EPA PROJECT REPORT NO. 74-STN-l
I
O
ARIZONA PORTLAND CEMENT
Rillito, Arizona
322
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
-------
PREFACE
The work reported herein was conducted by Valentine, Fisher &
Tomlinson Consulting Engineers (VF&T), pursuant to Task Order No. 16
issued by the Environmental Protection Agency (EPA), under the terms
of EPA Contract No. 68-02-0236. Mr. Wesley D. Snowden served as the
contractor Project Engineer and directed the VF&T field sampling
participation as well as the sample analyses performed at the VF&T
laboratories.
Mr. Clyde Riley, Office of Air Quality Planning and Standards,
Emission Measurement Branch, served as Project Officer and was responsi-
ble for coordinating the performance program.
Mr. Alfred E. Vervaert, Office of Air Quality Planning and Standards,
Industrial Studies Branch, served as Project Engineer and was responsible
for monitoring process operations.
VF&T submitted a draft document to EPA from which EPA personnel
prepared this final report (74-STN-l).
-------
Note: Mention of trade names or commercial
products in this, publication does not constitute
endorsement or recommendation for use by the
Environmental Protection Agency.
-------
TABLE OF CONTENTS
Page Number(s)
I. INTRODUCTION ; 1-3
II. SUMMARY OF RESULTS 4-23
III. PROCESS DESCRIPTION AND OPERATION 24-33
IV. LOCATION OF SAMPLING POINTS 34-38
V. TEST PROCEDURES 39-51
APPENDICES
APPENDIX A - Complete Particulate Results with
Sample Calculation
APPENDIX B - Particle Size Results
APPENDIX C - Visual Emissions Results
APPENDIX D - Field Data
APPENDIX E - High-Volume Sampling Data
APPENDIX F - Sampling Log
APPENDIX G - Process Operation Field Data
APPENDIX H - Sample Identification Log
APPENDIX I - Sample Handling Log
APPENDIX J - Analytical Data Sheets
APPENDIX K - Sampling and Analytical Procedures
APPENDIX L - Project Participants
-------
I. 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 which may contribute significantly 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 the stone crushing industry, the emission control systems (baghouses)
of the Arizona Portland Cement Company, Rillito, Arizona, were selected
by EPA for an emission testing program to provide a portion of the data
base to be used for developing "best technology" and for developing
emission standards for stone crushing operations. This report presents
the results of the testing performed during the weeks of June 2, 1974,
and June 9, 1974.
The Arizona Portland Cement Company crushes and processes lime-
stone for the manufacturing of cement. The limestone is taken from an
on-site open quarry and processed through several crushing and screening
steps prior to entering the kiln. Except for the original trucking
of the raw materials from the quarry, the materials are handled by an
automated system of enclosed conveyor belts and process buildings.
This is an extremely well-controlled stone crushing operation with
1
-------
participate (dust) control systems throughout. These control
systems consist of Mikropul* pulse jet baghouses located at the
primary crusher, primary screen, conveyor transfer points, and
secondary screening and crushing plant.
A total of fourteen particulate tests using EPA Method 5 were
conducted to determine outlet emissions from each of the four bag-
houses. These consisted of four tests at the primary crusher, four
tests at the primary screen, and three tests at the No. 2 overland
conveyor transfer point and three tests at the secondary screen and
crushing plant. Eight particle size determinations were performed
on the outlet stream from the primary crusher discharge and five
particle size determinations were performed on the inlet to the primary
screen baghouse. All tests were conducted using a Brinks Cascade*
impactor modified for in-stack collection. In addition, to compliment
the particle size data, samples of captured dust were collected from the
primary crusher and secondary screen and crusher plant baghouses for
subsequent analysis by a centrifugal classifier.
Visible emission readings were recorded at the exhaust of each of
the four baghouses mentioned above. Also five test runs using
experimental high-volume sampling equipment were conducted at the primary
screen baghouse exhaust to determine the comparability of the equipment
to EPA Method 5 equipment. The operational and emission data gathered
during these five runs are treated as an experimental effort and are not
to be used as data to support new source standards of performance.
*Mention of a specific company or product does not constitute endorsement
by EPA.
-------
The VF&T field team arrived at the plant on June 3, 1974, and
performed a series of velocity traverses on the exhaust stack of the
primary crusher and primary screen baghouses. Preliminary tests and
measurements were completed by June 4, 1974, and formal test runs on
these sources were conducted on June 4, 5, 10, 11, 12, and 13. The
remaining test runs were conducted at the No. 2 conveyor transfer site
and at the secondary screen and crushing plant on June 6, 7, 8, 10,
11, and 12.
Many of the runs were discontinuous due to process delays and
shutdowns encountered throughout the test program. As indicated in
Appendix F (Sampling Log), a two- or three-day sampling period was
required for each of the particulate runs conducted, except for the
three runs at the secondary plant and runs 2 and 3 at the conveyor
transfer site. This was due to the process delays and the fact that the
plant only operated approximately six hours each day. Also, the extremely
low concentration of emissions forced an unusually long sampling time
(400 minutes) to collect a sample which would permit an accurate analysis.
All particulate samples were returned to the VF&T Laboratories in
Seattle, Washington, for analysis. Particle size analyses were performed
on the collected baghouse dust by EPA at its Research Triangle Park
Laboratories.
The following sections of this report cover the summary of results,
process description and operation, sampling point locations, and test
procedures.
-------
II. SUMMARY OF RESULTS
A summary of the test results for the gas flows and particulate
results relative to each of the baghouse outlets is presented in
Table 1.
Primary Crusher
The results of the particulate analysis of the samples collected
at the primary crusher for each of the four test runs are presented in
Table 2. Included are pertinent data concerning sample volume and test
conditions. Averaging the results of the four runs indicates an average
concentration of particulate matter in the probe and filter catch of
0.00514 grains per dry standard cubic foot. The corresponding average
emission rate was measured as 1.01 pounds per hour. The average concen-
tration and emission rate based on the total catch is not reported because
the impinger water samples were inadvertently discarded after the
chloroform-ether extraction analysis.
The first test run at the primary crusher was above the desired
maximum of 110 percent isokinetic condition, so a fourth test run was
performed. The emission rates (pounds per hour) for this high isokinetic
run were determined by averaging the results of two independent
calculating procedures - the concentration method and the area-ratio method.
Additional information concerning the particulate testing at the primary
crusher baghouse is presented in Section VI, "Test Procedures", and Tables
A-I and A-II of the Appendices.
-------
TABLE 1
SUMMARY OF PARTICULATE EMISSIONS
Primary Crusher
(Average for
Four Runs)
Method
Gas Flow
Standard Cubic Feet/Minute, Dry
Actual Cubic Feet/Minute, Wet
Particulate (Probe and Filter)
Grains/SCF, Dry1 2
Grains/ACF, Stack Conditions
Pounds/HouK
Pounds/Ton
Particulates (Total Catch)
Grains/SCF, Dry
Grains/ACF, Stack Conditions
Pounds/Hour
Pounds/Ton
EPA-5
22,615
26,856
Primary Screen
(Average for
Four Runs)
EPA-5 H1-Volume
13,361
15,779
0.00514
0.00432
1.01
0.00101
0.00179
0.00151
0.22
0*. 00022
12,421
14,656
0.00132
0.00112
0.14
0.0014
Primary Transfer
Conveyor
(Average for
Three Runs)
EPA-5
1,935
2,346
0.00155
0.00128
0.03
0.00003
Secondary Screen and
Crusher
(Average for
Three Runs)
EPA-5
9,214
10,532
0.00062
0.00054
0.05
0.00031
Grains per Dry Standard Cubic Foot, Standard Conditions of 70°F, 29.92 in. Hg.
Grains per Actual Cubic Foot, Stack Conditions
BaseJ on Raw Materials Entering Primary Crusher
Average Data Not Presented Due to Impinger Water Being Erroneously Discarded
Calculated by Averaging the Concentration and Area Ratio Results
-------
TABLE 2
SUMMARY OF TEST RESULTS
PRIMARY CRUSHER *
Run Number
Date
Volume of Gas Sampled - DSCFa
Percent Moisture by Volume
Average Stack Temperature - °F
Stack Volumetric Flow Rate - DSCFM1
Stack Volumetric Flow Rate - ACFMC
Percent Isokinetic
Percent Excess Air
Percent Opacity ••"•-.•••:'*•
Feed Rate - ton/hr
Particulates - probe,
and filter catch
mg
gr/DSCF
gr/ACF
Ib/hr
Ib/ton feed
Particulates - total catch
mg
gr/DSCF
gr/ACF •
Ib/hr
Ib/ton feed
Percent impinger catch
1
6-4-74
286.20
2.4
79.0
23,469
27,198.
. • *'^ " * » i . *• - «»•-
.... -.Ji ;v^r?.V • •_. .' *'. - . -. ,
' :: "114.3
,-...,;,_.,,-,,
978.0
."..-. ' . " ". -»y ' ,' •-..:. "• ? -"v., - •
,.66.06
0.00355
0.00307
0.77d
0.00079 ;
72.61
0.00391
0.00337
0.85d
0.00087
2
6-10-74
245.71
2.5
81.0
22,351
.26,430.,,
109.1
" o :: '
994.0
75.13
0.00471
0.00398
0.90
0.00091
e
e
e
e
e
3
6-11-74
186.74
3.0
88.0
22,140
.26,653 ,_.
104.7
Y
1028.0
61.13
0.00504
0.00419
0.96
0.00093
72.34
0.00597
0.00495
1.13
0.00110
4
6-12-74
141.82
3.3
88.0
22,502
27.142
104.3
,..,,-=
1010.0
66.91
0.00727
0.00602
1.40
0.00139
77.25
0.00839
0.00695
1.62
0.00160
9.0
15.5
13.4
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 fact per minute ,,
Calculated by averaging the concentration and. area ratio results
e
Impinger water erroneously discarded
x ^
6
-------
Primary Screen
~* "Taulc 3 summarizes the results of the particulate sampling at
the primary screen baghouse outlet. The average particulate concen-
tration for the four test runs (based on the probe and filter catch)
was 0.00179 grains per dry standard cubic foot, and the average emission
rate was 0.22 pounds per hour. The primary screen test runs were
affected by several operational problems. The runs were discontinuous
due to process upsets and/or shutdowns. In addition, several other
problems were encountered with malfunctions of the testing equipment.
An incorrect metering orifice coefficient was used to determine the
sampling conditions. This error resulted in sampling conditions that
again exceeded the maximum of 110 percent isokinetic. The emission
rates (pounds per hour) were therefore computed by the average of two
metnods - the concentration method and the area-ratio method. Additional
information concerning the particulate testing at the primary screen
baghouse is presented in Section VI, "Test Procedures", and Tables A-III
and A-IV of the Appendices.
Primary Transfer Conveyor .
Table 4 presents a summary of the results for the three particulate
test runs at the No. 2 overland primary transfer conveyor baghouse out-
let. The average particulate concentration for the probe and filter
catch was 0.00155 grains per dry standard cubic foot, and the average
emission rate was measured as 0.03 pounds per hour. The concentrations
of particulate ranged from 0.00095 to 0.00207 grain per dry standard
cubic foot. These variations in the particulate results can possibly
7
-------
TABLE 3
'SUMMARY OF TEST RESULTS
PRIMARY SCREEN
Run Number 1
v " • •
Date
Volume of Gas Sampled - DSCFa
Percent Moisture by Volume
Average Stack Temperature - °F
Stack Volumetric Flow Rate - DSCFMb
Stack Volumetric Flow Rate - ACFMC
.i •;-.," ''•->:• •:-*•',: -:.-•'-.-.'£;
Percent Isokinetic "'-"'• '.'"' " ' ' "*'•'.
Percent Excess Air
Percent Opacity • •— - •••r--..-^;^^ • ••.>.-x •-.
Feed Rate - ton/hr
Particulates - probe;
and filter catch
ing
gr/DSCF
gr/ACF
Ib/hr
Ib/ton feed
Particulates - total catch
mg
gr/DSCF
gr/ACF
Ib/hr
Ib/ton feed
Percent impinger catch
6-4-74
328.07
1.7
82.0
13,636
115 ,.682
:-:-0'-vX
967.0
27.82
0.00131
0.00113
0.17d
0.00018
30.38
0.00143
0.00124
0.1 9d
0.00020
8.4
2
6-10-74
331.80
1.4
90.0
13,368
...... 15,79.7...
—
;•:•.. - Q --
965.0
37.94
0.00176
0.00149
0.22d
• 0.00023
e
e
e
e
e
e
3
6-11-74
257.81
2.1
90.0
13,246
^ 15, 771
' 111.5V
— _
0 v
1023.0
31.51
0.00188
0.00158
0.23d
0.00022
39.28
0.00235
0.00197
0.29d
0.00028
19.8
4
6-12-74
196.69
2.5
94.0
13,196
. 15,866
113.8
-..-,Q.,.,
1056.0
28.34
0.00222
0.00184
0.27d '
0.00026
40.11
0.00314
0.00261
0.39d
0.00037
29.3
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
Calculated by averaging the concentration and area ratio results
e Impinger water erroneously discarded
8
-------
TABLE 4
SUMMARY OF TEST RESULTS
PRIMARY TRANSFER CONVEYOR
Run Number
Date
Volume of Gas Sampled - DSCFa
Percent Moisture by Volume
Average Stack Temperature - °F
Stack Volumetric Flow Rate - DSCFM
Stack Volumetric Flow Rate - ACFMC
Percent Isokinetic
Percent Excess Air
Percent Opacity
Feed Rate - ton/hr
Parti culates - probe;
and filter catch
mg
gr/DSCF
gr/ACF
Ib/hr
Ib/ton feed
Particulates - total catch
mg
gr/DSCF
gr/ACF
Ib/hr
Ib/ton feed
Percent impinger catch
1
6-10-74
273.32
2.4
98.0
1,900.
2,303.
105.9
0
909.0
16.83
0.00095
0.00078
0.02
0.00002
d
d
d
d
d
2
6-11-74
223.12
2.4
101.0
1,902.
2,313.
107.9"
0
914.0
23.54
0.00162
0.00134
0.03
0.00003
27.59
0.00190
0.00156
0.03
0.00003
3
6-12-74
231.50
2.3
97.0
2,003.
2,422.
: '106.3
o •
873.0
31.14
0.00207
0.00171
0.04
0.00004
38.93
0.00259
0.00214
0.04
0.00005
14.7
20.0
Dry standard cubic feet at 70°F, 29.92 in. Hg.
Dry standard cubic feet per minute at 70°F, 29;92 in. Hg.
Actual cubic feet per minute
Impinger water erroneously discarded
-------
be attributed to process production rates and/or the inaccuracies that
accompany measurement of gas streams which have low concentration of
particulate emission. Additional information concerning the
particulate testing at the primary transfer conveyor baghouse outlet
is presented in Section VI, "Test Procedures", and in Tables A-V and
A-VI of the Appendices.
Secondary Screen and Crusher
The results of the particulate analysis of the samples collected
at the secondary screen and crusher plant baghouse outlet are presented
in Table 5. The average concentration of particulate matter (probe and
filter catch) for the three runs was 0.00062 grains per dry standard
cubic foot. The corresponding average emission rate was measured
as 0.05 pounds per hour. The stack volumetric flow rates were
reasonably uniform and the isokinetic sampling rates were within the
specified tolerances. The variation between the particulate loadings
can possibly be attributed to the process and flow variables encountered
during the testing. Additional information concerning the particulate
testing at the secondary screen and crusher plant baghouse outlet is
presented in Section VI, "Test Procedures", and in Tables A-VII and
A-VIII of the Appendices.
Particle Size
Particle size distribution tests were conducted on the baghouse
inlet ducts of the primary crusher and the primary screen. Tests 1
through 5 were on the primary screening device and tests 6 through 13
were on the primary crusher. The test equipment consisted of a cutting
10
-------
Run Number
TABLE 5
SUMMARY OF TEST RESULTS
SECONDARY SCREEN AND CRUSHER
Date
Volume of Gas Sampled - DSCFa
Percent Moisture by Volume
Average Stack Temperature - °F
Stack Volumetric Flow Rate - DSCFM
Stack Volumetric Flow Rate - ACFMC
Percent Isokinetic
Percent Excess Air
Percent Opacity
Feed Rate - ton/hr
Particulates - probe,
and filter catch
mg
gr/DSCF
gr/ACF
Ib/hr
Ib/ton feed
Particulates - total catch
mg
gr/DSCF
gr/ACF
Ib/hr
Ib/ton feed
Percent impinger catch
1
6-6-74
201.05
2.3
81.0
9,277.
10,579,
102.2
0
170.0
4.69
0.00036
0.00031
0.03
0.00017
6.12
0.00047
0.00041
0.04
0.00022
2
6-7-74
•173.87
2.2
77.0
8,711.
9,971.
99.8
0"
162.0
8.44
0.00075
0.00065
0.06
0.00034
12.25
0.00109
0.00095
0.08
0.00050
3
6-8-74
216.14
2.1
80.0
9,656.
11,045.
105.6
0
152.0
10.44
0.00074
0.00065
0.06
0.00041
d
d
d
d
d
23.4
31.1
Dry standard cubic feet at 70°F, 29.92 in. Hg.
Dry standard cubic feet per minute at 70°F, 29.92 in. Hg.
Actual cubic feet per minute
11
-------
cyclone, a Brinks* Model B cascade impactor, and a back-up glass
fiber filter. In addition, the cyclone catches from runs 7 through
9 were analyzed using a Bahco particle sizer. A summary of the
test data is given in Tables 6, 7, and 8 and Figures 1 through 4.
The test results are presented in the following three different forms.
1. Cumulative Mass Percent Less Than or Equal to Effective
Particle Diameter - All stages with cyclone (Tables 6-7, Figure 1).
This is the most widely used method for presenting particle size data.
It is based on the mass percentage of total particulate collected on
each stage.
2. Cumulative Mass Percent Less Than or Equal to Effective
Particle Diameter - All stages excluding cyclone (Tables 6-7, Figure 2),
The data is presented in this manner because the weight percentage
collected in the cyclone often varies widely from test run to test run.
This is especially true at a control device inlet. When working with
the small amounts of mass collected by the cascade impactor, a few
extremely large particles collected in the cyclone may make a large
difference in weight percentage. Using a cumulative basis of data
presentation, this variance in weight percentage in the cyclone is
propagated to the lower stages. Therefore, eliminating the cyclone
from the data reduction eliminates any bias introduced during data
reduction due to the fact that some extremely large and heavy particles
may have collected in the cyclone. A better correlation of the weight
percentages on the stages would tend to indicate that some of the
variance indicated when all five stages and the cyclone are plotted is
*Mention of a specific company or product does not constitute endorse-
ment by EPA.
12
-------
TABLE 6
PARTICLE SIZE
SUMMARY OF RESULTS - PRIMARY SCREEN
Collection Stage
Cyclone
1
2
3
4
5
Fi 1 ter
Dp 50
Microns
6.8
3.4
2.0
1.4
0.7
0.4
0.3
Run 1
97.69
1.53
0.47
0.14
0.07
0.00
0.10
Run 2
96.73
1.54
0.61
0.14
0.58
0.10
0.30
Run 3*
91.94
1.30
6.37
0.000
0.15
0.00
0.24
Run 4
97.76
1.59
0.50
0.03
0.10
0.02
0.00
Run 5
93.68
3.67
1.41
0.52
0.36
0.05
0.31
Ave.
96.46
2.08
0.75
0.21
0.28
0.04
0.18
Cumulative Percent
Less than Dp - Ave.
3.54
1.46
0.71
0.50
0.22
0.18
Results Omitting Cyclone
u* «
. 1
2
3
4
5
Filter
3.4
2.0
1.4
0.7
0.4
0.3
66.35
20.12
6.19
3.09
0,00
4.25
47.16
18.79
4.26
17.73
2.84
9.22
16.14
79.01
0.00
1.90
0.00
2.95
71.24
22.22
1.31
4.58
0.65
0.00
58.02
22.22
8.23
5.76
0.82
4.94
60.69
20.84
5.00
7.79
1.08
4.60
39.31
18.47
13.47
5.68
4.60
— — — —
*Results not used in average values
-------
TABLE 7
PARTICLE SIZE
SUMMARY OF RESULTS - PRIMARY CRUSHER
Collection
Stage
Cyclone
1 -
2 .
3
4
5
Fi 1 ter
Dp 50
Microns
6.8
3.4
2.0
1.4
0.7
0.4
0.3
Run 6
97.23
2.20
0.41
0.09
0.03
0.04
0.00
Run 7
97.43
2.13
0.33
0.06
0.00
0.01
0.04
Run 8
97.55
1.92
0.35
. 0.07
n n?
0.00
o.no
Run 9
97.31
2.12
0.49
0.03
0.01
0.00
0,04
Run 10
98.46
1.05
0.29
0.07
0.03
0.00
0.10
Run 11
98.22
1.29
0.37
0.05
0.03
0.00
0.03
^Run 12*
97.65
1.78
0.49
0.06
0.01
0.00
0.01
Run 13*
98.94
0.77
0.21
0.03
0.01
0.00
0.04
Ave.
97.70
1.79
0.37
0.06
0.02
0.01
0.05
Cumulative Percent
Less than Dp - Avg
2.30
0.51 .
0.14
O.OS
0.06
0.05
Results Omitting Cyclone
*• «
1
2
3
4
5
Filter-
3.4
2.0
1.4
0.7
0.4
0.3
79.26
14.88
3.34
1.17
1.34
0.00
82.85
12.66
2.29
0.00
0.18
2.02
78.34
14.44
?.q?
n.sn
0.00
3.50
78.79
18.27
1.05
0.37
0.00
1.52
68.10
19.01
4.46
1.82
0.00
6.61
72.09
20.71
3.63
1.65
0.18
1.74
75.88
20.80
2.35
0.53
0.00
0.44
72.48
19.42
3.20
0.80
0.40
3.70
76.57
16.66
2.95
0.97
0.28
2.57
23.43
6.77
3.82
2.85
2.57
_:
*Results not used in average values.
-------
to. o
')-o
*'.!>
7.0
QJ
1/1
O
U
"3
D-
II
0 Primary Crusher
Average Runs 6-11
A Primary Screen
Average Runs 1, 2, 4, 5
(Unit Density Assumed)
flTT'
Figure 1
Particle Size Distribution
Cumulative Percent Less Than Dp
-------
-^ - -^Kb^x.lA-w-^T— rfn>-, - tru ± Vtl-t ^
CTi
!_
CU
10
c
o
s_
u
0 Primary Crusher
Average Runs 6-11
A Primary Screen
Average Runs 1, 2, 4, 5
(Unit Density Assumed)
Particle Size Distribution
Omitting Cyclone
6.1
c.c-i
, 20 30 •'•) !>0 0(1 7U
Cumulative Percent Less than
-------
due to the bias introduced by abnormally large particles impacting
in the cyclone. Therefore, the data is presented in this manner
only to indicate whether or not a large amount of bias has been
introduced due to abnormally large particles collecting in the cyclone.
3. Mass Loading as a Function of Effective Particle Size (Table 8,
and Figure 3). Although particle size distribution data is usually
presented in a cumulative percent form, in some cases this presents a
biased view. This is because any error introduced either during sampling
or analytical procedures on any single collection stage is propagated to
the other stages by the nature of the data reduction. This is .due to
the fact that each data point is based on the total mass and consequently
introduces error to every other data point. Presenting the data in the
form of mass loading as a function of particle size permits the results
of each size range to be computed independently of the other size ranges.
The dry particulate material from the cyclone catches of test runs
7 through 9 were combined to provide a sufficient sample for a Bahco
particle size analysis. These results are illustrated in Figure 4.
Several problems concerning sampling and sample conditions were
experienced during the particle size testing. On test run #3, the
orifice of the second stage clogged with a small piece of gasket material.
This invalidated the run; however, the results are reported but are
not used to calculate the average values. During test run #12, the
first stage was overloaded; therefore, the results are reported but
are not used in the average values. For test runs #12 and #13 a large
17
-------
OD
TABLL 8
SUMMARY OF PARTICLE SIZE RESULTS
DIFFERENTIAL MASS LOADING BASIS
[dM/d Log D (gr/scf)3
PRIMARY SCREEN
Collection
Stage Run 1
Cyclone
1
2
3
4
19.16
1.87
0.75
0.32
0.094
5 0.000
Fi 1 ter
0.022
Run 2
6.87
0.69
0.36
0.11
0.28
0.060
0.024
Run 3*
7.Rfi
0.83
4.64
0.000
0.091
0.000
0.025
u' «
Run 4
Run 5 Avg.
3.69
0.38
0.15
0.013
0.026
i 0.005
0.000
2.22
0.55
0.27
0.14
0.058
0.011
0.008
7.99 '
0.87 i
0.38
0.15
0.11
0.019 !
0.029
••
Cyclone
1
2
3
4
5
Fi 1 ter
•
CRUSHER
Run 6 Run 7 Run 8
66.76
9.50
2.32
0.72
0.15
0.23
0.000
i
78.89
1
0.80
2.17
0.54
0.
0.
0.
000
033
048
94.94
11 .75
2.83
0.79
0.13
0.000
0.09
Run 9 Run 10 Run 11
126.97 6
17.40
5.27
0.42
0.'087 !
0.00 C
0.050 C
,7.53 191.09
4.51 15.72
1.64 5.90
0.53 1.43
0.13 0.39
1.000 ! 0.058
1.079 ! 0.068
Run 12* Run 13* Avg.
484.93 272.45 90.12i
55.36 13.39 11.611
19.82 4.67 3.36^
3.10 1.07 0.74
0.41 0.16 0.15'
0.000 0.11 0.053'
0.057 0.12 0.058
* Results not used in average values.
-------
Figure 3 • ;.
Particle Size Distribution
0 Primary Crusher
A Primary Screen
(Unit Density Assumed)
•3.0
cr>
o
"O
01
c
•r—
"O
US
O
M-
O
VI
-^,
in
c
i-
C71
.c-T
> O'/
-05.
i ,._ .
>( ^ i
2. 3 V>
Particle Diameter , D
(Microns)
19
-------
fei*%^^-'^Vt^'TVMSi>l^rtir:' m\
ro
o
Figure 4
Particle Size Distribution
(BAHCO Analysis)
Cyclone Catch; Runs 7-9
1.0 -
0.01
io ::u 30 10 1:1 r.n ;o . «u .. 90 _
Cumulative Percent Less Than Particle Size, D
-------
diameter nozzle was used resulting in extreme under-isokinetic
sampling. This was done intentionally for experimental purposes.
These values are reported for comparison, but are not included
in average values.
Eight representative samples of the captured baghouse dust
were collected. A Banco* centrifugal classifier was used by EPA
laboratory personnel to determine the terminal velocity distribution
of the samples. Graphical presentations of the data showing the
percent (by weight) of those particles in the dust samples with
terminal velocities less than various indicated values can be found
in Appendix B.
Visible Emissions
The summary of visible emission test results for each of the
4-hour test periods are provided in Appendix C for the four processes
evaluated. Opacity readings were recorded simultaneously by two
certified observers for 15-second intervals during the particulate
testing. At no time did either of the two observers note any visible
emissions (all runs had zero opacity) from either of the four baghouse
fan outlets. Additional information concerning visible emissions may
be found in the "Test Procedures" section of this report. Visible
emissions field data sheets are located in Appendix C.
High-Volume Sampling
The high-volume train used in these tests was developed within
EPA and is still in the experimental stage.
*Mention of a specific company or product does not constitute endorse-
ment by EPA.
21
-------
During the weeks of June 3 and June 10, 1974, five comparison
tests were conducted between the Method 5 pa.rticulate train and an
experimental high-volume sampler developed by the Emission Measurement
Branch. The purpose of these tests was to try and establish the
precision and accuracy of the high-volume train; i.e., whether or
not it yields grain loadings consistent (from one run to the next)
with those of a Method 5 train. The sampling location was the outlet
duct from the baghouse serving the primary screening operation.
A comparison summary of the particulate analysis for the two
sampling trains is presented in Table 9. The average particulate
concentration for the probe and filter catch was 0.00132 grains per
dry standard cubic foot, and the average emission rate was measured
as 0.14 pounds per hour. The concentrations of particulate for the
five runs ranged from 0.0010 to 0.0017 grains per dry standard cubic
foot. The results from each of the individual high-volume runs
compared favorably to EPA-5 test runs. The volumetric flow data
showed a slight discrepancy between the measured results of the two
trains. This can be explained by the fact that erroneous flow measure-
ments were recorded during the first two EPA-5 test runs and, subsequently,
averages were substituted for these data. Additional information
concerning the high-volume sampling may be found in Appendix E.
22
-------
TABLE 9
SUMMARY OF COMPARISON TEST
BETWEEN
HIGH VOLUME AND METHOD 5 SAMPLING
Dates
6/4-5/74
6/10-11/74
6/11-12/74
6/12-13/74
Run No.
EPA 5
]
2
3
4
Hi-vol
1
2
3
4^
5
Total Sampling
Time
EPA B
400
•
400
320
240
Hi-vnl
120
120
340
300
264
Percent
Isokinetic
FPA .R
116.8
115.1
111.5
113.8
Hi-vnl
104.1
99.8
98.1
98.8
98.8
Participate *
Grain Loading
FPA R
0.0013
0.0018
0.0019
0.0022
Hi-vnl
0.0010
0.0010
0.0017
0.0013
0.0016
Comparison between front half of EPA 5 and hi-vol; units are grains/dscf
23
-------
HI. PROCESS DESCRIPTION AND OPERATION
The Arizona Portland Cement Company installation at Rillito,
Arizona, was extensively modernized in 1972. Included in the moderni-
zation was the construction of an entirely new raw material processing
and handling system incorporating a new primary crushing and screening
system, surge storage and blending facilities, a four-mile overland
belt conveying system and a new secondary crushing and raw milling
circuit. In addition to production and energy considerations, air
pollution control has been highly emphasized. Control systems for
particulate emissions are evident throughout the plant.
Tests for particulate emissions were conducted on specific process
operations which could be considered similar to those found in a typical
crushed stone plant. These operations are illustrated in Figure 5 and
consist of the primary crusher system, primary screens, secondary
screening and crushing plant and a transfer point at the head of the
overland conveyor.
Process Description
The quarry and closed-circuit primary crushing plant are located
on a 100-acre site. Rock is blasted from narrow benches along a small
mountainside and from a small pit. Type of rock quarried and processed
consists primarily of limestone but is somewhat variable depending upon
the quarry location from which it is extracted and raw material blending
requirements. A rainbird watering system is used during excessively dry
periods to wet the broken stone prior to loading to reduce fugitive dust
emissions. The stone moisture is maintained at approximately 1.5 percent.
24
-------
D
Primary crusher and screen
To raw surge storage
ro
en
From surge
storac
From raw mill
feed bin
To main plant
surge pile
Transfer point at heat of overland conveyor
Legend
—>• Rock Flow
—c Ducting and pickup points
A Pan feeder
B Vibrating grizzly
C Impact crusher
D T-bar feeder
E Vibrating screen (primary)
To raw
milling
Secondary screening and crushing
F Vibrating screen (secondary
G Cone crusher
H Baghouse
I Baghouse
J Baghouse
K Baghouse
Figure 5. Process Flow Diagram
-------
Seventy-five-ton capacity haul trucks are used to transport the
broken stone to the primary plant (Figure 5). The stone is dumped
onto a hoppered 84-inch pan feeder which feeds a 7 x 12-foot vibrating
grizzly. Plus 2 1/2-inch material is fed to the primary crusher, a
Hazemag-1000* impactor driven by a 1000-hp motor. A 60-inch T-bar belt
feeder receives both the grizzly throughs and the crusher discharge and
supplies a 42-inch enclosed belt conveyor which transports the stone out
of the crusher building to an enclosed screening tower. The flow is
discharged to an 8 x 18-foot screen from which plus 2 1/2-inch oversize
is returned via a 30-inch enclosed conveyor to the primary crusher for
recrushing. Screen throughs (2 1/2-inch minus) are transferred via a
42-inch conveyor to a 7,800-ton capacity raw surge storage building and
are discharged to a shuttle belt which feeds 12 surge bins. A reclaim
tunnel under the bins houses a battery of 12 Eriez* vibrating feeders
which discharge to a 46-inch collecting/reclaim belt according to flow
and raw blending requirements.
Material on the reclaim belt is transferred to a 30-inch accelerating
transfer belt which transfers the flow to the first flight of a 30-inch
covered overland conveyor. In this way, the flow from the reclaim belt
is accelerated to the 700 fpm speed of the overland conveyor prior to
being discharged to it. The overland conveyor transports about 900 tons
of crushed stone per hour in two flights a distance of about four miles
to a storage structure at the terminus, capable of housing two 30,000-ton
linear piles.
*[-iention of a specific company or product does not constitute endorsement
by EPA.
26
-------
Reclaimed material from the storage structure is transferred
by 30-inch conveyor flights to a shuttle belt which feeds three of
six.300-ton raw mill feed bins. The other three contain higher lime
rock, iron oxide, and shale. The mix components are reclaimed and
conveyed by a 30-inch collecting belt to the secondary crushing and
screening building. Here the flow is discharged to a 5 x 10-foot
screen from which oversize (plus 3/4-inch) is conveyed to an El-Oay
56 Rollerone* for secondary reduction. The crusher product and
screen throughs (3/4-inch minus) are then conveyed directly to the
raw mill circuit and subsequently reduced to kiln feed.
Emissions Control Systems
Dust emissions are controlled by hooding emission points and
venting emissions via ducting to four baghouse units for collection.
All hoods are designed for a capture velocity of not less than 200 fpm
and ductwork sized for a conveying velocity of not less than 3,600 fpm.
Specifications for each collector are summarized in Table 10.
The primary crusher system incorporates a pan feeder, vibrating
grizzly, Hazemag* impact crusher, T-bar belt feeder, and primary belt
conveyor. Emissions are collected at various points and vented to a
Mikropul* pulse jet baghouse for collection. Collection points include
(a) the top of the impact crusher (almost blanked off); (b) the grizzly
throughs discharge chute; (c) the crusher discharge; and (d) the T-bar
feeder to primary belt conveyor transfer.
Emissions from the primary screen are vented to a second Mikropul
unit for collection. Dust emissions are collected at the top of the
^Mention of a specific company or product does not constitute endorse-
ment by EPA.
27
-------
ro
OD
TABLE 10
DUST COLLECTOR SPECIFICATIONS
Collector Designation
Manufacturer
Type
Model
Number of Bags
Bag Material
Capacity (c.f.m.)
Pressure Drop (in H^O)
Filtering Ratio
Cloth Area, (sq. ft.)
Expected Efficiency
Fan Horsepower
Primary
Crusher
MikroPul
Pulse Jet
2G3-96
528
Polypropylene HCE
26,856
1.5
5.4:1
4,973
99.99
75 @ 1800 rpm
Primary
Screen
MikroPul
Pulse Jet
IF3-24
240
Polypropylene HCE
15,779
2.6
7.0:1
2,262
99.99
40 @ 1800 rpm
Conveyor
Transfer
MikroPul
Pulse Jet
36S-8-30
36
Polypropylene HCE
2,346
1.5
6.9:1
339
99.99
7.5 @ 1800 rpm
Secondary
crusher & screen
MikroPul
Pulse Jet
1F3
216
Polypropylene HCE
10,532
1.8
5.2:1
2,036
99.99
40 @ 1800 rpm
-------
totally enclosed screen and two conveyor transfer points where the
oversize and screen throughs are discharged. A small baghouse unit
effectively controls the No. 2 transfer point at the head of the over-
land conveyor.
A fourth unit services the secondary screening and crushing plant.
Here, emissions are vented from the top of the screen enclosure and
the cone crusher discharge. Fines collected by this unit, as well as
the other three, are discharged by rotary feeders onto the belt leading
to the next process step.
Process Operation
Tests were conducted to determine particulate emission levels during
normal plant operation. Process conditions v/ere carefully observed and
tests performed only when facilities tested appeared to be operating
normally. Operating data relevant to the operation of the process
equipment and control units tested appear in Table n and Appendix G.
The rated capacity of the primary plant is 1000 tph. Throughput was
determined by the number of truck dumps per hour. Each truck is assumed
to carry 72 tons/load. The accuracy of this assumption should be within
+5 percent. It should be noted that this does not reflect the actual
impactor throughput but the amount of material dumped, scalped, crushed,
screened, and recrushed. No method was available to monitor the actual
tonnage of rock crushed. The operation was closely monitored and tests
conducted only when a minimum throughput of 900 tph was attained. The
rocks processed varied between a high limestone rock (90% Ca and MgCOo)
29
-------
TABLE 11. PROCESS WEIGHT RATES
o
LOCATION RUN DATE
1. Primary 1 6/4
Crusher 6/5
2 6/10
6/11
3 6/11
6/12
4 6/12
6/13
2. Primary 1 6/4
Screen 6/5
2 6/10
6/11
3 6/11
6/12
TIME
BEGAN
0803
0759
0822
0826
1134
0750
1220
0755
0839
0754
0827
0828
1117
0755
ENDED
1447
1020
1439
1011
1435
1109
1434
1407
1439
1100
1442
1008
1439
1050
ELAPSED
TIME (MIN)
259
141
400
295
105
400
136
184
320
120
120
240
214
186
400
300
100
400
148
172
TONS
4058
2465
6523
4987
1640
6627
2016
3465
5481
2030
2010
4040
3448
3001
6449
4867
1570
6437
2280
3175
PROCESS
WEIGHT (TPH
940
1049
978.
1015
937
994
889
1130
1028
1015
1005
1010
966
968
967
973
942
965
924
1108
320
5455
1023
-------
TABLE 11. PROCESS WEIGHT RATES (Continued)
CO
LOCATION RUN
4
3. Conveyor 1
Transfer
2
3
4. Secondary 1
Screen
and Crusher 2
3
DATE
6/12
6/13
6/10
6/11
6/11
6/12
6/6
6/7
6/8
TIME
BEGAN ENDED
1220
0755
0913
0746
0940
0733
1103
0727
0751
1434
1407
1440
0920
1442
1237
1635
1312
1331
ELAPSED
TIME (MIN)
120
120
240
315
45
360
288
288
320
320
320
TONS
2155
2070
4225
4750
703
5453
4389
4190
906
864
810
PROCESS
WEIGHT (TPH
1078
1035
1056
905
936
909
914
873
170
162
152
-------
and a low rock (70% carbonates - 10% silicates). The moisture content
of the stone was very low averaging 0.62% and ranging from 0.16 to
1.22% during the testing at the primary plant.
Although designed to transport 900 tph, the overland conveyor
currently handles only 800 tph under normal operating conditions. The
amount of material passing the conveyor transfer point at the head of
the conveyor is monitored by a load-cell belt scale upstream in the
raw surge storage building. Testing at this point was conducted only
when a minimum of 800 tph was being reclaimed and transferred to the
overland conveyor.
The secondary screening and crushing plant is rated at about 185 tph.
However, the plant is seldom operated in excess of 165 tph because its
output is transported directly to a raw mill rated at 165 tph. The feed
to the plant is monitored by a series of belt scales at the discharge
of the raw mill feed bins and consists of the actual blended cement mix
components. The amount of material processed through the plant ranged
from a low of 150 to a high of 170 tph during the three test runs. Again,
these figures do not reflect the amount of stone crushed by the secondary
cone crusher but rather the amount of material processed through the
secondary plant.
The pressure drop across each baghouse unit was monitored during
each test run to assure proper operation. With the exception of an
increase in the pressure drop across the baghouse unit servicing the
secondary plant during the latter part of run 2, all units operated
32
-------
properly throughout the tests. Also, visual observations were
constantly made at collection points throughout the test program.
. Essentially no visible dust emissions were observed with the
exception of one location at the primary screen. Although not severe,
puffs of dust were observed under the screen where "throughs" were
discharged to a belt feeder.
33
-------
IV. LOCATION OF SAMPLING POINTS
The outlet ducts on the primary crusher and primary screen operations
were somewhat eliptical in shape. The dimensions referred to in the
following are a representative average of the duct diameters.
Primary Crusher Baghouse
Particulate samples were collected on the baghouse outlet uuct
serving the primary crusher. The two sampling ports were located
113 inches downstream from the 90° baghouse outlet, and 126 inches
upstream from the 45° fan inlet. Two 4-inch I.D. sampling ports were
welded to the 33 7/8 inch I.D. sloping duct at right angles to each
other. The port locations did not meet the "eight diameters" criteria
as outlined in EPA Method la; consequently, 20 sampling points were
chosen according to method instructions for each traverse axis for a total
of 40 sampling points. Figure 6 shows the sampling port locations and
the dumensions of the outlet stack. Information concerning the particle
size sampling locations can be found in Appendix B.
Prj_mary^ Screen Baginouse
The primary screen 23 3/4 inch I.D. outlet duct was fitted with two
4-inch I.D. sampling ports in a manner similar to the primary crusher
site. The downstream distance of 129 inches and the upstream distance
of 147 inches again did not meet the Method 1 criteria; therefore a total
of 30 sampling points were selected to satisfy the minimum number of
traverse points. However, these were reduced to 20 points as specified
aEPA Standards of Performance for New Stationary Sources, FederaJ^ Register,
Volume 36, No. 247, December 23, 1971.
34
-------
in Method 1 (diameters less than 24 inches). The two high volume
sampling ports were located 135 inches downstream of the above referenced
EPA-5 test ports. Figure 7 shows duct dimensions and port locations.
Information concerning the particle size sampling locations can be found
in Appendix B.
Primary Screen Baghouse
The primary screen 23 3/4 inch I.D. outlet duct was fitted with
two 4-inch I.D. sampling ports in a manner similar to the primary
crusher site. The downstream distance of 129 inches and the upstream
distance of 147 inches again did not meet the Method 1 criteria there-
fore, a total of 30 sampling points were selected to satisfy the
minimum number of traverse points. However, these were reduced to
20 points as specified in Method 1 (diameters, less than 24 inches).
»
The two high volume sampling ports were located 135 inches downstream
of the above referenced EPA-5 test ports. Figure 7 shows duct dimensions
and port locations. Information concerning the particle size sampling
locations can be found in Appendix B.
Primary Conveyor Transfer Baghouse
The primary conveyor transfer baghouse was sampled at the outlet
of the baghouse and prior to the fan. The 8-3/4 inch I.D. duct was
fitted with two 4-inch I.D. sampling ports in a manner similar to the
other test sites. The downstream distance of 168 inches and the
upstream distance of 20 inches did meet the Method 1 criteria, but due
tc the diameter of the duct (8 3/4 I.D.) and the fact that several
35
-------
33 7/a
20 B /&
Millll I I I I Illlllllp
U> CAT/QMS
Figure 6
-------
points were within one inch of the duct wall, 8 sampling points were
selected for each traverse axis. However, due to the Method 1
instructions (24 inch diameters or less), the eight sampling points
were reduced to six per traverse axis for a total of 12 sampling
points. Figure 8 shows duct dimensions and port locations.
Secondary Screen and Crusher Baghouse
The secondary screen and crusher baghouse was sampled at the fan
outlet. In order to meet the sampling requirements of Method 1 the
outlet stack was fitted with a 180 inch stack extension. The 21 1/2
inch I.D. extension duct was fitted with two 4-inch I.D. sampling ports.
The downstream distance of 150 inches and the upstream distance of
30 inches met the Method 1 criteria and 24 sampling points were selected.
However, due to Method 1 directions the total.number of sampling points
%
were reduced to 16. Figure 9 shows duct dimensions and port locations.
37
-------
WBSfPQKT
-------
V. TEST PROCEDURES
Stack gas sampling equipment designed by the United States
Environmental Protection Agency (EPA), Office of Air Programs was
used on this evaluation. A schematic of the sampling equipment is
included in this report.
Sampling was performed according to the following: Sampling
port locations were selected and ports installed. The number of
sampling points were determined by considering the number of duct
diameters between obstructions in the duct upstream and downstream
distances of the sampling ports. Stack pressure, temperature, moisture
content and maximum velocity head readings were measured. EPA
designed nomographs were set up using this data and the correct nozzle
diameter was selected using these nomographs.
Many of the runs were discontinuous due to process upsets and/or
shut downs. The discontinuities are detailed in the field data sheets.
The sample times selected were quite long (up to 400 minutes) so process
delays, production reductions and process equipment malfunctions often
affected the continuity of sample collection. When a process upset
occurred, the sampling train was stopped and the probe removed from the
sampling duct.
Published average Cp factors of 0.90 to 0.85 were used for calcula-
tions of air flow and percent isokinetic on this project. Valentine,
Fisher, & Tomlinson's pitobes have been calibrated with and without
39
-------
nozzles in the 6" diameter wind tunnel. Calibration of pitches
without nozzles in the 6" diameter conforms to published EPA
Method 2 procedures. A February 1974 EPA Quality Assurance docu-
ment on velocity measurements states that a wind tunnel of 12"
diameter or greater should be used for the calibration of pitot
tubes. Valentine, Fisher & Tomlinson feels that the pitobes should
be calibrated with the nozzles in place and operating at isokinetic
velocities during calibration. Due to the above considerations,
accepted Cp factors of 0.90 to 0.85 have been used for calculations
and reporting of data.
A leak test was performed on the assembled sampling train. The
leak rate did not exceed 0.025 cfm at a vacuum of 22.5 inches HG.
The probe and filter were not heated as this was an ambient air source
with very low humidity. Crushed ice was placed around the impingers
at the beginning of the test with new ice being added as required to
keep the gases leaving the sampling train as much as possible below
70°F.
The train was operated as follows: The probe was inserted into
the stack to the first traverse point with the nozzle tip pointing
directly into the gas stream. The pump was started and immediately
adjusted to sample at isokinetic conditions. Equal time was spent
at selected points of equal elemental areas of the duct with the
U.S. Environmental Protection Agency, Office of Research and Develop-
ment, EPA - 650/4-74-005-a.
40
-------
pertinent data being recorded from each time interval, ihe EPA
nomograph was used to maintain isokinetic sampling throughout the
sampling period. At the conclusion of the run, the pump was turned
off, the probe was removed, and the final readings were recorded.
Clean-up of the EPA train was performed by carefully removing the
filter and placing it in a container marked "Run X, Container A".
Reagent grade acetone along with brushes were used to clean the
nozzle, glass probe, and pre-filter connections. The acetone wash
was placed in a container marked "Run X, Container B". The volume
of water in the impinger and bubblers (glassware) was weighed in
their respective containers to the nearest 0.1 gram, The original
weights which included approximately 100 ml. in the bubbler and 100
milliliters in the impinger were then subtracted and the difference
added with the water weight gain from the silica gel. This constituted
the amount of water collected during the run. The water from the
glassware and water rinse of the glassware were placed in a container
marked "Run X, Container C". An acetone rinse of the glassware and
all post-filter glassware (not including the silica gel container) was
performed and placed in a container marked "Run X, Container D".
Because of the possible insolvability of the particulate in acetone,
a water wash was conducted on the front half of the sampling train (prior
to filter). This was placed in a container marked "Run X, Container E".
The water samples after the ether-chloroform extraction were
inadvertently discarded on the first four runs analyzed. The runs
included Run No. Two primary crusher, Run No. Two primary screen, Run
41
-------
•No. One conveyor transfer point and Run No. Three secondary screen
and crusher.
Total particulate emissions were calculated using the water
clean-up following the acetone clean-up of the probe. Front half
particulate emissions were calculated using the filter and acetone
clean-up of the prefilter portions of the sampling train.
The amount of hydrocarbons collected in the sampling train on
this project appears higher than we have found for similar rock
handling operations. On similar baghouse emission collection systems,
we have discovered hydrocarbons entering the sampling train from the
oil seals of the air compressors which are used for bag cleaning. For
additional details concerning hydrocarbon analysis, please refer to
Appendix J.
The visible emissions observations were scheduled to coincide with
the particulate sampling test runs. Opacity readings were recorded
simultaneously by two certified observers during a period of four
hours on each of the four processes evaluated on this project. Zero
visible emissions were recorded for each 15-second interval contained
in the 32 total hours of observation. Opacity observations were made
at the fan outlets from the baghouse control systems. The procedures
adhered to EPA Method 9. Data sheets are provided in Appendix C.
Particle size distribution tests were conducted on the baghouse
inlet ducts of the primary stone crusher and the primary screen. Tests
1-5 were on the primary screening device and tests 6-13 were on the
42
-------
primary stone crusher. The test equipment consisted of a Brinks
Model B cascade impactor used with a cutting cyclone and a back-up
filter.
Prior to testing, a full velocity traverse was conducted at
each sample location. At both test sites, each test run was conducted
at a single point. The velocity at this point was measured immediately
preceeding each test run. The locations of the test points are
indicated in Appendix B.
The configuration of the sampling train is shown in Figures
B-7 and B-8. The impactor was placed in the stack and the nozzle
was then turned downstream and the vacuum pump started. The flow rate
through the impactor was kept constant throughout each run. This
was accomplished by operating at a constant vacuum. (The particulate
build-up on the filter is not great enough to cause a significant
increase in pressure drop across the filter; therefore, a constant
vacuum is indicative of a constant flow rate.)
At the termination of each run, the nozzle was turned upstream
and removed from the stack and the pump was turned off. This was
necessary due to high negative pressure in the stacks. After each test,
the impactor was disassembled and the collection plates and filter
were removed, placed in dessicators, and returned to the field laboratory.
The cyclone was washed with acetone, and the collected material was
returned to the laboratory for drying and weighing. For test runs 7-9,
the cyclone catch was collected, dried, and weighed. The dry particulate
43
-------
from the cyclone of the three test runs was then combined so that
a Bahco particle size analysis could be performed. The results are
shown in Figure B-4.
The impaction plates provided with the impactor each weighed
approximately 3 grams. In order to reduce this tare weight, aluminum
foil discs were inserted on the steel plates. The substrates were
treated in the field laboratory with silicon grease dissolved in
benzene and then baked at 110°C for 1 1/2 hours. The substrates were
then dessicated and weighed. After testing, the substrates were
returned to the field laboratory, dessicated, and reweighed. All
weighing was done on an electronic balance accurate to .001 ing.
Dry molecular weight and moisture content was obtained by
averaging the results of the EPA Method 3 and Method 5 tests conducted
by Valentine, Fisher, and Tomlinson.
Preliminary Velocity Traverse
A series of velocity traverses were conducted on June 3-4, 1974,
in accordance with procedures detailed in EPA Method 1 . Two testing
teams, provided with "S" type pitot tubes and inclined manometers,
performed velocity measurements at the primary crusher and screen bag-
house outlet ducts. A static pressure for each stack was obtained, and
stack gas temperature confirmed by utilizing a metal-steam dial thermometer.
Stack gas moisture content was measured by wet and dry bulk thermometers
Standards of Performance for New Stationary Sources, Federal Register,
Volume 36, No. 247, December 23, 1971.
44
-------
and determined from standard psychometric tables. U-tube manometers
were used to measure the pressure drop across each baghouse unit.
These data were recorded during the test periods and are presented
in Appendix G.
Particulate Sampling
Primary Crusher Baghouse
The sampling train designed to perform the participate sampling
of the emissions was a modified EPA Method 5 train. The modifications
consisted of the removal of the cyclone from the train and the omission
of heating the sampling probe and filter holder compartment. A 0.175"
I.D. stainless steel nozzle was attached to the 5/8" diameter pyrex
probe. The probe was connected directly to the glass filter holder
containing a pre~weighed glass fiber filter (MSA 1106-BH). A small
glass 90° elbow was used between the filter holder and the first of
four Greenburg-Smith impingers.
The first impinger was modified by replacing the orifice with a
1/2" open tube. The second impinger was a standard type Greenburg-Smith,
and the remaining two were modified. Each of the first two impingers
contained 100 ml. of distilled water, the third was dry, and the final
impinger contained 200 grams of pre-weighed dry indicator-type, 6-16
mesh silica gel. To complete the train, a control console provided a
leakless vacuum pump, a dry test meter, and a calibrated orifice connected
to an inclined manometer. Stack gas velocity measurements were accomplished
by means of a calibrated "S" type pitot tube attached to the sampling
45
-------
probe, and positioned so that the measurements were made at the
nozzle tip. The sampling train is illustrated in Figure 10.
Detailed information for the particulate testing are presented in
Appendices D and K.
Before the start of each test, leak checks were made on the
assembled sampling train (excluding the probe). The first and second
test samples at the primary crusher were withdrawn from the gas
stream for ten minutes per each of the 40 sampling points, providing
a total test time of 400 minutes. The gas velocity was observed
immediately after positioning the probe at each sampling point, and
sampling rates were adjusted to maintain isokinetic sampling conditions.
Temperature measurements were obtained of the stack gas of the impinger
gas stream, and at the inlet and outlet of the dry test meter. Test
data were recorded every five minutes throughout the sampling periods.
The duct was sampled at each point by moving the probe across and then
reversing the direction. This actually represents 20 sampling points
per traverse. It was felt that a more representative velocity profile
of the duct would be obtained over the 4-hour sampling period if this
procedure was followed.
Due to the number of process equipment malfunctions and/or shutdowns,
the project officer decided to reduce the sampling time to eight minutes
per each of the 40 sampling points, providing a total test time of 320
minutes for the third test sample. For the fourth test the sampling time
was reduced to six minutes per each of the 40 sampling points, providing
a total test time of 240 minutes.
46
-------
,/-
///
i
Knz7-LE "^^?xr
^
,-.—
1
./-STACK WALL ' ,_
r-
ppQgT? — y
V
f.'^sjT. ,
u i ;
JP ^
j/^. ., ; x
"S" Type pitot__^
integral with
sampling probe
^.
/•
^ — ;
U
n S
, H , . . i
u •
!
T I
' ^o^
^l_
xt
- MANOMETER |.
.
TEMPERATURE
ORIFICE
JiS4s4u
MANOMETER
I
TEMPERATURE
UMBILICAL
CORD
METER
BOX
Figure 10
VALENTINE, FISHER & TOMLINSON STACK GAS SAMPLING TRAIN
-------
The test procedures for sampling particulate conform to
Method 5 of the EPA Standards of Performance for New Stationary
Sources (See Appendix K). The procedures for performing velocity
traverses and determining volumetric flow rates were in conformance
with EPA Methods 1 and 2. The composition of the gas stream was
assumed to be air.
The first primary crusher sample was collected at an isokinetic
sampling rate of 114 percent and an additional sample was therefore
collected. This high isokinetic condition was the result of an
incorrect nomograph setting after the initial set point had been
determined and positioned.
Runs 3 on the primary crusher and primary screen operations
were found upon review of the data as presented in the Valentine,
Fisher, and Tomlinson draft report to have been cleaned up into mis-
marked clean-up containers. The particulate concentration stack gas
moisture, and comparison with the EPA high-volume data provided the
insight, evidence, and conclusion for the clean-up being placed in
mis-marked containers. This error was corrected and the data as
reported is correct.
Primary Screen Baghouse
The sampling train configuration utilized for particulate sampling
of the emissions from the primary screen baghouse was identical to the
Vederal Register, Volume 36, No. 247, December 23, 1971
48
-------
train used for the primary crusher tests. The test procedures also
conformed to EPA Methods 1, 2, and 5. All field data sheets can
be found in Appendix D.
Run Nos. One, Two, Three, and Four on the primary screening
operations were collected with a Valentine, Fisher, and Tomlinson
meter box which had just been calibrated. The metering orifice had
been modified during calibration. The new calibration coefficient
was not placed on the meter box before shipment to the field. Primary
screen Run Nos. One, Two, Three, and Four are therefore approximately
5% over the desired isokinetic nozzle velocity range. Atmospheric
emission calculations have been performed by two independent calcula-
tion procedures (concentration and area ratio) to determine the most
accurate emissions whenever the isokinetic velocities are outside the
110% range (see Appendix A).
The first two tests performed at the primary screen were plagued
by erroneous pi tot measurements at the manometer. Several discrepancies
were discovered in the values of AP recorded on June 4 and 5, 1974;
the values recorded on June 5 were substantially higher than those
recorded on June 4. It is believed that either the hot ambient tempera-
ture caused the pi tot tube lines to sag or that the Swagelok fittings
on the static pressure line were not sealed properly. This would
account for the unusually high velocity measurements. Run No. Two was
plagued by erroneous velocity data during the first 12 of 79 sampling
49
-------
points. Run Nos. One and Two air flows were calculated using the
average air flow readings from Run Nos. Three and Four. Four samples
were collected on the primary screen operation because the unusually
high velocity readings were recognized in the field to be unrepresenta-
tive. The sampling times were identical to those used at the primary
crusher.
A high volume sampling train was run for shorter time periods
collecting samples from the primary screen operation. Valentine, Fisher,
and Tomlinson provided the filters and analysis services on the high
volume train. The calculations and data are contained in Appendix E.
Conveyor Transfer Point Baqhou!S_e_
The sampling train configuration utilized for particulate sampling
of the emissions from the primary transfer conveyor baghouse was
identical to the trains used at the two other primary sites. The test
procedures also conformed to EPA Methods 1,2, and 5. Except for the
unusual small diameter duct (8 3/4" I.D.), no problems were realized
during the three test runs. The first sample was withdrawn from the
gas stream for 30 minutes per each of the 12 sampling points, providing
a total test time of 360 minutes. This was conducted by traversing each
axis twice during the sampling period. Test data were recorded every
five minutes throughout the sampling period. The remaining two runs were
conducted using 24 minutes per each of the 12 sampling points, providing
a total test time of 288 minutes. These runs were also conducted by
traversing each of the axes twice. Test data were recorded every four
minutes throughout the sampling periods. All field data sheets are presented
in Appendix D.
50
-------
Secondary Screen and Crusher Baghouse
The sampling train configuration utilized for particulate
sampling of the emissions from the secondary screen and crusher bag-
house was identical to the trains used for the other three sampling
locations. The test procedures also conformed to EPA Methods 1,2,
and 5. Some of the velocity measurements recorded during Run No. Two
indicate that the total baghouse volumetric flow was reduced because of
the dust (cake) buildup on the bag filters. (See secondary screen and
crusher run Appendix A). This is also noted in Appendix G under operating
data at the secondary plant that the baghouse pressure drop (AP) was
unusually high at the beginning of the work day (07:30/6-8-74). The
total sampling period for each of the three runs was 320 minutes or
20 minutes per sampling point. Test data were recorded every five
minutes throughout the sampling periods. Field data sheets are pre-
sented in Appendix D.
51
------- |