PROJECT NO. 75-CCL-7
AIR POLLUTION
EMISSION TEST
O
•Jm
FINAL REPORT
WESTMORELAND COAL COMPANY
QUINWOOD
WEST VIRGINIA
MAY 25 1976
UNITED STATES ENVIRONMENTAL PROTECTION AGENCY
Office of Air and Waste Management
Office of Air Quality Planning and Standards
Emission Measurement Branch
Research Triangle Park. North Carolina
-------
FINAL REPORT
on
COAL CLEANING PREPARATION PLANT EMISSIONS
AT WESTMORELAND COAL COMPANY,
QUINWOOD, WEST VIRGINIA
to
ENVIRONMENTAL PROTECTION AGENCY
May 14, 1976
by
P. R. Webb, J. M. Pilcher, and J. A. Gieseke
BATTELLE
Columbus Laboratories
505 King Avenue
Columbus, Ohio 43201
-------
TABLE OF CONTENTS
Page
INTRODUCTION 1
SUMMARY AND DISCUSSION OF SAMPLING RESULTS 5
Particle Size Results 5
Process Coal Sieve Results. 5
Particulate Collected by the Scrubber 7
Cyclone Catch Analyses 17
Scrubber Inlet Measurement Results 24
Scrubber Outlet Results 24
Total Solids in Scrubber Outlet Water 29
Opacity Measurements 30
Data Applicable to All Five Runs 39
Trace Metal Results 41
LOCATION OF SAMPLING POINTS 41
Inlet to Venturi Scrubber . . 41
Outlet of Mist Eliminator 44
PROCESS DESCRIPTION AND OPERATION 47
Process Description 47
Emission Control Equipment 49
Process Operation 49
SAMPLING AND ANALYTICAL PROCEDURES 51
Coal Sample Collection Method . 51
Scrubber Water Collection 53
Particle Sizing Methodology 55
Distribution Change Due to Drying 56
Effect of Ultrasonic Dispersion Compared With
Mechanical Dispersion 56
Ultrasonic Cleaning of Alundum Thimble 59
Relative Comparison—Particle Concentration ... 62
Representative Aliquots From Large
Electrolyte Volumes 63
Coal Particle Solubility 63
REFERENCES 65
-------
TABLE OF CONTENTS
(Continued)
Page
APPENDIX A
COMPLETE PARTICULATE RESULTS A-l
Sample Calculations—Outlet Run No. 4 A-6
APPENDIX Bl
VISIBLE EMISSIONS OBSERVATIONS DURING PARTICULATE SAMPLING ... Bl-1
Summary and Discussion of Visible Emissions Bl-1
Method 9 Report Bl-3
APPENDIX B2
TOTAL VISIBLE EMISSION DATA B2-1
APPENDIX C
FIELD AND LABORATORY DATA RELATED TO PARTICULATE SAMPLING ... C-l
Preliminary Data C-l
Sample Run Data C-7
Molecular Weight Calculations C-28
Field Analytical Data C-34
Sample Drying Data C-46
Visible Emission Field Data C-75
Cyclone Field Data C-166
APPENDIX D
SAMPLE COLLECTION LOG D-l
Sampling Task Log D-4
APPENDIX E
PROCESS OPERATION FIELD DATA E-l
APPENDIX F
SAMPLE IDENTIFICATION LOG F-l
APPENDIX G
CYCLONE SAMPLE LOG AND ANALYTICAL RESULTS G-l
APPENDIX H
DIAGRAM OF SAMPLE EQUIPMENT SERIES CYCLONES INCLUDING
CALIBRATION DATA H-l
APPENDIX I
DETAILED STANDARD SAMPLING AND ANALYTICAL PROCEDURES 1-1
Particle Size Analysis by Mine Safety
Appliance Particle Size Analyzer 1-4
ii
-------
TABLE OF CONTENTS
(Continued)
APPENDIX J.
PROJECT PARTICIPANTS AND TITLES J-l
LIST OF TABLES
Table 1. Sieve Analyses of Coal Samples From West-
moreland Coal Company 5
Table 2. Subsieve Particle Distribution of Westmoreland
Coal Company Samples by MSA Particle
Size Analyzer 8
Table 3. Size Distribution Comparison of Thimble
Catches and Scrubber Water Suspension by Coulter
Counter From Westmoreland Coal Company 16
Table 4. Test Data 18
Table 5. . Analytical Data For Particle Size Runs With
Multiple-Cyclone Sampler. .__. . . . ._ 19
Table 6. Summary of Results of Particle Size Runs 20
Table 7. Inlet Measurement Results Summary (English Units) 25
Table 8. Inlet Measurement Results Summary (Metric Units). 26
Table 9. Outlet Measurement Summary Results
(English Units) 27
Table 10. Outlet Measurement Results Summary (Metric Units) 28
Table 11. Scrubber Water Analysis 29
Table 12. Frequency Observation Averages Exceeding Stated
Opacity for Particulate Sampling Only 31
Table 13. Frequency Observation Averages Exceeding Stated
Opacity for Total Opacity Data 3~2
Table 14. Opacity Observation Summary 40
Table 15. Optical Emission Spectroscopy Analysis for
Trace Metals (Run No. 4) 42
Table 16. Traverse Port Locations—Inlet 45
Table 17. Traverse Port Locations—Outlet 46
-------
LIST OF TABLES
(Continued)
Page
Table 18. Production Rates (June 25, 1975) .......... 50
Table 19. Particles Generated From Ultrasonic Thimble
Cleaning, Coulter Counter Analyses ......... 61
Table A-l. Inlet Results ................... A-l
Table A-2. Outlet Results ................... A-3
Table A-3. Orsat Gas Composition Measured at Outlet Stack —
Westmoreland Coal Company, Quinwood, West Virginia. A- 5
LIST OF FIGURES
Figure 1. Process Coal Sieve Analysis ............ 6
Figure 2. Subsieve Size Distribution of Process Coal by
Coulter Counter ..... ............. 9
Figure 3. Subsieve Size Distribution of Process Coal by
Mine Safety Appliance Particle Size Analyzer ... 10
Figure 4. Comparison of Run 2 Thimble Catch and
Associated Scrubber Water Suspension, Coulter
Counter Analyses ................. 12
Figure 5. Comparison of Run 4 Thimble Catch and
Associated Scrubber Water Suspension, Coulter
Counter Analyses ................. 13
Figure 6. Comparison of Run 5 Thimble Catch and
Associated Scrubber Water Suspension, Coulter
Counter Analyses ................. 14
Figure 7. Comparison of Run 6 Thimble Catch and
Associated Scrubber Water Suspension, Coulter
Counter Analyses ................. 15
Figure 8. Particle Size Distributions of Cyclone Catches
Coulter Counter Analyses ............. 22
Figure 9. Sample Point Location ............... 23
Figure 10. Opacity Observation Sites ............. 33
Figure 11. -Run No. 1 Opacity Results as a Function of
Time of Day .................... 34
iv
-------
LIST OF FIGURES
(Continued)
Page
Figure 12. Run No. 2 Opacity Results as a Function of
Time of Day 35
Figure 13. Run No. 3 Opacity Results as a Function of
Time of Day 36
Figure 14. Run No. 4 Opacity Results as a Function of
Time of Day 37
Figure 15. Run No. 5 Opacity Results as a Function of
Time of Day 3£
Figure 16. Inlet Stack Special Adapter With Gate Valve
Showing Position of Alundum Thimble Prior
to Sampling 43
Figure 17. Inlet Stack Geometry Showing Sampling Location
and Sample Point Configuration 45
Figure 18. Outlet Stack Geometry Showing Sampling Location
and Sample Point Configuration 46
Figure 19. Exhausting-Type Fluidized-Bed Thermal Coal Dryer,
Showing Component Parts and Flow of Coal and
Drying Gases, Westmoreland Coal Company, Imperial
Smokeless Division, Quiiiwodd, West Virginia ... 48
Figure 20. Coal Sample Collection Method 52
Figure 21. Schematic of Venturi Scrubber Water Flow System. . 54
Figure 22. Distribution of Dried and Undried Coal Samples
From a Thimble Catch, Coulter Counter Analyses . . 57
Figure 23. Comparison Distribution of Wet and Dry Sieved
Process Coal Samples 58
Figure 24. Comparison of Ultrasonic and Medicine Dropper
Dispersion Techniques, Coulter Counter Analyzer. . 60
Figure 25. Representative Aliquot Check of Large Solution
Volume Having High-Particle Density, Coulter
Counter Analyses 64
Figure H-l. Series Cyclone Sampling Apparatus Block
Diagram H-l
-------
LIST OF FIGURES
(Continued)
Figure H-2. Cut-off Point Versus Flow Rate Calibration
Curves for EPA Series Cyclone H-2
Figure H-3. Cyclone Collection Efficiency Calibration
Data H-3
VI
-------
COAL CLEANING PREPARATION PLANT EMISSIONS
AT WESTMORELAND COAL COMPANY,
QUINWOOD, WEST VIRGINIA
by
P. R. Webb, J. H. Pilcher, and J. A. Gieseke
INTRODUCTION
In accordance with Section 111 of the Clean Air Act of 1970,
the Environmental Protection Agency is charged with the responsibility
of developing standards of performance for emissions from new stationary-
sources (New Source Performance Standards) which may contribute signi-
ficantly to air pollution. A standard of performance developed under
the Act for emissions of air pollutants must be based on the "best emission
reduction systems that have been adequately demonstrated, taking into
account economic considerations.
The development of standards of performance utilizes emissions
data for pollutant sources in the particular industry being studied. In
response to comments on the proposed emission standards by the coal clean-
ing industry, the emission control systems (venturi scrubbers) of the
Westmoreland Coal Company, Quinwood, West Virginia, were selected by
EPA for an emission testing program to provide test data by which the
effectiveness of that emission control system could be compared with other
facilities previously tested.
The Westmoreland Coal Company contested the proposed standards
as a result of their preliminary tests which indicated that their specific
type of coal could present an emissions control problem not previously
considered by EPA. The Westmoreland data implied that the particle size
of the particulate matter entering the venturi scrubber device was finer
than had previously been acknowledged. The data indicated that one type
of coal (Sewell) processed at the Quinwood plant presented different size
characteristics than coal on which the proposed standards were based.
-------
In order to adequately evaluate these allegations, EPA decided to conduct
an extensive test program at the Quinwood facility.
Goal delivered to the plant is crushed to a variety of sizes and then
washed, screened, and separated by both water and concentration tables.
The coal washing occurs in a continuous process which reduces ash content .
from 16 to 3 percent. The 0 x 3/8-inch size of coal of interest in this
study is dewatered by vacuum discs, then dried in a fluidized bed. Coal
drying is accomplished by use of flue gas (produced by a coal-fired
furnace) reduced in temperature by dilution air. The fluidized bed is
contained in a chamber loaded with 40 or more multiclones which separate
the coal from the fine dusts. Dust particles pass from the multiclones
of the dryer through a high-pressure blower to a venturi scrubber, then to
a large mist eliminator, and finally out an 81-inch-diameter stack.
EPA requested Battelie-Columbus to: sample particulate and gas-
eous emissions from the inlet and outlet stacks of the venturi scrubber
control device at the Westmoreland Coal Company, Quinwood, West Virginia.
The test program was conducted from June 18, through June 25, 1975
during the times the plant was processing Sewell coal. High mass loading
at the inlet required the use of an in-stack Alundum* thimble (in con-
junction with an EPA Method 5 train). Tests for particle size
distribution of the inlet particulate were made by EPA utilizing a special
multilpe-cyclone sampler. Grab samples of the process coal and ven-
turi scrubber water were taken simultaneously with the particulate and
gaseous measurements. The purpose of the aforementioned measurements
is to provide EPA with additional data for support of their standards
development efforts.
In order to evaluate the effect of particle size on the venturi
scrubber equipment it was necessary to determine the particle size dis-
tribution of the particulate matter entering the scrubber for each type
coal. Particle size comparison by sieve analysis was conducted on both
types of process coal (Sewell and Pocahontas). Particle sizes of the sus-
pended particulate at the scrubber inlet, suspended particulate in the
* Mention of trade names does not constitute endorsement or recommendation
for use by the Environmental Protection Agency.
-------
scrubber water, and <325 mesh process coal from the sieve analysis were
(I)*
determined by Coulter Counterv ' . A Mine Safety Appliance Particle Size
(2}
Analyzer was also used to determine the particle size of the <325 mesh
particulate in an effort to substantiate the test results as determined
by Coulter Counter. Particle size data, as presented in the figures
throughout this report , were determined from one aliquot per sample unless
otherwise noted.
Since the effectiveness of the venturi scrubber was in question rel-
ative to the process coal particle size, additional effort was placed on
determining venturi inlet particulate size distribution. A special EPA
series cyclone was used to futher evaluate the size cut of the inlet
particles.
Trace metals in Run No. 4 outlet emission samples were determined by
optical emission spectroscopy.
Present opacity of the outlet stack emissions was monitored by two
certified observers during the particulate sampling. The test program
was conducted to determine mass concentration levels and size distribu-
tions of particulate emissions during normal plant operation. Process
conditions were carefully observed and tests were performed only when the
process operations appeared to be operating normally. Process and/or
scrubber malfunctions occurred during Runs 2, 3, and 5 at which time the
particulate sampling was terminated. Opacity readings continued through-
out the entire period including times when high values were observed
because of scrubber malfunction.
A total of five outlet and six inlet runs were made for particulate
loading. Run No. 1 inlet was voided due to incorrect (Pocahontas) coal
being processed. Run No. 2 outlet was not considered valid due to an open
circuit in the probe heater and a leak in the filter which allowed some
particulate to pass through into the impinger catch. Run No. 3 inlet was
voided due to nonrepresentative sampling caused by incorrect placement of
the sampling nozzle and, therefore, incorrect sampling point location.
Inlet Runs No. 4, 5, and 6 and outlet Runs 3, 4, and 5 were essentially
*
References are given on page 65.
-------
simultaneous and considered valid. The following sections of this report
cover the summary of results, process description and operation, location
of sampling points, and sampling and analytical procedures.
-------
SUMMARY AND DISCUSSION OF SAMPLING RESULTS
Particle Size Results
Process Coal Sieve Results
The effectiveness of a venturi scrubber is, among other factors,
related to the particle size of the effluent to be scrubbed. To better
understand and evaluate the final outlet emission results of this study
it is pertinent to obtain a size distribution of the bulk process coal.
Figure 1 graphically depicts the sieve analysis results of two types of
coal processed by the Westmoreland Coal Company. (Run No. 1 is probably
an unknown composite of both Sewell and Pocahontas types of coal.) The
size distributions of the Sewell process coal used during outlet Runs 3,
4, and 5 show very good agreement. As indicated, the Sewell coal has a
mass mean diameter of 1000 pm with approximately 3 to 4 percent being less
than 325 mesh (44 pjn). The Pocahontas coal, as indicated in Figure 1,
is relatively smaller than the Sewell coal having a mass mean diameter
of 250 um, but with approximately 4 percent being smaller than 325 mesh.
Table 1 is a tabular summary of the Sewell and Pocahontas sieve analysis
data.
TABLE 1. SIEVE ANALYSES OF COAL SAMPLES FROM
WESTMORELAND COAL COMPANY
Sieve Size
U.S. Sieve Screen Opening,
Number \m
4
8
16
30
50
100
200
325
Pan (-325)
4,760
2,380
1,190
590
297
149
74
44
<44
Particle Size
Range Collected
per Stage, yjn
>4760
2380-4760
1190-2380
590-1190
297-590
149-297
74-149
44-74
<44
Accumulative Weight-
Percent Retained per
Stage and Smaller Stages
Run 1
100.0
90.3
72.5
54.0
38.5
25.2
fl 16.7
10.8
3.7
Run 3
100.0
92.9
75.8
57.5
41.7
27.9
17.9
11.6
3.1
Run 4
100.0
91.0
72.4
53.4
38.2
25.7
17.1
11.3
2.9
Run 5
99.9
91.5
73.1
53.9
38.3
25.2
16.4
10.5
3.9
Pocahontas
100.0
97.8
92.5
84.2
. 72.2
54.9
36.4
21.8
4.0
^ Run No. 2 discarded _.
(a) Sample size approximately" 900 grama (2~ Ib)
-------
y»
98
95
90
80
? 70
o>
* 60
>.
* 50
*-
§ 40
I 30
on
V- ^u
o
1 10
O
5
2
1
0.5
0.2
O.I
2
•*
/
f
L
¥
/
*
X
X
X
ft
'
x
t*
X
X
^
^
^
^
J.
^
^
_^
V\
H
^
T
^
^1
kx
?
X
X
X*
X
fX
X
I
(
i
\
_fS
^
/
&
-tlS^
JT"
^
X
>
^
^
3 Pocohantusi
3 Run 1 Composite
^ Run 3-1
\. Run 4 Sewell -
^ Run 5-J
0 30 40 60 80 100 200 300 400 600 800 1000 2000 3000 50(
Particle Diameter, micrometers
FIGURE 1. PROCESS COAL SIEVE ANALYSIS
-------
In order to evaluate the proportion of ultrafines (<2 |jjn) that is
of primary importance relative to venturi scrubber effectiveness, Coulter
Counter analyses of the subsieve fraction (<325 mesh) were performed for
the indicated runs. The results are shown in Figure 2. Again the size
distribution for the Sewell coal for outlet Runs 3, 4, and 5 show good
agreement, the mass mean diameter being approximately 17 pm and approxi-
mately 1 to 2 percent being less than 2 pm. The Pocahontas coal has
more fines than the Sewell coal, with a mass mean diameter of approximately
12 p,m and approximately 2.5 to 3 percent being less than 2 pjn. Therefore,
with all other factors equal, the Sewell coal emissions are no more dif-
ficult to collect with venturi scrubber equipment than the Pocahontas coal
emissions.
The Mine Safety Appliance Particle Size Analyzer results as depicted
in Figure 3 essentially substantiate the Coulter Counter results for the
Sewell samples but shows the Pocahontas and Sewell with essentially an
equal size distribution except in the relatively smaller size fractions
where scatter in the data occurs. If one averages the MSA data in the
smaller size fraction, 2 to-3 percent of the size distribution is less
than 2 am; this agrees with the Coulter Counter data of Figure 2. Table
2 is a tabular summary of the Sewell and Pocahontas results for size
measurements with the MSA Particle Size Analyzer.
Particulate Collected by the Scrubber
It is to be expected that the venturi scrubber used to clean partic-
ulate from the effluent gas stream will remove large particles more
efficiently than smaller particles. To directly measure such effects,
mass concentrations and size distributions are required for particles in
the gas stream entering the scrubber and in the scrubber water. Because
of the time averaging effect of scrubber water recirculation, a direct
measurement of particulate mass collected in the water over the inlet
gas sampling time was not possible. The scrubber water samples were col-
lected over a very short time interval and, hence, a time averaged value
for the particulate collected by the water prior to the sampling time
was determined.
-------
TABLE 2. SUBSIEVE PARTICLE SIZE DISTRIBUTION OF WESTMORELAND
COAL COMPANY SAMPLES BY MSA PARTICLE SIZE
ANALYZER(a)
Particle
Size, \m
60
50
40
35
25
20
15
12
10
8
6
4
2
1
0.8
Weight Percent of
Run 1
•• «
100.0
98.1
97.1
79.2
63.3
43.7
33.1
26.8
18.5
12.2
5.3
1.3
0
—
Run 3
— —
100.0
97.6
95.3
79.2
63.4
46.7
31.9
25.8
22.5
15.5
6.4
0.7
0
--
Run 4
« ••
100.0
97.1
94.9
76.1
62.5
47.5
39.1
29.4
21.3
12.9
9.2
5.9
0
--
<44 \m Portion(b)
Run 5
— —
--
100.0
97.2
78.3
65.0
48.3
38.5
30.8
21.7
14.0
8.4
4.9
0.7
0
Pocahontas
100
97
93
91
73
57
42
32
25
17
9
4
1
0
-
.0
.5
.7
.2
.3
.5
.9
.9
.0
.1
.6
.6
.3
-
Weight Percent of Total Sample^)
Run 1
— .
3.70
3.63
3.59
2.93
2.34
1.62
1.22
0.99
0.68
0.45
0.20
0.05
0
--
Run 3
__
3.10
3.03
2.95
2.46
1.97
1.45
0.99
0.80
0.70
0.48
0.20
0.02
0
--
Run 4
•. M
2.90
2.82
2.75
2.21
1.81
1.38
1.13
0.85
0.62
0.37
0.27
0.17
0
--
Run 5
— «•
--
3.90
3.79
3.05
2.54
1.88
1.50
1.20
0.85
0.55
0.33
0.19
0.03
--
Pocahontas
4.00
3.90
3.75
3.65
2.93
2.30
1.72
1.32
1.00
0.68
0.38
0.18
0.05
0
--
Note: Run No. 2 discarded.
(a) Sedimentation liquid—isopropyl alcohol; feed liquid--50 percent isopropyl alcohol, 50 percent heptane.
(b) Data are reported as less than indicated size.
00
-------
.c
*
C
0)
u
^
(U .
a.
0)
3
E
3
o
99.9
99.8
99.5
99
98
95
90
80
70
60
50
40
30
20
10
2
I
0.5
0.2
O.I
O Run 1
£ Run 3—
QRun 4
-J- Run 5-
Pocahontas
0.6 0.8 I
4 6 8 10 20
Particle Diameter, micrometers
40 60
FIGURE"2.' SUBSIEVE"SIZE DISTRIBUTION OF PROCESS COAL
BY COULTER COUNTER
-------
10
o»
"55
c
0>
u
k.
«
a.
0>
£
3
o
)9.8
9.5
99
98
95
90
80
70
60
50
40
30
20
10
5
2
1
0.5
0.2
O 1
.^
£
m
\
/£
^ L
»-
1
3
\
\
^A
/ I
r
j
j
' i
Y
J
c
'I
t>
J
^
i
S
*
/'
^
j«
x
1
r
]
Q Run 1
/^ Run 3-,
•^
V
a i
Run 4
Run 5-
Pocahoi
/
i
\
r
Composite
— SewelT
atas
0.6 0.8
4 6 8 10 20
Particle Diameter, micrometers
40
60
FIGURE 3. SUBSIEVE SIZE DISTRIBUTION OF PROCESS~~COAL BY"
MINE SAFETY APPLIANCE PARTICLE SIZE ANALYZER
-------
11
However, the samples were collected to determine if any differences be-
tween inlet suspended particulate and scrubber water particulate existed.
These differences could lead to significant inferences regarding scrubber
collection efficiency as dependent on particle size.
The size distribution curves given in Figures 4, 5, 6, and 7 allow
comparisons between suspended inlet particulate and scrubber collected
particulate for Runs 2, 4, 5, and 6, respectively. Inlet Runs 1 and 3 were
voided as explained in the Introduction. The results may appear to be
inconclusive because of the data in Figures 4 and 5» which show less partic-
ulate below about 2 \m in the scrubber water than in the inlet gas, and
Figures 6 and 7, which show the reverse. Table 3 is a tabulation of these
data. Within the error of the particle sizing methods, the data show
that the particle size distribution in the scrubber water is practically
the same as the distribution at the scrubber inlet. This result could
be predicted when the scrubber collects essentially all of the particulate
matter. Data of this type cannot be used to obtain size dependent col-
lection, efficiencies. Any major impact on control effectiveness would
occur only if a sufficiently large fraction of the particulate material
had sizes too small for efficient collection. Size distributions for the
suspended inlet particulate indicate that only about 5 percent of the particles
have sizes below about 2 pm. This low fraction in the small size range
makes comparison impractical since the small sizes constitute a nearly
negligible portion of the particulate for measurement purposes. If, how-
ever, the differences between the curves are attributable to normal
scatter in the experimental data, as is likely, it can be concluded that
the overall efficiency is not sensitive to low collection efficiency for
these small sizes. This is because such a small fraction of the total
entering particles are smaller than the size where efficiency drops off.
Therefore, valid conclusions that can be drawn from these data are that
the venturi scrubber essentially collects all sizes of particulate in
excess of 2 yjn. The amount of particulate matter less than 2 micrometers
that reduces the scrubber's collection efficiency is not significantly
different for Sewell coal as compared with Pocahontas coal.
-------
12
TH-2
Scrubber Water
0.6 0.8 I
4 6 8 10 20
Particle Diameter, micrometers
40 60
FIGURE 4. COMPARISON OF RUN 2 THIMBLE CATCH
AND ASSOCIATED SCRUBBER WATER SUSPENSION,
COULTER COUNTER ANALYSES
-------
13
TH-4
Scrubber Water
0.6 0.8 I
4 6 8 10 20
Particle Diameter, micrometers
40 60
FIGURE 5. COMPARISON OF RUN 4 THIMBLE CATCH AND
ASSOCIATED SCRUBBER WATER SUSPENSION,
COULTER COUNTER ANALYSES
-------
14
0.6 0.8 I
4 6 8 10 20
Particle Diameter, micrometers
40 60
FIGURE 6. COMPARISON OF RUN 5 THIMBLE CATCH AND
ASSOCIATED SCRUBBER WATER SUSPENSION,
COULTER COUNTER ANALYSES
-------
15
TH-6
Scrubber Water
0.6 0.8
4 6 8 10 20
Particle Diameter, micrometers
40 60
FIGURE 7. COMPARISON OF RUN 6 THIMBLE CATCH AND
ASSOCIATED SCRUBBER WATER SUSPENSION,
COULTER COUNTER ANALYSES
-------
TABLE 3. SIZE DISTRIBUTION COMPARISON OF PARTICLES ENTERING
THE SCRUBBER AND PARTICLES CAUGHT BY THE SCRUBBER
WATER(a), FROM WESTMORELAND COAL COMPANY
(Percent less than indicated size)
Particle
Size, (j,m
44.2
35.1
27.8
22.1
17.5
13.9
11.1
8.8
7.0
5.5
4.4
3.5
2.8
2.2
Run
Thimble
Catch
_-
--
--
100.0
91.87
91.13
81.52
70.98
63.11
53.00
41.33
28.89
17.68
9.48
2
Scrubber
Water
--
100.0
94.81
94.29
89.61
85.58
78.11
70.02
60.80
50.43
46.38
29.87
17.67
7.57
Run
Thimble
Catch
--
100.0
98.89
98.34
95.43
91.20
83.37
75.48
66.11
53.83
40.70
26.95
14.82
9.53
4
Scrubber
Water
100.0
97.88
96.82
94.70
91.25
85.94
79.47
70.03
60.75
51.00
38.41
26.71
14.15
5.93
Run
Thimble
Catch
--
—
100.00
98.27
97.11
92.78
86.78
79.04
69.67
57.55
40.95
24.51
12.36
5.54
5
Scrubber
Water
--
—
—
100.00
98.33
93.13
86.47
80.63
71.78
60.53
48.15
34.07
20.47
10.02
Run
Thimble
Catch
--
100.0
97.12
94.26
92.79
85.58
78.42
71.14
61.73
44.97
29.89
16.54
9.03
4.90
6
Scrubber
Water
--
100.0
98.96
93.24
90.90
81.54
74.45
67.52
58.60
48.65
38.03
27.83
14.83
7.61
(a) Size distribution of thimble particle catch and scrubber water particle suspension determined
by Coulter Counter using an aperture diameter of 100 pjn, and a manometer volume of 500 y&.
-------
17
Cyclone Catch Analyses
Tests were conducted using a sample train consisting of two aluminum
cyclones followed by a glass fiber filter. The purpose of the tests was
to determine the mass loading of particulates less than 2.5 \ua in diam-
eter. All tests were conducted at the inlet to the scrubber. The testing
cyclones were developed by an independent contractor for the Environmental
Protection Agency. Since the equipment had not been previously used by
the tester, numerous problems were encountered until a sampling technique
was established. A total of 13 runs was conducted. During the first six
runs, problems were encountered with the testing; Runs 7-13 were valid
tests. However, during Runs 8, 10, and 12, the process was operating under
a light load. Furthermore, Run 13 was conducted when Type 7 Sewell sur-
face-mined coal was being processed.
Tables 4 and 5 summarize the test data and analytical data. Table 6
summarizes the test results. The tests were conducted at a nominal flow
rate of 1 acfm, yielding cut-off points of 2.5 and 1.0 \aa for the first
and second cyclones, respectively. Therefore, the particulate less than
2.5-p,m diameter was calculated as the total mass minus the mass collected
in the first cyclone.
Two tests with a cascade impactor were attempted, but were not suc-
cessful. In one case, the stages were overloaded, and in the other case,
the collection substrates were knocked ajar.
Table 6 summarizes the test results. As previously noted, Tests 7-13
were considered the best tests as far as sampling procedures are concerned.
In performing the calculations, an average stack gas moisture (9.9 percent)
fUitA-
calculated from the five Alundum thimble/flwas used. The results presented in
Table 6 show a reasonable correlation between the Alundum thimble and cyclone
methods for determining particulate concentration of particles less than
2.5 ym diameter. For cyclone Runs 7-13 the mass percent less than 2.5 ym
ranged from 11.6 to 17.6 percent with an average of 14.5 percent. For
the Alundum thimble Runs 2, .4, 5, and 6, the mass percent less than 2.5 ym
ranged from 7 to 15 percent with an average value of 11.5 percent.
-------
TABLE 4. TEST DATA FOR PARTICLE SIZE RUNS WITH MULTIPLE-CYCLONE SAMPLER
Run
Number
1
2
3
4
5
6
7
8
9
10
11
12
13
Dace
6/18
6/19
6/19
6/20
6/20
6/23
6/23
6/25
6/25
6/26
6/26
6/26
6/26
Sample
Start Port Period
Time Identification (minutes)
1800
1415
1715
1115
1445
1700
2000
1005
1255
1650
1225
1420
1550
S
S
S
E
E
E
E
E
E
S
S
E
S
22
10
11
11
11
11
11
11
11
11
11
11
11
Meter
Volume
(acf)
22.345
8.000
10.520
7.230
9.050
8.675
8.600
8.520
8.525
—
7.895
8.600
8.715
Stack
Temp.
(F)
190
190
170
170
170
200
200
155
150
170
180
180
180
Nozzle
Diameter
(inches)
0.4375
0.4375
0.4375
0.4375
0.4375
0.1875
0.1875
0.1875
0.1875
0.1875
0.1875
0.1875
0.1875
Absolute
Stack
Pressure
(in. Hg)
29.68
29.80
29.80
29.80
29.80
29.81
29.81
29.74
29.74
29.74
29.63
29.63
29.63
Volume
Sampled
(dscf)
20.16
7.24
9.52
6.54
8.20
7.84
7.75
7.68
7.68
—
7.08
7.72
7.82
Sample
Rate , .
(acfm)(a)
1.3
1.0
1.1
0.8
1.0
1.0
0.9
0.9
0.9
—
0.9
1.0
1.0
Percent
Isokinetic
(b)
(b)
(b)
(b)
(b)
(b)
77.8
78.5
78.5
—
85.2
78.1
77.0
(a) Stack conditions.
(b) Used incorrect sample nozzle.
00
-------
TABLE 5. ANALYTICAL DATA FOR PARTICLE SIZE RUNS
WITH MULTIPLE-CYCLONE SAMPLER^)
Test
1
2
3
4
5
6
7
8
9
10
11
12
13
First
Cyclone
__
0.3349
1.7575
3.9237
3.6606
1.3006
1.1888
0.5013
1.2658
0.6590
1.6196
1.0512
0.5947
Second
Cyclone
__
N/A
N/A
N/A
2.2320
0.1659
0.1346
0.0535
0.1166
0.0752
0.2217
0.1151
0.0771
Filter
._
0.0620
0.1144
0.1266
0.0229
--
0.0450
0.0278
0.0383
0.0234
0.109
0.000
0.0386
First
Cyclone
Wash
__
N/A
N/A
N/A
N/A
N/A
N/A
N/A
0.0060
0.0035
0.0081
0.0059
0.0048
Second
Cyclone
Wash
__
N/A
N/A
N/A
N/A
N/A
N/A
N/A
0.0076
0.0062
0.0089
0.0172
0.0098
Filter
Assembly
Wash
_ _
N/A
N/A
N/A
0.0085
--
0.0073
0.0048
0.0048
0.0040
0.0086
0.0070
0.0028
First
Cyclone
Total
__
0.3349
1.7575
3.9237
3.6606
1.3006
1.1888
0.5013
1.2718
0.6625
1.6277
1.0571
0.5995
Second
Cyclone
Total
__
N/A
N/A
N/A
2.2320
0.1659
0.1346
0.0535
0.1242
0.0814
0.2306
0.1323
0.0869
Filter
Total
__
0.0620
0.1144
0.1266
0.0314
—
0.0523
0.0326
0.0431
0.0274
0.1176
0.0070
0.0414
Total
_-
0.3969
1.8719
4.0503
5.9241
--
1.3757
0.5874
1.4391
0.7713
1.9759
1.1964
0.7278
(a) All masses are reported in grams.
N/A Not Applicable.
-------
20
TABLE 6. SUMMARY OF RESULTS OF
PARTICLE SIZE RUNS-
Run Number
Cyclones:
1
2
3
4
5
6
7
8
9
10
11
12
13
Alundum Thimbles:
1
2
4
5
6
Mass
Total
---
.84
3.02
9.54
11.13
--
2.73
1.18
2.90
--
4.30
2.39
1.43
2.73
1.79
2.89
3.23
2.67
Loading, gr/scf
Less than
2.5 nm(a>
--
0.14
0.19
0.30
4.25
--
0.37
0.17
0.34
'
0.76
0.28
0.25
--
0.27
0.35
0.39
0.19
*Mass Less
than 2.5 pm,
percent
«• «
15.6
6.1
3.1
38.2
--
13.6
14.7
11.6
--
17.6
11.6
17.6
--
15
12
12
7
(a)
(a) Total mass, less mass collected in first cyclone.
-------
21
Figures 8 presents the particle size distribution of the cyclone
catches which were analyzed by electronic counter. Field data sheets are
presented in Appendix C. Cyclone sample identification and analytical
results are presented in Appendix G.
Figure 9 indicates the location of the sampling points. Because of
the design of the equipment only the first 22 points on each traverse
could be sampled. Each test (Runs 6-13) was conducted using two cyclones
in series with a back-up filter. The sampling rate was kept constant
during each test at a nominal flow of 1 acfm. Each test was conducted
across only one traverse and was only 11 minutes in duration. It was
necessary to limit each test to only one traverse because of the capaci-
ties of the cyclone collection cups.
Prior to each test the sample train was checked for leaks and then
inserted into the stack (with nozzle plugged) in order to allow the cyclones
to heat to stack temperature. After a 20-minute pre heat, the cyclones
were removed, the nozzle unplugged, the cyclones reinserted, and the test
begun.
The sample train was constructed so that the nozzle was inserted
directly into the first cyclone. The first cyclone was immediately fol-
lowed by the second cyclone and a back-up filter connected to the probe.
During sampling, both cyclones and filter were in-stack. A flexible hose
was used to connect the probe to a silica gel impinger, vacuum pump, and
dry gas meter, in that order. Figure H-l of Appendix H is a block diagram
of the cyclone and filter assembly. Figure H-2 shows the calibration
curves for the cyclones. Figure H-3 shows the calibration curves for each
cyclone and relates the particle cut-off size, |j,m, as a. function of flow
rate.
The particulate catch of each cyclone was emptied from the collection
cup into a weighing vessel. All residual material was brushed into these
vessels using a camel hair brush. In addition, for Runs 9-13, acetone
rinses were conducted on the cyclones and filter holder.
-------
60.0
First cyclone
Run 2
Run 7 •
Run 9 A
Run II «
Second cyclone
Run 8
Run 7 o
Run 9
Run II a
K)
to
0.01
O.I
0.5 I 2
10 20 30 40 50 60 70 80 90
Percent Mass Less Than D
95 98 99
99.9 99.99
FIGURE 8. PARTICLE SIZE DISTRIBUTIONS OF CYCLONE CATCHES,
COULTER COUNTER ANALYSES
-------
23
South
Port
Traverse points:
1
2
3
4
5
6
7
8
9
10
11
12
1 inch
2.25
4
5.75
7.5
9.5
11.5
14
16.5
19.5
23.25
28.75
[
13
14
15
16
17
18
19
20
21
22
*23
*24
East
Port
43.25
48.75
52.5
55.5
58
60.5
62.5
64.5
66.25
68
69.75
71
•X
*Not Sampled,
FIGURE 9 . SAMPLE POINT LOCATION
-------
24
For several cyclone catches, the particle size distribution of
the collected material was determined using an electronic counter. The
sizing procedures used were the same as those employed for the Alundum
thimble particulate catches.
Scrubber Inlet Measurement Results
Tables 7 and 8 show computer output tabular summaries (in English
and metric units, respectively) of results of Runs 1 through 6, Run 3 being
aborted because of operator error. Runs 1 and 2 are not considered representa-
tive relative to outlet comparison but are being presented only as support
data to aid in ascertaining a better understanding of inlet sampling conditions.
Weight concentration for Runs 4, 5, and 6 ranged from 2.68 to 3.23 gr/dscf
(6144 to 7407 mg/Nm3 ) resulting in a mean of 2.93 gr/dscf (6728 mg/Nm3 ).
Isokinetic sampling for Runs 4 to 6 varied from 100 to 106 percent, in spite
of the undesirable working conditions encountered, which is well within the
desired limits of 100 _ 10 percent for producing valid data.
Scrubber Outlet Results
Tables 9 and 10 show summarized data by computer (in English and metric
units, respectively) obtained at the outlet stack of the Westmoreland Coal
Company. Runs 1 and 2, as previously mentioned, are not considered representative
because of plant and/or BCL operator error, but have been presented as
additional support data.
For Runs 3 to 5, the dry catch (probe and filter) weight con-
centration results are 0.037, 0.059, and 0.035 gr/dscf (86.1, 134.5, and
80.6 mg/Nm3 , respectively. The impinger catch results for Runs 3 to 5
increased the weight concentrations an average of 25 percent, the total
catch (probe, filter, and impingers) being 0.054, 0.072, and 0.049 gr/dscf
(123.2, 164.6, and 112.9 mg/Nm3 ) , respectively.
-------
TABLE 7. INLET MEASUREMENT RESULTS SUMMARY (ENGLISH UNITS)
WESTMORELAND COAL COMPANY INLET RESULTS
RUN NO. ^
TEST DATE
VOLUME OE GAS SAMPLED, DSCF^'
PERCENT MOISTJRE BY VOLUME
AVERAGE STACK TEMPERATURE, E
STACK VOLUMETRIC FLOW RATE, DSCFM(C)
STACK VOLUMETRIC FLOW RATE, ACFM(d>
PERCENT ISOKINETIC
PERCENT EXCESS AIR
PERCENT OPACITY
FEED RATE, TONS/HR
PARTICULATES - PROBE, CYC, FILTER CATCH
MG
GR/nSCF -
GR/ACF -
LR/HR -
LB/TON FEED
__1
6/ifl
ua.i
It. 2
166
136913
183212
98.6
1072
NA
ND
181t»i*.0
2.729
2.018
3201. 3
ND
2
6/19
88.8
7.0
16«»
l«»8680
191155
9«».6
1785
NA
ND
9271.0 1
1.785
1. 387
2273.3
ND
ND - No Data (a) Run No> 3 aborted<
(b) Dry standard cubic feet at 70° F, 29.92 in. Hg.
(c) Dry standard cubic feet per minute at 70° F, 29,
(d) Actual cubic feet per minute.
(e) Average includes only Runs 4, 5, and 6.
J» _
6/20
90. t«
9.1
171
1<*(«0<»1
190091*
100.0
2«»11
NA
ND
5368.0
2.888
2.187
356<».l
ND
.92 in. Hg
5
6/23
li»0.0
10.7
202
138019
193875
106.1
2267
NA
ND
26167.0
3.225
2.29**
3812.9
ND ..
NA - Not
•
6
6/25
126.6
12.1
181
135899
1880Q(*
99.7
1595
NA
ND
20087.0
2.675
1.932
311U.2
ND
Applicable
(e\
Average v '
119.0
10.6
185
137986
190657
101.9
2091
NA
ND
20540.0
2.929
2.138
3497
ND
Ul
-------
TABLE 8. INLET MEASUREMENT RESULTS SUMMARY (METRIC UNITS)
WESTMORELAND COAL COMPANY
RUN NO.
TEST DATE
VOLUME OF GAS SAMPLED, NCM
PERCENT MOISTURE RY VOLUME
AVERAGE STACK TEMPERATURE,
STACK VOLUMETRIC FLOW RATE
STACK VOLUMETRIC FLOW RATE
PERCENT ISOKINETIC
PERCENT EXCESS AIR
PERCENT OPACITY
FEED RATE, MTON/HR
PARTICULATES - PROBE, CYC, F
MG
MG/NCM -
MG/CM -
KG/HR -
KG /MTON FEED
INLET RESULTS
124
6/18 6/19 6/20
(a) 3.3 2.5 2.6
11.2 7.0 9.1
C 74 73 77
/v \ _ .. _ . _ .
, NCMMV ' 3862 4194 4063
, CMM(C) 5168 5392 5362
98.6 94.5 100.0
1072 1786 2411
NA NA NA
ND ND ND
ILTER CATCH
18144.0 9271.0 15368.0
6269.3 ff099.4 5634.4
U 680 .7 1185^7 5022.5
1452.1 1031.2 1616.7
ND ND ND
5
6/23
3.9
10.7
94
3893
5469
106.1
2267
NA
ND
26167. 0
7407.2
5268^2
1729.5
ND
6
6/25 Average
3.6 3.4
12.1 10.6
82 84
3833 3930
5303 5378
99.7 101.9
1595 2091
NA NA
ND ND
20087.0 20540
6144.3 6728.6
4437.2 4909.3
1412.6 1586.3
ND ND
ND - No Data.
(a) Normal cubic meter (dry) 20.0 C, 760 mm Hg.
(b) Normal cubic meter per minute (dry) at 20.0 C, 760 mm Hg.
(c) Actual cubic meters per minute.
NA = Not Applicable
CT>
-------
27
TABLE 9. OUTLET MEASUREMENT SUMMARY RESULTS (ENGLISH UNITS)
HFSTKnPFLftMn. rOAL rOSPAMY flUTl FT O^SULTS
SUN NO. 1 2 •*
TEST TATE 6/18 6/19 6/?0
VOLUME OF GAS SAHPLFD, nSCF(b) 109.5 100.5 106.7
PERCENT MOIST'JRF BY VOLUME It. 8 1 1 . l« 11. 1.
avF^aGE srac< TEMPEPATU°E, F us IDS . no
STACK VOLU^ET = IC FLOW RATE, QSCFn(c> "~ 1325U3 125531. ~ 139320
STACK VOLUMFT°IC FLOW RATE, &CcM 20 ?5
FEFO P!\TE, TOf.'S/HP ND ND ND
PflPTICULATES - PROnF, CYC, FILTER CATCH
M3 . .. ._ _._. 2 '».'*.'« 32_P.l 239.6
G9/1SCF - .. O.Q"?8 0.055 0.037
GP/ACF - ..Q.JJ.29 O..0.if_2 n.0?9
LR/TON CEEH ND ND ND
PARTICIPATES - TOTAL CATCH
HG 299.7 Ui.a.2 31.2.9
GR/'JS'CP - O.di.7 0.076 0.051.
GP/ftCF - 0.035 0.058 O.Qi.1
LB/H9 - 5?. 9 81.5 6<*.0
LR/TON Fprr) ND jjp ND
PFDCFNT TMPTN:E° CATCH 18.5 26. a 30.1
«. 5
6/23 6/?5
109.2 10«..0
13.2 11.9
116 112
135983 137815
181.17? 182969
103. R 97.1
2?67 15^5
13 20
ND ND
_376.9 ..21 1». 2
0.05C> 0.035
0.0d3 P..026
68.2 Ul. if
ND !!D
'•fSl.if 300.2
0.07? O.DI.O
0.053 0.037
83.5 58.0
ND ND
18.3 28.6
Average
106.6
12.2
113.3
137629
183457
100.6
2091
19
ND
276.9
0.044
0.033
51.4
ND
36S.2
0.058
0.044
68.5
ND
" 25.7
ND - No Data.
(a) Dry standard cubic feet at 70° F, 29.92 in. Hg.
(b) Dry standard cubic feet per minute at 70° F, 29.92 in. Hg.
(c) Actual cubic feet per minute.
(d) Average includes only Runs 3, 4, and 5.
(e) Opacity measured only during portion of run.
-------
28
TABLE 10. OUTLET MEASUREMENT-RESULTS' SUMMARY (METRIC UNITS)
WESTMORELAND CO<'«L COMPANY OUTLET .". 30.1
1* 5
'" "6/23 " ~ ' 6V?5 " Average
"3.1 ~2.9 " 3.0
13.2 11.9 12.2
46 I* 1. 44
3836 38fl7 3S84
5195 5161 5171
103.8 97.1 100.6
2267 1595 2091
13 20 19
' ND ND ND
376.9 211*. 2 276.9
131*. 5 flO.6 100.4
99.2 60.6 75.1
30.9 13.8 23.3
ND ND ND
1.61.1* 300.2 368.2
16i*. 6 112.9 133.6
121.'. 85.0 100.1
37.9 26.3 31.1
ND ND ND
IB. 3 28.6
ND - No Data.
(a) Normal cubic meter (dry) at 20.0 C, 760 mm Hg.
(b) Normal cubic meters per minute (dry) at 20.0 C, 760 mm Hg.
(c) Actual cubic meters per minute.
-------
29
Inlet and outlet volumetric flow rates compare reasonably well
when considering the relatively poor inlet sampling location. Insufficient
downstream stack diameters from the inlet sampling location did not allow
sufficient time to establish a uniform flow profile; the end result being
a poor velocity profile which could be a reasonable explanation for the
inlet and outlet volumetric flow rate difference.
Moisture condensation in the outlet stack and reentrainment of water
droplets resulted in the effluent gases being supersaturated and, thus, the
stack gas moisture content of the gases was greater than the moisture cor-
responding to the dew point.
Complete particulate results by computer, which include additional
operational data, and sample calculations for outlet Run No. 4 can be seen in
Appendix A.
Total Solids in Scrubber Outlet Water
Solids in the scrubber water samples were determined by passing
a 50-ml aliquot of the scrubber water through a tared 45-mtn glass fiber filter.
The sample was then dried to a constant weight. Table 11 shows the results
for the indicated runs.
TABLE 11. SCRUBBER WATER ANALYSIS
Outlet
Run No.
2
3
4
5
Total Solids,
percent
7.64
8.69
7.74
6.94
-------
30
Opacity Measurements
Visible emissions were observed by two BCL certified smoke
readers simultaneous with particulate sampling. During the five runs,
opacity values ranged from 10-30 percent. Particulate sampling was
terminated during periods of time when the venturi scrubber was malfun-
tioning. Opacity readings for the most part were taken during all periods.
A complete summary of the data is presented in Appendix B-l and B-2. Tables
12 and 13 are computer summaries of Appendices B-l and B-2 and list the
frequency each observer exceeded the stated opacities.
In order to make it easier to interpret this large amount of
data a graphic presentation of each run has been prepared and is presented
in Figures 11 through 15. Each graph shows the percent opacity as a
function of time of day for the total reading period for each test. Runs
3 and 5 have been divided into A and B sections because of relatively long
interruptions in opacity observations and also to facilitiate computer
data reduction. Also shown on the graphs, by means of horizontal arrows,
are the time periods during which "Particulate Sampling" occurred. These
graphs make it possible to quickly determine the ranges of visible emissions
during particulate sampling times for each of the five test runs. Because
opacity readings were continued when scrubber malfunctions occurred, abnormal
values of visible emissions may have been observed. Emphasis should be
placed only on the visible emission data obtained during particulate sampling.
Figure 11 shows the relative positions of the observation sites with respect
to the outlet stack and adjacent structures. Site B was used for opacity
reading in the mornings and Site A in the afternoons.
-------
TABLE 12. FREQUENCY OBSERVATION AVERAGES EXCEEDING STATED
OPACITY FOR PARTICULATE SAMPLING ONLY
F^£r
~~uX I
30 X
25X
' 2CX
~ i ex
1CX
5X
OX
FRSO
Run No. 1
)U£NCY OBSERVE? AV,_°2GE
CUOS STATtO 0°ACITYI
PLCH"? SNYOP
0 u
S j
C 2
3 5
6 7
7 7
7 7
Run No. A
USSCY OiiScRVE0. AV-:RAGr
IfXCEKOS STfiTf H 0°ACITYI
"~3 0 X "
25X
" 2 OX"
16X
1CX
5X
ax
PLCHR SMYOR
" o ' ' " 3
0 Q
o i" "
f 7
21 20
21 22
21 22
Run No. 2 Run No. 3A Run No> 3B
F'JlHENcr r->s;^vr? AV'^AOC FI-^UCNCY rnr>i?vtR AV^^AGF FFraur,NCY ns'^v^' AVERAGE
t X E 10 2tX It 3 25X C 0
2CX 8 1*. 20X " 17 17 " 2cy. 0 3
15X 17 20 15X ifl 19 15X 5 i.
1C/. Z< M . IPX 18 20 iOX 5 U
i"< 22 22 5>X 18 20 5 X 5 t
OX 22 22 fX 18 20 " OX 5 <.
U>
l-1
Run No. 5A Run No. 5B
FRrQOfNCY OBSrPV"=> AU^PflGP FPcQUENCY ris:?v£.^ Ayc^AGt >
CXCiTLOS STATLn 0PACITYJ LXCc.c.OS STATED OPACITY!
PLCHP ' "st.-YO* PLCH^ SNYOR"
3QX 0 0 30X 0 0
20X 0 0 25 X 1 5
2 ax 01 20 x 6 ii
ItX 0 tt 15X 9 13
IOX 3 10 1CX 10 13
5X S 10 " 5X ~" 10' 13"
OX 6 10 OX 10 13
-------
TABLE 13. FREQUENCY OBSERVATION AVERAGES EXCEEDING
STATED OPACITY FOR.TOTAL OPACITY DATA
_
Run No. 1
FHECUENCY OBSErtvEH AVERAGE
EXCEEDS sTAtto OPACITY:
PLCHC SNYDH
30* 00
• "255 0 0
20% 0 2
15% 35
10% 6 7
5% 7 7
0% 7 7
Run No. A
FHECUE^CY OHSEHVER AVERAGE.
EXCEEDS STATED OPACITYI
PUC^R SNYDH
30% 3 0
20% 3 2
15% 4 9
10% 19 23
5» 3* 35
0% 35 35
Run No. 2
FHEUUENCY TOBSEHVEH AVERAGE
EXCEEDS STATED OPACITY:
PLCHR SNYDH
30% 2 6
25% 6 11
20% 9 17
15% 22 24
10% 27 26
5% 27 .26
0% 27 26
Run No. 5A
_jF£iQ«l- NCY Pjr-'PUF? AV^=?AG£
t XC-
3 OX
25X
2 1 X
1SX
10 X
5X
OX
•::os sia
fLCHr
c"
2
1<»
2ST
*9
29
29
T:J OPACITY:
SNVQR
i
c
25
26
30
30
30
Run No. 3A Run No> 3B
crIO"-iCY r^:r:v'= AVE^AG^ F^OU-.NCY r->s--^vfp Av^ar.f
f.xjrr..i5 s-u-:o CFACITYI LXJ.^OS STUID CFACITY:
^LCH-^ SMYOR PLCHR SNvnp
30 X 11 0 30 X 0 0
25X 15 i» 25X 0 o
2&X 23 19 20 X 6 ' 5
15'. ^s <;^ 15 x
Run No. 5B
FxcE^ns STATTH OPACITY:
PLC^? SNYDK
30 X 2 2
2 5 X 3 2
2CX "1 V
15X 6 7
10X 15 l«t
5X 15 1*«
C X 15 1 <,
-------
33
Observation
•Site A
Hillside
400
76-8"
120
5
^
0'
f i
i
f>r
k
1
41
1
Observation
Site'B
Hillside
Cross-Sectional View
2
3
Coal preparation;
cleaning and drying
process area
Venturi scrubber inlet stack
Effluent outlet stack
Observation
Site B
Top View
FIGURE 10. OPACITY OBSERVATION SITES
-------
34
50
40
§ 30
o>
CL
(J
o
Q.
O
20
10
Run No. I
Particulate sampling
Sundown - no further readings
of visible emissions
18
19
20 21
Time of Day, hours
22
23
FIGURE 11. RUN NO. 1 OPACITY RESULTS AS A FUNCTION OF TIME OF DAY
-------
35
50
40
30
0)
Q.
O
o
O.
20
10
Run No. 2
Particulate sampling
15
•((Run started at 14:48)
16
17 18
Time of Day, hours
19
FIGURE 12. RUN NO. 2 OPACITY RESULTS AS A FUNCTION OF TIME OF DAY
-------
50
40
§ 30
i_
0)
ex
o
o
O 2°
10
Particulate
samplings
Run No. 3A data
I
Run No. 3
Particulate
sampling
Run No. 3B data
10
II
12 13 14
Time of. Day, hours
15
16
17
FIGURE 13. RUN NO. 3 OPACITY RESULTS AS A FUNCTION OF TIME OF DAY
-------
37
50
40
u
i_
0)
a.
30
o
o
£ 20
10
15
16
Particulate samplings
_L
17 18
Time of Day, hours
Run No. 4
19
20
FIGURE 14. RUN NO. 4 OPACITY RESULTS AS A FUNCTION OF TIME OF DAY
-------
50
40
30
o>
Q.
O
Q- 20
O
10
J
10
Run 5A
Particulate
/samplings
KH
i
Run No. 5
II
12
Run5B
13 14 15
Time of Day, hours
Particulate
• i
sampling
16
17
00
18
FIGURE 15. RUN NO. 5 OPACITY RESULTS AS A FUNCTION OF TIME OF DAY
-------
39
Data Applicable to All Five Runs
The following data, required by the EPA "Guidelines for Pre-
paring Summary of Visible Emissions", apply to all five of the test runs.
Type of Plant; Coal cleaning
Type of Discharge; Emissions from thermal dryer
Location of Discharge; Stack above dryer
Height of Point of Discharge; 76 feet, 8 inches
above ground level
Distance from Observer to Discharge Point;
' Site A: 400 ft northwest of stack
Site B: 120 ft southeast of stack
Height of Observation Point;
Site A: 50 ft above base of stack
Site B: 40 ft above base of stack
Direction of Observer from Discharge Point;
Site A: northwest of stack (in afternoon)
Site B: southeast of stack (in morning)
The remaining data called for in the EPA guidelines are different
for each run and are presented in Table 14. An examination of Figures 11
through 15 show the opacity ranges as follows.
Test Run 1 - Opacity ranged from about 8 to 18 percent
during the 50 minutes of observations prior
to sundown.
Test Run 2 - Opacity ranged from about 14 to 31 percent
during time of particulate sampling which
extended from 1448 to 1736 hours.
Test Run 3A - Opacity ranged from 22 to 31 percent.
Test Run 3B - Opacity ranged from 15 to 20 percent.
Test Run 4 - Opacity ranged from about 7 to 18 percent
during the times of particulate sampling
which were 1552 to 1655 hours and 1811 to
1917 hours.
Test Run 5A - Opacity ranged from about 17 to 25 percent
during the first two particulate sampling
periods .
Test Run 5B - Opacity ranged about 12 percent at the
start of the third particulate sampling
period.
-------
TABLE 14. OPACITY OBSERVATION SUMMARY
Description Description Wind
Run Date 24 Hr. Clock Observation of of Wind Velocity,
No. (1975) Time Site Background Sky Direction mph Color of Plume
Comments
Duration of
Observation
1/18/75 1809-1859
Sky
20% clouds from NNW
2 6/19/75 1510-1850 A
3A 6/20/75 0940-1300 B
3B 6/20/75 1504-1628 A
4 6/23/75 1530-1930 A
5A 6/25/75 1004-1400
5B 6/25/75 1430-1715
Sky
50% clouds from SW
Sky Overcast from NE
(100% clouds)
Sky Overcast from NE
(100% clouds)
0-5 Black with steam
plume near stack
5-10 Black with steam
plume near stack
0-5
0-5
Sky Overcast from NW 0-5
(100% Clouds) but variable
Sky Slightly over- from NE 0-5
cast
(107. clouds) .
Sky Slightly over- from NE 0-5
cast
(10% clouds)
Black with steam
plume attached
to stack
Black with steam
plume attached
to stack
Black with steam
plume attached
to stack
Black with steam
plume attached
to stack
Black with steam
plume attached
to stack
Black plume obscured
by attached steam
plume sometimes ex-
tending 200 ft above
stack
Black plume obscured
by attached steam
plume up to 75 ft
above stack
Black plume obscured
by attached steam
plume extending 100
to 200 ft above stack
Black plume obscured
by attached steam
plume extending 50
to 100 ft above stack
Black plume obscured
by attached steam
plume extending 75
to 250 ft above stack
Steam plume extended
50 to 125 ft above
stack
Steam plume extended
50 to 125 ft above
stack
50 rain;
discontinued
at sundown
(7:00 p.m.)
3 hrs, 40 mia
3 hrs, 20 min
1 hr, 24 min
4 hrs
3 hrs, 56 min
2 hrs, 45 min
(a) Moved to Site A at 1310 hours.
-------
41
Trace Metal Results
The catch from Run No. 4 was analyzed for trace metals and the
results are shown in Table 15. For the most part these data are considered
to be representative of what one might expect from coal. The results must
be interpreted with consideration of mass for each sample. The ppm values,
as presented, tend to be misleading if one does not relate the indicated
ppm values to the associated sample mass(as presented at the bottom of the
table) to obtain the mass per sample of each element. The major constituents
based on the mass collected on probe and filter catch are iron, potassium,
sodium, and zinc. Relatively small amounts of other elemental constituents
were picked up in the acetone rinse of the impingers but are insignificant
when considering their relative mass. Silicon, being the predominant metal,
is most likely present due to its occurrence in coal ash. However, some
portion of the silicon may represent fragments of filter picked up during
the probe-and-filter-holder acetone rinse.
LOCATION OF SAMPLING POINTS
Inlet to Ventu'ri Scrubber
Particulate samples were collected from the scrubber inlet gases
using an Alundum thimble in conjunction with a standard EPA Method 5
rig. A rather large port consisting of a 6-inch-diameter nipple was
necessary to accept the special EPA series cyclone used to classify par-
ticle size. The cyclone had an overall height of 5-1/2 inches. Due to
the high pressure drop across the scrubber (30 inches water pressure),
it was necessary to fabricate a sealed holding chamber and sliding gate
valve assembly (Figure 16) so that the sampling probe could be inserted
into the stack against the high static pressure. The relatively large
holding chamber would not allow complete traversing of one stack diameter
-------
42
TABLE 15. OPTICAL EMISSION SPECTROSCOPY ANALYSIS
FOR TRACE METALS (RUN NO. 4), PPM
Element -441(1)
Hg
Be
Cd
As
V
Mn
Ni
Sb
Cr
Zn
Ca
Pb
Se
B
F
Li
Ag
Sn
Fe
Sr
Na
K
Ca
Si
Mg
Ba
(a)
--
1
<200
<200
30
10
50
<50
50
1,000
300
50
—
100
—
<600
<1
10
3,000
50
2,000
3,000
500
--
300
50
Descriptions of
Blank Nos. -493
-441(2)
—
1
<200
<200
30
10
50
<50
50
1,000
300
50
—
.200
—
<600
<1
10
3,000
20
2,000
3,000
500
--
300
50
Samples (a»
-442
—
<1
<200
<200
30
200
100
<50
300
300
50
<20
—
200
'
<600
150
<10
6,000
100
2,000
2,000
1,000
15,000
600
100
b)
-443
--
<1
<10
<10
<3
<3
<3
<10
<3
<20
<3
40
--
<3
--
<25
10
115
<3
10
115
45
370
100
45
3
Sample Nos. S75-001-441
through
-496 are as
-444
—
<1
<170
<170
<60
10
<60
<170
<60
600
<60
120
—
170
--
<570
30
570
400
<15
570
860
<60
1,700
<60
60
-445
--
<1
<285
<285
<95
70
190
<285
95
950
1,900
190
—
280
—
<950
50
1,900
570
70
2,900
2,400
1,900
1,900
1,600
285
through -445 and
follows:
Wt
(Solid) ,
-441
-442
-443
-4A4
-445
-493
-494
-495
-496
Glass fiber filters (2
Acetone, wash
Itnpinger catch
of front
each)
half
54
and water rinse
Chloroform/ether extraction
Final acetone
rinse of
Glass fiber filter (3
Acetone blank,
Lot #7150
Oemineralized
Burdick
back half
each)
and Jackson
me
Vol
(Liquid) ,
ml
.3/54.1
268
87
3
2
0
0
double distilled water, OSU 1
Chloroform/ether, Baker A.R.
1
.8
.0
.5
.1
.4
.6
.9
160
815
150
325
__
200
200
150 .
(b) Background Subtracted.
-------
43
6'
FIGURE 16. INLET STACK SPECIAL ADAPTER WITH GATE VALVE SHOWING
POSITION OF ALUNDUM THIMBLE PRIOR TO SAMPLING
-------
44
due to insufficient probe length; therefore, the second half of the stack
was traversed without the holding chamber. Working conditions were very
undesirable because of the hot exhaust gases (180-200 F) escaping through
the 6-inch-diameter nipple at approximately 300 cfm. During the second
half of each of the two stack traverses, the gate valve was opened, the
probe was inserted into the stack against the flow of hot exhaust gases,
the 6-inch pipe reducing coupling was threaded into the valve assembly,
and traversing continued.
Figure 17 shows the inlet stack geometry relative to sampling
(3)
locations. According to the December 23, 1971, Federal Register ,
Method 1, sampling two diameters upstream and less than one diameter
downstream from a disturbance requires at least 48 sampling points, 24 on
a diameter as indicated on Figures 16 and 17. Specific distances for traverse
points are given in Table 16.
Outlet of Mist Eliminator
Particulate and gaseous emissions were sampled at the outlet
of the mist eliminator. Figure 18 shows the geometry relative to the
sampling locations. According to the December 23, 1971, Federal Register,
Method 1, 22 sampling points per stack diameter were required relative to
the associated upstream and downstream stack diameters. Table 17 shows
specific distances for traverse point locations.
Prior to outlet particulate sampling, it was necessary to
install straightening vanes to reduce the cyclonic swirl caused by the
mist eliminator. This was accomplished by the insertion of two 8-foot
lengths of metal at right angles to each other as depicted in Figure 18.
-------
45
TABLE 16. TRAVERSE PORT LOCATIONS—INLET
Sample
location
Flow
o
12
Point
1
2
3
4
5
6
7
8
9
10
11
12
Percent
of
Diameter
1.1
3.2
5.5
7.9
10.5
13.2
16.1
19.4
23.0
27.2
32.3
39^8
Distance from
Outside of Sam-
ple Port (a), in.
41
42-1/8
44
45-3/4
47-5/8
49-1/2
51-5/8
54
56-5/3
59-5/8
63-1/4
68-5/8 '
Point
13
14
15
16
17.
18
• 19
20
21
22
23
24
Percent
of
Diameter
60.2 •
67.7
72.8
77.0
80.6
83.9
86.8
89.5
. 92.1
94.5
96.8
98.9
Distance from
Outside of Sam-
ple Port(b), in.
57-3/8
61-3/4
65-3/8
68-1/2
69
73-3/8
75-1/2
77-1/2
79-1/4
81
82-3/4
84
(a) Using "holding chamber".
(b) Using "short nipple".
'Venturi
scrubber
FIGURE 17. INLET STACK GEOMETRY SHOWING SAilPLING LOCATION AND
SAMPLE POINT CONFIGURATION
-------
46
6-9
Mist
eliminator
•Straightening
vanes
TABLE 17. TRAVERSE POINT LOCATIONS —OUTLET
22-9'
Point
1
2
3
4
5
6
7
8
9
10
11
Percent
of
Diameter
1.1
3.5
6.0
8.7
11.6
14.6
18.0
21.8
26.1
31.5
39.3
Distance from
Outside o£
Sample Port, in.
6
7-7/8
9-7/8
12
14-3/8
16-3/4
19-1/2
22-5/8
26-1/8
30-1/2
36-3/4
Point
12
13
14
15
16
17
18
19
20
21
22
Percent
of
. Diameter
60.7
68.5
73.9
78.2
82.0
85 .4
88.4
91.3
94.0
96.5
98.9
Distance from
Outside of
Sanple Port, in.
54
60-3/8
64-3/4
68-1/4
71-1/4
74
76-1/2
78-3/4
81
83
84-3/4
FIGURE 18. OUTLET STACK GEOMETRY SHOWING SAMPLING LOCATION
AND SAMPLE POINT CONFIGURATION
-------
47
PROCESS DESCRIPTION AND OPERATION
The Westmoreland Coal Company installation near Quinwood, West
Virginia in Nicholas County thermally dries Pocahontas No. 3 Sewell and
Dorchester-Imboden coal. The demand for thermal drying is due to freight
rate savings, the elimination of handling problems due to freezing, and
the needs of the customer's process (coke ovens must control bulk density
and power plants must control plugging of pulverizers). A venturi scrub-
ber is used to control particulate emissions from the thermal dryer.
Tests for particulate emissions were conducted on the thermal dryer ex-
haust only during periods of typical plant operations.
Process Description
Coal is delivered to the plant and is crushed to a variety of
sizes. It is then washed, screened and separated by both water tables
and concentration tables. The coal washing is a continuous process and
reduces ash content from 16 percent to 3 percent.
The coal that is thermally dried, the 3/8 x 0-inch coal or the
slack is dewatered by shaker tables and vacuum discs and then dried to
2.0 percent to 3.5 percent surface moisture in a thermal dryer. Coal
drying is accomplished by use of flue gas (produced by a coal-fired furnace)
reduced in temperature by dilution air. The emissions from the fluidized
bed thermal dryer are exhausted to a chamber loaded with 40 or more multi-
clones which separate much of the fine dust from the emissions. The finer
dust particles pass from the multiclones of the dryer through a high pres-
sure blower to a venturi scrubber and a demister where over 95 percent of
the particles are collected. The emissions to the atmosphere are ducted via
an 81-inch diameter stack. Figure 19 is a sketch of the thermal dryer
system. The design capacity of the thermal dryer is 270 ton/hr of coal at
approximately 250 F. The plant is usually operated at two-thirds to
three-fourths of this rate because of the moisture level. The actual
operating capacity of the dryer is limited by the tons of water evaporated
per hour rather than the amount of coal being processed. The plant runs
on a schedule of two 8-hour shifts and is typically closed on weekends.
-------
Denxister
A
I
I
Explosion Vent
KEY
-*• —— Gas flow
*C —— • Coalf low
ischarge\
Gate, - v
Rotary
Discharge
O O
Control
. Panel
Note: Sampling point
located at Dryer
feed conveyor belt
^X ru _ »v!
, vrr^
Feeder : •
Bedplate 'I
CD
Furnace
Tempering
Louvers
A exhaust temperature and fan
pressure
+ drying chamber pressure and
temperature
• bed pressure
a inlet pressure
FIGURE 19.. EXHAUSTING-TYPE FLUIDIZED-BED THERMAL COAL DRYER, SHOWING COMPONENT PARTS AND
FLOW OF COAL AND DRYING GASES, WESTMORELAND COAL COMPANY, IMPERIAL SMOKELESS
DIVISION, QUINWOOD, WEST VIRGINIA
-------
49
Emission Control Equipment
The thermal dryer exhausts into a cyclone separator where the
large particles are removed. Next, the exhaust is blown into the venturi
scrubber where the fine particles are impacted upon water droplets which
are removed from the effluent gases by the mist eliminator. From there,
the exhaust gases exit the mist eliminator and go out the stack. The
fan volume is rated at 134,000 cfm for the maximum design capacity of
270 ton/hr. The pressure drop across the venturi scrubber is 32 in. H20,
3 to 4 in. H20 attributable to the mist eliminator.
Process Operation
Samples were collected from the inlet and outlet gases to the
venturi scrubber control equipment. Opacity observations and dryer feed
*
samples were gathered concurrently. Each test run is described as follows.
June 18, 1975 - Refer Appendix E Run 1
Test went well with only one interruption that occurred
due to low feed. Also, Sewell No. 3 strip mined coal was used
for the majority of the time instead of Sewell deep-mined coal.
Though the data are not suitable for evaluation of Sewell No. 7,
they do provide an interesting comparison between the two coal
mining methods.
June 19, 1975 - Refer Appendix E Run 2
The test had only two interruptions but process operations
were unsteady. Feed to thermal dryer was discontinued, problems
with the water supply to the venturi scrubber were suspected, and
the supply of raw coal to the plant was inconsistent. These data
are not typical of routine operation. Repairs were made to the
venturi scrubber water supply system prior to commencing the next
test.
June 20, 1975 - Refer Appendix E Run 3
The test was very good, even though it was necessary to
temporarily suspend sampling twice due to process malfunctions.
Process on the whole operated very smoothly, and during actual
testing was very consistent. The test data are representative
of typical operations.
*Run numbers are relative to outlet.
-------
50
June 23, 1975 - Refer Appendix E Run 4
Test was very good with no delays at all. High drying
chamber temperatures occurred for one-half hour from 4:45 to
5:20 p.m. but were not extreme.
June 24, 1975
Process shut down every half hour due to broken equip-
ment. There was no continuity to process whatsoever. Testing
was cancelled for the day.
June 25, 1975 - Refer Appendix E Run 5
Test very good. Sampling was interrupted briefly during
two short process delays. Drying chamber temperatures were a
bit high from 12:30 p.m. to 1:05 p.m. and from 4:40 p.m. to 5:03
p.m. but were not abnormal.
June 26, 1975 - Refer Appendix E
Several special inlet tests were done by Roy Neulicht.
The first and second tests were good with no problems. The
third and fourth tests had high but not normal drying chamber
temperatures. All tests were conducted during the processing
of Sewell No. 7 deep-mined coal except the fourth test which
was Park Sewell or Sewell No. 3 strip-mined coal.
The production for the second half of Test No. 5 (June 25, 1975)
was determined to be 180 ton/hr. The average process rate for that day,
however, was only 137.7 ton/hr-because of production stoppages (see Table 18).
The rate during actual sampling was higher. The reason for this inconsistency
is that testing was never done during process malfunctions or slow ups be-
cause of the abnormally high drying chamber temperatures that occur. The
emission tests were conducted only during periods when conditions were
typical of routine operation and the data, therefore, are representative
of normal production.
TABLE 18.. PRODUCTION RATES
(June 25, 1975)
Recorder
Westmoreland Coal Co.
EPA
EPA
EPA
Amount Time
Time Period minutes
Both shifts
10:00 to 11:18
14:55 to 16:18
16:18 to 17:18
1035
438
83
60
Process
, Rate ,
ton /he
137.7
158.9
187.9
180.0
-------
51
SAMPLING AND ANALYTICAL PROCEDURES
Plant operations, by necessity, were monitored by EPA personnel
prior to and during all sampling to ensure the validity of the measure-
ments being presented in. this report.
Standard EPA sampling methods, as reported in the December 23,
1971, Federal Register, were followed to obtain samples of gaseous
emissions for C0~, CO, and 02• Samples of scrubber water and process
coal were taken under the direction of the EPA Project Officer. Sample
times are reported in the Sample Collection Log presented in Appendix D.'
A temporary on-site lab facility was set up for equipment clean-
up. Due to the nature of the process, it was important to isolate, as
much as possible, the cleanup and analytical area from plant operations.
Analytical balances were set up in the temporary lab and mass
determination of BCL inlet and outlet samples and for the EPA cyclone catches
were determined on site after each run. All calculations pertinent to
isokinetic sampling were determined after each run to ensure results
were within EPA guidelines.
Coal Sample Collection Method
Representative samples of the process coal were collected for
sieve analysis during each run. Figure 20 depicts the process feed
system used at the Westmoreland Coal Company and also gives a pictorial
diagram of how coal sample aliquots were collected.
The conveyor pushes the coal into a large hopper where it is
fed into the dryer. The bulk or large mass of the coal falls into the
hopper just as it passes the end of a stationary^ inclined ramp. As the
conveyor reaches the end of its travel and starts its return, additional
coal which has adhered to the conveyor surface is thrown off and appears
in the hopper as relatively fine particles.
-------
52
Coal samples
FIGURE 20. COAL SAMPLE COLLECTION METHOD
-------
53
The hopper outlet coal samples were collected from each of four
quadrants as depicted in Figure 20. The sample collector, fabricated ac-
cording to ASTM standards^ , was inverted and placed behind the stream-of
coal as in Position 1. The collector was then turned upright and moved into
and through the stream of coal as in Positions 2, 3, and 4. The sample
collector was emptied into a sample container and the collection method was
repeated for each of the other quadrants to obtain one composite sample.
Three samples were collected each 15 minutes during the entire particulate
sample run. The coal samples for each run were combined and a fraction was
removed for sieve analysis according to ASTM Standards .
Scrubber Water Collection
To develop a better understanding of the operation and effective-
ness of the venturi scrubber, samples of water which had passed through
•the venturi were collected at Che outlet of the scrubber sump. Figure 21
is a simple schematic depicting the basic flow pattern of the scrubber
system. As indicated, about 80 percent of the scrubber sump water is
recirculated back to the ventari for reuse while the other 20 percent is
sent to a static thickner where the particles collected in the scrubber
water settle by gravitational forces. The scrubber water sample to be used
for size distribution of the collected particles was taken from the 20
percent by-pass line as indicated. It is assumed that sufficient water
flow is maintained in the recirculation system to allow a representative
water sample to be collected relative to the associated outlet emission
sampling. If this assumption is valid then a comparison can be made of
the inlet particle size distribution and the associated particle size
measured in the scrubber water. However, it should be noted that the
particles collected in the scrubber water represent a time-averaged
representation of the particles removed from the gas stream. The volume
of the scrubber sump relative to the 20 percent discharge volume flow
would indicate the extent of time averaging involved relative to the gas
sampling time. Only if there exist steady-state scrubber operation and
-------
Straightening
vanes
Level
control
Makeup water
20% clarified
Scrubber sump
Overflow
to thick-
ener
80% recirculated water
About 20% of scrubber water flow
Water pump
Ui
Scrubber
water sample
To static thickener
FIGURE 21. SCHEMATIC OF VENTURI SCRUBBER WATER FLOW SYSTEM
-------
55
an unchanging gas borne participate feed to the scrubber can the parti-
cles found in the scrubber be directly compared with gas samples for
evaluating scrubber performance.
Particle Sizing Methodology
Particulate samples were collected by sampling the venturi scrub-
ber inlet stack gases at an isokinetic rate into an Alundum thimble. During
sampling, the thimble was located in the stack and was at stack gas temperatures,
thereby minimizing condensation. Size distribution for particles collected
in the thimbles, in a cyclone sampler, and from bulk coal samples were de-
termined with a Coulter Counter' . A Mine Safety Appliance (MSA) Particle
Size Analyzer' ' complemented Coulter Counter data relative to <325 mesh
samples. The following questions were considered relative to analysis tech-
niques which, if not applied properly, could alter the reported particle
size distribution:
(1) What effect would sample drying have on the particle
size distribution?
(2) Would particle dispersion by ultrasonic vibration
alter the particle size distribution?
(3) What particle concentrations would be "generated" from
thimble breakup during ultrasonic cleaning?
(4) If ultrasonic thimble cleaning were feasible because
of insignificant thimble particle concentration, would
it be possible to obtain a representative aliquot
from a relatively large electrolyte volume?
(5) Since the MSA particle size analyzer utilizes various
fluids to determine particle size from settling
velocity, might the coal particles be soluble enough
in the fluids to affect their sizes?
Experimental efforts were undertaken to answer these questions so that a
valid experimental procedure for size distribution measurements could be
established. The following particle size distribution results are an
average of three runs per aliquot, which is a routine procedure for Coulter
Counter analyses. One aliquot per sample was taken for Coulter Counter
analyses and one aliquot per sample was taken for analyses by the Mine
Safety Appliance Particle Size Analyzer.
-------
56
Distribution Change Due to Drying
(a)
It was determined by desiccator drying of samples collected
in thimbles at the scrubber inlet that they contained <1.0 percent mois-
ture; therefore, particle size change due to drying was expected to be
insignificant. This was verified by Coulter Counter size analyses as
shown in Figure 22. Only slight differences between the particle size
distribution curves of "undried" or "dried" samples are seen, these differences
are considered to be within experimental error and, hence, no effect of
coal sample drying was found.
The bulk process coal samples were wet when collected and,
therefore, a change in particle size distribution due to drying was
considered. An extra bulk feed coal sample was divided into two equal
portions according to ASTM standards'^. One half was dried at 100 F
and sieved dry. The other half was sieved wet to determine if. the size
distribution would show a difference. Figure 23 shows the results of
these measurements which indicate essentially no difference except in the
pan catch (<325 mesh or <44 M-m) which shows a slightly higher catch for
the wet sieving. This could be attributed to experimental error or
possibly the loss of the relatively fine catch during the drying process.
Effect of Ultrasonic Dispersion
Compared with Mechanical Dispersion
Before particle dispersion by ultrasonic techniques could be
utilized, it was necessary to determine to what extent the particle size
distribution would be altered by ultrasonic dispersion techniques and
what procedures were sufficient to assure adequate dispersion. A sample
(a) The effects of drying on size distribution were determined with coal
samples which may not have been representative aliquots and, therefore,
the data can only be considered relative with respect to size distri-
bution and are valid only for evaluating drying effects.
-------
57
99.9
99.8
99.5
99
98
95
90
•21 80
>s 70
c 60
0)
Sf 50
-------
33
98
95
90
80
4—
.? 70
o>
$ 60
^ 50
§ 40
I 30
-------
59
aliquot of a thimbl6 catch was placed in a dram vial containing the
wetting agent and electrolyte solution. Mixing was accomplished by
repeated filling and emptying of a medicine dropper (the usual technique
for dispersing particles prior to Coulter Counter sizing). A size dis-
tribution then was determined for an aliquot of this dispersion. The
dispersion was then subjected to ultrasonic vibration ; another aliquot
was removed, and a size distribution determined. Figure 24 depicts the
results of the two dispersion methods. Although slight differences
can be seen, the differences are within experimental error and it can be
concluded that either method of dispersion could be used without sub-
stantial error. (This conclusion applies only to the specific samples
of this task. Other materials could possibly respond differently depend-
ing on their morphology and wetting properties.) Particles from the
solutions used in the analysis by the Coulter Counter were collected on a
silver membrane filter and observed by means of a scanning electron
microscope for possible size differences in an additional effort to
verify that ultrasonic dispersion did not modify the particle size.
Electron photomicrographs were taken of representative areas of the
filters and it was observed from these photos that the particles are
essentially the same in number and size with and without ultrasonic
dispersion.
Ultrasonic Cleaning of Alundum Thimble
To determine the extent and nature of particle release from an
unused thimble, a new Alundum thimble was placed in 825 cc of 4.0 percent
sodium chloride (Coulter Counter electrolyte) and cleaned by ultrasonic
vibration for 30 minutes at a nominal frequency of 28 kHz + 1.5 kHz. This
procedure was repeated three times, each time using the same thimble in
clean electrolyte to determine if any change would occur in the rate of
particle generation. A size distribution by Coulter Counter was determined
from each 825-cc volume. Table 19 compares the total number of particles
generated each time by ultrasonic cleaning.
-------
60
99
98
95
90
80
70
i, so
'o>
$ 50
.0 40
1 30
0
& 20
o>
>
| 10
I '5
2
1
0.5
0.2
O.I
0.05
0.01
X:
6
j
4
»
^
,Ss
•<*?
'/
1
//
&
fc/
T
/
/
/
«»
/
^
/
1
y
L
/
^
/
y
/\ Ultrasonic Dispersed
^^ Medicine Dropper Dispersed -
).2 0.3 0.4 0.6 0.8 2 3 4 6 8 10 2C
Particle Diameter, micrometers
FIGURE 24. COMPARISON OF ULTRASONIC AND MEDICINE
DROPPER DISPERSION TECHNIQUES , COULTER
COUNTER ANALYZER
-------
61
TABLE 19. PARTICLES GENERATED FROM ULTRASONIC
THIMBLE CLEANING, COULTER COUNTER ANALYSES
Size,
M-m
22.0
17.5
14.0
11.0
8.8
7.0
5.5
4.4
3.5
2.8
2.2
1.7
Accumulative Number of
Than Indicated Size
Run 1
0
0
0
13
83
173
233
488
1,023
1,922
3,235
10,213
Run 2
2
3
16
38
51
130
231
393
690
904
1,548
3,765
Particles Larger
, x 10-3
Run 3
0
3
5
3
25
61
125
229
338
546
843
2,020
-------
62
It appears that the number of particles generated decreases with
each successive cleaning and begins to level off to some extent. One can
conclude from these data that thimble cleaning prior to use may reduce
error in the size-distribution data where relatively small sample catches
are expected.
Relative Comparison--
Particle Concentration
The importance of the relative concentration of thimble-generated
particles compared with collected coal particles in the particle population
being sized was determined. Sample contamination due to ultrasonic thimble
cleaning was found to be insignificant when determining size distribution if
a sufficient sample mass was collected. Table 20 shows the total number of
particles generated during initial thimble cleaning is of the order of 10 x
10 and is significantly reduced with successive cleanings. For comparison,
it has been determined, during this study, that the total number of particles
in a measured aliquot of 2.16 mg of coal is of the order of 12 x 10 . During
this study, the thimble sample catch range was from 9,271 mg to 26,167 mg
(see Table 7). By ratio (9,271 mg:2.16 mg), the thimble sample particle
population is greater by at least a factor of the order of 4,000. Therefore,
it can be concluded that the relative number of particles generated by thimble
cleaning is insignificant by comparison with the mass contained in any of
the thimble samples collected during this study. As previously mentioned,
if one expects a relatively small thimble catch it would be advisable to
preclean the thimble several times before use to eliminate the possiblity of
a significant background particle contribution.
-------
63
Representative Aliquots From
Large Electrolyte Volumes
Thimble cleaning and particle dispersion were both accomplished
via ultrasonic vibration. The thimble contents were emptied into a beaker
containing 825 cc of isopropyl alcohol, the alcohol being used because of
its wetting characteristics. The thimble was then placed into the alcohol
bath and ultrasonically vibrated at a nominal frequency of 28 kHz - 1.5
kHz for 30 minutes. Obtaining a aliquot from three different beaker levels
while the particles are in suspension would give some identification of
possible stratification due to nonuniform mixing and check the representa-
tive nature of aliquot collection for analysis. Figure 25 depicts the
size distribution indicating that a representative aliquot can be selected
from a relatively large volume of solution with a very dense particle con-
centration if the suspension is maintained by ultrasonic mixing. A further
conclusion from these results is that the measurement technique is repeat-
able and random errors are small.
Coal Particle Solubility
Solubility of coal particles in the solutions used with the Mine
Safety Appliance Particle Size Analyzer was found to be negligible. The
operating principle of this device is essentially to measure the particle
sedimentation rate as a function of time in an appropriate liquid media.
The particles need to be thoroughly wetted and are placed in a feed liquid
prior to being placed in the sedimentation liquid. The feed liquid consisted
of 50 percent isopropyl alcohol and 50 percent heptane. A solubility check
for coal particles in this mixture was made by measuring weight loss during
the washing of a coal sample with a large excess of the solution. A weight
loss of 0.08 percent was found which is considered to be insignificant.
-------
64
0.6 0.8 I
4 6 8 10 20
Particle Diameter, micrometer
40 60
FIGURE 25. REPRESENTATIVE ALIQUOT CHECK OF LARGE
SOLUTION VOLUME HAVING HIGH-PARTICLE DENSITY,
COULTER COUNTER ANALYSES
-------
65
REFERENCES
(1) Coulter Counter Industrial Model B Instruction Manual, Coulter
Counter Industrial Division, Franklin Park, Illinois 60131.
(2) Whitby, K. T., "A Rapid General Purpose Centrifuge Sedimentation
Method for Measurement of Size Distribution of Small Particles",
Mechanical Engineering Department, University of Minnesota, 1955;
presented at the 61st Annual Meeting of ASHRAE, Philadelphia,
Pennsylvania, January, 1955.
(3) Federal Register. December 23, 1971, Vol. 32, No. 247.
(4) 1974 Annual Book of ASTM Standards. Part 26, "Gaseous Fuels;
Coal and Coke, Atmospheric Analysis", American Society for
Testing and Materials, Philadelphia, Pennsylvania 19103, 1974.
(5) Federal Register. November 12, 1974, Vol. 39, No. 219.
------- |