables. Their inclusion here is to
emphasize the role of heat removal as a
control variable in an experimental apparatus
2.
Desiqn Variables
(Covered under fuel, air, bed, and products of combustion)
Response Variables
1.
Steam generated or superheated; i.e.,
from bed via heat exchange surface
Erosion of surfaces
Corrosion of surfaces
Slagging of surfaces
energy removed
2.
3.
4.
While the listing in Tables 1-6 is extensive, it cannot
include all of the variables which actually influence the
performance of a combustor nor all of the variables by which
this performance might be measured.
6.2.3
Selection of Most Important Variables
Of the variables listed in Tables 1 through 6 those
considered most significant are listed below:
As important control variables the following would be
included: Mass flow rate of air (l.A.l), Mass flow rate of
fuel (2.A.l), Composition of fuel (2.A.3), Particle size
distribution of fuel (2.A.4), Mass of bed (3.A.l), position
of cooling surfaces (6.A.l) and Coolant temperatures (6.A.2).
POPE. EVANS AND ROBBINS
INCORPORATED
-------
78
As important response variables would be included:
Nominal or average bed temperature (3.C.l), Superficial
velocity (4.C.l), Final gas composition (4.C.2), Final
gas-borne solids composition (5.C.2) and Energy removed
from bed heat exchange surfaces (6.C.l).
If all of these most important control and response
variables had been studied thoroughly, it might have been
possible to achieve a relatively basic understanding of the
process of fluidized-bed combustion of carbon in fly ash and
fluidized-bed heat exchange. Unfortunately it did not appear
possible to conduct such thorough fundamental studies with
the financial resources which were available. On the other
hand a basic understanding of the process was not required to
(a) prove that it would work and (b) produce a prediction tool
for use in design.
Even these limited goals could only be partially
achieved: (a) Definitive proof or operability must await the
construction of a large system, (b) No means were available
at the outset of the program with which bed temperature could
be predicted and no such technique was produced as a result
of the program described here. Instead of being a response
variable as anticipated beforehand, bed temperature had to.be
treated as a control variable. This anomalous behavior of
bed temperature will be discussed later.
6.2.4
Variables Actually Studied
In preparing the proposed work plan in August, 1969, it
was useful .to treat the variables somewhat differently than
indicated in the preceding paragraphs. Table 7 below lists
the variables to be studied just as they were listed in the
POPE. EVANS AND ROBBINS
INCORPORATED
-------
79
proposed work plan, the range originally suggested and the
range actually studied.
It is instructive to use this original format and to
explain how response variables, as listed in Tables 1 through
6, can appear as control, or, possibly a better term, pseudo-
control variables.
TABIE 7
CONTROL AND PSEUDO-CONTROL VARIABLES
SEIEC'IED FOR CARBON-BURNUP CELL INVESTIGATION
Variable
Range
Initially
Suqqested
1.
Carbon content of fly ash
Particle size of fuel
20% to 80%
as received
and 70%
through 200
mesh
2.
3.
4.
Static bed depth
Nominal bed temperature
6" to 18"
l5000F to
2l00oF .
5.
6.
Excess air
Superficial velocity
0% to 50%
5 to 20 fps
Range
Actually
Studied
28% to 65%
as received
10" to 22"
l7000F to
2l50oF
10% to 90%
6 to 15.3 fps
Temperature could be treated as a control variable
since it was possible to alter the bed temperature during
the course of an experiment by inserting or withdrawing a
heat sink in the form of a bayonet tube. True, the
temperature changed internally in response to an external
control, the position of the bayonet, but not in a way that
could be predicted. The tube was moved up and down until
the desired temperature was achieved.
-------
80
Excess air (actually residual oxygen), could more
readily be used as the control than air rate. Residual
oxygen was indicated and recorded at the FBC operating
station, whereas air rate was indicated via a manometer
which required interpretation.
Superficial velocity was treated as a control variable
in planning since the selection of this parameter determined
the sizing of the starter bed.
The other control variables listed in Table 7 --
carbon content of fly ash, particle size of this fuel and
the static bed depth -- were also not subject to absolute
control. The fuel would vary from test to test and quite
probably within tests since it was real fly ash; not even
homogenous, let alone "reagent grade". Static bed depth was
as much a response as a control. The bed could increase in
depth via ash agglomeration or decrease due to attrition.
Tests were never of sufficient duration for the bed depth
change to become important.
The statistical consultant, Professor Arthur E. Hoerl
of the University of Delaware, selected from among the
various controls, pseudo-controls and methods of expressing
the values the following:
Bed temperature, static bed depth, carbon feed rate,
inert feed rate and air feed rate. In addition to these
control variables, it was found desirable to add ,some special
experiments which might point up design considerations.
These special tests included the addition of coal and
operation with two bed particle-size distributions. The
effect of coal was tested since in an actual Carbon-Burnup
Cell coal will be used to maintain a constant bed temperature
POPE, EVANS AND ROBBINS
INCORPORATED
-------
81
despite variations in flyash flow and fuel value.
could
found
The effect of coal addition and of bed particle size
not be included in the modeling process. Neither was
to be of major importance over the range studied.
The effect of grinding the fly ash was omitted from
the test program since favorable results were obtained without
this step.
6.2.5
Preliminary Modelinq (A Simple Physical Model)
Before the experimental work began, a preliminary
physical model was prepared in order to aid in the statistical
design of the experiments. The model was intended to serve
only as a map of the operating region of the FBC. This model
is shown diagramatically in Figure 16.
It was assumed for the purposes of this model that the
heat release rate within the fluidized-bed was given by
mUltiplying together the carbon feed rate, the heat of reaction
for
C
+
02 :t
CO
2
and an arbitrary combustion efficiency, taken as 80%.
The heat loss rate from the bed through contact with
walls of the combustor was taken as 65,000 Btu/hr; a heat
loss through radiation was taken as proportional to the
fourth power of the bed temperature with emissivity-absorptivity
factors taken as 0.8 from a radiating surface of 1.57 square
feet; a sensible heat loss in the flue gas was assumed to be
proportional to the bed temperature, the air rate and the flue
gas C02 content; and heat removed by the variable cooling coil,
was taken to be the product of the heat flux (100,000 Btu/hr ft2
POPE. EVANS AND ROBBINS
INCORPORATED
-------
1000
800
~
.£:'
......... .
.a
.-i
.. 600
G)
~.
to
~
~
~
p..
~ 400
~
.~
~
200
SJ
Superficial
velocity
limit
'I. 15 fps
o
o
FIGURE 16
Bed Tempc~ature, 8F
1\
. I f:)() \
. ). b CO () f:) f:) f:)
. i,). /'}.f:)() i ').\.? 1
I , I I
IJ " h4 Isotherms for cooling
, I I with excess ir only
'I I /
/ I
20
40
Carbon Input Rate, lb/hr
Su face cool ng limit
Assumptions
1 CODbustion efficiency = 80%
2. 1\11. cor:1bustion occurs in bed
3. Bed cross-section = 0.9 'ft"
4. Loss to walls = 65,J)~ Btu/hr
5. Loss fro~ bed surface bv radiation
= 1.57 ft2x(0.B)-.q,(T4) I3tu/hr
6. Loss to coolinq Drobe
= 100,000 DtU/~t~ hr of surface.
Surface may be varied between
o and 4.0 ft2
60
80
100
PREDICTED AIR-CAR30:~ OPER.~TING REGIHE OF 'i'IlE FilC
POPE, EVANS A1'C) ROB8NS ~
. ..,;1 ::~I~A~'-: :;:'1:'~-,' :[J~/\n i~,:~:
-------
83
of surface at 2l50oF) times the exposed coil surface (varied
between 0 and 4.0 ft2). In this model bed temperature can be
reduced either by increasing excess air, inserting the cooling
coil or both.
With the equations thus derived, it follows that
the bed temperature would be predicted in this apparatus by
the control variables; carbon rate, air rate and exposed coil
surface. The baseline of Figure 16, line A-D, is for stoichio-
metric carbon/air feed. No excess air and no cooling, other
than the wall and radiation, is provided to control bed
temperature. The bed temperature is shown to increase from
point A (1600oF) to point B (1800oF) as the carbon and air
rates are simultaneously increased in stoichiometric proportions.
Further increase in both fuel and air, maintaining a constant
and stoichiometric fuel/air ratio, is shown to increase the bed
temperature to 20000F at point C and 2l50oF to point D.
In preparing the map, 2l50oF was considered maximum
feasible operating temperature since the bed could become sticky
and collapse if the temperature were increased further. However,
operation at 22400F had been demonstrated in one experiment with
coal. To limit the temperature to 2l50oF as the fuel and air
rates are increased beyond point D toward point E, additional
cooling must be provided. This additional cooling is provided
by the insertion of the cooling coil. At point D, 0 ft2 of
surface is exposed; at point E, the 4.0 ft2 of the coil are
fully exposed to the bed.
A heat balance shows a stoichiometric fuel-air ratio
is no longer possible yond point E, if the temperature is to
be maintained at 2l50oF, since the coil is fully inserted.
Therefore an increase in fuel rate must be accompanied by
POPE. EVANS AND ROBBINS
INCORPORATED
-------
84
more than a stoichiometric increase in air rate so as to
provide cooling. Line E-F is constructed by calculating
the excess air required to maintain a 2l50oF bed temperature
as the fuel rate is increased. At point F the temperature
and air rate (approximately 750 lb/hr) provide a superficial
gas velocity of 15 feet per second. This gas velocity was
arbitrarily chosen as the maximum permissible. An increase
in air rate beyond that at point F is not permitted until
the bed temperature is reduced below 2l500F. Point F also
represents the maximum permissible firing rate, equivalent
to a heat release rate of 950,000 Btu/hr per square foot of
grate area.
Line A-J is the l6000F isotherm. Along this line air
rate is increased more rapidly than the fuel rate so as to
maintain a constant bed temperature without the use of the
cooling coil; i.e., the excess air provides the required
cooling. At point J the air rate (approximately 900 lb/hr,
40 lb/hr fuel) and the l6000F bed temperature would give a
superficial velocity of 15 ft/sec. Point J therefore
represents the maximum permissible air rate. Isotherms
similar to line A-J can be drawn at other temperatures
(See lines B-1, C-H, D-G on Figure 16) for the case where
the required cooling is provided by excess air.
Line J-G represents a constant velocity curve
connecting those points on each of the isotherms where the
air rate and bed temperature result in the 15 fps gas
velocity set as the maximum. For example, at point I the
fuel rate has been increased to 40 lbs/hr while the excess
air was decreased sufficiently to maintain the superficial
gas velocity at 15 fps while the bed temperature rose to
l8000F. The cooling coil is not in use in this area of the
POPE, EVANS AND ROBBINS
INCORPORATED
-------
85
figure. Further decrease in excess air by increasing fuel
and decreasing the air rate past point I maintains the 15
fps limit through point H (20000F) to point G (2l500F), the
maximum permissible temperature. Between points G and F
the cooling coil must be inserted as the fuel rate is
increased so that the bed temperature can be maintained
at 2l500F. The superficial velocity remains 15 fps.
Within the area enclosed by lines A-E, E-F, F-G, G-J
and J-A, the Carbon-Burnup Cell was assumed to be stable.
Assuming a combustion efficiency different from 80%
would result in an area of similar shape. A three-dimensional
body could then be constructed from the isoefficiency maps
within which a Carbon-Burnup Cell would be assumed to be
operable. At 100% combustion efficiency, the map would be
displaced downward and to the left. At about 35% combustion
efficiency, the regime of operation would appear as a single
point at approximately 900 lb/hr air rate, 90 lb/hr-carbon
and l6000F.
Additional dimensions might be added to the regime of
operation representing other variables; for example, the
temperatures of the incoming air and fuel feel could be
included in the physical model.
When the experimental work aimed at developing a
statistical model was begun some of the test results actually
fit the simple physical model depicted in Figure 16; most did
not. As will be shown later it was not possible to predict
the bed temperature as a function of fuel rate, air rate,
combustion efficiency and heat removal from the combustion
chamber.
Three reasons will be offered for the deficiencies in
POPE, EVANS AND ROBBINS
INCORPORATED
-------
86
the physical model.
In summary, these are as follows:
1) The process is bi-stable; i.e., two apparently
identical experiments will yield two different results
depending, for example, on whether the steady state bed
temperature was approached from an initially higher or from
an initially lower value.
2) The physical model assumed that all energy released
by the burning fuel occurred within the dense phase of the
fluidized bed. An unknown fraction of the burning actually
occurs in the freeboard above the bed.
3) The mechanisms by which the turbulent fluidized-
bed loses energy to heat transfer surfaces above the dense
phase surface are not fully understood. Radiation from the
refluxing bed particles is one important mechanism for losing
energy from the bed.
Despite these deficiencies, no attempts were made to
improve the physical model since it served its mapping
purpose adequately. However, fundamental studies of
heterogenous combustion within a fluidized-bed would improve
the model and have been advocated by Pope, Evans and Robbins
(19) as well as others (15, 20). Fundamental studies on the
course of combustion in a fluidized-bed might also be useful
for air pollution control
(g) .
6.2.6
The Desiqn of the Experiments: Selection of Test
Conditions
The simple physical model and the predictions based
on previous experience were reviewed with the statistical
consultant. Because of the numerous parameters the consultant
suggested that it would be desirable to run well over 100
POPE. EVANS AND RbBBINS
INCORPORATED
-------
87
tests.
This was not possible with the available resources.
A total of 55 tests were run, 38 of which were used
in producing the statistically derived performance model.
All of the tests are described in Appendix E; while those
used by the consultant are listed in his report, included
here as Appendix F.
At the consultant's request, tests were run in small
blocks, and the data reduced and the results reviewed follow-
ing each block. The consultant would then select the
conditions for the next small block of tests.
The design of the experiments was thereby conducted
intermittently during the test program and did not actually
precede it as the term design implies.
One of the first tasks assigned by the consultant was
to explore the boundaries of the Carbon-Burnup Cell process;
i.e., the regime of stable operation. This topic is discussed
in the next section.
6.3
General Results: Explorinq the Reqime of Stable
Operation
The physical model of F~gure 16, and its extensions to
a third dimension, combustion efficiency, indicated the outer-
most limits in which the Carbon-Burnup Cell system would be
operable.
It should be recalled that the FBC only
Carbon-Burnup Cell and problems encountered in
of this small system might not exist in a much
the destruction of the uncooled metal walls is
simulates a
the operation
larger device;
an obvious
example.
Conversely, problems in fuel distribution do not
POPE, EVANS AND ROBBINS
INCORPORATED
-------
88
exist in a small bench scale system but must be considered in
the commercial design.
An experimental tool, such as the PBC, is capable of
answering questions about a process if these questions are
properly posed. It is often impossible to determine if the
apparatus has answered a question about its own limitations
or about a limitation inherent in the process itself. One
obvious example of this kind of ambiguity are the upper and
lower temperature limits at which the process was stable.
This and several other topics are discussed below.
A.
Low Temperature
It was found that operations below about l8000p
were not feasible with fly ash as the fuel. TIle minimum
temperature l6000p used in the preliminary model had been
chosen as a lower limit since coal combustion was stable at
l6000p and, more important, because a high ash ( ~ 50%) char
had been burned at 1600oP. Operation at low air rates (say,
equivalent to a superficial velocity of 0.5 fps) should have
been possible even at 13000F in an almost adiabatic apparatus;
this is impractical however. Only a few tests were conducted
at l8000p and below.
B.
Superficial Velocity
It was known, and confirmed in the tests, that a
fluidized bed of coarse particles has a relatively small
value for the Operable Velocity Ratio:
where
V
max
Operable Velocity Ratio = ~
mln
POPE, EVANS AND ROBBINS
INCORPORATED
-------
89
where
V
max
=
Superficial velocity at which
substantial elutriation occurs
v .
mln
=
Superficial velocity at which
substantial stagnation occurs
For particles in the size ranges -8 +16 and -16 +22, this
ratio is on the order of 2.
A precise value for the ratio depends on the
reactor configuration (e.g., higher values would be found
for conical reactors), the operating temperature, and the
nature of the particles. This last factor, the nature of
the particles, relates to the definition of substantial
stagnation. It is defined here as being that point of gas
velocity and bed temperature at which the particles not only
stagnate but also sinter into a large mass. Goldberger (17)
has provided a curve that indicates, for his apparatus, a
relationship between superficial gas velocity and bed collapse
tempera ture.
C.
High Temperature
Sintering of the particles was fo~nd to occur at
a bed temperature above 20000F at the higher superficial
velocities. This sintering was reflected by an increase in
the particle Slze as fine matter released by the burning fuel
adhered to the bed particles. The rate of growth appeared very
rapid above 2l00oF.
Agglomeration can be a potentially useful effect;
it has been proposed that this particle capture effect be
utilized to reduce the dust load to the pOllution control
system. However, it was not desirable here. Agglomeration
in the FBC would require that either a system be provided to
POPE. EVANS AND ROBBINS
INCORPORATED
-------
90
remove particles continuously, with separation of the larger
fraction and return of the finer fraction, or that the air
rate be increased continuously as the test progressed and the
particles grow.
The former action was beyond the scope of the
current effort. Continuous increase in air rate would have
been useful if the apparatus had been instrumented so that
the results could have been properly interpreted. Studies of
both size maintenance and dynamic testing would appear to be
justified, on the basis that the first may eventually be
required in a commercial fluidized-bed boiler, and the latter
is required if substantive insights on the nature of fluidized-
bed combustion are to be achieved.
D.
Bed Particle Size vs. Temperature and Air Rates
In the tests performed, the operating regime of the
Carbon-Burnup Cell system was limited to a lower temperature
of l8000F, an upper limit of 20S00F, and two bed particle size
regimes. Air rates in the range below 800 lb/hr could be
accommodated with a bed of particles sized -16 +22 U.S.
Standard sieve size. Air rates above 600 lb/hr could be
accommodated with a bed of particles sized -8 +16 U. S.
Operation at 600 lb/hr was possible with either bed so that
the effect of particle size on combustion efficiency could
be determined.
E.
Bed Depth
The effect of bed depth was also explored. At the
outset of the program, it was thought that operation with beds
as shallow as 6" could be useful since coal had been burned in
4" deep bed.
It was found, however, that below about 14",
POPE, EVANS AND ROBBINS
INCORPORATED
-------
91
combustion efficiericy deteriorated badly at the higher air
rates.
The dependence of combustion stability on a minimum
bed depth was assessed by a theoretical analysis of the rate
of particle heating. It was assumed that initially the
particles of fly ash would be at the temperature of the
injection air. Further, the particle temperature rise would
be equivalent to the air temperature rise. This was supported
by an analysis of the rate at which energy is transferred to
the particle, and the rate at which energy received by the
surface of the particle would lead to a temperature rise in
the center of the particle. For particles of carbon, less
than 200 microns in diameter, the convective heat transfer
rate to the particle and the rate of internal heat transfer
were sufficiently high to insure that the particles and gas
would be at essentially the same temperature during the
heating-up period. The problem was reduced, therefore, to
one of predicting the rate at which the gas temperature would
rise as it entered the hot fluidized-bed. The analysis,
presented as Appendix G, shows that within about 1/2" of the
injection the fly ash particles should be close to the bed
temperature.
One possible reason for the poor performance of
a shallow bed might have been inadequate distribution of the
fly ash by the fuel injector. The FBC feeder was not modified.
Instead, beds of 10" and above were studied. The CBC fly ash
feeder was changed to the mushroom arrangement as noted in
the description of the equipment.
F.
Other Limitations
There were a number of other limitations which may
POPE. EVANS AND ROBBINS
INCORPORATED
-------
92
relate to the process but are probably related to the
apparatus. For example, it was not possible to fire fly
ash with a carbon content below about 30%. Fly ash with
a high moisture content performed poorly. The overall system
heat balance did not permit operation at above 10% residual
oxygen in the flue gas (high excess air). The heat removal
rate was also limited by heat balance considerations. When
the ratio (heat removal from the chambers via the walls,
hood and cooling probes plus dust) : (heat released)
exceeded about 35% to 45% the combustion process would begin
to deteriorate. It is believed that the fire would go out
if this ratio is > 50%.
6.4
StatisticallY Derived Model for Fly Ash Carbon
Combustion in a Fluidized-Bed Combustor
6.4.1
Description of Model
Based on the data from 38 (Appendix F) of 55 (See
Appendix E) combustion tests in the FBC, an empirical
regression model was developed with which the combustion
efficiency of fly ash in the FBC may be accurately predicted
if the following are known: (a) the bed temperature,
(b) the air rate, (c) the static bed depth, (d) the fuel rate
and (e) the carbon content of the fly ash.
Of the 17 tests not used in the model development,
about one-half were performed after the model was complete,
and one-half were too deviant to use or were conducted with
coal as well as fly ash feed.
The general function is as follows:
POPE, EVANS AND ROBBINS
INCORPORATED
-------
93
y = be + bl (T) + b2 (A) + b3 (H) + b4 (C)
+ bS (I) + b6 (I x C)
(6)
where
y = a particular response
be to b6 = constants determined in the analysis
for the particular response
and
o
T = bed temperature, F
A = air rate, lbs/hr
H = height of static bed, inches
C = carbon rate, lbs/hr
I = inert rate, lbs/hr
(I x C) = cross product of carbon rate and inert
rate, Ibs2/hr2
The parameters C and I are related to the fuel feed rate
by the expressions
C = F x (c)
I = F - C = F x (~-c)
(7)
(8)
where
F = fuel feed rate, lbs/hr
c = carbon content of fuel, lb carbon/lb fuel
It has been noted previously that the bed temperature,
T, appearing in Equation 6 as an independent variable, would,
preferably have been treated as a response variable. However,
temperature could not be treated as a response variable.
When the heat removal rate, expressed in Btu/hr, by the
cooling surfaces and in dust' (the preferred independent
POPE, EVANS AND ROBBINS
INCORPORATED
-------
94
variable) was employed in Equation 6, replacing the temperature,
and when a correlation was then attempted to describe T as the
response variable (y), no correlation was possible. This lack
of correlation between T and the independe~t variables is
attributed to the hypothesized bivariance of temperature (see
Sec. 6.5.3).
The use of bed temperature as an independent variable
in the performance model is acceptable since, in the commercial
design, the bed temperature and its rate of change, will be
used to vary the supplementary coal feed rate, or the preheat
of CBC
air.
were:
The important response variables used for modelling
the combustion efficiency, CE, and the residual oxygen,
The other gas constituents which were treated as response
RO.
var iables were C02' S02' and NO.
6.4.2
Results
The results of the tests, as described in Appendix F,
were utilized to:
(a) determine the form of the model which best related
all of the data (Equation 6 was the result, and
(b) compute the regression coefficients, bO through b6'
for the models. The values in Table F-2 (Appendix F, page
F-9) were the result. The coefficients for CE, RO and C02
are also listed in Table 9.
In developing the form of the model, the principal
response, CE-combustion efficiency, served as the primary
criterion for evaluating alternatiye forms.
The reliability of the models expressed in terms of
POPE. EVANS AND ROBBINS
INCORPORATED
-------
95
. F . d R2 .
their respectIve -ratIos an IS given below in Table 8.
TABLE H
RELIABILITY OF CARBON-BURNUP CELL
PERFORMANCE MODEL (EQUATION b)
RESPONSE, Y
F-RATIO*
R2**
CE (Combustion Efficiency)
CO (02)
C02
S02
NO
74.5
26.3
23.0
9.2
6.0
.935
.836
.816
.642
.535
*The F-ratio is a statistical measure of the variance in the
response data which is explained by the model divided by the
residue. (unexplained) variance. If there were no re 1ation-
ship between the independent variables and the response, the
F-ratio would tend to take on a value of one. In general,
the higher the F-ratlo, the more reliable the model is in
explaining the trends in the data.
**The parameter R2 is defined as the explained v~riation
divided by the total observed variation. As R approaches
unity the model becomes more reliable.
It may be seen from Table 8 that
and C02 as responses are reliable while
are not.
the models for CE, RO
those for S02 and NO
Another method of demonstrating the reliability of a
model is to plot the observed results against the calculated
results. Figures 17 and 18 present such plots for CE and RO
POPE. EVANS AND ROBBINS
INCORPORATED
-------
~ .'
96
. ..
.....
a
i
~
-
x x
~
10
SYMBOL
. TEST USED IN DERIVATION OF MODEL
x TESTS NOT USED IN DERIVATION OF MODEL
(SEC. 6. 4 . 3 )
D TES""S IN WHICH COAL \vAS ADDED TO
FLY ASH FEED (SEC. 6.4.5)
6 TEST IN WHICH LIMESTONE BY PRODUCTS
WERE PRESENT
l!&
8
...
tJ
z
fq
H
U
H
f1c
rz.
fq
z
o
H
E-t
U)
::>
~
o
u
70
.
60
60 70 80
COMBUSTION EFFICIENCY, % (OBSERVED)
90
100
FIGURE 17
COMPARISON OF OBSERVED AND CALCULATED COMBUSTION
EFFICIENCY IN FBC TESTS
POPE. EVANS AI'JD ROf3BI\JS
I~'r~ ~~: :~:/,;;:::~' :.):I;~~,J~Y~ ~
-------
97
...J
o
(f)
w
0::2
. 1/
.
. )<
. ;(
.
/ .
.
i/ .
.
7 .
.
/
9
-
o
w
>
0::
W 0
(f)
m
o
-
o
0'7
..
(f)
0
...J
LL
z5
z
w
~4
x
o
I
2
3 4 5 6 189
RES I D U A LOX Y G E N I NFL U E GAS t % (C A L C U LA TED)
FIGURE 18 COMPARISON OF OBSERVED AND CALCULATED
RESIDUAL OXYGEN LEVEL
POPE. EVANS AND ROBBINS
:j:.~ ;',-~. ~}.+-'j '---':~;:II::-~~::::'~T'.:~ ':']1
-------
98
respectively.
The 450 line represents an ideal correlation.
The regression coefficients for CE, RO and C02 listed
ln Table 9 will be used in sample calculations which follow
immediately. (S02 and NO are discussed later in this section).
6.4.3
Test of Model in Same Apparatus
To illustrate the use of the basic model equation,
Equation 6, for predicting combustion efficiency, residual
oxygen and carbon dioxide, the data for FBC tests C-3l9-l
and C-3l9-2 will be evaluated. FBC tests C-3l9-l and C-3l9-2
were performed after the derivation of the model was complete
and were not used in the model preparation. In addition, the
apparatus had been rebuilt to reduce the heat loss to the
enclosure. So that the meaning of each parameter will be
clear, a sample calculation is provided on page D-l.
TABLE 10
ILLUSTRATION OF CARBON-BURNUP CELL
PERFORMANCE MODEL (EQUATION 6)
Test No.
319-1
319-2
Combustion Efficiency, %
Calculated 93.3 94.0
Observed 89.0 90.6
Residual - 4.3 - 3.4
Carbon Dioxide, %
Calcula ted 13.5 12.9
Observed 11.8 11.8
Residual - 1.7 - 1.1
Residual Oxygen, %
Calculated 7.3 7.9
Observed 8.2 8.7
Residual +0.9 +U.8
POPE, EVANS AND ROBBINS
INCORPORATED
-------
TABLE 9. REGRESSION COEFFICIENTS FOR USE IN RESPONSE FUNCTION (EQUATION 6)
TERM TO WHICH
COEFFICIENT IS APPLIED DE SIG-
AND APPLICABLE RANGE NATION DIMENSIONS COEFFICIENTS*
For For For
Combustion Carbon Residual
Efficiency Dioxide Oxyqen
Constant bO None - 13.78 - 2.498 + 22.91
~ Bed Tempe ra ture b1 0 -1 0.05193 + 0.007547 0.007353
F +
(1,750 < T2-2,140oF)
~~
Air Rate b1 hr/1b + 0.0462 - 0.01190 + 0.01300
~(f) (330< A 2-.820 1b/hr)
g ~ Bed Height -1 0.3831
~ 0 b3 ln + + 0.1382 0.1390
~;o (10< H< 22 in)
~
Carbon Rate (C) b4 hr/1b 0.8737 + 0.1589 0.1769
~ (26.::.. C 2-.96 1b/hr)
Inert Rate (I) b5 hr/1b 0.1905 + 0.01259 0.01756
(31.:. I .::. ~06 1bs/hr
Cross Product (C) (I) b6 hr2/1b2 + u.00270 - 0.0002862 + u.0003587
*Coefficients are listed here as derived from FBC data. One method of applying the
model to other apparatus is to assume that cross-sectional area is the appropriate
scaling factor. This topic is discussed in Section 6.5.2.
I.D
I.D
-------
100
The residuals for combustion efficiency are higher
than those of the majority of tests used in preparing the
model, but are not the most deviant (see Appendix F). This
illustrates that the model is probably a useful engineering
tool for predicting the performance of a Carbon-Burnup Cell
from a knowledge of the ingredient rates, composition and
the bed temperature.
A second test of the model was made when the results
of the statistical analysis were reviewed. As presented in
Appendix F there was a residual of +2.7% between the observed
and calculated CE's for Test C-3l6-2. When the correct data
were used the calculated value matched the observed value
exactly.
6.4.4
Relative Siqnificance of Parameters in Model Equation
In order to illustrate the relative importance of the
control variables a simple parametric study was done using
Equations 6, 7, and 8 and the regression coefficients in Table
9 for combustion efficiency.
6.4.1.1 Effect of Bed Temperature, T
Table 9 shows that for each 19.30F rise in bed
temperature a 1% increase in combustion efficiency is
predicted by Equation 6, all other independent variables
remaining constant. Figure 19 shows the bed temperature at
which a 90% combustion efficiency will be achieved as a
function of energy release rate for several bed depths and
with A/C = 14 (21% excess stoichiometric air), c = 0.5. One
curve is shown for A/C = 17.3 (50% excess stoichiometric air).
Solid lines are used where the result is reasonable, and
POPE, EVANS AND ROBBINS
INCORPORATED
-------
1.6
1.4
1.2
\0
I
o
r-f 1.0
>C
D::
:x:
~ 0.8
~
"-
~
8
CQ
.. 0.6
r.:I
8.
~
~ 0.4
z
H
D::
H
~
0.2
101
A/c= 17.3
,
Alc = 14 . 0
\
\
20"
"
CE = 90%
c = 0.5
SEE TEXT
FOR DESCRIP-
T ION 0 F
CALCULATIONS
,
/
/
/
1700
FIGURE. 19
I
,
I
. /
USEFUL RA.NGE
---OUTSIDE USEFUL
RANGE
1900
2100
2300
BED TEMPERATURE, T, of
EFFECT OF BED TEMPERATURE ON CARBON-BURNUP
CELL PERFORMANCE MODEL PREDICTIONS
POPE. EVANS AND ROBBINS
-------
102
dotted lines show Equation 6's predictions which are probably
outside the equation's useful range. The curvature shown in
Figure 19 is inherent in the Equation 6 because a square term
enters as the cross-product of carbon rate and inert rate and
this becomes significant at very high throughputs.
Figure 19 clearly illustrates that as the throughput
is increased, the bed temperature (or bed depth) must be
increased to maintain the desired 90% combustion efficiency.
It is also seen that the air to carbon ratio may also be
increased in lieu of bed depth or temperature to maintain the
desired combustion efficiency.
6.4.4.2 Effect of Bed Depth. H
Table 9 shows that for each 3.6" increase in bed depth
a 1% increase in combustion efficiency is predicted. Figure
20 shows the bed depth at which 80, 90 and 100% combustion
efficiency will be predicted by Equation 6 as a function of
throughput for a stoichiometric air-to-carbon ratio. The bed
temperature will be maintained constant at 20000F and again,
c = 0.5.
6.4.4.3 Effect of Air Rate. A
Table 9 shows that the combustion efficiency, predicted
by Equation 6 for the FBC, will increase by 1% tor each 22 lb/
hr increase in air rate, equivalent to 25 lbs/hr ft2 of grate
area. This increase in air flow will also increase energy loss
from the bed (in the form of sensible heat of the flue gases)
by roughly 10,000 Btu/hr while the 1% increase in efficiency
re~eases an additional 5000 Btu/hr. Therefore, to maintain a
constant bed temperature, the heat transfer surface must be
POPE, EVANS AND ROBBINS
INCORPORATED
-------
103
1.4
CE= 80% 90% 100~
I ! .
I , T
I , I
I / I
,
I I
1 '
I I I
I I I
,
I I '
, ,
I ,
/ I '
/ / I
I ASSUMPTIONS
I T = 20000F
c = 0.5
I / I SEE TEXT
I I FOR ADDITIONAL
I DATA
I
/ / I
I '
1
I II /
/1 /
1
1 /
I j / USEFUL RANGE
--, ---OUTSIDE USEFUL
/ I RANGE
/ /
~,/ /
/
1.2
\0
,
~. 1.0
><
P:
:z:
8 0.8
f:L.
"
::>
E-4
~ 0.6
..
~
E-4
~
C> 0.4
z
H
P:
H
~
0.2
o
20
40
60
80
100
STATIC BED DEPTH, H, INCHES
FIGURE 20
EFFECT OF STATIC BED DEPTH ON CARDON
BURNUP CELL PERFO:'!':.U.nCE HODEL PREDIC7ImJS
POPE, EVANS AND ROBBINS
".1...'.
-------
104
reduced by approximately 0.05 ft2 (See Section 6.2.5).
A more useful means of describing the effect of air
rate is to discuss the excess stoichiometric air as given by
the ratio A/C. For C + 02 = C02 the stoichiometric oxygen
requirement is met with 11.53 pounds of air per pound of
carbon. Figure 21 shows the excess stoichiometric air
required for Equation 6 to predict 90% combustion efficiency
as a function of throughput. Bed temperature is set at 2000oF,
and bed depth at 24". The carbon content of the fuel is varied
in this example.
6.4.4.4
Effect of Carbon Content of Fly Ash Fuel, c
To prepare Figure 21, Equation 5 was rewritten:
Y = b +
o
b1 (T) + b2 (AjC) (C) = b3 (H)
(C) + b5 (I/C)(C) + b6 (C)(IjC)(C)
(6A)
+ b4
From Equations 7 and 8 it may be seen that:
I/C = (l.O-c)
c
(9)
where
o
-------
\D
I
o
r-f
~
~ 0.8
N
8
~
..........
::>
8 0.6
.~
..
~
8
~ 0.4
t!J
z
H
p::
H 0.2
~
105
1.4
"
,
, ' r
, \ I
, \ I
ASSUMPTIONS " \ I
\ I
CE = 90% '. ,
T = 20000F I' \ T
H = 24" \
" \ ,
' \ \
'~
USEFUL It \
RANGE , I \
---OUTSIDE ' ""
/ '
USEFUL \
RANGE \
\
IJ \
\
\
}
;,' / /
.,~
,,,,, "
., //
,. j
1.-'"
-," ./ /
;'
.JI'
C = 0.75
T/c= O~ 33 c = 0.5 c = 0.25
l/c = 1. 0 I/C= 3.0
1.2
1.0
-40
-20
40
80
20
60
o
EXCESS STOICH:OMETRIC AIR, %
FIGt,JRE 21
EFFECT OF EXCESS STOICHIGrv1ETRIC AIR ON
CARI30N-BURNUP CELL PERFORH1\NCE HODEL
PREDICTIONS
POPE, EVANS AND .ROBBINS
-------
106
~
tI:
.......
tJ)
fQ
~
..
tJ) 100
f~
8
~
tJ)
tJ)
0
~
Z
0
~
6
Q
~
Q
~
~
~
~
fI.1
Z
H
..
Q 10
fI.1
fI.1
~
5
~
6
5
FIGURE 22
500
o
\
\
\
\
\
CARBON FEED RATE, C
INERT RATE, I
I = C (~ - C)
\
\
\
\
\
....
'-- CARBON
LOSS RATE, C~ =C{1-i~o
COMBUSTION EFFICIENCY, CE
0.1
0.3
0.4
0.5
0.9
0.6 0.7
0.8
0.2
c, CARBON CONTENT OF FLY ASH FEED, LBS C/LB FEED
PREDICTED EFFECT OF CARBON CON'rENT OF FLY ASH
FEED ON PERFOR:1A.~CE OF CARBON-BUHNUP CELL
POPE. EVANS AND ROBBINS
: H'fX:: Rf"':"R:'TEO
o
20
30 clP
..
40 ~
u
50 t1
H
608
~
~
~
70
~
o
H
8
U)
~
80 r~
'::'
o
U
..
fi1
U
90 ~
95
1.0
-------
107
ash, C ,
a
predicted by the
CE
Ca = C (1-100)
model equation and
(10)
is plotted as is the inert rate, I.
Figure 22 reveals that the smallest carbon loss is
realized if the fuel is relatively rich in carbon; i.e.,
c > 0.3. The figure also reveals why a fuel low in carbon;
i.e., c~0.2 will not be a suitable feed to the Carbon-Burnup
Cell of a Multicell Fluidized-Bed Boiler. As c~ 0 the inert
rate becomes so high that it would be impossible to maintain
combustion; the ash sensible heat would remove more energy
than the fuel could supply. For the air rate and temperature
chosen in the example, the amount of heat transfer surface
becomes zero at c'" 0.2. Below c = 0.2, a temperature of
20000F is no longer possible even in an adiabatic chamber.
The carbon content of the fly ash feed is not a true
control variable in a Multicell Fluidized-Bed Boiler since the
quantity and fuel value of the fly ash fed to the Carbon-Burnup
Cell depends on the operation of the primary cells.
Although the topic of design will be discussed later,
it will be useful to show here, by way of a sample calculation,
how the combustion efficiency of the primary cells affects the
operation of the Carbon-Burnup Cell by affecting the carbon
content of the fly ash feed.
Assume that a coal having the following ultimate
analysis is to be fed to a Multicell Fluidized-Bed Boiler:
C = 72.0%, H = 5.5%, N = 1.5%, 0 = 10.0%
S = 3.0%, Ash = 8.0% where the values are in weight
% for the elements shown. In this example, the hydrogen and
oxygen values have been adjusted to include the moisture.
POPE, EVANS AND ROBBINS
INCORPORATED
-------
108
This is a deviation from the normal analysis format.
Dulong's formula
Using
o
Btu/1b = 145.44C + 620.28 (H - 8 ) + 40.50 S
the higher heating value of this coal is 13,229 Btu/1b.
Assume that all of the mineral matter in the coal appears as
fly ash and that 99.5% of the fly ash is captured in an
electrostatic precipitator. This fly ash will be fed to the
Carbon-Burnup Ce11(s). Using the methods of the ASME Power
Test Code (21) and the model equations, Figure 23 shows the
combustion efficiency for various values of "combustible
content of refuse"; i.e., loss-on-ignition of fly ash. Two
values are shown for combustion efficiency in the primary
cells: Curve "A" for the combustion efficiency of all the
fuel constitutents of the coal C, H, S; and Curve "B" for the
(11)
combustion efficiency of the carbon content of the coal alone.
Note the loss-on-ignition of fly ash is assumed to be almost
entire 1y carbon. Curve "C" shows the combustion efficiency
for fly ash carbon fed to the Carbon-Burnup Cell, obtained
from Figure 22). Curve "D;' shows the overall combustion
efficiency assuming that the Carbon-Burnup Cell fires the fly
ash under the conditions used to generate Figure 22.
From Figure 23 it may be seen that the overall
combustion efficiency of a Mu1tice11 Fluidized-Bed Boiler is
relatively insensitive to the carbon content of the fly ash
from the primary cells. This follows from the fact that as
the primary cells become less efficient (Curves A and B) the
Burnup-Ce11 can become more efficient (Curve D). Optimization
is discussed in Section 8.
( POPE. EVANS AND ROBBINS
INCORPORATED
-------
100
dP
...
~
u
z
~
H
U
.H 80
.~
~
~
Z
o
H
8
t/)
o
t:Q
~
o
u
109
90
70
60
o
CURVE "D"
OVERALL co~mUSTION EFFICIENCY
OF MULTICELL FLUIDIZED-BED BOILER
SEE TEXT FOR
ASSUHPTlm~s
"CURVE "A"
. COMBUSTION EFFICIENCY
IN PRD1ARY CELL FOR
ALL FUEL CONSTITUELJTS
IN COAL ec, HAND S)
CURVE "13"
COMBUSTION EFFICILNCY
IN PRIMARY CELL FOR
CARDON CO:'ITENT OF '
COAL ONLY
CURVE "C"
COHI3USTION EFFICIENCY
OF CARI3ml- I3URNUP CELL
10
20
60
70
80
30
40
50
c, "COMBUSTIBLE CONTENT OF REFUSE"
. FI~URE 23
PREDICTED COHBUSTION EFFICIENCY vs. CARBON CONTENT OF
FLY ASH FROH P~Il11'.R.Y Fl,UIDIZED-BED BOILER CELL
POPE. EVANS AND ROBBINS
INCORPORATEO
90
-------
110
6.4.4.5
Effect of Flrinq Rate, C
In Equation 6A, the parameter C appears in four terms.
To illustrate the significance of this parameter, Figure 24
was prepared which plots as "Y" the combustion area per 106
Btu of heat release, and as ''X'' the firing rate. X and Y here
are defined by:
C, 1b/hr
X = 0.86 ft£ x 14,100 Btu/1b = 16,400C, Btu
(fired)/ft2 hr
CE 2 6
Y = 60.99 C (100)' ft /10 Btu (released) per hr
o
For this example I/C = 1.0, H = 24" and T = 2000 F. Two
values for A/C are used; A/C = 16.2 (40% excess stoichiometric
air) and A/C = 11.6 (0% excess stoichiometric air).
It may be seen that the reduction in combustion
efficiency which results from increasing throughput has little
effect on the required unit size; the over-riding factor is
throughput. Halving the firing rate (Btu/ft2hr) would require
doubling the area of the Carbon-Burnup'Ce11. It was not
meaningful to plot combustion efficiency on Figure 24, instead
10Q-CE was plotted. It may be seen that combustion efficiency
does decrease with increasing throughput.
6.4.4.6
Conclusion Reqardinq Siqnificance of Parameters
The preceding paragraphs have discussed the effect of
bed temperature, bed depth, excess air, firing rate and carbon
content on the combustion efficiency of the Carbon-Burnup Cell
as this efficiency is predicted by the model equation. It was
noted that carbon content is not a true control variable
although it is a design consideration. Each of the following
changes will increase the combustion efficiency by 1%, all other
POPE, EVANS AND ROBBINS
INCORPORATED
-------
~
=
~
f1:I
P.
o
rz~
U)
~
f1:I
H
~
I
::>
8
t:Q
\0
o
....
"-
N
[...
f<..
,tIP
~
~ci
tJ
~ I
00
[...0
U)~
::>-
~
::E:O
8~
II
:>4
FIGURE 24
100
'.
111
10
1
20,000
RANGE
Alc, LBS AIR/LB CARBON
(EXCESS STOICIIIO~mTRIC
r
AIR,%)
,
~Y, RELATED TO
, COMBUSTOR AREA
,
,
,
,
,
,
,
11.6(0%)
",.,....--
;,,'"
./
/'
/'
;" "
"
,,;,,"
ASSU!.1!'TIo:m
YTc = 1. 0
T = 2,00ooF
II = 24"
SEE TEXT FOR DISCUSSION
"-
"-
, ,
, ,
, ,
. ,,,-
16.2(40%)~ "-
, "
, "
, -
100,000 1,000,000
2
X = FIRING RATE, BTU-FIRED/FT -HR
EF:"ECT OF FIRING RATE ON PREDICTED PLAN AREA
OF CARBON-BUP~UP CELL
POPE. EVANS AND ROBBINS
.~, f ;i,,';~:-;-;,:: ~:I;:;,c.:q.:?,Tr::::.-=;-1
-------
112
parameters held constant:
a. An increase in air rate of 25 1bs/hr ft2
b. An increase in bed depth of 2.6 inches
c. An increase in bed temperature of 19.30p
As noted previously, each of these changes, a-c,
may have an undesirable effect on either capital costs,
operating costs or both. It should also be clear that in
a real system any change in one parameter will affect all
other parameters. Still, viewed in isolation, the following
conclusions may be drawn:
a. The air rate (or really air to carbon ratio)
would be increased until the gain in combustion efficiency
is matched by the loss as sensible heat in the extra products
of combustion.
b. The bed depth of the Carbon-Burnup Cell would
either be fixed at the value of the primary cells (one fan
for whole system) or, if a second fan were used, the bed depth
in the Burnup-Cel1 would be increased until the gain in steam
production due to increased combustion efficiency equaled
the value of the additional fan power.
c. The firing rate (actually superficial velocity)
must be roughly equivalent to that in the primary cells or a
isolated Carbon-Burnup Cell would be required; e.g., if the
Burnup-Ce11's firing rate were 1/4 that in the primary cells
the primary cell bed material could not be properly fluidized
when it flowed into the Carbon-Burnup Cell.
d. The bed temperature would be maintained
highest level consistent with preventing excessive
agglomeration. Excessive agglomeration" may occur at
at the
POPE, EVANS AND ROBBINS
INCORPORATED
-------
113
temperatures above about 20500F where bed particle growth
begins. If the bed particle growth were accommodated in the
design (See Section 6.6.5) the limiting temperature would be
the bed collapse temperature, possibly as high as 2400oF.
It will be shown later that the simple linear forms
of the parameters are not appropriate for optimization.
What should be obvious from the preceding discussion
is that a high bed temperature ( > 1900oF) is desirable. In
the next section it will be shown how a high bed temperature
is maintained.
6.4.5
Maintaininq the Bed Temperature for Hiqh Combustion
Efficiency: Effect of Coal Addition to the FBC
Figure 19 shows that if the Carbon-Burnup Cell is to
operate efficiently at high throughputs with a shallow bed
and moderate levels of excess stoichiometric air, the bed
temperature must be maintained at a relatively high level.
There are problems in maintaining the bed temperature at a
high level in view of possible variations in the fly ash
supply and its heating value which will occur in a commercial
system.
One convenient method of maintaining a constant bed
temperature is to add coal to the fuel fed to the burnup-cell
at a rate sufficient to maintain the bed temperature at the
desired steady state value. One important question to be
answered was how coal addition would affect the combustion
efficiency of the fly ash. A second question was how useful
the mathematical model would be when part of the fuel was coal.
Therefore, some experiments were run in which coal was fed
with the fly ash. Four tests were run: C-3l4-4, C-3l9-3, 4
and 5. In test C-3l9-5, no fly ash was fed, only coal, to
POPE, EVANS AND ROBBINS
INCORPORATED
-------
114
see if the mathematical model would work in the extreme.
Table 11 below presents the values of carbon
combustion efficiency predicted by the mathematical model
and the observed efficiency for these. tests.
TABLE 11
PREDICTED AND OBSERVED CARBON COMBUSTION EFFICIENCY
FOR CARBON-BURNUP CELL OPERATION WITH COAL IN FEED
Ratio Carbon Combustion Efficiency, %*
Test Fly Ash Carbon/Coal
No. Carbon Observed Calculated Residual
j 14 -4 3.03 89.5 88.5 +1
319-2 00 90.6 94.0 -3.4
319-3 6.97 92.0 95.5 -3.5
319-4 1.14 92.9 99.8 -6.9
319-5 0 95.5 103.6 -8.1
*Resu1ts not adjusted for hydrogen content of fuel.
It is seen that the actual carbon combustion
efficiency is maintained at approximately 90% over a wide
range of the ratio of fly ash carbon to coal carbon.
The mathematical model is less useful in predicting
carbon combustion efficiency when coal is added than when
fly ash alone is the fuel. However, the calculated and
observed values may be adjusted by including the hydrogen
in the fuel in the re.su1 ts. The adjustment consisted of
treating two moles of hydrogen as if they were one mole of
carbon, and using the adjusted carbon rate to predict an
POPE, EVANS AND ROBBINS
INCORPORATEO
-------
115
adjusted combustion efficiency.
in Appendix I.
This is described in detail
TABLE 12
ADJUSTED* VALUES OF COMBUSTION EFFICIENCY
WHEN COAL IS FED TO A CARBON-BURNUP CELL
Test
No.
Ratio
Fly Ash Carbon/Coal
Carbon
Predicted and Observed Values of
the Adjusted Combustion
Efficiency for a Carbon-Burnup Cell
Observed Calculated Residual
o
90.5 86.7 +3.8
92.4 94.5 -2.1
93.7 97.0 -3.3
96.3 98.5 -2.2
314-4
319-3
319-4
319-5.
3.03
6.97
1.14
*See Appendix I for adjustment technique.
Table 12 shows that the statistically derived model
is capable of predicting the combustion effi~iency of a
Carbon-Burnup Cell when coal is added to the fly ash, or
even when coal is the sole fuel. Results for Test 314-4
should be disregarded since the 10" static bed used is too
low for efficient fly ash combustion.
The requirements for accurate prediction are there-
fore that the carbon rates be adjusted and that combustion
efficiency be defined on a heating value basis (See Appendix
I). A further requirement is apparently that the bed depth
be on the order of 18" or more. This latter requirement
follows from the significance of bed depth to the combustion
of fly ash and its lack of importance in coal combustion.
POPE. EVANS AND ROBBINS
II\ICORPORATED
-------
116
The use of coal as a method for maintaining Carbon-
Burnup Cell bed temperature has been shown to be effective.
High temperature in turn leads to high combustion efficiency
in the Carbon-Burnup Cell. However, as the ratio (Coal fed
to the Carbon-Burnup Cell):(Coal fed to Multicell Fluidized-
Bed Boiler) 1ncreases, the overall combustion efficiency must
drop. This follows from the fact that the coal fed to the
burnup-cell has only one pass through the system. This may
be understood by way of an example:
Assume that the data for tests C-3l9-2, 3, 4 and 5
represent the operation of a Carbon-Burnup Cell burning the
fly ash collected from the flue gas of the primary cells of
a Multicell Fluidized-Bed Boiler. Test 319-2 is assumed to
represent full load operation; i.e., the Burnup Cell is.
operating at design capacity, and no coal need be fed to
this cell. In tests 319-3, 4 and 5, the load on the pr1mary
cells is assumed to drop while the FBC (burnup-cell) operates
at an essentially fixed air rate, 700 lbsjhr; thus coal must
be fed to maintain the Burnup Cell bed temperature. If it
is assumed that the primary cells operate at 90% combustion
efficiency at all loads the overall combustion efficiency for
this particular case would be as represented by the bottom
curve in Figure 25. Note that the data points on Figure 25
apply to the case when the burnup cell is not capable of
turndown. Other curves on Figure 25 are plotted for a
burnup cell with turndown (i.e., for the case when coal
feed is not necessary when fly ash feed to the Burnup Cell
1S reduced).
Tne use of high-sulfur coal in the Carbon-Burnup
Cell could lead to relatively high levels of 802 in the
Carbon-Burnup Cell off-gas. This higher emission might not
POPE. EVANS AND ROBBINS
INCORPORATED
-------
RATIO:
COAL TO CARDON-BUru~UP CELL/COAL TO BOILER
117
.012 0
100
1.0
0.45
0.110.074 .048
.028
0.26
0.17
PROJECTED PERFORHANCE IN LARGE SCALE APPARATUS) (ESTIIIZ\TE)
- - - - ---- - - - - - - - - - - - - -
BASED ON FDC D~~~
99
---
~';
98
97
'\
TURNDm.m FOH CjmDOi~-BURNUP CELL
96
95
94
100
o
10
60
80
90
40
50
70
20
30
MULTICELL FLUIDIZED-BED BOILER LOAD,
PERCENT OF rULL LOAD
FIGURE 25 PREDICTED CQ!.1BUSTION EFFICIENCY OF HULTICELL
FLUIDIZED-BED BOILER
POPE, EVANS AND ROBBINS
INCORPORATED
-------
118
be acceptable if when the CBC off-gas is mixed with the
larger stream of boiler off-gas, the tota1,S02 content of
the combined gas stream exceeds the limit established by
the local air pOllution control authorities.
Two alternatives are available to insure that a
fluidized-bed boiler in which fine limestone injection is
used does not exceed the required 802 emission level.
1. Carbon-Burnup Cell bed temperature maintenance
may be accomplished without the use of coal by the inclusion
of a high temperature air heater in the boiler circuit.
Under full-load conditions the air from the high-temperature
heater would flow into the boiler circuit ana the Carbon-
Burnup Cell would recelve ambient air. As Carbon-Burnup Cell
bed temperature tended to drop, the high temperature air
stream would be partially diverted to the Burnup-Ce11 in order
to maintain the desired bed temperature. At the lowest fly
ash feed levels, the Burnup-Ce11 would be receiving only high
temperature air (approximately 1000oF).
2. Alternatively, Carbon-Burnup Cell off-gas could
be wet scrubbed and the residual 802 removed by the lime also
contained in the gas stream. The cold wet Carbon-Burnup Cell
gas would then be mixed with the hot, dry and relatively,
sulfur-free boiler gas. It may be shown by a heat and mass
balance that the mixed gas stream would be only 2SoF cooler
than the primary cell gas and gain very little additional
humidity. The gas buoyancy would therefore be relatively
unaffected.
A different set of alternatives are appropriate if
sulfur emission control is achieved by using coarse limestone
as the bed material and sulfur control agent, and/or if
sintered ash by-products are to be produced in the Carbon-
POPE, EVANS AND ROBBINS
INCORPORATED
-------
119
Burnup Ce 11.
1. The fly ash carbon plus coal might be used as
the reductant in a limestone regenerator. This alternative
is presently under study and will be discussed in a future
report.
2. If sintered ash by-products are to be produced,
it is likely that collected fly ash would be stored so that
the Carbon-Burnup Cell would be operated at a constant load
regardless of the primary cell load. A wet scrubber would
very probably be used for both S02 and particulate control.
6.4.6
The Effect of Fine Limestone Sorbent on Combustion
Efficiency in a Carbon-Burnup Cell
Fine limestone may be injected into the p~imary cells
of a fluidized-bed boiler to reduce emissions of sulfur
dioxide. Limestone by-products would then appear in the fly
ash feed to the Carbon-Burnup Cell. One original concern was
that the presence of limestone by-products in the feed would
adversely affect the combustion efficiency of the Carbon-
Burnup Cell. At the beginning of the program, it was believed
that the limestone would act as an inert diluent to the fuel
and would have less adverse effect on combustion efficiency
than the natural ash content of the fly ash -- if the excess
oxygen in the gas were above about 3%.
A second concern was that the CaS04 in the fly ash
feed might decompose at residual oxygens below 3 to 4% (See
Section 6.6.4); the calcium sulfate produced in the boiler
would tend to be decomposed at the high temperatures of a
Carbon-Burnup Cell. In addition to releasing S02' the
decomposition might be unfavorable to high combustion
POPE, EVANS AND ROBBINS
INCORPORATED
-------
120
efficiency since the reaction
1
C + Ca804 + "2
02 t CaO + 802 + C02
(12)
is endothermic. Note, however, that calcium sulfate
represents an oxidant in the reaction and might increase
carbon consumption. If substantial decomposition of calcium
sulfate occurred in the Carbon-Burnup Cell, the flue gas
from the Carbon-Burnup Cell would have to be treated to
maintain the overall sulfur capture efficiency of the boiler.
Two tests, 320-1 and 320-2, were conducted in the
FBC in which fly ash containing limestone by-products was
the fuel. The fly ash for both tests had been collected
during FBM test B-16 in which fine limestone had been
injected. The collection of this material and its use in
the FBC are potentially dangerous*, and so the number of
tests was limited to two. Test 320-1 was run under strongly
oxidizing conditions (02 = 4.3%) while test 320-2 was run
under only mildly oxidizing conditions to encourage sulfate
decomposition. This test is discussed later.
Table 13 below presents the key results of Test
320-1, the oxidizing test.
Table 13 shows that the 802 emission is only 350
ppm (about 6% of the sulfur input). Therefore, calcium
sulfate does not decompose under the normal conditions of a
Carbon-Burnup Cell; i.e., high temperature, high residual
oxygen. It also appears that the mathematical model over-
predicts the lowering of combustion efficiency by treating
*The fresh fly ash contained CaO, quicklime. It had to be
removed hot from the hopper and placed in an air tight 55
gallon drum. When it was used in the FBC it is poured into
an open hopper. During both these steps fines become air-
borne. Lab personnel wore respirators, gloves, etc. during
these operations and were very uncomfortable.
POPE, EVANS AND ROBBINS
INCORPORATED
-------
TABLE 13
RESULTS OF FBC TEST 320-1 WITH FLY ASH
CONTAINING LIMESTONE BY-PRODUCTS
121
Air Rate, lb/hr
Bed Temperature, OF
Bed Height (static), in.
Fuel Rate, lb/hr
Carbon Content, Wt. %
Carbon Rate, lb/hr
Inert Rate, lb/hr
Sulfur Content, Wt. % (Total organic & sulfate)
02 in Flue Gas (RO), vol. %
Calcium Content, Wt. %
Calculated Combustion Efficiency, %
Observed Combustion Efficiency, %
Residual, %
S02 in Flue Gas, ppm
Value
+
620
1980
20
167
27.9
46.5
120.5
2.42
4.3
19.4
77
87
10
350
the large limestone by-product load as ash. The combustion
efficiency is approaching the desired 90% level and appears
to be relatively unaffected by the high lime by-product load.
It is possible to adjust the inputs to the model by
determining the inert load due to limestone by-products and
deducting that amount from the total load of 120.5 1bs/hr
indicated in Table 13 for the inert rate.
A detailed chemical analysis of the fly ash feed
was not made to determine the amount of the various forms of
calcium (CaO, CaC03 and CaS04).
Instead the following is
POPE, EVANS AND ROBBINS
INCORPORATED
-------
122
assumed:
(a)
95% of
the sulfur is in CaS04; (b)
(not in CaS04) is in CaC03;
(not in CaS04 and CaC03) is
20% of
and (c)
the remaining calcium
the remaining calcium
CaD.
These are reasonable assumptions. When this is done
the inert rate, I, put into the model equation is 60.5 lbs/
hr and the predicted combustion efficiency is 80.8%. The
residual, +6.2%, is more deviant than any of the 38 tests
used to generate the model.
These results make it appear (based on one test)
that limestone by-products enhance combustion efficiency of
carbon in a Carbon-Burnup Cell. What is more likely however
considering the arduous conditions of this test, is that an
error was made in recording either the fuel rate or the air
rate. These two items of data are obtained at the FBC
operating station. A second possibility is that the measured
carbon content of the feed material (27.9%) includes some
carbon as CaC03. If in fact 20% of the calcium not in CaS04
were in CaC03 the observed combustion efficiency of the
elemental carbon would drop to 86.5% and the calculated value
would rise to 82.3%. Unfortunately, a detailed analysis of
the fly ash feed to determine the forms of carbon and cacium
was not done.
For design the model may be used even with fine
limestone present in the fly ash feed if the limestone is
ignored in the calculations; i.e., treat it as containing
two streams - a normal stream of boiler fly ash and another
stream of fine sorbent ignored in employing the model.
The results of test 320-1 reinforce our finding that
the inherent inert content of a normal fly ash is a measure
of the availability of the carbon for further reaction, also
POPE, EVANS AND ROBBINS
INCORPORATED
-------
123
see Curve D on Figure 23. The carbon of a fly ash high in
carbon content is seen as being more reactive than a fly ash
low in carbon content. Despite tne possible flaws, the
results of tests 320-1 lead to the conclusions that limestone
by-products do not adversely affect the operation of a Carbon-
Burnup Cell and of equal importance -- under sufficiently
oxidizing conditions CaS04 does not decompose at the high
temperature of a Carbon-Burnup Cell.
6.4.7
The Effect of Particle Size of Input Fly Ash
Relationships for burning time
to 100Q~m) particles of carbon take the
tb = Kd~
for small (loo~m
form
(13)
where
=
burning time, seconds
tb
K
=
an empirical constant, 300 to 2000
sec/cm2
initial particle diameter, cm
<.g)
d
o
=
Other, more complex forms, have been suggested in
which terms for temperature and oxygen partial pressure are
added.
For a Carbon-Burnup Cell, the time-for-burning is
the time the carbon particle spends in the dense phase
fluidized-bed plus the time the particle spends in the hot,
oxygen-rich freeboard. Tnis time will be about one second
in a commercial Carbon-Burnup Cell.. In the FBC the time-for-
burning is somewhat less. On the ba sis of a one --second time-
for-burning Equation 13 indicates, depending on the value for
K, that particles over about 200~m to 600~m may not burn
out in a Carbon-Burnup Cell. Although experimental data for
POPE, EVANS AND ROBBINS
INCORPORATED
-------
124
part icles less than 100~ID are not ava ilable, part icles under
10~m should burn almost instantaneously in a Carbon-Burnup
Cell.
In all FBC runs, the fuel feed was fly ash captured
by a cyclone-type dust collector, since neither fly ash source
was equipped with an electrostatic precipitator. The efficiency
of a typical cyclone collector is relatively low for particles
below 10 microns in diameter and approaches zero for particles
under one micron. Therefore, the feed material to the FEC was
depleted in the finest size fraction of the gas-borne dust
produced by the furnace. This is shown in Figure 26 in which
the distribution of the fly ash fed to the FBC Test in C-3l2
is shown as Curve A. An efficiency curve for a typical medium
efficiency cyclone is also plotted on the figure as Curve B.
in the collected sample 8 micron particles made up less than
1% of this sample although particles between 7 }lID and 9J,1ID
probably made up 2% of the gas-borne dust entering the fly
ash source's collector.
The shaded area on the figure, marked 1, represents
material that was not fed to the FEC, but which would have
been had the Capitol Power Plant and the FBM been equipped
with electrostatic precipitators.
The effect of this depletion on the performance model
is to understate slightly the combustion efficiency which will
be achieved in a commercial boiler where the entire primary
cell dust stream is fed to the Carbon-Burnup Cell. The
assertion that the combustion efficiency of the Carbon-Burnup
Cell would increase had the fine fraction been included was
not demonstrated experimentally.
based on the assumption that the
out more effectively, as implied
Instead the assertion was
finer material would burn
by Equation 13.
POPE, EVANS AND ROBBINS
INCORPORATED
-------
100
U)
Z
g
U
H
~
..
~
~
8
E:l
,~
H
a
~
H
U
H
E-t
~
~
~
10
700
2'
.. ". . . .. ~, ",.". ... . .
125
'SEE TEXT FOR DISCUSSION I
CURVE "A", INPUT FLY ASH FEED
},':TUAL SIZE DISTRIBUTION
- '
OF COLLECTED DUST
--PROBABLE SIZE DISTRIBUTION
OF GAS-BOillIE DUST
CURVE "B"
EFFICIENCY CURVE,
TYPICAL MBDIUM EFFICIENCY
CYCLONE COLLECTOR
SHADED AREAS G) AND G)
REPRESENT DEPLETION IN FINER
SIZE FRACTION DUE TO LOSS OF
COLLECTOR EFFICIENCY t'lITH
DECREASING PARTICLE SIZE
5
10
15 20
70
30
40
50
60
85
90
80
COLLECTION EFFICIENCY, &~D PERCENTAGE UnDER STATED SIZE,
,BY WEIGHT
FIGURE 26
PARTICLE SIZE DISTRIDUTIONS FOR FBC TEST C-312
.'
POPE. EVANS AND ROBBINS
INCORPORATED
95
98
-------
126
Data from an FBC test which might support this
conclusion is shown below in Table 14, in which the feed
and output for test C-3l4-2 were analyzed for carbon
content versus particle size. Data similar to that used
in preparing Figure 26 is not available for this test.
TABLE 14
CARBON CONTENT OF INPUT AND OUTPUT
FLY ASH AS A FUNCTION OF PARTICLE SIZE
(Data from Test C-3l4-2)
Carbon Content, % by Weiqht
Sample Particle Size
+ 295 Jlm
-295Jlffi + 210 ]lffi
-210 JJm + 125 ]lffi
-125 Jjffi + 38 pffi
38 ]lffi
Feed Fly Ash
69.1
50.6
67..7
54.8
31.4
Average Carbon Content of
Entire Sample
-56.0
18.6
Output Fly Ash
33.5
14.0
21.5
21.7
~
con te nt
ash was
Note that in the input material the lowest carbon
is found in the part icles under 38]lID. After the fly
burned in the FBC the finest fraction had the lowest
carbon content, but this is not the same material that entered
the FBC as fines; i.e., the larger particles become small as
they are burned.
that the
+295JjID ,
The important information provided by Table 14 is
carbon content of large output fly ash particles,
is roughly twice that of the average for the entire
TI~e definitive tests which might have answered the
sample.
POPE, EVANS AND ROBBINS
INCORPORATED
-------
127
question: How does mean particle size affect combustion
efficiency?, were not run since there was no real incentive
to do so. It was already apparent from the data available
that an overall combustion efficiency of 99% could be
achieved by firing primary cell fly ash, without grinding
or other preparation, into a properly designed Carbon-Burnup
Cell. For test 314-2, the FBC operated with a 10" deep bed
which is obviously too shallow, and achieved a combustion
efficiency of 82%.
It would appear from Table 14 that particles above
about 300 ~m elutriating from the Burnup-Cell (these can be
removed from a gas stream simply by a change in direction;
e.g., by a heat exchanger baffle as in the FBM) might be
recycled to the Carbon-Burnup Cell.
Size distribution data for the collected fly ash
FBC test 312-1 is also plotted on Figure 26 as Curve C.
Although carbon content versus size data is not available
from this test, it is reasonable to assume that it would be
similar to that listed above for test 314-2; the overall
from
combustion efficiencies were comparable.
are valid then the fraction of particles
which is about 10% of the mass, contains
unburned carbon for the effluent dust.
If these assumptions
larger than 295 JIm,
about 20% of the
The material not collected by the FBC dust collector
is also depleted in fines shown by the shape of Curve C on
Figure 25. The size distribution of this material, indicated
by the shaded region marked 2, was determined by sampling the
gas leaving the cyclone with an isokinetic probe, and is
plotted on Figure 27. The amount of material collected during
test 312 on the isokinetic probe was sufficient only for the
size analysis shown on Figure 27 and so carbon content data
POPE, EVANS AND ROBBINS
INCORPORATED
-------
128
100
tI)
z
~ --1
u
H
~
.. 0
~
f~
8 I
~ 10 --.~
H -i
Q
~ -',
H ~
U
H
~
,:(
A4 0
-(
-I
1
2%
5
30
50
70
95
:;
PERCENTAGE UNDER STATED SIZE, BY WEIGHT
FIGURE 27
PARTICLE SIZE DISTRIBUTION OF FLY ASH
DISCHARGED TO ATMOSPHERE - FBC TEST C-312
POPE. EVANS AND ROBBNS
INCORPORATED
-------
129
are unavailable for the particulate emisslon during the test.
On the basis of the material presented above the
following conclusions are drawn:
(1) The performance model understates slightly,
the combustion efficiency for a Carbon-Burnup Cell since
the finest fraction of fly ash generated in the source
furnace was not fed to the FBC.
(2) The coarsest fraction may not burn out in a
single pass through a Carbon-Burnup Cell.
(3) If the primary cells are equipped with an
electrostatic precipitator and if the coarsest fraction,
(+300~rn ) of dust from the Carbon-Burnup Cell is recycled,
the combustion efficiency of the Carbon-Burnup Cell would
increase by about 1 to 2% and of the Multicell Fluidized-Bed
Boiler by 0.1 to 0.2%. An obvious alternative to recycling
the coarse fraction would be to grind the fly ash feed to a
fine powder so that all entering particles would be - 100~rn.
The beneficial effect of grinding the feed fly ash would have
to be balanced against operating a pulverizer on a material'
which is 50% (or more) inert. Since for RunC-3l2, the rup
for which the size data were analyzed, an 84% combustion
efficiency was achieved with an untreated feed (mean particle
size 100 microns), grinding of the entire fly ash stream does
not appear warranted.
It is unfortunate that data on particle size
distribution and carbon contents for each fraction were not
obtained for each of the 38 tests used in mOdeling. Had such
data been obtained it might have been possible to reduce the
scatter in Figure .16; i.e., the model would be more accurate.
But, size distribution, just like overall carbon
POPE, EVANS AND ROBBINS
INCORPORATED
-------
130
content, are not real control variables. They result from
primary cell operating conditions; e.g., coal size consist.
It might therefore be argued that the less precise model,
that in which the effect of particle size is ignored, is the
more useful design tool in commercial boilers.
6.4.8
Conclusions Reqardinq StatisticallY Derived Model
for Carbon-Burnup Cell Performance
It has been shown that a model derived by the
statistical analysis of the results of 38 fly ash combustion
tests effectively predicts the carbon combustion efficiency of
the FBC operating as a Carbon-Burnup Cell. The model is use-
ful over the following range of values:
1. Bed temperature 1750 to 2l40oF
2. Bed depth ( s ta t 1 C ) 10 to 22 inches
3. Fuel rate (fLY ash) 60 to 300 lbs/hr
4. Air ra te 330 to 860 lbs/hr
5. Fuel carbon content 28 to 65% by weight
6. Two bed particle size
consists: -16 +22 and -8 +16 mesh
It was further shown that the model may be extended
to include coal as part of the feed by applying the techniques
indicated.
When large quantities of limestone by-products are
included in the fuel stream, the model is less useful even
after the feed parameters are adjusted for the diluent.
However, even here the difference between the adjusted
calculated value and the observed value is within the range
required for a design correlation. It was not found possible
to produce a correlation when heat removal was substituted
for bed temperature as an independent variable in the model.
POPE, EVANS AND ROBBINS
INCORPORATED
-------
131
However, it will be shown later that there is a reasonably
well-defined relationship between heat removal and combustion
efficiency.
The a~curacy of the model
improved if particle size data had
to the model.
would probably have been
been obtained and input
The trends for the parameters are clear and agree
with intuition. Combustion is more complete at hiqher
temperatures, with deeper beds, with lower throughputs and
with higher ratios of air to carbon.
It is possible to use the model equation to design
a Carbon-Burnup Cell if plan area is the appropriate scaling
factor (this is discussed in the next section). It does not
appear possible to use the equation to optimize the design.
This flaw follows from the lack of curvature terms* in the
equation. The model will therefore predict a combustion
efficiency greater than 100% (See Appendix F for a discussion
of the statistical consultant's attempts to correct this
flaw) .
The combustion efficiency of carbon, in fly ash,
ln a Carbon-Burnup Cell does appear to increase linearly
with each parameter from low values to the point where
CE '\, 90%. Beyond where CE ~ 90%, the effect of increasing bed
temperature (or bed depth, etc.) should be diminished by a
square term with a negative coefficient, thus
2
CE = bO + bl (T-alT ), etc.
*Curvature terms comprise square terms plus second order
interaction terms other than C x I.
.. POPE. EVANS AND ROBBINS
INCORPORATED
-------
132
Such a form could not be derived from the small
amount of data in the high CE range. The FBC was simply
not big enough, adiabatic enough and tall enough to explore
the regime near CE -+ 100% at the throughputs which are of
commercial interest.
without curvature terms, there is no basis for trade-
offs. For example, if CE = 90% with a bed depth, H, of 22
inches, the model indicates that CE = 100% will be achieved
if the depth were increased to 48", all other parameters
constant. It may be shown that for reasonable values of
A/c, the air-to-carbon ratio, more energy is made available
by burning the extra 10% of the carbon than is used in
driving the bigger fan required by the deeper bed. But
this is unreasonable; the relationship between combustion
efficiency and bed depth cannot remain linear as CE-+ 100%.
Despite the flaws listed in the
the Carbon-Burnup Cell performance model
useful design tool.
preceding paragraphs,
is judged to be a
6.5
Applyinq the Performance Model in Desiqn
6.5.1
General
The mathematical model which predicts carbon
combustion efficiency was derived from the statistical
analysis of 38 tests conducted in the FBC. While the range
of parameters studied was remarkably wide (only bed temperature
did not vary by a factor of two or more), only one apparatus
was used. There is no absolute certainty that some change
in the design of the apparatus would not have affected the
results markedly.
POPE, EVANS AND ROBBINS
INCORPORATED
-------
133
If the apparatus had been conical rather than
straight-sided, the coefficient for bed depth might have
been totally different. While bed particle size was not
found to have a profound effect on the results, only
sintered coal ash was evaluated, and so a change to, say,
alumina or sulfated lime might have affected the results.
A possible problem in any statistical correlation,
such as this one is that its extension to conditions and/or
apparatus not studied may prove unsuccessful.
The alternative would be a physical model which
describes, analytically, the entire process of fluidization
and combustion. Given the current status of fluidization
theory and heterogeneous combustion theory, a useful physical
model is not available.
Despite the above, the statistical model is felt to
be valid and extendible because of the range of parameters
over which it was derived and because the apparatus was large
enough (and air rates high enough) to minimize wall effects.
The first step in applying the model in design is
to normalize the model, derived in the FBC, so that it may
be used for the much larger systems.
6.5.2
Normalizinq the Model for Use in Other Apparatus
The Carbon-Burnup Cell performance model, Equation
6, was derived from tests in the FBC. The data used in the
derivation were presented in the simplest possible format;
i.e., as they were measured in units such as pounds per hour,
degrees fahrenheit and inches of static bed depth. The
performance model as given by Equation 6 requires information
to be input in this same format. It should be obvious that
POPE. EVANS AND ROBBINS
INCORPORATED
-------
134
each parameter has the implicit dimension -- per "FBC". 80
the air rate should be expressed as pounds of air/hour -
"FBC"; bed depth as inches of depth per "FBC" and possibly
o h.
even temperature as F per "FBC". In t e regresslon co-
efficients presented in Table 9 there is also the implicit
dimension -- "FBC".
Thus for CE; bO = -13.78
"FBC" ,
bl =
+0.05193 "FBC,,/oF, b2 = +0.0462 "FBC" -hr/lb, etc.
To normalize the Carbon-Burnup Cell performance
model it is necessary to express the dimension, "FBC", in
units of mass, length and time. TI1is is the common problem
of determining scaling factors and a different scaling factor
may be required for each parameter.
This is a serious problem, and although a scaling
factor was selected, as described below, the basis was
arbitrary. For this reason the model equation and co-
efficients were left in their original format until this
point in the study.
The basic scaling factor selected is the simplest,
plan area, and affects each regression coefficient as follows:
For
bO (a constant) an "FBC" = 1 (dimensionless)
bl (for bed temperature) an "FBC" =
1 (dimensionless)
2
b2 (for air rate) an "FBC" = 0.86 ft
b3 (for bed depth) an "FBC" = 1 (dimensionless)
2
b4 (for carbon rate) an "FBC" = 0.86 ft
b5 (for inert rate) an "FBC" = 0.86 ft2
b6 (for carbon rate inert rate cross product)
an "FBC" = 0.76 ft4
It is believed that these scaling factors will glve
valid results if it is understood that the carbon rate given
POPE, EVANS AND ROBBINS
INCORPORATED
-------
135
in pounds of carbon per hour per square foot of grate area
must be evenly distributed over this area. It is believed,
but not proven, that the model will be valid for a point in-
. b 2
jector SupplYlng a grate area of a outn ft .
The normalized coefficients will now be derived for
the performance model. It will be useful to do this by way
of a sample calculation for a test in the FBM. The FBM test
used is Test B-14, for which the air rate to the CBC was
1227 lb/hr ft2. The sample calculation is given on page D-2.
No adjustment is required for bed depth or temperature.
The normalized form of the performance model,
presented differently in the summary, is
CE = bO + bl
,
+ b5 (I') +
, ,
(T) + b2 (A) + b3 (H) + b4 (C')
,
b6 (C' x I')
(lA)
where
bO = -13.78 (Dimensionless)
bl = +0.05193 (oF-I)
b; = +0.03973 (ft2hr/lb)
b3 = +0.3831 (in-I)
b~ = -0.7514 (ft2 hr/lb)
b~ = -0.1638 (ft2 hr/lb)
b~ = +0.0020 (ft4 hr2/lb2)
Applying the normalized data to the normalized
performance model (Equation lA), the predicted combustion
efficiency for FBM test B-14 is 95.8%. The observed value
was 95%. The residue is -0.8%.
When FBM test B-15 was analyzed, in which the air
rate to the CBC was 1880 lb/hr, the predicted value of
POPE, EVANS AND ROBBINS
INCORPORATED
-------
136
combustion efficiency was 117% against an observed value of
93%. (The model contains no provision to prevent values
over 100%, See Appendix F). It appears, therefore, that the
model, based on tests with a maximum air rate of 1000 lb/hr
per ft2 (860 lb/hr into the 0.86 ft2 FBC), may be extended
to an air rate on the order of 1250 lb/hr ft2 (25% extrapo-
lation), but it is useless at air rates of 1700 lb/hr ft2
(70% extrapolation); i.e., the model needs a parabolic term
for air rate so that it is not linear at high air rates.
A comparison of observed and calculated values for
CBC tests with lower air rates is given on Figure 37 in
Section 7. Some of the residuals, especially for the earlier
tests, are high; e.g., for test B-5 the residual = +16, an
absurd value. For test B-12 on the other hand where the air
rate was 750 lb/hr ft2 the residual = O.
The following conclusions are drawn regarding the
application of the Carbon-Burnup Cell performance model to
larger apparatus and conditions outside the range used in
the derivation:
1.
The model is useful for devices 50% larger
than the FBC.
2. The model is probably useful for rates* 25% to
50% higher than used in the FBC.
6.5.3
The Bed-Temperature Anomaly
In the preceding section a normalized form of the
performance model was discussed. In the next section the
use of this model in design will be illustrated. But first
*Air, fuel, heat removal, etc.
POPE, EVANS AND ROBBINS
INCORPORATED
-------
137
the problem of bed temperature prediction must be discussed.
It was noted in Section 6.2.5 that it is not
possible to predict the bed temperature of a Carbon-Burnup
Cellon the basis of the three parameters (1) fuel rate,
(2) air rate and (3) a heat loss term expressed as Btu/hroF
which includes heat transfer coefficients and areas; or at
least the attempt made here to do so was unsuccessful. Even
when a fourth parameter, combustion efficiency, is added it
is still not possible to predict bed temperature. It was
further noted that this problem probably arises from three
factors:
(1) the combustion process in the Carbon-Burnup
Cell is bi-stable
(2) not all of the energy release, implied by a
knowledge of the combustion efficiency and fuel rate, is
released within the dense phase fluidized-bed and
(3) the heat loss to surfaces above the bed may be
higher than anticipated.
These three problems are discussed below.
6.5.3.1
Bi-Stable Combustion
During testing (See Table 15, A) it was found
possible to perform two experiments, the first in which the
o
bed temperature was 2100 F and a second at a bed temperature
of l8000F. When the two experiments were complete and the
data reduced it was found that the air rate, carbon rate,
inert rate and bed depth were almost identical in both
experiments. Since no cooling coil was used and the loss
of heat to the enclosure was almost identical in both
experiments, the bed temperature would have been expected
POPE, EVANS AND ROBBINS
INCORPORATED
-------
138
TABLE 15
ILLUSTRATION OF BI-STABLE COMBUSTION IN FBC
A.
Variation in Bed Temperature at Constant Feed Rate(l)
Air Rate, Carbon Rate, Bed Tempera ture, Combustion
Test No. lbs/hr lbs/hr of Efficiency, %
C-306-l 600 41.5 2100 85.5
C-306-4 600 43.0 1800 72.7
C-307-l 600 40.6 1800 76.0
C-307-3 600 44.0 2140 91.0
B. Variation in Carbon Feed Rate at Constant Bed Temperature
C-308-1 600 37.8 1900(2) 81.5
C-308-2 600 42.4 1900(3) 81.3
C-308-3 600 43.0 1900(4) 80.6
1In the two test pairs there was a 3.5 to 7.5% variation
in carbon feed rate but note that in 306 the higher
temperature was at the lower carbon rate while in 307
the higher temperature was found at the higher carbon rate.
2Bed temperature reached by slow rise to steady state.
3Bed temperature reached after 15 minutes at 2100oF.
4Bed temperature reached after 15 minutes at 1500oF.
to be the same. That the system may perform differently
under apparently identical conditions appears at first to
be anomalous.
One possible explanation for the apparent anomaly
is that the data were incorrect, but in a second set of
experiments the anomalous results were again observed. It
POPE, EVANS AND ROBBINS
INCORPORATED
-------
139
was also suggested that the bed might be "conditioned" by its
previous temperature history. Tests 308-1, 2 and 3 (see
Table 15 B) with different preconditions showed that this was
probably not a valid hypothesis. A more suitable explanation
is that a stable "flame" could be achieved at more than one
bed temperature, all other system parameters kept constant.
This may be understood by way of a hypothetical illustration.
Curve A in Figure 28 plots the energy lost by the bed as a
function of the bed temperature. Curve A is a straight line
based on the assumptions that: (1) the major fraction of
energy lost by the bed is the sensible heat in gas and dust
and that the enthalpy rise of these substances is proportional
to the bed temperature and (2) the heat lost by the ,bed to the
walls of the combustor is proportional to the bed temperature.
Curve B of Figure 27 is a plot of the energy released
within the bed, also as a function of bed temperature. The
shape of the curve is taken as the top portion of the classic
S-shaped curve for a combustion process. The shape can be
explained as follows:
(1) At some temperature below T1' ignition occurs,
the rate of combustion of carbon then increases exponentially
with increasing temperature in the region where the process
is reaction-rate limited. At higher temperatures, above that
at which the process becomes diffusion limited, the rate curve
increases more slowly with temperature since the effect of
temperature on the diffusion rate is not so great. (The
temperature at which the combustion process changes from
being reaction rate limited to being mass transfer limited
depends on a number of factors), or,
(2) Similar to (1) above except that the curve
flattens at relatively high temperature simply because the
POPE. EVANS AND ROBBINS
INCORPORATED
-------
140
CUHVE "C", HAXIMUM POSSIBLE
HEAT GENERATION RATE, CE = 100%
Qin
_._~--- -----
- ---------
t
CURVE "B", RATE OF
HEAT GENERATION
o
~~
::r:
z""'
0::>
H8
8a:1
~ '
~U)
~~
t!>...:1
88
~~
r4~
::r:::r:
~~
00
'r4~
88
~~
1
I POINTS 1
I AT NHICH
I
I
I
I
I
I 2
I
I
CURVE "Art, RATE OF HEAT
LOSS
I
I
I
I
I
I
I
I
I
I
AND 2 REPRESENT T''10 OPERATING CONDITIOHS
THER\fAL EQUILIBRIUH IS ACHIEVED
I
I
I
I
" Tl
T2
FLUID BED TEHPERATURE, of --..
RATE OF HEAT GENERATION
= COHBUSTION RATE (LBS OF CARBON PER HOUR/FT2 OF CARBON
SURFACE IN BED) X CARBON SURFACE (FT' OF CARBON SURFACE
IN BED) X HEAT OF REACTION (BTU/LB OF CARBON BUffi1ED)
RATE OF HEAT
.. ENTHALPY
OF INERT
AND FROM
LOSS
RISE OF PRODUCTS OF
PASSING THROUGH BED
BED SURFACE
CO!-1BUSTION + ENTHALPY RISE
+ HEAT LOSS THROUGH WALLS
FIGURE 28
ILLUSTRATION OF BI-S'fABLE COHBUSTION PROCESS
POPE. EVANS AND ROBBINS
:rJf";,rRFC,n,.',Tl:p
-------
141
rate of heat generation cannot exceed the input rate Q. .
ln
The rate of heat generation in a fluidized bed may
be seen to depend on the combustion rate expressed as pounds
of carbon consumed per hour per square foot of carbon
surface, the surface area of carbon existing in the bed at
any time, and the heat of the reaction. The surface area of
carbon existing in the system depends, on the one hand, on
the rate at which carbon is being fed, and on the other, on
the rate at which it is being lost by combustion and
elutriation, and also on the size distribution of the carbon-
containing particles within the bed. The temperature of the
bed is that temperature at which the rate of heat generation
is equal to the rate of heat loss. Two such temperatures are
seen to exist on Figure 28. Point 1, represents the the
thermal balance at a "low" temperature and point 2 at a "high"
tempe ra ture.
In a pulverized coal furnace point 1 of Figure 28
would represent an unstable condition. Note that if a
perturbation slightly increased the rate of heat generation
the temperature would tend to increase. This in turn would
increase the heat release rate and the system would move to
point 2. If, on the other hand the combustion rate dropped
slightly the temperature would also have to drop decreasing
the heat generation rate such that the fire must go out.
Point 1 may be stable in a fluidized-bed combustor
because:
(1)
The thermal inertia of the bed allows the
system to override minor perturbations, and
(2) Figure 28 is drawn only for the Bed, not the
entire system which includes combustion in the freeboard.
POPE, EVANS AND ROBBINS
INCORPORATED
-------
142
When the combustion rate in the bed drops as it might at
Point 1, the combustion rate may increase in the space above
the bed. Energy released in the freeboard is returned to
the bed most probably in the form of the sensible heat
content of the refluxing bed particles. The energy partition
between bed and freeboard is discussed below.
6.5.3.2
Effect of Enerqy Release Patterns on Temperature
of Fluidized Bed
It was shown in the preceding section that the
fluidized-bed of a Carbon-Burnup Cell could achieve a thermal
balance at two distinct bed temperatures. At the higher bed
temperature, a larger fraction of the fuel which enters the
bed burns within the bed than would burn there at the lower
temperature.
It follows therefore that at the lower bed
temperature more fuel appears in the freeboard as unburned
carbon or carbon monoxide. Assuming that the air rate
remains constant, as it did in the construction of Figure 28,
oxygen must also exist in the freeboard and so some fuel will
burn in the freeboard.
In a test run in the FBC it would not be possible to
determine what fraction of the carbon burned within the bed
and what fraction of the carbon was consumed in the freeboard.
The combustion efficiency, derived from a carbon balance, was
for the combustor as a whole.
The partition of energy release, between bed and
freeboard may have a profound effect on the temperature of
the bed as will be shown by the simple illustration which
follows. Figure 29 has been prepared for a hypothetical
Carbon-Burnup Cell and experiment to which the following
POPE, EVANS AND ROBBINS
INCORPORATED
-------
2200
2100
2000
1900
~
o
...
~ 1800
D
8
~
~
~
:E:
~
8 1700
1600
1500
..
143
CURVE "B"
FREEBOARD EXIT TEMPERATURE
c:( : ENERGY RELEASED ~'1ITHIN THE BED
- TOrrAL ENERGY RELEASED \'lI'rHIN COl.lBUSTOR
CURVE "A"
BED TEHPERATURE
ASSUMPTIONS
AIR RATE = 700 LBS/HR
FUEL R7\TE = 75 LBS/HR
CARBON cm1TENT.= 50%
HEAT LOSS FRO~.l:
BED-rrO-l'1ALLS = 70,000 BTU/HR
FREEBOARD-TO-~'lALLS = 0 BTU!IIR
CE = 90%
SEE TEXT FOR DISCUSSION
0.7
.0
0.8
0(
0.9
..
FIGURE 29
EFFECT OF ENERGY REI.F.7\SF; PARTITION, cX.ON TEf.1PEru\TURBS
,
HYPOTlffiTICAL FLUIDIZED-BED COMBUSTOR
POPE. EVANS AND ROBBINS
INr,ORPORl\Tf'n
-------
144
data will apply:
Carbon Content
Heating value of carbon
Combustion efficiency
700 lbs/hr
75 lbs/hr
0.5 lb carbon/lb
14,100 Btu/lb
90%
fuel
Air Ra te
Fuel Rate
The combustor has the following characteristics
(a)
Heat loss from bed to walls, = 70,000 Btu/hr
(This was typical of the FBC where heat loss to
walls was almost independent of bed temperature)
(b)
Heat loss from freeboard to enclosure or bed =
o Btu/hr.
(This assumption is invalid but is useful for the
sake of the illustration)
(c)
Size of combustion chamber = Variable
(The
chamber changes as required to permit CE = 90%
regardless of bed temperature).
Curve A in Figure 29 represents the bed temperature,
Curve B the temperature at the exit from the freeboard for
values of a parameter, a , where
a
=
Enerqy released within the bed
Total energy released within combustor
(14)
The system described in Figure 29 may be viewed as
the extreme case in which the temperature of the bed is
totally independent of the temperature in the freeboard.
It is possible for such a system to exist, for example if
fine coal were fed to a single point at the base of a low
velocity (~2 fps superficial velocity) fluidized-bed
combustor contained in a large refractory enclosure. In
such a case volatiles and coal fines would enter the freeboard
POPE. EVANS AND ROBBINS
INCORPORATED
-------
145
and burn there.
Exit gas temperatures might exceed 25000F.
Data obtained in the FBC tests cannot be used to
calculate values fora. The marked temperature differences
between bed and freeboard implied by Figure 29 were not
recorded. If it had been possible to determine the average
gas composition at the "top" of the expanded fluidized bed
it might have been possible to determine values fora as a
function of the control variables. This was beyond the
scope of the test program described in this report.
The discussion here was not intended to obtain values
fora but to show a few of the implications of such a parameter.
It can also be seen that for the FBC where the high air rates
cause large quantities of bed materials to be thrown into the
freeboard that a mechanism exists for returning energy released
in the freeboard to the bed; i.e., in the sensible heat of the
refluxing particles.
6.5.3.3
Enerqy Loss to Cool Surfaces Above the Dense Phase
Fluidized Bed
In the preceding section it was noted that the bed of
a Carbon-Burnup Cell can receive energy as sensible heat of
particles which return to the bed at a higher temperature than
the temperature at which they were initially expelled from the
turbulent bed. This might occur if the freeboard were well
insulated and if a fraction of the carbon feed burned in the
freeboard instead of within the dense phase bed.
freeboard is not so well insulated, the bed may
to the cool surfaces above the dense phase bed.
If the
lose energy
Part of the inability to predict bed temperature for
the FBC follows from the inability to measure or compute the
energy lost, by the bed, to cooler surfaces above the bed.
POPE, EVANS AND ROBBINS
INCORPORATED
-------
146
In the FBC the cool surfaces above the bed included the upper
part of the bayonet cooling tubes, the hood and the exhaust
duct. In the FBC, it was not even possible to measure the
energy lost from the dense phase bed to the walls in contact
with the bed; the walls received energy from both the bed and
the freeboard and only the total energy input to the walls was
measured.
The transfer of heat from a f1uidi~ed-bed to heat
transfer surfaces above the top of the bed does not appear
to have been treated in the fluidization literature; one
reason being the difficulty of getting useful data with
small scale apparatus. But, despite the lack of quantitative
data on the loss of energy by the bed to cool surfaces above
the bed it is necessary to demonstrate here that such a loss
does sometimes exist, and to indicate how its magnitude has
been estimated. For purposes of illustration, assume that a
fluidized-bed combustor is contained within a tall cyc1indrica1
column.
The wall of the column will be maintained at a
temperature lower than the bed temperature. Under these
conditions energy will be removed from the bed according to
Q r = Ab a (E b T: - ex w T~ )
(15)
where
Q = Radiative heat loss rate from fluidized bed
r
to walls, Btu/hr
Ab = Surface area of top of fluidized bed
cr
= Stefan-Bo1tzman's constant = 0.1713xlO-3
Btu/hr ft20R4
Eb = Emittance of fluidized bed~ 1.0
ex =
w
Absorptance of enclosure walls; 0.8< ex < 0.95
w-
for typical surface walls
Tempera ture of bed near surface ~ average bed
o
tempera ture , R
Tb =
POPE, EVANS AND ROBBINS
INCORPORATED
-------
147
T
o
= Temperature of enclosure wall, F
When (Tb/Tw)4
Equation 15 becomes
« 1 and assuming e: b
=a
W
= 1
4
Qb = Ab a Tb
(15A)
For
Tb = 20000F (24600R)
Btu 2
Qb = 62,600 hr ftL x (Ab' ft )
The energy loss due to radiation to surfaces above
the top of the bed could exceed the loss to submerged surfaces
in a number of practical situations. To use Equation (15A) ln
predicting the bed temperature, it is necessary to estimate
Ab' the area of the top of the fluidized bed.
If the fluidized bed were relatively quiescent, the
surface area of the top of the bed, for the vertical
cylindrical chamber would be
2
~ = Ap = 1T R
(16)
where
A = Area of flat plate
p
R = Radius of vessel
Where the bed is fired at high rates; e.g.,
approaching 106 Btu/ft2hr, the surface would be far from
. quiescent. It may be shown that the greatest surface area,
for radiation, of the most irregular surface is given by
1T
A = ( -) x A .
e 2 p'
A /A = 1T /2
e p
= 1.57
(17)
where
A = Extended surface area (for radiation)
e
....
POPE, EVANS AND ROBBINS
INCORPORATED
-------
148
If the diameter of the fluidized bed were small
enough, the bed would slug at high throughputs and the
extended surface area of the top of the bed would approach
the area of the absorbing enclosure; i.e.,
A
e
(in the limit) ~ /271 RH
slug
(18)
where
H = Effective height to which slugs of bed
slug
material are thrown from the bed and are
still close to the bed temperature.
Even where the bed does not slug, there may be
sufficient material thrown into the freeboard to require the
extended surface to be computed thus:
Ae = 271 R Heff
= P Heff
(19)
where
Heff = Effective height to which particles are
thrown from the bed in sufficient
concentration to approximate a black
radiating at the bed temperature
= Perimeter of vessel, ft.
body
P
In preparing Figure 16, Equation 17 was used to
compute the radiative loss from the fluidized bed. In
previous (unpublished) work it has been found that the
measured bed temperature of the FBM could be reproduced by
an empirical model in which Equation 19 was used when Heff~
3.5 ft and the perimeter follows from the rectangular
dimensions of the enclosure.
It follows from the definition of
very small R, A /A ~ 1, and for very large
e p
At intermediate values of R, A /A reaches
e p
Heff that for
R, A /A < 71 /2.
e p
a maximum. For
POPE, EVANS AND ROBBINS
INCORPORATED
-------
149
the FBC, A /A may have exceeded 10; no attempt to measure a
e p .
real value could be made in the FBC.
The anomalous behavior of bed temperature stems in
part from the inability to partition the energy absorbed by
the hood of the FBC into two components; one primarily
convective having small effect on bed temperature and a second
primarily radiative having large effect on bed temperature.
6.5.4
An Example: Usinq the Carbon-Burnup Cell
Performance Model in Desiqn
Assume a Multice1l Fluidized-Bed Boiler is to be
designed to provide 250,000 1b/hr of steam at 600 psig,
7500F for the sample design calculation which follows.
The performance specification contains an efficiency
guarantee which will require 99% combustion efficiency at
full load.
The coal to be used, at the rate of 27,500 1b/hr,
has the analysis shown below in Table 16. It will be
assumed that just enough of the coal ash will be retained
in the bed to compensate for attrition; i.e., all the ash
eventually appears in the fly ash. Fine limestone will be
injected into the primary boiler cells to reduce emissions
of 802 by 80%; based on past work this level of sulfur
removal requires a Ca/8 ratio of 2.5 and this value is
assumed here. The primary cells will operate so that 3%
residual oxygen appears in the stack gas. At this residual
oxygen level, about 10% of the input fuel energy value
normally appears as carbon in the fly ash. The value 10%
will be used here. The dust collected in the primary cyclone
and precipitator is to be fired in the Carbon-Burnup Cell of
the Mu1tice11 Fluidized-Bed Boiler.
POPE, EVANS AND ROBBINS
INCORPORATED
-------
150
TABLE 16
COAL ANALYSIS FOR FUEL TO FWIDIZED-BED
BOILER, AN EXAMPLE OF CARBON-BURNUP CELL DESIGN
ULTIMATE ANALYSIS
Constitutent
Carbon
Hydrogen
Nitrogen
Oxygen
Sulfur
Water
Ash
Total
Btu/lb as fired = 12,550
%, By Weiqht
69.4
4.7
1.3
7.3
3.2
5.0
9.0
100.0
To insure that no calcium sulfate is decomposed, the
Carbon-Burnup Cell will operate at a residual oxygen level
of 6%. Other cell design parameters are as follows:
Static bed height, H = 24", Bed temperature,
T = 2050oF, Cell length, L = 10' (based on design arrange-
ment) .
Equations 1 and 2, restated below, and C + 02 = C02
will be used in these computations in addition to a straight-
forward heat balance.
CE, % = -13.78 = 0.05l93(T) +
+ 0.3831 (H) - 0.7514
+ 0.0020 (C' I')
.03973 (A')
(C') - 0.1638 (I').
(lA)
RO, % =
+22.91 -0.007353 (T)
-0.1390 (H) - 0.1521
+.000265 (C' I')
+ 0.01118 (A')
(C') - 0.0151 (I')
(2A)
C + 02 = C02 is rewritten below as a mass balance
POPE, EVANS AND ROBBINS
INCORPORATED
-------
151
assuming the oxygen is supplied as air and using the nomen-
clature of Equation 1 and 2.
C(s) +
CE/100
(1.0-4.76x(RO/100)
(CE/100) x C02 (g) + (1.0 -
CE/100
(1.0 - 4.76 x (RO/100) x
+ 3.76 N2(g)
x 02(g) + 3.76 N2(g)
=
CE/100) x C(s) +
(20)
4.76 x (RO/100) x )2(g)
where
C(s) = moles of carbon fed as a solid
02(g) + 3.76 N2(g) = molar composition of air
From the information given above, the following is
known or may be computed:
CE = (99-90)/0.1 = 90, RO = 6, T = 20500F
H = 24 It, L = 10', A' = A/(L x W)
It = I/(L x W), I = 27,500 x (.09) = 2475 1bs/hr
C' = C/(L x W), C = 27,500 (12,500)(0.1)/14,100
= 2,447 1bs/hr
Equations 1 and 2 may be solved simultaneously for
two unknowns, A' and C'. Then the cell width, W, may be
readily computed from W = C/(C' x L). Using this procedure
C' = 69 Ibs/hr ft2, A' = 1,053 1bs/hr ft2.
Alternatively Equation 1 and 20 may be solved
simultaneously. If instead Equation 20 is used as a check
for the air rate substitute C(s) = 69/12 = 5.75 m01es/hr and
the A' (Equation 20) = 1,012 Ibs/hr ft2; a reasonable agreement.
Next compute the cell width:
2447
W = 69x10
= 3. 54 f t
POPE, EVANS AND ROBBINS
INCORPORATED
-------
152
The final task is to perform a heat
Burnup Cell and determine energy surplus (t.o
tubes inserted in the bed) or deficit (to be
added coal).
balance for the
be removed by
made up by
The following information will be assumed as given:
(1) The walls of the cell will be made up of boiler
tubes filled with boiling water at 645 psig (4850F).
(2 )
bed depth.
The expanded bed depth is 1.2 times the static
(3)
The heat flux into wetted surface is given by:
2 0
q = U, Btu/ft hr F
w
= 78,200 Btu/hr ft2
x (ld', of) = 50(2050-485)
(4 )
is given by
(5)
The heat loss due to radiation to the enclosure
2
Qr = 66,500 Btu/hr ft x Ab
The radiating top surface of the fluidized-bed,
Ab = 7T/2 (L x W) = 55.6 ft2
(6)
in that Q = 3,700,000 Btu/hr
r
The combustion air is supplied at 6000F.
(7) Limestone by-products nre fed to the Burnup Cell
with the normal fly ash = 5,815 1bs/hr (assuming limestone is
97% CaC03' that all is calcined in primary cells, Ca/S = 2.5
and 80% of coal sulfur is removed as CaS04)'
(8) Gas leaving the bed at 20500F has an enthalpy
value of 515 Btu/lb.
(9) Dust leaving the bed has an average enthalpy
rise of 365 Btu/lb based on an average specific heat of 0.22
Btu/1b of and an inlet temperature of 4000F.
POPE. EVANS AND ROBBINS
INCORPORATED
-------
153
(10)
The a factor, defined by Equation 15, = 0.9.
A.
Input
1,000's of
Btu/hr
1.
From combustion within bed
( a = 0.9)
=
28,000
4,700
32,700
2. From preheated combustion air =
Total energy input
B.
Output
2.
3.
In gaseous products of
combustion
In dust leaving bed
To boiler tubes contacted
by bed
=
20,800
3,200
1.
=
=
5,100
4.
To boiler tubes viewed by
bed, Q
Total energy output
=
3,700
32,800
C.
Net surplus (deficit)
=
(100)
Based on the assumptions there is an energy deficit
of 100,000 Btu/hr if the Carbon-Burnup Cell is to operate at
o
2,050 F. The deficit may be made up by feeding 10 pounds of
coal per hour with the fly ash.
The sensitivity of this balance to some of the
assumed parameters is as follows:
(1) If a = 0.8, the coal feed rate would have to
be 250 1bs/hr.
If a = 1.0, the energy balance would indicate
an energy surplus, so that 76 linear ft of 2" O.D. boiler
tube would be added to the bed; or the air rate would be
increased to about 1,260 1b/hr ft2.
POPE, EVANS AND ROBBINS
INCORPORATED
-------
154
(2) If A /A . (See Equation 17) = 1.0; air rate
e p 2
would be increased to about 1,170 lbs/hr ft .
If Heff = 5 ft (See Equation 19), the coal feed
rate = 425 lbs/hr or some refractory would replace the heat
transfer surface on one wall.
6.6
Other Results from Tests in the FBC
6.6.1
Effect of Heat Removal
The effect of heat removal on combustion efficiency
is shown in Figure 30.
The simple quadratic
CE = 100 x (1 - 2.3 X2); standard deviation
S = 4.5%
(21)
was used between points
3 the fire would ~o out
curve must fall rapidly
1 and 3 on the figure. Beyond point
if more heat were removed, so the
toward CE = O.
20000F,
by point
apply.
The curve is felt to represent data for T'" 1900 to
A ~ 600 to 800 lbs/hr. For low air rates represented
C, (A = 330 lbs/hr) an entirely different curve would
Equation 21 is not useful as a prediction tool since
the parameter "X" cannot be determined unless CE is already
known, thus
X - Enerqy lost to coolinq surfaces and dust, Btu/hr (22)
- Fuel rate, lbs/hr x Carbon content x CE x 14,100 Btu/hr
The significance of Figure 30 can be understood by
pointing out that the parameter "X" represents the fraction of
the energy release not removed in the flue gas. So the
POPE. EVANS AND ROBBINS
INCORPORATED
-------
E-4
&3
u
~
(1:1 70
Po.
~
~
u
~
~
u
Z 60
(1:1
H
U
H
~
~
(1:1
Z
o 50
H.
8
tI)
::>
~
o
u
100
90
100 (1-2. 3X2)
(j)C
DATA POINTS FROM
FBC TESTS C-306
THROUGH C-319
..
o
155
3
\
\
\
\
\
. \
\ 4
o 0.1 0.2 0.3 0.4 0.5
X = ENERGY RE~OVED VIA. ~~ALLS , HOOD, PROBE At"lD DUST
- ENERGY RELEASED IN COHBUSTION CHi\iiBER
80
40
~.
o
A
.
NOTES: DOTTED LINE FROM
1 TO 2; ASSU:.1ED APPROACH
TO CE = 100%.
DOTTED LINE FROM 3 TO 4
BASED ON LOSS OF FIRE AT
X'I\, 0.5.
POINTS NEAR 2 NITH COAL,
NEAR 3 U~STA.BLE OPERATIO~
POINTS LABELED A, D, AUDC
REPRESENT ATYPICAL TESTS
.f
FIGURE 30
EFFECT OF HEAT REMOVAL ON COMBUSTION EFFICIENCY
IN FBC
POPE. EVANS AND ROBBINS
~ . J: I-;-~Ii~';: -'}: -.:':.:~;: ~~:;.~..-:':-~ :-!I
-------
156
combustion should be most complete when only excess air is
used for cooling. For the sample calculation in the preceding
section the parameter X:::. 0.3, and from Figure 30, CE:::'" 81% for
this value of X, not the 90% used in the previous section.
The model equations, are believed to be more reliable than
the curve sketched on Figure 30.
6.6.2
Nitric Oxide Emissions
It has been shown in our past studies (~) that
nitric oxide (NO) is the predominant form in the NO group.
x
This predominance was demonstrated by comparing IR measure-
ments (specific to NO) with the PDS method which measures
total NO. During the course of the current study, no attempt
x
was made to measure total NO. It is assumed that the NO level
x
recorded by the infrared analyzer cont inued to represent total
NO. In past studies it was shown that when coal was burned
x
in the fluidized bed, NO increased with increasing residual
x
oxygen in the range of 1 to 5% RO (02) . NOx showed no clear
correlation with increasing bed temperature over the range
15000F-18500F.
When the FBC was modified to simulate a Carbon-
Burnup Cell and operated on fly ash in the range 3 to 10% RO
and 17500 to 21400F, the dependence of NO on residual oxygen
x
was no longer apparent; in fact there was no correlation.
The statistical test series did not produce as useful
a correlation between NO and the control variables studied
x
as for combustion efficiency, RO and C02' During the present
tests, a general trend of increasing NO with increasing bed
x
temperature is shown in Figure 31. It may be seen that the
highest values of NO were recorded at the highest temperatures,
and the lowest values of NO were recorded at the lowest bed
POPE, EVANS AND ROBBINS
INCORPORATED
J
-------
800 157
.
.
700
'. .
.
600 . .
..
. . .: Ge
. . .
~ . o.
~ . . 0
~ . .
.. 500 . .8
tI) ..
~. .
. . .
r..1
tJ
~ .
~ .
z 400 .
H
r..1 .
Q .
H
:<
0
~ 300
~
8
H
Z
200
100
1700
2000
2100
1800
1900
BED TEMPERATURE, of
FIGURE 31
NITRIC OXIDE CONCENTRATION IN FLUE GAS'
FROM FBC VERSUS TEMPERATURE
POPE, EVANS AND ROBBINS
2200
-------
158
temperature.
However, the data are extremely scattered.
Although a clear correlation is not evident in
Figure 31, an examination of the individual test results
shows that within each test an increase in temperature
almost always resulted in an increase in NO emission.
typical results are plotted in Figure 32A.
Some
When coal was added as in Run C-3l9, the NO level
dropped with increasing coal content despite the high
temperature and increasing residual oxygen (See Figure 32B).
The results, in which coal and fly ash were interchanged,
suggest that the major determinant in the level of NO
emissions may be the combustion pattern within the bed,
although a number of conflicting hypotheses have been
advanced by other investigators. It is likely that NO was
removed as formed, by reaction with hydrocarbons from coal.
The correlation between the nitric oxide level and
the control variables is as follows:
Nitric oxide, ppm = -464.3
+ 0.4257 (bed temperature, of)
+ 0.329
in. )
- 2.337
(air rate, 1b/hr ft2) + 4.468 (bed height,
rate,
rate,
(carbon rate, 1b/hr ft2) - 0.7829 (inert
1b/hr ft2) + U.005372 (carbon rate x inert
1b2/hr2 ft 4) .
(23)
A comparison between the observed and calculated
values is shown in Figure 33.
6.6.3
Sulfur Dloxide Emissions from FBC without Presence
of Limestone Sorbents
Past studies (1) had shown that 90 to 95% of the
POPE, EVANS AND ROBBINS
INCORPORATED
-------
i-~-
:E:
g: 500
...
W
Q
H
><
o
~ 400
p:;
E-t
H
Z
~
~
~
...
W
Q
H 500
><
o
u
~
Eof
H
~
A. RUNS 307 AND 308
600
SEE APPENDIX E FOR
TEST DETAILS
300
1700
1800
1900
2000
2100
159
c
2200
90
100
BED TEMPERATURE, of
B.
TESTS 319-2, 3, 4, and 5
'. FIGURE 32 TYPICAL NITRIC OXIDE ElUSSION RESPONSES IN FBC
I . '
600
400
o
10
20
30 40 50 60 70
PERCENT OF FUEL AS COAL
80
POPE. EVANS AND ROBBINS
'I~i.. I}
-------
160 800
.
.
Q 700
.~
~
ra1
(I)
s:Q
0
-
~ 600
~
.
~ .
[. ~ 500 .
~ .
~ .
H .
~
Q
H .
)< .
0 400
u
H
~
~
H
~
300
300
400
500
600
700
800
NITRIC OXIDE IN FLUE GAS, PPM (CALCULATED)
'FIGURE 33
COMPARISON OF OBSERVED AND CALCULATED
NITRIC OXIDE EMISSION FROM FBC
PCA::, EVANS AN:) ROBBINS
I'::' :': c,C'~c':'2.C':-:.~,'r ,'.:.
-------
161
sulfur in coal would appear as S02 in the flue gas when the
coal was burned in a fluidized-bed combustor without lime-
stone addition. At the same tlme, a smaller fraction of the
carbon would appear as carbon dioxide; typically, only 80
to 85% of the carbon was consumed. Fluidized-bed combustion
of coal in a sorbent-free bed produces a fly ash which
contains 5 to 10% of the sulfur and 15 to 20% of the carbon
that was in the feed coal; hence, the requirement for a
Carbon-Burnup Cell.
In burning the high-carbon fly ash during the present
test program, the remaining sulfur was consumed. Table 17
below summarizes the sulfur balances for these tests and
demonstrates that the sulfur and carbon in fly ash are
consumed with about the same efficiency. Results for Runs
302 and 303 are presented, but are suspect and should be
disregarded.
A ratio of combustion efficiencies defined by
CR' % sulfur burned
E atlO = % carbon burned
(24)
was computed for each test in Table 16. Excluding the
results for Tests 302 and 303 the mean value was:
CE Ratio
= 1.0003
with standard deviation, s = 0.193
It is concluded that in the absence of a sulfur sorbent the
sulfur and carbon are burned with equivalent efficiency in
a Carbon-Burnup Cell.
6.6.4
Behavior of Limestone and Sulfur in a Carbon-Burnup
Cell
In Section 6.4.6 the effect of limestone (in the
POPE, EVANS AND ROBBINS
INCORPORATED
-------
162
TABLE 17
SULFUR BALANCE SUMMARY FOR FBC TESTS
Fly Ash Fly Ash Flue Gas Total Input
Sulfur Sulfur Sulfur Sulfur Less Sulfur Carbon
Test Input Output Output Output Output Burned Burned
No. 1b/hr 1b/hr 1b/hr 1b/hr 1b/hr % %
C-302-1 .605 .248 .286 .534 .071 47.2 90.0
-2 .860 .310 .640 .950 -.090 74.0 83.0
-3 .565 .240 .448 .686 .121 78.0 86.0
C-303-1 .410 .310 .110 .420 -.010 27.0 6~.0
-2 .435 .234 .210 .444 -.011 48.3 66.0
-3 ..615 .402 .180 .582 .033 29.4 65.0
-4 .578 .320 .266 .586 -.011 46.0 74.0
-5 .655 .262 .384 .646 -.009 58.7 78.0
C-306-1 .520 .091 .415 .506 .014 80.0 86.0
-2 .698 .104 .620 .724 -.026 89.0 70.0
-3 .635 .082 .530 .612 .023 83.0 89.0
-4 .535 .070 .450 .520 -.015 84.0 73.0
C-307-1 .510 .194 .306 .500 .010 60.0 76.0
-2 .292 .097 .210 .307 .015 72.0 90.0
-3 .518 .070 .420 .490 -.028 81.0 91.0
-4 .344 .092 .248 .330 .014 72.0 63.0
C-308-1 .470 .085 .398 .483 .013 85.0 81.5
-2 .550 .086 .450 .536 .014 81.0 81.3
-3 .560 .075 .470 .545 .015 84.0 80.6
C-310-1 .515 .041 .497 .538 .023 96.5 84.6
-2 .600 .060 .537 .597 .003 89.0 75.8
C-311-1 .600 .130 .448 .578 -.082 68.0 77.0
-2 .800 .280 .504 .784 -.016 63.0 66.7
-3 .950 .350 .525 .875 -.075 55.0 58.0
-4 .100 .500 .600 .600 .000 83.0 75.8
-5 .860 .173 NA NA
-6 1.11 .275 NA NA
C-312-1 .645 .135 .530 .665 -.020 82.0 84.0
-2 .625 .150 .505 .655 -.030 80.0 83.8
-3 .670 .175 .510 .685 -.015 76.0 76.0
NOTE: In Tests C-302 and C-303 the fly ash feed was wet and
not homogenous. The input data were probably not valid.
POPE, EVANS AND ROBBINS
INCORPORATED
-------
163
TABLE 17
(Continued)
Fly Ash Fly Ash Flue Gas Total Input
Sulfur Sulfur Sulfur Sulfur Less Sulfur Carbon
Test Input Output Output Output Output Burned Burned
No. 1b/hr 1b/hr 1b/hr 1b/hr 1b/hr % %
C-313-1 .87 .08 .76 .84 .03 87.3 91.0
-2 1.03 .10 .87 .97 .05 85.0 87.2
-3 1.17 .21 .98 1.19 -.02 83.6 86.0
C-314-1 .90 .22 .67 .89 .01 74.5 84.8
-2 1.01 .23 .76 .99 .02 75.0 82.0
-3 1.33 .29 .90 1.19 .14 68.0 76.0
-4 1.16 .14 .98 1.12 .04 84.0 89.5
C-315-1 .95 .25 .82 1.07 -.12 86.2 83.3
-2 1.01 .20 .89 1.09 -.08 88.0 84.5
-3 1.09 .21 .97 1.18 -.09 89.0 85.2
C-316-1 .95 .23 .80 1.03 -.08 84.5 74.5
-2 1.11 .23 .95 1.18 -.08 85.5 76.5
-3 1.29 .29 1.04 1.13 -.04 80.5 72.0
C-317-1 1.16 .20 .84 1.02 .14 72.0 82.2
-2 1.13 .28 .90 1.06 .07 79.0 81.5
C-318-1 1.04 .35 .78 1.13 -.09 75.0 72.5
-2 .97 .26 .75 1.01 -.04 77.5 79.8
POPE, EVANS AND ROBBINS
INCORPORATED
-------
164
fuel fed to a Carbon-Burnup Cell) on performance was outlined.
It was concluded, on the basis of data presented there that:
(a) Limestone, or more precisely calcium oxide plus
some sulfate, does not reduce the combustion efficiency of
a Carbon-Burnup Cell.
accurate
rich in
(b) The performance model, Equation 6, was less
in predicting performance when the fly ash feed is
limestone.
(c) Under sufficiently oxidizing conditions, calcium
sulfate will not decompose at the high temperatures
experienced in a Carbon-Burnup Cell. This latter conclusion
is discussed in more detail here and data presented on the
conditions which lead to decomposition of calcium sulfate.
As noted in previous sections limestone may be fed,
with the coal, to a fluidized-bed boiler in order to reduce
air pOllution due to sulfur oxides. Some, if not all, of the
limestone would be carried out of the furnace by the flue gas
and be removed by the dust collection equipment. The
collected dust would contain unreacted carbon, ash which
originated as mineral matter in the coal, and particles which
contain calcium oxide, sulfate and possibly carbonate. Since
no means exists for separating the carbon, the entire mass
would be fed to the Carbon-Burnup Cell.
Some 0 f the chemical reactions which could take
place within the Carbon-Burnup Cell include:
2C(s) + 02(g) ~ 2CO(g) (25)
2CO(g) + 02(g) t ~C02(g) (26)
S(s) -+-
+ 02(g) + S02(g) (L7)
CaO(s) + SOL (g) + 1/2 02(g) t CaSO (s) (28)
q
CO(g) -+- CaO(s) SOL (g) + CO~(g) (29)
+ CaS04(s)+ +
POPE. EVANS AND ROBBINS
INCORPORATED
-------
165
The performance model, Equation 6, is a tool with
which Equations 25 and 26 are linked to control variables.
The material in the preceding section shows that, on the
average, sulfur and carbon are burned equally well and so
Equation 27 is linked to the control variables. Little
data on Equation 28, the removal of S02 by lime, and Equation
29, the decomposition of calcium sulfate, were produced in
the current test series. Data collected in earlier work (~)
are of little value here since the conditions were so
different in that work.
Sufficient information was collected,
however, to draw reasonable qualitative conclusions and to
propose some interesting theories. These are discussed below.
6.6.4.1 Removal of S02 by Lime in a Carbon-Burnup Cell
In FBC Test C-3l8-2, the fly ash feed was precisely
1 wt. % sulfur. The fly ash contained no limestone. The
ratio of sulfur to carbon was '\, l:bO, by weight, while in the
original #~ coal, fed to the FBM, the ratio had been'\, 1:20.
About 80% of the sulfur and carbon was burned in this FBC test
( 0 ) .
T = 1950 F, RO = 4.5% and the S02 in the flue gas was 1050
ppm.
In FBM Test B-16, this same #~ coal was burned in
the FBM and fine limestone injected. The composition of the
fly ash which was then used in FBC Tests C-320-l and 2, was
given in Table 13. The ratio of sulfur to carbon in this
material is ~ 1:10, since much of the sulfur originally in
the coal was present in the feed to the FBC as CaS04.
In Test C-320-l, almost 90% of the carbon was burned
and it would be anticipated that 80 to 90% of the burnable
sulfur would also burn. However, in this test the S02 in the
FBC flue gas was only 350 ppm; or 1/3 the level in Test C-3l8-2.
POPE. EVANS AND ROBBINS
INCORPORATED
-------
166
The test conditions were comparable (T = 1980oF, RO = 4.3%)
to those in 318-2.
Two hypotheses are offered for the relatively low
802 level in Test 320-1:
1) Lime is capable of capturing 802 at temperatures
approaching 2,000op under the conditions of a Carbon-Burnup
Cell.
2) When limestone is added to a fluidized-bed burning
coal, the limestone acts to reduce the quantity of sulfur in
the fly ash which is subject to oxidation; i.e., limestone
converts fly ash sulfur into calcium sulfate.
The first hypothesis is supported by the findings of
many investigators that the kinetics of 802 capture by lime are
enhanced by increasing temperature in conventional furnaces and
with artificial flue gas. But, these results are in direct
contradiction to our earlier work (~) and that of almost every
other worker (23) ln fluidized-bed combustion.
One major difference between a Carbon-Burnup Cell and
a coal-burning fluidized-bed combustor is that, in the former,
little hydrogen is contained in the fuel; i.e., hydrogen or
methane from the volatiles in coal might reduce the efficiency
of sulfur capture in the PBM by:
4 Ca804 (s) + CH4(g)
+ 2 H20(g) + 4 802(g)
-:. 4 CaO(s) + C02 (g)
(30)
Reaction 30 is a global one and does not specify whether the
intermediate calcium - sulfur compound is sulfide or sulfite.
If sulfide, the intermediate could be oxidized by oxygen in
the gas or by adjacent molecules of sulfate in the solid. A
debate on mechanisms is not useful at this point.
POPE, EVANS AND ROBBINS
INCORPORATED
-------
167
The second hypothesis is simply a conjecture that
the history of sulfur in coal might be as follows:
Assume that most, or even all, of the sulfur in coal
leaves the coal particle as it is being pyrolized, and that
the resulting sulfurous gas may attack lime or residual carbon
(or oxygen and so form 502). In the absence of lime, some of
the sulfur recombines with the carbon to appear in the fly ash
as a residual organic sulfur. When this fly ash is subsequently
burned, this organic sulfur is oxidized and appears as 502 in
the flue gas.
When this fly ash
Ca504' remains in
flue gas.
If lime is present, in sufficient excess when the coal
is being pyrolized, sulfur is fixed by the lime rather than by
carbon and enters the fly ash stream as Ca504' not C 5 .
x Y
is subsequently burned, the sulfur, as
that form and does not appear as 502 in the
The obvious experiments and analyses which might have
proven which hypothesis was correct, if either really was,
were not done; the program was essentially over and the reason
didn't matter -- it had been definitely established that 502
emissions from a Carbon-Burnup Cell would be low if limestone
were fed to the primary cells, and the burnup cell were
designed and operated properly.
6.6.4.2
Release of 502 by Lime
Equations 28 and 30 explain, in a straightforward
manner how the sulfur in lime can be released as 502. Other
paths are as follows:
Ca504(s)
Ca503(s)
+ CO(g)
+
+
Ca503(s) + C02 (g)
CaO(s)+ 502(g)
(at To > l3000F)
(31)
+
+
+
+
(32)
POPE. EVANS AND ROBBINS
INCORPORATED
-------
168
or
4CaS03(s)
CaS(s) + 3CaS04(s)
-+
+
CaS(s) + 3CaS04(s)
4CaO(s) + 4S02(g)
(33)
(34)
-+
+
A re-examination of previous S02 control results
(~) was made in an effort to explain the release of S02
by lime, and these results were compared with predictions
based on PUblished thermodynamic data. These published data
are highly suspect, since data on calcium compound thermo-
dynamics have never been rigorously reviewed and do not
appear in the JANAF Tables.
The difference between predicted and observed results
led to the postulation of a pseudosulfate as an intermediate
between S02 capture and the CaS04 product. The sulfur bound
as the pseudosu1fate was assumed less tightly held than in
the true sulfate, and would decompose more readily than the
true final sulfate. Appendix H contains the theoretical
analysis and the results of an experiment intended to determine
if psuedosu1fate played a major role in the release of sulfur
from partially sulfated lime.
The experiment, C-320-2 and C-320-3, was carried
out by firing, in the FBC, fly ash from the FBM containing
the lime by-products and then switching to an artificial
mixture of coal fines, ash, limestone, and natural gypsum.
The composition of the natural fly ash is described in Table
13. The test revealed that both feed materials produce
increasing S02 as residual oxygen is reduced (See Figure H-2),
and that the pseudosu1fate theory was not necessary to
explain the observed results. It was noted that it is
possible to retain the sulfur in the artificial mix and
in the FBM fly ash by operating the FBC at residual oxygen
levels above about 3%.
POPE, EVANS AND ROBBINS
INCORPORATED
-------
169
Table 18 shows the level of S02 in the FBC off-gas
as a function of residual oxygen with a fly ash/lime by-
product feed.
TABLE 18
S02 APPEARING IN EXHAUST OF FBC FIRING
FLY ASH CONTAINING 19.4% Ca, 2.4% S
Bed temperature = 1970-l980oF, Air rate = 620 lb/hr
Fuel Rate = 167-174 lb/hr, Superficial velocity = 12.1 fps
Residual 02' vol. %
S02' ppm
0.1
0.4
0.5
0.8
1.0
4.3
4100
2400
2150
750
550
350
It may be seen from these results that fine lime, in
fly ash, cannot be effectively regenerated under the conditions
existing in a Carbon-Burnup Cell.
1) The residence time of the fine particles is
apparently too short to permit more than partial desulfuriza-
tion, or
2) S02 may be re-absorbed in the flue as the gas
and lime particles travel toward the dust collector together.
Whatever the reason, only about one-half the calcium
sulfate was decomposing at 0.1% 02.
6.6.5
Particulate Emissions
The fuel fed to a Carbon-Burnup Cell will be
POPE, EVANS AND ROBBINS
INCORPORATED
-------
170
relatively high in ash. Without the addition of limestone,
the fuel would typically be 50% ash or about 35 pounds of
ash per 106 Btu fired. With limestone, ash (or inert) input
would be on the order of 240 Ibs per 106 Btu fired. These
values may be compared with a typical coal (9% ash, 13,500
Btu/lb) where the ash input is about 7 Ibs per 106 Btu fired.
In tests in the FBC the particulate emissions, PE --
that is the fly ash passing through the bed as well as the
FBC's cyclone collector -- were
1 < PE' < 8 Ibs Per million Btu of fuel fired.
'\, '\,
The emission data, in this form, were plotted against
reciprocal bed temperature in Figure 34. These data were too
scattered to produce a useful correlation, and none should be
expected considering the number of control variables. (These
are all considered later). It is clear however that emissions
decline with increasing bed temperature.
The reduction in emissions with increasing bed
temperature is expected because:
a) At high temperature more of the carbon burns and
so less is available as particulate. This has been
demonstrated previously.
b) At high temperature, some of the ash matter is
retained by the bed particles via agglomeration. Also,
agglomeration makes those particles which leave the bed
larger and hence more collectable by the cyclone. These
factors are discussed below.
of sand,
retained
Goldberger (17, 23), burning coal in
has demonstrated that some of the ash
by the bed particles according to:
a fluidized-bed
would be
POPE. EVANS AND ROBBINS
INCORPORATED
-------
It'I
."'IP
.
0
-
.-4.
I "'IP
P: "'IP
.
.
... 0
~
.......
0
0 M
O"'IP
.-4 .
- 0
...
~
::> 0
8
.~ N
'<3'
f1I
p,. 0
~
f1I
8
Q
f1I ..-i
aI ~
CD - 0
:;1 0
U 00 0
o
p:; 0
p,.
H 0
U '<3' 0
~ 0
o
00
o
o
o
0\
M
o
.
o
o
1
3 4 5 6
PARTICULATE EMISSION, LBS PER 106 BTU
2
171
00
7
rIGURE 34
PARTICULATE EMISSION FROM FBC AS A FUNCTION OF BED
TEMPERATURE
PCFE, EVANS AND ROB8NS
o
o
8
-------
172
C
I In
n-
C
out
w
= Kc G
(35)
where
c. , C
In out
= fly ash concentration entering the
system and leaving the fluidized-bed,
grains/scf
k
c
=
capture coefficient, grains captured/
min x lb-bed x grains per scf
w
= bed weight per unit cross section,
Ib-bed/ft2
G
= gas flow rate. per unit cross section,
sCfm/ft2
Goldberger showed that
~. ~5l,300 (~,)
In k
c
+ 20.47
(36)
where
T'
o
= bed temperature, F
From Equations 35 and 36 it
quantity of dust leaving the bed and
collector will decline as the:
follows, that the
entering the dust
1)
2)
3)
4)
quantity of entering mineral matter is reduced
bed temperature increases
bed depth (or particle density) increases
air rate is reduced
Also, the fraction of dust entering a cyclone dust
collector which passes through without capture will decline
as the:
1)
2)
quantity of entering dust is increased
bed temperature is increased (because this acts
the gas temperature, hence velocity entering the
to increase
POPE, EVANS AND ROBBINS
INCORPORATED
-------
173
collector)
3)
air rate is increased
Goldberger finally suggests that
total ash collected by
Es' % = total ash in feed
would be given by
system x 100
(37 )
E , % = 100 -(exp
s
-k (WIG) x (lOO-E )
c c
(38 )
where
E
c
= dust collection efficiency of cyclone,
% (considered independent of T' and G)
Attempts to reproduce the experimental data derived
from the tests in the FBC with Equations 35 and 36 were not
successful. This may be because Golderger's studies were
carried out in the range 1 ~ WIG ~ 3.5, ,,,hile most of PER's
work was carried out with W/G;f, .1.
A second reason may be that the ash matter in fly ash
has a lower propensity to agglomerate than the ash in coal.
However, there was some indication that at the
highest temperatures the ash was captured. For example,
during tests C-302-2 and C-306-3, where the bed temperatures
exceeded 2,100oF, the bed mass was increasing so rapidly that
material had to be removed to maintain a constant reading on
the low bed manometer which indicates bed' depth. The amount
of material removed was not noted.
The agglomeration effect at high bed temperature may
also be noted in Figure 35. This figure presents the results
of a sieve analysis of bed samples taken before and after run
C-308, in which the bed spent a short time (about 15 minutes)
at 2,1000F and about two hours at 1,9000F. The original top
POPE, EVANS AND ROBBINS
INCORPORATED
-------
174
2400
2200
2000
1800
1600
tI)
z
~
u 1400
H FINAL BED'
~
..
~ 1200
~
8
~ 1000
H
Q
~
H
U 800
H.
8
~
~
p. 600 0
400
200
o
.0
SEE TEXT FOR
DESCRIPTION OF
TEST CONDITIONS
20
40
60
80
100
PERCENT SMALLER THAN STATED SIZE
. FIGURE 35
ASH BED PARTICLE SIZE DISTRIBUTION
BEFORE AND AFTER HIGH TEMPERATURE
OPERATION IN FBC TEST C-308
POPE. EVANS AND ROBBINS
= ;'"~\~':<~~i /,\';:-[.2 ::;1
8
10
12
.
o
z
14 00
. .
::I
..
16 ~
H
Z
~
18 g;
20 ~
~
tI)
25
30
35
40
-------
175
size before the run was 14 mesh.
It is seen that 30% of the
bed particles after the run were larger than the original
top size. Excluding a sampling error, this result can only
be explained on the basis of agglomeration.
a)
The elutriation of bed fines could not have
increased the top size
b)
The fly ash fed had a top size less than 1000~m
Also, when the bed particles were examined they had
a coating of ash matter.
It is clear that the ability to agglomerate
ln a Carbon-Burnup Cell may be a marked advantage of
Multicell Fluidized-Bed Boiler.
fly ash
the
In certain tests where limestone was present (to be
discussed in a future report), fly ash formed uniform spheres,
up to about 1/2" in diameter. At the least, ash in this form
could be disposed of at no cost to the boiler operator since
it would be a useful fill material.
The particulate emission data listed in Appendix E
was fit to the performance model (See pages F-ll through F-13).
Equation 3, based again on the assumption that cross
sectional area is the appropriate scaling factor, was the
result.
PE = Pounds of particulate emission per hour per
ft2 of bed cross section =
+ 15.57
+ 0.0236(H)
+ 0.00034 (c' x
0.00664(T) - 0.00134(A')
0.03034(C') + U.00461 (I')
I' )
(3A)
The signs of the regression coefficients and their
relative importance appear logical, except for that modifying
POPE. EVANS AND ROBBINS
INCORPORATED
-------
176
H, the static bed depth.
would be negative. This
magnitude of the entire
It would be expected that coefficient
is not a serious flaw since the
term is low.
The derivation of this relationship is discussed in
Appendix F. The calculated and observed data are compared
in Table F-4.
6.6.6
Hydrocarbon and Carbon Monoxide Emissions
Hydrocarbons were measured continuously by a flame
ionization analyzer, while carbon monoxide was measured
intermittently by Orsat analysis. During the entire FBC
test series, the highest hydrocarbon value was 200 ppm, at
a time when coal fines were used with gypsum to test the
pseudosulfate hypothesis at a residual oxygen level of 0.8%.
The next highest value was 50 ppm, and most readings (reported
in Appendix E) were 30 ppm or less. Orsat CO readings were
zero. However, one scale division on the apparatus is 0.2%.
Therefore, under carbon burnup cell conditions (high
temperature and high residual oxygen, without coal feed),
little if any hydrocarbon or CO emission is anticipated.
POPE, EVANS AND ROBBINS
INCORPORATED
-------
177
7.0
DEVELOPMENT OF AN INTEGRATED FLUIDIZED-BED BOILER
AND CARBON-BURNUP CELL: TESTS WITH THE FBM/CBC
7.1
Genera 1
As described in Section 5.2.2, a simulated Carbon-
Burnup Cell was added to the modular boiler cell (FBM). As
was noted, this unit, designated the CBC, was added to
evaluate the Carbon-Burnup Cell concept in an integrated
system and to gain insights on the operation of a Mu1tice11
Fluidized-Bed Boiler.
The CBC's air distributor had the dimensions 10-1/2"
x 15-1/2". Compared with the FBC, the CBC was 28% larger in
plan area. Unfortunately, the effective height for disentrain-
ment of particles splashed above the bed of the CBC was
considerably less than that of the FBC. This was caused by
the existing FBM steam drum arrangement.
The FBM and CBC shared a common fluidized bed, in
that an opening existed in the common wall shared by the two
units. Bed particles would therefore circulate between the
two reactors at a rate established by a diffusion-like driving
force. The interchange was sufficiently rapid to permit the
CBC to be ignited by the hot bed material which flowed in from
the operating FBM, without the use of a separate gas-fired
ignition system. This feature, and several others, characterize
the Mu1tice11 Fluidized-Bed Boiler.
If the exchange of bed material were too rapid, the
desired temperature differential (300°F to 600°F) required
for efficient operation could not be maintained. One purpose
of the test effort with the FBM was to determine what the
interchange rate was and how it was controlled.
POPE, EVANS AND ROBBINS
INCORPORATED
-------
178
In adding the CBC to the FBM, it was necessary that
the coal and additive injection systems be changed from those
described in Reference 1. A technique of vertical coal
injection was seen as the preferred mode for commercial
application; so, coal formerly injected via the front of
the unit at operating floor level was now to be injected via
the top of the unit. This also cleared the operating level
for the equipment required for the CBC.
The coal injection method and the efficiency of
sulfur capture by limestone are related through the oxygen
requirement of reaction 28 and the deleterious effect of
reducing gases implied by reactions 30, 31, and 32. If the
quality of the coal distribution is poor, regions will exist
in the combustion bed in which there is insufficient oxygen
to carry reaction 28 to the right; i.e., to CaS04' and 802
will escape from the bed into the flue gas. If grossly
reducing conditions were to exist, CaS04 might tend to
decompose and 802 escape the bed into the flue gas. Therefore,
another purpose of the test program was to determine the
effectiveness of coal distribution in a large fluidized-bed
combustor.
Because of the limited space above the bed in the CBC,
particles of bed material which were thrown into the freeboard
were swept out with the exhaust gas and appeared in the unit's
dust collector. Although this effect was very undesirable,
compactness was seen as a desirable feature of a fluidized-bed
boiler. Therefore, another purpose of the test effort was to
determine whether bed material loss could be inhibited by a
baffle screen, and how effective heat exchange would be in the
freeboard of a fluidized-bed combustor.
Fine limestone injection had been studied in the
POPE, EVANS AND ROBBINS
INCORPORATED
-------
179
program reported. in Reference 1. The addition of the CBC to
the FBM and the changes in the coal and additive injection
systems required that the efficiency of fine lime injection
be re-eva1uated. There was some concern that the high
temperatures of the CBC would cause the calcium sulfate formed
in the FBM to decompose in the CBC. This effect was potentially
advantageous--if the fine lime could be reused in the FBM or,
for example, if Portland Cement and sulfuric acid were to be
produced from coal ash and partially sulfated lime. More
likely, decomposition of fine limestone by-products at the
conditions of a Carbon-Burnup Cell could not be exploited.
Therefore, another purpose of the FBM test program was to
assess fine limestone effectiveness in an integrated system
and to determine the fate of sulfur carried into the CBC with
the fly ash.
7.2
Coal Feedinq to the FBM
7.2.1
General
As noted earlier, coa1.had previously been fed into
the FBM via a horizontal pneumatic injector which rested on
the air distribution grid. This method is described in detail
in Reference 1. Other coal injection methods tested are
described in Reference 4. As part of the current program,
it was necessary to revise the coal injection technique to
accommodate the addition of the CBC to the FBM. This provided
the opportunity to test a vertical coal injection technique,
which was felt to have a number of potential advantages. The
revised coal handling system was shown schematically in Figure
10. The portion of the injection system internal to the
boiler was shown in Figures 8 and 9. It was this internal
portion that most directly influenced performance.
POPE. EVANS AND ROBBINS
INCORPORATED
-------
180
Briefly, the external system was as follows: Coal
is carried from the weigh hopper to the pneumatic injector
via an inclined screw. This inclined screw provides an
effect ive pressure seal between the 35" w.g. pressure at the
injector inlet and the atmospheric pressure above the open
weigh hopper. The coal dropping off the end of the screw
passes through a flexible connection (required by the weighing
arrangement) and enters the vertical pneumatic injector.
The vertical injector consists of a section of 1-1/4"
square tubing of 304 stainless steel which penetrates the
upper boiler wall just below the steam drum. The feeder
divides into two parallel tubes, also 1-1/4" square, which
go down through the fluid bed and which terminate with a
horizontal turn, one feeder pointing forward and the other
toward the back of the FBM. The feeder assembly is centered
in the boiler and the coal is discharged about I" above the
grid plate. Each outlet tube services about 4 square feet
of grid area. As described in Reference 4, a single point
injector servicing 8 square feet would be less effective.
7.2.2
Results
For a fluidized-bed boiler, the quality of fuel
injection is determined by the presence or absence of
temperature gradients within the heavily cooled fluidized
bed, and by mapping residual oxygen distribution above the
bed. Symptoms of poor fuel distribution would be: large
horizontal temperature gradients, large horizontal oxygen, CO
and hydrocarbon gradients, poor sulfur capture by limestone,
and under some operating conditions, the agglomeration of large
quant ities of bed material. In early work described in Reference
4, horizontal gradients of 2500F over a 6 foot span were not
unusual.
POPE, EVANS AND ROBBINS
INCORPORATED
-------
181
Poor fuel distribution within the bed may result from
inherent flaws or design errors. As an inherent flaw we
might include, for example, feeding gas, oil, solid waste, or
pulverized coal into the bed of a fluidized-bed boiler
through point injectors. As design errors, we might include:
arranging the heat-exchange surfaces so that they interfere
with the movement of coal through the bed, or providing an
inadequate number of discrete coal injection points; i.e.,
over-estimating the mixing ability of the fluid bed or under-
estimating the reactivity of the fuel. Several other potential
design errors are discussed below.
The initial shakedown test of the modified FBM system
gave evidence of extremely poor coal distribution. In this
test, Feeder "a" in Figure 36 was used. Over the 6-foot span
of the unit, residual oxygen, measured 86" above grid level,*
was negligible directly above the coal injector and over l~~ at
the two ends of the unit. (l~~ is the limit of the recorder).
Investigation revealed that the injector had partially plugged
with coke, a symptom of the coal overheating in its passage
through the feeder, possibly from inadequate injection velocity,
or from excessive residence time. For transport air at ambient
temperature, the temperature rise of the coal traveling through
the feeder had been calculated to be less than 300oF. This
calculation indicated that the coal would not be overheated--
that no devolatilization or coking would take place. The
orig inal 1" x 2" in jector was replaced with the 1-1/4" in jector
to increase velocity, and the tube above the bed wrapped with
insulation to reduce heat gain from the bed and freeboard.
Th i s mod i fica t ion is shown as "b" in Figure 36. The symptoms
of poor coal distribution; i.e., horizontal oxygen gradients,
continued during subsequent shakedown tests. A water jacket was
*In more recent tests a probe 4 ft. above grid was used.
POPE, EVANS AND ROBBINS
INCORPORATED
-------
l-s-
.COAL
!- X 2-
1-1/4-X !-1/4-.
s. ---J
G. FIRST DESIGN
WATER
COOLED JACKET
---WATER IN
~' 5" =1:
58
d. FOURTH DESIGN
!-1/4- X 1-1/4-
INSULATION
1-1/4. X 1-1/4.
L5":I: 5" ~
b. SECOND DESIGN
WATER IN
KICKER
e. FIFTH DESIGN
182
WATER
COOLED JACKET
-WATER IN
.L5"~ 5"~
c. THIRD DESIGN
WATER IN
~ TOWARD REAR OF FBM
TOWARD FRONT OF FBM':-
-~ >
f. SIXTH a FINAL DESIGN
FIGURE 36
FBM COAL INJECTOR MODIFICATIONS
POPE, EVANS AND ROBBINS
1:,. CORPORa TIO
-------
183
added to the coal injector as shown in "c" of Figure 36; and
still no improvement was noted in the oxygen map. It was then
apparent that the calculation of coal temperature rise had
probably been correct and that the poor coal distribution
was the result of some other, as yet unknown, design error.
It was suggested that the poor performance of this
new vertical injector could be explained if the larger coal
particles were being disentrained in the 900 bend between the
vertical and horizontal sections of the injector. Once the
large coal particles had been separated from the injection air
stream, they would not enter the fluidized bed with a velocity
near that of the injection air. Instead, it was visualized that
the injection air, and fine coal left the feeder at the top of
the tube at high velocity, while the coarse fraction would
"slide" out of the bottom of the tube without the momentum
required to deeply penetrate the fluidized bed.
To cause the
coarse coal to be re-entrained, a "kicker" plate was tack-
welded into the horizontal s,ection, as shown in "d" of Figure
36.
The addition of this kicker was not a complete
solbtion, but the extreme oxygen gradient previously encountered
was reduced to a 5% 02 differential between the center of the
boiler and either end. A typical set of 02 readings here would
be a 1% 02' at center, 6% at either end. The average gas
composition, measured downstream, would then range from 3.5
to 4.5%.
Extens ion of the horizontal legs to 14", as in "e"
of Figure 36, gave a favorable oxygen map. The typical maps
with this feeder configuration is shown in Figure 37 for FBM
Test B-3. ~ 02 is about 1 percentage point. Figure 37 shows
that the fuel/air ratio was slightly higher in the rear of the
POPE, EVANS AND ROBBINS
INCORPORATED
-------
...
at
PROBE NO. AND LOCATION .
1 2 4
4
~
~ ~ AVERAGE 02 MEASURED
0
~ DOWNSTREAM OF PROBES
~ V) ~= HIGH 02 TEST
ct ~
C) .. = LOW 02 TEST
TAKEN 86" ABOVE ~
~ ~. LLJ
:> AIR DISTRIBUTION
..J GRID
~ (J) t1.2 COAL FEEDER.
~~ z
a~ z I C7
LLJ I
C)
>- ..
~I
..J AVERAGE 02 MEASURED
ct .. DOWNSTREAM OF PROBES
::>
c
en
LLJ ..
a:::
0 9 18 27 36 45 54 63 72
f t
FRONT DISTANCE FROM FRONT WALL OF 801 LER, INCHES REAR.
WALL WALL
FIGURE 37 . TYPICAL FBM RESIDUAL OXYGEN MAPS- FBM TEST B-3
-------
185
unit.
Wi th the feeder further modified as shown in "f II of
Figure 36, the oxygen gradients were reduced to as low as
20.2% 02 over the 6-foot span.
7.3
Feedinq Fly Ash to the CBC
7.3.1
General
The use of fine solids « 20 mesh) in a fluidized bed,
is not unusual. Most commercial fluid-bed processes, in fact,
make use of fine solids, and in order to avoid excessive loss
of these particles by elutriation, limit superficial gas
velocities to low levels. For particles the size of fly ash,
about 100 microns, the terminal velocity, the highest velocity
at which particles would remain fluidized and not elutriate,
is on the order of one foot per second.
Operation of a Carbon-Burnup Cell at air rates that
would result in a superficial velocity of one foot per second
was undesirable for two reasons. First, it was necessary to
build the system compactly and this requires that relatively
high air rates be used. Second, for a Carbon-Burnup Cell which
is an integral part of a fluidized-bed boiler, the minimum
permissible superficial velocity, set by the size and density
of the bed particles in the coal-burning regions, will
normally be far in excess of one foot per second. But high
air rates are incompatible with long residence times for fine
particles in the fluidized bed of the Carbon-Burnup Cell.
If the fine particle residence time is short, the time
available for reaction within the bed will also be short. For
a bed one foot deep, operating at an air rate and temperature
which yield a superficial gas velocity of 10 feet per second,
the average residence time of the gas and of the fine particles
POPE. EVANS AND ROBBINS
INCORPORATED
-------
'86
may be on the order of 0.1 second. This would be true if the
fluidized bed had no capacity to retain and distribute fine
particles. If the fuel were not evenly distributed and
retained by the bed so that it could react with the air
passing evenly through the grid plate, the fuel would burn
in the freeboard with the undesirable results shown in Figure
29.
It was shown in Section 6.5.3 that at a low bed
temperature (T < 1,8500F), large quantities of relatively hot
carbon may appear in the freeboard. If this carbon then reacts,
a temperature rise will occur in the gas space of an adiabatic
combustor. In a sense, the fluidized bed itself is cooled in
heating the fuel, and a large temperature difference will
exist between the fluidized bed and the space above the bed.
Despite an extensive search, no usable correlation
was found which would predict the fine particle residence
time in a high velocity fluidized bed. The work required to
produce such a correlation was beyond the scope of this study.
Even with such a correlation, performance predictions would
still be difficult, since the fly ash particles are continually
becoming smaller as they are consumed. An additional and more
practical problem is that the fine solids may be pneumatically
injected at a single point, or series of discrete points.
Therefore, an average superficial velocity may not adequately
describe the actual gas velocity in the vicinity of the
injector, and it is probably this local velocity which most
influences the fine particle residence time. There is an
obvious need for basic research on the behavior of fine
particles in a fluidized bed composed of relatively large
particles.
It is apparent that an effective Carbon-Burnup Cell
POPE, EVANS AND ROBBINS
INCORPORATED
-------
187
design requires that the bulk of the fly ash burn within the
bed and not in the freeboard. Burning within the burn up cell
bed is favored by high temperatures and even fuel distribution
across the cross-section of the bed. These two factors are,
of course, related. If the fuel is poorly distributed, its
combustion will not be complete in the bed and the bed must
cool. On the other hand, if the burnup cell bed is heavily
cooled--by heat exchange surface, by excess air, or by
circulation of bed material from the cooler primary cell
region -- combustion within the bed will also be incomplete
and the bed will cool further. Carried to the extreme, poor
fuel distribution and/or excessive cooling will result in the
fire being extinguished.
7.3.2
Fly Ash Feeder Desiqns Tested
During the first shakedown tests with the FBM, the
combustion efficiency of the CBC was extremely poor; i.e.,
CE '" 25%. During these tests, the CBC was excessively cooled
by circulation of 1,500oF bed material from the FBM. An
additional cause of the poor performance may have been the
use of a horizontal stub fly ash injector similar to that used
in the FBC (See Figure 4). We found that the temperatures near
the base of the bed (3" above the grid) were lower than those
9" above the grid, (near the center of the expanded bed), and
that the temperatures in the gas space directly above the bed
(28" above the grid) were higher still.
A change was made to the ''mushroom'' feeder described
in Section 5.2.2 and shown in Figure 11, and the undesirable
gradient was effectively removed. Table 19 lists the
temperatures in a test with the horizontal stub feeder and a
test with the mushroom feeder.
POPE, EVANS AND ROBBINS
INCORPORATED
-------
188
TABLE 19
EFFECT OF FLY ASH FEEDER DESIGN ON
TEMPERATURE DISTRIBUTION IN A CARBON-BURNUP CELL
Feede r Type
Thermocouple locations
(Distance above air
distribution qrid)
. 0
Temperatures, F
Horizontal Stub Mushroom
3" (Base of bed)
9" (Center of expanded bed)
28" (10" above expanded bed)
1,770
1,810
1,950
1,910
1,930
1,900
The CBC has a total cross-sectional area of 1.1 ft2.
Table 19 indicates that the single mushroom feeder effectively
services this area. As noted earlier, it is not known whether
a greater area could be served via a single mushroom. It
seems reasonable to attempt to place similar feeders on about
a two-foot triangular pitch to feed fly ash to a large Carbon-
Burnup Cell.
7.4
Interchanqe of Bed Material
7.4.1
General
Under "Apparatus", Section 5, the constructional
details of the FBM and CBC were described. These two
sections of the integrated FBM system shared a common
fluidized bed. The common perforated wall permitted bed
particles to interchange between the FBM and CBC regions.
The FBM region contained the only direct-contact
cooling surface, the water walls.
The FBM had the sole
POPE. EVANS AND ROBBINS
INCORPORATED
-------
189
igniter and also the only means for adding or removing bed*
material from the system.
The interchange of bed material through a barrier
has not received attention from investigators in fluid-bed
technology, although it might be a useful tool in the design
of catalytic reactors. Perforated barriers may be an
important design feature of fluidized-bed boilers for the
following reasons:
1) The barriers will permit a large boiler to be
started with one or just a few gas/oil fired burners. A
special starting zone may be provided without numerous
cooling tubes; once fired, hot bed material and burning
coal would flow through the barrier and ignite adjacent
regions.
2) When the steam required from the boiler is
about 50% of the full load capacity, the barriers make it
possible to turn-off sections of the boiler without causing
the bed depth in the active sections to decline.
3 )
The barriers will permit the use of a common
bed for both primary and burnup cells, even though these
two sections may operate with a 5000F difference in bed
temperature.
In the FBM/CBC system, some of the proposed functions
of the semi-permeable barrier were explored, within the limits
of the apparatus.
It was found possible to ignite the CBC by allowing
hot fluidized bed material from the FBM to interchange with
*
A CBC bed sampling outlet pipe and valve were added later.
POPE, EVANS AND ROBBINS
INCORPORATED
-------
190
the cold fluidized bed of the CBC.
o
about 800 F, coal could
operating temperature.
be allowed to continue
When the CBC bed reached
be fed and the CBC brought to
Alternatively, the interchange
until the CBC had reached about
could
13000F
at which point fly ash could be used as a fuel to further
raise the CBC temperature. This procedure was as follows.
First the FBM was ignited using a propane-fired burner while
the CBC was still static. Then the CBC was fluidized. It
was also found possible to shut down the CBC, continue
operation of the FBM, and then restart the banked CBC.
All of the features, and potential problems, of
semi-permeable barriers should some day be rigorously
explored. However, the one of major importance in the study
of a Carbon-Burnup Cell is the rate of bed material inter-
change between the FBM and CBC. It is important to determine
how this rate would affect the required temperature
differential between the coal-burning and fly ash-burning
regions of a 'fluidized-bed boiler. Using limestone as the
sulfur control agent, it is 1ilce1y that the primary cells
will operate at 1500oF; the burnup cells might well operate
at 2050oF.
7.4.2
Measurement of Rate of Interchanqe
No direct method exists for measuring the exchange
rate.* Therefore an indirect method was required. One such
indirect method would have involved adding a tracer material
to the FBM bed, in a pulse, and then, by sampling the CBC
bed, determining how rapidly the tracer concentration in the
*
Radioactive tracer techniques might be useful.
POPE. EVANS AND ROBBINS
INCORPORATED
-------
191
CBC reached the level in the FBM. Unfortunately the tracer
experiment would have been costly to perform and was beyond
the scope of this program. In observing the startup of the
system, an inexpensive tracer-like experiment suggested
itself. By igniting the FBM and bringing it to a steady-
state temperature while the CBC was inactive, a source of
hot bed material would be available on only one side of the
barrier. Then, by fluidizing the CBC and determining how
rapidly its bed temperature increased, a measure of the
interchange rate could be obtained.
The derivation of equations describing the transient
heating and the details of the experiments are given in
Appendix J. In two experiments, in which the area of the
opening between regions was different, values of the inter-
change rate were obtained.
A value of = 12,000 pounds per hour of bed material
interchange per square foot of open area
2
of 10,000 pounds per hour per ft
the time to raise the temperature
of an adjacent hot bed.
was found.
A va 1 ue
should be used in estimating
of a cool bed to near that
7.4.3
Effect of Interchanqe of Bed Material on Carbon-
Burnup Cell
As noted earlier, Section 5.2.2, the initial open
area in the wall separating the FBM from the CBC was 90 square
inches. With this open area, it was not possible to achieve
a CBC bed temperature more than about 2000F above that in the
FBM. By reducing this open area, finally to 2 square inches,
it was possible to achieve as large a differential as desired.
For example, in FBM test B-IO a differential of 4500F was
ach ieved .
POPE. EVANS AND ROBBINS
INCORPORATED
-------
192
As the tests in the FBM proceeded, the carbon steel
plate, used to reduce the open area, deteriorated and by Test
B-13 a 300°F differential was the maximum possible. The plate
was changed to ASTM 446 and the differential restored; for
o
example, in Test B-13 the FBM operated at 1590 F and the CBC
at 20000F.
Since an open area of 2 in2 between the FBM and
CBC was found acceptable, it is assumed that for a larger
Carbon-Burnup Cell, rapid ignition and equal bed depths may
be achieved with a suitably larger opening.
The most straight forward scaling factor is to keep
the ratio open area: plan area constant. For the CBC this
was
2 . 2
In
1.1 ft2
=
1.82 in2/ft2
Thus, for the Carbon-Burnup Cell, described in
2
Section 6.5.4, with a plan area of 35.4 ft , an open area in
the tube wall between primary cell and burnup cell would be
64.5 in2 (0.45 ft2). This open area would then result in
the interchange of 5400 1b/hr of bed material between the
primary and burn up cells. For a primary cell temperature
of 15000F and burn up cell bed .temperature of 20500F this
interchange would remove 650 K-Btu/hr from the burnup cell
which is negligible when compared to the total energy loss
of 32,800 K-Btu/hr.
7.5
Prevention of Carryover of Jetted Particles from
a Hiqh Velocity Fluidized Bed
The addition of the CBC to the existing FBM
required compromises in the location and design. The
decision was made to locate the CBC at the rear of the
unit
POPE, EVANS AND ROBBINS
INCORPORATED
-------
193
so that access to the larger FBM would not be more difficult
and to avoid cutting pressure parts. This location limited
the effective height of the CBC to 48" (the height above the
air distributor at which the exhaust duct was placed, See
Figure 9).
Although circumstances forced a small freeboard
in this instance, it is desirable that a Multicell Fluidized-
Bed Boiler have an open freeboard that is far smaller than
the open freeboards most often used in fluid-bed roasters,
etc. The test work in the CBC therefore afforded an
opportunity to explore the potential for small freeboard.
While the low height of the CBC did not interfere
with the combustion process within the bed itself, it did
cause the loss of bed material. The bed material would be
carried over and would appear in the dust collector discharge.
Since the CBC bed was in communication with the FBM bed, the
bed depth in the smaller CBC would remain relatively constant
despite the depletion of bed material. The cause of this
carryover has been discussed in detail in Appendix L.
Prevention of carryover in fluidized beds is normally
accomplished by insuring that the exhaust is above the point
in the vessel where a relatively high density of jetted
particles exists. In the case of the CBC, this was not
possible without a major change to the FBM's steam drum.
(This change was subsequently made and will be discussed in
a future report).
An alternative to a large, open freeboard is to
install some form of baffle at the top of the vessel. If
the particles strike this baffle and thus lose their vertical
momentum, those particles large enough to return to the bed
against the average gas velocity will return. The use of a
POPE. EVANS AND ROBBINS
INCORPORATED
-------
194
louvered baffle in this application was described by Lashe
and Commander (12) in reporting on the fluidized-bed calciner.
for radioactive wastes built at the National Reactor Testing
Station. The installation of such a louvered baffle in the
CBC was being contemplated when a visit to BCURA Industrial
Laboratories, under OAP sponsorship, revealed BCURA's
application of an array of horizontal heat-exchange tubes
above the bed of a fluidized-bed combustor which accomplished
the same knockout function. Such an array was installed in
the CBC.
By providing a surface on which large particles
may lose their kinetic energy, the baffle screen allows
these particles to fall back into the bed. The baffle screen
does not prevent elutriation of fine material, that is
particles whose density and diameter give them a terminal
velocity below the actual gas velocity. However, if the
baffle screen is made up of heat exchange surfaces, the gas
temperature will drop and the gas velocity, which is
proportional to gas temperature and inversely proportional
to free area, will also drop, permitting some of the marginal
particles to return to the bed.
The first knockout array, studied in the current
program, consisted of seventeen one-inch O.D. water-cooled
tubes arranged in four rows on a triangular pitch. The
arrangement was "opt ically dense"; i.e., such that, when
viewed from above the baffle screen, the bed could not be
seen. Therefore there was no line-of-sight path for a
jetted particle to travel into the exhaust duct without
striking a tube. The tube arrangement is shown in Figure 9.
After the installation of this baffle screen, large
bed particles (-8+22 mesh) no longer appeared in the cyclone
POPE, EVANS AND ROBBINS
INCORPORATED
-------
195
discharge. The size distribution was then similar to that
in Figure 26.
The cooled screen was such an effective heat
exchanger that the CBC bed had excessive cooling and the
combustion efficiency suffered. The heat exchanger
performance of the cooled screen is discussed below.
A second baffle screen was also tested.
This
baffle was identical in arrangement to the first, but was
made up of uncooled ASTM 446 rods. This screen also
prescribed the carryover of ash bed partiCles. The uncooled
screen prevented chilling the bed.
The designs tested and found serviceable consisted
of four rows of staggered tubes. Additional rows, more widely
spaced, may be used to avoid erosion by the heavy dust loading
in the Carbon-Burnup Cell exhaust. A maximum velocity of 30
feet per second is recommended for this application based on
conservative, conventional boiler practice. For a high-
velocity Carbon-Burnup Cell, operating at a superficial gas
velocity of 15 fps, a free area reduction of 5~~ due to the
baffle screen would yield the desired result.
We recommend that an open freeboard space of no less
than 36 inches be provided between the top of the expanded
bed and the entrance to the convection bank or cooled baffle
screen. In the CBC, the comparable distance was only 10
inches. The 36" distance, provided in the FBM, is considered
the minimum for completion of combustion occurring in the
freeboard.
We recommend that any fluidized-bed boiler make use
of the baffle-screen concept so that the freeboard height
will not be as great as that required for disentrainment of
bed particles by gravity alone.
POPE, EVANS AND ROBBINS
INCORPORATED
-------
196
7.6
Performance of a Baffle Screen As a Heat Exchanqer
As described in the previous section, an array of
water-cooled tubes was installed in the CBC to reduce
carryover of bed material. The screen, consisting of 1"
tubes arranged on a triangular pitch, was located between
the 32" level (i.e., 32" above the air distribution grid)
and the, 39" 1eve 1. (See Figure 9). The expanded bed depth
was about 20" for most tests. Therefore, there was, on the
average, a 10-12" gap between the dense phase fluidized bed
and the bottom of the baffle screen. As noted earlier, there
is considerable splashing in the dense-phase bed so that hot
particles would hit the tubes.
In designing the screen, it was anticipated that
the effective average heat transfer coefficient would be no
higher than about 18 Btu/ft2hroF, calculated for convection
from the gas and radiation from the bed to the viewed tubes.
If the coefficient were 18, a heat absorption rate of
~ 150,000 Btu/hr would be anticipated for the tube array.
When actual absorption was determined by measuring
the coolant flow and temperature rise, a value of over
300,000 Btu/hr was indicated. Overall coefficients of ~ 35
Btu/ft2hroF were calculated, using the bed temperature as
the gas side temperature in the basic heat transfer equation.
If actual gas temperatures entering and leaving the tube
bundle were used~ the coefficients compare to ~ 50 Btu/hr ft20F
to a single tube immersed in the fluidized bed.
This result demonstrates that a heat exchanger
placed above the dense phase of a high-velocity fluidized bed
will experience a heat flux almost as great as if it had been
immersed in the bed. This is significant for the design of a
Mu1tice11 Fluidized-Bed Boiler, since a much lower fan power
POPE, EVANS AND ROBBINS
INCORPORATED
-------
197
penalty is paid for surface located above the bed than in
raising the bed level so as to immerse more surface. While
the height of this baffle screen above the bed could not be
varied, it was found in more recent tests that the heat
transfer coefficient decreased as the distance above the
bed increased.
In the absence of tests with a movable heat transfer
. 2 0
bundle,* a film coefficient'" 35 Btu/ft hr F should be used
for heat exchangers located as in the CBC ( '" 12" above the
expanded bed). This coefficient should apply only to about
the first four rows in the exchanger. Beyond the fourth row,
the fluidized-bed should no longer affect the heat exchange
coefficient and normal convective transfer coefficients may
be calculated and used in design.
7.7
Performance of Inteqrated FBM/CBC System
The performance of the FBM and the appended CBC
which was developmental in nature was poor during most of
the test program described here. Test data, listed in
Appendix K, are summarized below in Table 20. It may be
seen that in only two tests did the system operate at an
integrated combustion efficiency of about 99%, and this was
when both the FBM and CBC burned carbon efficiently. Burnup
calculations are uncertain when the ash output is not weighed.
Because of a number of problems, some of which were
outlined in the preceding sections, it was not found possible
*Tests with single heat-transfer probes are not considered
adequate to provide engineering data for design of heat
exchangers made up of an array of tubes. One reason is
that aerodynamic similarity is inadequate.
POPE. EVANS AND ROBBINS
INCORPORATED
-------
198
TABLE 20
PERFORMANCE OF FBM, CBC AND SYSTEM IN CURRENT TEST SERIES
COMBUSTION EFFICIENCY, %
TEST NO. IN FBM 1 IN CBC1 IN SYSTEM 2 .
B-5 82
B-7a 75
B-7b 78
B-8 76 75 94
B-9 76 74 94
B-10 78.4 72 94
B-11 75 53 89
B-11b 61 73 90
B-12 78.5 67 93
B-12b 67 57 87
B-12e 73 35 84
B-13c 83.2 60 93.3
B.-14 81.5 95 99.1
B-15 80.5 93 98.6
B-16b 79
B-17 83
1CE = (Carbon Burned/Carbon In) x 100
2CE = (Heat Released in FBM and CBC/HHV of Coal Feed) x 100
to perform the numerous parametric studies with the FBM which
might have provided additional useful data. The desired
statistically designed test program was not conducted with
the FBM.
POPE, EVANS AND ROBBINS
INCORPORATED
-------
199
The operational problems found in the FBM during
this test series were as follows:
1) Debugging of the vertical coal feeder (Described
in Sect ion 7.2).
2) The new coal handl ing system,
ten-foot length of inclined screw, produced
in the coal feed. The coal delivered at the
which included a
excessive fines
outlet of the
feeder was captured in a cold test and found to contain 60
wt. % of -16 mesh material. This is roughly twice the fines
content of the coal used in past work (1:., .1). In ope rat ion,
the fines would be carried out of the FBM without completely
burning; thus, the inability to reach the 85% carbon
combustion efficiency expected from the FBM. The coal
handling system could not be repaired or replaced within the
time and funds available.
3) There was evidence that the grid plate had
warped in the vicinity of the CBC and that some air was by-
passing the bed. This was indicated by signs of erosion.
The symptom of bypassing of this type, would be a lower carbon
combustion efficiency, higher hydrocarbons in the flue gas and
poorer sulfur removal by limestone at coal/air ratios that in
the past had produced better results. The plate was caulked
but this was only a temporary repair. Recently (1971) a
metal seal was installed that may have reduced the leakage.
The operational problems found in the CBC during
this test series were as follows:
1) The excessive cooling due to the high rate of
interchange of cool bed material from the FBM (See Section
7.4).
2 )
The poor performance of the original fly ash
POPE. EVANS AND ROBBINS
INCORPORATED
-------
200
in jector (See Sect ion 7.3).
3) The excessive carryover of bed materials
caused by the inadequate freeboard (See Section 7.5).
4) The excessive cooling due to the unexpected
high heat flux into the water-cooled baffle screen (See
Se ct ion 7.6).
As noted above, the last two tests in the integrated
system were highly efficiently, and when most of the
operational problems had been eased both the FBM and CBC
performed acceptably well.
The effectiveness of the performance model, derived
in the FBC, was tested with the results of CBC tests. The
observed and calculated combustion efficiencies are shown in
Figure 38. It is seen that the results of only two tests
were predicted precisely. It is unfortunate that time was
not available to perform a series of tests with the properly
operating apparatus over the entire efficiency range (75% to
+90%) to fully test the performance model.
7.8
Sulfur Capture with Fine Limestone Injection
Reference 1 describes the procedures and results of
a number of tests in which fine limestone was added to the
FBM before the CBC was appended. The details of these tests
will not be repeated here. The addition of the CBC to the
FBM required major changes to the apparatus. Among these
changes was a new coal injector; another was a new limestone
injector. These have been described in Section 5. The
procedures used in the current test series were identical to
those described in ClJ. Presumably then, only the equipment
changes could have affected the sulfur removal performance
of the FBM.
POPE, EVANS AND ROBBINS
INCORPORATED
-------
120
110
-
Q
riI 100
~
u
~
-
... 90
tIP
...
>t
U
:zoo
tiI
H
U. 80
H
f&"
i:&.
riI
:zoo
0
H'
E-4
U) 70
::>
~.
0 .
U
60
50
50
.
AIR RATE TOO HIGH!
FOR PERFO~~CE MODEL
.
o
.
.
POOR PERFOID1-
ANCE DUE TO
EXCESSIVE
COOLING OF
CBC. SEE
TEXT FOR
DISCUSSION
POPE. EVANS AND ROBBINS
INCORPORATED
90
COl1BUSTION EFFICIENCY, %, (OBSERVED)
60 '
70
80
rIGURE 38
COMPARISON OF OBSERVED N~D CALCULATED
COMBUSTION EFFICIENCY OF CBC
201
100
-------
202
When fine sorbent was injected into the FBM, it
would pass through the unit to the FBM dust collector.
Then, along with the carbon-bearing fly ash, the partially-
sulfated limestone would be injected into the CBC, pass
through the CBC, and be collected by the CBC dust collector.
From the CBC dust collector hopper, this mixture of "burned
out" fly ash and lime by-products could be wasted or
reinjected into the boiler.
Figure 39 compares the data previously obtained
with some of the data obtained in this program. It may be
seen that more limestone was required in the current test
series to achieve the same level of S02 reduction. No
reason for these poor results was apparent, and no attempt
was made to alter the system so that the formerly high
efficiency could be restored. It is believed that the
currently installed additive injectors are less effective
than those used in (1) and that much of the additive leaves
the bed in one location without reacting with sulfur. Poor
coal distribution might also account for poor sulfur removal.
Emphasis is now being given to a regenerative lime-
stone process (Section 4.3.3) instead of once-through fine
limestone injection. Therefore, no additional effort in
this area could be justified.
POPE, EVANS AND ROBBINS
INCORPORATED
-------
20
~ 30
~
u
~ 40
~
p.
...
~ Z 50
0
I H
8
U
~~ 0 60
CI
~
~U> ~ 70
g~ CI 0
~~ H
~
0
~ 80
~
m ::J
~ 90
~
0
en
1 2
o PREVIOUS FBM RESULTS (REF. 1)
~CURRENT FBM RESULTS
DATA FOR NO. 1359 SORBENT, -325 MESH
RESULTS FOR NO. 1359 SORBENT, -325 MESH
3
4
5
6
7
LI~mSTONE: COAL RATIO AS Ca/S MOLE RATIO
FIGURE 39
8
SULFUR DIOXIDE REDUCTION WITH FINE LIMESTONE INJECTION INTO FBM
N
«:)
w
-------
204
8.
ECONOMICS OF THE CARBON-BURNUP CELL
8.1
Pe r spec t i ve
The detailed design and pricing of a Mu1tice11
Fluidized-Bed Boiler were beyond the scope of the current
program (See Reference 4 for detailed cost data on an early
Mu1tice11 Fluidized-Bed Boiler design). A cost estimate is
not provided.
In the current study, devoted almost entirely to
the Carbon-Burnup Cell portion of the Mu1tice11 Fluidized-Bed
Boiler, the most significant question of economics concerns
optimizing the Carbon-Burnup Cell as a part of the boiler.
Unfortunately, as noted in Section 6.4.8, the performance
model is not SUfficiently sophisticated to permit a meaningful
evaluation of the comparative capital or operating costs of
various alternative Carbon-Burnup Cell designs. More
specifically, for example, it is not meaningful to compare
the costs for operation with a bed depth of 24" and a bed
depth 48" since Equation 1 would give an incorrect result
for the 48" bed depth. For this reason it is not possible,
using the information available, to optimize the Carbon-
Burnup Cell alone.
Despite the performance model's deficiency, it is
not recommended that any additional effort be devoted to
extending its range to permit such optimization. The existing
model describes with sufficient precision how a Carbon-Burnup
Cell should be designed.
The material which follows will show how the Carbon-
Burnup Cell performance model plus data on the combustion
efficiency in the primary cells of the multi-cell boiler may
be used to optimize the design of the entire unit. The analysis
POPE, EVANS AND ROBBINS
INCORPORATED
-------
205
is based on the premise that the unit is optimized when fuel
costs are minimized and that fuel costs are minimized when
the thermal efficiency of the boiler is maximized.
in this statement are assumptions that:
Implicit
1) The capital cost of burnup cell and primary
cell, expressed in dollars per Btu of heat release rate, are
equal. Since the primary cells may be operated at heat
release rates of from 800,000 to 1,200,000 Btu/hr ft2 of
grate area, while the burnup cells would be operated at
500,000 to 800,000 Btu/hr ft2, this assumption is only
approximately true.
2) The auxiliary power requirements (mainly for
the forced draft fans) expressed as KWH per Btu of heat
release are equal for the primary and burnup cells. This
assumption is also not strictly true since from 10 to 2~~
more air, per Btu of heat release, will be used in the burn up
cells than in the primary cells.
The specific capital and auxiliary power costs of
the burnup cells are therefore somewhat higher than the
same costs in the primary cells. However, it is felt that
the 10 cell - burnup cell system can be optimized without
regard to these differences in cost because:
1) The capital costs which are heat-release-rate
and excess-air-rate dependent are in the air distributor and
in fans, ducts, breeching and dust removal equipment. Heat
transfer surface is not included in this category. A 5~~
change in heat release may result in only about a l~~ change
in the capital cost, $/Btu of heat release, of these items.
2) The auxiliary power required for fans by the
Multicell Fluidized-Bed Boiler will be on the order of 1.0
POPE. EVANS AND ROBBINS
INCORPORATED
-------
206
to1 0 5 KWH per 106 Btu of heat released. For a 40% efficient
power station, 1 KWH - 8550 Btu of fuel input or about 0.85%
6
of 10 Btu. The difference in specific auxiliary power
requirements of burnup cell and primary cell are therefore
neg1 igib1e .
3) It will be shown that an optimum design is
achieved when the primary cells burn about 9~~ of the fuel
and the Carbon-Burnup Cells the remaining 1~~. The curves,
to be developed, are flat and no sharply defined optimum is
established.
On the basis of these three arguments it is felt that
the simplifications do not detract from the essential validity
of the premise that the system is optimized when the boiler's
thermal efficiency is maximized.
802
Analysis
In numerous tests in fluidized-bed coal combustors,
it has been found that a sizable fraction of the fuel value
of the input coal may be lost as unburned carbon in fly ash.
This loss may range from 5% to 20% of the input carbon, or
about 5% to 15% of the input energy value of the fuel when
the hydrogen content of the coal is accounted for.
Two basic methods of recovering the fuel value in
the fly ash have been proposed. In one design concept, the
boiler is operated at low air rates, (~2 to 4 fps superficial
gas velocity) and the collected fines which contain unburned
carbon are recycled back to the fluidized bed from which the
fines originated. In a second design concept, which is
favored here, the fluidized-bed boiler operates at high air
rates (~10-15 fps). Here the collected, carbon-bearing fines
are recycled back to a special region of the fluidized bed in
POPE, EVANS AND ROBBINS
INCORPORATED
-------
207
which a relatively high residual oxygen and temperature
( > 1950oF) are maintained by providing less heat-transfer
surface per unit fuel burned than in the coal-burning regions.
The data gathered in an earlier study (~) established
that the carbon in fly ash could not be effectively burned in
a fluidized bed at l500-l700oF with 3% residual oxygen in the
flue gas at high (~ 6 fps) air rates. Pending the disclosure
of an invention which provides an alternative to the Carbon-
Burnup Cell, such a region must be provided if high levels of
combustion efficiency are to be achieved in an economically
viable fluidized-bed boiler. As examples of boilers which we
do not consider economically viable, consider those with high
excess-air levels and those with unrealistically low gas
velocity.
Figure 40 shows the response, carbon combustion
efficiency, for the carbon in coal, as a function of the major
pseudo-control variable, residual oxygen in the flue gas.
Although residual 02 is itself a response to fuel rate, air
rate, fuel reactivity, etc., in actual operation, residual
02 is treated as a control. Figure 40 shows how combustion
efficiency increases with increasing residual 02.
Note that these data, obtained in different apparatus
over the past several years, represent coal burning tests,
not tests in which fly ash was the fuel. The combustors were
operating in the manner of a primary cell, not as a carbon
burn up cell. The fuel was 1/4" x 0, not -1 mm.
A power curve was fit to the data which appears
intuitively correct although its mathematical form is not the
most appropriate. This curve has the form
CE = a (RO) b
POPE. EVANS AND ROBBINS
INCORPORATED
-------
208
95
o
0
Ci""O
><
-
Q 90
~
~
~
~
U
~ 85
H
z
0
~
u
~ 80
o
t/)
o
5
o
Po.
,
Q
~ 75
~
p
a3
~
o
~ 70
~
o
Q
5
o
~ 65
811
~
U
o
o 0
o
o
o
o
o
o
o
o
o
o
o
60
1
3
o
o 0
o
2
o
o
o
.0.
o
o
CE = 71.2(RO)0.145
FOR:
1. CARBON IN COAL
2. BED TEHPERATURES:
1500 TO 1750°F
3. SUPERFICIAL GAS VELOCITIES;
6 fps TO 15 fps
NOTE:
THIS DATA IS
APPLICABLE FOR THE
CARBm~ IN COAL ~'lIO
RECYCLE OF COLLECTED
FLY ASH
4
5
6
7
RESIDUAL OXYGF.N IN FLUE GAS, RO, ,
. FIGURE 40: CARBON COHBUSTION EFFICIENCY IN A COAL-FIRED
FLUIDIZED-BED CmmUSTOR
POPE. EVANS AND ROB8I\IS
-------
209
whe re
CE = (weiqht of carbon burned)
weight of carbon in coal fed
x 100
(41)
RO = residual 02' vol. %
with coefficients
a = 71.2 and
b = 0.145
for RO between 1.3 and 7.0, with standard deviation s = 4.1%.
The relationship
100
100-CE
=
2.6 + e(0.39 x 00)
(42 )
could be used for extrapolation to high values of RO. For
RO between 1.0 and 8.0, Equations (41) and (42) give identical
results.
With this expression and other data obtained in this
study, it is possible to determine:
bo i 1e r
recycle
1. The optimum operating point for a fluidized-bed
containing primary cells only and without fly ash
of any kind.
2. The optimum division of duty between the coa1-
burning primary cells and fly ash-burning Carbon-Burnup Cells
of a Mu1tice11 Fluidized-Bed Boiler for minimum fuel costs.
8.2.1
Evaluation of a Fluidized-Bed Boiler Without Fly Ash
Cycle
Example 1.
Problem: Using Equation (41), determine the optimum
level of residual oxygen for a fluidized-bed boiler consisting
only of primary cells and in which the fly ash is not recycled.
POPE, EVANS AND ROBBINS
INCORPORATED
-------
?10
Solution: Detennine the residual oxygen level at
which the sum of the energy loss in unburned carbon plus
that in the flue gas is minimized.
Data:
a. The coal described in Table 16 (Page 150) will
be burned having an as-fired heating value of 12,500 Btu/lb.
b. Flue gas will be discharged at 4000F with an
enthalpy of 81 Btu/lb of gas.
c.
Carbon in fly ash has a fuel value of 14,100
Btu/lb.
d.
The latent heat loss, 49,000 Btu/lOO lbs of
coal,
is a constant and need not be considered in finding the
opt imum.
Results:
Figure 41 shows (1) the heat loss due to unburned
carbon as a function of residual oxygen (RO) level calculated
from Equation 41; (2) the sensible heat loss in 4000F flue
gas as function of RO; and the total heat loss being the sum
of (1) and (2).
The results plotted in Figure 41 reveal that the
total loss does not reach a minimum even at RO = 8%. (The
computations showed a minimum at about RO = l~/o but the loss
is not much different than at 8%). The total loss for a typical
highly efficient pulverized coal (p.c.) furnace, also with
4000F exhaust temperature, is also shown on the figure. The ~
range shown is for 15 to 20% excess air and a 0 to l.~/o
combustible loss typical of p.c. units.
Note that the minimum loss for the fluidized-bed
boiler is higher than that for the pulverized coal boiler.
POPE. EVANS AND ROBBINS
INCORPORATEO
-------
~ 400
0
u
~ (1)+(2) ASSUMPTI.ONS
0
tI) 1. LATENT HEAT LOSS CONSTANT AND
t:Q 300 NOT SHOvlN
H
~. 0 2. COAL FROH TABLE 16
0
.-f 31 CURVE (2) BASED ON EQ. 41
.......
~~ 0
E-f
co
~U> ~ 200
g~ 0
~; V) ~OTAL LOSS TYPICAL UNBU&~ED
-
0 -x:5 OF HIGHLY EFFICIENT
0
0 PULVERIZED COAL nOILERS
.-f
m .. 100 .l.
tI)
U)
0
H (1) G~S LOSS AT 4000F
~
~
~
r::I
Z 4 5 6 7 8
w 0 1 2 3
RO = RESIDUAL OXYGEN IN FLUE GAS, PERCENT.
FIGURE .41
THERMAL LOSSES FROM FLUIDIZED-BED BOILRR OPERATING WITHOUT RECYCLE OF
CARBON-BEARING FLY ASH
N
....
....
-------
212
Including the latent heat loss, the fluidized-bed boiler
achieves a maximum thermal efficiency of 83.4% while the
unit is 86.5% efficient.
p.c.
A Carbon-Burnup Cell is therefore required if the
fluidized-bed boiler is to operate with a thermal efficiency
comparable to that of the p.c. unit.
With no fly ash recycle, the minimum loss point
moves toward even greater values of residual oxygen as the
exhaust temperature is reduced.
8.2.2
Evaluation of a Multicell Fluidized-Bed Boiler:
Optimum Duty Split Between Primary Cells and
Carbon-Burnup Cells
Example 2.
Problem:
In the case of a Multicell Fluidized-Bed
Boiler in which the carbon lost by the primary cells is
burned in the Carbon-Burnup Cells, determine if an optimum
exists in the division of duty between the coal-burning cells
and the fly ash burning cells.
Solution:
Determine the residual oxygen level in
the coal burning regions at which the total energy losses
are minimized.
Data:
-
a .
The coal described in Table 16 will be burned.
b. Flue gas will be discharged at 4000F with an
enthalpy of 81 Btu/lb from the coal-burning regions (5%
moisture) and 79 Btu/lb (~l% moisture) from the fly ash
burning regions.
c.
Carbon in fly ash has a fuel. value of 14,100
Btu/lb.
POPE, EVANS AND ROBBINS
INCORPORATED
-------
213
d. The latent heat loss, 49,000 Btu/100 pounds of
coal, is a constant and need not be considered in finding an
opt imum .
e.
The Carbon-Burnup Cell has the following
parameters:
(1 )
Fuel input rate = Ct(14,000) = 500,000
Btu/ft2hr
Bed temperature, T = 20000F
(2 )
(3 )
(4 )
Bed depth, H = 24"
Combustion efficiency, CE = 85, 90, and 95%
Procedure:
1. Assume a value of RO from the primary cells,
calculate CE for carbon in coal using Equation 41. From
this compute the composition of the primary cell fly ash,
% C, % inert to be fed to the burnup cells.
2. Compute the weight of air required by the primary
cell at the assumed value for RO. For these calculations
assume complete combustion of the hydrogen and sulfur content
of the coal. The air for carbon and residual oxygen is
computed from Equation 21. Compute the total weight of flue
gas produced and the sensible heat loss at 4000F.
3. Using Equation 1 and for the conditions given
above in the data for Carbon-Burnup Cells, compute the air-to-
carbon ratio At/Ct to produce the specified burn up cell
combustion efficiency. Compute the flue gas weight and
sensible heat loss at 4000F. Compute the heat value for
unburned carbqn.
4.
For one value of RO in the primary cell flue
POPE. EVANS AND ROBBINS
INCORPORATED
-------
214
gas and one value of burnup cell efficiency compute the
total heat loss (neglecting the latent heat loss) which is:
(a) the sensible heat loss of primary cell
flue gas; plus,
(b)
the sensible heat loss of burnup cell
flue gas; plus,
(c)
the heating value of unburned carbon.
5. Repeat computations for a range of values of
primary cell RO and burnup cell CE's.
Results: The sum of thermal losses, dependent on
the residual oxygen level, from the two regions of a Multicell
Fluidized-Bed Boiler are shown graphically in Figure 42 (add
49,000 Btu to the values shown to include the latent heat loss).
It may be seen that an optimum does exist at about 2-1/2 to
3-1/2% residual oxygen in the primary cell flue gas. This
range in residual oxygen may be related to the duty on the
Carbon-Burnup Cell. At RO = 2-1/2% in the primary cell flue
gas, the Carbon-Burnup Cell fires 14.5% of the total fuel
input; at RO = 3-1/2%, the Carbon-Burnup Cell has 11% of the
total duty. Total fuel input is 1,255,000 Btu/IOO pounds of
coal. The total. loss shown for the p.c. unit of typical
efficiency is the same as in the preceding example.
A second case (more heat recovery equipment used so
the flue gas exhausts at 2500F) is shown in Figure 43. A
slight shift in duty to the coal-burning regions is indicated
here. In the case of a 2500F gas discharge temperature, the
Carbon-Burnup Cell should burn 9 to 12% of the fuel. However,
the curves are so flat that little penalty in efficiency is
paid in optimizing on criteria other than duty division.
Thermal loss for a p.c. unit is not shown for a
2500F exhaust since low temperature corrosion, due to H2S04'
POPE, EVANS AND ROBBINS
INCORPORATED
-------
Z = FRACTION OF TOTAL DUTY O~ CARBON-BU~~UP CELL
~ 140
o
u
~.
0
~ CJ)
co
...:I 130
o
Z~ 0
M
n . "
0 . ::>
~(J) E-1
g~ co
~
~O 0 120
a; en
0
m 0
0
r-I
-. 110
CJ)
CJ)
0
...:I
)4
~
P::
~
z 100
~
o
0.30 0.25
'"
~95
0.20
0.125
0.075
0.10
0.15
OVE?ALL CO:1BUSTION EFFICIENCY = 100 Z (100- .E)
CE = co:mUSTI()~1 EFFICIENCY OF CARBON-BURNUP
/
5
90
SEE TEXT FOR DATA
&~D ASSm.lPTIONS
CURVES REP~ESE~T THEID1AL LOSS
IN SEnSIBLE HEAT OF FLUE. GAS
AT 400°F PLUS HEATING VALUE OF
u:.mURJ.'1ED CARBON
1
2
3
4
5
6
7
RO = RESIDUAL OXYGEN IN FLUE GAS OF PRI~~RY CELLS, %
N
....
VI
FIGURE 42
THERr1AL LOSSES. FRO~1 HULTICELL FLUIDIZED-BED BOILER -
EXHAUST TEt.lPER1\TURE 400°F
-------
~
~~
~U>
o}>
]I Z
~;
~
N
....
C1'\
Z:: FRACTION OF TOTAL DUTY ON CARBON-BURNUP CELL
0.30 0.25 0.2 0.15 0.125 0.10 0.075
J
<[
890
1.1..
o
en
m
J
080
o
.......
:)
t-
m
1.1..
070
-'"
o
o
o
-
en
en 60
o
J
>-
C)
a::
w
z
w 50
o
COMBUSTION EFFICIENCY = lOO-ZOOO-CE)
/CE=COMBUSTION EFFICIENCY OF CARBON-BURNUP
CELL, - %
85 LOCUS
90
"
95
SEE TEXT FOR DATA AND
ASSUMPTIONS
2
3
CURVES REPRESENT THERMAL
LOSS IN SENSIBLE HEAT OF
FLUE GAS AT 250°F PLUS
HEATING VALUE OF UNBURNED
CARSON
4
5
7
8
6
RO= RESIDUAL OXYGEN IN FLUE GAS OF PRIMARY CELLS, 0/0'
FIGURE 43 THERMAL LOSSES FROM MULTICELL FLUIDIZED-BED BOILER-
EXHAUST TEMPERATURE 250°F
-------
217
is avoided by operating at an exhaust temperature in excess
of 300oF.
Conclusions: It is seen that with a Carbon-Burnup
Cell, fuel consumption in a Multicell Fluidized-Bed Boiler
will be comparable to that of a conventional pulverized-fuel
boiler at a gas exit temperature of 400oF. Since lower
grade, and hence lower cost, fuels may be burned in the
fluidized-bed boiler, fuel costs will be lower than in the
pulverized-fuel boiler.
If the gas discharge temperature is relatively low,
220 > Tgas ~ 300oF, the optimum duty division between the
primary cells and the burnup cells covers a wide range and
the selection should be based on other criteria. For hot
discharge, ~350oF, a slightly more pronounced optimum occurs
between 10-15% duty on the Burnup Cell. If fuel costs are
relatively high and sulfur control effective, it may be
economically feasible to build Multicell Fluidized-Bed
Boilers with an exhaust temperature of 2500F or even lower.
The overall thermal efficiency at 2500F exhaust would be
90.5 to 91.5% depending on the efficiency of the Carbon-
Burnup Cell for the coal shown.
POPE. EVANS AND ROBBINS
INCORPORATED
-------
218
9.
SUMMARY OF DESIGN CRITERIA FOR CARBON-BURNUP CELL
The bulk of this report has been devoted to a
discussion of the design of a Carbon-Burnup Cell. The
purpose of this section is to summarize the conclusions on
opt imum des ign.
Combustion efficiency and minimum particulate
emissions are achieved by operating a high bed temperature.
With reliable controllers, consistent operation at 20500F
appears feasible. With the development of a technique for
continuously removing larger agglomerates, operation at bed
temperature exceeding 20500F would appear desirable.
The model equations of Section 6 may be used to
optimize the Carbon-Burnup Cell for a particular application.
Removal of energy by contacted and viewed heat
exchange surface should not exceed about 25% of the energy
re le a se .
Air rates consistent with high superficial
velocities (10 to 15 fps) may be used and are desirable for
a compact design.
The unit should be designed to fire more fly ash
than is anticipated so that under no reasonable circumstances
will carbon-rich fly ash accumulate in the hoppers. The
energy deficiency, which follows from this design arrangement
is to be made up with coal. No need exists to operate with
the Carbon-Burnup Cell as a separate system. It is an
POPE, EVANS AND ROBBINS
INCORPORATED
-------
219
integral part of a Multicell Fluidized-Bed Boiler and should
share a common bed with the primary cells. If fine lime were
injected to reduce sulfur oxide emissions, the Carbon-Burnup
Cell should operate with a r~sidual oxygen level> 3% to
minimize decomposition of partially sulfated limestone.
A static bed depth of
reasonably efficient operation;
is an appropriate design range.
over 10" is required for
a bed depth of 18" to 24"
POPE. EVANS AND ROBBINS
INCORPORATED
-------
220
10.
REFERENCES
(1) Characterization and Control of Gaseous
Emissions from Coal-Fired Fluidized-Bed Boilers, Interim
Report, October 1970, for Division of Process Control
Engineering, National Air Pollution Control Administration,
Department of Health, Education and Welfare, by Pope, Evans
and Robbins, PB198 413.
Report covers mostly fine limestone injection
into fluidized bed combustors.
(2) U. S. Patent No. 3,508,506, Process and
Apparatus for Reduction of Unburned Combustible in Fly Ash.
The term, "Carbon-Burnup Cell" was coined to
identify the methods and apparatus described in this patent.
Since the words "Carbon-Burnup Cell" refer to a specific
design concept they should always be capitalized.
(3) Serial No. 66,724 filed U. S. Patent Office
August 25, 1970.
The "S02 Acceptor Process" described in this
patent application is a method of controlling 802 emissions
from any boiler but most likely fluidized-bed boiler in
which a gas containing more than about 5% S02 is produced
for use or neutralization. The patentable features pertain
to the methods of making this gas at relatively low cost
and free of potential pollutants (other than 802).
(4) Development of Coal-Fired Fluidized-Bed
Boilers prepared for Office of Coal Research, Department
of the Interior, by Pope, Evans and Robbins, Interim Report,
Office of Coal Research, Research and Development Report No.
36, Contract No. 14-01-0001-478. Available from Superintendent
of Documents.
Report covers experimental and design studies
for compact fluidized-bed boilers carried on by Pope, Evans
and Robbins from 1965 through 1969. Describes apparatus
used in work reported here.
(5) U. S. Patent No. 1,687,118 (October 9, 1923 and
D.R.P. 437,970 (1922).
Winkler describes the invention of fluidized
bed combustion and gasification.
POPE. EVANS AND ROBBiNs
INCORPORATED
-------
221
(6) U. S. Patent No. 1,984,340 (December 18,
1934, reissued August 6, 1940).
Odell describes the invention of f1uidized-
bed catalytic reactors. Odell stated recently (1970) in a
personal communication that he had considered the feasibility
of a coal-fired fluidized-bed boiler in the 1930's but had
thought that coal ash clinkers would form and make such a
system unworkable.
(7)
U. S. Patent No. 2,509,866 (May 30, 1950).
Hemminger describes how a coal-fired f1uidized-
bed boiler would operate. He recognizes that it would not
be possible to burn all of the coal in a single stage.
(8) Gode1, A. A., "Ten Years of Application of
Fluidized-Bed Combustion of Coal" Rev. General de Thermique,
1966, 5 (52), 349-359.
Gode1 explains the insights which led to the
success of his Ignif1uid stoker for the fluidized-bed
combustion of low grade coal.
(9) Bishop, J. W., et a1, Status of Direct Contact
Heat Transfering Fluidized-Bed Boiler, American Society of
Mechanical Engineers, Winter Meeting, December, 1968, Paper
68-WA/FU-4.
Bishop outlines the development of the Mu1ti-
cell Fluidized-Bed Boiler.
(10) Jonke, A. A., et a1, Reduction of Atmospheric
Pollution by the Application of Fluidized-Bed Combustion,
Annual Report, July 196R-June 1969, ANL/ES-CEM-1001.
In Figure 21 a reduction in NO content, in
flue gas, is shown when fine limestone is added to a fluidized
bed combustor. NO is shown to decline as S02 declines. (See
11 be low) .
(11) Hammons, G. A. and Skopp, A., A Regenerative
Limestone Process for Fluidized Bed Combustion and Desu1furi-
zation, Final Report, February 28, 1971 for Process Control
Engineering Program, Air Pollution Control Office by Esso
Research and Engineering.
POPE. EVANS AND ROBBINS
INCORPORATED
-------
222
In Figure 25 a reduction in NO content in flue
gas, is shown when a bed of coarse limestone becomes inactive
for sulfur removal. NO is shown to decline as S02 increases.
(12) Fine, D. H., et a1, The Importance of Nitrogen
in Coal as a Source of Nitrogen Oxide Emission from Furnaces,
Fuels Research Laboratory, Massachusetts Institute of
Technology. Delivered 64th Annual AIChE meeting. Undated
(received December 1971).
Fine notes that the experimental finding that
NO emissions from a fluidized-bed combustor are higher than
anticipated is "due to the intimate contact between the
part ic1es and entra ining a ir stream" (p. 19). He may be
right. .
(13) Hoy, H. R. and Roberts, A. G., Fluidized
Combustion of Coal at High Pressure. Paper delivered at
64th Annual Meeting of the American Institute of Chemical
Engineers, 28 November 1971.
Hoy shows in Figure 5 the following relationship
between excess air and (overall combustion efficiency) l~/o
(96%), 20% (98%), 35% (99%), 33% (99.5%). These data are
believed to be at 2-3 fps, 5-10 atmospheres, 1500oF, 2-3 foot
static bed depth.
(14) Demming, L. F., Trip report conta ined in
Progress Report No.3, under Office of Coal Research Contract
14-01-0001-478, for May 1965.
Demming reports about his meeting at the
Marchwood Laboratories of the British Central Electricity
Generating Board. CEGB had found that operation of a
fluidized bed combustor was good in a bed 3 ft deep with a
superficial gas velocity of 0.5 fps. At 1 fps carryover
becomes objectionable.
(15) The Federal Research and Development Plan for
Air Pollution Control By Combustion-Process MOdification,
Final Report, January 11, 1971 for Air Pollution Control
Office, Environmental Protection Agency by Battelle Memorial
Institute, Columbus Laboratories.
(16) Rice, R. L., and Coates, N. H. Fluid Bed
Combustion: Suitability of Coals and Bed Materials, Power
Engineering, December 1971, 36-38.
POPE. EVANS AND ROBBINS
INCORPORATED
-------
223
Rice and Coates point out that in their
apparatus lignite is the only coal which can be burned
completely without recycle.
(17)
U. S. Patent No. 3,171,369 (March 2, 1965).
Stephens and Goldberger show that high
temperatures and high excess air are required for complete
combustion of carbon in finely ground coal.
(18) U. S. Patents 2,818,049; 2,842,102; 2,976,853;
2,983,259; 2,997,031; 3,048,153 and 3,119,378.
First U. S. Patents for fluidized-bed boiler
based on experimental work; concerned mainly with heterogenous
combustion catalysts.
(19) Ehrlich, S., The Fluidized-Bed Boiler, paper
presented at the 1968 Technical Meeting, The Eastern Section
of the Combustion Institute, Amherst, Massachusetts, October
22, 1968.
Describes problem areas in which basic
research on fluidized bed combustion and heat transfer are
required.
(20) Essenhigh, R. F., Professor - Department of
Fuel Science, Pennsylvania State University. Various
personal communications, February 1967 through November 1968.
Dr. Essenhigh describes areas in which useful
and interesting basic research might be carried out in the
field of fluidized bed combustion and heat transfer.
(21) American Society of Mechanical Engineers,
Power Test Codes, Steam Generating Units PTC 4.1 - 1964 and
Addenda.
This code presents the accepted accounting
method for defining the efficiency of a boiler. PTC 4.1 -
1964, Section 9.4, recommends the use of 14,500 Btu/lb for
the heating value of carbon in refuse. Note that this report
has used 14,100 Btu/lb of "loss-on-ignition" which mayor
may not be all carbon.
(22) Essenhigh, R. F. and Fells, I., Discussions
Faraday Society, Volume 31 (1961) page 208.
POPE. EVANS AND ROBBINS
INCORPORATED
-------
224
(23) Second International Conference on Fluidized
Bed Combustion sponsored by Division of Process Control
Engineering, National Air Pollution Control Administration
at Hueston Woods Lodge, College Corner, Ohio on October 4-7,
1970.
Figure 15 of paper by Anas~asi3' et aI, shows
an optimum sulfur capture temperature of 1450 F to l5500F
(oxidizing conditions).
Figure 2 of paper by Moss shows optimum .
sulfur capture temperature of 8750C (oxidizing conditions).
(24) Goldberger, W. M., Collection of Fly Ash
in a Self-Agglomerating Fluidized Bed Coal Burner, American
Society of Mechanical Engineers, Winter Meeting, November,
1967, paper 67-WA/FU-3.
(25) Loshe, G. E. and Commander, R. E., Initial
Operation of the Idaho Waste Calcining Facility with Radio-
active Feed, Proceedings of the Symposium on the
Solidification and Longterm Storage of Highly Radioactive
Wastes, Richland, Washington, February 1966.
POPE, EVANS AND ROBBINS
INCORPORATED
-------
A-l
APPENDIX A.
FBC SPECIFICATIONS
1.
Air Supply
Two centrifugal fans in series for 300 cfm at 30" w.g.
connected to a smooth 4" diameter conduit 20' long.
Air flow is controlled with a gate valve and monitored
with a venturi, pitot pressure, static pressure, and
temperature measurements.
2.
Plenum
Mild steel, 1/4" thickness, 21" x 18" x 12" outside
dimensions with 8" diameter air inlet.
3.
Water Column
Mild steel, 1/4" thickness, 24" x 20" x 36" outside
dimensions with 16" x 12" x 36" inside dimen~ions.
A.
Wall on inlet air side contains:
a)
One nominal 3" diameter pipe for lightoff burner.
b)
One nominal I" diameter instrument port.
B.
Left wall (facing air inlet) contains:
a)
One nominal 2" diameter pipe with valve for
removal of bed material.
b)
Eight nominal I" diameter instrument ports at
various levels.
c)
One nominal I" diameter water outlet.
d)
One nominal 2" diameter pressure relief port.
C.
Right wall (facing air inlet) contains:
a)
One rectangular 2" x I" coal feed port.
b)
One nominal 3/4" diameter cooling water inlet.
D.
Wall opposite the air inlet contains:
Three nominal 1-1/2" diameter ports.
POPE, EVANS AND ROBBINS
INCORPORATED
-------
A-2
E. Insulating liner consisting of 1/4" ASTM 446
stainless steel extending from the grate to
connection with hood (37-3/4"). Internal
dimensions are 9-3/8" x 13-1/4". Kaowool
insulation is placed in the nominal 1" annular
space.
4.
Air Distribution Grid
The grid contains 130 stainless steel air distribution
buttons spaced on 1-1/4" centers, each containing eight
drilled ports, .089" diameter. The air is discharged
downward at an angle of 15° to the horizontal. with the
insulating liner installed, 42 of the air distribution
buttons are not in service.
5.
Water-cooled Hood
The hood is a truncated pyramid 24" x 20" at the bottom
and 17" x 17" at the top, with a height of 24" and a
flue opening 12" diameter. Material is #10 gauge mild
steel. One 4" diameter observation port is provided
with 1" diameter water ports and a 2" diameter pressure
relief port.
6.
Flue System
From the FBC-l hood, the flue system is run in 12" diameter
#10 gauge steel pipe to the induced draft fan. From the
fan the pipe is continued at 6" diameter, again #10 gauge
steel. All connecting sections are welded. The induced
draft fan may be bypassed.
7.
Dust Collector
The collector contains two 8" diameter centrifugal
collector units with a dust hopper, rotary feeder, and
a valve for fly ash removal.
POPE, EVANS AND ROBBINS
INCORPORATED
-------
NO.
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
A-3
LOCATION OF THERMOCOUPLES FOR FBC
LOCATION
Forced draft fan air
Plenum air
Bed 1-1/2"
Bed 3"
Bed 7"
Bed II"
Bed 15"
Flue gas in hood
Air distribution grid
Dust collector outlet
Gas sample line inlet
Flue gas exit
Cooling water in
Cooling water out-hood
Cooling water out-water walls
Isokinetic probe (above lab)
Sample gas discharge
NO sampling line
x
Flue gas exit (over lab)
POPE. EVANS AND ROBBINS
INCORPORATED
-------
B-1
1.
2.
3.
APPENDIX B.
FBM SPECIFICATIONS
Air Supply
One centrifugal fan at 2500 cfm at 50" w.g. connected
to 12-inch square duct which expands to full width of
plenum at inlet.
Air is controlled by means of a
damper and monitored by an orifice.
Plenum
Mild steel, J.,j:" thickness, 72" x 20J.,j:" x 12" inside
dimensions with a 6' x l' air inlet.
Boiler Construction
a.
Single 20" steam drum
b.
Dual 6" lower headers
c.
2~" risers on 4" centers for side walls
d.
4" downcomers
(external)
e.
5'4" distance from grid to uninsulated bottom of steam drum
f.
Combustion space = 53 ft3
Projected heating surface = 80 ft2
g.
h.
Average direct contact surface = 30 ft2
i.
Boiler capacity = 5000 lb/hr excluding convection heat
transfer; 7000 lb/hr including convection heat transfer
j .
k.
8.75 ft2 of bed area
Heat release rate:
800,000 to 1,200,000 Btu/ft2hr
1.
Pressure rating:
300 psi design, 200 psi normal operating
POPE, EVANS AND ROBBINS
INCORPORATED
-------
B-2
APPENDIX B.
FBM SPECIFICATIONS
(Continued)
4 .
Air Distribution Grid
The grid contains 815 stainless steel air distribution
buttons spaced on l~" centers, each containing eight
drilled ports, .089" diameter.
The air is discharged
downward at an angle of 15° to the horizontal.
5.
The flue system is fitted first with a water-cooled tube
array for temperature quenching, and a two-pass, 104 tube
(1" x 6'), 6000 air preheater; this is followed by a
second water-cooled gas cooler dust collector, which
exits to a 16" duct.
The system is drawn by a 4000 cfm,
5" w.g. static pressure, induced draft fan.
6.
Dust Collector and Fly Ash Reinjection
The dust collector contains twelve 10-inch diameter
centrifugal collector units with a dust hopper, a
4" Allen-Sherman-Hoff rotary feeder for fly-ash reinjec-
tion and a valve for fly-ash removal.
7.
Coal Input
700 - 900 Ibs per hour
8.
Thermocouple locations are listed for both the FBM and
CBC in Appendix C.
POPE, EVANS AND ROBBINS
INCORPORATED
-------
C-l
APPENDIX C.
CBC SPECIFICATIONS
1.
Air Supply
Shares air supply with FBM from centrifugal fan rated
at 2500 cfm, 50" w.g.
Air is supplied via a 4" nominal, Schedule 40 pipe,
and is controlled by a gate valve.
2.
Plenum
Mild steel, 1/4" thickness. Approximate dimensions
are 10" wide, 10" deep, and 18" high. The plenum rests
between the FBM's cross-headers.
3.
Column
A.
ASTM 446, stainless steel, 1/4" thickness,
10-5/8" x 15-5/8" inside cross-section 56" high.
B.
Gas outlet - 8" ID diameter starting at 48" level.
C.
Front wall contains 18"
X 24"
access plate.
D.
Left wall (facing toward common wall with FBM
contains
E.
Right wall (facing toward common wall with FBM
contains
F.
Back, or common wall, was varied. Open area, achieved
by cutting holes in a steel baffle, varied from
90 square inches to 2 square inches.
4.
Air Distribution Grid
Grid contains 96 ASTM 304 stainless steel air distribution
buttons spaced on 1-1/4" centers, each containing eight
drilled ports, 0.089" diameter. The air is discharged
downward at an angle of 15° to the horizontal.
5.
Flue System
From the CBC exhaust, the gas is carried in an 8"
to a 4" dust collector unit with a dust hopper,
feeder, and a valve for recirculating fly ash or
discharging it to waste.
pipe
rotary
POPE, EVANS AND ROBBINS
INCORPORATED
-------
NO.
10
11
12
13
14
15
16
17
18
19
20
LOCATION OF THERMOCOUPLES INFBM/CBC SYSTEM
LOCATION*
1
Inlet air
2
FBM 1-1/2" bed, 9-1/2"
3
FBM 1-1/2" bed, 28-1/2"
4
FBM 1-1/2" bed, 45-1/2"
5
FBM 1-1/2" bed, 63-1/2"
6
FBM 9-1/2" bed, 9-1/2"
7
FBM 9-1/2" bed, 28-1/2".
8
FBM 9-1/2" bed, 45-1/2"
9
FBM 9-1/2" bed, 63-1/2"
FBM 29-1/2" bed, 9-1/2"
FBM 29-1/2" bed, 28-1/2"
FBM 29-1/2" bed, 45-1/2"
FBM 29-1/2" bed, 63-1/2"
CBC 3" bed
CBC 9-1/2" bed
CBC 25-1/2" bed
CBC 42-1/2" bed
CBC stack after cooling coil
H.V.T. below steam drum
H.V.T. above steam drum
POPE. EVANS AND ROBBINS
INCORPORATED
C-2
-------
C-3
21 R.V.T. after convection bank
22 R.V.T. after air preheater
23 R.V.T. after economizer
24 Ash reinjection air
*
Location in FBM described as height above grid, distance
from front of unit.
POPE. EVANS AND ROBBINS
INCORPORATED
-------
D-1
APPENDIX D.
SELECTED CARBON BURNUP RESULTS
TABLE D-1
DATA FROM TEST C-319-1
Bed temperature, T = 19800F
Air Rate, A = 700 1b/hr
Fuel Rate, F = 61 1b/hr
Carbon Content, c = .65 1b carbon/1b fuel
Bed Height, H = 22 inches
Calculate from data and Equations 7 and 8
=
61 (.65) = 39.7 1b/hr
21.3 1b/hr
Carbon Rate, C = F(c)
Inert Rate, I = F-C
=
Using the above data, Equation 6 and coefficients listed in
Table 9, compute the combustion efficiency (CE), carbon
dioxide (C02) and residual oxygen (RO):
CE, % = -13.78 + 0.05193 (1980) + 0.0462 (700)
+ 0.3831 (22) - 0.8737 (39.7) - 0.1905 (21.3)
+ 0.00270 (39.7) (21.3) = 93.3
C02' % = -2.498 + 0.007547 (1980) - 0.01190 (700)
+ 0.1382 (22) + 0.1589 (39.7) + 0.01259 (21.3)
RO, %
-0.0002862 (39.7) (21.3) = 13.5
= 22.91 - 0.007353 (1980) + 0.013 (700)
-0.1390 (22) - 0.1769 (39.7) - 0.01756 (21.3)
+ 0.0003587 (39.7) (21.3) = 7.3
POPE. EVANS AND ROBBINS
INCORPORATED
-------
D-2
TABLE D-2
CBC DATA FOR FBM TEST B-14
Bed Temperature = 20000F
Air Rate = 1350 1b/hr
Fuel Rate = 139 1b/hr
Carbon Content = .48 1b carbon/1b fuel
Bed Height (static) = 14 in.
For the CBC with a plan area of 1.1 ft2 the normalized
data (normalized to unit plan area) are:
Air Rate, A'
Carbon Rate, C'
=
1227
60.6
65.7
1b/hr ft2
1b/hr ft2
1b/hr ft2
=
Inert Rate, I'
=
POPE. EVANS AND ROBBINS
INCORPORATED
-------
APPENDIX E
SUMMARY OF FBC DATA FOR CARBON-.-BURNUP CELL TESTS
POPE. EVANS AND ROBBINS
INCORPORATED
E-l
-------
B-2
TABLE E-1.
SUMMARY OF FBC DATA FOR CARBON-BURNUP CELL TESTS
Test Number
Reactor Data
Air rate, 1b/hr
Bed temperature, of
Starter bed
a. Static depth, in.
b. Sieve size, USS
4. Fly ash input, 1b/hr (dry)
a. Water content, % (a.r.)
b. Carbon content, % dry
5. Coal input, 1b/hr
a. Carbon content, %
6. Total carbon input, 1b/hr
7. Fly ash output, 1b/hr
a. Carbon content, %
8. Carbon burned, 1b/hr
9. Combustion efficiency, %
10. Superficial velocity, fps
11. Cooling probe position
1.
2.
3.
Flue Gas Data
12. C02, %
13. 02, %
14. CO , %
15. S02, ppm
16. NO, ppm
17. Hydrocarbons, ppm
18. Particulate emission
a. To stack, Ib/hr
b. Carbon content, %
Heat Balance Data
19. Losses, kBtu/hr
a. Probe
b. Wall
C. Hood
d. Gas
e. Dust
f. Total
20. Input, kBtu/hr
From 8.
C-302-1* C-302~2* C-302-3* C-303-1*
760 760 820 330
1800 2140 1780 1750
18 18 18 18
. . . . . . . . . . . . . . -16+22. . . . . . . . . . .
73 105 72 41
12.6 10.0 4.3 26.9
54.9 50.5 60.5 50.9
o 0 0 0
40.4 53.5 43.7 20.8
37 62 35 28
11. 2 15.5 17.2 28.3
36.4 44.0 37.8 13.4
90.1 82.2 86.5 64.4
14.6 17.0 15.6 5.7
. . . . . . . . . . . . . hot used............
11. 5
9.1
o
350
560
14.2
6.4
o
800
670
10.8
9.7
o
550
570
9.8
10.8
o
320
470
. . . . . . . . . . . Ne g 1 i g i b Ie. . . . . . . . . . . .
. . . . . . . . . . . . No data..............
o 0 0 0
40 44 27 36
36 60 45 34
390 478 401 182
14 28 13 10
480 6TO 486 262
513 620 533 189
*Test results are suspect.
Presented for record purposes only.
POPE, EVANS AND ROBBINS
INCORPORATED
-------
TABLE E-1.
(Continued)
E-3
Test Number
Reactor Data
Air rate, 1b/hr
Bed temperaturp., of
Starter bed
a. Static depth, in.
b. Sieve size, USS
4. Fly ash input, 1b/hr (dry)
a. Water content, % .(a.r.)
b. Carbon content, % dry
5. Coal input, 1b/hr
a. Carbon content, %
6. Total carbon input, 1b/hr
7. Fly ash output, 1b/hr
a. Carbon content, %
8. Carbon burned, 1b/hr
9. Combustion efficiency, %
10. Superficial velocity, fps
11. Cooling probe position
1.
2.
3.
Flue Gas Data
12. C02, %
13. 02, %
14. CO , %
15. S02, ppm
16. NO, ppm
17. Hydrocarbons, ppm
18. Particulate emission
a. To stack, 1b/hr
b. Carbon content, %
Heat Balance Data
19. Losses, kBtu/hr
a. Probe
b. Wall
c. Hood
d. Gas
e. Dust
f. Total.
20. Input, kBtu/hr
From 8.
*Tests results are suspect.
C-303~2* C-303-3* C-303-4* C-303-5*
330 500 500 500
1980 1750 1860 1970
18 18 18 18
. . . . . . . . . . . . ~ . -16+22. . . . . . . . . . .
46 68 67 71
24.9 26.4 . 25.2 26.0
52.5 57.0 55.5 54.5
o 0 0 0
24.1 38.7 37.0 30.7
30 43 39 41
27.5 32.0 24.4 20.9
15.8 25.0 27.3 30.1
65.6 64.6 73.8 77.8
6.9 9.4 9.9 10.5
. . . . . . . . . . . . . Not used............
11. 5
9.0
o
600
480
12.0
8.5
o
320
410
14.5
6.2
o
700
600
13.2
7.4
o
500
570
. . . . . . . . . . . Negligible. . . . . . . . . . . .
. . . . . . . . . . . . No data..............
o 0 0 0
27 36 43 50
36 36 41 17
193 281 299 322
13 16 15 17
- - - -
269 369 398 406
223 353 385 424
Presented for record purposes only.
POPE. EVANS AND ROBBINS
INCORPORATED
-------
E-4
TABLE E-1.
(Continued)
Test Number
C-306-4
Reactor Data
Air rate, Ib/hr
Bed temperaturp., of
Starter bed
a. Static depth, In.
b. Sieve size, USS
4. Fly ash input, Ib/hr (dry)
a. Water content, % (a.r.)
b. Carbon content, % dry
5. Coal input, Ib/hr
a. Carbon content, %
6. Total carbon input, Ib/hr
7. Fly ash output, Ib/hr
a. Carbon content, %
8. Carbon burned, Ib/hr
9. Combustion efficiency, %
10. Superficial velocity, fps
11. Cooling probe position
1.
2.
3.
Flue Gas Data
12. C02, %
13. 02, %
14. CO, %
15. S02, ppm
16. NO, ppm
17. Hydrocarbons, ppm
18. Particulate emission
a. To stack, Ib/hr
b. Carbon content, %
Heat Balance Data
19. Losses, kBtu/hr
a. Probe
b. Wall
c. Hood
d. Gas
e. Dust
f. 'l'ota1
20. Input, kBtu/hr
From 8.
C-306-1
600
2100
18
C-306-2
860
1750
18
C-306-3
775
2120
18
600
1800
........... -8+16..............
18
106
0.4
39.2
o
41. 5
70
8.5
35.5
85.5
11. 8
134
0.6
41. 7
o
56.0
95
17.5
39.2
70.0
14.5
130
0.5
40.3
o
52.5
82
7.5
46.5
88.7
15.2
105
0.5
40.9
o
43.0
74
16.0
31. 2
72.7
10.5
. . . . . . . . . .. Not
us ed ............
14.3
6.2
o
800
720
11. 0
9.3
o
750
560
14.6
5.8
o
800
770
12.4
8.2
o
750
520
. . . . . . . . . . . N e g 1 i g i b Ie. . . . . . . . . . . .
1. 27
9.0
o
38
72
354
31
495
501
4.15
10.5
o
32
68
406
35
541
553
1. 71
9.3
o
33
79
461
37
610
656
2.97
5.3
o
36
75
294
28
433
440
POPE, EVANS AND ROBBINS
INCORPORATED
-------
TABLE E-1.
(Continued)
E-5
Test Number
C-307-4
Reactor Data
Air rate, Ib/hr
Bed temperaturp., of
Starter bed
a. Static depth, In.
b. Sieve size, USS
4. Fly ash input, Ib/hr (dry)
a. Water content, % (a.r.)
b. Carbon content, % dry
5. Coal input, Ib/hr
a. Carbon content, %
6. Total carbon input, Ib/hr
7. Fly ash output, Ib/hr
a. Carbon content, %
8. Carbon burned, Ib/hr
9. Combustion efficiency, %
10. Superficial velocity, fps
11. Cooling probe position
1.
2.
3.
Flue Gas Data
12. C02, %
13. 02, %
14. CO , %
15. S02, ppm
16. NO, ppm
17. Hydrocarbons, ppm
18. Particulate emission
a. To stack, 1b/hr
b. Carbon content, %
Heat Balance Data
19. Losses, kBtu/hr
a. Probe
b. Wall
c. Hood'
d. Gas
e~ Dust
f. 'i'otal
20. Input, kBtu/hr
From 8.
C-307-1
600
1800
17
C-307-2
330
2100
18
C-307-3
600
2140
. . . . . . . . . . . . -16+22
18
............
18
106
0.2
39.2
o
40.6
81
11. 9
30.9
76.0
9.8
65
0.4
39.5
o
25.7
42
6.0
23.2
90.0
6.6
110
0.1
40.0
o
44.0
70
6.2
39.7
91.0
11. 9
330
1750
73
0.2
40.3
o
29.4
54
19.6
18.7
63.0
5.7
. . . . . . . . . . . . Not
used. . . . . . . . . . . .
13.2
7.5
o
500
410
o
3.0
8.3
o
50
53
293
31
427
436
16.8
3.5
o
600
500
o
0.4
o
48
55
197
19
319
327
16.0
4.8
o
650
550
o
1.5
4.7
o
47
68
361
32
508
560
13.4
6.9
o
700
370
o
4.1
9.4
o
32
58
157
20
267
264
POPE. EVANS AND ROBBINS
INCORPORATED
-------
E-6
TABLE E-1.
(Continued)
C-308-2
Test Number
React.or Data
Air rate, 1b/hr
Bed temperaturp., of
Starter bed
a. Static depth, in.
b. Sieve size, uss
4. Fly ash input, 1b/hr (dry)
a. Water content, % (a.r.)
b. Carbon content, % dry
5. Coal input, 1b/hr
a. Carbon content, %
6. Total carbon input, 1b/hr
7. Fly ash output, Ib/hr
a. Carbon content, %
8. Carbon burned, Ib/hr
9. Combustion efficiency, %
10. Superficial velocity, fps
11. Cooling probe position
l.
2.
3.
Flue Gas Data
12. C02, %
13. 02, %
14. CO, %
15. S02, ppm
16. NO, ppm
17. Hydrocarbons, ppm
18. Particulate emission
a. To stack, Ib/hr
b. Carbon content, %
Heat Balance Data
19. Losses, kBtu/hr
a. Probe
b. Wall
c. Hood
d. Gas
e. Dust
f. 'riotal
C-308-1
C-308-3
600
19001
600
19002
600
19003
18 18 18
............-16+22 ....
87 100 102
000
43.5 42.4 42.2
000
37.8 42.4 43.0
56 66 68
12.0 12.3 12.4
31.0 34.3 34.6
81.5 81.3 80.6
10.2 10.3 10.2
. . . . . . . . Not used........
13.4 14.5 15.0
7.1 6.2 5.5
o 0 0
600 750 800
520 500 520
30 30 30
2.7 2.5 3.1
8.5 8.6 9.4
o 0 0
55 53 53
69 90 92
310 311 311
23 26 27
457 480 483-
437 484 488
20. Input, kBtujhr
From 8.
Notes: 1 Temperature stabilized with slow rise to 1900°F
2 After temperature held at 2100°F for 15 min.
3 After temperature held at 1500°F for 15 min.
POPE, EVANS AND ROBBINS
INCORPORATED
-------
TABLE E-1.
(Continued)
E-7
Test Number
Reactor Data.
Air rate, 1b/l1r
Bed temperaturp., of
Starter bed
a. Static depth, in.
b. Sieve size, USS
4. Fly ash input, 1b/hr (dry)
a. Water content, % (a.r.)
b. Carbon content, % dry
5. Coal input, 1b/hr
a. Carbon content, %
6. Total carbon input, 1b/hr
7. Fly ash output, 1b/hr
a. Carbon content, %
8. Carbon burned, 1b/hr
9. Combustion efficiency, %
10. Superficial velocity, fps
11. Cooling probe position
1.
2.
3.
Flue Gas Data
12. C02, %
13. 02, %
14. CO, %
15. S02, ppm
16. NO, ppm
17. Hydrocarbons, ppm
18. Particulate emission
a. To stack, 1b/hr
b. Carbon content, %
Heat Balance Data
19. Losses, kBtu/hr
a. Probe
b. Wall
c. Hood
d. Gas
e~ Dust
r. 'rotal
20. Input, kBtu/hr
From 8.
C-310...:.1
C-310-2
700
2040
700
2000
18 18
.. .-16+22....
105 125
0.5 0.5
44.0 44.0
o 0
46.2
65
11. 0
39.2
84.6
12.9
OUT
55.0
83
16.0
41.7
75.8
13.2
IN
13.2 13.8
7.1 6.4
o 0
700 720 .
460 420
. . . Negligible. . .
0.88
13.7
1.10
15.2
6 67
55 57
30 32
397 390
28 35
516 581
553 588
POPE, EVANS AND ROBBINS
INCORPORATED
-------
E-8
TABLE E-1.
(Continued)
Test Number
C-311-3
Reactor Data.
Air rate, Ib/hr
Bed temperature, of
Starter bed
a. Static depth, in.
b. Sieve size, USS
4. Fly ash input, Ib/hr (dry)
a. Water content, % (a.r.)
b. Carbon content, % dry
5. Coal input, Ib/hr
a. Carbon content, %
6. Total carbon input, Ib/hr
7. Fly ash output, Ib/hr
a. Carbon content, %
8. Carbon burned, Ib/hr
9. Combustion efficiency, %
10. Superficial velocity, fps
11. Cooling probe position
1.
2.
3.
Flue Gas Data
12. C02, %
13.02, %
14. CO, %
15. S02, ppm
16. NO, ppm
17. Hydrocarbons, ppm
18. particulate emission
a. To stack, Ib/hr
b. Carbon content, %
Heat Balance Data
19. Losses, kBtu/hr
a. Probe
b. Wall
c. Hood
d. Gas
e. Dust
f. 'llotal
20. Input, kBtu/hr
From 8.
C-311-1
700
2000
18
C-311-2
700
1930
18
600
2000
...... .-16+22.... ~....
18
158
1.2
33.0
o
52.0
118
10.2
40.0
77.0
11. 3
OUT
16.0
4.5
o
700
570
o
1. 32
9.1
14
55
32
393
50
544
564
190
2.0
33.0
o
63.0
148
14.2
42.0
66.7
11.1
IN
16.6
3.6
o
800
470
50
2.78
14.0
68
68
32
379
60
607
592
250
1.5
31. 3
o
79.5
205
16.8
45.0
58.0
11.1
IN
18.1
2.1
o.
850
470
o
7.5
15.5
76
68
33
346
87
610
635
POPE. EVANS AND ROBBINS
INCORPORATED
-------
TABLE E-l.
(Continued)
Test Number
Reactor Data.
Air rate, Ib/hr
Bed temperaturp., of
Starter bed
a. Static depth, in.
b. Sieve size, USS
4. Fly ash input, Ib/hr (dry)
a. Water content, % (a.r.)
b. Carbon content, % dry
5. Coal input, Ib/hr
a. Carbon content, %
6. Total carbon input, Ib/hr
7. Fly ash output, Ib/hr
a. Carbon content, %
8. Carbon burned, Ib/hr
9. Combustion efficiency, %
10. Superficial velocity, fps
11. Cooling probe position
Flue Gas Data
l.
2.
3.
12. C02, %
13. 02, %
14. CO, %
15. S02, ppm
16. NO, ppm
17. Hydrocarbons, ppm
18. Particulate emission
a. To stack, Ib/hr
b. Carbon content, %
Heat Balance Data
19. Losses, kBtu/hr
a. Probe
b. Wall
c. Hood
d. Gas
e. Dust
f. 'l;otal
20. Input, kBtu/hr
From 8.
*
Declining temperature
C-311-4
600
1980
18
C-311-5
800
1950
18
C-311-6
800
1800*
....... .-16+22........
18
158
1.5
32.0
o
50.5
120
10.5
38.0
75.8
10.5
OUT
15.9
4.5
o
820
500
Neg1.
1. 77
12.3
16
68
45
334
50
513
536
POPE, EVANS AND ROBBINS
INCORPORATED
232
1.2
32.0
o
74.5
186
13.9
48.2
64.5
13.7
OUT
15.8
4.7
o
Inst.
510
Inst.
5.3
17.4
17
68
65
438
77
665
680
302
1.8
32.0
o
96.0
250
16.9
52.0
56.2
13.2
IN
16.4
3.9
o
Failure
350
Failure
7.1
15.6
67
65
68
406
95
701
733
E-9
-------
E-IO
TABLE E-1.
(Continued)
Test Number
Reactor Data
Air rate, 1b/hr
Bed temperatur~, of
Starter bed
a. Static depth, in.
b. Sieve size, USS
4. Fly ash input, lb/hr (dry)
a. Water content, % (a.r.)
b. Carbon content, % dry
5. Coal input, lb/hr
a. Carbon content, %
6. Total carbon input, 1b/hr
7. Fly ash output, lb/hr
a. Carbon content, %
8. Carbon burned, 1b/hr
9. Combustion efficiency, %
10. Superficial velocity, fps
11. Cooling probe position
1.
2.
3.
Flue Gas Data
12. C02, %
13. 02, %
14. CO, %
15. S02, ppm
16. NO, ppm
17. Hydrocarbons, ppm
18. Particulate emission
a. To stack, 1b/hr
b. Carbon content, %
Heat Balance Data
19. Losses, kBtu/hr
a. Probe
b. Wall
c. Hood
d. Gas
e. Dust
f. rrotal
20. Input, kBtu/hr
From 8.
C-312-1
C-312-2
C-312-3
700 700 700
2000 1990 1980
18 14 10
....... .-16+22........
140 138 156
0.2 0.3 0.2
33.0 33.0 32.8
o 0 0
46.5 45.7 51. 3
97 96 117
7.5 7.7 10.8
39.2 38.3 . 37.4
84.2 83.8 76.0
13.6 13.6 13.5
........ Not used ......
13.8 13.5 13.2
7.0 7.2 7.5
o 0 0
900 900 900
500 480 480
. . . . .. Negligible ......
2.2 2.3 1.5
o 0 0
67 58 47
27 28 29
389 385 383
41 40 49
524 511 508
553 540 527
POPE. EVANS AND ROBBINS
INCORPORATED
-------
TABLE E-1.
(Continued)
E-ll
Test Number
Reactor Data
Air rate, Ib/hr
Bed temperaturp., of
Starter bed
a. Static depth, in.
b. Sieve size, USS
4. Fly ash input, Ib/hr (dry)
a. Water content, % (a.r.)
b. Carbon content, % dry
5. Coal input, Ib/hr
a. Carbon content, %
6. Total carbon input, Ib/hr
7. Fly ash output, Ib/hr
a. Carbon content, %
8. Carbon burned, Ib/hr
9. Combustion efficiency, %
10. Superficial velocity, fps
11. Cooling probe position
1.
2.
3.
Flue Gas Data
12. C02, %
13. 02, %
14. CO, %
15. S02, ppm
16. NO, ppm
17. Hydrocarbons, ppm
18. Particulate emission
a. To stack,lb/hr
b. Carbon content, %
Heat Balance Data
19. Losses, kBtu/hr
a. Probe
b. Wall
c. Hood
d. Gas
e. Dust
f. 'l'otal
20. Input, kBtu/hr
From 8.
C-313-1
C-313-2
C-313-3
600 700 800
1980 1980 1980
18 18 18
....... .-16+22........
63 74 84
o 0 0
58.0 58.0 57.5
o 0 0
36.6 43.2 48.2
29.6 36.5 45.2
11. 5 16.0 19.0
33.4 37.5 41. 5
91. 0 87.2 86.0
11. 5 13.5 15.2
. . . . . . . Not used...... .0..
13.3 12.9 12.8
7.5 8.0 8.1
000
1160 1160 1160
630 610 600
.......Negligible.......
1.4
5.7
1.0
1.7
6.2
o 0 0
75 84 89
25 33 33
328 382 436
12 15 19
440 514 577
471 529 585
POPE. EVANS AND ROBBINS
INCORPORATED
-------
E-12
TABLE E-1.
(Continued)
Test Number
Reactor Data
Air rate, Ib/hr
Bed temperature, of
Starter bed
a. Static depth, in.
b. Sieve size, USS
4. Fly ash input, Ib/hr (dry)
a. Water content, % (a.r.)
b. Carbon content, % dry
5. Coal input, Ib/hr
a. Carbon content, %
6. Total carbon input, Ib/hr
7. Fly ash output, Ibjhr
a. Carbon content, %
8. Carbon burned, Ib/hr
9. Combustion efficiency, %
10. Superficial velocity, fps
11. Cooling probe position
1.
2.
3.
Flue Gas Data
12. C02, %
13. 02, %
14. CO , %
15. S02, ppm
16. NO, ppm
17. Hydrocarbons, ppm
18. Particulate emission
a. To stack, Ib/hr
b. Carbon content, %
Heat Balance Data
19. Losses, kBtu/hr
a. Probe
b. Wall
c. Hood
d. Gas
e. Dust
f. Total
20. Input, kBtu/hr
From 8.
C-314':'1
600
1980
10
C-314-2
700
1960
10
C-314-3
800
1980
10
C-314-4*
700
1980
. . . . . . . . . . . . ~ -16+22. . . . . . . . . . . .
10
70
o
56.0
o
82
o
56.0
o
39.2 45.8
36.6 44.5
16.5 18.6
33.2 37.5
84.8 82.0
11.5 13.3
. . ... . . . . . . . . . . . Not
13.4
7.4
o
1050
560
13.0
7.6
o
1050
560
102
o
56.0
o
57.0
58.3
25.5
43.3
76.0
15.3
53
o
56.0
15
64.0
39.6
29.6
14.6
35.4
89.5
13.4
used. . . . . . . . . .
13.0
7.8
o
1050
560
12.2
8.5
o
1500
530
. . . . . . . . . . . Negligible. . . . . . . . . . . .
0.9
o
84
27
328
15
454
468
0.5
18.6
o
85
29
378
18
510 -
529
0.6
o
88
31
436
24
---s79
611
0.5
o
78
27
394
12
511
499
*He~t balance data adjusted for hydrogen content of coal.
POPE. EVANS AND ROBBINS
INCORPORATED
-------
TABLE E-l.
(Continued)
E-13
Test Number
Reactor Data
Air rate, Ib/hr
Bed temperaturp., of
Starter bed
a. Static depth, in.
b. Sieve size, USS
4. Fly ash input, Ib/hr (dry)
a. Water content, % (a.r.)
b. Carbon content, % dry
5. Coal input, Ib/hr
a. Carbon content, %
6. Total carbon input, Ib/hr
7. Fly ash output, Ib/hr
a. Carbon content, %
8. Carbon burned, Ib/hr
9. Combustion efficiency, %
10. Superficial velocity, fps
11. Cooling probe position
1.
2.
3.
Flue Gas Data
12. C02, %
13. 02, %
14. CO, %
15. S02, ppm
16. NO, ppm
17. Hydrocarbons, ppm
18. Particulate emission
a. To stack, Ib/hr
b. Carbon content, %
Heat Balance Data
19. Losses, kBtu/hr
a. Probe
b. Wall
c. Hood
d. Gas
e. Dust
f. 'rotal
20. Input, kBtu/hr
From 8.
C-315-1
600
2010
18
C-315-2
700
2000
18
C-3l5-3
8QO
1980
....... .-16+22........
18
81
o
57.5
o
46.3
42.4
18.2
38.6
83.3
11. 7
IN
15.4
5.5
o
1250
560
86
o
57.0
o
48.8
44.8
17.0
41. 2
84.5
13.5
IN
15.0
5.8
o
1250
560
92
o
56.0
o
51. 8
47.8
16.0
44.2
85.2
15.3
IN
14.6
6.0
o
1250
540
.......Negligible.......
0.7
5.3
31
100
22
337
18
508
544
0.7
7.4
31
113
24
389
19
576
581
0.8
8.2
31
115
26
437
20
629
623
POPE, EVANS AND ROBBINS
INCORPORATED
-------
E-14
TABLE E-1.
(Continued)
Test Number
Reactor Data
Air rate, Ib/hr
Bed temperaturp., of
Starter bed
a. Static depth, in.
b. Sieve size, USS
4. Fly ash input, Ib/hr (dry)
a. Water content, % (a.r.)
b. Carbon content, % dry
5. Coal input, Ib/hr
a. Carbon content, %
6. Total carbon input, Ib/hr
7. Fly ash output, Ib/hr
a. Carbon content, %
8. Carbon burned, Ib/hr
9. Combustion efficiency, %
10. Superficial velocity, fps
11. Cooling probe position
1.
2.
3.
Flue Gas Data
12. CO 2, %
13. 02, %
14. CO, %
15. S02, ppm
16. NO, ppm
17. Hydrocarbons, ppm
18. Particulate emission
a. To stack, Ib/hr
b. Carbon content, %
Heat Balance Data
19. Losses, kBtu/hr
a. Probe
b. Wall
c. Hood
d. Gas
e~ Dust
f. '!'otal
20. Input, kBtu/hr
From 8.
C-316-1
C-316-2
C-316-3
600
2000
700
1990
800
1980
10 . 10 10
. . . . . . .. -16+22.. . . .. ..
86 94 109
000
58.0 58.0 58.0
000
49.6
49.1
26.3
36.9
74.5
11. 6
IN
54.4
52.4
24.4
41.6
76.5
13.5
IN
63.5
63.3
28.1
45.7
72.0
15.3
IN
14.8 14.4 13.6
5.8 6.0 6.4
o 00
1250 1300 1250
520 510 520
.......Negligible.......
0.9
4.0
0.8
5.4
0.9
6.2
31 31 31
92 99 100
26 28 29
334 386 438
21 22 26
504 576 624
520 587 644
POPE, EVANS AND ROBBINS
INCORPORATED
-------
TABLE E-1.
(Continued)
Test Number
Reactor Data
1.
2.
3.
Air rate, Ib/hr
Bed temperaturp., of
Starter bed
a. Static depth, in.
b. Sieve size, USS
4. Fly ash input, Ib/hr (dry)
a. Water content, % (a.r.)
b. Carbon content, % dry
5. Coal input,lb/hr
a. Carbon content, %
6. Total carbon input, Ib/hr
7. Fly ash output, Ib/hr
a. Carbon content, %
8. Carbon burned, Ib/hr
9. Combustion efficiency, %
10. Superficial velocity, fps
11. Cooling probe position
F'lue Gas Data
12. C02,: %
13. 02, %
14. CO, %
15. S02, ppm
16. NO, ppm
17. Hydrocarbons, ppm
18. Particulate emission
a. To stack, Ib/hr
b. Carbon content, %
Heat Balance Data
19. Losses, kBtu/hr
a. Probe
b. Wall
c. Hood
d . Ga s .
e. Dust
f. 'l'otal
20. Input, kBtu/hr
From 8.
C-317-1
700
1980
22
C-317-2
700
1980
. . . . -8+16. . . .
18
95
o
54.4
o
51.7
51.0
18.0
42.5
82.2
13.5
94
o
54.4
o
51. 4
51. 3
18.8
41.7
81. 5
13.5
C-318-1
700
1900
E-15
C-318-2
700
1950
22 18
... .-16+22...
104 97
o 0
58.6 58.2
o 0
61.0
60.8
28.4
44.0
72.5
13.0
56.0
52.0
21. 8
44.6
79.8
13.3
. . . . . . . . . . . N at
us ed. . . . . . . . . . . . . .
15.5
4.8
o
1200
570
15
0.68
7.4
o
129
19
385
21
554
599
15.3
5.3
o
1275 .
570
15
0.60
6.1
o
123
20
384
21
548
588
15.8
5.0
o
1050
550
20
0.72
9.7
o
172
22
365
24
583
620
16.0
4.5
o
1050
550
20
0.75
5.7
o
150
27
380
21
578
629
POPE. EVANS AND ROBBINS
INCORPORATED
-------
E-16
TABLE E-l.
(Continued)
Test Number
Reactor Data
Air rate, Ib/hr
Bed temperaturp., of
Starter bed
a. Static depth, in.
b. Sieve size, USS
4. Fly ash input, Ib/hr (dry)
a. Water content, % (a.r.)
b. Carbon content, % dry
5. Coal input, Ib/hr .
a. Carbon content, %
6. Total carbon input, Ib/hr
7. Fly ash output, Ib/hr
a. Carbon content, %
8. Carbon burned, Ib/hr
9. Combustion efficiency, %
10. Superficial velocity, fps
11. Cooling probe position
1.
2.
3.
Flue Gas Data
12. C02, %
13. 02, %
14. CO, %
15. S02, ppm
16. NO, ppm
17. Hydrocarbons, ppm
18. Particulate emission
a. To stack, Ib/hr
b. Carbon content, %
Heat Balance Data
19. Losses, kBtu/hr
a. Probe
b. Wall
c. Hood
d. Gas
e. Dust
f. 'l'otal
20. Input, kBtu/hr
From 8.
C-319-1
700
1980
C-319-2
700
2000
22 18
.... -8+16...
61 59
o 0
65.0 65.0
o 0
39.8
26.2
17.0
35.3
89.0
13.5
. . . Not
11. 8
8.2
o
1150
520
12
0.60
6.5
38.4
24.3
15.0
34.8
90.6
13.5
used. . . .
11. 8
8.7
o
1100
575
12
0.61
5.5
o 0
47 50
13 15
381 386
11 10
452 461
498 491
POPE, EVANS AND ROBBINS
INCORPORATED
-------
TABLE E-l.
(Continued)
/
E-17
Test Number
Reactor Data
Air rate, 1b/hr
Bed temperature, of
Starter bed
a. Static depth, in.
b. Sieve size, USS
4. Fly ash input, 1b/hr (dry)
a. Water content, % (a.r.)
b. Carbon content, % dry
5. Coal input, 1b/hr
a. Carbon content, %
6. Total carbon input, 1b/hr
7. Fly ash output, 1b/hr
a. Carbon content, %
8. Carbon burned, 1b/hr
9. Combustion efficiency, %
10. Superficial velocity, fps
11. Cooling probe position
l.
2.
3.
Flue Gas Data
12. C02, %
13. 02, %
14. CO , %
15. S02, ppm
16. NO, ppm
17. Hydrocarbons, ppm
18. Particulate emission
a. To stack, 1b/hr
b. Carbon content, %
Heat Balance Data
19. Losses, kBtu/hr
a. Probe
b. Wall
c. Hood
d. Gas
e. Dust
f. Total
20. Input, kBtu/hr
From 8.
C-319-3~ C~319-4* C-319-5*
700
1980
18
700
2000
18
700
1980
18
. . . . . . . . . - 8+ 16. . . . . . . .
o
o
48
o
65.0
6.8
66.0
35.6
18.8
15.0
32.8
92.0
13.5
27
o
65.0
23.2
66.0
32.4
15.3
15.2
30.1
92.9
13.5
42.0
66.0
27.7
7.6
17.0
26.4
95.5
13.5
.........Not used.......
10.9
9.5
o
1300
490
15
.47
7.0
\
o
49
13
386
8
456
10.2
9.8
o
1600
440
22
.19
8.3
o
51
16
403
6
476 ' '
489
10.0
9.8
o
2000
420
35
.17
o
51
16
412
3
482
489
481
*Heat balance data adjusted for hydrogen content of coal.
POPE. EVANS AND ROBBINS
INCORPORATED
-------
E-18
TABLE E-1.
(Continued)
Test Number
Reactor Data
Air rate, 1b/hr
Bed temperature, of
Starter bed
a. Static depth, in.
b. Sieve size, USS
4. Fly ash input, 1b/hr (dry)*
a. Carbon content, %
5. Carbon input, 1b/hr
6. Fly ash output, 1b/hr
a. Carbon content, %
b. Sulfur content, %
c. Carbon output, 1b/hr
7. Carbon burned, 1b/hr
8. Combustion efficiency, %
9. Superficial velocity, fps
10. Cooling probe position
1.
2.
3.
11. Fuel heat input, kBtu/hr
12. Input sulfur content, %
13. Sulfur input, 1b/hr
14. Sulfur output, 1b/hr
Flue Gas Data
15. C02' %
16. 02' %
17. CO, %
18. S02, ppm
19. HC, ppm
20. particulate emission
a. To stack, 1b/hr
b. Carbon content, %
Heat Balance, kBtu/hr
21. Probe loss
22. Wall loss
23. Hood loss
24. Flue gas loss
25. Ash loss
26. Total losses
C-320-1
C-320-2**
C-320-3***
620 620 620
1980 1970 1970
20 20 20
. . . . . . . . . . . -8xO. . . . . . . . . . . . .
167 174 208
27.9 27.5 19.2
46.5 47.9 40.0
63 159 215
9.46 17.0 3.59
2~49 2.02 2.66
6.0 26.9 7.7
40.5 21.0 32.3
87.1 43.8 80.7
13.1 12.7 12.9
OUT 1 tube OUT
lowered
570 296 455
2.42 2.42 2.92
4.0 4.2 6.1
1.6 3.2 5.7
15.0 20.0 16.0
4.3 0.5 0.8
o 0 0
350 2150 4400
8 7 200
.......... No data ...........
4.4 5.3 3.6
62
5
91
342
26
526
177
13
31
330
66
617
92
18
68
366
89
6IT
See footnotes next pag~.
POPE, EVANS 'AND ROBBINS
INCORPORATED
-------
E-19
*
Feed material C-320-1 and C-320-2: Fly ash containing Ca804
from Run FBM B-16. C-320-3: 8ynthetic fly ash: gypsum 10.6%,
ash 26.4%, CaC03 34%, coal 29%.
Also at 02= 1.0,802=550; 02=0.8, 802=750; 02=0.4, 802=2400;
02=0.1, 802=4100.
**
*** Also at 02=1.7, 802=4300; 02=1.9, 802=3800; 02=2.0, 802=3400.
POPE. EVANS AND ROBBINS
INCORPORATED
-------
F-l
APPENDIX F.
STATISTICAL ANALYSIS AND MODELLING OF
FBC FLY. ASH CARBON-BURNUP DATA
By Arthur Hoerl, University of Delaware
PART 1*
INTRODUCTION
A series of definitive sequential experiments were carried
out to generate response data at near-optimum conditions
with the exception of several exploratory runs used to
evaluate various alternatives such as coil cooling and bed
particle size distribution.
Table F-l includes .asummary of the operating conditions
observed during the series of runs and the measured response
levels.
For the purposes of an experimental test program, most of
the variables included in Table F-l are well defined. However,
in the case of bed temperature it is, in effect, a pseudo-
independent variable. Its set point and control can be
achieved by varying fly ash (at specified percent carbon
levels) and air rates. In addition, maximum temperature
limitations (~2000°F) can be maintained at higher carbon feed
rates through the use of cooling coils. A theoretical heat
balance relationship between carbon and air rates related to
temperature level was developed by PER and summarized in their
Progress Report Nos. 2, 3, and 4.
REGRESSION MODELLING
The independent variables which were considered in various
alternative regression modes are given in Table F-l. The
principal response, CE-combustion efficiency, served as the
primary criterillnfor evaluating the alternative models.
The other response variables, included in Table F-l, were
also fitted to the models but given secondary priority for
selecting the final model form. Empirical regression models
of the type
y = bo + bl xl + --- + bk xk
were successively fitted to the data and evaluated through
their corresponding F-ratios. Here the F-ratio is a statis-
tical measure of the variance in the response data which is
explained by the model divided by the residue (unexplained)
variance. If there were no relationship between the inde-
pendent variables and the response, the ratio would tend to
take on a value of one. In general, the higher the ratio,
the more reliable the model is in explaining the trends in
the data.
*
Editor's Note: The report is reproduced here as submitted by
the authC'.r July 1970, with minor editorial corrections.
POPE, EVANS AND ROBBINS
IrJCORPORJlTFn
-------
F-2
As previously d~scussed, bed temperature is, in effect, a
pseudo-independent variable. Ord~nar~ly, ~t would not be
~ncluded ~n a model ~n combination with air rate, feed rate
and percent carbon, since its level is primarily a function
of these variables. However, with the implementation of
coil cooling, alternative combinations of the causative vari-
ables, in general, can be specified for a specified tempera-
ture level. Without inclusion of the temperature variable,
no satisfactory model could be developed. This is a direct
consequence of the fact that no satisfactory way could be
developed to quantify the alternate variable-coil cooling.
In structuring the empirical regression model, two alternate
groupings of the fly ash variables were considered. These
included:
GROUP I - FLY ASH FEED RATE
PERCENT CARBON
GROUP 2 - CARBON FEED RATE
INERT FEED RATE
The basic nature of the alternate groupings tends to take on
different interpretations when they are expanded into cross-
product and required terms.
Based on exploratory regression analysis, it was found that
the second grouping resulted in higher F-ratios than the first
group. In particular, it was found that a cross-product term
(interaction) between the two variables in group 2 was highly
significant and added appreciably to the reliability of the
regression model. Its counterpart through the group I combina-
tion is not directly definable. Because of the improved re-
liabili ty with the group 2 variables, only those corres;pond-
ing results will be reported.
Initially, all of the experimental data, 47 test results,
were included in the regression analysis. However, as antici-
pated, it was found that the data collected on Runs C-302 and
C-303 significantly reduced the reliability of the model.
The first
two runs were carried out with high moisture fly ash. This
moisture seems to have adversely influenced the response data
and trends.
An additional point, C-314-4, was carried out, as planned,
w!th auxiliary coal feed. There was
no simple way of including this run point, and it was there-
fore excluded from the analysis. The remaining 38 points
were used to develop the final model form.
POPE. EVANS AND ROBBINS
INCORPORATED
-------
F-3
Various alternative forms, for representing the combustion
efficiency response, were evaluated. These included
CE (as reported)
LOG(l - CE)
a
LOG (CE)
The purpose here was to develop a response form consistent
with theoretical considerations. That is, a response model
which could not predict above the theoretical limit of 100%
combustion efficiency. Therefore, the alternative response
form
LOG(l - CE)
a
[with a = 95-100% corresponding to the theoretical limit of
operation] would be a more satisfactory response. A summary
of the respective F-ratios for CE, LOG (CE), and various
values of a indicate that the form (CE) results in the best
compromise~ This is a consequc~ce of the fact that as ~
gets relatively large LOG(l - --) is inversely proportional
to (CE) a
FORM F-RATIO
CE 74.5
LOG(l - CE) 46.2
95
LOG(l CE 69.2
- 101)
LOG(l CE 72.1
- 110)
CE 74.2
LOG(l - 120)
CE 75.1
LOG(l - 145)
LOG (CE)
73.8
POPE. EVANS AND ROBBINS
INCORPORATED
-------
F-4
As reflected in the above table, relatively large a values
of 120% + are required to attain the reliability level of
the CE response. Therefore, this form of the response,
LOG(l - CE), still does not satisfy the theoretical limit
95-l00%,aand its form is not felt to be justified.
From the standpoint of interpretation, .the form CE is not
too unsatisfactory, since over the range of data unrealistic
response values are not defined.
The ultimate form of the model, which best related all of
the data, was found to be
RESP = bo + bl (TEMP) + b2(AIR RATE)
+ b3(BED DEPTH) + D4 (CARB. RATE)
+ b 5 (INERT RATE) + b6 (CARB. R) (INERT R)
Over the range of the observed data, curvature terms were
found to be insignificant.
A summary of the reliabilities of the models expressed in
terms of their respective F-ratios and R2 (the explained
variation divided by the total observed variation) is given
in the following table:
RESPONSE F-RATIO R2
CE 74.5 .935
C02 23.0 .816
02 26.3 .836
S02 9.2 .642
NO 6.0 .535
The computed regression coefficients for the respective
models are given in Table F-2. Observed and calculated
values of combustion efficiency are shown in Table F-3.
POPE, EVANS AND ROBBINS
INCORPORATED
-------
F-5
TEMPERATURE ,RESPONSE
tero;perature isa xunction of
As derived by PER, ,
temperature level
is extremely sensitive to air and carbon rates (at a par~,
ticular combustion efficiency rate). In addition, based'
on an analysis of the test data, it is also quite sensi~
tive to the percent carbon in the fly ash and bed condi-'
tioning. ' This latter effect was observed in runs prior to
C-308. For this reason, Run C-308 '
was carried out under varying conditionings involving
initial bed temperatures prior to the run. In particular,
it was found that carbon rate could be increased 'from 38
to 42 lb/hr (all other variables at the same level) through
nominal bed conditioning.
As previously discussed, bed
other independent variables.
An attempt was ~ade to develop a temperature model in terms
of the observed operating conditions (with those runs in
which there was no coil cooling). However, no satisfactory
model could be developed. The inherent sensitivity of bed
temperature to measured and unmeasured variables (such as
conditioning discussed above) is best exemplified by com-
paring ,several of ,the runs:
RUN NO.
, TEMP
AIR
RATE
CARBON
RATE
C-306-1
-4
C-307-l
-3
2100
1800
1800
2140
600
600
600
600
41
43
41
44
(CE)
(86)
(73)
(76 )
( 91)
During runs C-306 and C-307, a 300°F temperature range was
observed at essentially the same operating conditions.
The above data also illustrates the separate effect of tem-
perature on combustion efficiency. The observed large dif-
ferences in efficiency are totally caused by the temperature
differenti~l since all other variables are at eisentially
the same level (at a temperature of 2l00oF"a slight inc~ease
in carbon rate, with coil cooling, would ha~e ~ negli~ibre
effect on the efficiency.
POPE, EVANS AND'R0Bf;3INS
INCORPORATED
-------
F-6
RECOMMENDATIONS AND CONCLUSIONS
Based on an analysis of the data, it is apparent that a
temperature control loop is required for final modelling
and optimization. It is recommended that a fly-ash feed
rate controller, adapted to temperature set point (at the
maximum allowable level ~2000°F), be installed with the
CBC. This will be required to achieve maximum performance.
With a built-in controller and operations data over
rate range and carbon content in fly ash, a refined
can be developed to attain the fine balance between
put, efficiency, and oxygen level.
an air
model
through-
Optimization of the present model is not justified at the
present time due to the difficulty of handling the tempera-
ture variable. As previously discussed, given the operating
conditions of the CBC, the model will predict (over the
range of the data) the combustion efficiency and oxygen
level quite well. However, there does not seem to be any
ready method for predicting bed temperature at a specified
air and fly-ash rate at a particular carbon level (% carbon
in the fly ash). This difficulty is a consequence of the
extreme sensitivity of temperature level to operating con-
ditions.
In general, most of the directional effects of the operating
variables are consistent. For example, a deeper bed
is appreciably more desirable than a shallow bed. This re-
sults in a higher combustion efficiency and lower oxygen
leve 1 .
Coil cooling increases carbon throughput but because of de-
creased combustion efficiency, results in a negligible in-
crease in rate of carbon burned. There is a decrease in
oxygen level, but it is not sufficient to justify the de-
creased combustion efficiency. Therefore, it does not seem
advisable to utilize coil cooling.
Two levels of average particle size were included in the
analysis. These results indicate a slightly better combustion
efficiency with the -16 +22 particle size.
POPE. EVANS AND ROBBINS
INCORPORATED
-------
TABLE F-l. EXPERIMENTAL TEST CONDITIONS AND RESULTS '"Ij
I
-.J
COOL-
RUN AIR BED FEED CARBON INERT PART ING
NUMBER TEMP RATE DEPTH RATE % C RATE RATE SIZE COIL CE C02 02 S02 NO
C-302-1 1800 760 18 73 55 40 33 + 90 11.5 9.1 350 560
-2 2140 760 18 105 51 53 52 + 83 14.2 6.4 800 670
-3 1780 820 18 72 61 44 28 + 86 10.8 9.7 550 570
303-1 1750' 330 18 41 51 21 20 + 65 9.8 10.8 320 470
-2 1980 330 18 46 53 24 22 + 66 11.5 9.0 600 480
-3 1750 500 18 68 57 39 29 + 65 12.0 8.5 320 410
~ -4 1860 500 18 67 55 37 30 + 74 13.2 7.4 500 570
-5 1970 500 18 71 55 39 32 + 78 14.5 6.2 700 600
306-1 2100 600 18 106 39 41 65 86 14.3 6.2 800 720
111
~ ~ -2 1750 860 18 134 42 56 78 70 11. 0 9.3 750 560
-3 2120 775 18 130 40 53 77 89 14.6 5.8 800 770
~ (f) -4 1800 600 18 105 41 43 62 73 12.4 8.2 750 520
o »
:II z 307-1 1800 600 17 106 39 41 65 + 76 13.2 7.5 500 410
~ 0
~ ::u -2 2100 330 18 65 39 26 39 + 90 16.8 3.5 600 500
~ -3 2140 600 18 110 40 44 66 + 91 16.0 4.8 650 550
-4 1750 330 18 73 40 29 44 + 63 13.4 6.9 700 370
~ 308-1 1900 600 18 87 43 38 49 + 81.5 13.4 7.1 600 520
-2 1900 600 18 100 42 42 58 + 81.3 14.5 6.2 750 500
-3 1900 600 18 101 42 43 59 + 80.6 15.0 5.5 800 520
310-1 2040 700 18 105 44 46 59 + 84.6 13.2 7.1 700 460
-2 2000 700 18 125 44 55 70 + IN 75.8 13.8 6.4 720 420
BED PARTICLE SIZE CODE
+ -16 +22 (fine)
-8 +16 (coarse)
-------
TABLE F-l. (Continued)
COOL-
RUN AIR BED FEED CARBON I NE RT PART ING
NUMBER TEMP RATE DEPTH RATE % C RATE RATE SIZE COIL CE C02 02 S02 NO
C-311-1 2000 700 18 158 33 52 106 + 77.0 16.0 4.5 700 570
-2 1930 700 18 190 33 63 127 + IN 66.7 16.6 3.6 800 470
-3 2000 600 18 250 31 79 171 + IN' 58.0 18.1 2.1 850 470
-4 1980 600 18 158 32 51 107 + 75.8 15.9 4.5 820 500
-5 1950 800 18 232 32 75 157 + 64.5 15.8 4.7 (820) 510
-6 1800 800 18 302 32 96 206 + IN 56.2 16.4 3.9 (820) 350
~ 312-1 2000 700 18 140 33 47 93 + 84.2 13.8 7.0 900 500
-2 1990 700 14 138 33 46 92 + 83.8 13.5 7.2 900 480
-3 1980 700 10 156 33 51 105 + 76.0 13.2 7.5 900 480
~~ 313-1 1980 600 18 63 58 37 26 + 91.0 13.3 7.5 1160 630
-2 1980 700 18 74 58 43 31 + 87.2 12.9 8.0 1160 610
~(j) -3 1980 800 18 84 57 48 36 + 86.0 12.8 8.1 1160 600
o »
J] z 314-1 1980 600 10 70 56 39 31 + 84.8 13.4 7.4 1050 560
~ 0 -2 1960 700 10 82 56 46 36 + 82.0 13.0 7.6 1050 560
~; -3 1980 800 lQ 102 56 57 45 + 76.0 13.0 7.8 1050 560
315-1 2010 600 18 81 57 46 35 + IN 83.3 15.4 5.5 1250 560
m -2 2000 700 18 86 57 49 37 + IN 84.5 15.0 5.8 1250 560
-3 1980 800 18 92 56 52 40 + IN 85.2 14.6 6.0 1250 540
316-1 2000 600 10 86 58 50 36 + IN 74.5 14.8 5.8 1250 520
-2 1990 700 10 94 58 58* 36* + IN 76.5 14.3 6.0 1300 510
-3 1980 800 10 109 58 63 46 + IN 72.0 13.8 6.4 1250 520
317-1 1980 700 22 95 54 52 43 82.2 15.5 4.8 1200 570
-2 1980 700 18 94 54 51 43 81.5 15.3 5.3 1275 570
318-1 1900 700 22 104 59 61 43 + 72.5 15.8 5.0 1050 550
-2 1950 700 18 97 58 56 41 + 79.8 16.0 4.5 1050 550
* These data in Use data from Appendix E.
Editor's Note: error. IT!
I
co
-------
TABLE F-2. REGRESSION COEFFICIENTS
>-rj
I
I.D
TERM CE C02 02 S02 NO
CONSTANT -13.78 -2.498 22.91 -44.12 -464.3
TEMP .05193 .007547 -.007353 .4095 .4257
AIR RATE .04620 -.01190 .01300 .1199 .3825
BED DEPTH .3831 .1382 -.1390 -8.986 4.468
~ CARB. R -.8737 .1589 -.1769 12.60 -2.717
8~ INERT R -.1905 .01259 -.01756 -7.608 -.9104
(CR) (IR) .002700 -.0002862 .0003587 .03046 .007263
~(J)
0):>
]I z
~~
m
-------
TABLE F-3. OBSERVED AND CALCULATED COMBUSTION EFFICIENCY
RUN OBS CALC RES RUN OBS CALC RES
C-306~1 86 88.9 -2.9 C-312-1 84.2 82.3 1.9
-2 70 71.7 -1.7 -2 83.8 81.0 2.8
-3 89 89.1 -.1 -3 76.0 75.1 .9
-4 73 72.1 .9 313-1 91.0 89.0 2.0
307-1 76 73.3 2.7 -2 87.2 88.4 -1.2
-2 90 90.0 0 -3 86.0 88.8 -2.8
-3 91 88.8 2.2
-4 63 69.0 -6.0 314-1 84.8 83.9 . 9
~ -2 82.0 81.6 .4
308-1 81.5 82.0 -.5 -3 .76.0 78.4 -2.4
-2 81.3 78.3 3.0
~~ -3 80.6 77.5 3.1 315~1 83.3 82.7 .6
-2 84.5 84.4 .1
~(J) 310-1 84.6 87.3 -2.7 -3 85.2 85.5 -.3
0» -2 75.8 78.3 -2.5
JJ Z 316-1 74.5 76.0 -1.5
~ 0
~~ 311-1 77.0 78.6 -1.6 -2 76.5 73.8* 2.7*
-2 66.7 68.0 -1.3 -3 72.0 73.8 -1.8
-3 58.0 59.5 -1.5
m -4 75.8 73.4 2.4 317-1 82.2 82.2 0
-5 64.5 67.7 -3.2 -2 81.5 81.4 .1
-6 56.2 53.8 2.4
318-1 72.5 71.2 1.3
-2 79.8 76.2 3.6
*
Editor's Note: These results in error; read 76.5 for CALC and 0 for RES.
t-rj
I
t-'
a
-------
F-ll
PART 11.1 REGRESSION FOR PARTICULATE EMISSION
A regression model was developed using particulate emission
data furnished by Pope, Evans and RQbbins. The same run data,
previously reported Runs 306 through 319, were used for this
analysis.
Several alternative forms for the response and different com-
binations of the independent variables were tried. These
alternatives led 20 the following results in terms of an
F-ratio criteria.
MODEL RESPONSE F-RATIO
Type 1 P.E. 28.0
Type 1 Log (P.E.) 18.3
Type 2 P.E. 20.5
Type 3 P.E. 16.5
TYPE 1
bo + bl
+ b2
+ b3
+ b4
+ b5
+ b6
(Bed Temperature - OF)
(Air Rate - Ibs/hr
(Static Bed Depth, incites)
(Carbon Feed Rate, Ibs of C/hr
(Inert Feed Rate, Ibs of I/hr
(C X I)
TYPE 2
Same as Type 1 but excluding bl' b2' and b3.
1. Editor's Note: The report is reproduced here as submitted
by the author in January 1972, with minor editorial cor-
rections.
2. The F-ratio is the explained variation or variance (as
explained by the associated assumed model) divided by the
unexplained variation.
POPE, EVANS AND ROBBINS
INCORPORATED
-------
F-12
TYPE 3
Same as Type 1 but with Log (Carbon F.R.) and Log (Inert
F.R.) and excluding b6.
On the basis of these calculations it may be concluded that
both carbon and inert feed rates and their interaction are
extremely important variables as they affect particulate.
emissions (as well as the other responses that were prevlously
reported.
The regression coefficients for Type 1 model with the P.E.
response are:
b
o
= 15.57
b = .006639
1
b = .001571
2
b - .02362
3
b - .03528
4 -
b = .005357
5
b = .0003935
6
The residual standard deviation for this model is s = 0.75.
Table F-4 includes ~ summary of the reported P.E. determina-
tions and the corresponding model predictions.
The model is quite adequate but could be improved without
too much additional work if it were required. This would
necessitate a more careful screening of the reported P.E.
data to screen out gross errors. It was reported that the
P.E. results are not highly reliable and this is reflected
by the relatively large residual variance s = .75. There-
fore, additional refinement of the model is not sufficient
by itself, but rather a combination of model development and
data screening would be required to improve the overall
reliability. This would only be justified if an improved
model were needed. .
POPE, EVANS AND ROBBINS
INCORPORATED
-------
F-13
TABLE F-4.
OBSERVED AND CALCULATED PARTICULATE EMISSIONS
RUN NO. REPORTED CALCULATED RUN NO. REPORTED CALCULATED
306-1 1. 27 1. 06 312-1 2.2 2.18
306-2 4.15 3.19 312-2 2.3 2.13
306-3 1. 71 .85 312-3 1.5 2.43
306-4 2.97 2.97 313-1 1.4 1.12
307-1 3.0 3.05 313-2 1.0 .93
307-2 . 4 1. 23 313-3 1.7 .77
307-3 1.5 .79 314-1 .9 .99
307-4 4.1 3.57 314-2 .5 .92
308-1 2.7 2.09 314-3 .6 .65
308-2 2.5 2.23 315-1 . 7 .91
308-3 3.1 2.22 315-2 . 7 .80
310-1 .88 1.11 315-3 .8 .79
310-2 1.1 1. 57 316-1 .9 .72
311-1 1. 32 2.52 316-2 . 8 .66
311-2 2.78 3.69 316-3 . 9 .57
311-3 7.50 5.22 317-1 .68 1.12
311-4 1. 77 2.83 317-2 .6 1. 05
311-5 5.3 4.62 318-1 .72 1. 49
311-6 7.1 8.29 318"""2 .75 1.10
319-1 .60 .88
319-2 .61 .70
POPE, EVANS AND ROBBINS
INCORPORATED
-------
G-1
APPENDIX G.
GAS TEMPERATURE GRADIENT IN A HOT FLUIDIZED BED
The problem of measuring the gas temperature gradient at the
base of a hot fluidized bed is complicated by radiation heat
gain from bed particles in the vicinity of the thermocouple.
The problem is analagous to that of measuring temperature in
very hot gas streams in a non-adiabatic duct, except here
heat is lost by radiation from the thermocouple.
In the
latter instance, the usual practice is to reduce the radiation
loss by shielding and increase the convective heat gain by
high gas velocities.
In the fluidized bed operating at
above 1600°F, it is not possible to shield out the radiation
field effectively without disturbing the factors which
influence the local temperature.
An estimate was made of the gas temperature gradient as it
enters the bed, based on convective exchange of heat between
the gas and the bed particles.
It was assumed in the analysis
that the gas moved in plug flow through a back mix reactor
of particles and that the particle temperature remained
constant at the total bed temperature.
This latter assumption
seems valid in view of the fact that no temperature gradients
are found when burning coal.
The following function was determined for the variation of
gas temperature with distance:
1) ~ 'b - (Tb - ~ ) e=-kX )
k ~ h ~t
C:J C.p.L
(G-l)
POPE, EVANS AND ROBBINS
INCORPORATED
-------
G-2
where Tg and Tb are the gas and bed temperatures respectively,
h is the partic1e-to-gas heat transfer coefficient, At/L the
particle surface per unit bed depth, G the mass gas flow,
C the gas heat capacity, and x the distance above the grid,
p
T is the initial gas temperature.
o
The function is plotted in Figure G-1 for a 16-mesh bed
particle size at a 1600°F bed temperature.
The particle
surface was estimated from that of close packed spheres with
a 50% bed expansion.
The mass flow was taken at 780 1b/hr
and the partic1e-to-gas transfer coefficient was estimated from
the relation*:
hd (bG)V~(c..~).l)o.b7
--=001' - -
k . 1e)1 k
(G-2)
*
Frantz, J. F., Chern. Eng. prog. 57, 35(1961)
POPE. EVANS AND ROBBINS
INCORPORATED
-------
2000
u..
0
- Tb
LIJ
~ ~ 1500
....
~
g~ ~ 1000
~; en
-------
G-4
The temperature gradient or the .rate of rise with
di~tance at x = 0 is: -
-------
0.4 2000
(/)
,LIJ h= 15.0 BTU/HR FT2 oR
z
0 Dp =.0039 FT
z -Tb -
G= 780LB/ HR FT2
~ .. 0.3 A~ = 1500 FT2/FT --- 1500
c ,---
- TEMPERATURE........ ~
a::
~~ C> Tb = 16000 F /
"., -4
LLJ T = 70°F ./ fTI
~U> > go ./ 3:
g~ g 0.2 .,/ 1000 -0
/' rTI
~; ~ ::0
/ :t>
LLJ /' -i
U /' c
Z ::0
~ /' CT1
..
~ 0.1 /' 500
/' ,0
- ."
C /'
I ;/
x ;/
~
o 0.1
0.2 0.3
0.4 0.5
0.6 0.7 0:8 0.9
1.0
1.1
1.2
1.3
TIME- MILLISECONDS
C')
I
VI
FIGURE G-2 TIME VARrATION OF GAS TEMPERATURE AND
DISTANCE TRAVELED THROUGH FLUIDIZED-BED
-------
H-l
APPENDIX H.
A THEORETICAL ANALYSIS OF ,SULFUR RETENTION
PHENOMENA (THE PSEUDOSULFATE THEORY)
A reexamination of previous S02 control results was made.
A theoretical explanation of the performance of a fluidized-
bed boiler in once-through and regenerative modes of lime
desulfurization is described.
The factors that limit the limestone removal of S02 from
combustion gases in a coal-fired boiler may be understood
in the context of the following considerations. Assume
that the following equilibria and rates are involved:
kl
CaC03 -+
+-
CaO
+ C02
Kl
(H-l)
kl
CaO + S02 + 1 02
2"
k2
-+ CaS04 ,
+-
k2
K2
(H- 2)
CaC03 + S02 + ! 02~ CaS04 + C02,K3 = KIK2
2
The K's are the equilibrium constants, and the k's are the
rate constants.
(H-3)
The concentration of S02 in equilibrium with solid CaC03
and CaS04, in the presence of known concentrations of C02
and 02' may be calculated from the published equilibrium
constant for reaction (H-3). Thus:
[S02] in ppm = [C02]
eq. K 3 [0 2 ] ~
where [S02] is in ppm (by volume), [C02] and [02J are
the partial pressures of these ~es in atm, and K3 is
the equilibrium constant in IjlaEffi. The results of such
calculations for concentrations of C02 and 02comparable
to those present in the fluidized-bed systems are shown
in curves A of Fig. H-l.
x 105
(H-4)
POPE, EVANS AND ROBBINS
INCORPORATED
-------
z
Q
..J
..J
:i
a:
ILl
0.
en
t-
a:
~
~
N
o
en
II..
o
z
Q
!;(
a:
t-
Z
ILl
~
8-
IOOPOO
10,000
1000
B-2
CURVES D
'0.2 %0,
~I'%O,
'\ 3% 0,
CALCULATED EQUILIBRIUM
OBSERVED IN COMBUSTOR
WITH LIMESTONE PARTICLE
ADDITION
DESORPTION PEAKS FROM
LIMESTONE BED
LEGEND
A
8-----
C-Q-&-9-
D--<>-:-O-O-
100
. .
'9'9
.c& CoCO
J
FIGURE H-I
CALCULATED EQUILIBRIUM CONCENTRATIONS OF S02
IN FLUE GASES COMPARED WITH OBSERVATIONS
IN FLUIDIZED-BED COMBUSTOR
10
CURVE B
.10
~TEMP (OF)
.Of
2000
1700 1600
tIT, °K-I~
,
.9XI0-3
I
1.0:<10-3
1500
1400
1900
1800
I
.7 XIO-J
,
.8X.0'J
,',,' ",'.:.
.., 'a"
POPE. EVANS A1'CJ ROBBINS
-------
H-3
The partial pressure of C02 is taken as 0.15 atm, and
two concentrations of excess oxygen are shown, 1% and 10%.
The equilibrium constant is taken from the tabulation pre-
pared for NAPCA by the FMC Corporation, "Applicability of
Inorganic Solids... etc." PB 184 751. These calculations
are in fair agreeme~t with the calculations of W.T. Reid
(Paper No. 69-Pwr-5, ASME - IEEE Power Generation Confer-
ence Sept. 1969), which are shown in curve B. Reid's curve
is calculated from K2 only, with an assumed concentration
of 2.7% 02. Since there is fair agreement, the equilibrium
condition does not appear to be too sensitive to the pres-
ence of reaction H-l at least at the higher temperatures.
The measured S02 concentrations from the PER experiments in
coal-fired, fluidized-bed boilers a~e ~hown in the curves
marked C taken from Reference 1. The
data are for stoichiometric ratios Ca/S of 2.5 - 2.7 and
for decreasing particle sizes.
The comparison indicates that the observed S02 concentra-
tions (Curves C) are from two to three orders of magnitude
higher than the assumed equilibrium constraints. The
system is closer to equilibrium at the higher temperatures
than at the lower temperatures. It is also closer to
equilibrium for the smaller size additives than for the
larger diameter additives. .
The departure from the indicated equilibrium could be
caused by one or a combination of the following factors:
(i)
The process is rate-limited by a slow readsorption-
recombination rate constant, k2, for the formation
of the sulfate.
(ii) The system is controlled by equilibria other than
those indicated.
(iii) Non-isothermal temperature distributions between
the solid surfaces of the bed and the gas phase.
The observations generally indicate that the first factor,
(i), the rate limitation for the forward reaction rate-
constant, k2, may be the reason for the observed reduc-
tions in S02 being well below those expected on the basis
of thermochemical equilibrium. The data are closer to
equilibrium from the smaller particle size additives,
which have. a larger area and hence a higher k2. The data
are closer to equilibrium for higher Ca/S ratios, which
also have larger sorbent areas. These two effects are
to be expected for a rate-limiting, heterogeneous reaction
POPE, EVANS AND ROBBINS
INCORPORATED
-------
H-4
in which k2 is expected to be directly proportional to
some power of the surface area of solid reactant. In
fact, the data presented in Reference 1 can be
used to infer that k2 varies as d~ or d~ where d is the
particle diameter. This indicates that k2 is controlled
only mildly by the adsorption rate of 02 and 802' and
that major control is exerted by diffusion processes on
the surface or interior to the solid. 'This viewpoint is
consistent with the observation that limestone, CaC03,
is more effective than 'pre-calcined lime, CaO. ~he carbon-
ate decomposes readily, and the desorbing C02 creates fresh,
internal pathways, and a large surface area, that are
readily available to the adsorbing 802 and 02.
It is possible that factor (ii) is a contributing cause.
The following additional equilibria may influence the
system:
1
Ca(OH)2 + 802 + 2 02 t Ca804 + H20
(H-S)
CaO + 8i02 + Ca8i03
+-
(H-6)
1
Ca8i03 + 802 + 2 02 t 8i02 + Ca804
(H-7)
Reaction H-5 is probably not too significant above 800°F
according to Reid, and reactions H-6 and H-7 would probably
not be controlling in a steady-state system where fresh
limestone is being added continuously, since. it involves
the slow solid-solid phase slagging reaction H-6. Reac-
tions H-6 and H-7 may however become important in a
static, capture-regenerative system that is cycled at
high temperatures where the formation of a slag on the
fresh lime surface would tend to poison the absorbent.
Factor (iii) is a real effect, but is probably not too
significant in magnitude.
The data obtained for combustion with a limestone bed, (1),
which naturally contains an enormous excess of limestone~
is consistent with this general interpretation. The data
show a wide scatter in the measured 802 concentra.tion for
given conditions of temperature and excess oxygen. In the
low temperature range, l450-1700°F, referred to as the
sorption cycle, the measured 802 concentrations are below
the curves C (for the much lower Ca/8 ratios in the parti-
cle addition); however, they are generally still above the
POPE, EVANS AND ROBBINS
INCORPORATED
-------
H-5
equilibrium curves. In the higher temperature ranges
1900-2000oF, referred to as the desorption cycle, the
peak S02 concentrations are indicated as the curves D
in Fig.H-I. These concentrations are substantially larger
than those that would be expected for the limestone parti-
cle additives of curves C.
Nevertheless, the data can be understood in terms of
factor (i), and particularly in terms of the kinetic
mechanism for reaction H-I, the sorption-desorption
reaction. As indicated earlier, the limestone decompo-
sition by reaction H-I is rapid, and the desorbing C02
from the interior of the solid creates fresh pathways
that are available for the absorbing S02 and 02' The
kinetic mechanism of the absorption process is probably
a two-step process; the first process, ~, involving the
chemisorption of S02 and 02 on the surface of the freshly
calcined lime~tone; the second process, S, involving the
conversion of the chemisorbed species to the stable sulfate.
Thus k2 is a composite rate-constant for the two sequential
reactions:
I ~ S
CaO + S02 + _2 02 + CaO'[S02.0] d + CaS04
a s +
Furthermore, process 8 is itself rather complex, involving
first the mobility of S02 and ° along the surface to active
sites where they react with each other and with the
Ca++O= lattice to form an initial sulfate film. Secondly,
once the initial film is formed, the formation of more
sulfate is limited by the interdiffusion of reactants
through 'the existing sulfate film. [Ca++ ions diffuse
from the interior and/or S02"0 diffuse from the surface.]
Thus the overall rate constant, k2' is probably limited
by the process S which is much slower than~. This view-
point is essentially equivalent to the prevailing view
that sorbent utilization is limited by the formation of
a sulfate shell on the sorbent particles.
However, this overall viewpoint also implies that the
reactive lime surface is both a chemisorbing surface
and a reactant. Thus, in a limestone bed containing a
large surface area, a significant fraction of the S02
captured could be in the form of a chemisorbed inter-
mediate CaO.[S02.0]ads, or in the form of a pseudosulfate
intermediate CaS04.[Ca++O='S03]. It is plausible to
postulate that it is these intermediates which decompose
readily during the desorption cycles at 1900-2000°F,
POPE. EVANS AND ROBBINS
INCORPORATED
-------
H-6
generating much higher 502 concentrations than would be
expected from the stable sulfate. The TVA report of
August 1969 indicates that the measured decomposition of
the true sulfate Ca504 is small even at 2200°F, as the
equilibrium curves predict. Thus the behavior of the
limestone-bed system is also explainable in terms of the
process being limited by the rate constant k2. And, in
particular, the behavior of the system in the sorption-
desorption cycles is consistent with 'the existence of a
chemisorbed or pseudosulfate intermediate which is less
stable than the true sulfate and which undergoes its
sorption-desorption equilibrium in a lower temperature
range than the stable sulfate.
The observed dependence of 502 concentrations in the flue
gas, on the excess 02 in the system, is as would be
predicted by the chemisorption rate process a.
The difficulty of obtaining steady-state equilibrium in
the experiments with a limestone bed is associated with
the large mass of sorbent used relative to the sulfur
present in the combusting coal. With its large surface
area that is reactive to the combustion products, the
limestone bed exerts a stronger influence on flue gas
composition than the limestone particulates. The limestone
bed is no longer the inert thermal matrix of the conventional
bed material, but is rather an active reactant. If com-
bustion conditions are altered, it takes longer for the
system to reach steady-state due to the sluggishness of
process S, and the large mass of reactant.
To test the possibility that pseudosulfate formation in
boilers would explain the discrepancy observed by other
investigators between:
1.
Equilibrium 502 partial pressure at a given temperature
over calcium sulfate by thermodynamic calculations, and
2.
502 desorptions actually observed in sulfated limestone
regeneration under low oxygen conditions,
a test was performed.
The approach taken was to save all fly ash from an FBM run,
which was mixed and analyzed: 27% carbon, 2.42% sulfur,
19.4% Ca. A separate batch of "synthetic fly ash" was then
prepared as follows: inert fine ash 26.4%, limestone 34%,
gypsum 10.6%, coal fines 29%. A test was then made in
the FBC at 1970°F to
POPE. EVANS AND ROBBINS
INCORPORATED
-------
H-7
compare S02 desorption from sulfated fly ash with that
from the synthetic mixture where Ca and S are present
at the appropriate ratio. Note that the "synthetic mix"
sulfur content is 1.97% as CaS04.2H20 plus 0.95% in the
coal fines. C02 and H20 are generated by decomposition
of the limestone and gypsum. The coal fines equivalent
S02 in the flue gas from burning the synthetic mix is
2900 ppm. The excess above this value represents gypsum
decomposition.
The FBC unit was successfully operated on fly ash contain-
ing the low value of 19% to 27% carbon. At this high dust
loading there is, from the material balances, some holding
of fine material in the bed. At about 2000°F and low 02
levels in the flue gas, some of the S02 captured by lime-
stone in the FBM desorbed. The S02 equilibrium ranges
from 4100 ppm at 0.1% 02 to 350 ppm at 4.3% 02. Most of
the points shown in Figure H-2 represent flash readings
without separate material balances, necessitated by the
low supply of fly ash. Substitution of the synthetic mix
at four levels of 02 leads to S02 values of 3400 to 4400
depending on 02 level; all of the S02 levels, after
correction for the coal S content, are at or above the
corresponding levels previously obtained with real sulfated
fly ash. It appears unlikely that calcining gypsum could
lead to "pseudosulfate" formation.
It would be desirable to repeat this experiment with
"synthetic mix" containing coal fines of low S content,
such as anthracite fines, to minimize the needed correction.
Fly ash of lower Ca and higher C content would also be
desirable to reduce the dust loading. Nevertheless the
results indicate the "pseudosulfate" hypothesis is unnecessary
to explain the observations. Also, the existing high
temperature CaS04 thermodynamic data, on which the equili-
brium calculations are based, need to be redetermined.
POPE, EVANS AND ROBBINS
INCORPORATED
-------
JI-!
4000
FBM FLY ASH
(C, C20, CaS04' INERT)
SYNTHETIC MIX
(C, S, CaC03' CaS04' 2 H20,INERT)
CORRECTED FOR PPM S02
DUE TO COAL S CONTENT
BED TEMPERATURE = 1,980° F
SEE TEXT FOR DETAILS
3000
:E
a..
. a..
.
en
~
C)
~ 2000
::::>
iI..
iI..
o
t-
Z
i&J
t-
Z
o
o
c\
\
\ .
\ a
\
\
\ 0
\
\
\
\
c
N
g 1000
o
I 2 3
RO = 02 IN FLUE GAS, 0/0
FIGURE H-2 S02 DESORPTION IN FBM FROM SULFATED
FLY ASH AND SYNTHETIC MIX
4
5
PCPF, A~NS AND RQAAlN.c;
= ;c 1';- ,~. II .: ;:
-------
1-1
APPENDIX
I.
ADJUSTMENT PROCEDURE FOR CARBON-BURNUP CELL
PERFORMANCE MODEL WITH COAL IN FUEL
The Carbon-Burnup Cell Performance model (Equation 1) based
on tests in which carbon-bearing fly ash 1vas the sole fuel
will predict performance when coal is part or all of the
fuel if the following adjustments are made:
1 .
The carbon content of the coal fed was increased by
treating two moles of hydrogen as if they were one mole
of carbon.
This is based on the oxygen requirement of
the two reactants for complete combustion.
The adjusted carbon rate, C1' is given by:
( : lH'I ~,~~~{/ 'I( ~.7../l~rh~1\ * lCl~~ R&.~e W\.t,c.~r~- 1-
'1 1DO
~\;'~ C~)1 J
(I-I)
POPE. EVANS AND ROBSINS .
INCORPORATED
-------
1-2
2.
The observed combustion efficiency is adjusted to
account for the total oxidation of hydrogen under the
conditions of a Carbon-Burnup Cell.*
The adjusted
value, CE' is given by:
(1- 2 )
(t~
(c.;. - (0 ),..14,1Q()
(c..~)x 14,1. DO
+
-r
c.c.l \(\H.. 7.., ~\ ) '" '51/(.,.00
C( \. 'N \.. i. I 1-\ ))( 5 i I ~ 0 ()
where
Ci = Fly ash rate x (wt %, Carbon)
+ Coal rate x (wt %, Carbon)
C = Ash collected in cyclone x (wt %, Carbon)
o
Cc
=
Coal rate
All mass flows are in pounds per hour.
Note also that
CE'
CE because the adjustment is small and is added
both above and below the line.
* The hydrogen content of the gas is not actually measured.
It is assumed that under strongly oxidized conditions none
would exist. This assumption is supported by the low
hydrocarbon values in the FBC exhaust.
POPE, EVANS AND ROBBINS
INCORPORATED
-------
J-l.
APPENDIX J.
BED MATERIAL FLOW RATE BETWEEN REGIONS
When a fluidized-bed boiler is designed in such a way that
various regions of the boiler are to operate at different
temperatures, the interchange of bed material becomes a
design consideration.
For example, the flow of relatively
cool particles into a hotter region acts as a heat sink.
To compensate for the "cool-particle heat sink," either more
fuel must be burned or less energy extracted by heat
exchange surfaces, excess air, etc.
The flow between adjacent regions separated by a "semi-
permeable" barrier may be the result of diffusion or a
pumping action resulting from pressure differences.
Wha teve r
the motive power, the interchange rate should be proportional
to the area of the openings between regions.
To establish the interchange rate, two tests were conducted
in the FBM in which the boiler was fired with coal while the
CBC remained cold and unfluidized.
When stable boiler operation
was achieved the burnup cell was fluidized, without fuel being
added, and the CBC bed temperature rise followed.
Both the
transient rise and steady-state values were noted, as both might
be instructive.
It was found that the temperature of the bed
rose more rapidly than that of the thin chamber walls.
This
meant that the equations describing the heat-up rate could
not be integrated.
POPE, EVANS AND ROBBINS
INCORPORATED
-------
J-2
The transient heating rate is described by
M,~~
-------
J-3
Combining Eqs. 1 and 2
M ~f dT-L -=
b b dt
GCf ~ ~ - 'b) t UtA1 (',AI -~ )- GqCp ('6 -It.) - L\A...JT\>-I", ')-Ui\\\b -Ie. ')
b '
Since T , the temperature of the chamber wall, is not a
constanE and is related to T in a complex manner,
Equation 3 cannot be integra€ed. Despite this difficulty,
it was found that the interchange rate, G, could be determined
by evaluating Equation 1 at Tb = T. The result thus obtained
also satisfied the equation 0
~J-3 )
'B:= Kk - -l k~ - tcJe- ol--t
\. k!
(J -4 )
whe re
Kl -= c:, LPb 1- l\}\l + Gc.. CPt.. t- 1J...f\1-
~ l- -::0 L, Cr b'1 "\- Ui t\;c, + Gc.. Cp~ +- lJ2- t\.l-I:
cJ..-...==
k.1
M ~ (PL
+ MbLPb
For the steady state case, t - "'"
the bed temperature should be given
by
\;,
-=
~~
Kj
(J-5)
POPE, EVANS AND ROBBINS
INCORPORATED
-------
J-4
RESULTS:
Table J-l below gives the flow rate, G, obtained by
evaluating Eq. J-l at Tb = To' and dT - 0 and Eq. J-5.
dt -
TABLE J-l.
RESULTS OF INTERCHANGE EXPERIMENTS
Indicated Flow Rate Flow Rate
Open Area, Pounds Per Hour Per Unit
Test Sq. Inches Eq. J-l Eq.. J-5 Area, Eq. J-l
A. 8 690 570 12,400
B. 12 1,000 1,390 12,100
It is believed that the results of Eq. J-l would be most
reliable since fewer constants must be evaluated.
For larger openings, which would be used in commercial designs,
a value of 10,000 lb/hr per square foot of opening is recommended.
This data may be roughly compared with the diffusion coefficient
suggested by Argonne (ANL/ES/CEN-F017, page 29) for -1/16 inch
particles diffusing in a bed fluidized at 3 ft/sec.
The value
given there was D = 300 ft2/hr based on experiments at the
Coal Research Establishment.
D =
kdp(V - Vmf)
Vmf
(J-6)
Assume
which has been given for a "packed-fluidized bed" where
k = an empirical constant, ft/sec
d = particle diameter, ft
p
V = superficial gas velocity, ft/sec
Vmf = minimum fluidization velocity, ft/sec
POPE, EVANS AND ROBBINS
INCORPORATED
-------
J-5
Comparing a -1/8 inch bed fluidized at 12 fps (HV) with
a -1/16 inch bed fluidized at 3 fps (LV), a high-
velocity diffusion coefficient may be estimated from
\)~~ = b~V (~) ( ~~U -. \J:~ J / V::
\~ (VLV - V~~)/ V~:
(J -7)
where
l d;/ d;\I) -= 2 i \jj(~ = l2 s,r~/ V LV = ~~P$
D~" = 3~~ \f/hr .
J
\J ,,~.- 1-1/ 1 r
'>1t -=c :) ~f<;' ) Vrn~ -= tp~
Substituting these values into Eq. J-7, D may be evaluated
t\~J -. ( (Ii.. - -::. )!?> \ - Q. r 1'2(,
~ -=: 5 aD (1.) \ l-:. -1 )/I--HO o'''-J - ..Jt)C) ,\, fW'
A "pseudo" diffusion coefficient may also be computed based
on the data of the interchange experiments. The interchange
rate, R, may be equated with a diffusion coefficient and
concentration gradient
-~~) ( ~~ I )"D( H~ \ J CJ~~,t -
'it t)v . hv- I d >\ ( ~.t )
(J-8)
Assume that dC describes the concentration of hot particles,
i.e., the density of the expanded bed, = 50 lb/ft>.
Assume dX is the distance over which the hot particles travel,
i.e., the distance between the point at which the main bed
temperature is not depressed by the cold CBC interchange and
the end of the CBC, dX ~ 3 ft
\-2 ::
\ 'I) b()O
(16- \ ~ \) (W\ '5~~~L~_+~L-
~r ~v ) ~r J ~> (~+)
b ~ 7~ 0 +rL/h~
l +~y 0 U) ~
G..
b Cl.. ~, (e.. V""" )
POPE, EVANS AND ROBBINS
INCORPORATED
-------
K-l
APPENDIX K
FBM-CBC TEST SUMMARY DATA
POPE, EVANS AND ROBBINS
INCORPORATED
-------
TABLE K-l. FBM-CBC TEST SUMMARY DATA
~
I
N
Test Number B-5 B-7a B-7b B-8 B-9* B-10*
Coal Type E.Ky. Pike Cty. Perf. 8 Pittsburgh #8 Seam
% S .9 .9 . 9 3.3
% Ash 10.7 13.9 13.9 13.5
Overall Heat Balance
Fuel heat input, kBtu/hr 7470 7600 6800 7300 6600 7100
Boiler steam gain 3960 4060 3660 3760 3550 3760
Circulated H20 absorption 2226 2008 1887 1907 1764 1674
Flue gas loss 979 1253 964 1124 937 1134
~ Combustible 10ss** 265 344 279 402 422 406
Radiation & Unaccounted 100 52 20 123 57 124
Thermal efficiency, % 83 80 82 78 82 76
~ ~ Overall Combustion efficiency, % 94 94 94
CaC03 feed rate, Ib/hr
~(j) FBM
0»
]I Z Air rate, Ib/hr 5550 7150 5890 5725 5470 5685
~~ Coal rate 615 645 575 680 580 610
Bed temperature, of 1580 1650 1620 1650 1600 1600
Fly ash output, Ib/hr 160 194 174 201 172 170
~ Fly ash carbon content, % 59 54 54 54 53 51
Flue gas concentrations:
;.02' % 2.6 3.0 2.5 2.7 3.0 3.0
f?02, ppm 650 660 680 3100 3300 3350
NO, ppm 480 420 430 400 450 540
HC, ppm 90 45 80 365 150
Cb/C. 76 76 78.4
In
CBC
Air rate, Ib/hr 905 985 880 925 910 800
Coal rate, Ib/hr 10.5
Fly ash input, Ib/hr 150 184 166 190 165 160
Bed temperature, of 1960 1950 1980 1960 2000 2050
Fly ash output, Ib/hr 78 100 96 113 100 102
Fly ash carbon content, % 21 23 20 23 22 23
Cell combustion efficiency, % 82 75 78 75 74 72
1
-------
~
~~
~cn
0»
:II z
~~
~
TABLE K-l.
(Continued)
Test Number B-5 B-7a B-7b B-8 B-9* B-I0*
Flue gas concentrations:
02' % 5 6 5.4 6 5 4.5
S02' ppm 800 620 600 1100 1300 750
NO, ppm 570 490 510 520 480 440
HC, ppm 30 20 30 30 30 20
*
Initial condition before additive addition.
** Includes C ash heat, blqwdowp and HC.
Cb/Cin = (Carbon burned / Carbon ~n Coal) x 100
~
I
W
-------
TABLE K-l. (Continued)
"
I
Test Number B-ll* B-llb B-12* B-12b B-12e ~
Coal Type Perfect 8 Pittsburgh #8 Seam
% S 3.3
% Ash 13.5
Overall Heat Balance
Fuel heat input, kBtu/hr 7120 7330 7500 7600
Boiler steam gain 3670 3760 3860
Circulated H20 absorption 1486 1775
Flue gas loss 1002 1190
~ Combustible 10ss** 626 422
Radiation & Unaccounted 285 129
Ther~al efficiency, % 72.5 76
m nver~ll Combustion efficiency, % 89 90 93 87 84
~~
CaC03 feed rate, lb/hr 118 124 215
~ U>
0:t> FBM
JlZ
~ 0 Air rate, lb/hr 5710 5660
8::0 Coal rate 620 620 635 650 660
~ Bed temperature, of 1620 1610 1640 1570 1600
Fly ash output, lb/hr 200 364 174 338 361
~ Fly ash carbon content, % 51 45 52 42 33
Flue gas concentrations:
02' % 3.2 2.8 3.8 2.4
S02, ppm 3450 1900 3200 1950 1700
NO, ppm 370 380 390 430 390
HC, ppm 120 75 125 70
CB/CIN 75 61 78.5 67 73
CBC
Air rate, lb/hr 780 820
Coal rate
Fly ash input 182 331 158 308 329
Bed temperature, of 1980 1970 1950 1910 1900
Fly ash output, lb/hr 133 222 104 235
Fly ash carbon content, % 33 18 26 24 24
Cell cumbustion efficiency, % 53 73 67 57 35
-------
~
~~
~(J)
g~
~~
6j
TABLE K-l.
(Continued)
Test Number B-ll* B-llb B-12* B-12b B-12e
Flue gas concentrations:
02' % 3.5 2.4 2.4 2.3 0.6
502' ppm 1420 940 2500 2000 3950
NO, ppm 560 540 510 430 580
HC, ppm 20 35
*
Initial condition before additive addition.
** Includes C, ash heat, blowdown and HC.
~
I
Ul
-------
TABLE K-1. (Continued)
"
I
(j\
Test Number B-13c B-14 B-15 B-16b B-17b
Coal Type Powhatan
% 8 4.5
% Ash 12.1
Overall Heat Balance
Fuel heat input, kBtu/hr 7600 6750 6030 5472 11,065
Boiler steam gain 3810 3500 3050 2690 4365
Circulated H20 absorption 1699 1426 1388 1138 1807
Flue gas loss 1159 1535 1324 628 1469
~ Combustible 10ss** 439 107 104** 780
Radiation & Unaccounted 466 182 167 1860
Thermal efficiency, % 73 73 73 70 56
m
~~ Over~ll CQrnbustion efficency, % ['\,93.3] 99.1 98.6
CaC03 feed rate, lb/hr Bed 240 140
~(f)
o~ FBM
JJZ
~O Air rate, lb/hr 5440 5830 4821 4699 8400
~::u .Coal rate 666 592 529 480 846
~ Bed temperature, of 1600 1700 1590 1570 1640
~ Fly ash out, lb/hr 165 153 140 300 318
Fly ash carbon content, % 45 48 49 22 33
Flue gas concentrations:
02' % 3.3 3.5 3.3 6 3.4
802' ppm 1500 3100 2600 700 1700
NO, ppm 290 370 310 300
HC, ppm 200 70 20 0
CB/CIN 83.2 81. 5 80.5 79 83
CBC
Air rate, lb/hr 840 1350 1880 1194 0
Coal rate coal feed
Fly ash input 150 139 127 0 0
Bed temperature, of 1800 2000 2000 1860 170
Fly ash out, lb/hr 109 76 68 0
Fly ash carbon content, % 4.3 7
Cell combustion efficiency, % ['\,60] 95 93
-------
rn
~~
~(J)
~~
~~
~
TABLE K-l.
(Continued)
Test Number B-13c B-14 B-15 B-16b B-17b
Flue gas concentrations:
02' % 1 8.5 10 9
S02' ppm 4500 1000 1800 1850
NO, ppm 350 460 470
HC, ppm 600 10 20 0
~
* Initial condition before additive addition.
** Includes C, ash heat, blowdown and HC.
~
I
--.J
-------
L-l
APPENDIX L.
DISCUSSION OF FLUIDIZED-BED COMBUSTION VARIABLES
A.
1.
COAL RELATED VARIABLES
The description of any particular coal may include a large
number of chemical and physical parameters, some of general
interest, others of interest only in a limited number of
applications.
The following basic coal variables are presently believed
to influence, to some extent, either the design or opera-
tion of a fluidized-bed combustor: . .
1-
2.
3..
4.
5.
6..
Moisture content
Ash content
Calorific value
Sulfur content
Ash fusibility
Volatile content
The first three variables determine the theoretical mass
flow rate of fuel required to deliver the desired steam
flow. Hydrogen content might also be added to this list.
The sulfur content determines the additive requirements.
Variables 5 and 6 "suggest" the upper and lower limits for
operating temperature of a fluidized-bed combustor.
In addition to these properties, for which standard tests
have been suggested, size consist, density and hardness of
the ash is of significance to the fluidized-bed boiler
designer and operator.
Definitions and Methods for Determination of Coal Related
Variables
a.
Moisture in coal is defined in ASTM D121-62 as, "Essen-
tially water, quantitatively determined by definite
prescribed methods which may vary according to the
nature of the material". It is further noted the
methods suggested may not determine all of the water
present and that ASTM D271 describes the prescribed
method for coal and coke. It may be assumed that the
same methods will apply to fly ash or char.
Within the limits of accuracy required for design and
operation ASTM D27l adequately measures moisture in
coal. .
POPE. EVANS AND ROBBINS
INCORPORATED
-------
L-2
In this method a specially prepared sample is heated
in dry air at 104 - 110°C for one hour and the moisture
content is defined as the loss of sample weight. The
method does not determine some of the water chemically
bound which then reports as hydrogen in the ultimate
analysis. An additional source of error is that due
to oxidation weight gain by the coal in the heated air.
An inert gas drying method has been tentatively
suggested to overcome this difficulty.
Moisture may affect the combustion process in two ways.
First, it may be assumed that ignition does not occur
until essentially all moisture has been driven from
the coal. However, heat-up time in a fluidized bed is
small (ten's of milliseconds) when compared with time
spent in the bed (hundreds to thousands of milli- .
seconds).
The second manner in which moisture affects the combus-
tion process is in the energy removed from the combus-
tion bed in heating, vaporizing and superheating water.
Assuming a partial pressure of 1 psi for water vapor
in the flue gas a pound of water in the fuel removes
~ 1800 Btu from a 16000F bed. This effect may be
anticipated in the design, if the moisture content of
the coal to be burned is known, or in changes in
appropriate operational parameters (bed height primar-
ily) if the source of coal is changed.
Q.
Ash in coal is defined in ASTM D 121-62 as the "Inor-
ganic residue remaining after ignition of combustible
substances, determined by definite prescribed methods".
In addition it is noted that ash so determined may not
be iuentical, in composition or quantify, with the
inorganic substances present in the material before
ignition. It is further noted that the prescribed
method for ash content of coal is that of ASTM D 271.
Within the limits of accuracy required for the design
and operation ASTM D 271 adequately measures the
quantity of ash in coal. Except for calcium oxide
content the chemical composition, per se, of ash is of
no interest to the designe~ or operator of a fluidized-
bed boiler. Where increased accuracy is desired as to
the quantity and composition corrections may be made
using the methods of Parr and Rees.
POPE, EVANS AND ROBBINS
INCORPORATED
-------
L-3
The effect of ash content on the combustor is similar
to that of moisture in that the ash removes energy
from the bed. One pound of ash elutriating from the
bed at 16000P removes approximately 400 Btu.
Ash in coal, as fired in a fluidized-bed
be retained in the bed or may be carried
gas. The retention or expulsion depends
of the ash, primarily its particle size,
hardness.
boiler, may
out with the
on the nature
density and
Coal ash is believed to have originated in two ways.
First, some ash is that mineral matter absorbed by the
vegetation which gave rise to coal. Second, the bulk
of the ash arose from the silt washed in as the precur-
sors of the coal beds were forming. It is the sedi-
mentatous matter which would be expected to provide
the ~articles which might be ~etained by the bed. A
final source of material which is defined as ash are
impurities taken with the coal in mining, rocks or
earth, and impurities which intrude the coal in handl-
ing and shipping i.e., railroad bed gravel, rust and
scale, nuts and bolts, etc.
c.
No standard tests have been devised for the property
of coal ash which affects a fluidized-bed boiler since
it is superfluous to conventional firing methods;
The calorific value or heating value of the fuel is
determined following the methods of ASTM D 271. In
this method a sample of the coal is burned with oxygen
in a bomb calorimeter following the prescribed
procedure.
The value obtained in this procedure is the gross or
higher heating value of the fuel burned at constant
volume with the moisture in the fuel and water vapor
resulting from the hydrogen in the fuel being cooled
to about 200C. .
In an actual fluidized-bed boiler the fuel will be
burned at constant pressure and the gases discharged
at above 100°C. Two corrections are therefore required,
one to account for the combustion mode and the second
to account for the discharge of water vapor. The
latter correction is made in the procedures prescribed
for determination of boiler efficiency wherein the
accepted convention is to charge the boiler with the
enthalpy contained in the discharged water vapor.
The correction for constant pressure combustion.,
general]y small, is derived at the close of thi~ ~ection~,
POPE, EVANS AND ROBBINS
INCORPORATED
-------
L-4
The effect of calorific value of the coal may be treated
in a straightforward manner in the design and operation
of a fluidized-bed combustor.
d.
Sulfur in coal is defined as the total elemental sulfur
equivalent of the organic and inorganic sulfur forms.
ASTM D 271 prescribes the methods for determining this
sulfur. ASTM D 2492 describes the methods for differ-
entiating between forms of sulfur; sulfate, pyritic
and organic.
Sulfate sulfur in coal is of some interest in that it
may be retained in the ash after combustion in a fluid-
ized bed. However, the amount of sulfate sulfur is
generally small and its retention or release as S02
does not effect the design and operation of a fluid-
ized-bed boiler.
The sulfur content of coal has been treated as a
control variable and was found not to affect signifi-
cantly the stoichiometric limestone requirement within
the range evaluated, Le., 2.6% S to 4.5% S.
e.
Ash fusibility denotes the temperatures at which coal
ash deforms and flows when heated in certain atmo-
spheres. The tests which are empirical are described
in ASTM D 1857.
Four temperatures are noted: An initial ash deform-
ation temperature (IT) at which the apex of the
specially prepared ash cone rounds; a softening temper-
ature (ST) at which the cone is spherical and height
equals base width; a hemispherical temperature (HT) at
which the cone is hemispherical and the height is one
half the base width and a fluid temperature (FT) at
which the cone height is 1/16" or less. These four
temperatures are noted when the test furnace atmosphere
is reducing and also when it is oxidizing. Eight
values are therefore normally provided for any coal.
It has been empirically determined that it is possible
to operate a stable coal-fired fluidized bed at the
softening temperature (ST) of the coal ash as deter-
mined in a reducing atmosphere.
POPE. EVANS AND ROBBINS
INCORPORATED
-------
L-5
This appe~rs to establish the upper limit for operation
of such a system.
Although the empirical limit appears to exist at ST
(reducing) a correlation of the bed collapse temperature
and the fusibility of coal ash has not been developed.
f.
Volatile matter is defined in ASTM D 121 as "those
products, exclusive of moisture, given off by a
material as gas or vapor, determined by definite
prescribed methods which may vary according to the
nature of the material". It is noted that for coal
and coke the prescribed method is that of ASTM D 271.
Although the amount of weight lost in heating coal in
a fluidized bed will differ from that lost following
ASTM D 271 it has been determined empirically that a
correlation exists between ignition temperature and
volaLile content as determined in the ASTM test.
Ignition temperature is defined as the temperature at
which more heat is generated than is lost to the
surroundings. It is a function of both fuel properties
and reactor design.
The correlation between ignition temperature and
volatile matter although not quantified is sufficiently
well understood for purposes of fluidized-bed boiler
design and operation.
A similar correlation exists for ignition stability
temperature, a temperature somewhat higher than the
igpjtion temperature. This too, although not quanti-
fied is sufficiently well understood for design
purposes.
Although the temperature at which coal ignites and
burns stabily in a fluidized bed is related to its
volatile content other tests are required if predi~a-
tions of the reaction rate are desired. These tests
would, by necessity, be empirical and conducted in a
test system almost identical to the fluidized-bed
boiler.
One tes~ procedure has been suggested in which a "burn-
ing profile" is plotted for various coals and is cited
as being capable of predicting carbon loss from a
pulverized-fuel boiler. This test however omits the
comminution which occurs in the fluidized bed.
POPE. EVANS AND ROBBINS
INCORPORATED
-------
L-6
B.
COMBUSTION AIR RELATED VARIABLES
Air for combustion is, by convention, considered to consist
of "dry air" and water vapor. Dry air consists, again by
convention, of oxygen (21.0% by volume) and a mixture of
inert gases which are considered nitrogen (79.0% by volume).
The pressure and temperature of the air also influence .
design and performance.
1.
J1ater Vapor
o
The water vapor content may be eXpressed in a number of ways
but must be reduced to pounds of water vapor per pound of
dry air (specific or absolute humidity) for use in a combus-
tion calculation. Determining the wet and dry bulb tempera-
ture with a'sling psychrometer and the barometric pressure
provides data of sufficient accuracy for calculating the
effects of water vapor content. .
The water ',Y3.por content of the comb..,stion air affects bed
temperature and boiler efficiency in predicable ways. ~
pound of water vapor entering the bed at 600°F and leaving
at 1600°F removes approximately 522 Btu from the bed. A
pound of water vapor entering the air heater at 80°F and
exiting with the flue gas at 350°F is responsible for an
energy loss of approximately 120 Btu. A precise value for,
efficiency loss due to moisture in combustion air may be
derived from a calculation of enthalpies at the precise
partial pressures and temperatures. The convention, however,
is to assume the exiting water vapor at 1 psia.
2.
Pressure and Temperature
Combustion air becomes available to the boiler system at
ambient or atmospheric conditions. In order to make it
available to the combustion process, it must be compressed
and accelerated. In passing through the air distributor and
the bed, some of the potential energy in the air is .
transferred to the bed. The effect on bed temperature is
small, although stagnation temperature equations could be
used to calculate the effect of various grid designs and bed
_h d~pths on the bed energy balance. . '
Combustion air temperature is an important parameter since
the inert components and unreacted oxygen remove energy from
the bed. One pound of nitrogen entering the bed at 600°F
and exiting at l600°F removes approximately 271 Btu from the
bed. Unreacted oxygen has a similar effect, 254 Btu/lb
required to go from 600°F to l600oF. Use of air in excess of
that required for complete combustion acts to cool the bed
in a way generally predicatable on the basis of the energy
rise of these two gases and the water vapor content.
POPE, EVANS AND ROBBINS
INCORPORATED
-------
L-7
3.
Discussion
The combustion air related variables are design variables,
not control variables. The design convention is to assume
the unit 0 feet above sea level, air at 800P and 60% rela-
tive humidity. In designing a unit for a specific installa-
tion, the user must specify local conditions. Although a
slight increase in oxygen partial pressure is obtained under
a potential mode of operation, i.e., pressurized vs. balanced
draft, its effect on combustion intensity could not be
measured.
Combustion air temperature is also a design variable. Air
passing into the bed achieves bed temperature so near the
bottom of the bed that the effect of inlet air temperature
alone on combustion is probably not detectable.
c.
BED MATERIAL RELATED VARIABLES
1.
General
The bed material of a fluidized-bed boiler is defined as that
inert, solid, granular substance which makes up the bulk of
the fluidized bed. The gases, burning carbon particles,
and dusts (limestone or fly ash) are excluded from this
definition. The bed material supplies ignition energy to
the incoming coal, transfers energy to the boiler tubes,
dilutes the burning coal so as to prevent clinkering,
provides an effective radiator to the freeboard and convec-
tion bank, attrites burning coal particles and may act to
"accept" sulfur dioxide. It also may decrepitate and
require replacement, erode the boiler tubes and air distrib-
utor and change with time in a way that disturbs the process.
2. .Source or Name
It is necessary to first identify the material as to its
source. Materials being studied today include: natural.
coal ash, sintered coal ash, alumina, limestone, zirconia
and refractory grog.
3. .Particle Density
Particle density effects fluidization velocity requirements,
elutriation rate and heat transfer properties.
\
Por most substances that would be suitable as a bed material,
the pycnometer methods of ASTM Standards DI53-54, D792-60
and D1480-62 will be appropriate. .
POPE, EVANS AND ROBBINS
INCORPORATED
-------
L-8
Since particle density is temperature dependent, it is neces-
sary to correct the density found at room temperature to the
temperature at operating conditions.
The pycnometer method requires wetting the material with a
liquid, normally water. For materials which will react with
water, such as calcined limestone, an organic liquid, such as
carbon tetrachloride or hexane, should be used. .
The particle density which is significant to fluidization and
elutriation phenomena will probably be that determined with a
fluid which did not wet the particles or drive air from any
pores. Viscous or a non-wetting liquid, such as mercury,
might be used here.
A factor such as shape, which influences the particle behavior
profoundly, is so poorly understood that small errors in
particle density description cannot be considered serious.
The use of a wetting liquid, appropriate to the chemical
nature of the particle, provides an adequate measure of parti.-
cle density.
4.
Bulk Density
The fan power requirements are a strong function of the
total bed weight. Bed weight is in turn proportional to
the bed depth and its bulk density. Settled bed density is
also an appropriate term for this parameter and is probably
more meaningful. The methods pre~r.ribed in ASTM D29l-60
will provide data of sufficient accuracy for design purposes.
5.
Size Consist or Particle Size Distribution
The behavior of the operating bed appears to be strongly
influenced by the particle size distribution or size consist.
Certainly, the ability to turn down without any portion of
the bed becoming static is related to the size consist. Again,
particle shape which is poorly understood appears to be of
equal importance.
POPE, EVANS AND ROBBINS
INCORPORATED
-------
L-9
The methods of ASTM CI36-67, a test for aggregates, appear
adequate for particle size distri~ution.. A mean ~r.a~erage
particle diameter may then be der1ved uS1ng a def1n1t10n by
Reboux.
~:'I~
(L-I)
where
D = average or mean particle diameter
p
x = weight fraction in a certain narrow cut size range
d = simple mean diameter in the narrow cut
p
Generally, the particle size distribution and ave7a~e p~rti-
cle diameter have been ignored in favor of a spec1f1cat1on of
the top particle diameter which is simple to establish. Top
size is useful since it clearly differentiates between the
high velocity and low velocity systems. A s~ecification of
both top size and mean size is probably requ1red for a good
description.
6.
Particle Shape or Rugosity
Natural ash beds tend to be made up of particles which are
irregular flat discs. Most other materials tend toward
spheres. Rugosity is the ratio of the actual external surface
area to the surface area of a hypothetical sphere. A shape
factor has been defined for use in various correlations but
appears not to be very useful and is. very tediously deter-
mined. If fluidized-bed boilers come into widespread use,
an investigation of particle shape on performance may be
justified. .
7.
Friability.or Resistance to Abrasion
Bed particles impact heat transfer surfaces, the air distrib-
utor and one another and are constantly being diminished in
size. Although the only conclusive test is at temperature
and in a device similar to the boiler, methods are available
for gross screening of materials as candidate substances.
The methods of ASTM D441-45, Tumbler Test for Coal, or
ASTM CI31-66, for aggregates, describe adequate procedures
and apparatus which might be adapted for use as a standard
method for fluidized-bed materials. Where the material to be
evaluated is natural coal ash, a sample of coal would be
ashed at low temperatures to preclude sintering and tested.
Limestones would be difficult to test properly since, in
operation, they would be calcined and partially sulfated.
POPE, EVANS AND ROBBINS
INCORPORATED
-------
L-10
8.
Loss-on-Charging
Bed .materials other than natural ash would normally be
purchased on a weight basis (or by a volume measure, cubic
yards, which is essentially equivalent). It may be antici-
pated that the as-received particle size distribution will
differ from that required in the boiler unless a double
screened, and therefore more costly, material is specified.
The selected material may suffer a loss on ignition due to
: organic component, moisture, and carbonate or water of hydra-
tion. In preparing a material for charging by crushing,
fines are produced which will elutriate on charging to the
boiler. On charging, the sharp edges attride rapidly and
additional mass is lost.
Although no standard test exists for "loss-on-charging,"
it might be constructed from existing tests for loss-on-
ignition, resistance to abrasion and screen analysis.
Certain potential bed materials may also decrepitate on
charging due to the rapid heating. Some limestones or
materials ,. i th water of hydration 1'ight literally explode
on charging. It appears that actual use in a fluidized-bed
boiler is the most meaningful test of "loss-on-charging."
9.
Discussion
The choice of bed material appears to be a design variable.
It is doubtful that bed material affects the combustion
process to any significant degree. Although not listed
above, it should be obvious that a high melting point,
inertness and low cost are basic requisites of any candi-
date material.
D.
BED TEMPERATURE
The temperature of an oxidizing fluidized bed has been
defined as the temperature indicated by a thermocouple
immersed in the bed. While the precise type of couple
its heat loss by conduction may affect the temperature
cated, errors above 5% are unlikely.
and
indi-
The temperature so indicated represents the average tempera-
ture of the energy sources and sinks which make up the bed.
It has been assumed that this average temperature closely
approximates the average temperature of the inert particles
which make up the bulk of the bed.
POPE, EVANS AND ROBBINS
INCORPORATED
-------
L-ll
It is possible by use of thermocouples to detect the
macroscopic gradients, both vertical and horizontal, which
result from the combined effects of combustion and heat
loss gradients.
It is not possible by means of thermocouples to determine
the microscopic gradients which must exist due to second
law considerations. The maximum temperature to which a
particle has been exposed might be determined by analysis
of samples removed from the bed.
While bed temperature is a response variable, it is actually
one of the fundamental design parameters. Bed temperature
has been found to affect combustion intensity, carbon carry-
over, 802 removal by limestone, and NO generation.
E.
FUEL ENERGY INPUT RATE
The fuel energy input rate is defined as the higher heating
value of the fuel (Btu/lb) multiplied by the fuel rate
(pounds/hr or pounds/hr for each square foot of bed surface).
When the fuel enters the combustor at a temperature above
the reference temperature, usually the ambient air tempera-
ture or 80°F, the sensible heat in the fuel is added. By
convention, coal is assumed to have a mean specific heat of
0.3 Btu/lboF. More precise values for specific heat may be
used although the fraction of the total energy carried into
the combustor as sensible heat in the raw fuel is very small
in any real case. When the fuel is recycled particulate,
the sensiblp- heat may be important. .
Fuel energy input rate, coal rate, fuel rate and firing rate
are synonymous terms. While this control variable is related
to the heat release rate or combustion rate (a response
variable), it is not identical. Not all the fuel entering
the bed actually burns (10-20% of the energy in the fuel may
appear as unburned combustible), and not all of the fuel
. which burns in the combustor 'actually burns in the bed.
POPE, EVANS AND ROBBINS
INCORPORATED
-------
L-12
F.
SUPERFICIAL VELOCITY
The specification of an approximate superficial velocity in
a fluidized-bed reactor is useful in distinguishing between
kinds of fluidized beds, that is, true fluid beds and what
has been termed "teeter"* beds. Superficial velocity is
normally considered a control variable. In fact, it is
response variable. Superficial velocity is defined as the
velocity of the fluid which would exist at a specific point
in the reactor if it were empty of both particles and inter-
nals. The specific point which is usually designated is the
top of the expanded bed. It may be specified knowing the
total mass flow rate, the specific volume at temperature and
the cross sectional area of the reactor at the point of
interest. It may be useful as a tool in determining, for
fluid beds, from correlations available in the literature,
minimum fluidization velocity, bed expansion, elutriation
rates and particle residence times. Unfortunately, many of
the correlations proposed for the fluid bed do not hold for
teeter beds. .
For a combustion system, which may be packed with internals
and in which all the combustion does not occur in the bed,
specification of a superficial velocity is of relatively
little value as a design tool.
In lieu of superficial velocity, it would be more appro-
priate to specify the air rate and the average bed tempera-
ture. If the superficial velocity is of interest, an approx-
imate value may be computed from a knowledge of these two
parameters.
*
Beds ~ade up of large particles ( + 20 mesh) permitting high
mass gas flows
POPE. EVANS AND ROBBINS
INCORPORATED
-------
L-13
G.
CARBON LOSS
Comparing the combustion conditions of a fluidized-bed boiler
with a pulverized coal furnace, bed temperature and oxygen
partial pressure in the fluidized bed are lower and coal
particle sizes greater. As a result, it has not been found
possible to burn all of the coal as it is fed" to the bed.
Some particles leave the bed and are carried out of the
region in which burning may occur with the flue gas before
combustion is complete.
The potential loss of fuel value due to carry-over of this
unburned particulate combustible matter has been termed
"carbon loss" since the major combustible component is
carbon. The particles, however, still contain some hydrogen
which increases its heating value. While the use of the
term carbon loss is a useful shorthand, since it describes
the major potential inefficiency of fluidized-bed boilers,
it is more appropriate to specify the combustion efficiency
as defined below:
CE
-
1-~
- "~
(L-2)
where
CE"
= combustion efficiency, dimensionless
Ho = heating value of unburned fuel components
leaving, etc., Btu
H. = heating value of fuel fed to system, Btu
J.
The term Hi may be considered as the integrated higher heat-
ingvalue of fuel fed over a specified period.
The term Ho is the sum of all combustible matter leaving
the system. It would include the heating value of:
a.
b".
Combustible in
Combustible in
Combustible in
bottom ash, primarily carbon
fly ash, primarily carbon
flue gas as hydrogen, as carbon
monoxide and as hydrocarbons
c.
While a precise determination of the heating value of the
solid matter and gas might be preferable, conventional
practice has been to assume the following values:
solid combustible = 14,500 Btujlb
carbon monoxide = 10,160 Btujlb
hydrogen = 319 Btujft3 (STP)
POPE. EVANS AND ROBBINS
INCORPORATED
-------
L-14
No convention exists for the heating value of hydrocarbons,
but it might be assumed that the hydrocarbons as reported
by a flame ionization chamber, as methane, do describe the
gas. Therefore, the heating value, the psuedo methane,
might be taken as 1030 Btu/ft3 (STP).
POPE, EVANS AND ROBBINS
INCORPORATED
-------
L-15
CORRECTION OF CONSTANT VOLUME HEATING VALUE TO CONSTANT PRESSURE
HEATING VALUE.
Nomenclature:
C = Specific heat, Btu/lb oR
H = Weight of hydrogen in fuel, %
J = Mechanical equivalent of heat = 778.26 ft lb per Btu
M = Pound moles of gaseous constituents per pound of fuel
N = Weight of nitrogen in fuel, %
P = Pressure, pounds per ft2
Q = Quantity of heat absorbed by system from surroundings,
Btu/lb of fuel
R = Universal gas constant = 1545 ft lb/lb mole oR
T = Calorimeter temperature, oR
U = Internal energy, Btu
V = Volume, ft3/lb of fuel
~ = Net change in quantity following
Subscripts
v = constant volume
p = constant pressure
1 = start of process
2 = end of process
(1) = liquid
(s) = solid
(g) = gaseous
The derivation of the correction is based on the first law of
thermodynamics and the perfect gas law. In burning a sample of
coal in a bomb, no work is performed and there is no kinetic
energy change and the first law may be expressed as
Qv = U2 - Ul =
But Cp~T = Qp
Rearranging
Qp = Qv +
Cp~T - ~PV/J
( L- 3 )
(L- 4)
~PV/J
(L- 5 )
The perfect gas law states,
~PV = ~MRT
(L-:-6)
POPE. EVANS AND ROBBINS
INCORPORATED
-------
L-16
Therefore,
Op = Ov + 6MRT/J
(L-7 )
To determine the 6M, the change in the number of moles of
gaseous constituents, the combustion reactions in the calorim-
eter are examined.
C ( s ) + 02 ( g ) + CO 2 ( g) ,
1 solid mole + 1 gaseous mole + 1 gaseous mole, no net change
H (s) + ~02(g) + H20 (1)
1 solid mole + ~ gaseous mole + 1 liquid mole, -~ gaseous mole
S (s) + 02 (g) + S02 (g)
1 solid mole + 1 gaseous mole ~ 1 gaseous mole, no net change
N2 (s) + N2 (g)
1 solid mole + 1 gaseous mole, + 1 gaseous mole
Other fuel constituents are assumed to undergo no net change,
i.e. ,
H20 (1) + H20 (1)
02 (s) + 02 (s)
Ash (s) + Ash (s)
-N
6M = 28.016
~H
+ 2.016
(L- 8)
Inserting into Eq. (L-7)
Op = Qv + (4.~32 - 28.~16)
RT
778.26
(L-9)
Examples:
N = 1%
H = 4%
T = 68°F
Ov = 13,000 Btu/lb ;
H
4.032
=
.04
4.032
=
.00993
-N
28.016
=
.01
28.016
= -~00036
6
=
.00957
Qp = 13,000 + .00957(1050) = 13,010 Btu/lb
POPE. EVANS AND ROBBINS
INCORPORATED
-------
M-l
APPENDIX M.
WHY THE CARBON BURNUP CELL?
M .1
Perspect ive
While the body of the report discusses
in October 1970, this section must properly
reflect what is known in January, 1972.
Yi"Ork completed
be written to
To answer, "Why the Carbon-Burnup Cell" it is necessary
to explain, a bit more fully, what a fluidized-bed is, and
how particles leave the bed, how coal burns and how coal burns
in a fluidized-bed, why a fluidized-bed boiler is put together
the way it is and why all of these factors could have led to
the invention of the Carbon-Burnup Cell.
M.2
Why Particles of Coal Leave a Fluidized Bed
A fluidized-bed was defined in Section 4.1.2 and in
Section 4.1.3 it was noted that some particles added to,
or made, in the bed, could be too large and would sink to
the bottom of the bed. Other particles added to, or made,
in the bed are too small to remain in the bed and are carried
out and away from the bed by the gases which fluidized the bed
or passed through the bed.
In practical fluidized-beds, some gas passes through the
bed, either as bubbles or in channels, without participating
in the fluidization of the particles. These channels or
bubbles may cause the bed to be violently agitated. And,
just as a violently boiling pot of water throws off large
droplets of water, the violently fluidized-bed shoots up,
into the space above the bed, even the largest particles
which make up the fluidized-bed. The support of the upward
flow of gas and the kinetic energy of the particles, imparted
as they left the bed, tend to keep the particles moving upward.
POPE, EVANS AND ROBBINS
INCORPORATED
-------
M-2
Gravity acts to return the particles to the bed.
The he ight of the freeboard; i.e., the space above the
bed, detenmines which of the large particles will finally
lose their initial upward momentum and fall back into the
bed.
If the chamber, which holds the fluidized-bed in its
bottom, is tall enough, every particle which is too heavy
to be supported by the upward flow of gas will fall back
into the bed. If the pipe or duct which allows the gas to
leave the chamber is too close to the top of the fluidized-
bed; i.e., there is little freeboard, many of the large
particles will be swept into the outlet pipe and will not
fall back into the bed.
The small particles, whose terminal velocity is less
than the actual gas velocity, cannot be made to return to the
bed unless they are first separated from the gas in a dust
collector. Fine particles appear in the bed because: (a) they
are fed to the bed as fine particles, (b) small pieces are
broken from large pieces, (c) large particles are consumed by
chemical reactions until the particles become small, (d)
chemical reactions change gases into fine particles. The
large particles can be made to return to the bed by providing
a baffle in the freeboard which the particles may strike,
thus losing their initial upward momentum. These large
particles cannot be supported, solely, by the upward flow of
gas and so fall back into the bed.
Particles of coal, like the other particles which make
up the fluidized-bed, will appear in the freeboard either
because they are too small to remain in the bed or because
they are thrown out of the bed by the violent agitation.
POPE. EVANS AND ROBBINS
INCORPORATED
-------
M-3
From what has been said above it should be clear that
if particles of coal, fed to the bed, have not been totally
consumed before they come near the top of the bed, some coal
particles will appear in the freeboard. It should also be
clear that the freeboard is an integral part of a fluidized-
bed system. It also follows from what has been said that
unless the coal particles return* to the bed, or are totally
consumed in the freeboard, unburned fuel values will leave
the furnace. These matters will be discussed in following
sect ions.
Despite a greal deal of effort by researchers around
the world, no general equations have been derived which can
be used to predict how much material will enter the freeboard
above a fluidized-bed. Some of the major chemical and
petroleum companies have developed useful prediction tools
for some types of fluidized-beds made up of fine particles
which tell them how tall a vessel (i.e., its disengaging space)
should be. These tools, carefully guarded trade secrets,
are of no relevance here. No one, to our knowledge, has
developed a prediction tool for the high gas velocities used
in fluidized-beds made up of the large particles used in a
fluidized-bed boiler.
M.3
Coal Combustion
It was shown, in the preceding paragraphs, that the entire
size spectrum of bed particles will appear in the freeboard,
and under some conditions, in the outlet pipe, of a fluidized-
bed boiler's furnace. Some of these will be coal derived
particles which are commonly referred to as "char".
The material which follows is taken, in part from
*e.g., by internal cyclones with dip legs.
POPE, EVANS AND ROBBINS
INCORPORATED
-------
M-4
Reference 12, sent by the OAP Project Officer in December
1971. Its useful explanation of pulverized coal (p.c.)
combustion, with changes necessary to emphasize the
differences between fluidized-bed combustion and p.c.
combustion, has been used.
The coal particle undergoes complex physical and chemical
changes during combustion which may be characterized roughly
by the sequence of preheat, ignition and char burnout. The
details, and time taken, in each phase will vary greatly
among different flames as a consequence of differences in
coal type, the coal particle size spectrum, the aerodynamic
patterns in the combustion zone and beyond, and the temperature
at which the burning occurs.
A number of the variables applicable to pulverized coal
combustion have been studied in a pilot-scale furnace at
IJmuiden by the International Flame Research Foundation. While
fluidized-bed combustion is not identical to pulverized coal
combustion, they both involve burning in a suspension, and the
same basic processss.
After injection into a furnace, the coal particles are
heated both by radiation and by convection. Time for
temperature equibriation in a conventional P.F. furnace for
-6 -2
1.JJm particles is about 10 seconds, 10 seconds for 100.}.Im
particles and 10 seconds for 1000 "Jlm particles.
For bituminous coal with a high volatile content, ignition
is achieved rapidly and the volatiles are released and burn as
a diffusion flame in fuel rich pockets. It is possible that
sufficient fuel gas is released in the vicinity of a coal
feeder of a fluidized-bed boiler to form bubbles which travel
up through the bed with oxygen diffusing into the bubble.
POPE, EVANS AND ROBBINS
INCORPORATED
-------
M-5
For large particles, volatiles may be released at a
rate high enough to prevent oxygen from approaching the
developing char until several seconds after the particle
enters the furnace.
The burnout of the char is relatively well understood,
occurring by the diffusion of oxygen to the surfa~e, reaction
at the surface to carbon monoxide, and subsequent' gas phase
, .
oxidation of the outwardly diffusing carbon monoxide to
carbon dioxide.
The rate of char burning depends on the rate at which
oxygen can approach the surface, the reaction rate of oxygen
with carbon and the rate of escape of the carbon monoxide.
The very thin boundary layers associated with fluidized beds
aid char burnup.
The relatively low temperatures and large particles
typical of a fluidized-bed boiler act to reduce the reaction
rate at the surface and also to slow the ma~s transfer.
Times to burnout for large particles of char in fluidized
beds may be on the order of tens or even hundreds of seconds.
M.4
Combustion of Coal in a Fluidized-Bed'
The chemical history of a coal particle in a fluidized-
bed boiler is similar to, altho~gh not i~entical to, a
particle in a pulverized coal flame. Both methods of burning
keep the fuel in suspension. In the fluidized-bed, the
suspension is 'stationary as long as the particles are too
large to be swept out of the bed. Particles of char which
leave the bed experience the same phenomena as the char
particies in a pUlverized coal furnace, except that the
relatively low temperature in the freeboard of the fluidized-
bed boiler is less conducive to high burning rates.
POPE. EVANS AND ROBBINS
INCORPORATED
-------
M-6
If the burning process is either reaction-rate or
desorption-rate controlled, as it appears to be, an increase
in absolute pressure will not decrease the burnout time.
Experiments reported by Hoy (13) confirm this.
From what has been described to this point, it should
be clear that combustion of the solid portion of coal residue,
the char, is relatively slow; too slow to be completed within
the fluidized-bed before some of the particles are swept or
thrown out of the bed into the freeboard. It would be possible
to provide conditions in the freeboard which would permit the
combustion of char to be completed. That these conditions are
incompatible with the design goals of a fluidized-bed boiler
is shown below.
M .4
Why a Fluidized-Bed Boiler Is Put Toqether the Way It Is
In a pulverized-coal boiler, the firing rate rarely
exceeds about 20,000 Btu/hr per cubic foot of furnace volume.
In a fluidized-bed boiler, the rate can easily be ten times
as great. The main cause of this difference is that in the
fluidized-bed a heat transfer medium is provided and it is
possible to remove the heat generated by the burning coal with
tubes that are, in a sense, immersed in the flame. The ratio
square feet of heat transfer surface per cubic foot of heat
release volume, is 10 to 15 times higher in the fluidized-bed
boiler than in the pulverized coal furnace.
One of the first test systems built by Pope, Evans and
Robbins provided the desired surface-to-vo1ume ratio by
placing boiler tubes uniformly in the combustion space. The
problems found in that design have been described in Reference
4. A design concept then evolved in which heat-transfer
surface was banked so that the tubes then formed natural
POPE, EVANS AND ROBBINS
INCORPORATED
-------
M-7
barriers and repeating elements. The resulting configurations
was termed a "unit cell". A unit cell contained a fuei supply
and sufficient heat transfer surface (boiler tubes) to absorb
a major fraction of the energy released within the cell.
Since tubes were not arranged close to and directly above
the fuel outlets, the problem of erosion due to high velocity
jets was avoided. It was also possible for a man to enter a
cell and make repairs without dismantling the boiler. The cell
lent itself to the development of effective light-off and
turndown techniques, as well as facilitating shop fabrication
of the boiler moduler. And most important here, cells could
be used to perform specialized functions.
All Pope, Evans and Robbins design concepts for fluidized-
bed boilers contain cells, and some designs proposed by others
appear to contain cells. Two names have been given to this
basic concept, which is such a radical departure from both
conventional boiler practice and what might be called
conventional fluidized-bed practice. The names to be used
are: Pope-Bishop Boiler or Multicell Fluidized-Bed Boiler.
The modern pulverized coal furnace exceeds 90% combustion
efficiency, not because there is an over-riding economic
incentive to do so but because the furnace size is set by
heat transfer needs of the process; the very high combustion
efficiency is an "extra" that follows from this large furnace
size. A fluidized-bed boiler furnace could also be designed to
provide 99¥~ combustion efficiency. For example, the following
could be done: (a) the size of the fluidized-bed boiler's
furnace could be increased by a factor of ten, and (b) the
operating temperature could be increased from 1,5000F, now
considered adequate, to about 2,2000F, and (c) the excess air
could be increased from about 15% to the 50% common in stoker
POPE, EVANS AND ROBBINS
INCORPORATED
-------
M-8
fired units and (d) carbon-bearing fly ash could be recycled
back to the furnace or in lieu of (a) to (d), (e) the air
rate could be reduced so that the superficial velocity
dropped from 15 feet per second, now used, to 0.5 feet per
second at which carryover is negligible (14). Why these
approaches to complete single-stage combustion are undesirable
is discussed shortly.
A fluidized-bed boiler is put together the way it is
because it is possible to build an extremely compact primary
cell ("furnace") if it is permiss ible to solve the problem of
fuel carryover by an adjunct means such as the Carbon-Burnup
Cell. The sections which follow will outline some of the
unsuccessful efforts at reducing loss of fuel as high-carbon
fly ash by more direct techniques.
To close this section, it is appropriate to note here
that the problem of fuel loss from a fluidized-bed boiler has
been considered in "The Federal Research and Development Plan
for Air-Pollution Control by Combustion-Process Modification"
by Battelle Memorial Institute (15). The Battelle plan
recommends, on page 111-34, basic research which ". . .shou1d
permit identification of conditions under which co can be kept
to levels not exceeding 0.1 percent and carryover of unburned
fuel can be lowered to the equivalent of carryover in existing
combustion systems." The work to be described in this report
indicates that the goals set in the Battelle plan have been
met without requiring a basic alteration in the configuration
of the fluidized-bed boiler.
4.6
Methods for Reducinq Fuel Loss Considered Prior to
Invention of Carbon-Burnup Cell
A number of techniques for reducing fuel loss from a
POPE, EVANS AND ROBBINS
INCORPORATED
-------
M-9
fluidized-bed boiler were considered, and some tested, before
the invention of the Carbon-Burnup Cell. None provided the
degree of combustion efficiency desired without at the same
time possessing an over-riding liability. These are discussed
below under the heading Time, Temperature and Turbulence, the
three factors known to increase the rate of combustion. The
discussion is qualitative and brief.
M.6.l
Time
(a) We learned in May 1965 (14) that studies at the
Marchwood Laboratories of England's Central Electricity
Generating Board (C.E.G.B.) indicaOted that at a superficial
velocity much above 0.5 feet per second, carryover of char
would be excessive in beds up to three feet deep.
Operation at low velocities was judged unfeasible,
although Pope, Evans and Robbins conducted one low velocity
test in a bed composed solely of fly ash.
(b) Increasing the depth of the bed, which would also
act to increase the time for char particles to be burned,
was considered and rejected. The power required by the high
pressure blowers needed to fluidize a deep bed, over about
three feet, would cause a fluidized-bed boiler to be
uneconomical. There was also no evidence that such deep beds
would provide the desired burnout.
CEGB's results in a three-foot bed at 0.5 fps might
be duplicated in a 90 foot deep bed at 15 fps. This calculation
is an oversimplification but the implication may be correct.
(c)
The volume of the freeboard could not be increased
to the extent required for complete combustion (at least a
factor of 10) without losing one of the major advantages of the
POPE, EVANS AND ROBBINS
INCORPORATED
-------
M-lO
fluidized-bed boiler; i.e., compactness.
(d) Recycle of collected fines* back to the bed from
whence they initially arose was tested a number of times at
high superficial velocities and found unsatisfactory.
Experiments at low superficial velocities, 2-4 fps, indicates
that this technique will work (16). However, operation at 2-4
fps is uneconomical. In addition, if the dust collectors used
for recycle are relatively efficient, an extremely high dust
load is imposed on the system. If the collector is 90%
efficient, the quantity of circulating dust may exceed, by
a factor of ten, the ash being fed with the coal.
(e) In one test, collected fines were pelletized with
coal fines into 1/2" lumps resembling charcoal briquettes.
Although the loss of fuel was reduced by the increase in
residence time, the improvement did not justify the
complication of a pelletizing operation.
(f) Theoretical studies had indicated that a pulsing
air supply could be used to pass relatively large quantities
of air through a fluidized-bed without excessive carryove.r.
One test was made of this approach but the results were not
promising. Coal feeder plugging might result.
The conclusion drawn from this work and study was
that no feasible method existed for increasing the time that
the particles of char would spend in the furnace, at least
to the extent required.
M.6.2
Temperature
(a )
Goldberger has shown (17) the effects of temperature
*Which must compete with fresh coal for available oxygen.
POPE, EVANS AND ROBBINS
INCORPORATED
-------
M-ll
on combustion efficiency of coal fed to a fluidized-bed
combustor. It was shown that for a bed depth of about three
feet, operated at 4-7 fps, feeding relatively fine coal, a
temperature above 19000F was required to achieve complete
combustion. Goldberger did not study coarse coals or propose
a method for feeding fine coals to a large diameter bed, since
these did not require consideration for the application proposed
by the sponsors of his work.
(b) Tests by Pope, Evans and Robbins at bed temperature
up to 2,lOOoF with coarse coals, shallow beds and high super-
ficial velocities did not indicate that 99~~ combustion
efficiency could be achieved by operation in this mode.
We also determined that high bed temperatures were inconsistent
with sulfur-oxide control using limestone.
(c) The use of a two-stage combustor was considered.
Coal would be fed to a fluidized-bed operating at, say 1,500oF,
at superficial velocities exceeding 15 fps. The freeboard
would be large and refractory lined. Sufficient fuel would
be expelled, by choice of the velocity and certain other
parameters to increase the freeboard temperature to as high
as 2,500oF. This was not studied in any depth and was not
tested.
(d) Combustion catalysts may properly be treated under
temperature, since they increase the rate of the process just
as an increase in temperature would.
Combustion Engineering, Inc. (CE) has a number of
patents (18) in which combustion catalysts were used in a
fluidized-bed to promote complete combustion. Pope, Evans and
Robbins did not evaluate the methods of CE and in a number of
discussions CE has not ~uggested application of their techniques.
POPE. EVANS AND ROBBINS
INCORPORATED
-------
M-12
It is likely that CE's methods are not compatible with coal
firing.
A combustion catalyst has also been studied by
Pope, Evans and Robbins and will be discussed in the report
which will follow this one. Although the catalyst markedly
increased the combustion efficiency, to the extent that the
CBC was starved of fuel, it did not appear to make the Carbon-
Burnup Cell unnecessary.
The conclusions drawn from this work and study was
that it was not useful to attempt to increase combustion
efficiency by operating the coal-burning fluidized-bed
at a temperature higher than that required to provide a
stable process and burn out CO and hydrocarbons.
M.6.3
Turbulence
Turbulence implies any
flux to the burning particles
well as mixing techniques.
means for increasing the oxygen
and includes high excess air as
(a) The action of the fluidized-bed appears to provide
sufficient turbulence to reduce the thickness. of the boundary
layer for heat transfer to tubes. It should also have a
similar action on the boundary layer of carbon monoxide leaving
a burning char particle. No means for increasing this action
is apparent.
(b) A number of studies have shown that high combustion
efficiencies could be achieved by operating at high levels of
excess air but also at relatively low superficial velocities.
The most recent to become available is that by Hoy (13).
Studies by Pope, Evans and Robbins indicate that
even at high levels of excess air ( ~50%), complete char
POPE. EVANS AND ROBBINS
INCORPORATED
-------
M-13
burnout is not achieved in a low-temperature fluidized bed,
operating at high velocities. Referring to Equation 42 an
excess air requirement of 90 to 100% (RO = 12.7%) is indicated
for a 99-1/2% combustion efficiency.
(c) Some early tests were made using over-fire air
jets so as to provide turbulence in the freeboard. The
results were not promising. Even if the tests had been
successful, over-fire jets do not appear consistent with
good fluidized-bed boiler design practice.
(d) Members of the staff of Combustion Engineering
suggested that collected fines might be finely ground before
recycle so as to increase the carbon surface area, the
equivalent of increasing turbulence. This suggestion was
never evaluated.
The conclusion drawn from this work and study was
that it was not useful to attempt to increase mass transfer
by increasing turbulence or increasing excess air.
M.7
Invention of the Carbon-Burnup Cell
It appeared from all of these efforts that it was not
feasible to provide the conditions required for single-pass
char burnout and still retain enough of the potential
advantages of a fluidized-bed boiler to make the development
worthwhile. For reasons which are unknown, the obvious
answer, the Carbon-Burnup Cell, was not obvious. Once .it
was accepted that the collected char had to be recycled, it
made little sense to return it to the same bed from which it
had originally come.
Possibly the fact that the FBM and FBC were next to
one another in 1968 made the experiment self-suggesting.
POPE, EVANS AND ROBBINS
INCORPORATED
-------
M-14
Therefore, the water cooled FBC was internally insulated by
a carbon-steel liner backed with high-temperature insulation.
The FBM and FBC were then operated in tandem; the FBM at a
low bed temperature and moderate excess air. The FEC, firing
o 0
the FBM's collected fly ash, was operated at 1900 -2000 F
with excess air being used to prevent exceeding 2000oF. A
few ounces of coal were added each time the bed temperature
fell below 1900oF. This test indicated that a combustion
efficiency exceeding 99% could be achieved in a fluidized-bed
boiler which contained a Carbon-Burnup Cell.
of the
to the
The Office of Air Programs asked that implications
Carbon-Burnup Cell be more fully explored. This led
work described in this report.
POPE, EVANS AND ROBBINS
INCORPORATED
------- |