EPA-340/l-77-014a
August 1977
Stationary Source Enforcement Series
SOURCE TESTING
OF A
STATIONARY
COKE-SIDE
ENCLOSURE
GREAT LAKES CARBON CORPORATION
ST. LOUIS, MISSOURI PLANT
VOLUME I
U.S. ENVIRONMENTAL PROTECTION AGENCY
Office of Enforcement
Office of General Enforcement
Washington, D.C. 20460
-------
STUDY OF COKE-SIDE COKE-OVEN EMISSIONS
(Volume 1 of 3)
Great Lakes Carbon Corporation
St. Louis, Missouri
Contract No. 68-02-4108
Task No. 14
Prepared for:
Technical Support Branch
Division of Stationary Source Enforcement
U.S. Environmental Protection Agency
Washington, D.C. 20460
Prepared by:
Clayton Environmental Consultants, Inc,
25711 Southfield Road
Southfield, Michigan 48075
August 31, 1977
-------
DISCLAIMER
This report was furnished to the Environmental Protection
Agency by Clayton Environmental Consultants, Inc. in fulfillment
of Contract No. 68-02-1408, Task Order No. 14. The contents of
this report are reproduced herein, as received' from the contractor.
The opinions, findings, and conclusions expressed are those of the
authors and not necessarily those of the Environmental Protection
Agency.
-------
ACKNOWLEDGEMENTS
This report was prepared under the direction of Mr. John
Mutchler with the assistance of principal authors Thomas Loch,
Fred Cooper, and Janet Vecchio of Clayton Environmental Consult-
ants, Inc. The Project Officer for the U.S. Environmental Pro-
tection Agency was Mr. Kirk Foster. The authors are grateful to
Mr. Foster for his recommendations, comments, and review through-
out the execution and report-development phases of the study. The
authors also appreciate the valuable contributions of Bernard Bloom
and Louis Paley of EPA to this project. Finally, the assistance
of individuals from the following offices at the field-study site
is very gratefully acknowledged: the management of Great Lakes
Carbon Corporation, St. Louis, Missouri; the City of St. Louis
Division of Air Pollution Control; and the Region VII Enforcement
Division (Kansas City, Missouri), the Region V Surveillance and
Analysis Division (Chicago, Illinois), and the National Environ-
mental Investigation Center (Denver, Colorado) of the Environmen-
tal Protection Agency.
-------
TABLE OF CONTENTS
VOLUME 1 Pa8e
LIST OF APPENDICES iii
LIST OF FIGURES V
-LIST OF TABLES vl
GLOSSARY OF TERMS vii
1.0 INTRODUCTION
1.1 Ba ckground ......................................... 1
1.2 Purpose and Scope .................................. 2
1.3 Limitations ........................................ 4
2.0 SUMMARY AND CONCLUSIONS ................................. 4
2.1 Particulate Emission Factors and Rates ............. 5
2.1.1 In-Duct Emissions During Pushing Cycle ...... 5
2.1.2 In-Duct Emissions During Non-Pushing Cycle.. 5
2.1.3 Overall Emissions Due to Pushing Operation
Only ...................................... 6
2.1.4 Overall Emissions Due to Door Leaks Only.... 6
2.2 Particulate Capture Efficiency of the Shed ......... 6
2.3 Composition of Particulate Emissions ............... 7
2.4 Particle Size Distribution ......................... 8
2.5 Emission Rates of Other Materials .................. 8
2.6 Dustfall Measurements .............................. 8
2.7 Indices of Visible Emissions ....................... 9
2.8 Process and Emissions Correlations ................. 10
3.0 PROCESS AND OPERATIONS DESCRIPTION ...................... 11
3.1 Description of the Coking Process .................. 11
3.2 Description of the Shed Capture System ............. 14
4.0 SAMPLING AND ANALYTICAL METHODS ......................... 20
4.1 Location of Sampling Points ........................ 20
4.2 In-Duct Particulate Emissions ...................... 22
4.3 Fugitive Emissions ................................. 26
4.4 Particle Size Distribution ......................... 27
4.5 Emissions of Other Materials ....................... 28
-------
ii
4.5.1 Sulfur Dioxide and Sulfur Trioxide 28
4.5.2 Gaseous Contaminants by Charcoal Tube
Collection ' 28
4.5.3 Polynuclear Aromatic Compounds 28
4.5.4 Gaseous Contaminants by Collection in Gas
Burette- - 29
4.5.5 Gaseous Contaminants by Collection in
• • Aqueous Sodium Hydroxide 29
4.6 Dustfall Measurements 30
4.7 X-ray Fluorescence and Microscopic Analysis 32
4.8 Visible Emissions Monitoring 32
4.8.1 Degree-of-Greeness Ratings 32
4.8.2 Stack Opacity Rating >. : 34
4.8.3 Transmissometer Data 34
4.8.4 Door Leak Inspection Data ' 37
4.9 Calibration of Sampling Equipment and
Example Calculations 37
4.10 Quality Assurance and Chain of Custody 38
5.0 PRESENTATION AND DISCUSSION OF RES-ULTS 39
5.1 Comparison of Pushing-Cycle and Non-Pushing-Cycle
Particulate Tests 39
5.2 Calculation of Emission Factors 46
5.2.1 Emission Factor for'Coke Oven Pushing 46
5.2.2 Emission Factor for Door Leaks 46
5.2.3 Overall Emission Factor 48
5.3 Significance of Fugitive Leaks 49
5.4 Chemical and Physical Characteristics of
Particulate Emissions 53
5.5 Particle Size Analysis 60
5.6 Door Leak Rates 74
5.7 Emission-Related Correlations 74
5.7.1 Correlations Between Pushing-Cycle
Filterable Particulate Emission Factors and
Operating Data 74
5.7.2 Correlations Between Pushing-Cycle Filterable
Particulate Emission Factors and Indices
of Visible Emissions • 80
5.7.3 Correlations Among Visible Emissions
Parameters 84
5.8 Significance of Emissions of Other Contaminants 86
-------
iii
Page
5.9 Assessment of the Shed's Impact Upon Dustfall
in the Work Environment 89
5.10 Impact of the Shed Upon Airborne Agents Within 103
5.11 Precision of Test Results 103
REFERENCES 106
VOLUME 2
A Emission-Measurement Project Participants
B Pushing-Cycle Particulate Test Results
C Non-Pushing-Cycle Particulate Test Results
D Particle Size Distribution Data
E Emission Results of Gases and Other Materials
F Sampling Summary Sheets
G X-Ray Fluorescence Spectrometer Analysis of
Coke-Side Particulate Emissions
H Microscopic Analysis of Coke-Side Particulate Emissions
I EPA Report of Continuous Opacity Measurements Using a
Transmissomete r
VOLUME 3
J Method of Pitot Tube Calibration
K Pitot Tube Calibration Data
L Method of Meter and Orifice Calibration
M Meter and Orifice Calibration Data
N Method of Temperature Sensor Calibration
0 Method of Nozzle Calibration
P Particulate Sampling Train Data Sheets
Q Sampling and Analytical Method — Sulfur Dioxde and
Sulfur Trioxide
-------
iv
R Determination of Various Emissions Adsorbed on Activated
Carbon from Coke-Oven Pushing
S Determination of Various Emissions Absorbed in Cyclohexane
from Coke-Oven Pushing
T Determination of Various Emissions Captured in a Glass.
Gas Burette During Coke-Oven Pushing
U Determination of Various Emissions Absorbed in
Sodium Hydroxide from Coke-Oven Pushing
V Coke Pushing Evaluation Data
W Quench Tower Opacity Data
X Transmissometer Chart Recordings
Y Door Leakage Data
Z Example Calculations
AA Chain of Custody
BB Wind Observation Data Taken at GLC Coke Plant Site
CC Wind Roses — Lambert Field
-------
LIST OF FIGURES
FIGURE 3.1
FIGURE 3.2-1
FIGURE 3.2-2
FIGURE 3.2-3
FIGURE 4.1
FIGURE 4.2
FIGURE 5.5-1
FIGURE 5.5-2
FIGURE 5.6
FIGURE 5.7.1-1
FIGURE 5.7.1-2
FIGURE 5.7.2-1
FIGURE 5.7.2-2
FIGURE 5.7.2-3
FIGURE 5.7.2-4
FIGURE 5.7.3
Page
Location of Great Lakes Carbon Plant in
Southern St. Louis, Missouri 13
Configuration of North End of Shed 16
Diagram of Side View of Shed 17
Schematic View of Sampling Site 18
Location of Sampling Points — Coke-Side
Shed Exhaust Duct 21
Particulate Analysis Flowchart 23
Particle Size Distribution
(Brink Tests 1-9) 61
Particle Size Distribution
(Andersen Tests 10-14) 62
Coke-Side Door Leaks After Oven Charging 76
Effects of Coking Time on Particulate
Emissions — Coking Time Versus
Filterable Particulate Emissions 78
Average Oven Temperature Versus Filterable
Particulate Emissions 79
Degree-of-Greenness Versus Filterable
Particulate Emissions 81
Maximum Attenuation Coefficient Versus
Filterable Particulate Emissions 82
Plume Attenuation Coefficient Versus
Filterable Particulate Emissions 83
Filterable Particulate Emissions Versus
Average Quench Tower Opacity 85
Degree-of-Greenness Versus Maximum
Attenuation Coefficient 87
-------
vi
TABLE 5.1-1
TABLE 5.1-2
TABLE 5.2.1
TABLE 5.3-1
TABLE 5.3-2
TABLE 5.4-1
TABLE 5.4-2
TABLE 5.4-3
TABLE 5.4-4
TABLE 5.5-1
TABLE 5.5-2
TABLE 5.8
TABLE 5.9-1
TABLE 5.9-2
TABLE 5.9-3
TABLE 5.9-4
LIST OF TABLES
Page
Summary of Farticulate Emissions 40
Push Characteristics (Particulate Tests) 42
Summary of Particulate Emission Factors 47
Summary of Fugitive Emission Estimation—
itorth End of Shed 50
Shed Particulate Capture Efficiencies 51
l
Emission of Particulate Contaminants—
Pushing Cycle 54
Emission of Particulate Contaminants—
Non-Pushing Cycle ' ' - 55
Characterization of Particulate Weight 56
Summary of Water Soluble pH and' Acidity/
Alkalinity on Particulate Samples 59
Push Characteristics (Particle Size Tests) 63
Concentration of Particulate Matter Calculated
from Particle Sizing Samples 75
Summary of Contaminant' Emission Rates 88
Summary of Dustfall Measurements 90
Chemical Characterization of Dustfall 97
Dustfall Summary • 98
Format Used for Analyses of Dustfall Data 100
-------
vii
GLOSSARY OF TERMS
1. Atypical operating conditions
Extremely infrequent major process change (or upset).
2. Average rate of coke-side particulate emissions
The sum of the particulate emissions captured by the shed and
the emissions which are fugitive from the shed.
3. Coke side
That side of a coke-oven battery from which the ovens are
emptied of coke.
4. Degree of greenness of a coke oven push
A subjective, visual estimate of the quantity of particulate mat-
ter released during a single coke oven push by estimation of the
apparent visibility of the plum immediately above the quench car,
5 . Door leakage
Any visible emissions observed emanating from coke-side oven
doors, push-side oven doors, or push-side chuck doors.
6. Filterable particulate
Material captured on or before the front filter of a particu-
late sampling train.
7. Fugitive particulate emissions
Particulate emissions which escape capture from the shed and
pass unrestrained into the atmosphere.
8. Green coke
Coke which, when pushed from an oven, produces copious quan-
tities of visible emissions, particulate matter, and/or flame
on the coke side of the battery.
9 . Net coking time
The elapsed time in minutes between the charging of a coke
oven with coal and the pushing of that same oven.
-------
viil
10. Non-pushing cycle
That portion of the repetitive coke pushing operation outside
the pushing cycle. This period includes the time during which
no push has occurred beneath the shed (A Battery) for 30 minutes.
(During this period coke ovens on the C Battery were normally
being pushed.)
\
11. Precision of a test result
The statistical confidence interval associated with the mean
value of a series of replicate measurements at a risk level
of five percent.
12. Pushing cycle
That portion of the coke pushing operation during which ovens
beneath the shed (A Battery) were being pushed at a regular
interval of approximately one oven every 23 minutes up to 30
minutes beyond the time of the most recent push.
13. Settleable particulate
That material collected in a cylinder whose height is two to
three times its diameter and which passes through a No. 18
(1 mm) sieve, ASTM Method 1739-70.
14. Total particulate
Filterable particulate plus that material captured in imp Ingers
containing distilled water immediately following the filter
in the sampling train.
15. Transmissometer
A device, utilizing a light source and a light detection cir-
cuit, which provides a measurement of the transmittance of
stack gas passing between the light source and the detector.
16. Typical operating conditions
Any process operating conditions which are not atypical.
-------
1.0 INTRODUCTION
1.1 Background
The Division of Stationary Source Enforcement, United States
Environmental Protection Agency (EPA) retained Clayton Environ-
mental Consultants, Inc. to conduct a study of coke-side emissions
at coke-oven batteries producing foundry coke at Great Lakes Car-
bon Corporation (GLC) in St. Louis, Missouri. One of three bat-
teries of GLC was equipped with a shed-type enclosure designed to
contain particulate and gaseous emissions produced on the coke
side of the battery during coking and coke pushing. An induced
draft fan exhausts the shed enclosure through ductwork to the
quench tower for discharge to the atmosphere. At the time of
the study, no control device other than improvised spray headers
in the ductwork and quench tower .was included in the control system.
to abate emissions in the shed exhaust gas.
Foundry coke is produced by three batteries of ovens at the
GLC plant. The south battery ("A") is equipped with the coke-
side shed. The center battery ("B") and north battery ("C") were
not equipped with a functional shed at the time of the study.
During this study, B Battery was being rebuilt; only the 40-oven
A Battery and 35-oven C Battery operated during.the testing pro-
gram. All three coke batteries at GLC are similar in construction,
capacity, and operation. Furthermore, all three are served by a
single work crew using, a single set of charging equipment and a
single quench car.
At the time of the study, construction of a shed over the B
and C Batteries was in progress. Nevertheless, coke-side emis-
sions from C-Battery ovens escaped directly to the atmosphere and
-------
- 2 -
were not captured by the shed, nor did they affect results of the
sampling in the exit gases of the A-Battery shed.
Exhaust gas sampling was conducted primarily in the A-Battery
shed exhaust duct using EPA standard source testing methods, or
similar methods modified to suit this particular source, to
measure particulate and gaseous emissions during the test program.
Additionally, the particulate emissions from the coke side of the
coke-oven battery which escaped capture by the shed were measured.
Process operating conditions were monitored to assure the collec-
tion of representative samples with respect to "typical" operating
conditions. The source testing results were correlated with
process data and other, secondary observational data auxiliary to
the emission measurements. The results of the field study and
analysis of the data are presented in Volume 1, while all raw data
and background data are provided in Volumes 2 and 3.
The field study was conducted during the week of April 21,
1975 by the staff of Clayton Environmental Consultants. Messrs.
Kirk Foster, Louis Paley, and Bernard Bloom of the Division of
Stationary Source Enforcement, U.S. EPA, and Messrs. Edward Roe
and George Shell of Great Lakes Carbon Corporation provided coordi-
nation with the plant operation. A listing of project participants
and their respective roles in the study is included in Appendix A
(Volume 2).
1.2 Purpose and Scope
The purpose of this study was to provide basic engineering
data concerning the quantities and characteristics of air-contami-
nants emitted from the coke side of the A-Battery coke ovens,
-------
- 3 -
and further, to evaluate the performance of the shed in capturing
coke-side emissions.
The scope of the study included the measurement of emissions
from the A-Battery shed and the monitoring of process parameters
which may affect or be related to emission rates. In addition to
the emission tests, dustfall measurements were collected beneath
the shed and in similar locations near the ovens of the nearby
unshedded C-Battery. Additional emission parameters were moni-
tored by EPA personnel, including the "degree-of-greenness" of
each push beneath the shed during sampling, visual opacity of the
quench tower exit gases (the ultimate point of discharge to the
atmosphere from the coke-side shed), and optical density, measured
with a transmissometer installed temporarily on the exhaust gas
duct and located between the shed and the quench tower.
The EPA emission testing program focused primarily on the measure-
ment of gaseous and particulate emission rates, and characterization
of the chemical species and size distribution of particulate contami-
nants in the duct exhausting the emissions from the shed capture
system. Measured contaminants included:
1. Particulate during the coke pushing cycle;
2. Particulate during the non-pushing cycle;
3. Particle size distribution during the pushing cycle;
4. Sulfur dioxide;
5. Sulfur trioxide;
6. Polynuclear aromatic hydrocarbons;
7. Carbon monoxide;
8. Gaseous hydrocarbons; and
9. Phenolics.
-------
- 4 -
1.3 Limitations
. This comprehensive emission study was neither intented nor
designed to include an evaluation of the effect of the coke-side
shed pn th.e occupational environment.. With the exception of the
dustfall, measurements collected beneath the shed, .the ,study effort
dea.lt mainly with the quantity and characteristics of contaminator
presen.t,,in, the shed, exhaust,. Thus, any definitive evaluation of
related occupational exposure wi.thin this, or any coke-side shed
would 'be supplementary to the study .reported herei,n.
2.0 SUMMARY AND CONCLUSIONS
To fulfill the purpose of this study, and therefore provide
basic engineering data concerning process emissions, fugitive
emissions from the shed, and capture efficiency of the shed, the
measured findings in the study and the field data have been ana-
lyzed with respect to emission factors and emission rates attrib-
utable to: pushing and non-pushing cycles, fugitive particulate
emissions, door leaks, and the overall pushing operation. Deter-
mination of these emission data required estimation and calculation
of the shed's capture efficiency for filterable particulate emis-
j •
sions. Additionally, other basic engineering data necessary for
the specification of (future) retrofitted collectors installed on
the shed exhaust were collected and included the measurement of
particulate emissions'composition ,r particle size 'distribution, '
and the determination of exhaust stream composition as affe'cted by
other species of contaminants detected in the shed exhaust. Finally,
in an attempt to relate these measurements to process conditions and
thereby enable cautious application of these results to other coke-
oven batteries, correlations were attempted between various' process
parameters and the computed emission factors.
-------
- 5 -
2.1 Particulate Emission Factors and Rates
2.1.1 In-Duct Emissions During Pushing Cycle
Filterable particulate emission measurements made in
the duct evacuating the shed during the time when ovens were
pushed beneath the shed indicated that the average emission
factor is 0.38 pound per ton of dry coal charged to the ovens
pushed (+0.24 pound per ton).* The corresponding average
emission rate during the time when pushing of ovens was occur-
ring beneath the shed indicated that an average of 16.7 pounds
per hour (+8.8 pounds per hour) of filterable particulate were
emitted. These estimates inherently exclude fugitive emissions
due to shed leakage and inherently include door leakage emis-
sions.
2.1.2 In-Duct Emissions During Non-Pushing Cycle
Particulate emission measurements made when no ovens
were being pushed beneath the shed indicated an average emis-
sion factor due to door leaks of 0.36 pound per ton of dry
coal charged to all ovens beneath the shed (+0.40 pound per
ton). The corresponding emission rate occurring during the
time when no ovens were being pushed beneath the shed averaged
6.9 pounds per hour of filterable particulate (+_7.6 pounds
per hour). These estimates inherently exclude fugitive emis-
sions due to shed leakage and inherently include only door
leak emissions.
* The notation (+0.24 pound per ton) is an estimate of the sta-
tistical precision of the average value based upon a 95-percent
level of confidence. Although the precision is +^0.24, the con-
fidence interval for a concentration, emission rate, or emis-
sion factor is always bounded by a minimum value of zero. Like-
wise, the corresponding confidence interval for a percentage is
always bounded by a maximum value of 100 percent. (See Section
5.11)
-------
- 6 -
2.1.3 Overall Emissions Due to Pushing Operation Only
The push-only emission factor for filterable particulate
emissions, including estimated fugitive emissions but excluding
door leaks, averaged 0.25 pound of filterable particulate per
ton of dry coal fed to the ovens pushed (+0.35 pound per ton).
The corresponding overall emission rate of filterable particu-
late due to the pushing operation (including fugitive emis-
sions) averaged 10.7 pounds of filterable particulate per hour
(+14.5 pounds per hour).
2.1.4 Overall Emissions Due to Door Leaks Only
Because the shed capture efficiency was estimated to
be 100 percent for door leak emissions, the overall emission
factors (i.e., including fugitive emissions) and emission
rates for door leak emissions are identical to those presented
in Section 2.1.2 where in-duct measured emissions during the
non-pushing cycle are documented. (See Section 2.2 for shed
capture efficiencies.) No fugitive emission measurements
were conducted during non-pushing periods; rather, the esti-
mated non-pushing capture efficiency is based upon visual
determination.
2.2 Particulate Capture Efficiency of the Shed
The efficiency of the shed in capturing and exhausting coke-
side emissions from pushing ranged from 81 to 98 percent, and aver-
aged 91 percent (+12 percent). Fugitive emissions during periods
when ovens were not being pushed were not measured, but were esti-
mated visually to be minimal. Assuming that no fugitive particulate
-------
- 7 -
escapes the shed during the non-pushing cycle, the overall effi-
ciency of the shed in capturing particulate ranged from 92 to 99
percent, and averaged 96 percent (+5 percent).
Wind speed and direction affected the location and extent of
end leaks (smoke emissions that escape from under the shed) from the
Great Lakes Carbon coke-side shed. Particulate emissions from the
pushing of coke ovens are less likely to be collected by the shed
capture system if the oven being pushed is located on the downwind
end of the shedded coke battery. End-leak measurement estimates of
particulate materials escaping the collection system from the north
end of the shed on April 23, 1975, ranged from 2 to 19 percent of
the overall (duct plus fugitive) particulate emissions during pushing,
and averaged 9 percent (+12 percent).
2.3 Composition of Particulate Emissions
Eighty-seven percent (+9 percent) of the total particulate was
captured as filterable particulate, the remaining 13 percent (+9
percent) was captured in the impinger (back-half) portion of the
sampling train. Cyanide, chloride, and sulfate accounted
for minor portions of filterable and total particulate during both
pushing-cycle and non-pushing cycle particulate tests. For both
the pushing and non-pushing-cycle particulate tests, 87 percent
(+7 percent) of the filterable particulate was inorganic, that is,
insoluble in cyclohexane or acetone. However, only 22 percent (+_16
percent) of the impinger catch material was inorganic. Although
carbonaceous material apparently constituted the majority of fil-
terable particulate, x-ray fluorescence indicated that chlorine,
sulfur, silicon, and aluminum were also present in the filterable
particulate.
-------
- 8 -
2.4 Particle S-ize Distribution
Variation in particle size distribution measured during each of
several tests correlated poorly with net coking time, possibly due
to multiple pushes being captured in each particle size test: A
statistically significant correlation was found, however, between
oven temperature and the percentage of particles less than five
microns in diameter. Size distributions measured by the Brink and
Andersen impactor methods indicated that 10 percent (+3 percent)
and 13 percent (+4 percent), respectively,' of the particulate was
submicron in diameter as emitted during pushing-cycle tests. ;
-,
2.5 Emission Rates of Other Materials
• . \* '
Coke-side emission rates of gaseous substances and other contami-
nants from this source were minor. Polynuclear aromatic compounds and
those with similar structures (such as pyrene) were not found- in
detectable quantities. Sulfur dioxide plirs sulfur trioxide emis-
sion rates ranged from 1.7 to 4.2, and averaged 2.8 pounds per
hour (+3.2 pounds per hour). The emission rate of carbon monoxide
at the peak during the push ranged from 8 to 24, and averaged 14
pounds per hour (+_21 pounds per hour). Total light hydrocarbon
emissions during-peak emissions- averaged seven pounds per hour
(+6 pounds per hour.) .
2.6 Dustfall Measurements
i
For two of the three locations considered, dustfall (settle-
able particulate) rates beneath the shed were statistically greater
than those at corresponding locations in the unshedded C Battery.
As expected, greater dustfall rates were experienced at the A Battery
-------
- 9 -
near the shed wall than at locations nearer the bench. In contrast,
greater dustfall rates at the C Battery areas were found at the bench
location than at the site equivalent to the wall location on A Battery.
2.7 Indices of Visible Emissions
Statistical analyses indicate that pushing-cycle filterable
particulate emission factors were statistically significantly corre-
lated with the average "degree-of-greenness" rating for pushes
observed during the pushing-cycle particulate tests. No statisti-
cally significant linear correlation could be established, however,
between quench tower plume opacity (the discharge stack for the shed
exhaust) and pushing-cycle filterable particulate emission factors.
This lack of correlation may have been due to the small number of
particulate tests available as well as the limitations involved with
reading the plume opacity in the presence of the steam plume from
the quenching operation.
One index used to characterize the optical density of the shed
exhaust in the duct as it varied during the course of the push was
the average of the maximum attenuation coefficients of the pushes
included in the multi-push particulate test. No statistically
significant linear correlation was apparent between this index and
the pushing-cycle filterable particulate emission factor, likely due
to the limited number of particulate tests available. The attenuation
coefficient integrated over time, however, was found to be signifi-
cantly correlated with the pushing-cycle filterable particulate
emission factor. It is therefore concluded that increased optical
density (manifest by integrated attenuation coefficient or the degree-
of-greenness rating) accompanied elevated filterable particulate
emission factors measured during the four pushing-cycle particulate
tests in this study.
-------
- 10 -
Correlations were also examined among the four indices of
visible emissions monitored independently in the project: degree-
of-greenness, maximum attenuation coefficient, Integra ted. a ttenua-
tion coefficient, and quench tower opacity. Statistical analyses
between various combinations of these variables suggest that all
k. >
combinations are highly interrelated. These results indicate that,
for example, the integrated attenuation coefficient is statistically
*1
significantly correlated to the quench tower opacity, the degree-of-
greeness rating is statistically significantly correlated to the
quench tower opacity, and the maximum attenuation coefficient is
statistically significantly correlated to degree-of-greenness .
2.8 Process and Emissions Correlations
Observations of coke-side door.leaks indicated that door
leaks more likely occurred during the.initial coking period, after
oven charging, than in the later hours of the coking period.
Pushing-cycle filterable particulate emission factors were
found to be significantly correlated with average net coking time
but were not significantly correlated with average oven tempera-
ture.
Temperatures of ovens pushed during particle sizing tests
were found to be significantly correlated with the percentage of
particles less than five microns in diameter but not with the
percentage of submicron particulate. No correlation could be
found between cue particle size distribution and the net coking
time of ovens pushed during particle sizing tests.
-------
- 11 -
3.0 PROCESS AND OPERATIONS DESCRIPTION
3.i Description of the Coking Process
Coking is a process by which coal is destructively distilled
in an atmosphere of low oxygen content to produce volatile gases
and a residue of relatively non-volatile coke. In the byproduct
coke production process (constituting more than 90 percent of the
coke produced in the United States), the gases and volatile matter
distilled from the charged coal are recovered throughout the coking
cycle, processed, and partially recycled to the ovens for use as
fuel.
A contiguous series of rectangular chambers, coke ovens,
separated by heating flues placed between the ovens, constitutes
a coke "battery." Based upon production requirements and hard-
ware available at a given battery, ovens are charged, coked, and
pushed according to a relatively fixed schedule. Coking times
for the production of foundry coke can range from 25 to 32 hours,
with the ovens being maintained at a temperature between 1800 and
2400°F throughout the period.
During the coking cycle, volatiles are driven from the charged
coal beginning at the oven walls and proceeding toward the center
of the charge. When the charge is "fully coked out," a ram opera-
ting from the "push side" of the oven forces the coke through the
oven and out the "coke side" of the oven where the incandescent
coke passes through a temporarily-aligned coke guide and falls
into a quench car. The incandescent coke is subsequently quenched
using water sprays in a quench tower generally positioned at or
near the end of the battery.
-------
- 12 -
Great Lakes Carbon Corporation, a producer of foundry coke,
is located on the south side of St. Louis, Missouri, adjacent to
the Mississippi River and River DesPeres. Figure 3.1 shows the
location of the plant relative to its immediate surroundings. The
' i ' •
GLC plant produces coke from coal which is unloaded from river
barges or from railroad cars. The coke product is transported
from the GLC plant by rail.
The ovens in the three batteries are serviced by one larry
car, two pushing machines, two door machines on the push side,
two door machines with coke guides on the coke side, and one
quench car. Again, "B" Battery was inoperative during the study
because it was being rebuilt. Therefore, the availability of
charging and pushing machines to the other two batteries was
somewhat more optimal than normal operations.
At GLC, the charge car is filled with approximately 13.7 tons
of dry coal per charge. During the testing program, charging of
an oven normally occurred 15 to 20 minutes after that oven had been
pushed and the doors replaced. Net coking times averaged approxi-
mately 28 hours. Thus, the 75 operating ovens were pushed at an
average interval of about 23 minutes.
The normal sequence of oven pushing usually resulted in five
or six ovens being pushed beneath the shed (Ovens 1 through 55),
followed by five ovens being pushed north of the production office
in the unshedded C-Battery area. A typical sequence of oven push-
ing was: 2, 12, 22, 32, 42, 52 (A Battery); 92, 102, 112, 122,
132 (C Battery); 4, 14 ...
Sources of emissions which contribute to the materials cap-
tured by and exhausted from the shed include:
-------
c-:>
FIGURE 3.1
LOCATION OF GREAT LAKES
CARBON PLANT IN SOUTHERN
ST. LOUIS, MISSOURI
April 21-24, 1975
<--4/
K .X:. -----
^f^^y
i •-.
••\^
„• ' ."^
1/&
;*%&
=./•/ :-:••/. /&;•••'••?.
*V ' •••$*•• X..:.;..„.
^e^c. :.-.j!- */.,7:.>s
3^
/ -•/ "
\ • •// ; \
*a v^
Great Lakes
Carbon Corporation-
f/^-SK^jf
•: •• K ^ vnv^-f-Uv^'''C-;^;"s - V/^
p A /••" .--' - /.<•'A»//' x/ V"^
v';^,r^
' -- Sfc. TJ ' x.
>J^
-'r^s^?>'
• K K • J
TV. A' •'-.-' - />"'"»/•/* x/ X
v^./v^'-- .--.'• i4'.<-/
~^>C •,(</,/ ' .'
A IfffJ'Z:'
\\ £//•', srt >
fef'^
; ••//• .*.< /' •
/ -'-* S?
', /i
\ .^ X''^'
\
^CiO _•
:> -^ 0
• v-0-
w
" r\ I
);7
vl.' ••',
o •/
//-/ /
^ /
•' '
1 /
/• /
' /• /
/
••/ . /
vw- ^AH
!:M-V;-Vu /-
mo
^^
v 0\w,
-,^ /•».--• »•" .
i/11
5
-\, •'-,-;~ • ~"-" \\
".' .;-''^Vrx.--'r*?Vr-\
--. -!i; ' ^,:':'^v-^-
5\^^i ^
!AMs
•o.
--^
•v\ -i;
'.-R^- ''iV'"' • .-•?
'/V-
'Vj.'
VvovV
;,v :l
'i
^
^
s
'--| |
;" /.i^-^V V.;-;::/1^rU
- r:.-, .-••- C'^T: V ' .••-"' •'••• -%tV
•;-^:^^M-^
,/
•/
Jk>
^'"•'^1-t/
K; "^,
/I
o
p
f/J
ry//.\'
«vW
;i ay t on E nvi:
-;•/ /-^...;/-.;>'
• '• '"•.• ;.'.•••''/ /' .•-.,,. ] L
-------
- 14 -
1. Pushing Operations:
a. Emissions from the coke side of an oven whose door has
i
been removed, .before and after pushing;
b. Emissions from.and during pushing of the hot'coke
thrcugh the coke guide into the quench car;
c.. Emissions from the .newlyrfilled quench car immedi-
ately following pushing and before the car leaves
the shed as it. t.ravels to the quench tower;
2. Door Leaks:
Emissions from leaking coke-oven doors after the oven
• ' ' '
has been charged with coal and placed under positive
pressure during the conversion of the'coal, to coke.
The A Battery at the GLC plant contains 40 Simon-Carves ovens south
• ' i '
of the control room (Ovens 1 through 55), and the .C Battery contains
an additional 3.5 Wilputte ovens north of the office (Ovens 83
through 132). As indicated', ovens 56 to 82 (B Battery) were under-
going repair' and were not coking at the time of the study .(coke-
oven numbering system at GLC excludes 8's, 9's,.and O's in the
last digit).
Plant personnel at GLC , indicated that during this study, coke
I
batteries A and C operated -at typical condition*. Clean, as well as
green, pu.-.hee were experienced during the sampling phase of the
study.
3 .2 Description of the Shed Capture System
. The shed capture system on Battery A is constructed of corru-
gated metal on a steel frame. It covers the coke-side bench and
-------
- 15 -
part of the quench car tracks and extends from approximately 25
feet beyond Oven 1 to approximately 15 feet beyond Oven 55. The
shed does not extend to the ground or bench level on the side or
at the ends because the coke guide car and the quench car must
move in and out of the structure during the production cycle. The
side of the shed extends vertically down to approximately 10 to 11
feet above grade, slightly below the top of the outer wall of the
quench car. A sketch of the north face of the shed, which must
allow clearance for the coke guide car and quench car, is shown in
Figure 3.2-1. A detailed drawing of the side view of the shed is
shown in Figure 3.2-2 to give an overall perspective of the general
appearance of the structure.
Exhaust gases are evacuated from the shed through a variable
cross-section, rectangular duct that extends the entire length of
the shed immediately beneath the shed's peak (Figure 3.2-3). Air
scoops are located along the sides and bottom of this duct in such
a way as to allow the duct to capture the exhaust gases along its
entire length inside the shed. A vane-axial fan draws the exhaust
gas from the shed through a rectangular duct with a slight downslope
to the front face of the quench tower. The ultimate point of exit
to the atmosphere of the shed exhaust gas is from the top of the
quench tower. During normal operations, water is sprayed from
nozzles placed along the length of this rectangular duct downstream
of the fan (additional water is sprayed from the top of the quench
tower). Emission samples were collected in this rectangular duct
during the test program. Therefore, to allow better measurement
of the coke-side emissions as captured by the shed, the water to
-------
-16-
FIGURE 3.. 2-1
CONFIGURATION OF NORTH END OF SHED
Great Lakes Carbon 'Corporation
St. Louis, Missouri
April 21-24, 1975
"Overhead"
Dust fall
Sampling
Location
"Wall"
Dustfall
Sampling
Location
Coke
Guide
Car
Q
Quench
Car
Oven
"Bench"
tpust£ail
Sampling
Location
-------
FIGURE 3.2-2
DIAGRAM OF SIDE VIEW OF SHED
Great Lakes Carbon Corporation
St. Louis, Missouri
April 21-24, 1975
ri
/
7
•™=^— _ji
-^-^n Duct
^ —
%r^
Shed
^
X
X
X
X
X
X
X
X
XIX
X
1
-------
Sampling Site A
Ports
Sampling Site B
Gaseous
Sampling
Port
FIGURE 3.2-3
SCHEMATIC VIEW OF SAMPLING SITE
Great Lakes Carbon Corporation
St. Louis, Missouri
April 21-24, 1975
.Platform
(railing
not shown]
90"x90"
Duct
Vane Axial Fan
Transmissometer Location
Quench car
Tracks
Quench
Tower
Scaffolding
-------
- 19 -
the spray nozzles in the rectangular duct was turned off during
the sampling period.
Because the ends and the sides of the hood are not completely
enclosed, so as to permit the door machine and quench car to enter
and exit, the capture efficiency of the hood is less than 100 per-
cent. During the pushing of an oven, a black plume was seen to
rise to the upper portions of the shed and some of the particu-
late emission was seen to escape from the side and the ends of
the shed. Wind speed and direction obviously affected the rate
of emissions escaping from the shed system. A southerly wind
likely results in particulate emissions from the north end of the
shed, especially when the oven being pushed is near the north end.
Similarly, northerly winds enhance particulate escape at the
south end of the shed.
-------
- 20 -
4.0 SAMPLING AND ANALYTICAL METHODS
4. 1 Location of Sampling Points
Sampling of particulate, particle sizing, and measurement of
exhaust gas velocity and flowrate were conducted at the uniform
airflow profile located at cross section A shown in Figure 3.2-3.
Sampling for substances other than particulate, such as sulfur
oxides, polynuclear aromatics, etc., was conducted in the more
turbulent airflow stream located at cross section B (Figure 3,2-3).
near the inlet o,f the vane-axial fan. The dimensions of the duct
at location A were 89.-3/411 by 84" with an equivalent duct diameter
of 8.16 feet. This location is therefore six equivalent diameters
downstream of any bend or obstruction and 1.5 equivalent diameters
upstream of the quench tower. An independent velocity traverse at
this location indicated no spiraling airflow patterns in the rec-
tangular cross section at location A as might result from the nearby
vane-axial fan. Figure 4.1 indicates the location of sampling
points in the duct cross section. These points were accessible
through two sets of four ports located on the west side of the
duct, one set of ports for each of the two particulate test modes
(pushing and non-pushing).
Velocity pressure measurements taken at sampling cross section
A were made using a stardard S-type Pitot tube. Temperature meas-
urements were made using an iron-constantan thermocouple attached
to a calibrated Mini-mite potentiometer. All calibrations are in-
cluded in Appendices J through 0 (Volume 3) and discussed further
in Section 4.9.
-------
- 21 -
FIGURE 4.1
LOCATION OF SAMPLING POINTS
COKE-SIDE SHED EXHAUST DUCT
Great Lakes Carbon Corporation
St. Louis, Missouri
April 21-24, 1975
.0"
•>
52.2"
i
-1- H
+ H
;
1
I
h + H
h 4- H
^r~ J / » J
h H
i_ i
r "1
;
\f ?2 4" V
«;- i. Z. . ' 1 ,(»
•~^i7»5 "'
'
h -f
^—
4 A A A
10.5"| \ ;
\ !
; j
31.5"! :
i
. ! '
ty i i
1 I
52.5";
84"
+
89-3/4"
+ +- + 4- 4- -h
+• 4- -h
77.5"
i
Y .
-------
- 22 -
4.2 In-Duct Particulate Emissions
Particulate sampling methods follow the guidelines outlined
in EPA Methods 1 through 5.'*' Deviations from these procedures
included the following:
1. An abbreviated number of sampling -points was chosen in
order to complete one particulate test per day for each
of the two modes (pushing-cycle and non-pushing-cycle).
2. An integrated sample of the stack gas was not analyzed
for each particulate test by the standard Orsat procedure,
Before the testing began, however, an Orsat analysis of
stack gas collected during a coke oven push indicated
that the composition of the stack gas was essentially
that of air.
3. Collected particulate samples were not simply weighed
but were analyzed as well for other components as out-
lined in the particulate analysis flowcharts (Figure
4.2). Particulate captured by the impingers was in-
cluded in "total particulate," whereas "filterable par-
ticulate" only included the probe and cyclone washes
plus the filter catch.
4. Filter and probe temperatures were not maintained at
250CF. Temperatures were adjusted to slightly above
stack temperatures to assure that no moisture condensa-
tion occurred in the train upstream of the filter.
The "pushing-cycle" particulate tests refer to samples ac-
quired during those times when the ovens beneath the shed capture
-------
- 23 -
FIGURE 4.2
PARTICULATE ANALYSIS FLOWCHART
Great Lakes Carbon Corporation
St. Louis, Missouri
April 21-24, 1975
Pushing-Cycle Particulate Test 1 &
Non-pushing-Cycle Particulate Test 1
Probe
Wash
dry & weigh
dry & weigh
combine
extract with acetone
filter through tared filter
Acetone
Insolubles
' extract with cyclohexane
filter through tared filter
I
weight by difference
I
extract with cyclohexane
Cyclohexane
Solubles
I
Cyclohexane
Insolubles
Cyclohexane
Solubles
Cyclohexane
Insolubles
dry & weigh weight by difference
extract with hot water
dry & weigh weight by difference
extract with hot water
Water
Solubles
Water Water
Insolubles solubles
Water
Insolubles
1 1 1 1 1
weight by difference dry & weigh weight by difference dry & weigl
Q © © ©
yOfc Aliquot 1
weight by sum
| so4=
1
1 I 101 Aliquot
1 1 1 -
1 pH 1 Acidity CN~
Cl"
-------
- 24 -
FIGURE 4.2 (continued)
PARTICIPATE ANALYSIS FLOWCHART
Great Lakes Carbon Corporation
St. Louis, Missouri
April 21-24, 1975
Pushing-Cycle Particulate Tests 2, 3, and .4 &
Non-pushing-Cycle Particulate Tests 2 and 3
Probe Wash
Filter
dry & 'weigh
dry & weigh
I
combine
extract with acetone
filter through tared filter
extract with cyclohexane
Acetone or
Cyclohexane
Solubles
I
dry & weigh
T
extract with hot water
Acetone or
Cyclohexane
Insolubles
I
weight by difference
extract with hot water
Water
Solubles
1 Water
Insolubles
1 1
ght by difference dry & weigh
© ©
90% Aliquo
weight by s
so4~
1
combine
I
t 1
um
Water | Water
Solubles Insolubles
1
weight by difference dry & weigh
[ PH [
10% Aliquot
1 1
Acidity
CN~ Cl"
-------
- 25 -
FIGURE 4. 2 (continued)
PARTICULATE ANALYSIS FLOWCHART
Great Lakes Carbon Corporation
St. Louis, Missouri
April 21-24, 1975
All Particulate Tests
^O Impinger Catch
extract witt
Aqueous
cyclohexane
Organic
Phase Phase
107 Aln nil of 1 907 AH
I
tquot cc
.... d rv
Acetone
Impinger Rinse
1
jmb ine
& weigh
s^\.
ci-
CN~
Acidity
PH
BB]
extract with acetone
filter through tared filter
Acetone
Solubles
extract with water
Acetone
Insolubles
extract with water
Water
Solubles
Water
Insolubles
Water
Solubles
1
Water
Insolubles
I
weight by
difference
I
dry & weigh ****ht by dry &,weigh
3 difference '
I
combine
weight by sum
SO,
-------
- -26 -
system were being pushed sequentially. During the normal pushing
cycle, five or six ovens beneath the shed would be pushed at approx-
imately 20- to 30-minute intervals (e.g., ovens 3, 13, 23, 33, 43,
and 53). Pushing-cycle particulate tests commenced as a push be-
neath the shed began, and the test terminated no more than 30 min-
utes after the racst recent push occurred beneath the shed. By
following this timing procedure, the approximate average pushing
rate beneath the shed was reflected in the pa.rticulate samples.
"Non-pushing-cycle" particulate tests were conducted only
when the pushing-cycle particulate samples were not being col-
lected (i.e., when oven pushing was occurring on the C Battery,
which is not under the shed). Therefore, these tests measured
the particulate generated from door Ieak3 only. To further in-
sure that pushing emissions were not captured during the non-
pushing-cycle test, that test was discontinued temporarily for
one-minute intervals each time the quench car, filled with hot
coke from the C Battery, traveled beneath the shed on its journey
to the quench tower. Also, the nor-pushing-cycle particulate
tests were discontinued as an oven beneath the shed was prepared
for pushing. Field data sheets for pushing and non-pushing-cycle
particulate tests are included in Appendix F (Volume 3). Summaries
of calculated sampling volumes, etc., are included in Appendix F
(Volume 2).
4 . 3 Fugitive Emissions
End leaks of emissions resultant from coke oven pushes at the
north end of the shed were estimated and later compared with emis-
sions exhausted through the shed capture system. A series of
four filterable particulate measurements was conducted on April
-------
- 27 -
23, 1975, to measure end leaks from the 12' x 15' rectangular
area over the bench on the north end of the shed (Figure 3.2-1).
The test used a 47-mm diameter glass-fiber filter, a probe, and
a dry-gas test-meter assembly similar to that used for filterable
particulate emissions from the shed capture system. A vane ane-
mometer measured exhaust gas velocities as the instrument was
passed slowly over representative portions of the rectangular
cross section from which particulate emissions were observed to
escape from the shed. The probe-filter assembly was swept over
this area during each of four tests.
4.4 Particle Size Distribution
Particle size Tests 1 through 9 were collected with a Brink
impactor which included the use of five separation stages plus
a cyclone pre-separator and a 47-mm type A back-up filter, fol-
lowing the procedure outlined in the instructions.'2' The entire
unit was placed in the stack and samples were drawn isokinetically
through an appropriately sized nozzle preceding the cyclone.
After sufficient pushes (one to eight pushes) were sampled to
collect a weighable portion of material on each stage, the col-
lection plates and cyclone collector were rinsed with acetone
and the sample transferred to glass sample containers. In the
laboratory, the acetone from the sample was evaporated and the
samples weighed on a laboratory balance capable of resolving 0.1
milligram.
Particle size sampling with an Andersen impactor was con-
ducted similarly, following the procedure outlined in the instruc-
tions. ^ ^ No filter paper was used in the collection plates
and the cyclone pre-separator was not used during this evaluation.
-------
. 28 _
4.5 Emissions of Other Materials
4.5.1 Sulfur Dioxide and Sulfur Trioxide
Sulfur dioxide and sulfur trioxide samples were col-
lected by the Shell method. Filtered, sampled gas was passed
through isopropyl alcohol to collect the sulfur trioxide and
then through 3-percent hydrogen peroxide to collect sulfur
dioxide in Greenburg-Smith impingers. Each sample was col-
lected from the shed exhaust gas during at least one coke
oven push. A description of the sampling and analytical
procedure is included in Appendix Q (Volume, 3).
4.5.2 Gaseous Contaminants by Charcoal Tube Collection <
Emissions of benzene, the homologues of benzene, and
pyridine were measured by adsorption of these gases from
the stack gas on activated charcoal. Later the charcoal was
desorbed with an appropriate eluant which was then analyzed
by gas chromatographic techniques.
A description of the sampling method used for these,
measurements is indicated in Appendix R (Volume 3).
4.5.3 Polynuclear Aromatic Compounds
Polynuclear aromatics, including benzo(a+e)pyrene,
chrysene, fluoranthene, and pyrene were measured with a
sampling train consisting of a probe, a filter, and impin-
gers containing cyclohexane. Filterable emissions included
the probe wash and filter catch, whereas total emissions
also included the impinger catch. Analysis of each frac-
tion was performed independently.
-------
- 29 -
Sampling for these contaminants was conducted over a
minimum of one hour and included sampling during at least
one coke oven push to assure that collected emissions re-
presented both door leaks and coke oven pushes beneath the
shed. A detailed description of the sampling method is
found in Appendix S (Volume 3).
4.5.4 Gaseous Contaminants by Collection in Gas Burette
Emission concentrations of carbon monoxide, total
light hydrocarbons, methane and homologues, ethene and homo-
logues, and acetylene were measured by collection in gas
burettes. This "grab sample" was analyzed in the laboratory
by extraction of a small sample from the burette with a hy-
podermic needle and syringe followed by injection into a
gas chromatograph. A detailed description of the sampling
method is indicated in Appendix T (Volume 3).
4.5.5 Gaseous Contaminants by Collection in Aqueous Sodium
Hydroxide
Cyanide, chloride, nitrogen oxides, sulfite, sulfate,
and phenolic materials were collected in impingers containing
a 0.1 N solution of sodium hydroxide after the exhaust gas
materials had previously been passed through a filter.
Cyanide and chloride ion concentrations were measured with
ion selective electrodes. Sulfite and sulfate were meas-
ured turbidimetrically. Oxides of nitrogen were measured
by the phenoldisulfonic acid spectrophotometrie method.
Phenolic materials were measured by distillation followed
by gas chromatography. A detailed description of the sampling
and analytical procedures is indicated in Appendix U (Volume 3)
-------
- 30 -
4.6 Pus t fa11 Mea su rements
Settleable particulate was measured at various locations
beneath the shed and in geometrically similar locations near the
unshedded C Battery. Settleable particulate was measured by
placing dustfall buckets (with 6-inch diameter openings) at vari-
ous locations and transferring these samples at approximately
12-hour intervals. Approximately one inch 'of distilled water
was placed at the b'ottom of the dustfall bucket at the beginning
of the sampling period. The location of the dustfall'bucket was
indicated by the oven nearest the dustfall sampling location and
by the terms "bench," "wall," "overhead," or "coke guide car."
For example, the "No. 12 Bench" site indicates that the dustfall
bucket was nearest oven No. 12 and was located al'ong the coke-side
bench.
Locating the dustfall buckets was difficult because the buck-
ets had to be placed at a point where the coke guide car and the
quench car would not interfere with the bucket. Buckets at the
"bench" site were located approximately five feet above the ground
level and approximately one foot away from the bench wall (Figure
3.2-1). At this location, coke passing through the coke guide
passed directly over the dustfall-buckets en route to the quench
car when nearby ovens were being pushed;
Dusti'all buckets at the 'wall'1 site were located inside the
shed wall approximately one foot above the bottom of the wall
and approximately one foot inward from the wall. At this loca-
tion, the quench car passed not beneath but approximately one
r
foot to the side of the bucket en route to the quench tower.
-------
- 31 -
Dustfall buckets were placed on the No. 1 car (operating beneath
the shed) and the No. 2 car (operating at the north end of the
battery outside the shed) approximately 15 feet north of the coke
guide at an elevation approximately three feet higher than the
bottom of the coke guide. Both buckets were located north of
the coke guide. Dustfall buckets at the "overhead" location were
suspended from the supporting steel work at the upper portion of
the shed. The buckets were located immediately above the quench
car at an elevation slightly above the top of the oven.
In the laboratory, the material captured in each bucket was
passed through a No. 18 sieve (1-mm square holes) and the weight
captured on the sieve was determined first by drying the collected
material and then weighing the material on an analytical balance
capable of resolving 0.1 milligram. The material passing through
the sieve was further filtered to separate the water-soluble from
the water-insoluble dustfall portions. Materials captured on the
filter were dried and weighed on an analytical balance and the
water-soluble materials passing through the filter were placed
in a beaker in an oven operated at 105°C where the water was
evaporated from the sample. The dry residue was then weighed on
an analytical balance. Dustfall materials were then divided into
three categories:
1. That composed of particles which were collected on the
No. 18 yieve:
2. Materials passing through the No. 18 sieve which were
not water soluble; and
3. Materials passing through the No. 18 sieve which were
water soluble.
-------
- 32 -
Settleable particulate was calculated from the second and third
categories above.
The materials captured in category No. 2 (water-insoluble
smaller particles) were further characterized by acetone solu-
bility, cyclohexane solubility, and pH for six of the samples.
The samples were divided into three weighed portions. Acetone
was added to .the first portion and the resulting slurry was
passed through a filter after which the acetone solution was
evaporated to produce a residue of constant weight. This indi-
cated the percent of acetone solubles. Similarly, the second
weighed portion was treated with cyclohexane to indicate the
percent of cyclohexane solubles.. Water was added to the third
portion and the pH of the resulting slurry was measured with a
pH meter.
4.7 X-ray Fluorescence and Microscopic Analysis
Samples of filterable particulate were captured over brief
sampling periods during coke-oven pushing on a cellulose acetate
filter for subsequent'X-ray analysis. The description of the
procedure and the computer results of the evaluation are indi-
cated in Appendix G (Volume 2).
The same filter samples were also examined using light micro-
scopy and scanning electron microscopy •techniques to determine
particle morphology, size, and physical characteristics. The
analysis technique and results are presented in Appendix H (Volume 2)
4.8 Visible Emissions Monitoring
4.8.1 Degree-of-Greenness Ratings
During the testing program each individual coke-oven
push was observed visually and rated according to the opacity
of the plume immediately above the quench car. Observations
-------
- 33 -
were made and recorded by EPA-certifled visible emissions
observers in all cases (Appendix V, Volume 3). The results
of this subjective, opacity-type rating technique were
labeled "degree-of-greenness." A high rating indicates an
opaque plume resulting from the pushing of "green" (insuf-
ficiently carburized) coke. Each push was divided into three
approximately equal parts and each third of the push was
classified according to greenness by giving it a separate
rating number. Faint or light plumes were given a "1"
rating, and opaque plumes usually accompanied by flames in
the plume were classified as "4." Ratings of "2" or "3" were
subjective interpolations between the number "1" and number
"4" conditions.
A plume whose three-part rating was, for example,
"1-2-4" indicated that the first third of the push was fairly
clean, the middle segment of the push resulted in a moderately
clean plume, and the last third of the push was extremely
dirty. The sum of the three digits (7 in this example) is
an indication of the overall greenness as a function of the
plume appearance. The duration, in seconds, of each push
varied somewhat; therefore, the time-weighted product of the
duration (D) and the sum of degree-of-greenness ratings (S)
yielded a parameter which characterized each push in terms
of a plume appearance above the quench car. The degree-of-
greenness rating accounts for the emissions generated during
the falling of coke into the quench car as well as those
arising from the coke in the quench car. Emissions data
presented in Section 5.0 are accompanied by these degree-of-
-------
- 34 -
greenness records for the pushes which occurred during
emission measurements.
4.8.2 Stack Opacity Rating
During the source testing program, the opacity of
the plume emitted from the shed capture system, which entered
the atmosphere above the quench tower, was observed and re-
corded by EPA Method 9 (40CFR60) at 15-second intervals
(Appendix W, Volume 3). (Minor portions of the plume were
sometimes observed to exit to the atmosphere through the
quench car door of the quench tower.) Between the pushes
occurring under the shed, this source produced a plume of
zero- or five-percent opacity. As a direct result of pushing
under the shed, however, the opacity above the quench tower
would increase to 25 to 30 percent. Immediately following
the elevated plume opacity readings, the steam plume from
the quenching operation masked the plume from the shed
capture system; thus, the duration of elevated stack opacity
could not be determined by visual methods. Observations of
quench tower plume opacity, including average and maximum
percent opacity of the quench tower emissions during each
push for the particulate emission and particle sizing tests,
are presented in Section 5.0.
4.8.3 Transmif-someter Data
During the test program, a transmissometer was
installed in the shed exhaust duct at the rectangular section
immediately downstream of the shed and upstream of the exhaust
-------
- 35 -
fan. The transmissometer continuously monitored the opacity
levels of the exhaust air discharged from the shed capture
system by transmitting a beam of light across the duct and
measuring the amount of attenuation. A description of the
transmissometer system, and its operation, including an
analysis of the opacity (optical density) measurement data
obtained during the test period is set forth in the report
prepared by EPA shown in Appendix I (Volume 2).
The transmissometer strip chart records show that dur-
ing the period when no pushing was occurring under the shed,
the optical density of the stack exhaust gas was only very
slightly above the background opacity line of the strip
chart recording due to door leaks under the shed. The in-
strument was zeroed during this time when the shed appeared
to be relatively "clean"; therefore, an absolute zero opac-
ity base line was not established. During a push, the opac-
ity density of the stack would increase, reach a maximum,
and then decrease gradually until the shed was evacuated of
the plume produced by that oven-pushing operation. Normally,
the optical density would return to near the zero base line
within two minutes after the push had begun, thus providing
a measure of the pushing emissions clearing time.
The optical density of the exhaust gas sometimes
increased beyond the zero base line st times other than
during pushing. A noticeable increase was evident when
excessive door leaking occurred or when the quench car,
returning from the quench tower, passed beneath the shed,
-------
- 36 -
resulting in a steam plume which was detected by the
transmissometer.
Two characteristic parameters were determined for each
coke-oven push from the transmissometer data: maximum or
peak optical density during the push and total optical den-
sity. The second parameter is a relative measure of the
total area beneath the optical-density-versus-time curve
produced by the strip chart recording.
For the purposes of comparing the opacity^levels meas-
ured by transmissometer with opacity readings made by trained
observers at the GLC plant and other coke plants and for
developing correlations with mass emissions measurements
and process variables, the maximum optical density and
optical-density-time values were converted to equivalent
values of attenuation coefficients by the formula:
r- _ optical density _ ln(l/T)
^- path length path length
where: £. = attenuation coefficient; and
, /opacity \
T = transmittance = 1 - I-— 1QQ I
Correlations between the particulate emission factors
and various indices of visible emissions, including maximum
and total optical density as measured and calculated from
the transmissometer strip chart recordings for each push
occurring during the parciculate and particle sizing tests,
are presented in Section 5.0. Reproductions of the strip
charts themselves are contained in Appendix X (Volume 3).
-------
- 37 -
4.8.4 Door Leak Inspection Data
During the sampling study, door leaks were observed
and recorded as they occurred around the oven doors on
the push and coke sides of the battery. If it was visually
apparent that a door beneath the coke shed was leaking, the
oven number of that door was noted at the time of the door
leakage survey. Sometimes an oven could not be observed
because it was obscured by the coke guide car; this was so
noted on the field data sheets that are presented in
Appendix Y (Volume 3).
4.9 Calibration of Sampling Equipment and Example Calculations
Before and after the field study was conducted, several key
pieces of the sampling equipment were calibrated, including
Pitot tubes, dry-gas meters, orifice meters, sampling nozzles,
and thermocouple potentiometers. Where correction factors are
applicable, the average of pre- and post-study calibration
correction factors was applied.
The S-type Pitot tube used to measure stack gas velocities
was calibrated over a range of velocity pressures and compared
with velocity pressures measured with a standard-type Fitot tube.
Appendix J (Volume 3) contains a description of the procedure
used for Pitot tube calibration; Appendix K (Volume 3) contains
the Pitot tube calibration data used for this study.
The dry-gas tas«- meters and orifice Jieters used to measure
sample volume were calibrated against a wet-test meter accord-
ing to the procedure found in Appendix L (Volume 3). Pre- and
post-study calibration data are presented in Appendix M (Volume 3)
-------
- 38 -
Thermocouple potentiometers were calibrated according to
the procedure outlined in Appendix H (Volume 3), and accuracy
to within five degrees 'Fahrenheit was assured over a wide range
of stack gas temperatures. ' • >
Sampling nozzle diameters were measured with a micrometer
before and after the study, This calibration procedure is
described in-Appendix 0 (Volume 3). • • '
Appendix Z (Volume 3) contains sample calculations for
particulate emissions,< gaseous emissions, particle size distri-
bution, and dustfall.
4.10 Quality Assurance and 'Chain of Custody
To insure the integrity of all samples, the chain of custody
procedure (Appendix AA, Volume 3) was followed conscientiously.
At all time's, either one member* of the Clayton test team was
with the samples or the samples -were locked securely in storage.
-------
- 39 -
5.0 PRESENTATION AND DISCUSSION OF RESULTS
5.1 Comparison of Pushing-Cycle and Non-Pushing-Cycle Particu-
late Tests
In this test program, particulate samples were collected
during each of two cycles of the coke-pushing operation at the
Great Lakes Carbon plant. Samples collected during the "pushing
cycle" were collected continuously during the time that the pro-
duction schedule called for the pushing of ovens beneath the
shed. When the schedule called for the pushing of ovens at the
C Battery (those ovens not beneath the shed), no pushing was
occurring beneath the shed. Therefore, particulate emissions
captured during this time were labeled "non-pushing-cycle"
particulate tests. Sampling during each of these two different
types of operational cycles was an attempt to quantify the rela-
tive contribution of door leaks and oven pushes to the particu-
late emissions.
Table 5.1-1 Summarizes the particulate emissions occurring dur-
ing the pushing cycle (oven pushes plus door leaks - Appendix B,
Volume 2) and the non-pushing cycle (door leaks only - Appendix C,
Volume 2). The difference in the particulate emissions during the
two cycles is an indication of the relative contribution of coke-
oven pushing to the total particulate emissions from the coke sife..
This calculation inherently assumes that th.2 average door leak rate
during non-pushing-cycle tests is the same as that during pushing-
cycle particulate tests. From Table 5.1-1 it is evident that the
pushing of coke ovens accounts for an average of 56 percent of
the filterable particulate emissions captured by the shed during
-------
TABLE 5.1-1
SUMMARY OF PARTICULATE EMISSIONS
Coke Shed
Great Lakes Carbon Corporation
St. Louis, Missouri
April 21-24, 1975
Test
Condition
Pushing Cycle
Non-Pushing
Cycle
Push-Only
(Pushing Cycle'-
Non- Pushing
Cycle )
Tc st No .
1
2
3
4
Avg(l-3)
Avg(l-4)
1
2
3
Avg
1
2
3
Avg
Stack Gas
Conditions
Temp
(°F)
75
85
74
88
78
80
69
85
70
75
;;-
Flowrat e
(DSCFM)
129,000
119,000
123,000
121,000
124,000
123,000
128,000
125,000
132,000
128,000
::::
Particulate Concentration
(gr/DSCF)
Filterable
0.019
0.013
0.015
0.028
0.016
0.019
0.006
0.009
0.003
0.006
0.013
0.004
0.012
0.010
Back half
0.001
0.002
0.003
0.003
0.002
0.002
0.0003
0.001
0.001
0.0008
0.0007
0.001
0.002
0.001
Total
0.020
0.016
0.018
0.031
0.018
0.021
0.007
0.010
0.005
0.007
0.013
0.006
0.013
0.011
Particulate Emission Rate
(Ibs/hr)
Filterable
20.6
13. 7
15.7
29.0
16.7
19.8
6.9
10.0
3.9
6.9
13.7
3.7
11.8
9.7
Back half
1.1
2.5
3.4
3.2
2.3
2.6
0.37
1. 1
1.4
0.96
0.7
1.4
.. 2.0
«w
1-4
Total
21.7
16.2
19.1
32.2
19.0
22.3
7.2
11.1
5.3
7.9
14.5
5. 1
13.8
11. 1
*•
0
i
-------
- 41 -
the pushing cycle. This conclusion is dependent upon the charac-
teristics of the pushes occurring beneath the shed during the
pushing-cycle particulate tests. Table 5.1-2 displays the data
necessary to characterize these pushes.
To determine the relative contribution of oven pushing to
the filterable particulate emissions during the entire cycle, the
relative duration of each of the two cycles in the overall pro-
duction schedule must be established. Because 40 ovens are
beneath the shed and 35 are outside of the shed, the pushing
operation is in the pushing-cycle mode approximately 12.8 hours
per day (40/75 times 24 hours per day). Similarly, the non-
pushing cycle is in operation 11.2 hours per day (35/75 * 24
hours per day). The contribution of oven pushing to the overall
filterable particulate emissions from the overall operation is
43 percent, as shown in the following time-weighted average cal-
culation:
9.7 Ibs/hr * 12.8 hrs/day ^___ _ 42.7%
16.7 Ibs/hr * 12.8 hrs/day + 6.9 Ibs/hr * 11.2 hrs/day
This indicates that the continuous leaking of smaller
amounts of particulate matter from coke-oven doors accounts for
a greater portion of the filterable particulate emitted by the
shed capture system (57 percent) than the infrequent but more
concentrated emissions resultant from the pushing of coke ovens
at GLC's A Battery. A similar time-weighted calculation, using
the back-half emissions listed in Table 5.1-1, indicates that 60
percent of the back-half emissions at GLC's A Battery may be
attributed to door leaks.
-------
TABLE 5.1-2
PUSH CHARACTERISTICS
Great Lakes Carbon Corporation
St. Louis, Missouri
April 21-24, 1975
Test Pushing-Cycle Particulate 1
Date
4/21/75 & 4/22/75
Time
10:40
10:52
12:30
12:46
13:04
16:30
16:40
16:47
17:00
17:10
17:30
09:50
10:04
10:46
11:25
Oven
Pushed
7
17
27
37
47
2
12
22
32
42
52
5
15
45
55
AVERAGE
Net
Coking
Time
26:10
26:02
26:56
27:15
27:19
28:07
28:04
27:56
27:48
27:35
27:44
27:50
27:46
27:35
26:45
27:23
Degree of Greenness
Rating
232
332
322
212
311
221
444
331
442
432
111
131
221
212
221
--
Sum
(S)
7
8
7
5
. 5
5
12
7
10
9
3
5
5
5
5
7
Duration
(D)
28
29
34
38
34
29
26
28
29
27
28
32
29
38
38
31
S*D
196
232
238
190
170
145
312
196
290
243
84
160
145
190
190
199
Quench Tower
Opacity
Average
Percent
18.0
16.0
9.0
4.0
7.5
8.0
42.5
9.0
16.7
_ _
17.0
7.5
14.6
— —
14.2
Maximum
Percent
30
25
15
5"
15
15
80
. 10 .
30
30
15
25
—
25
Plume
Attenuation
Coefficient
(sec.rmeters~l)
29.92
32.52
5.88
5.0,0
9.51
3.25
54.64 '
9.11
32.52
4.55
4.55 ' -
6.50
2.60
8.46
5.85
14.32
r
Maximum
Attenuation
Coefficient
(me ters~ I )
0.885
0.976
0.156
0.137
; 0.286
0.091
1.626
0.260
0.976
0.117
0.117
0.195
0. 104
0.247
0.163
0.422
ISJ
I
-------
Test
TABLE 5.1-2 (continued)
PUSH CHARACTERISTICS
Great Lakes Carbon Corporation
St. Louis, Missouri
April 21-24, 1975
Pushing-Cycle Particulate 2
Date
4/22/75
Time
14:08
14:21
14:32
15:20
15:35
16:15
16:25
18: 15
18:22
18:32
14:45
Oven
Pushed
7
17
27
37
47
14
34
2
12
22
32
AVERAGE
Net
Coking
Time
27:12
26:01
25:40
26:05
26:06
48:30
48:20
25:32
25:2?
25:25
25:20
29;58
Degree of Greenness
Rating
232
432
221
111
211
221
211
221
444
211
432
--
Sum
(S)
7
9
5
3
4
5
4
5
12
4
9
6
Duration
(D)
32
35
33
31
26
26
35
32
34
31
36
32
S*D
224
315
165
93
104
130
140
160
408
124
324
199
Quench Tower
Opacity
Average
Percent
15.0
26.7
8.0
5.6
8.3
5.6
8.6
5.0
52.5
6.3
40.0
16.5
Maximum
Percent
30
50
10
10
15
15
20
10
80
10
60
30
Plume
At tenuation
Coefficient
(sec . -meters- 1 )
17.85
25.50
7.99
4.55
9.76
3.50
4.55
3.25
32.52
6.50
29.92
13.26
Maximum
Attenuation
Coefficient
(meters-1)
0.650
0.703
0.217
0.130
0.286
0.072
0.124
0.078
0.976
0.195
0.885
0.392
-------
Test
TABLE 5.1-2 (continued)
PUSH CHARACTERISTICS
Great Lakes Carbon Corporation
St. Louis, Missouri
April 21-24, 1975
Pushing-Cycle Particulate 3
Date
4/23/75
Time
08:45
08:59
09:14
10:08
13:12
13:27
13:42
14:00
14:14
14:31
Oven
Pushed
23
33
43,
53
5
15
25
35
45
55
AVERAGE
Net
Coking
Time
26:02
26:04
26:02
26:38
27:01
27:02
26: 52
42:33
27:02
27: 17
28:15
Degree of Greenness
Rating
221
321
" 432
211
222
211
322
212
211
121
—
Sum
(S)
5
6
9
4
6
4
7
5
4
4
5
Duration
(D)
34
35
38
38
37
35
35
40
38
40
37
S*D
170
210
342
152
222
140
245
200
152
160
199
Quench Tower
Opacity
Average
Percent
5.0
6.0
15.8
11.7
11.7
5.8
14.0
4.2
5.8
2;7
8.3
Maximum
Percent
10
. 10
25
25
25
10
25
10
10
. 5
15
Plume
At t enua't ion
Coefficient
1 (sec . -meters- 1 )
5.20
4.55
19.52
6.50
18.21
5.20
22.77
3.25
7.16
4.55
9.69
Maximum
Attenuat ion
Coefficient
(meters" * )
0.137
0-.130
0.585
0'Jl74
0.546
0.150
0.664
0.098
0.208
0.130
0.282
-------
TABLE 5.1-2 (continued)
PUSH CHARACTERISTICS
Great--Lakes Carbon Corporation
St. Louis, Missouri
April 21-24, 1975
Test
Pushing-Cycle Particulate 4
Date
4/24/75
Time
10:45
11:12
11:35
12:45
13:12
C"v\en
Pu sn e d
13
23
33
43
53
AVERAGE
Net
Coking
Time
25:56
26:06
26:1C
26:31
26:43
26:17
Degree of Greenness
Rating
432
312
212
421
344
--
Sum
(S)
9
6
5
7
11
8
Duration
x^D)
38
> 34
40
36
45
39
•«S*D
342
204
200
252
495
299
Quench Tower
Opacity
Average
Percent
--
- —
--
-Maximum
Percent
--
- -
--
Plume
Attenuation
Coefficient
(sec. -meters" 1 )
24.04
6.30
8.25
34.02
95.36
33.59
Maximum
At tenuat ion
Coefficient
(metiers" * )
0.533
0.141
0.195
1.067
2.602
0.908
in
I
-------
- 46 -
5.2 Calculation of Emission Factors
5.2.1 Emission Factor for Coke-Oven Pushing
Because oven pushing accounted for a majority of the
particulate captured during pushing-cycle particulate tests,
the process weight rates, used in the calculation of emission
factors during these tests (Appendix B, Volume 2), were based
on the weight of dry coal fed to those ovens pushed and the
weight of coke produced during the given test. For example,
Table 5.1-2 indicates that 15 ovens were pushed during
Pushing-Cycle Particulate Test No. 1. Assuming that each
oven was charged with 13.7 tons of dry coal and that 10.5 tons
of coke were produced during each push, the process weight
represented in this test was 205.5 cons of dry coal or 157.5
tons of coke. Appendix F (Volume 2) indicates that the net
test duration was 288 minutes. Therefore, the feed rate was
calculated to be 42.8 tons of dry coal per hour, or 32.8
tons of coke per hour.
Using these feed rates and the emission rates from
Table 5.1-1, the emission factors for the pushing-cycle and
push-only emissions are calculated in Table 5.2.1. By includ-
ing the contribution of fugitive emissions (see Section 5.3
for documentation of fugitive emissions), the push-only emis-
sions measured by the shed capture system are increased by
10 percent and included Ln Table 5.2.1.
5.2.2 Emission Factor for Door Leaks
Process weight rates for the non-pushing-cycle particu-
late tests, which measure door leak emissions, could not be
-------
TABLE 5.2.1
SUMMARY OF PARTICULATE EMISSION FACTORS
Great Lakes Carbon Corporation
St. Louis, Missouri
April 21-24, 1975
Test
Condition
Pushing
Cycle
Non-
Pushing
Cycle
Push-
Only*
Push-Only
Including
Fugitives
Test
No.
1
2
3
4
Average (1-3)
Average (1-4)
1
2
3
Average
1
2
3
Average
1
2
3
Average
Particulate
Emission Rate
(Ibs/hr)
Filter-
able
20.6
13.7
15.7
29.0
16.7
19.8
6.9
10.0
3.9
6.9
13.7
3.7
11.8
9.7
15.1
4.1
13.0
10.7
Total
21.7
16.2
19.1
32.2
19.0
22.3
7.2
11.1
5.3
7.9
14.5
5.1
13.8
11. 1
16.0
5.6
15.2
12.3
Process
Weight
tons dry
coal/hr
42.8
47.1
41.7
32.9
43.9
41.1
19.1
19.3
20.0
19.5
42.8
47.1
41.7
43.9
42.8
47.1
41.7
43.9
Rate
tons
coke/hr
32.8
36.1
32.0
25.2
33.6
31.5
14.6
14.8
15.4
14.9
32.8
36.1
32.0
33.6
32.8
36.1
32.0
33.6
Particulate Emission Factor
Filterable
Ibs/ton
dry coal
0.48
0.29
0.38
0.88
0.38
0.51
0.36
0.52
0.20
0.36
0.32
0.079
0.28
0.23
0.35
0.087
0.31
0.25
Ibs/ton
coke
0.63
0.38
0.49
1.2
0.50
0.68
0.47
0.68
0.25
0.47
0.42
0. 10
0.37
0.30
0.46
0.11
0.41
0.33
Total
Ibs/ton
dry coal
0.51
0.34
0.46
0.98
0.44
0.57
0.38
0.58
0.26
0.41
0.34
0. 11
0.33
0.26
0.37
0.12
0.36
0.28
Ibs/ ton
coke
0.66
0.45
0.60
1.3
0.57
0.75
0.49
0.75
0.34
0.53
0.44
0.14
0.43
0.34
0.49
0.16
0.48
0.38
* Emission factors for push-only emissions (i.e.
tracting emission rates and subsequently divid
two different process weights for pushing and
subtracted directly.
, no door leaks included) are
ing by the "process weight."
non-pushing cycles, emission
computed by sub-
Due to the use of
factors cannot be
-------
- 48 -
calculated from pushing data because these test periods
inherently excluded coke pushing. Process weights were
established by dividing the. total weight of dry coal fed to
all of the ovens beneath the shed by the average coking time
for those ovens containing the coal charge associated with
the emissions occurring during that non-pushing-cycle
particulate test (Appendix C, Volume 2).
Door leak emission factors, shown in.Table 5.2.1,
averaged 0.36 pound of filterable particulate per ton of dry
coal fed to all ovens producing door leak emissions, or 0.47
pound per ton of coke produced.
5.2.3 Overall Emission Factor
The average overall emission factor for filterable
particulate emissions from the coke side of the A Battery
is the sum of the emission factor for particulate originating
from door leaks and that from coke-oven pushing. Therefore,
0.61 (0.25 + 0.36) pound of filterable particulate per ton
of dry coal fed or 0.80 (0.33 + 0.47) pound per ton of coke
produced, was emitted from the coke side of the battery.
Although the process weights used in computing the two compo-
nents of the summed emission factor are different (i.e., coal
fed to pushed ovens for pushing emissions and coal fed,to all
ovens for door leaks) , the sum is a meaningful indicator of
coke-side overall emissions because all leaking ovens are
pushed eventually. Thus, the emission factor depends upon
the characteristics of the pushes occurring during the
testing as well as the degree of door maintenance practiced
at the time of field measurements.
-------
- 49 -
5.3 Significance of Fugitive Leaks
Table 5.3-1 summarizes the measurement of particulate emis-
sions escaping from the north end of the shed. During these four
tests on April 23, 1975, fugitive particulate emissions from this
source ranged from 0.0081 to 0.090 pound per ton of dry coal fed to
the coke ovens, or 0.011 to 0.12 pound per ton of coke produced.
When these emission factors are compared to the emission factor for
filterable particulate measured during Pushing-Cycle Particulate
Test 3 on April 23, 1975 (0.38 pound of particulate per ton of dry
coal fed), the end leakage ranges from 2 to 19, and averages 9 per-
cent of the emissions from coke-oven pushing. Thus, the average
capture efficiency of the shed during pushing was 91 percent, as
shown in Table 5.3-2. Then, the total emissions from pushing only were
about (TJ— ST) or 110 percent of the emissions captured by the shed
and measured in the shed exhaust.
The sum of the degree-of-greenness ratings for the five pushes
represented in the four fugitive emission estimation tests averaged
5.8. The 10 pushes constituting Pushing-Cycle Particulate Test 3
had an average degree-of-greenness sum of 5.4 (Table 5.1-2). These
results indicate that the five pushes represented in the fugitive
emission estimation were of the same approximate degree-of-greenness
rating as those measured in the particulate test.
Subjectively, there was no visible evidence that door leaks
contributed to fugitive emissions; therefore, the total non-pushing-
cycle emissions were emitted through the shed capture system. Con-
sidering both the pushing and non-pushing cycles, the overall average
percent capture efficiency of the shed thus appears to be about
96 percent, as shown in Table 5.3-2.
-------
TABLE 5.3-1
SUMMARY OF FUGITIVE EMISSION ESTIMATION
NORTH END OF SHED
Great Lakes Carbon Corporation
St. Louis, Missouri
April 23, 1975
Fugitive
Particulate
Test Number
I
2
3
4
Time
14:14 - 14:56
17:10 - 17:14
17:18 - 17:23
17:26 - 17:32
Oven (s)
Pushed
45,55
27
37
47
AVERAGE
Flowrate
(SCFM)
39,960
39,960
39,960
39,960
(39,960)
Particulate
Concentration
(gr/SCF)
0.0103
0.0048
0.0143
0.0091
0.0096
Particulate
Emission
Rate
(Ibs/hr)
3.53
1.66
4.91
3.10
3.30
Particulate
Emission Factor
Ib/ton
dry coal
0.090
0.0081
0.030
0.023
0.038
Ib/ton
coke
0.12
0.011
0.039
0.030
0.050
Ui
o
-------
- 51 -
TABLE 5.3-2
SHED PARTICULATE CAPTURE EFFICIENCIES
Great Lakes Carbon Corporation
St. Louis, Missouri
April 23, 1975
Fugitive
Particulate
Test
Number
1
2
3
4
Average
Particulate Capture
Efficiency of the Shed
During Coke Oven
Pushing
(Percent)
81
98
93
94
91
Overall Particulate
Capture Efficiency
of the Shed*
(Percent)
92
99
97
97
96
Overall Efficiency
Efficiency during pushing^
100
* 0.43 + 0.57
where the factors, 0.43 and 0.57, represent the fractions of
time corresponding to pushing and non-pushing operational modes,
respectively, occurring under the shed, and the capture efficiency
during non-pushing is estimated to be 100 percent.
-------
- 52 -
Table 5.3-1 indicates that greater emission factors were
calculated for emissions leaking from the north end of the shed
when ovens at the- north'end of the shedded A Battery were being
pushed. Fugitive Particulate Test 1 (which resulted in a higher
emission factor of leaks at the north end of the shed) represented
the pushes of ovens 45 and 55, located at the north end of the
shed. The smallest fugitive emission factor was estimated during
Test No. 2.when Oven No. 27 (center of shed) was pushed.
Although the end leak measurements made on April 23 only
included the pushing of five ovens, it was noted that the charac-
teristics of these pushes were similar to those observed during
the tests which measured particulate exhausted through the shed
capture system. The wind blew from the southeast quadrant during
the measurement of fugitive emissions, a direction which is expected
to result in maximum emissions from the north end. Wind speeds
were approximately 10 miles per hour (Appendix BB, Volume 3), which is
somewhat above the annual average wind speed (Appendix CC, Volume 3).
In summary, from these data we conclude that, based on visual
evaluation, the shed is less than 100 percent efficient in capturing
coke-pushing emissions. An increased leakage rate is observed
as a result of higher wind speeds, as a result of pushing occurring
near the end of the shed, especially the downwind end, and as a
result of pushes with a high degree of greenness. During conditions
which were relatively conducive to leakage, average emissions es-
caping the shed ranged from 0.0081 to 0.090 pound per ton with
an average of 0.038 pound per ton of dry coal fed, or 0.050 pound
per ton of coke produced.
-------
- 53 -
5.4 Chemical and Physical Characteristics of Particulate Emissions
Tables 5.4-1 and 5.4-2 indicate that the distribution of total
particulate catch, for both pushing-cycle and non-pushing-cycle
particulate tests, averages 87 percent as filterable particulate
and 13 percent as materials captured in the impingers following
the filter in the front half of the sampling train.
Cyanide, chloride, and sulfate accounted for minor por-
tions of filterable particulate during both pushing- and non-
pushing-cycle particulate tests. Table 5.4-3 indicates that 87
percent of the filterable particulate matter is neither soluble
in acetone nor cyclohexane, indicating that a minor portion of the
filterable particulate is organic in composition (Figure 4.2). On
the other hand, only 22 percent of the particulate captured in
the impingers is composed of materials insoluble in cyclohexane,
\
indicating that a majority of this particulate material is of
organic composition for both the pushing- and non-pushing-cycle
particulate tests. Table 5.4-4 indicates that nearly all particu-
late fractions were slightly acidic.
X-ray fluorescence analysis of filterable particulate emis-
sions produced during the pushing of a coke oven indicated that
the non-carbon portion of the collected particulate contained
the elements chlorine, sulfur, silicon, and aluminum with minor
amounts of calcium and iron (Appendix G, Volume 2).
Microscopic examination of filterable particulate emissions
produced during the pushing of coke revealed a variety of particles
which, -for the purpose of the analysis, were classified into 8 dif-
ferent categories based on particle morphology, color, birefringence,
-------
- 54 -
TABLE 5.4-1
EMISSION OF PARTICULATE CONTAMINANTS (LBS/HR)
PUSHING CYCLE
Coke Shed
Great Lakes Carbon Corporation
St. Louis, Missouri
April 23, 1975
Test
No.
1
2
3
4
Average
(1-3)
Filterable
Back-half
Total
Filterable
Back-half
Total
Filterable
Back-half
Total
Filterable
Back-half
Total
Filterable
Back-half
Total
Particulate Fraction
Particulate
20.6
1.1
21.7
13.7
2.5
16.2
15.7
3.4
19.1
29.0
3.2
32.2
16.7
2.3
19.0
Cyanide
<0.87
<1.2
<2.1
<0.72
<0.91
<1.6
<0.69
<1.3
<2.0
<1.0
<2.0
<3.0
<0.76
<1.1
<1.9
Chloride
0.04
0.45
0.49
0.10
0.62
0.72
0.12
0.62
0.74
0.21
0.74
0.94
0.09
0.56
0.65
Sulfate
0.25
1.2
1.4
0.07
0.37
0.44
0.17
0.64
0.81
0.15
0.78
0.93
0.16
0.74
0.88
Organics
1.9
0.60
2.5
0.43
1.9
2.4
3.2
3.1
6.3
5.1
2.4
7.5
1.8
1.9
3.7
Inorganics
18.7
0.51
19.2
' 13.2
0.59
13.8
12.5
0.27
12.8
23.9
0.78
24.7
14.8
0.46
15.3
-------
- 55 -
TABLE 5.4-2
EMISSION OF PARTICDLATE CONTAMINANTS (LBS/HR)
NON-PUSHING CYCLE
Coke Shed
Great Lakes Carbon Corporation
St. Louis, Missouri
April 23, 1975
Test
No.
1
2
3
Average
Filterable
Back-half
Total
Filterable
Back-half
Total
Filterable
Back-half
Total
Filterable
Back-half
Total
Particulate Fraction
Particulate
6.9
0.37
7.2
10.0
1.1
11.1
3.9
1.4
5.3
6.9
0.96
7.9
Cyanide
<2.3
<0.95
<0.86
<0.76
<0.95
<0.94
Chloride
0.11
0.33
0.44
0.19
0.24
0.43
0.07
0.36
0.43
0.12
0.31
0.43
Sulfate
<0.06
0.12
0.12
0.15
0.41
0.56
<0.03
<0.04
<0.07
105-0.08
0,18-0.19
01 23-0.25
Organics
0.80
0.37
1.2
1.5
0.50
2.0
0.61
1.1
1.7
0.97
0.66
1.6
Inorganics
6.1
<0.06
6.1
8.5
0.63
9.1
3.3
0.32
3.6
6.0
0.32-0.34
6.3
-------
TABLE 5.4-3
CHARACTERIZATION OF PARTICIPATE WEIGHT
(Referenced to Flow Diagram in Figure 4.2)
Great Lakes Carbon Corporation
St. Louis, Missouri
April 21-24, 1975
Test
Non-pushing
1
Pushing
1
Particulate fraction, weight in grams
A
0.0546
0.2602
B
0 .00113
0.02482
C
0.0052
0.0160
D*
0.0505
0.2690
'E
0.0033
0.0122
F*
0.0019
0.0038
G
0.0013
0.0104
H*
0.0492
0.2586
J*
0.0009
0.0019
K
0.0010
0.0019.
L*
0.0059
0.0349
M
0.0433
0.2237
N&
0.0068
0.0368
* By Difference
6 By Sum
-------
TABLE 5.4-3 (continued)
CHARACTERIZATION OF PARTICULATE WEIGHT
(Referenced to Flow Diagram in Figure 4.2)
Great Lakes Carbon Corporation
St. Louis, Missouri
April 21-24, 1975
Test
Son-pushing
2
Non-pushing
3
Pushing
2
Pushing
3
Pushing
4
Particulate fraction, weight in grams
P
0.0416
0.0124
0.1099
0.1361
0.1440
Q
0.05236
0.02474
0.01965
0.01311
0.03058
R
0.0139
0.0058
0.0041
0.0307
0.0305
S*
0.0801
0.0313
0.1254
0.1185
0.1441
T*
0.0069
0.0016
0.0029
0.0072
0.0073
U
0.0070
0.0042
0.0012
0.0235
0.0232
v*
0.0487
0.0077
0.0761
0.0067
0.1309
W
0.0314
0.0236
0.0493
0.1118
0.0132
X"
0.0556
0.0093
0.0790
0.0139
0.1382
* By Difference
By Sum
-------
TABLE 5,4-3 (continued)
CHARACTERIZATION OF PARTICULATE WEIGHT
(Referenced to Flow Diagram in Figure 4.2)
Great Lakes Carbon Corporation
St. Louis, Missouri
April 21-24, 1975
Test
Non-pushing
1
Non-pushing
2
Non-pushing
3
Pushing
1
Pushing
2
Pushing
3
Pushing
4
Farticulate fraction, weight in grams
AA
<0.0005
0.0059
0-.0030
0.0071
0.0056
0.0026
0.0047
BB
0.0030
0.0047
0.0102
0.0083
0.0183
0.0292
0.0144
CC
<0.0005
0.0008
<0.0005
0.0030
0.0012
<0.0005
0.0009
DD*
<0.0005
0.0051
0.0030
0..0041
0.0044
0.0026
010038
EE*
<0.0005
0.0008
<0.0005
0.0030
0.0012
<0.0005
0.0009
FF
<0.0005
<0.0005
<0.0005
<0.0005
<0.0005
-------
- 59 -
TABLE 5.4-4
SUMMARY OF WATER SOLUBLE pH
AND ACIDITY/ALKALINITY ON PARTICULATE SAMPLES
Great Lakes Carbon Corporation
St. Louis, Missouri
April 21-24, 1975
Sampling
Cond itions
Non-pushing
Cycle
Pushing
Cycle
Test
Number
1
2
3
1
2
3
4
Portion
of
Sampling
Train
Front
Back
Front
Back
Front
Back
Front
Back
Front
Back
Front
Back
Front
Back
PH
6.8
5.0
6.5
3.7
7.5
5.8
5.0
4.0
6.2
3.8
6.2
4.3
6.8
4.0
Acidity
(meq/gm)
<0.0001
<0.007
<0. 00003
0.01
<0. 00005
<0.001
0.00006
0.008
<0. 00002
0.004
<0. 00001
0.002
<0. 00001
0.002
-------
- 60 -
and surface characteristics. The bulk of the particles were
naturally coke and partially-coked coal. Significant amounts of
coal and mineral particles were also present. The size of the
particles ranged from sub-micron to about 100 microns depending
on the type of particle, with a typical size mode of about three
microns. Green pushes seemed to generate a greater amount of
submicron particles (0.5 - 1 micron) than clean pushes, although
it was difficult to draw conclusions on the difference between
normal pushes and green pushes when so few samples were available
for examination. Particle characterizations for each of the five
filter samples analyzed are summarized in the letter report pre-
pared by the IIT Research Institute, Chicago, Illinois, shown in
Appendix H (Volume 2).
5.5 Particle Size Analysis
The size distributions of particulate, as measured by the
Brink and Andersen impactor methods (Appendix D, Volume 2), are
presented graphically in Figures 5.5-1 and 5.5-2, respectively.
A statistical comparison (chi-square test for independence) of the
percentage of particulate less than one micron and the percentage
less than five microns shows no significant differences among the
14 particle size distributions. The average of the nine Brink-
method tests indicates 10 percent of the particulate to be less
than one micron, whereas the average of the five Andersen-method
samples shows 13 percent of the particulate to be submicron. Thus,
overall, an average of 12 percent of the particulate was submicron.
Table 5.5-1 displays the characteristics of the pushes occur-
ring beneath the shed during the particle size tests.
-------
10 .0 iii
Effective
Particle
Diameter
(microns)
FIGURE 5.5-1
PARTICLE SIZE DISTRIBUTION
(Brink Tests 1-9)
Great Lakes Carbon Corporation
St. Louis, Missouri
April 21-24, 1975
i
cr-
0.01 O.OS 0.1 0.2 0.5 1 Z 5 10 20 3d 40 50 60 70 80 90 95
Cumulative Percentage Less Than Indicated Diameter
98 99
-------
10.0
Effective
Particle
Diameter
(microns)
FIGURE 5.5-2
PARTICLE SIZE DISTRIBUTION
(Andersen Tests 10-14)
Great Lakes Carbon Corporation
St. Louis, Missouri
April 21-24, 1975
0.01
0.05 0.1 0.2 0.5 1 2 6 10 20 30 40 SO 60 70 10 90
Cumulative Percentage Less Than Indicated Diameter
98 99
-------
TABLE 5.5-1
PUSH CHARACTERISTICS
Great Lakes Carbon Corporation
St. Louis, Missouri
April 21-24, 1975
Test Pushing-Cycle Particle Size 1
Date 4/22/75
Time
09:50
10:04
10:21
Oven
Pushed
5
15
25
AVERAGE
Net
Coking
Time
27:50
27:46
27:45
27:47
Degree of Greenness
Rating
131
221
332
—
Sum
(S)
5
5
8
6
Duration
(D)
32
29
33
31
S*D
160
145
264
190
Quench Tower
Opacity
Average
Percent
17.0
7.5
33.0
-
19.2
Maximum
Percent
30
15
60
35
Plume
Attenuation
Coefficient
(sec. -meters- 1)
6.50
2.60
29.92
13.01
Maximum
Attenuation
Coefficient
(me ters~ *•)
0.195
0.104
0.885
0.395
-------
TABLE 5.5-1 (Continued)
PUSH CHARACTERISTICS
Great Lakes Carbon Corporation
St. Louis, Missouri -
April 21-24, 1975
Test Pushing-Cycle Particle Size 2 & 3
Date
4/22/75
Time
14:08
14:21
14:32
15:20
15:35
16:15
16:25
Oven
Pushed
7
17
27
37
47
14
34
AVERAGE
Net
Coking
Time
27:12
26:01
25:40
26:05
26:06
48:30
48:20
32:33
Degree of Greenness
Rating
232
432
221
111
211
221
211
--
Sum
(S)
7
9
5
3
4
5
4
5
Duration
(D)
32
35
33
31
26
26
35
-
31
S*D
224
315
165
93
104
130
140
167
Quench Tower
Opacity
Average
Percent
15.0
26.7
8.0
5.6
8.3
5.6
8.6
11.1
Maximum
Percent
30
50
10
10
15
15
20
20
Plume
At tenuat ion
Coefficient
(sec . -meters- 1 ) '
17 .85
25.50
7.99
4.55
9.76
3.50
4.55
-
10.53
Maximum
Attenuation
Coefficient
(meters" 1 )
0.650
0.703
0.217
0.130
0.286
0.072
0.124
0.312
-------
Test
TABLE 5.5-1 (Continued)
PUSH CHARACTERISTICS
Great Lakes Carbon Corporation
St. Louis, Missouri
April 21-24, 1975
Pushing-Cycle Particle Size 4 & 5
Date
4/22/75
Time
18:15
18:22
18:32
18:45
Oven
Pushed
2
12
22 ,
32
AVERAGE
Net
Coking
Time
25:32
25:27
25: 25
25:20
25:26
Degree of Greenness
Rating
221
444
211
432
--
Sum
(S)
5
12
4
9
8
Duration
(D)
32
34
31
36
33
S*D
160
408
124
324
254
Quench Tower
Opacity
Average
Percent
5.0
52.5
6.3
40.0
26.0
Maximum
Percent
10
80
10
60
40
Plume
Attenuation
Coefficient
(sec . -meters- 1)
3.25
32.52
6.50
29.92
18.05
Maximum
Attenuat ion
Coefficient
(meters"1)
0.078
0.976
0.195
0.885
0.534
-------
Test
TABLE 5.5-1 (Continued)
PUSH CHARACTERISTICS
Great Lakes Carbon Corporation
St. Louis, Missouri
April 21-24, 1975
Pushing-Cycle Particle Size 6 & 7
Date
4/23/75
Time
09:14
10:08
13:12
13:27
13:42
14:00
14:14
14:31
Oven
Pushed
43
53
5
15
25
35
45
55
AVERAGE
Net
Coking
Time
26:02
26:38
27:01
27:02
26:52
42:33
27:02
27:17
28:48
Degree of Greenness
Rating
432
211
222
211
322
212
211
121
—
Sum
(S)
9
4
6
4
7
5
4
4
5
Duration
(D)
38
38
37
35
35
40
38
40
»
38
S*D
342
152
222
140
245
200
152
160
202
Quench Tower
Opacity
Average
Percent
15.8
11.7
11.7
5.8
14.0
4.2
5.8
2.7
9.0
Maximum
Percent
25
25
25
10
25
10
10
5
15
Plume
At te nu at ion
Coefficient
(sec . -meters-1)
19.52
6.50
18.21
5.20
22.77
3.25
7.16
4.55
•
10.90
Maximum
Attenuation
Coefficient
(meters'1)
0.585
0.174
0.546
0.150
0.664
0.098
0.208
0.130
0.319
-------
TABLE 5.5-1 (Continued)
PUSH CHARACTERISTICS
Great Lakes Carbon Corporation
St. Louis, Missouri
April 21-24, 1975
Test
Pushing-Cycle Particle Size 8 & 9
Date
4/23/75
Time
16:50
16:55
17:10
17:20
17:27
Oven
Pushed
7
17
27 .
37
47
AVERAGE
Net
Coking
Time
26:25
26:19
25:40
•25:25
25:22
25:50
Degree of Greenness
Rating
214
342
342
221
421
--
Sum
(S)
7
9
9
5
7
7
Duration
(D)
31
30
30
30
29
30
S*D
217
270
270
150
230
227
Quench Tower
Opacity
Average
Percent
20.8
8.3
15.8
9.0
6.7
12.1
Maximum
Percent
45
20
35
20
15
25
Plume
Attenuation
Coefficient
(sec. -meters- * )
33.76
14.27
75.07
15.35
14.49
30.59
Maximum
At tenuat ion
Coefficient
(me t er s~ 1 )
1.236
0.390
1.952
0.585
0.546
0.942
-------
TABLE 5.5-1 (Continued)
PUSH CHARACTERISTICS
Great Lakes Carbon Corporation
St. Louis, Missouri
April 21-24, 1975
Test Pushing-Cycle Particle Size 10
Date
4/23/75
Time
13:12
13:27
13:42
14:00
14:14
Oven
Pushed
5
15
25
35
45
AVERAGE
Net
Coking
Time
27:01
27:02
26:52
42:33
27:02
30:18
Degree of Greenness
Rating
222
211
322
212
211
--
Sum
(S)
6
4
7
5
4
5
Duration
(D)
37
35
35
40
38
37
S*D
222
140
245
200
152
192
Quench Tower
Opacity
Ave rage
Percent
11.7
5.8
14.0
4.2
5.8
8.3
Maximum
Percent
25
10
25
10
10
15
Plume
At tenuation
Coefficient
(sec . -meters-1 )
18.21
5.20
22.77
3.25
7.16
11.32
Maximum
Attenuation
Coefficient
(meters" *• )
0.546
0.150
0.664
0.098
0.208
0.333
00
-------
TABLE 5.5-1 (Continued)
PUSH CHARACTERISTICS
Test
Great
Pushing-Cycle Particle Size 11
Lakes Carbon Corporation
St. Louis, Missouri
April 21-24, 1975
Date
4/23/75
Time
17:10
17:20
17:27
Oven
Pushed
27
37
47
AVERAGE
Net
Coking
Time
25:40
25:25
25: 22
25:29
Degree of Greenness
Rating
342
221
421
--
Sum
(S)
9
5
7
7
Dura tion
(D)
30
30
29
30
S*D
270
150
203
208
Quench Tower
Opacity
Average
Percent
15.8
9.0
6.7
10.5
Maximum
Percent
35
20
15
25
Plume
At tenuat ion
Coefficient
(sec. -mete r s~ 1 )
75.07
15.35
14.49
34.97
Maximum
Attenuation
Coefficient
(meters ~ * )
1.952
0.585
0.546
1.028
VD
I
-------
TABLE'S.5-1 (Continued)
PUSH CHARACTERISTICS
Great Lakes Carbon Corporation
St. Louis, Missouri
April 21-24, 1975
Test
Pushing-Cycle Particle Size 12
Date
4/24/75
Time
08: 25
Oven
Pushed
51
AVERAGE
Net
Coking
Time
26:50
26 :50
Degree of Greenness
Rating
— —
--
Sum
(S)
--
Duration
(D)
--
S*D
--
--
Quench Tower
Opacity
Average
Percent
--
--
Maximum
Percent
—
--
Plume
Attenuation
Coefficient
(sec. -meters' 1 )
--
--
Maximum
At tenuat ion
Coefficient
(meters' * )
--
--
o
I
-------
TABLE 5.5-1 (Continued)
PUSH CHARACTERISTICS
Great Lakes Carbon Corporation
St. Louis, Missouri
April 21-24, 1975
Test Pushing-Cycle Particle Size 13
Date
4/24/75
' Time
10:45
11:12
Oven
Pushed
13
23
AVERAGE
Net
Coking
Time
25:56
26:06
26:01
Degree of Greenness
Rating
432
312
--
Sum
(S)
9
6
8
Duration
(D)
38
34
36
S*D
342
204
273
Quench Tower
Opacity
Average
Percent
--
--
Maximum
Percent
--
--
Plume
At t enuation
Coefficient
(sec . -meters" 1 )
24.04
6.30
15.17
Maximum
Attenuation
Coefficient
(meters" 1 )
0.533
0.141
0.337
-------
TABLE 5.5-1 (Continued)
PUSH CHARACTERISTICS
Great Lakes Carbon Corporation
St. Louis, Missouri
April 21-24, 1975
Test
Pushing-Cycle Particle Size 14
Date
4/24/75
Time
12:45
13:12
Oven
Pushed
43
53
•
AVERAGE
Net
Coking
Time
26:31
26:43
26:37
Degree of Greenness
Rating
421
344
--
Sum
(S)
7
11
9
Duration
(D)
36
45
40
S*D
252
495
374
Quench Tower
Opacity
Average
Percent
--
--
Maximum
Percent
--
--
Plume
Attenuation
Coefficient
(sec . -meters- 1 )
34.02
95.36
-
64.69
Maximum
At tenuat ion
Coefficient
(me t er s~ 1 )
1.067
2.602
•
1.834
-------
- 73 -
No correlation could be found between v.rl.tlon. In .1*.
Distribution (fraction, 1... th.n one and five »tcrons) measured
duri.8 each of the individual te.ts and average net coHi»« time
for oven, pushed during each test (Tab!e 5.5-1). Oven tempera-
tures, however, were found to be statistically signif ipan.ly
correlated with the percentage of particles less than five .i.,p«
in diameter, but not with the percentage 1... th» one »icrpn -in
ciameter. (Oven temperature data were considered propriety infor-
mation and are not included in this report.)
The percentage of organic material (i.e., soluble In ,.c,t,one
.na cvclohe«ne) present in particle size samples was determined
* .extracting the r«id.e collected in the .vcl.one .or the ,,«
and on . ,c«,bi«.tlon of
«•» Brin* .sample,., th. mean or.ga,n-t.c cpnt.en* .of
„ ,,10.e «.
tk. p.a.t^-.t. .matter ce.u.6h, in the ,y,lone ,or
, was Sound - b.e .significantly l.e,,s ,th._n ,he «.. .organic
tor the .c-ombinstion of ..11 .other stages., A4,6 percent. The
«« «... -t-otf wa. A.6 »ic.Mns for thM.e
-suit was ob,.i.»ed for .the A,nd,r..en «mpXe... The
content *f «. Combined residua for t,he ,,.r,th .,t.,g, ***»
final filter .was *»* « *• •W**'"* —^ thM '*" "*
6he combined residues rrom a«.ges ,0 end .1 and .th.e .combined -r,,ld«e.
..tages 1 -and ». .The ;c»t-off for .this «-l .portion ,o,f
..verged 3.« mi,r.n. «* -he mea,
percent .
The ,co.centr.r*o,, -of f «t«.W -particulate .«.«er
-------
- 74 -
displayed in Table 5.5-2, indicate a range from 0.007 to 0.089
gr/DSCF. In consideration of the relatively short sampling period
used for these tests, the results, on a whole, compare favorably
with those obtained during the pushing-cycle particulate tests.
5 .6 Door Leak Rates
The leaking of coke-side oven doors is more apparent immediately
after the charging of an oven than late in the coking period. Figure
5.6 shows the frequency of oven leaks at various times after oven
charging. These data were accumulated from those found in Appendix Y
(Volume 3) and proprietary production data. Appendix Y indicates
which ovens were observed to be leaking at various observation times
during the study. The 75 leaking-door observations indicated -in
Figure 5.6 show a gradual decay of frequency of door leaks as a
function of the residence time of the coal in the charged oven.
These data suggest that volatile materials from the coked coal are
emitted at greater rates at the beginning of the coking period than
later in the coking period.
5.7 Emission-Related Correlations
5.7.1 Correlations Between Pushing-Cycle Filberable
Particulate Emission Factors and Operating Data
The source testing in this project yielded four pushing-
cycle particulate tests for which emission factors have been
computed in terms of pounds per ton of dry coal fed and pounds
per ton of coke produced. One of the objectives of the project
was to identify the operational variables which may affect the
level of the emission factor, such as net coking time and
average oven temperature. Another was to identify optical
-------
- 75 -
TABLE 5.5-2
CONCENTRATION OF PARTICULATE MATTER
CALCULATED FROM PARTICLE SIZING SAMPLES
Great Lakes Carbon Corporation
St. Louis, Missouri
April 21-24, 1975
Test
No.
Brink-1
Brink-2
Brink-3
Brink-4
Brink-5
Brink-6
Brink-7
Brink-8
Brink-9
Andersen- 10
Andersen- 11
Andersen- 1 2
Andersen-13
Andersen- 14
1975
Date
4-22
4-22
4-22
4-22
4-22
4-23
4-23
4-23
4-23
4-23
4-23
4-24
4-24
4-24
Sampling
Period
Start
09:43
14:05
14:05
18:09
18:09
09:16
09:16
16:40
16:40
13:02
17:12
08:28
10:34
12:46
Stop
10:33
16:28
16:28
18:59
18:59
15:00
15:00
17:35
17:35
14:22
17:24
08:41
11:24
13:26
Sampled
Volume
(DSCF)
4.32
10.8
11.1
4.27
3.89
16.9
15.2
4.31
4.31
69.0
9.33
10.4
41.9
35.0
Sample
Weight
(mg)
7.6
13.7
10.0
6.9
7.3
10.9
7.2
7.3
9.5
94.3
32.0
18.7
40.4
202.3
Particulate
Concentration
(gr/DSCF)
0.027
0.020
0.014
0.025
0.029
0.010
0.007
0.026
0.034
0.021
0.053
0.028
0.015
0.089
-------
Number
of
Oven
Leaks
12-1
11
1 '
FIGURE 5.6
COKE-SIDE DOOR LEAKS AFTER OVEN CHARGING
Great Lakes Carbon Corporation
St. Louis, Missouri
April 21-24, 1975
"vl
ON
10 12 14 16 18 20 22
Time Since Charging (hours)
24
26 28
40 42
-------
- 77 -
emission characteristics which correlate with the emission
factors. Because of the very limited data set, it is diffi-
cult to examine the effect of several variables acting
simultaneously. Therefore, the relationships between the
particulate emission factors, operating parameters, and
optical plume characteristics are examined by consideration
of only one "independent" variable at a time.
The emission factors, in terms of pounds of filterable
particulate per ton of dry feed, were plotted as a function
of the average net coking time for each of the four pushing-
cycle particulate tests (Figure 5.7.1-1). As expected, the
more fully-coked product (longer net coking time) results in
reduced filterable particulate emission factors. A log-log
relationship was found to yield a superior statistically
significant relationship for the four sets of data available
from this study.
Figure 5.7.1-2 displays the filterable particulate
emission factors as a function of average oven temperature.
(Again, oven temperature data were considered proprietary
and not included in this report.) Based upon the limited data
acquired, the correlation is not significant. This may be
due to the fact that oven temperatures are recorded only once
per shift by plant personnel; thus, the single reading may not
represent the actual range of temperatures for those ovens
pushed during a particulate test. Additionally, only three
particulate samples are available during which oven temperatures
were recorded.
-------
FIGURE 5.7.1-1
1.0
0.9H
0.8
EFFECTS OF COKING TIME ON PARTICULATE EMISSIONS
COKING TIME VERSUS FILTERABLE PARTICULATE EMISSIONS
PUSHING-CYCLE PARTICIPATE TESTS 1-4
Great Lakes Carbon Corporation
St. Louis, Missouri
April 21-24, 1975
Correlation Coefficient
-0.964
Filterable Particulate
(Ib/ton of feed)
ln(emission factor) =
60.2 - 8.2 * ln(coking time)
0.6
0.5
0.4
0.3
0.2
0.1-
CO
I
1500
I ^
1550
1600 1650
Average Net Coking Ti
1700
e (minutes)
1750
• i
1800
-------
FIGURE 5.7.1-2
1.0
0-. 9
0.8-
AVERAGE OVEN TEMPERATURE VERSUS FILTERABLE PARTICULATE EMISSIONS
PUSHING-CYCLE PARTICULATE TESTS 1-3
Great Lakes Carbon Corporation
St. Louis,Missouri
April 21-24, 1975
Filterable Particulate
(Ib/ton of feed)
0.6
0.5-
0.4-
0.3
0.2-
0.1
VO
I
1900 1920 1940 1960 1980 2000 2020 2040
Average Oven Temperature (°F)
2060
2080
3000
-------
- 80 -
5.7.2 Correlations Between Pus.kiag-Cycle Filterable Particulate
Emission Factors and Indices of Visible Emissions
Filterable particulate emission factors exhibited a
statistically significant correlation with the average degree-
of-greenness ratings for the tests (Figure 5.7.2-1). Unfortu-
T
nately, as shown in Table 5.1-2, Pushing-Cycle Particulate Tests 1,
2, and 3 resulted in identical average degree-of-greenness
ratings, an occurrence which limits the usefulness of correla-
tion analysis when so few data are available; nevertheless, the
empirical correlation is dramatic in this case.
The two parameters of optical density measured with
the transmissometer were each plotted against the filterable
particulate emission factor, pounds per ton of dry feed, for
the four pushing-cycle particulate tests. Figure 5.7.2.-2
shows the filterable particulate emission factor as a function
of the average maximum attenuation coefficients for the pushes
included in a particulate test. Although the linear relation-
ship between these two variables appears to be reasonable, a
statistically significant correlation was not found, due to the ,
limited amount of data. Figure 5.7.2-3 presents emission factors
as a function of the plume attenuation coefficient integrated
over time. The correlation in this case was found to be statis-
tically significant. Both plots indicate that Test 2 resulted
in a somewhat lower particulate emission than would be expected
from the results of the other three tests, based on transmissometer
data. Nevertheless, increased optical density obviously
accompanied elevated filterable particulate emission factors
during the four pushing-cycle particulate tests.
-------
FIGURE 5.7.2-1
1.0
0.9
0.8
DEGREE OF GREENNESS VERSUS FILTERABLE PARTICULATE EMISSIONS
PUSHING-CYCLE PARTICULATE TESTS 1-4
Great Lakes Carbon Corporation
St. Louis, Missouri
April 21-24, 1975
Filterable Particulate
(Ib/ton of feed)
0.6
0.5
0.4
0.3
0.2
0.1
i
oo
Correlation Coefficient = 0.954
190 200 210 220 230 240 250 260 270 280 290 300
Average Degree of Greenness (S * D)
-------
FIGURE 5.7.2-2
1.0
0.9
0.8
MAXIMUM ATTENUATION COEFFICIENT VERSUS FILTERABLE PARTICULATE EMISSIONS
PUSHING-CYCLE PARTICULATE TESTS 1-4
Great Lakes Carbon Corporation
St. Louis, Missouri
April 21-24, 1975
Filterable Particulate
(Ib/ton of feed)
0.6
0.5
0.4
0.3
0.2
0.1
I
CD
0.2
0.3
0.4 0.5 0.6 0.7 0.8
Maximum Attenuation Coefficient (meters"*-)
0.9
1.0
-------
FIGURE 5.7.2-3
PLUME ATTENUATION COEFFICIENT VERSUS FILTERABLE PARTICULATE EMISSIONS
PUSHING-CYCLE PARTICULATE TESTS 1-4
Great Lakes Carbon Corporation
St. Louis, Missouri
April 21-24, 1975
1.0
0.9
0.8
Filterable Particulate
(Ib/ton of feed)
0.6
0.5 •
0.4
0.3
0.2
0.1
®
Correlation Coefficient = 0.952
oo
to
8 10 12 14 16 18 20 22 24 26 28 30 32 34
Plume Attenuation Coefficient (sec.-meters"M
-------
- 84 -
Quench tower opacity data were not found to be statis-
tically significantly correlated with filterable particulate
emission factors for pushing-cycle tests. This lack of
correlation may be attributable to the small amount of data
and the limitation of reading opacities in the presence of
the steam plume from the quenching operation. Figure 5.7.2-4
displays a graph of quench tower opacity as a function of
filterable particulate emission factor.
5.7.3 Correlations Among Visible Emissions Parameters
Four optical emission characteristics were monitored
independently in the project: degree-of-greenness (DOG),
maximum attenuation coefficient (MAC), integrated attenuation
coefficient (IAC), and quench tower opacity (QTO). These four
variables can be paired such that six two-variable combinations
can be examined. Statistical .analysis shows clearly that all
combinations are highly interrelated, as shown below in order
of decreasing linear correlation coefficients:
Combination
MAC and IAC
IAC and QTO
MAC and QTO
DOG and QTO
DOG and IAC
MAC and DOG
Number of
Observe tions
41
33
33
33
41
41
Correlation
Coefficient
0.996
0.818
0.814
0.810
0.8C9
'0.797
-------
FIGURE 5.7.2-4
FILTERABLE PARTICULATE EMISSIONS VERSUS AVERAGE QUENCH TOWER OPACITY
PUSHING-CYCLE PARTICULATE TESTS 1-3
17.0-,
16.0-
Great Lakes Carbon Corporation
St. Louis,Missouri
April 21-24,1975
Average Quench Tower Opacity
(percent)
J
14.0-
13.0-
12;0
11.0-
10.0
9.0
8.0
oo
in
0.1 0.2 '0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0
Filterable Particulate (Ib/ton of feed)
-------
- 86 -
Individual degree-of-greenness ratings for the ovens
observed during the four particulate tests were plotted
against the maximum optical density measured by the trans-
missometer for each push (Figure 5.7.3). As expected, high
degree-of-greenness ratings resulted in generally higher
optical densities. The scattering of data points observed is
likely due to the subjectivity of the degree-of-greenness
rating as well as the dispersion of the coke-pushing plume
beneath the shed following the degree-of-greenness observa-
tion but prior to the plume passing the transmissometer beam.
5.8 Significance of Emissions of Other Contaminants
Emission testing for gases and other contaminants during th'is
sampling program indicated that minor quantities of all gaseous
constituents were found for all tests (Table 5.8 and Appendix E,
Volume 2). The polynuclear aromatic compounds and those with
similar structures (pyridine, phenolic compounds, benzo(a+e)pyrene,
chrysene, fluoranthene, and pyrene) were not found in detectable
quantities. The average emission rates of benzene and benzene and
its homologues were less than one pound per hour, while sulfur
dioxide plus sulfur trioxide emissions averaged less than three
pounds per hour.
The emission rate of carbon monoxide, resultant from the in-
complete conbustion of the freshly-pushed coke, averaged 14 pounds
per hour. It should be noted that this emission ra'te and those of
the light hydrocarbon compounds were instantaneous rates measured
during a push, which is likely the peak emission period in the
overall cycle. Total light hydrocarbon emissions averaged only
seven pounds per hour.
-------
FIGURE 5.7.3
2.6
2.4
2.2
DEGREE OF GREENNESS VERSUS MAXIMUM ATTENUATION COEFFICIENT
PUSHING-CYCLE PARTICULATE TESTS 1-4
Great Lakes Carbon Corporation
St. Louis, Missouri
April,21-24, 1975
2.0
1.8
Correlation Coefficient = 0.797
Maximum Attenuation. Coefficient
(meters'1)
I
00
50
100
150 200 250 300 350
Degree of Greenness (S * D)
400
450
500
-------
- 88 -
TABLE 5.8
SUMMARY OF CONTAMINANT EMISSION RATES (LBS/HOUR)
Coke Shed
Great Lakes Carbon Corporation
St. Louis, Missouri
April 21-24, 1975
Contaminant
Acetylene*
Benzene
Benzene & Homologues
Filterable Benzo (a ) pyrene
Total Benzo(a)pyrene
Carbon Monoxide*
Gaseous Chloride
Filterable Chrysene
Total Chrysene
Gaseous Cyanide
Filterable Cyclohexane Solubles
Total Cyclohexane Solubles
Filterabl'e Cyclohexane Insolubles
Total Cyclohexane Insolubles
Ethene & Homologues*
Filterable Fluoranthene
Total Fluoranthene
Total Light Hydrocarbons (as CH^)*
Methane & Homologues*
Gaseous Nitrogen Oxides (as N02)
Gaseous Phenolics
Filterable Pyrene
Total Pyrene
Pyridine
Gaseous Sulfate
Gaseous Sulfite
Sulfur Dioxide
Sulfur Trioxide
Test No.
/
1
0.48
0.35
0.48
<0.11
<0.16
24.2
0.63
<0.06
<0.09
0.002
<15.4
15. 4
<9.3
<10.8
2.4
<0.05
<0.08
7.9
6.7
0.09
<0.54
<0.05
<0.07
<0.03
3.5
<0.08
0.42
3.8
2
0.10
0.57
0.73
<0.08
<0.13
10.7
0.43
<0.05
<0.07
0.008
12.1
1 3_ 1
206
206
1.2
0.05
<0.07
9.1
8.0
0.06
<0.43
<0.07
<0.08
<0.03
0.51
1.1
1. 1
1.3
3
0.25
1.0
1.3
<0.10
<0.15
8.1
0.40
<0.06
<0.09
0.002
14.3
«•_•; — — -• • — '
28.7
28.7
1.1
<0.05 -
<0.07
4.7
4.2
0.05 '
<0.34
<0.06
<0.08
<0.03
1.2
0.11
0.84
0.84
Average
0.28
0.64
0.84
<0.10
<0.15
14.3
0.49
<0.06
<0.08
0.004
R . R_i •*_ o
"303
•**/•*/
78.2-81.3
78.2-81.8
1.6
0.02-0.05
<0.07
7.2
6.3
0.07
<0.44
<0.06
, <0.08
<0.03
1.7
0.40-0.43
0.79
2.0
*Emission rates are maximum (short-term) rates measured during oven pushing.
-------
- 89 -
Results of the caustic solution absorption tests indicated
that cyanide was emitted at an average of 0.004 pound per hour.
Fluoride, nitrogen oxide compounds, and sulfate and sulfite com-
pounds were also present in minor amounts.
5.9 Assessment of the Shed's Impact Upon Dustfall in the Work
Environment
Table 5.9-1 presents the dustfall data collected at the various
sites within the shed and in similar locations in the unshedded
C Battery. Chemical characteristics of selected dustfall samples
are presented in Table 5.9-2.
In order to identify how the shed affects measurable dustfall
•rates, other potentially-influential factors were first evaluated.
These other variables were: a) greenness of the pushes, b) pushing
rate, and c) location of the dustfall bucket. The data used for
the analyses are summarized in Table 5.9-3. All statistical analy-
ses were performed using the logarithms of the dustfall rates since
(4)
dustfall rates are known to be log-normally distributed.
As shown in Table 5.9-3, nine pairs of simultaneous samples
were collected. An initial test for statistical outliers was
performed using these paired data. To determine the precision of
each pair of samples, the difference in the logarithms of the
paired values was divided by the geometric mean of the pair. These
nine precision values, expressed as percentages, were then evaluated
to determine if any pair could be considered an outlier. The
pair of samples taken on the No. 12 Bench on April 23 was classi-
fied as an outlier in this manner and was not used in further
analyses. The precision value for this pair was 26.0 percent,
while those for the other eight pairs ranged from 0.1 to 2.6 percent.
-------
TABLE 5.9-1
SUMMARY OF DUSTFALL MEASUREMENTS
Great Lakes Carbon Corporation
St. Louis, Missouri
April 21-24, 1975
Site
No. 12 Bench
No. 12 Bench
No. 12 Bench
No. 12 Bench
No. 12 Bench
No. 12 Bench
No. 12 Bench
No. 12 Bench
No. 12 Bench
No. 12 Bench
No. 31 Bench
No. 31 Bench
No. 31 Bench
No. 31 Bench
Sampling Period
Start ] Stop
1975
Date
4/21
4/21*
4/22
4/22*
4/22
4/22*
4/23
4/23*
4/23
4/23*
4/21
4/22
4/22
4/23
Time
15:20
17:55
09:30
09:30
18:25
18:25
09:48
09:48
18:25
18:25
15:30
09:45
18:25
09:48
1975
Date
4/22
4/22
4/22
4/22
4/23
4/23
4/23
4/23
4/24
4/24
4/22
4/22
4/23
4/23
Time
08:20
08: 20
18:02
18:02
08: 15
08:15
17:45
17:45
08:50
08:50
09:30
18:02
08: 15
17:46
Settleable
Farticulate
gm/m^/wk
8470
- 8550
1570
1120
1250
Void
21,000
2120
5880
5280
4220
1200
Void
2100
tons /mi ^ /mo
104.000
105,000
19,200
13,700
15,200
Void
257,000
25,900
71,900
64,600
51,600
14,700
Void
25,700
Weight Collected
On No. 18 Sieve
Weight
(gm)
4.9669
2.8720
0.1624
0.3010
0.8763
Void
10.1142
1.4853
6.5937
7.0882
4.5982
0.1959
Void
0.9013
Percent
of Total
24.1
17.7
10.1-
22.4
31.9
Void
35.8
44.8
41.7
46.2
35.8
15.3
Void
33.1
Water Soluble
Dustfall
Percent of
Settleable
Part icula te
0.6
0.7
0.5
0.6
7.1
Void
0.7
1.3
0.2
0.2
0.5
0.4
Void
0.5
*Duplicate, Sample
-------
TABLE 5.9-1 (continued)
SUMMARY OF DUSTFALL MEASUREMENTS
Great Lakes Carbon Corporation
St. Louis, Missouri
April 21-24, 1975
Site
No. 31 Bench
No. 37 Bench
No. 37 Bench
No. 37 Bench
No. 37 Bench
No. 37 Bench
No. 46 Bench
No. 46 Bench
No. 85 Bench
No. 85 Bench
No. 85 Bench
No. 85 Bench
No. 94 Bench
No. 94 Bench
Sampling
Start
1975
Date
4/23
4/21
4/22
4/22
4/23
4/23
4/21
4/22
4/22
4/22
4/23
4/23
4/21
4/22
Time
18: 29
15:32
10:00
18: 25
09:48
18: 29
15:35
10:08
10:39
18:48
09: 20
18:50
15:44
10: 20
Period
Stop
1975
Date
Void
4/22
4/22
4/23
4/23
4/24
4/22
4/22
4/22
4/23
4/23
4/24
4/22
4/22
Time
Void
09:45
18:02
08: 15
17:48
08:48
10:01
18:02
18: 18
08: 20
18:50
08:44
10: 12
18: 18
Settleable
Particulate
gm/m^/wk
Void
1540
1080
4260
2570
Void
1580
1490
498
3100
1960
5720
495
695
n
tons/mi /mo
Void
18,800
13 ,200
52 ,100
31 ,500
Void
19,400
18,200
6090
37,900
24,000
69,900
6060
8500
Weight Collected
On No. 18 Sieve
Weight
(gm)
Void
3.3320
2.7908
6.9827
2.0489
Void
2.6439
2.4716
0.6736
12.4319
0.5892
36.5175
0.7866
0.7572
Percent
of Total
Void
52.3
74.7
52.2
47.9
Void
45.5
65.9
61.9
73.2
22.6
80.9
44.2
55.8
Water Soluble
Dustfall
Percent of
Settleable
Particulate
Void
0.9
2.3
0.9
0.8
Void
0.9
0.8
3.8
3.3
0.6
1.2
0.9
3.2
^Duplicate Sample
-------
TABLE 5.9-1 (continued)"
SUMMARY OF DUSTFALL MEASUREMENTS
Great Lakes Carbon Corporation
St. Louis, Missouri
April 21-24, 1975
Site
No. 94 Bench
No. 94 Bench
No. 94 Bench
No. 106 Bench
No. 106 Bench
No. 106 'Bench
No. 106 Bench
No. 106 Bench
No. 117 Bench
No. 117 Bench
No. 2 Wall
No. 2 Wall
No. 2 Wall
No. 2 Wall
Sampling Period
Start
1975
Date
4/22
4/23
4/23
4/21
4/22
4/22
4/23
4/23
4/21
4/22
4/21
4/22
4/23
4/23
Time
18:48
09:20
18:53
15:45
10:25
18:48
09: 20
18:58
15:57
10:35
16: 17
11:50
10: 06
18:34
Stop
1975
Date
4/23
4/23
4/24
4/22
4/22
4/23
4/23
4/24
4/22
4/22
4/22
4/22
4/23
4/24
Time
08:20
18:50
08:42
10:20
18:18
08:20
18:50
08:40
10:27
18: 18
11:40
18: 04
18:53
08:55
Settleable
Particulate
gm/m /wk
2390
1770
1830
1480
936
2280
1590
1440
1650
1240
8560
14,900
3020
8190
t\
tons/mi^/mo
29 ,200
21,600
22,400
18,100
11,400
27,900
19,400
17,600
20,200
15,100
105,000
182,000
36,900
100,000
Weight Collected
On No. 18 Sieve
Weight
(gm)
1.6997
0.6911
1.0243
7.6222
0.2083
1.8557
0.6449
2.3869
4.9051
0.6003
0.8860
0.3769
0.1629
1.3156
Percent
of Total
32.7
27.5
27.1
71.9
20.6
35.7
28.2
52.8
59.7
36.7
4.7
3.6
5.4
9.3
Water Soluble
Dustfall
Percent of
Settleable
Particulate
1.2
1.4
0.7
1.1
2.9
2.2
1.0
1.3
2.6
1.4
0.1
0.2
0.3
0.2
^Duplicate Sample
-------
TABLE 5.9-1 (continued)
SUMMARY OF DUSTFALL MEASUREMENTS
Great Lakes Carbon Corporation
St. Louis, Missouri
April 21-24, 1975
Site
No. 25 Wall
No. 25 Wall
No. 25 Wall
No. 25 Wall
No. 25 Wall
No. 36 Wall
No. 36 Wall
No. 36 Wall
No. 36 Wall
No. 36 Wall
No. 46 Wall
No. 46 Wall
No. 46 Wall
No. 46 Wall
Sampling
Start
1975
Date
4/21
4/22
4/22
4/23
4/23
4/21
4/22
4/22
4/23
4/23
4/21
4/21*
4/22
4/22*
Time
17:38
12:20
19:18
10:06
18:32
17:44
11:35
19:16
10:12
18:36
18:03
18:03
11:20
11: 20
Period
Stop
1975
Date
4/22
4/22
4/23
4/23
4/24
4/22
4/22
4/23
4/23
4/24
4/22
4/22
4/22
4/22
Time
12:15
18:06
08:40
18:55
09: 00
11:25
18:08
08:45
18:57
09:04
11:05
11:05
18: 10
18: 10
Settleable
Particulate
gm/m^/wk
10,700
21,200
11,400
11,000
9120
14,600
9480
9980
12,300
12,700
6720
8510
6720
6980
tons/mi^/mo
131.000
260,000
139,000
135,000
112,000
178,000
116,000
122,000
151,000
155,000
82,200
104,000
82,200
85,400
Weight Collected
On No. 18 Sieve
Weight
(gm)
0.6533
0.0121
1.0135
0.7194
119715
0.0523
0.4321
1.0076
1.3069
4.0224
0.5757
0.6036
0.3075
0.1829
Percent
of Total
2.9
0.1
5.8
6.4
12.1
0.2
6.0
6.5
10.0
16.8
4.4
3.7
5.8
3.4
Water Soluble
Dustfall
Percent of
Settleable
Particulate
0.1
0.01
0.2
0.2
0.2
0.2
0.2
0.2
0.2
0.1
0.2
0.2
0.5
0.3
^Duplicate Sample
-------
TABLE 5.9-1 (continued)
SUMMARY OF DUSTFALL MEASUREMENTS
Great Lakes Carbon Corporation
St. Louis, Missouri
April 21-24, 1975
Site
No. 46 Wall
No. 46 Wall
No. 46 Wall
No. 46 Wall
No. 46 Wall
NO. 46 Wall
No. 85 Wall
No. 85 Wall
No. 85 Wall
No. 85 Wall
No. 106 Wall
No. 106 Wall
No. 106 Wall
No. 106 Wall
Sampling
Start
1975
Date
4/22
4/22*
4/23
4/23*
4/23
4/23*
4/21
4/22
4/22
4/23
4/21
4/22
4/22
4/23
Time
19:14
19:14
10: 12
10: 12
18:45
18:45
17:08
13:20
19:05
19:20
17:09
13:17
19:05
19: 25
Period
Stop
1975
Date
4/23
4/23
4/23
4/23
4/24
4/24
ft/22
4/22
4/23
4/24
4/22
4/22
4/23
4/24
Time
08:45
08:50
18:58
18:58
09:06
09:06
13: 15
18:54
18:50
08:33
13: 13
18:52
18:50
08:31
Settleable
Particulate
f\
gm/m^/wk
8000
9260
6130
5940
4530
4380
78.4
106
104
378
606
502
553
841
tons/mi/mo
97,800
113,000
75,000
72,700
55,400
53,600
958
1290
1280
4630
7410
6140
6760
10,300
Weight Collected
On No. 18 Sieve
Weight
(gm)
0.7842
0.6834
0.1806
0.3742
0.9768
0.9736
0.0086
1.1405
0.0163
0.0184
0.2106
0.0052
0.0541
0.0712
Percent
of Total
6.3
4.8
3.0
6.2
12.2
12.5
4.8
94.7
5.7
3.3
13.8
1.7
3.7
5.6
Water Soluble
Dustfall
Percent of
Settleable
Particulate
0.2
0.2
0.3
0.2
0.4
0.4
1.8
38.0
6.2
<0.02
0.3
0.9
0.4
0.1
*Duplicate Sample
-------
TABLE 5.9-1 (continued)
SUMMARY OF DUSTFALL MEASUREMENTS
Great Lakes Carbon Corporation
St. Louis, Missouri
April 21-24, 1975
Site
No. 117 Wall
No. 117 Wall
No. 117 Wall
No. 117 Wall
Car No. 1
(South)
Car No. 1
(South)
Car No. 1
(South)
Car No. 1
(South)
Car No. 1
(South)
Car No. 2
(North)
Car No. 2
(North)
Car No. 2
(North)
Car No. 2
(North)
Car No. 2
(North)
Sampling Period
Start
1975
Date
4/21
4/22
4/22
4/23
4/21
4/22
4/22
4/23
4/23
4/21
4/22
4/22
4/23
4/23
Time
17: 11
13: 14
19:05
19:30
16:52
13:35
19: 12
09:43
19:10
15:59
10:50
19: 15
09: 22
19: 15
Stop
1975
Date
4/22
4/22
4/23
4/24
4/22
4/22
4/23
4/23
4/24
4/22
4/22
4/23
4/23
4/24
Time
13:08
18:50
18:50
08:30
13:30
18:45
08:35
19:00
08:45
10:42
19:05
08:40
19:00
08:47
Settleable
Particulate
ey
gm/m /wk
841
587
1010
458
3190
1920
4720
3750
2490
1340
1310
3120
2240
810
f\
tons/mi^/mo
10,300
7180
12,300
5600
39,000
23,500
57,700
45,800
30,500
16,400
16.000
38.100
27.400
9900
Weight Collected
On No. 18 Sieve
Weight
(gm)
0.1430
0.1996
0.1766
0.0254
0.3565
0.1091
0.4887
0.6354
0.2818
1.1897
0.1511
0.4066
0.2495
0.1435
Percent
of Total
7.3
35.9
6.4
3.8
4.8
9.2
6.7
14.4
7.1
30.4
11.4
8.2
9.6
10.8
Water Soluble
Dustfall
Percent of
Settleable
Particulate
0.2
1.0
0.4
<0.02
0.2
0.5
0.4
0.2
0.3
0.3
0.3
0.5
0.3
<0.008
^Duplicate Sample
-------
TABLE 5.9-1 (continued)
SUMMARY OF DUSTFALL MEASUREMENTS
Great Lakes Carbon Corporation
St. Louis, Missouri
April 21-24, 1975
Site
Overhead
No. 14
Overhead
No. 14
Overhead
No. 26
Overhead
No. 26
Overhead
No. 26
Overhead
Sampling
Start
1975
Date
4/22
4/23
4/23
4/22
4/23
4/23
Time
*?;35
09:32
19:00
19:30
09:32
19:00
Period
Stop
1975
Date
4/23
4/23
4/24
4/23
4/23
4/24
Time
09:00
19:00
08:48
09:00
19:00
08:48
Settleable
Part iculate
gm/m^/wk
3750
4280
5800
2760
2930
5000
tons/mi^/mo
45,800
52.400
71,000
33.800
35.800
61,200
•
Weight Collected
On No. 18 Sieve
Weight
(gm)
Oi3123
0.2719
0.4548
0.2131
0.2184
0.7365
Percent
of Total
5.4
5.8
5.0
5.0
6.8
8.9
Water Soluble
Dustfall
Percent of
Settleable
Particulate
0.4
0.3
0.1
0. 7
0.3
0.2
^Duplicate Sample
-------
TABLE 5.9-2
CHEMICAL CHARACTERIZATION OF DUSTFALL*
Great Lakes Carbon Corporation
St. Louis, Missouri
April 21-24, 1975
Sampl ing
Site
No. 12
Bench
No. 31
Bench
No. 46
Bench
No. 85
Bench
No. 94
Bench
No. 46
Wall
Sampling Period
Start
Date
4/23
4/22
4/21
4/23
4/23
4/22
Time
09:48
09:45
15:35
09: 20
18:53
19: 14
Stop
Date
4/23
4/22
4/22
4/23
4/24
4/23
Time
17:45
18:02
10:01
18:50
08:42
08:45
Acetone Solubles
Weight
(gm)
0.0176
0.0014
0.0127
0.0142
0.0389
0.0038
Percent
of Total
1.0
0.1
0.4
0.7
1.4
0.03
Cyclohexane Solubles
Weight
(gin)
0.0007
<0.0006
0.0010
<0.0006
0.0029
0.0013
Percent
of Total
0.04
<0.06
0.03
0.03
0.1
0.01
PH
5.70
5.80
6.48
4.88
4.92
6.80
* These data represent the portion of each dustfall sample which passed through the
No. 18 sieve.
-------
- 98 -
TABLE 5.9-3
DUSTFALL SUMMARY (gm/m /wk)
Great Lakes Carbon Corporation
St. Louis, Missouri
April 21-24, 1975
Sampling Location
No. 12 Bench
No. 12 Bench*
No. 12 Bench
Geometric Mean
No. 31 Bench
No. 37 Bench
No. 46 Bench
Shedded Bench
Geometric Mean
No. 85 Bench
No. 94 Bench
No. 106 Bench
No. 117 Bench
Unshedded Bench
Geometric Mean
No. 2 Wall
No. 25 Wall
No. 36 Wall
No. 46 Wall
No. 46 Wall*
No. 46 Wall
Geometric Mean
Shedded Wall
Geometric Mean
No. 85 Wall
No. 106 Wall
No. 117 Wall
Unshedded Wall
Geometric Mean
South Car
(Shedded)
North Car
(Unshedded)
No. 14 Overhead
No. 26 Overhead
Shedded Overhead
Geometric Mean
1975 Sampling Period
4/21-22
8,470
8,550
8,510
4,220
1,540
1,580
3,060
495
1,480
1,650
1,070
8,560
10,700
14,600
6,720
8,510
7,560
10,000
78.4
606
841
342
3,190
1,340
--
4/22
1,570
1,120
1,330
1,200
1,080
1,490
1,270
498
695
936
1,240
796
14,900
21,200
9,480
6,720
6,980
6,850
9,890
106
502
587
315
1,920
1,310
--
--
4/22-23
1,250
VOID
1,250
VOID
4,260
2,310
3,100
2,390
2,280
2,570
11,400
9,980
8,000
9,260
8,610
9,930
104
553
1,010
387
4,720
3,120
3,750
2,760
3,220
4/23
21,000**
2,120**
--
2,100
2,570
2,320
1,960
1,770
1,590
1,770
3,020
11,000
12,300
6,130
5,940
6,030
7,050
--
_ _
3,750
2,240
4,280
2,930
3,540
4/23-24
5,880
5,280
5,570
VOID
VOID
5,570
5,720
1,830
1,440
1,620
8,190
9,120
12,700
4,530
4,380
4,450
8,060
378
841
458
526
2,490
810
5,800
5,000
5,390
Geometric
Mean
_ _
2,980
2,200
2,070
1,530
2,420
1,450
1,220
1,490
1,430
1,380
7,490
10,500
11,700
6,540
8,800
134
613
691
385
3,060
1,580
4,530
3,430
3,940
* Duplicate Sample.
**Statistical tests indicate that these values are suspect.
not used in further statistical analyses.
They were
-------
- 99 -
In all additional evaluations, the geometric mean dustfall rate
was then used for the remaining paired samples.
An average greenness for ovens pushed during each daytime
dustfall sample was determined by averaging the value of S*D for
the pushes that occurred during the sampling period. For dustfall
samples taken within the shed, pushes at Ovens 1 through 55 were
used; for unshedded samples, pushes at Ovens 83 through 132 were
used. These average greenness values were then arranged in as-
cending order to determine a median value, 160. All greenness
values below 160 were labeled "low" and all above 160 were labeled
"high." It is interesting to note that none of the values for un-
shedded samples were associated with a "high" average greenness.
Also, only six of the 43 shedded samples for which greenness data
were available had average greenness values considered to be "low."
A pushing rate for each dustfall sample was determined and
normalized by counting the number of either shedded or unshedded
ovens, as applicable, that were pushed during a sample, dividing
by the time duration of the sample, and then dividing by the num-
ber of ovens in the shedded or unshedded area. Again, the values
were arranged in ascending order and the median was found to be
0.037 push per hour per oven. All pushing rates below this value
were considered "low" and all rates equal to or above this value
were considered "high."
The dustfall data were then arranged into several cells in
order to best eliminate any confounding effect of the multiple
variables. These cells, shown in Table 5.9-4, were defined by
first dividing the data into that applicable to shedded and unshedded
-------
- 100 -
TABLE 5.9-4
FORMAT USED FOR ANALYSES OF DUSTFALL DATA (gm/m /wk)
Great Lakes Carbon Corporation
St. Louis, Missouri
April 21-24, 1975
Battery
Greenness
Pushing Rate
L
0
A
T
1
0
N
Bench
Wall
Car
Overhead
Shedded
Low
Lew
11,400
9980
8610
4720
3750
2760
High
High
Low
1250
1200
1080
4260*
1490
14,900
3020
10,700
21,200
11,000
9480
12,300
6850
6030
3190
3750
4280
2930
High
8510
1330
4220
2100
1540
2570
1580
8560
14,600
7560
1920
Unshedded
Low
Low
695
936
106
502
587
1310
High
498
1960
495
1770
1480
1590
1650
1240
78.4
104
606
553
841
1010
1340
2240
•
High
Low
High
Within this combination of variables, this value was judged to be
an outlier and was not included in further analyses.
-------
- 101 -
areas. Each area was subdivided into one of four locations: "bench,"
"wall," "car," or "overhead." The next two subdivisions were those
of "low" and "high" pushing rates and "low" and "high" greennesses.
A single cell now contained the most homogeneous subset of data
available. Tests for statistical outliers were then conducted
within each cell using the logarithms of the dustfall rates. Only
a single value, as indicated in Table 5.9-4, was found to be an
outlier at this stage and was not used in subsequent analyses.
In order to determine whether greenness and dustfall rate
were correlated, the number of subdivisions was reduced by one
so tha t greenness was no longer used as a basis of subdivision.
In each of the remaining 16 cells, the logarithm of dustfall rate
for each sample was paired with its average greenness value. The
linear correlation coefficient for the pairs in each cell was then
determined. Since none of the correlation coefficients was found
to be significant at the five-percent level, it was concluded that
greenness and dustfall rate were not correlated for this set of
data.
Since greenness and dustfall rate were not found to be corre-
lated, those dustfall rates which did not have a greenness rating
associated with them could now be included in further analyses.
Thus, these values were added to their respective cells determined
in the previous analysis, and the tests for outliers were repeated.
No additional suspect values were found.
The correlation between pushing rate and dustfall rate was
evaluated next. The number of subdivisions was reduced by one
by eliminating pushing rate as a basis of division. In each of
the eight remaining cells, the logarithm of dustfall rate was then
-------
- 102 -
paired with its associated pushing rate. The linear correlation
coefficient was determined for each cell, and only two of the values
were found to be significant at the five-percent level. These
were the 10 data pairs for shedded and unshedded car locations.
On the basis of the fact that five of the seven correlation coef-
ficients were not significant, it was concluded that pushing rate and
dustfall rate were not significantly related for the overall data set.
Two factors remained to be considered — the location of the
dustfall'bucket and the shed effect, i.e., shedded versus unshedded
areas. To determine whether location was a significant variable,
two separate one-way analyses of variance were performed. The
wall-bench-car-overhead location samples were compared to one another
for the shedded area and for the unshedded area. For the samples
taken under the shed, the geometric mean of the wall samples was
found to be significantly higher than the geometric means of the
bench, car, and overhead samples. In addition, the geometric mean
of the overhead samples was found to be significantly greater than.
that of the bench samples. For the samples taken in the unshedded
area, the geometric mean of the wall samples was found to be sig-
nificantly lower than the geometric means of the bench samples and
the car samples. In both areas the geometric mean dustfall rates
for the bench and. car samples were essentially the same.
Since location of the dustfall bucket appeared to be a signifi-
cant factor, a one-way analysis of variance was done for each of the
three locations common to both areas to determine whether the shed
was a significant factor. At two of the three locations — the
wall and the car — the geometric mean dustfall rates under the
-------
- 103 -
shed were found to be significantly higher than those samples taken
outside the shed. For the bench location, however, the geometric
mean dustfall rate under the shed was not statistically different
from that found at the corresponding unshedded location. It can
thus be concluded that both the presence of the shed and the loca-
tion of the dustfall container have a significant influence upon
measured dustfall rate in this study.
5.10 Impact of the Shed Upon Airborne Agents Within
A semi-enclosed shed adjacent to a coke-oven battery could
have a significant effect upon the quality of the work environment
within the shed. The shed enclosure tends to confine the coke-
oven emissions both during and between coke-pushing operations, and,
by restricting the dispersion and dilution that would occur by
direct discharge to the atmosphere, elevates the magnitude and
duration of concentrations of suspended dust and the myriad of
chemical substances present in the coking emissions. During non-
push conditions, however, it is possible that the steady flow of
ventilation air into the shed hooding might act to reduce
concentrations.
This evaluation of coke-side emissions, however, was intended
neither to document nor interpret the exposures of coke-oven
operators to coke-side emissions within the shed. Two studies
by the National Institute for Occupational Safety and Health (NIOSH),
however, did address this issue. '
5.11 Precision of Test Results
Although the terms "precision" and "accuracy" are often regarded
as synonymous, each has a specific technical meaning. The accuracy
-------
- 104 -
of a measurement signifies the closeness with which the measurement
approaches the true value. Precision, on the other hand, charac-
terizes the repeatability of the measurements. Thus, the precision
of a measurement denotes the closeness with which a given measure-
ment approaches the average of a. series of measurements taken
under similar conditions. Clearly, if the bias is large, a measure-
ment may be very precise but very inaccurate.
Many techniques exist to evaluate the precision of a result.
Ideally, simultaneous replicate samples are taken and the coeffi-
cient of variation, the standard deviation expressed as a percentage
of the mean, is used as a measure of precision. In this study,
a replicate sampling technique was used only for nine pairs of
dustfall samples. The precision of these paired samples is dis-
cussed in Section 5.9.
When the sample at hand is the only measure of the variability
of data at given conditions, a confidence interval can be used to
bracket the true mean of the population. This interval may be
regarded as a first estimate of the precision of the results. In
this study, such confidence intervals were constructed (using the
t-statistic and assuming normality) at the 95-percent level, imply-
ing a five-percent risk of not bracketing the true mean value of
a series of test measurements. This confidence interval is ex-
pressed in the Summary and Conclusions (Section 2.0) as m (+ r),
where m is the arithmetic mean and 2r is the confidence interval.
This technique was used in the evaluation of particulate emission
rates, shed capture efficiencies, composition of particulate mat-
ter, particle sizing data, and emission rates of gases. Although
the statistical precision is expressed as (+ r) with a confidence
interval of 2r, any confidence interval for the mean of a percentage
-------
- 105 -
is necessarily bounded by a maximum value of 100. Likewise, the
confidence interval for a concentration, emission rate, or emis-
sion factor is limited by a minimum value of zero.
This report prepared by:
Fred I. Cooper
Thomas A. Loch, Ph.D., P.E
Janet L. Vecchio
John E. Mutchler, P.E.
-------
- 106 -
REFERENCES
1. "Standards of Performance for New Stationary Sources," Federal
Register, 40CFR60, June 14, 1974.
2. "BMS-11 Sampler — Sampling Instructions," Monsanto Enviro-
Chem Systems, Inc.
3. "Operating Instructions for Andersen Stack Sampling Equipment,"
Andersen 2000, Inc.
4. TR-2 Air Pollution Measurement Committee, "Recommended Standard
Method for Continuing Dustfall Survey (APM-1, Revision 1),"
Journal of Air Pollution Control Association, 16:7, pp 372-
377, 1966.~
5. National Institute for Occupational Safety and Health,
Division of Technical Services, Industrial Hygiene Services
Branch, "An Industrial Hygiene Survey of the Bethlehem Steel
Corporation (Burns Harbor Facility) Coke Side Emission
Collecting Shed," Project No. 75-32, April 28, 1975.
6. National Institute for Occupational Safety and Health,
Division of Technical Services, Industrial Hygiene Services
Branch, "An Industrial Hygiene Survey of the Great Lakes
Carbon Corporation (Missouri Coke ft Chemical Division) Coke
Side Emission Collecting Shed," Project No. 75-31,
April 28, 1975.
-------
TECHNICAL REPORT DATA
(flcait read fnsZuctions on the revtrte before comt'letiiig)
I REPORT NO
EPA/l-77-014a
3. RECIPIENT'S ACCESSlOV.-.O.
4 TITLE A.\O SUBTITLE
Study of Coke-Side Coke-Oven Emissions
Volume I
5. PPPORTDATE
August 31. 1977
6. PERFORMING ORGANIZATION COO3~
7. AUTHOHIS)
John E. Mutchler, Thomas A. Loch,
Fred I. Cooper, Janet L. Vecchio
3. PERFORMING ORGANIZATION rtb?3ar :.
------- |