United States Off ice of Air Quality EMB Report No. 80-OSP-6
Environmental Protection Planning and Standards J ' yo
Agency Research Triangle Park NC 27711
SEPA
On-Shore Production of
Crude Oil and Natural Gas
Sulfur Plants
Emission Test Report
Shell Oil Company
Thomasville, Mississippi
-------
DCN 81-222-018-04-31 EMB Report No. 80-OSP-6
EMISSION TEST REPORT
S02 'TESTING
AT THE
SHELL OIL THOMASVILLE PLANT
THOMASVILLE, MISSISSIPPI
Prepared by:
Jay R. Hoover
RADIAN CORPORATION
8501 Mo-Pac Boulevard
Austin, Texas 78759
Prepared for:
Wlnton Kelly
U. S. Environmental Protection Agency
ESED/EMB (MD-13)
Research Triangle Park, North Carolina 27711
EPA Contract No. 68-02-3542
Work Assignment No. 4
July 1981
-------
CONTENTS
Section Page
1 INTRODUCTION 1
2 SUMMARY OF RESULTS 2
3 PROCESS DESCRIPTION 5
4 LOCATION OF SAMPLING POINTS 7
5 SAMPLING AND ANALYTICAL METHODOLOGY 10
6 QUALITY ASSURANCE/QUALITY CONTROL 13
7 COMPLETE RESULTS AND EXAMPLE CALCULATIONS 18
11
-------
SECTION 1
INTRODUCTION
This report presents the results of testing for sulfur dioxide,
reduced sulfur, and nitrogen oxides emissions from the Glaus incinerator stack
at the Shell Oil Thomasville Plant. The testing was performed by Radian
Corporation on May 8 through May 18, 1981. This work was funded and adminis-
tered by the Emission Measurement Branch of the U. S. Environmental Protection
Agency. The results of this testing will be used to develop New Source
Performance Standards for onshore production facilities.
The following sections present a summary of results, a description of
the process configuration, location of sampling points, the testing
methodology, quality assurance/quality control procedures, and complete re-
sults with example calculations. A full listing of the data and other sup-
plemental information are included as appendices.
-------
SECTION 2
SUMMARY OF RESULTS
This section presents a summary of the results of testing at the
Shell Oil Thomasville Plant. The complete test results and example calcula-
tions are presented in Section 7. All of the supporting data sheets are
included as Appendix A.
The results for the S02, H2S, and TRS tests are summarized in
Table 2-1. This table also presents the liquid sulfur production data, the
calculated sulfur emission rate (S02 plus TRS), and sulfur recovery effi-
ciencies. Figure 2-1 graphically presents on a daily basis the sulfur re-
covery efficiency, the SOZ emission rate, and the liquid sulfur production.
-------
U>
TABLE 2-1. SUMMARY OF RESULTS - SHELL THOMASVILLE PLANT
S02(ppm)
Date
5/12
5/13
5/14
5/15
5/18
TEST
PERIOD
Range
8,764-9,161
8,792-9,352
8,615-8,990
8,522-8,704
9,091-10,125
8,522-10,125
Average
8,907
9,136
8,779
8,616
9,671
9,022
H2S(ppm)
Range
13-30
20-47
43-48
16-23
23-33
13-48
Average
19
.36
46
20
28
30
TRS
(ppm)
186
314
330
245
291
273
S02a Out
Stack(lb/hr)
6,984
7,386
6,972
6,786
7,662
7,140
Sulfur(s)b Out
Stack(LTPD)
37.4
39.6
37.4
36.4
41.0
38.4
Liquid Sulfur
Make(LTPD)
1221
1098
1173
1170
1111
1155
Plant %
Efficiency
97.0
96.6
96.9
97.0
96.4
96.8
alncludes S02 plus TRS expressed as S02
Includes SOz plus TRS expressed as S
-------
1300 T
Liquid Sulfur 1200
Production
(LTPD)
1100
1000
8000
Average Stack
SOa Emissions
(lb/hr)
(Includes
and TRS) 6000
7000 ..
98 T
Sulfur Recovery
Efficiency (%) 97
96 ••
-t-
j ,_
12 13 14 15
May Test Date
18
Figure 2-1. Summary of SOz emissions and sulfur recovery
at Shell Oil's Thomasville Plant.
-------
SECTION 3
PROCESS DESCRIPTION
The Shell Oil Thomasville Plant is a gas processing facility engaged
in natural gas sweetening with no natural gas liquids removal. Feed gas to
the plant is newly produced from six local gas wells. The gas is sweetened
in a Shell Sulfinol unit. The Sulfinol unit has a design capacity of 100
MMSCFD of raw natural gas, and was treating about 87 MMSCFD of gas during the
test period. The Sulfinol unit creates an acid gas stream that is relatively
rich in H2S (about 83 volume percent) to be fed to a Glaus plant. The Glaus
plant has a capacity of 1275 long tons per day (LTPD) of liquid sulfur and
was operating between about 1100 and 1220 LTPD during the test period.
A simplified schematic of the process is shown in Figure 3-1. Acid
gases are removed from the raw natural gas by absorption into Sulfinol
solution. During the regeneration of this Sulfinol solution, acid gases are
released and are sent to a Glaus sulfur plant for recovery of elemental
sulfur. The Glaus plant is a three-stage catalytic unit. A slipstream of
acid gas bypasses the reaction furnace and is used to fuel the three in-line
burners. Liquid sulfur from the Glaus plant is collected in a below-ground
storage tank and sold. The acid gases remaining in the Glaus plant tail gas
are routed to an incinerator to convert the H2S to S02 prior to emission to
the atmosphere.
Testing of the Glaus plant incinerator stack was performed to
determine the level of sulfur emissions (502, H2S, and total reduced sulfur)
and NO in the stack gas. In addition, the liquid sulfur production was
monitored to allow estimation of the efficiency of the sulfur recovery plant.
-------
SWEET
GAS
B
S
0
R
B
E
R
X
<
y-
ACID
CAS
LIQUID
SULFUR
SULFUR
CONDENSERS
Figure 3-1. Simplified flow diagram.
-------
SECTION 4
LOCATION OF SAMPLING POINTS
Gas-phase samples were collected on the incinerator stack that services
the offgases from the Glaus unit. Sampling was performed on the 360° sampling
platform located approximately 115 feet off the ground. Four three-inch (3")
flanged ports were available for sampling. The location and orientation of
these ports are shown in Figure 4-1.
All four ports were used to perform velocity traverses due to the
relatively large stack diameter (9.5 ft). The 12-point velocity traverse
layout is also shown in Figure 4-1. The distances into the stack for point 1
is 5.0 inches, point 2, 16.6 inches, and point 3, 33.7 inches. The proposed
three-point traversing technique used to collect the various gas-phase samples
was eliminated in favor of a single point of average velocity. This decision
allowed the use of a shorter probe. Since the sampling ports are located over
10 stack diameters from the nearest upstream or downstream disturbance and the
velocity profile was relatively constant across each diagonal, the gas stream
should be homogeneous at the sampling points. Figure 4-2 shows the location
of the sampling points with respect to upstream and downstream stack
disturbances.
All sets of gas-phase samples except one were taken using the south
port. Most were taken at point S-2, which was 16.6 inches into the stack.
Some samples were taken at point S-3, which is 33.7 inches into the stack.
One set of gas samples was taken at point E-2, which was 16.6 inches into the
stack through the east port. The field data sheets in Appendix A indicate
which point was used for individual samples.
-------
w
Ladder
Figure 4-1. Location of sampling ports
and velocity traverse points.
-------
Total Height
^ 300 ft.
Incinerated
Tail Gas
ID = 9.5 ft.
Sampling Platform
3 Height ^ 115 ft.
Figure 4-2. Location of upstream and downstream
disturbances from sampling ports.
-------
SECTION 5
SAMPLING AND ANALYTICAL METHODOLOGY
To meet the objectives of this project, the following gas-phase
parameters were measured at the incinerator stack sampling platform:
volumetric gas flow rate,
molecular weight,
• moisture content,
• S02,
' N0x-
A
• H2S, and
total reduced sulfur (TRS).
In addition to the above parameters, the liquid sulfur production rate was
also monitored. Whenever possible, referenced source sampling and analysis
methods were used during testing at the Thomasville Plant. Table 5-1 lists
the various parameters measured and the sampling and analysis methods used
to monitor these parameters. A description of the sampling and analytical
methodology is provided in Appendix B of this report.
TABLE 5-1. SAMPLING/ANALYSIS PARAMETERS AND METHODOLOGY
Parameter to be Measured Methodology
Volumetric Gas Flow Rate EPA Method 2
Gas-Phase Molecular Weight EPA Method 3
Gas-Phase H20 EPA Method 4
Gas-Phase S02 EPA Method 6
Gas-Phase NOX EPA Method 7
Gas-Phase H2S EPA Method 11
Gas-Phase TRS EPA Method 16A
Liquid Sulfur Production No Reference Method
10
-------
The moisture content of the incinerator flue gas at the Thomasville
Plant was determined using the methodology specified in EPA - Method 4.
Attempts were made early in the sampling program at the Warren Petroleum -
Monument Plant to collect both S02 and moisture samples with the same
sampling train (EPA Method 6). By weighing the impingers before and after
sampling the mass of water collected during sampling could then be related
to the moisture content of the gas. But, because the gas volume collected
during S02 sampling was only 20 to 30 liters, the total mass of water col-
lected during sampling was 2.5 to 4.0 grams based on 15% H20 in the gas.
Small losses (M..O grams) in the recovery of the collected H20 could have a
large effect in the apparent moisture content of the gas. To alleviate this
problem, a separate sampling train (EPA Method 4) was set up by using larger
impingers at a higher gas flow rate to collect the H20 samples. This sampling
arrangement allowed a larger volume of gas (^150 liters) and a larger mass of
water (15 to 20 grams) to be collected. Small losses in the recovery of the
collected water did not have as significant effect on the moisture determination.
The proposed three-point traversing technique used to collect the
various gas-phase samples was eliminated during this testing period. Instead,
a single point of average velocity was used to collect a majority of the gas-phase
samples. The decision to eliminate the three-point traverses was based upon two
facts. First, the sampling ports are situated approximately ten stack diameters
upstream and downstream from the nearest disturbance. Second, the velocity
profile is relatively consistent across each diagonal. To help insure that the
gas sample was homogeneous by the time it reached the sampling ports, four sets
of gas samples were collected at a separate point within the stack. Analysis
of these gas samples for SOz, H2S, 02, C02, and N2 indicated that the concen-
tration of these constituents were the same (within experimental error) at both
points in the stack.
Because of the very low particulate concentration in the incinerator
gas, a decision was made to eliminate the glass wool plug from the probe liner.
This decision eliminated the systematic placement and removal of a glass wool
11
-------
plug from a 1120°F probe in-between SOz and HaS sample collection runs. The
glass plug is designed to remove particulate from the gas-phase during SOz
(EPA Method 6) and N0x (EPA Method 7) sample collection. However, during H2S
(EPA Method 11) and TRS (EPA Method 16A) sampling, the glass wool plug is
eliminated to minimize sorption losses of these gas species across a particu-
late cake. By eliminating the glass wool plug, the probe did not have to be
removed from the stack in-between each sample. This minimized the time that
the extremely hot (1120°F) probe had to be handled resulting in increased per-
sonnel safety with a minimum of down time in-between runs.
12
-------
SECTION 6
QUALITY ASSURANCE/QUALITY CONTROL
A comprehensive quality assurance/quality control (QA/QC) program
(Radian DCN 81-222-018-04-09) was designed and implemented during this pro-
gram. The objective of this QA/QC program was to assess and document the
precision, accuracy, and adequacy of emission data developed during sampling
and analysis.
A summary of the QA/QC results obtained during activities at the
Thomasville Plant are presented in this section. A brief discussion of the
precision, accuracy, and data capture are also presented in this section of
the report. Copies of the equipment calibration forms and reagent preparation/
standardizarion forms are presented in Appendix A.
ASSESSMENT OF DATA QUALITY
Table 6-1 summarizes the estimated and measured precision, accuracy,
and data capture for each of the parameters monitored at the Thomasville Plant.
The measured precision and accuracy for each of the parameters fall within
the original estimates. Deviations from estimated data capture are discussed
later in this section.
DISCUSSION OF QA/QC RESULTS
During this project, the precision and accuracy of a particular measure-
ment was determined by one or more of the following methods:
13
-------
TABLE 6-1. SUMMARY OF ESTIMATED AND MEASURED PRECISION, ACCURACY, AND DATA CAPTURE
FOR THE DATA COLLECTED AT THE SHELL OIL THOMASVILLE PLANT
Measurement Parameter
(Method)
Volumetric Gas Flow
Rate (EPA 1 and 2)
Molecular Weight
(Orsat)
H20
(EPA 4)
S02
(EPA 6)
NOX
(EPA 7)
H2S
(EPA 11)
Total Reduced Sulfur
(EPA 16A)
Liquid Sulfur
Production
Estimated
(RSD)
20%
10%
11%
10%
10%
10%
51%
10%
Precision
Measured . i
TIonlinMli rv I7Pnp-ir jl>1 H t u Accuiacy DataCd]
(RSD) (RSD) Estimated2 Measured Estimated
20%" 20%" ± 11% ± 11% " 90%
<2% <2% ± 25% <± 10% 90%
11%" 11% * + 10% ± 10% "" 90%
0.14% 1.20% ± 20% <±0.5% 90%
— 6 — 6 ± 20% <±2.5% 90%
N/A7 N/A7 ± 20% N/A 7 90%
0.14%5 1.20% ± 15% <±0.5% 90%
± 5% — 100%
Dture
Measured3
100%
100%
100%
100%
100%
100%
100%
100%
'Accuracy is based upon QA/QC Field Audit Performed at Monument Plant.
2Expected range for bias of method.
3The valid data percentage of the total tests required in the scope of work.
''The Monument Plant field performance audit showed no deviation from accepted procedure. Precision and accuracy
should be within the estimated values.
5TRS analyses use same analytical method (BaCl2 titration, Thorin indicator) as S02 analyses, as
well as comparable Citrant volumes. Since the whole sample for TRS had to be analyzed, precluding
duplicates, no independent replicability data exists.
6NO precision cannot be properly calculated when all values are below the limit of quantitation (LOQ).
7Refer to the text.
RSD = Relative Standard Deviation
RSD = Mean Relative Standard Deviation
-------
• performance audit,
• system audit, and
quality control procedures.
Precision is defined here as a measure of mutual agreement among individ-
ual measurements of the same property. Precision can be qualified with respect
to the replicability and repeatability of a particular parameter. Replicability
is a measure of variability between measurements of the same parameter by the
same analyst using the same apparatus on the same day and in the same labora-
tory. Routine duplicate analyses were used to measure replicability during
the course of the project.
Repeatability is similar to replicability but requires that one or more
of the following be different:
• analyst,
• apparatus, or
the day.
Daily analysis of quality control standards by different analysts provided
a measure of repeatability.
Accuracy is defined here as the degree of agreement of a measurement (or
average of measurements of the same sample) with an accepted reference of true
value. The accuracy data presented in Table 4-1 represents the relative
accuracy of the measured value, X, with respect to the reference value, T,
of a field audit sample. Results obtained during the field performance audit
at the Monument Plant were also used to determine the accuracy of the data
collected at the Thomasville Plant.
15
-------
All definitions and procedures used in calculating precision and accuracy
were taken from Appendices A and C of the EPA document 600/9-76-005, Quality
Assurance Handbook for Air Pollution Measurement Systems, Volume I, Principles
(4-1).
Data capture can be calculated by several different techniques. The data
capture reported in Table 6-1 represents the valid data percentage of the total
tests required in the scope of work.
The following list summarizes the deviations, exceptions, and special
cases with respect to the precision, accuracy, and data capture data presented
in Table 6-1. These include:
The titrations of the peroxide impinger in the H2S trains are
included in the data base used in calculating precision for
502 analyses.
The precision and accuracy reported for the TRS analyses is
based on data from the S02 analyses since the analytical
procedure (BaCl2 titration, thorin indicator) is the same
and comparable titrant volumes were used.
Because of the lack of a suitably stable sulfide standard,
no sulfide QC standard was analyzed. Precision data for
H2S analysis is not presented because duplicate analysis
were not performed. The referenced analytical procedure
requires that the whole sample be titrated, precluding
duplicate analysis.
Accuracy data for H2S analysis are not presented. A "certified
standard" bottle of H2S gas is on order and will be used to
determine the accuracy of the methodology.
NO precision cannot be properly calculated when all values
except one are below the "Limit of Quantitation" (LOQ).
The reported accuracy data for molecular weight, S02, N0x, and
TRS reflect only the analytical phase of the measurement, as
discussed in Sections 3.2.8, 3.5.8, and 3.6.8 of EPA document
600/4-77-027b(4-2). The performance audit activities address
the sampling procedures.
16
-------
Both the precision and accuracy of the flue gas flow rate and
moisture determination are based upon the performance audit
performed at the Monument Plant. Since the field performance
audit showed no deviation from accepted procedure, both the
precision and accuracy are expected to lie within the estimated
values. Further information concerning the field systems audit
will be discussed in the separate QA/QC report.
17
-------
SECTION 7
COMPLETE RESULTS AND EXAMPLE CALCULATIONS
This section presents the complete results and example calculations
for testing performed at the Shell Oil Thomasville Plant. All of the support-
ing data sheets are included as Appendix A of this report.
The results for the velocity, gas composition, and SOz tests are
shown in Table 7-1. This table also presents the calculated flow rates and
S02 emission rates. Table 7-2 presents the test results for HzS, TRS, and
NO along with the calculated emission rates.
x
This section also presents example calculations which show how the
test results were used to obtain flow rates, emission rates, and sulfur
plant efficiencies.
18
-------
TABLE 7-1. COMPLETE RESULTS: S02, FLOW - SHELL THOMASVILLE
DATE
5/12
5/13
RUN
1
2
3
AVG
1
2
TIME3
0845
1402
1700
0845
1550
3U0843,5/14
5/14
5/15
5/18
AVG
1
2
3
AVG
1
2
3
AVG
1
2
3
AVG
1203
1545
1840
1255
1530
1935
0825
1010
1300
02
2.8
2.6
2.5
2.6
2.7
3.9
3.0
3.2
2.6
2.4
2.5
2.5
2.5
2.6
2.4
2.5
1.9
2.1
1.9
2.0
C02
7.9
8.3
8.2
8.1
8.2
7.4
8.0
7.9
8.2
8.5
8.4
8.4
8.4
8.3
8.3
8.3
8.6
8.5
8.6
8.6
N2
88.4
88.7
88.6
88.6
88.6
88.4
88.1
88.4
88.6
88.5
88.3
88.5
88.3
88.6
88.8
88.6
88.5
88.8
88.9
88.7
STACK
TEMP(°F)
1130
1115
1109
1118
947
1119
1119
1062
1118
1121
1113
1117
1138
1136
1136
1137
1136
1138
1139
1138
STACK
PRESS ("Hg)
29.94
29.94
29.88
29.92
29.51
29.78
29.80
29.56
29.45
29.45
29.45
29.45
29.54
29.54
29.46
29.51
29.79
29.79
29.79
•29.79
VELOCITY
%H20
28.06
28.44
28.67
28.39
27.98
26.46
29.00
27.81
28.78
28.13
29.05
28.65
26.75
28.17
28.91
27.94
29.46
29.33
28.88
29.22
(FT/S)
76.6
74.5
75.5
75.5
71.5
74.2
76.5
74.4
77.2
76.5
75.6
76.4
75.8
77.1
76.7
76.5
78.6
78.0
76.9
77.8
FLOW
(ACFM)
326000
317000
321000
321000
304000
316000
326000
315000
329000
326000
322000
325000
323000
328000
326000
326000
334000
332000
327000
331000
FLOW
(DSCFM)C
76800
75000
75600
75800
79800
76200
75600
77200
76200
76200
74400
75600
76200
75600
74400
75400
76200
76200
75600
76000
S02
(ppm.dry)
8764
8797
9161
8907
9263
9352
8792
9136
8990
8732
8615
8779
8704
8623
8522
8616
9091
9798
10125
9671
S02
(Ib/hr)
6840
6660
7020
6840
7500
7200
6720
7140
6960
6720
6480
6720
6720
6600
6420
6580
7020
7560
7740
7440
FOOTNOTES :
a - Time reported is when velocity profile was begun.
b - This run was actually performed on the morning of 5/14.
c - DSCFM at 60°F and 14.7 psia.
-------
TABLE 7-2 COMPLETE RESULTS: REDUCED SULFUR, NOX- SHELL THOMASVILLE
l-o
O
DATE
5/12
5/13
5/14
5/15
5/18
RUN TIME
1 0845
2 1402
3 1700
AVG
1 0845
2 1550
3 0843,5/14
AVG
1 1203
2 1545
3 1840
AVG
1 1255
2 1530
3 1935
AVG
1 0825
2 1010
3 1300
AVG
FLOW
(DSCFM)
76800
75000
75600
75800
79800
76200
75600
77200
76200
76200
74400
75600
76200
75600
74400
75400
76200
76200
75600
76000
CONC,
13
14
30
19
20
40
47
36
48
43
47
46
16
22
23
20
24
26
33
28
H2S
ppm LB/HR
10.2
10.8
22.8
14.4
16.2
30.6
36.0
28.2
37.2
33.0
35.4
35.4
12.6
16.8
17.4
15.6
18.6
19.8
25.2
21.6
TRSb
CONC, ppm
186
186
314
314
330
330
245
245
291
291
LB/HR
144
144
246
246
252
252
186
186
222
222
N0x°
CONC, ppm LB/HR
3.4
<3.0d
6.2
<4.2C <2.3
<3.0
<3.0
<3.0c
<3.0
<3.0
3.7
<3.2e <1.7
3.3
4.7
<3.0
<3.7e <2.0
<3.0
<3.0
5.4
<3.8e <2.1
FOOTNOTES :
a -
b -
c -
d -
e -
Time reported is when velocity profile
LB/HR reported as J
NOX LB/HR reported
Detection limit is
NOV samples taken c
302 equivalent.
as N02.
3ppm, averaged
luring run 3.
in as
was begun.
3ppm.
-------
EXAMPLE CALCULATIONS
FLOW RATES
•Actual Cubic Feet per Minute (ACFM)
ACFM = Velocity x Stack Cross - Sectional Area
Example: Based on 5/12 averages -
2
ACFM =75.5 ft x 9.5 (ft) x n x 60 sec
sec 4 min
ACFM = 321,000 ft3/min
•Dry standard cubic feet per minute (DSCFM) @ 60 °F and 29.92 in.Hg
_ AfFM Barometric Pressure Standard Temp /,_ mole
Standard Pressure Stack Temp. I fraction
Example: Based on 5/12 averages-
H20
DSCFM = 321,000(ACFM) x ' *} ,*( x "o \°^ x (1-0.2839)
/y.y/(in.Hg;
DSCFM = 75,800 ft3/min
EMISSION RATES
•Emission Rates- S02, NOX, H2S, TRS (.lb/hr)
„ . „ Concentration of Compound (ppm,dry) TM-^™*
Emission Rate = —r-g —c u-c— •*- x DSCFM x
compound mole wt
molar volume
Example: Based on 5/12 averages-
S02 Emission Rate = I907 *> x 75,800 DSCFM x
106 379 SCF
60 min
hour
Emission Rate = 6840 lb/hr
Note: H2S and TRS Emission Rates are expressed as equivalent S02;
NOX Emission Rate is expressed as N02;
Total S02 Emission Rate is the sum of S02 and TRS Emission Rates
21
-------
SULFUR PLANT EFFICIENCY
•Plant Efficiency = Sulfur Recovered x 100%
Sulfur Recovered + Emitted
where:
Sulfur recovered = liquid sulfur production (LTPD)
Sulfur emitted = S02 + TRS Emission Rates (expressed as
elemental sulfur) (LTPD)
Example: based on 5/12 averages-
Plant Efficiency =
1221 LTPD x 100%
m f(6840 + 144) Ib S02 32 Ibs __LT 24 hr
1221 (LTPD) + x x x
Plant Efficiency = 97.0%
22
------- |