EPA REPORT NUMBER 74-KPM-11
WEYERHAEUSER MILL
Vailiant, Oklahoma
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
-------
EPA REPORT FOR
WEYERHAEUSER MILL
VALLIANT, OKLAHOMA
No. 74-KPM-ll
Task No. 8
Contract No. 68-02-1406
Submitted to
Environmental Protection Agency
Office of Air Quality Programs and Standards
Emission Measurements Branch
Submitted by
Engineering-Science, Inc.
7903 Westpark Drive
McLean, Virginia 22101
February 1975
-------
TABLE OF CONTENTS
SECTION I
SECTION II
SECTION III
SECTION IV
SECTION V
INTRODUCTION
SUMMARY AND DISCUSSION OF RESULTS
Scrubber Inlet
Scrubber Outlet
PROCESS DESCRIPTION AND OPERATION
Process Description
Lime Kiln and Venturi Scrubber
Process Operation
LOCATION OF SAMPLING POINTS
Scrubber Inlet
Scrubber Outlet
SAMPLING AND ANALYTICAL PROCEDURES
Scrubber Inlet
Scrubber Outlet
2
2
8
11
11
13
13
16
16
18
20
20
21
APPENDIX - Not included this copy.
-------
SECTION I
INTRODUCTION
Under Section 111 of the Clean Air Act of 1970 as amended, the Environ-
mental Protection Agency (EPA) is charged with the establishment of perfor-
mance standards for new stationary sources which may contribute significantly
to air pollution. These performance standards are based on the best emission
reduction systems which have been shown to be technically and economically
feasible. In order to set realistic standards, accurate pollutant emission
data are normally gathered from the particular stationary source category
under consideration.
The lime kiln at the Weyehaeuser Company's containerboard mill in Valliant,
Oklahoma was designated as a well controlled stationary source in the kraft
pulp industry and, thereby, was selected by the Office of Air Quality Planning
and Standards for emission testing. The Valliant facility, one of the world's
largest, has a design production capacity of 1,200 tons per day (tpd) of
linerboard and 400 tpd of corrugating medium. The required pulp for the paper
and paperboard production is produced on-site from wood chips. Two types of
pulp are used; kraft pulp for the linerboard and NSSC pulp for the corrugating
medium.
A single causticizing system serves both pulping processes. The rotary
lime kiln, an integral part of the causticizing system, was manufactured
by Allis-Chalmers for a design capacity of 440 tpd of lime. The kiln is
fired with No. 6 fuel oil and is equipped to incinerate noncondensables
from an on-site turpentine system, as well as those from the hot well of
the kraft evaporators and the NSSC digester. Air pollutant emissions from
the kiln are controlled by a Chemico venturi scrubber.
Particulate tests of the lime kiln were conducted by Engineering-Science,
Inc. (ES) personnel during November 13 through 16, 1973. Three tests were
conducted at both the inlet and the outlet of the lime kiln venturi scrubber
to determine filterable and total particulate. The tests were designed to
measure the average emission rate and control device removal efficiency
during normal plant operation.
-------
SECTION II
SUMMARY AND DISCUSSION OF RESULTS
The field data and calculated results are summarized and presented in
Tables II-l through II-4. More detailed data summaries for each individual
test are contained in Appendix A. Data from the scrubber inlet tests is
given in the first two tables; Table II-l expresses the data in English
Units, while Table II-2 contains the same data expressed in Metric Units.
The scrubber outlet data is expressed in English Units in Table II-3 and in
Metric Units in Table II-4. These latter two tables also present the cal-
culated scrubber removal efficiency, expressed as the percentage of the
inlet particulate mass flow rate that is removed by the scrubber.
SCRUBBER INLET
Formidable sampling difficulties were experienced at the inlet test
location. These problems were the direct result of the adverse sampling
conditions. The high positive pressure, temperature, moisture content, and
particulate loading of the inlet stream all combined to make sampling diffi-
cult if not dangerous. Frequent shut downs were required to change filters
and occasionally to empty the cyclone. The pitot tube clogged repeatedly.
The stack was not equipped with valved ports, thus insertion and removal
of the pitobe required extreme caution to protect the team members from
being burned by the 500°F stack gases. Propelled by the 16 in. H_0 stack
static pressure, the hot gases escaped from the open ports with consider-
able velocity.
These difficulties forced the first run to be terminated after 28 of
the required 40 points had been sampled. Failure of the kiln burner caused
termination of the second run after 20 points had been sampled. Reducing
the sampling time at each point from five minutes to three minutes enabled
the team to sample all 40 points during the third run. This run was not
without problems, however. The bond between the silicone rubber seal and
the fritted glass filter support failed about three-fourths of the way
through the test. This allowed the sampled gases to by-pass the filter.
The failure was accompanied by a loss of vacuum in the sampling system
which alerted the operator of the problem. The test was stopped almost
-------
TABLE II-l
PARTICULATE TEST DATA SUMMARY
WEYERHAEUSER COMPANY, VALLIANT, OKLAHOMA
LIME KILN VENTURI SCRUBBER INLET
(English Units)
Run Number
Date
Volume of Gas Sampled - DSCFa
Percent Moisture by Volume
Average Stack Temperature - °F
Stack Vol. Flow Rate - DSCFMb
Stack Vol. Flow Rate - ACFMC
Percent Isokinetic
Pulp Production Rate - ton/hr
Particulates - probe, cyclone,
& filter catch
mg
gr/DSCF
gr/ACF
Ib/hr
Ib/ton ADPd
Particulates - total catch
mg v.
gr/DSCF
gr/ACF
Ib/hr
Ib/ton ADP
Percent Impinger Catch
1
11/14/73
77.81
32.9
498
29,620
78,850
117.30
62.5
55,110
10.91
4.098
2,769
44.3
55,380
10.96
4.118
2,782
44.5
0.5
2
11/15/73
44.00
33.7
514
26,490
72,490
105.90
64.6
46,970
16.44
6.009
3,732
57.8
47,110
16.49
6.026
3,743
57.9
0.3
3
11/16/73
56.13
26.8
516
29,160
72,220
100.30
60.0
97,910
26.86
10.85
6,714
111.9
99,640
27.34
11.04
6,832
113.9
1.7
Average
59.31
30.3
509
28,420
74,520
107.80
62.4
66,660
18.07
6.99
4,405
71.3
67,380
18.26
7.06
4,452
72.1
0.8
Dry standard cubic feet at 70°F, 29.92 in. Hg.
3Dry standard cubic feet per minute at 70°F, 29.92 in. Hg.
^
"Actual cubic feet per minute.
3Air Dried Pulp (ADP)
-------
TABLE II-2
PARTICULATE TEST DATA SUMMARY
WEYERHAEUSER COMPANY. VALLIANT. OKLAHOMA
LIME KILN VENTURI SCRUBBER INLET
(Metric Units)
Run Number
Date
3 a
Volume of Gas Sampled - Nm
Percent Moisture by Volume
Average Stack Temperature - °C
3 b
Stack Vol. Flow Rate - Nm /min
3 c
Stack Vol. Flow Rate - m /min
Percent Isokinetic
Pulp Production Rate - Mton/hr
Particulates - probe, cyclone,
& filter catch
mg
mg/Nm
tng/m
kg/hr
kg/Mton ADPd
Particulates - total catch
mg
mg/Nm
/ 3
mg/m
kg/hr
kg/Mton ADP
Percent Impinger Catch
1
11/14/73
2.203
32.9
258.8
838.7
2,233
117.3
56.7
55,110
24,960
9,378
1,256
22.2
55,380
25,080
9,422
1,262
22.3
0.5
2
11/15/73
1.264
33.7
267.8
750.1
2,053
105.9
58.6
46,970
37,620
13,750
1,693
28.9
47,110
37,730
13,790
1,698
29.0
0.3
3
11/16/73
1.589
26.8
268.9
825.7
2,045
100.3
54.4
97,910
61,480
24,820
3,045
56.0
99,640
62,560
25,260
3,099
57.0
1.7
Average
1.685
31.1
265.2
804.8
2,110
107.8
56.6
66,660
41,350
15,980
1,998
35.7
67,380
41,790
16,160
2,020
36.1
0.8
*J
Dry normal cubic meter at 21.1°C, 760 mm Hg.
Dry normal cubic meters per minute at 21.1°C, 760 mm Hg.
CActual cubic meters per minute.
dAir Dried Pulp (ADP)
-------
TABLE II-3
PARTICULATE TEST DATA SUMMARY
WEYERHAEUSER COMPANY, VALLIANT. OKLAHOMA
LIME KILN VENTURI SCRUBBER OUTLET
(English Units)
Run Number
Date
Volume of Gas Sampled - DSCFa
Percent Moisture by Volume6
Average Stack Temperature - °F
Stack Vol. Flow Rate - DSCFM
Stack Vol. Flow Rate - ACFMC
Percent Isokinetic
Pulp Production Rate - ton/hr
Particulates - probe, cyclone,
& filter catch
mg
gr/DSCF
gr/ACF
lb/hr
Ib/ton ADPd
Particulates - total catch
mg
gr/DSCF
gr/ACF
lb/hr
Ib/ton ADP
Percent Impinger Catch
Scrubber Removal Efficiency
front-half catch
total catch
2
11/15/73
63.41
37.3
165
29,450
56,830
108.9
64.6
1121.5
0.272
0.141
68.8
1.07
1226.3
0.298
0.154
75.2
1.16
8.5
98.2
98.0
3
11/16/73
71.37
26.6
151
33,390
53,640
108.1
60.0
1010.1
0.218
0.136
62.4
1.04
1029.7
0.222
0.138
63.6
•1.06
1.9
99.1
99.1
Ave.
67.39
31.95
158.0
31,420
55,235
108.5
62.3
1065.8
0.245
0.138
65.6
1.06
1128
0.260
0.146
69.4
1.11
5.2
98.6
98.6
aDry 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.
dAir Dried Pulp (ADP)
Equilibrium saturation conditions..
-------
TABLE II-4
PARTICULATE TEST DATA SUMMARY
WEYERHAEUSER COMPANY. VALLIANT. OKLAHOMA
LIME KILN VENTURI SCRUBBER OUTLET
(Metric Units)
Run Number
Date
3 a
Volume of Gas Sampled - Nm
Percent Moisture by Volume
Average Stack Temperature - °C
Stack Vol. Flow Rate - Nm /min
3 c
Stack Vol. Flow Rate - m /min
Percent Isokinetic
Pulp Production Rate - Mton/hr
Particulates - probe, cyclone,
& filter catch
mg
mg/Nm
mg/m
kg/hr
kg/Mton ADPd
Particulates - total catch
mg
mg/Nm
mg/m
kg/hr
kg/Mton ADP
Percent Impinger Catch
Scrubber Removal Efficiency
front-half catch
total catch
2
11/15/73
1.795
37.3
73.6
833.9
1,609
108.9
58.6
1121.5
623.3
323.1
31.2
0.53
1226.3
681.6
353.2
34.1
0.58
8.5
98.2
98.0
3
11/16/73
2.021
26.6
66.2
945.4
1,519
108.1
•54.4
1010.1
498.7
310.5
28.3
0.52
1029.7
508.4
316.5
28.8
0.53
1.9
99.1
99.1
Avg.
1.908
32.0
.69.9
889.65
1,564
108.5
56.5
1065.8
561.0
316.8
29.8
0.52
1,128
595
334.8
3l74
0.56
5.2
98.6
98.6
^ry normal cubic meter at 21.1°C, 760 mm Hg.
Dry normal cubic meters per minute at 21.1°C, 760 mm Hg.
cActual cubic meters per minute.
dAir Dried Pulp (ADP).
^Equilibrium saturation conditions.
-------
immediately after the failure, thereby saving the sample. The percentage
of the total particulate catch that was collected in the impingers increased
from about 0.5 percent for the first two runs to 1.7 percent for the third
run. The fact that the impinger catch was still relatively small indicates
that the filter was not by-passed for any appreciable length of time.
The measured particulate loading at the inlet showed wide variation
although the kiln feed rates were relatively constant. The actual weight
of calcium carbonate mud charged to the kiln is not measured by the plant.
Flow rate and percent solids of the mud fed to the vacuum filter (located
directly upstream of the kiln) are measured. Percent solids determinations
of the kiln feed are also made. During the particulate testing, the mud
flow rate to the filter ranged from 345 to 460 gpm. Solids content of the
kiln feed varied from 64.3 to 75.0%. Copies of the plant process data,
recorded by the EPA Process Engineer, are contained in Appendix C. When
averaged over the period of each test, there is very little variation.
The mud flow rate to the filter and percent solids to the kiln averaged
430 gpm/69.7%, 453 gpm/69.8%, and 386 gpm/69.7% for the three test periods,
respectively.
The plant also provided the total weight of pulp produced each day during
the test program. These figures were expressed in units of tons of equiva-
lent air dried pulp produced per day. This is a standard method of expressing
pulp production in the kraft industry. The weight of calcium carbonate mud
fed to the kiln is directly proportional to the weight of pulp produced.
Therefore, calculated emission factors were based on the equivalent air dried
pulp production rate. As expected, equivalent pulp production rates were
nearly identical for the three test days, ranging from 60 to 65 tons/hr.
Measured particulate concentrations ranged from 11 to 27 gr/DSCF.
There was no apparent reason for this wide variation. The back-half catch
was negligible for all three runs. Measured stack flow rates for the three
tests varied less than 10%. The isokinetics of the first run were high at
117%, but the second and third runs were well within the desired range at
106 and 100%, respectively. Sampling at a rate that is greater than the
-------
isokinetic rate generally yields calculated concentrations that are lower
than the true concentrations. The amount of bias introduced depends on
the particle size distribution of the particulate matter. The worst case
occurs where only large particles are present. Under these conditions, a
sampling rate that is 17% too high has been estimated to result in a cal-
culated particulate concentration that is about 16% lower than the true
concentration. Applying a 16% correction to the calculated concentration
obtained from the first test yields an estimated particulate concentration
of approximately 13 gr/DSCF. This agrees relatively well with the test
2 results of 16 gr/DSCF, but test 3 is still quite high at 26 gr/DSCF.
The total mass of particulate material charged to the kiln is the pro-
duct of the particulate concentration and the gas flow rate. The measured
gas flow rates for the three tests ranged from a low of 26,500 DSCFM to a
high of 29,600 DSCFM. Since the gas flows were nearly constant, the calcu-
lated particulate mass flow rates exhibited the same wide variation found
in the concentration determinations. Similarly, the calculated emission
factors, which are reported as pounds of particulate matter per ton of
equivalent air dried pulp, were subject to the same variation.
SCRUBBER OUTLET
Considerable sampling difficulties were encountered at the outlet
test site and most of those occurred during the first test run. During the
first 20-point traverse, the filter clogged three times and the impingers
had to be emptied once. These problems were caused by the combination of
the high moisture content of the sampled gas stream (about 27%) and inability
of the probe and sample box heaters to maintain the sampled gas temperature
above the dew point at the high sampling rate necessary (about 1 CFM) for
the chosen 0.375 in. diameter probe nozzle. At the end of the first 20-
point traverse, the nozzle was changed to a 0.250 in. diameter. The remain-
ing half of the first test was completed without further problems, thus the
0.250 in. nozzle was retained for the second and third tests.
(1) "Method of Interpreting Stack Sampling Data," Smith, W.S., et al, National
Mr Pollution Control Association, Public Health Service, U.S. Department
of Health, Education and Welfare.
-------
Because the nozzle was changed midway through the first test, each
half of the test was calculated separately. The calculated results for
both halves of the first run did not compare well with each other or
with the results of runs 2 and 3. It is believed that the numerous
problems encountered during run 1 resulted in the introduction of
enough error to bias the results and for this reason, run 1 is not
summarized in Tables II-3 and II-4. The results are available as
particulate emission data reports in Appendix A.
The scrubber outlet gas was saturated with water. The percent moisture
values calculated from the amount of condensate collected in the sampling
train were all in excess of the equilibrium saturation concentration. Any
water present in the stack gas in excess of the saturation concentration
must exist as a liquid. In calculating the volume of gas sampled at stack
conditions, the water that was sampled as a liquid must not be considered
to have been a gas when it was sampled. To correctly perform the particulate
calculations, it was necessary to calculate the saturation concentration
from vapor pressure data and then calculate the volume of condensate that
corresponded to that saturation concentration.
Calculations for all the outlet tests were performed using the calculated
saturation condensate volume rather than the actual collected condensate
volume. That is why the computer print-out sheets in Appendix A have a
different value for the volume of water collected than do the field clean-
up sheets.
During clean-up of the sample train after run 1A, the silica gel impinger
was broken and part of the silica gel was lost. Since the calculations were
subsequently based on saturation rather than condensate measurement, this
potential problem caused no adverse effects on test results.
The measured particulate concentrations at the outlet varied even more
widely than those at the inlet. With the exception of run 3, the percentage
of the total particulate that was collected in the impingers showed good
agreement, ranging from 8 to 14%. Run 2 impinger catch was only 2% of the
total catch however. Whether based on total catch or front-half catch only,
the calculated particulate concentrations varied by a factor of approximately
4 to 1. On the basis of total catch, concentrations ranged from 0.076 to
0.298 gr/DSCF with the average being 0.186 gr/DSCF. Similar values based
on front-half catch were 0.065 to 0.272 with an average of 0.173. Since gas
-------
flow rates and pulp production rates both varied less than 10%, the calculated
particulate mass emission rates and emission factors exhibited the same
fluctuations as the particulate concentrations. The cause of the variation
from test to test is unknown.
The performance of the scrubber as a particulate removal device was cal-
culated and is reported as a removal efficiency. These calculations are
based on the measured inlet and outlet particulate mass flow rates. The
removal efficiency is defined as:
PF Inlet Ib/hr - Outlet Ib/hr .„„
K.1L — ., - . --—/, X xUU/o
Inlet Ib/hr
The calculated efficiencies differed little between runs 2 and 3 whether based
on front-half catch or total catch. The values ranged from 98.0 to 99.1%.
Based on either front-half catch or total catch, the average removal efficiency
was 98.6%.
As mentioned, there is no obvious reason to believe the variations in
measured particulate concentration are the result of sampling errors. Although
the exact hourly feed rates to the kiln could not be determined, there is no
reason to believe the reported values are not reasonably accurate. Compari-
son of measured flow rates at the inlet and outlet of each test show good
agreement with all corresponding value being within 15%. Weights of parti-
culate collected and volumes of gas sampled were sufficiently large that
errors in their measurement should have been negligible. Thus, it appears
that the measured variations in particulate concentration are at worst
unexplainable and at best real variations.
10
-------
SECTION III
PROCESS DESCRIPTION AND OPERATION
The Weyerhaeuser Company mill at Valliant, Oklahoma.produces 1300 tons
of kraft pulp per day. The pulp is made into linerboard in the adjoining
paper mill.
PROCESS DESCRIPTION
General
The process for making kraft pulp from wood is shown in Figure III-l.
In the process, wood is chipped into small pieces and then cooked in two
continuous digesters at elevated pressure and temperature. The cooking
chemicals, called white liquor, are sodium hydroxide and sodium sulfide
in water solution. The white liquor chemically dissolves lignln, leaving
wood cellulose (pulp) which is filtered from the spent liquor and washed.
The pulp is made into linerboard.
The balance of the pulping process is designed to recover the cooking
chemicals. Spent cooking liquor and the pulp wash water are combined for
treatment to recover chemicals. The combined stream, called weak black
liquor, is concentrated in multiple-effect evaporators, including a
special effect called a concentrator. The strong black liquor leaving
the evaporators is burned in a recovery furnace.
Combustion of the organics in the black liquor provides heat needed
to generate process steam. Inorganic chemicals from the black liquor are
recovered as a molten smelt from the bottom of the furnace. The smelt,
consisting of sodium carbonate and sodium sulfide, is dissolved in water
and transferred to a causticizing tank. Lime added to this tank converts
sodium carbonate to sodium hydroxide, completing the regeneration of
white liquor which is then recycled to the digesters. The calcium car-
bonate mud that precipitates from the causticizing tank is recycled to
a kiln to regenerate lime.
11
-------
FIGURE III-l
THE KRAFT PULPING PROCESS AT THE
WEYERHAEUSER MILL
VALLIANL OKLAHOMA
O-
WOOD
WHITE LIQUOR
(NaOH + Na2S)
STACK
O
Ul
a:
WATER
t
WHITE LIQUOR
(RECYCLE TO
DIGESTER)
DIGESTER
SYSTEM
PULP
PULP
WASHERS
• PULP
WATER
WEAK BLACK LIQUOR
RECOVERY
FURNACE
SYSTEM
HEAVY
BLACK
LIQUOR
n
AIR
MULTIPLE
EFFECT
EVAPORATOR
SYSTEM
SMELT
(Na2C03
Na2S)
SMELT
DISSOLVING
TANK
GREEN LIQUOR
LIME KILN
CAUSTICIZING
TANK
LIME
CALCIUM
CARBONATE
MUD
12
ENGINEERING-SCIENCE, INC.
-------
LIME KILN AND VENTURI SCRUBBER
The lime kiln was designed by the Allis-Chalmers Company to produce
440 tons of lime per day. This is equivalent to a pulp production rate
of about 1650 tons per day. This rotary kiln is 300 feet long, with an
inside diameter of 12 feet. Bunker C oil (No. 6 fuel oil) is fired in
the kiln.
The feed to the kiln is the calcium carbonate mud that precipitates
from the causticizing tanks. The mud is washed and then dried on a
rotary vacuum drum filter, as shown in Figure III-2. The dried cake is
removed from the drum on a knife edge and conveyed to the kiln. In the
kiln, the calcium carbonate mud is roasted and carbon dioxide is driven
off, leaving calcium oxide (lime) as product. The hot gases flow counter-
current to the material in the kiln. Temperatures of 2000°F are usually
maintained in the hot end of the kiln to properly calcine lime.
Noncondensable gases from the turpentine system, the multiple-effect
evaporators, and the No. 3 (NSSC) digester are also burned in the kiln.
These gases enter with the primary air directly into the flame.
The combustion gases from the kiln are scrubbed in an adjustable
throat venturi scrubber. The gas stream after it has contacted the
water in the venturi enters a cyclonic separator where the particle laden
liquid and gases are separated by centrifugal force to complete the clean-
Ing operation. Liquid from the bottom of the separator is recycled to
the venturi. A portion of the recycled water is purged to prevent ex-
cessive accumulation of solids, and used in the lime mud washers. Make-up
to the scrubber system is either fresh water or weak wash liquor. The
venturi scrubber system was manufactured by Chemico and has a design
pressure drop of 25 inches of water.
PROCESS OPERATION
The purpose of the tests was to measure emission levels during normal
mill operation. Process conditions were carefully observed, and testing
was done only when the test facility appeared to be operating normally.
During the tests, important operating parameters were monitored and
recorded on data sheets. Readings were taken about once every half-hour.
These data are in Appendix C.
13
-------
FLOW DIAGRAM OF THE LIME KILN AT THE
WEYERHAEUSER MILL
VALLIANL OKLAHOMA
LIME MUD
MUD
WASHER
VACUUM
FILTER
ROTARY KILN
STACK
AIR
NO. 6 OIL
D
Z
m
m
3)
Z
9
CO
o
m
Z
O
m
*
Z
O
LIME
(PRODUCT)
BLEED
CD
cr
•ya
i
ro
-------
As far as is known from the process information and conversations
with the operators, the lime kiln and scrubber operated normally during
the tests. The lime mud feed rate to the filter ranged between 386 and
450 gallons per minute (gpm). The solids content of the lime mud to
the filter ranged from 25.2 to 28.9 percent during the tests. The pres-
sure drop across the venturi scrubber ranged between 19.2 and 20.8 inches
of water. The average of the major process parameters for each test are
listed in Table III-l.
During tests 1 and 2, more fuel oil than normal (10-11 gpm) was
burned in the kiln because the high mud load (450 gpm) was causing the
temperatures in the kiln to decrease. The operator stated that it is
hard to maintain proper temperatures at these high mud flow rates. Normal
mud flow rate to the filter is between 400 and 450 gpm.
TABLE III-l
SUMMARY OF LIME KILN PROCESS DATA DURING PARTICULATE SAMPLING
Test
1
2
3
Date
11/14/73
11/15/73
11/16/73
Fuel
Type
Oil
Oil
Oil
Fuel Flow
(gpm)
13.6
13.3
10.7
Mud Flow
To Filter
(gpm)
430
453
386
Solid
Content
(%)
26.8
25.8
26.2
Scrubber
Pressure Drop
(in. H20)
20.4
20.3
20.0
15
-------
SECTION IV
LOCATION OF SAMPLING POINTS
Guidelines for locating sampling sites and for determining the number
and location of sampling traverse points are contained in EPA Method 1,
"Sample and Velocity Traverses for Stationary Sources." This method, as
published in the Federal Register (Vol. 36, No. 247, Part II, December 23,
1971), was used to locate the sampling traverse points for both inlet and
outlet tests. Location of the sampling sites (ports) was beyond the control
of the test team, however, because the plant had previously installed perma-
nent work platforms and sampling ports. Although these existing ports were
located in shorter than ideal duct sections, there were no longer duct
sections available.
The ports were not positioned at the optimum location with respect
to the length of ducts. However, to reposition the ports would have required
relocation of the work platforms as well. This was not only impractical,
but would have been prohibitively expensive. .The nonideal sampling sites
were compensated for, as directed in Method 1, by sampling more points at
each site.
SCRUBBER INLET
All gases exiting the lime kiln travel through a 62-inch duct that
connects the kiln to the venturi scrubber inlet. The longest undisturbed
straight length of this duct was a vertical section approximately 21 feet
long. The disturbances at either end of this straight section were 90°
elbows. The two existing sampling ports were located 135 inches from the
upstream elbow and 117 inches from the downstream elbow. The two ports
were identified as Port A and Port B and were positioned on the duct circum-
ference such that the port centerlines formed a 90° angle. Figure IV-1
shows the relative position of the sampling ports.
The distance from the sampling ports to the upstream disturbance was
the limiting factor in determining the recommended minimum number of sampling
points. As specified by EPA Method 1, a total of 48 was recommended. In
determining the straight length of duct upstream of the sampling points, the
16
-------
FIGURE IV-1
SCRUBBER INLET SAMPLING SITE
Scrubber
Structure
62" I.D.
Duct
To Scrubber
Ports £— •
Mud Feed Pipe
TRAVERSE POINT LOCATIONS
Point No.
1 & 21
2 & 22
3 & 23
4 & 24
5 & 25
6 & 26
7 & 27
8 & 28
9 & 29
10 & 30
11 & 31
12 & 32
13 & 33
14 & 34
15 & 35
16 & 36
17 & 37
18 & 38
19 & 39
20 & 40
Distance from
Stack Wall (in.)
2-1/8
3-3/4
5-3/8
7-1/4
9
11-1/8
13-1/2
16-1/8
19-1/2
24-3/8
37-5/8
42-1/2
45-7/8
48-1/2
50-7/8
53
54-3/4
56-5/8
58-1/4
59-7/8
From Kiln
17
ENGINEERING-SCIENCE. INC.
-------
test team incorrectly measured to the centerline of the elbow rather than
to the elbow outlet. The resulting 166-inch measurement yielded a minimum
requirement of 44 traverse points. Four of these points were located within
one inch of the duct wall and therefore were discarded. The remaining 40
points were used for all sampling traverses. The distance from the stack
inside wall to each traverse point is given in Figure IV-1.
The fact that 40 rather than 44 traverse points were used is not expec-
ted to significantly affect the accuracy of the test results.
SCRUBBER OUTLET
The cleaned gases exiting the scrubber flowed through a demister and
a vertical discharge stack from which they were discharged to the atmosphere.
The discharge stack was 79.5 inches I.D. and nearly 31 feet tall. The up-
stream disturbance was a transition section to the larger diameter demister.
The downstream disturbance was the stack outlet. The two existing sampling
ports were located 20 feet-7 inches downstream from the transition section
and 10 feet-2 inches upstream from the stack outlet. The ports were identi-
fied as Port A and Port B and were positioned on the duct circumference
such that the port centerlines formed a 90° angle. The relative position of
the sampling ports is shown in Figure IV-2.
The limiting factor in determining the recommended minimum number of
traverse points was the distance from the transition section to the ports.
In accordance with Method 1, a total of 44 traverse points were recommended.
As with the inlet traverse point locations, the four outer points were
within one inch of the stack wall and were not used. The location of the
remaining 40 traverse points is given in Figure IV-2.
18
-------
FIGURE IV-2
SCRUBBER OUTLET SAMPLING SITE
rVi
Wood Platform Built
at Each Port to
Support Sample Train /
V /
10'
20'
1
-2"
t
\
-7"
t
\
'•
=5=-^
r
r~
^
A Hi 1 1
A ^
To Atmosphere
t
1
|
1
t
From Scrubber
Demister
B
•^^^Nv 79-1/2" I.D.
90^ \ Duct
TRAVERSE POINT LOCATIONS
Distance from
Point No. Stack Wall (in.)
1 & 21 2-3/4
2 & 22 4-3/4
3 & 23 6-7/8
4 & 24 9-1/4
5 & 25 11-5/8
6 & 26 14-1/4
7 & 27 17-3/8
8 & 28 20-3/4
9 & 29 2S
10 & 30 31-1/4
11 & 31 48-1/4
12 & 32 54-1/2
13 & 33 58-3/4
14 & 34 62-1/8
15 & 35 65-1/4
16 & 36 67-7/8
17 & 37 70-1/4
18 & 38 72-5/8
19 & 39 74-3/4
20 & 40 76-3/4
19
ENGINEERING-SCIENCE, INC.
-------
SECTION- V
SAMPLING AND ANALYTICAL PROCEDURES
Particulate samples were collected at the inlet and outlet of the
scrubber. The sample collection procedures conformed as nearly as possible
to EPA Method 5, "Determination of Particulate Emissions from Stationary
Sources." Necessary exceptions made at each sampling site are discussed
below in the applicable site subsection.
In support of the particulate sampling, the stack gas velocity, compo-
sition, and moisture content were measured, as specified by EPA Method 2,
"Determination of Stack Gas Velocity and Volumetric Flow Rate"V EPA Method 3,
"Gas Analysis for Carbon Dioxide, Excess Air, and Dry Molecular Weight";
and EPA Method 4, "Determination of Moisture in Stack Gases," respectively.
In addition to Method 1 mentioned previously, Methods 2, 3, 4, and 5 are
also published in the Federal Register (Vol. 36, No. 247, Part II, December 23,
1971).
Sample recovery and analyses of all particulate runs followed the proce-
dures specified in EPA Method 5 with two additional requirements. First,
the front-half of each sample train (that portion of the train from the
nozzle to and including the front-half of the filter holder) x^as washed
with water prior to the normal acetone wash and these water washings were
collected and analyzed for particulate content. Secondly, the impinger
contents were collected and analyzed for particulate content. The weight
of particulate matter collected in the front-half water wash was determined
by weighing the residue remaining after evaporating the water on a steam
bath. The procedure for recovery and analyses of the impinger contents was
in accordance with proposed EPA Method 5, published in the Federal Register
(Vol. 36, No. 159, Part II, August 17, 1971).
SCRUBBER INLET
Access to the sampling ports at the scrubber inlet was partially obstruc-
ted by process piping and the building structure. There was not sufficient
clearance for a standard EPA Method 5 train. Therefore, a modified train
that had separate containers for the filter/cyclone box was attached directly
to the outlet of a standard glass-lined pitobe. A heated flexible Teflon
20
-------
sample line connected the filter box to the impinger train which was located
on the work platform floor.
Numerous problems were encountered at the inlet site, most of which
resulted from the adverse stack conditions. Stack static pressure was approx-
imately 16 in. H-0, positive; the stack gas temperature was approximately
500°F. Insertion and removal of the pitobe assembly was difficult and
time-consuming. Particulate loading was so high that, even with the cyclone,
frequent filter plugging was encountered. Additional shutdowns were required
to clean particulate material from the pitot tubes. Because of the lengthy
testing delays encountered during the first two runs, the tests had to be
terminated before all 40 points could be sampled. For the third run, samp-
ling time was reduced from the standard five minutes to three minutes per
point to insure that all 40 points were sampled within the time available.
SCRUBBER OUTLET
During the first run at the scrubber outlet, the filter had to be
changed three times during the first traverse. While changing ports, the
probe nozzle was changed from a 0.375 inch diameter to a 0.250 inch diameter.
The lower sampling rate required with the smaller nozzle minimized the
filter plugging problem for the remainder of the tests. When the nozzle
was changed, the sample train was also changed. Thus for the first test,
each traverse was calculated separately.
21
------- |