CD
o.
74-KPM-IO
AIR POLLUTI
EMISSION TEST
WEYERHAEUSER COMPANY
KRAFT MILL
EVERETT WASHINGTON
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
-------
PARTICULATE EMISSION MEASUREMENTS
PROM KRAFT PULP MILLS
EMB Projects Report No.
7-4-KPM-10
Plant Tested
Weyerhaeuser Company, Everett, Washington
October 9-12, 1973
Prepared for
Environmental Protection Agency
Office of Air Quality Planning and Standards
Emission Measurement Branch
Research Triangle Park
North Carolina 27711
by
W. R. Feairheller
D. L. Harris
M. T. Thalman
MONSANTO RESEARCH- CORPORATION
DAYTON LABORATORY
1515 Nicholas Road
Dayton, Ohio 45^07
Report Reviewed by Robert M. Martin
Contract No. 68-02-0226, Task No. 11
-------
TABLE OP CONTENTS
Page
I. INTRODUCTION 1
II. SUMMARY AND DISCUSSION OF RESULTS . 3
III. PROCESS DESCRIPTION AND OPERATION 6
IV. LOCATION OP SAMPLING POINTS 1?
V. SAMPLING AND ANALYTICAL PROCEDURES 21
ill
-------
LIST OF FIGURES
Page
1. The Kraft Pulping Process 7
2. Smelt Dissolving Tank and Scrubber at the
Weyerhaeuser Mill at Everett, Washington 10
3. Inlet and Outlet of Weyerhaeuser Company
Scrubber 19
4. Traverse Point Locations 20
LIST OF TABLES
Page
1. Summary of Particulate Results 5
2. Summary of the Recovery Furnace and
Packed Tower Process Data 12
3. Summary of Calculations of Equivalent Pulp
Production Rate 16
iv
-------
I. INTRODUCTION
Under the Clean Air Act of 1970, the Environmental Protection
Agency is given the responsibility of establishing perform-
ance standards for new installations or modifications to
existing installations in stationary source categories. As
a contractor, Monsanto. Research Corporation (MRC), under the
Environmental Protection Agency's "Field Sampling of
Atmospheric Emissions" program, was asked to provide emis-
sion data from the Weyerhaeuser Company, Everett, Washington.
The field test work was directed by John Snyder, Field Test-
ing Section, Emission Measurement Branch. The sampling was
performed by Monsanto Research Corporation with Darrell L.
Harris as Team Leader.
This report tabulates the data collected at the outlet of
the wet scrubber system controlling the exhaust of the smelt
tank. The smelt tank is part of the overall recovery system
of the kraft pulp mill. Spent digestion chemicals (black
liquor) are burned in a recovery furnace and smelt is formed
in the bottom of the furnace. This smelt is. continuously re-
moved from the furnace and fed into the smelt dissolving tank
where it is dissolved in scrubbing liquor or makeup water to
make the uncausticlzed green liquor. Gaseous emissions from
the smelt dissolving tank are scrubbed in a packed tower
scrubber and emitted to the atmosphere.
-------
Sampling was performed on the outlet of the packed tower
scrubber. Particulate loadings were determined by Method 5,
"Determination of Particulate Emissions from Stationary
Sources." Other procedures that were required for the
Method 5 tests were: Method 1, "Sample and Velocity Tra-
verses for Stationary Sources;" Method 2, "Determination "of
Stack Gas Velocity and Volumetric Plow Rate (Type S Pitot
Tube);" Method 3, "Gas Analysis for Carbon Dioxide, Excess
Air and Dry Molecular Weight;" and Method 4, "Determination
of Moisture in Stack Gases." All of the above methods are
given in the Federal Register, Vol. 36, No. 247, Dee. 23,
1971.
The modification required at the sampling location was to
provide a support surface for the sampling equipment.
This report presen'ts a summary and discussion of sampling
results, a description of the process, the location of sam-
pling points an.d a description of the sampling and analyti-
cal procedures. Appendices contain all field and analytical
data generated from this project.
-------
II. SUMMARY AND DISCUSSION OF RESULTS
The outlet of the packed tower scrubber system of the smelt
dissolving tank at the Weyerhaeuser Company's kraft pulp mill
was sampled for particulate loading. An attempt was made to
sample the Inlet of this scrubber, but the flow was erratic
and Irregular at the only available sampling location. At
some point locations.flow direction was actually reversed.
After discovering the flow discrepancies, the inlet sampling
was aborted.
A total of six runs were attempted on the outlet, but only
three of the six were valid runs. Prior to the first run a
Method 4 moisture determination was performed, as well as a
velocity traverse according to Methods 1 and 2. The mois-
ture analysis indicated that the stack gases contained ap-
proximately 18.5% water by volume. This moisture value did
not seem reasonable in view of the fact that the effluent
was the outlet of a scrubber and saturated gas .was expected.
To recheck this calculation a wet bulb-dry bulb moisture
check was made and it indicated that the moisture level was
17% by volume. ' Because of the agreement of the two methods,
it was assumed that they were correct, and a value of 18$
was used to calculate the nomograph factor. During the first
half of the run the sampling train plugged several times and
had to be shut down and restarted. At the half way shutdown
for port change, the excess water was removed from the im-
pingers, the silica gel replaced and the run continued.
-------
During the second half of the run rough calculations were
made for the first half, and they indicated a moisture con-
tent of approximately 40$ and an isokinetic percentage of
about 145$. This run was aborted and the sample was not
saved.
Runs 2 and 3 were made the next day and both were calculated
to be super isokinetic. It was determined that the stack
gases were saturated, as suspected, and the stack tempera-
ture was changing during the run, causing wide variations in
the moisture content. It was decided that an average stack
temperature and average moisture content could not be used
to set the nomograph for the run. Instead, the stack tem-
perature would have to be closely monitored, and adjustments
made in the nomograph 'Calculation when a temperature change
was noted. The nomograph was therefore adjusted to each new
temperature for a moisture level equivalent to a saturated
gas at the new temperature. The next three runs (Nos. 4,
5, and 6) were made using this procedure and turned out to
be isokinetic within the ± 10% requirement.
Table 1 summarizes the results of particulate sampling, giv-
ing particulate loadings and several of the more important
variables. Run 1 is not listed since the partially collec-
ted sample was not saved. Runs 2 and 3, while not within
the isokinetic specification, are given for reference. Runs
4, 55 and 6 are the acceptable runs.
-------
Table 1
Summary of Particulate Results
Run Number:
Date:
Method Tvoe:
Volume of gas sampled - (Nm3)" - (DSCF)1
Percent Moisture by Volume
Average Stack Temperature - °C - (°P)
Stack Volumetric Flow Rate - Nrn'/sec - (DSCFM)2
Stack Volumetric Flow Rate - mVsec - (ACFM)3
Percent Isoklnetic
Feed Rate - M tons/hr of Equivalent
Pulp' Production - (tons/hr)
Duration of Run - Minutes
Particulates
Probe, Cyclone, and Filter Catch, mg
mg/Nm3 - (grains/DSCF)6
Kg/hr - (Ib/hr)
Kg/M ton of Equivalent Pulp
Production - (Ib/ton)
Total Catch - mg
mg/Nm3 - (grains/DSCF)6
Kg/hr - (Ib/hr)
Kg/ M ton of Equivalent Pulp
Production - (Ib/ton)
Percent Impinger Catch
2
10/10/73
EPA- 5
2.28 (80.5)
13.5
80.6 (177)
6590 (3880)
13800 (8110)
116.5
It. 3 (15.8)
220
559.6
215 (0.107)
1.61 (3.56)
0.113 (0.225)
613.6
270 (0.118)
1.77 (3.900)
0.124 (0.247)
8:8
3
10/10/73
EPA-5
2.10 (74.2)
47.6
81.7 (179)
5790 (3410)
13200 (7750)
121.9
17.0 (18.7)
220
628.8
300 (0.131)
1.73 (3.820)
0.102 (0.204)
678.5
323 (0.141)
1.87 (4.120)
0.110 (0.220)
7.3
4
10/11/73
EPA-5
1.93 (68.2)
44.8
78.9 (174)
6300 (3710)
13600 (8020)
104.2
15.9 (17.5)
220
503.4
261 (0.114)
1.64 (3.610)
0.103 (0'.206)
533.3
277 (0.121)
1.73 (3.820)
0.109 (0.218)
5.6
5
10/12/73
EPA-5
1.95 (69.0)
47.4
82.2 (180)
6120 (3600)
14000 (8220)
108.4
17.1 (18.8)
220
631.4
323 (0.141)
1.97 (4.350)
0.115 (0.23D
664.9
341 (0.149)
2.08 (4. 580J
0.122 (0.244)
5.0
6
10/12/73
EPA-5
1.79 (63-2)
47.8
80.6 (177)
5810 (3420)
13300 (7840)
104.6
17.3 (19.1)
220
.530.1
295 (0.129)
1.72 (3.780)
0.0994 (0.198)
556.2
311 (0.136)
1.80 (3.970)
0.104 (0.208)
4.7
'Dry-Standard Cubic Feet % 70°P, 29.92 in Hg
2Dry Standard Cubic Feet per Minute 6 70°F, 29.92 in Hg
3Actual Cubic Feet per Minute - Stack Conditions
"Normal Cubic Meters at 21.1°C, 760 mm Hg
5Metric Tons per Hour (1 metric ton - 1000 Kg)
6Grains per Dry Standard Cubic Feet
-------
III. PROCESS DESCRIPTION AND OPERATION
PROCESS DESCRIPTION
A. General
Kraft pulp Is produced from wood as shown in Figure 1. In
the process, wood is chipped into small pieces and then
cooked (digested) under pressure at elevated temperatures
in {'white liquor"—a water solution of sodium hydroxide and
sodium sulfide. The white liquor chemically dissolves
lignin, leaving wood cellulose (pulp) which is filtered from
the spent liquor and washed. The pulp produced is usually
made into paper.
The balance of the process is designed to recover the cook-
ing chemicals. Spend cooking liquor and the pulp wash water
are combined for treatment to recover cooking chemicals.
The combined streams, called weak black liquor, are normally
concentrated to 60-65% solids in evaporators and then fired
in a recovery furnace. Combustion of the organics in the
black liquor provides a significant portion of the heat re-
quired for generating process steam. Inorganic chemicals
from the black liquor are recovered as a molten smelt from
the bottom of the furnace. The smelt, consisting of so-
dium carbonate and sodium sulfide, is dissolved in water and
transferred to a causticizing tank. Lime added to the tank
converts sodium carbonate to sodium hydroxide, completing
-------
=}
Q_
I
§2
8
-------
the regeneration of white liquor, which is then recycled to
the digesters. The calcium carbonate mud which precipitates
from the causticizing tank is recycled to a kiln to regen-
erate lime.
B. Smelt Dissolving Tank
The smelt dissolving tank receives a stream of molten smelt
which is continuously tapped from the bottom of the recov-
ery furnace. The molten smelt is subsequently dissolved in
the smelt dissolving tank to form "green liquor"—a water
solution of sodium sulfide and sodium carbonate. The green
liquor is then sent to the causticizing department for fur-
ther processing. Weak wash is used as the dissolving liq-
uid in the smelt tank. Weak wash is the liquid stream that
results from washing the lime mud.
The recovery furnace in which the smelt is formed was man-
ufactured by the Babcock and Wilcox Company. This furnace
is designed to operate at an equivalent pulp production of
318 air dried metric tons (350 air dried tons) per day, but
it is presently operating around 363 to 408 air dried metric
tons (400 to 450 air dried tons) tons per day. Hot black
liquor is sprayed into the furnace through a nozzle near
the furnace bottom Occasionally, when extra heat is needed
or when the black liquor supply is temporarily interrupted,
natural gas or oil is burned.
The smelt dissolving system consists of a smelt spout, a
dissolving tank, agitators, circulating pumps, and a system
of pumps and piping for transferring the green liquor to the
causticizing department. The smelt dissolving tank is nor-
mally supplied with a closed top, a vent pipe and steel
housing which enclose the smelt spouts projecting from the
front of the furnace hearth.
8
-------
A smelt spout is basically a water cooled trough. It is
necessary to disperse or otherwise break up the stream of
smelt before it hits the surface of the liquor in the tank.
This is accomplished by discharging a heavy stream of re-
circulated green liquor into the stream of molten smelt
through nozzles located behind and slightly below the smelt
spout.
Breaking up the smelt stream before it hits the liquor sur-
face in the tank prevents serious explosions which might be
caused by the reaction between the hot smelt, consisting
primarily of sodium sulfide and sodium carbonate, and the
green liquor in which it is to be dissolved. In effect,
there is a continual series of small explosions taking place
at the contact point of the molten smelt and the recircu-
lated green liquor. This contact with the hot smelt thor-
oughly agitates the receiving water, and causes the for-
mation of large amounts of steam. The steam is vented to
the atmosphere through a scrubber to remove particulates.
The smelt dissolving tank and scrubber are shown in Figure
2.
C. Air Pollution Control System
The particulate emissions from the smelt tank are controlled
by a packed tower scrubber. This scrubber is designed to
handle 50,900 m3/sec (30,000 acfm). The scrubber is a 2.4
meter (8 foot) diameter steel tank 6.7 meters (22 feet)
high containing a scrubbing section in the bottom packed
with 7.6 centimeter (3 inch) polypropylene intalox saddles,
and an entrainment separator (demister) in the top packed
with 5.1 centimeter (2 inch) polypropylene intalox saddles.
Between sections weak wash is sprayed downward .through a
distributor.to contact the gases which are drawn up through
-------
•Sampling ports
Stack damper
(normally closed)
Smelt
(from recovery
furnace)
Green liquor
(to causticizing)
Stack
Entrainment
separator
Weak wash
JA
Smelt dissolving tank
Figure 2. Smelt Dissolving Tank and Scrubber
at the Weyerhaeuser Mill at
Everett, Washington
10
-------
the packing by an induced draft fan. One stream of water
leaves the scrubber. A portion of this stream is bled off•
and drained to the smelt dissolving tank. The rest of the
water is recycled and blended with the incoming makeup water
(weak wash).
PROCESS OPERATION
A. General
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 process conditions were monitored
and recorded on data sheets. Readings were taken about once
every half-hour. These data, and a key to the entries, are
given in Appendix B and summarized b'elow.
B. Smelt Dissolving Tank - Recovery Furnace
As far as can be determined from the process data (Appendix
B) and conversations with the operators, the equipment
operated normally during the tests. The observed ranges of
major operating conditions during the tests are summarized
in Table 2. Black liquor flow rate varied between 606 and
772 liters (160 and 204 gallons) per minute; the normal rate
is approximately 720 liters (190 gallons) per minute. The
solids contents in the black liquor during these tests
range between 59.4 and 64.4$.
The recovery furnace operation was steady throughout the
tests. The black liquor flow rate did change, however,
11
-------
Table 2
Summary of the Recovery Furnace and Packed Tower Process Data
.Range during test
Operation
condition
Black liquor
feed rate
Black liquor
solids
content
Steam
production
Units
1pm
gpm
%
1000 kg/hr
1000 Ib/hr
#1
757-776
200-205
59.4-61.8
57-64
. 126-141
#2
606
160
60.6-62.6
64-68
140-150
#3
719
190
60.2-62.0
64-68
140-150
#4
606-719
160-190
60.6-62.2
57-64
125-140
#5
704-731
186-193
60.8-62.8
64-67
142-148
#6 '
712-719
188-190
61.0-64.4
64
. 140-142
Total recir-
culation
rate to
packed tower
Makeup water
to packed
tower
1pm 1018-1136 1124-1143 1113-H54 1090-1168 1101-1136 1113-1136
gpm 269-300 297-302 294-305 288-308 291-300 294-300
1pm
gpm
303
80
299
79
299
79
299
79
299-303
79-80
303
80
-------
toward the end of Test No. 4. A larger size nozzle was sub-
stituted so that a greater black liquor flow could be
accommondated. The changeover only required about one min-
ute and, according to the operators, did not interrupt the
smelt flow to the smelt dissolving tank. Once the nozzle
was changed the black liquor flow and other furnace param-
eters were steady.
During the last test (No. 6) there was a problem with the
ash return system from the furnace's precipitator. The ash
is dissolved and ducted to the black liquor system prior to
the furnace. This duct became plugged and had to be washed
out, with the result that only water entered the black liq-
uor system and diluted the black liquor. This was. indicated
by the reduced black liquor density as recorded by the meter
in the control room.
However, according to the operators, this did not have an
affect on the total solids flow to the recovery furnace .
since they were adding extra salt cake makeup to compensate
for the loss in ash normally returned to the black liquor
system. Analyses of the black liquor during this period
indicated solids contents of 6l and 64.4$.
C. Air Pollution Control System
The total liquid recirculation flow rate to the scrubber's
sprays ranged between 1018 and 1166 liters (269 and 308
gallons) per minute during the tests.
The scrubber bypass damper was closed throughout the testing,
The damper did open slightly several times because of large
eruptions in the smelt dissolving tank. These eruptions are
typical of smelt dissolving operation, and cause a buildup
13
-------
of pressure in the system. The bypass damper is designed to
open when necessary to relieve the pressure. These openings,
however, lasted only a few seconds.
The gases from the black liquor oxidation (BLO) system at
this mill are vented to the smelt dissolving tank stack.
This BLO system was shutdown during our testing, however, so
that the BLO gases would not dilute the controlled emissions
from the smelt dissolving tank. The BLO gases are normally
added between the packed tower scrubber and the outlet sam-
pling location.
D. Equivalent Pulp Production Rates
The operations of the smelt dissolving tank and recovery
furnace are quantitatively related to the pulp production
rate in the digesters. As a result, pollutant emission
rate can be expressed on the basis of equivalent pulp pro-
duction, as shown below:
/ \ / \ /Equivalent pulp\
(Emission rate,] = (Emission rate,].J production rate] (1)
Ikg/ton pulp I \ kg/hr . 7 V tons/hr /
To use Equation 1, the equivalent pulp production was cal-
culated from the black liquor charged during the tests, as
shown below:
Equivalent pulp
production,
M. tons
Black
liquor
charged,
liters
Pulp-to-liquor ratio
at avg. % solids
M tons/liter
% Solids
(test avg.)
% Solids
(base avg.)
The last term in Equation 2 corrects for the actual percent
solids (in the black liquor charged) compared to the average
percent solids on which the pulp-to-liquor ratio is based.
(2)
-------
The pulp-to-liquor ratio in the above equation was determine-
from mill records for the month of September 1973. The total
pulp production during this time was 10,626 unbleached air
dried metric tons (11.713 tons). Flow meter readings showed
that 26,925,400 liters (7,113,700 gallons) of black liquor
were charged during the same period. By division, the pulp-
to liquor ratio was found to be 0.0004 M ton per liter
(0.00165 ton per gallon).
The average solids content of the black liquor charged during
the same time period (base average) was determined from the
furnace operator's two-hourly records of solids content. The
average of all the readings was found to be 61.4$.
Substitution of the above determined values into Equation 2
gives:
Equivalent pulp
production,
M tons
Black liquor
. charged,
liters
0.000-4
M tons/liter
"% Solids
(test avg.)
617T
(3)
Equation 3 was used to calculate the equivalant pulp pro-
duction rate during each test on the smelt dissolving tank.
The calculations are summarized in Table 3-
In summary, emission rates in units of kg per metric ton
(pounds per ton) of pulp are calculated from Equation 1.
Equivalent pulp production rates to use in this equation are
also given in Table 3.
15
-------
Table 3
Summary of Calculations of Equivalent Pulp Production Rate
CT\
Date-1973
Oct.
Oct.
Oct.
Oct.
Oct.
10
10
11
12
12
Average black liquor
flow rate during test,
1pm. (gpm)
•606
719
666
715
715
(160)
(190)
(176)
(189)
(189)
Test' Time, Avg. %
mln Solids
220
220
220
220
220
61.7
61.1
62.2
61.8
62.7
Black liquor
charged
liters
x 10s (gallons)
1.33
1.58
1.17
1.57
1.57
(35,
(11,
(38,
(11,
(11,
200)
800)
720)
580)
580)
Equivalent
pulp
production
M ton (tons)
52.5
62.2
58.2
62.6
63.6
(57.9)
(68.6)
(61.2)
(69.0)
(70.1)
Equivalent pulp^ '
production rate,
M ton/hr (tons/hr)
11.3
17.0
15.9
17.1
17.3
(15.8)
(18.7)
(17.5)
(18.8)
(19.1)
Calculated from Equation 2.
^Calculated by dividing equivalent pulp production by test time.
-------
IV. LOCATION OP SAMPLING POINTS
INLET TO SCRUBBER
The ports for the scrubber inlet sampling were located in
the original vertical stack at a distance of 4.17 m (164 in.)
from the smelt tank outlet and 0.97 m (31 in.) from the
junction of the scrubber inlet pipe and original stack.
A damper was located in the stack downstream from the
scrubber take-off line and could be opened if necessary to
bypass the scrubber. The ports in the 1.37 m (54 in.)
stack were 5.1 cm (2 in.) diameter flanged half unions
covered with a blank flange.
OUTLET OF SCRUBBER
The scrubber outlet sampling location was in the original
1.37 m (54 in.) diameter stack at a point 3-73 m (147 in.)
from the junction of the stack and scrubber fan outlet line
and 1.73 m (68 in.) from the stack outlet. This was the
center line location of one port, and the other port was
7.6 cm (3 in.) higher or 1.65 m (65 in.) from the atmospheric
outlet. Ports were 10.2 cm (4 in.) ID flanged half unions
with a blank flange for closing. The critical dimension
in terms of diameters was 2.72 diameters downstream from
the nearest upstream disturbance (the fan outlet). The
dimensions required a 44 point traverse (22 points per
direction).
17
-------
A diagram of the total scrubber system showing the location
of the sampling ports is given in Figure 3. Figure ^ shows
the location of the traverse points at both the inlet and
outlet ducts.
18
-------
t
LOCATION "A" 65"
1
r
— o (
."• "«^.
t
68"
1 PORTS SUITABLE FOR SAMPLING
~y~ TWO 4" PORTS @90°
45"
t ROOF
8' 6"
1 ,
•^
f
—
\
-^ FAN
LOCATION "B"
DAMPER
PORTS 2" I.D.
O C
-— 54"
^
TO SMELT TANK
12'
-SCRUBBER
FLOOR
Figure 3. Inlet and Outlet of Weyerhaeuser Company Scrubber
-------
•25 ,
5.625
7.125
8.625
21.5
32.5
36.5
39.25
41.5
43.5
45.25
46.875
52.250
Figure
Traverse Point Locations
20
-------
V. SAMPLING AND ANALYTICAL PROCEDURES
SAMPLING PROCEDURES
The outlet of the scrubber was sampled generally in accord-
ance with Federal Register methods. Some exceptions to
the methods are listed below.
1. The required number of traverse points for this
sampling location was 44 according to Method 1.
However, when these were calculated, the 4 points
nearest the wall were within 1 inch. The calcul-
ation for 48 points was used and the 4 extreme
points omitted resulting in a complete 44 point
traverse.
2. Method 4 and the wet bulb-dry bulb moisture
technique were used to determine initial moisture
content.
3. The nomograph calculations were changed during the
Method 5 runs to compensate for the change in
moisture content as the stack gas temperature
changed. Also, due to the high moisture content
of the stack gas, the excess water in the impingers
was removed and the silica gel replaced during the
port change of each run.
21
-------
4. Method 5 procedures were modified to collect "back
half" samples as well as the standard filter and
probe samples as described in the section below.
ANALYTICAL PROCEDURES
Analytical procedures for the Method 5 samples followed
the Federal Register guidelines, with one exception. Con-
tainer No. 3 as indicated in the method contains water from
the impingers and washing of,the glassware of the train.
The solution was extracted with chloroform and ether, and
then the extracted portion was dried to constant weight,
as specified. In addition, the water remaining after ex-
traction was evaporated to dryness at 100°C to constant
weight.1 Both weights were included in the total mass of
particulate. Sample weight from the Method 5 samples were
reported as "front half" (probe washings and filter col-
lection weights) and "total" (front half plus water, chloro-
form-ether extract and impinger acetone washing weights).
1See "Specifications for Incinerator Testing at Federal
Facilities," U.S. Department of Health, Education, and
Welfare publication, October 1967, page 31.
22
------- |