EPA/600/A-96/1
92-142.05
National Estimate of Methane Emissions from Compressors in the U.S. Natural
Gas Industry
Donna Lee Jones, Lisa M, Campbell, Clint E. Burklin, and Mahesh Gundappa
Radian Corporation, Research Triangle Park, North Carolina 27709
Robert A. Lott
Gas Research Institute, Chicago, Illinois 60631
INTRODUCTION
The combustion of natural gas emits much less carbon dioxide per unit of
energy generated than other fossil fuels. For this reason, one strategy that
has been suggested for reducing global warming is to encourage switching from
other fossil fuels to natural gas. However, methane is currently thought to
be a more potent greenhouse gas than carbon dioxide; if so, leakage of natural
gas (which is approximately 90 percent methane) could reduce or eliminate the
advantage of using natural gas because of its lower carbon dioxide emissions.
Two major issues must be addressed before the consequences of the fuel
switching strategy can be evaluated. First, there is a need to better define
the impact of methane relative to carbon dioxide on global warming. Second,
it is important to better define methane emissions from the gas industry.
Because of the latter issue, the Gas Research Institute (GRI) and the U.S.
Environmental Protection Agency (EPA) have developed a cooperative program to
quantify methane emissions from U.S. gas operations. Currently, estimates of
methane emissions from the gas industry range from 0.5 to 4.0 percent of
production. The GRI/EPA program is a comprehensive program to determine
methane emissions from the wellhead to the customer's meter. The goal is to
determine emissions to within approximately 3 billion cubic meters (m3), or
100 billion standard cubic feet (scf), which is approximately 0.5 percent of
U.S. production. To achieve this overall accuracy, an accuracy target has
been established for each source category in the natural gas industry.
One source category is the exhaust from compressor engines that are used
to move the gas through the system. Compressors are located in production
fields, processing plants, gas storage facilities, and along transmission
lines. A preliminary study indicated that the exhaust from compressor engines
might account for more than 50 percent of the industry's methane emissions.
Because the uncertainty in this early estimate was quite large, Radian
Corporation conducted a study to determine methane emissions from both
reciprocating and turbine compressor engines. The accuracy target established
by the GRI/EPA program was to determine the emissions for this source category
within 30 billion scf.
In this study, methane emissions were estimated by multiplying the
methane emission rate of the unit by the activity of the unit. The emission
rate expresses the amount of emissions per operating characteristic,
independent of the other features of the unit. The activity expresses the

-------
92-142.05
operating characteristic of the unit which was, in this case, either a
reciprocating engine or gas turbine. For example, when the operating
characteristic is fuel usage, the activity is annual fuel used by compressors.
When the operating characteristic is operating time, the activity is hours of
operation (per year).
The emission rates used in this study were in the form of methane
emissions per hour. The activity of each unit was the annual operating hours
for that unit. A national estimate of annual methane emissions was determined
from emissions estimates using available data for compressors in the natural
gas industry.
APPROACHES
The Ideal Emission Estimate
Ideally, the annual methane emissions from compressors in the natural
gas industry would be determined from the sum of the annual emissions from
each unit in the industry, where the emission rate may vary over time and from
unit to unit. This relationship is expressed by:
where ER£(t) is the instantaneous emission rate at time (t) for compressor^
Ai(t) is the instantaneous activity at time (t) for compressor^ and N is the
total number of compressor units in the industry. To reflect annual
emissions, the integral in Equation (1) is evaluated over the time interval tj
to t2, where t2 - tx is equal to 1 year. This approach produces no
uncertainty in annual emissions.
Emission Estimate Using Test Data
The ideal approach was considered impractical because it would require a
massive data-gathering effort by the industry. Fortunately, a substantial
amount of data was available for both the emission rates and activity factors
of compressor engines. For the emission rates, data for reciprocating engines
and gas turbines were available for a number of models that were tested during
relatively short time periods. Values for the annual activity of compressors,
in terms of operating hours, were also available for some models. If data for
all compressors in the industry were available, the following equation could
be used to estimate methane emissions:
N
i-r
(i)
N
Emissions = 2 [ER. |
i-1
"i I At°test
X MC> Ut-lyear
(2)
2

-------
92-142.05
where ERt is the emission rate of compressor (in grams of methane per hour)
evaluated over the time period of the test; where the activity, Ai( of
compressor corresponds to the annual operating hours of compressor; and
where N is the total number of compressor units in the industry. This
equation assumes that the emission rate of the compressors during the test
period represents the emission rate of the compressors during the year, on the
average.
Methane emissions from compressors will vary from manufacturer to
manufacturer and model to model. Assuming that the compressors are properly
maintained, differences in methane emissions, even for the same model, also
will be caused by operating the compressors at different horsepower levels or
speeds, or for longer or shorter periods of time in order to satisfy the
operational needs of the system.
Figure 1 illustrates the relationship between operating load and
emissions for a Cooper Bessemer two-cycle engine.1 The figure shows the
total hydrocarbon emission rate at varying loads, where speed is used to lower
horsepower. Total hydrocarbon emissions can be considered a surrogate for
methane emissions, since methane is expected to comprise a large portion (over
90 percent) of total hydrocarbon emissions from compressors. The data show
that the emission rate for a Cooper Bessemer engine increased at lower loads,
with a 40 percent increase in emissions at 50 percent of rated horsepower
(full load), and a possible 100 percent increase if loads close to 30 percent
of full load are used. A similar increase in methane emissions can be
expected for other engine and turbine models, although the magnitude of the
increase for other models is not known at this time.
The emission rate of any compressor, then, is a function of the rated
horsepower and operating horsepower. The equation below summarizes this
relationship:
ERi - f (ERi, Hpi", Hpj)	(3>
where ERt is the operating emission rate of compressor; ER^0 is the emission
rate of compressor at the rated horsepower, Hp^; and Hpi is the operating
horsepower. The data from the Cooper Bessemer illustrate that a linear
relationship can be used to represent emissions for this engine as a function
of horsepower, within 3 percent accuracy, over the range of horsepower from
50 to 100 percent of rated horsepower. In this situation, the average
horsepower at which the engine operates over the year [as in Equation (2)]
could be used to evaluate the integral in Equation (1) without any loss of
accuracy.
If the information needed to develop Equation (3) was known for all
engines and the relationship between horsepower and emissions for each engine
was linear, the following equation could be used to more accurately estimate
annual methane emissions from compressor engines from test data:
3

-------
12
11
10
9
8
7
I 6
f» 5
4
3
2























\










"S












She




























A
































ft t »
20	30	40	SO	60	70	80	90	100 110 120 130
Parctri Rated Horsepower
-+- 196 RPM	231RPM	2S4RPM -Q 297HPM H> 330HPM
Figure 1. Effect of horsepower and speed on total hydrocarbon emissions for a
Cooper Bessemer 2 cycle engine1
to
s>
to
o

-------
92-142.05
Annual Emissions
(grams)
N
= £ [ ER (Hp). x Annual Hours (Hp)i ]
i = l
(4)
where ER (Hp)t is the emission rate of compressor as a function of horsepower
(in grams of methane per hour); Annual Hours (Hp)x are the annual operating
hours of compressor at a specified horsepower; and N is the total number of
compressor units in the industry.
Scaled-up Emission Estimate Using Test Data
Although emission and activity data for all compressors in the industry
were not known, an estimate of methane emissions for the industry was made by
proportioning the emissions for a subset of the industry (with known emission
and activity data) according to the proportion of horsepower contributed by
this subset. This approach assumes that the emission rates for the test
engines during the test period and the operating hours estimated for the
subset represent the average operation of compressors in the industry. A more
detailed description of this approach is found in the Method section of this
paper.
It was found that for the emission test data, on the average, the
compressor emission rates were measured at virtually the same horsepower as
the operating horsepower that was recorded for the compressors with the
activity data (where the average horsepower of the compressors for the test
data was equal to 0.985 times the average of the horsepower for the activity
data). Therefore, knowledge of the relationship between horsepower and
emissions was not needed for the analysis presented here.
An estimate of methane emissions using test data for a subset of the
industry may underestimate actual industry emissions if many compressors are
operated at lower loads for a significant period of time, since methane
emissions are likely to increase when compressors are operated at less than
full load. Future work in this area should include an analysis of the
relationship between horsepower and methane emissions, so that individual
compressor emission rates can be estimated with more accuracy.
DATABASES
Description of the Databases
A number of databases were obtained for this study, that provided
information on compressor emissions and operations. The databases were part
of GRI's TRANSDAT compressor module.2 Two databases in the compressor module
contain information about the distribution of compressors in the natural gas
industry. One of the databases is an almost complete listing of engines and
turbines in the gas industry, accounting for 16,2 million out of the
16.7 million total horsepower that was reported by the American Gas
Association (AGA) for this industry in 1989.3 This "Industry Database"
5

-------
92-142.05
contains	information concerning 8,282 compressors in the natural gas industry.
However,	individual compressor horsepower is not reported in the database;
only the	total horsepower for all compressors of a specific type is reported
for each gas company.
A second smaller database in the compressor module (the "Operating
Database") is a subset of the Industry Database (with 1,515 units)
corresponding to 3.2 million total horsepower. The Operating Database is the
only database in TRANSDAT that contains information on annual operating hours
for each unit. The operating horsepower of each compressor is also recorded
in the database. The data in the Operating Database were obtained from an AGA
survey of 112 companies in ozone nonattainment areas, where nitrogen oxides
(N0X) emissions were minimized. Because emissions of N0X and hydrocarbons are
inversely related, the operation of compressors to minimize N0X will likely
increase emissions of hydrocarbons and, hence, of methane. Consequently, the
data in the Operating Database, such as the operating horsepower, may
represent maximum methane emission conditions. The use of these data, then,
presents a conservative estimate of industry emissions, if the data are
representative of the industry as a whole.
Data from emissions tests performed by Southwest Research Institute are
contained in a third database in TRANSDAT (the "Test Database"). During the
emission tests, the compressors were operated at close to full load (rated
horsepower). Methane emissions, fuel use, fuel use rate, and horsepower were
recorded for each emissions test for 241 models of engines and turbines.
Since there was some variation in horsepower in the multiple emissions tests
for the same compressor, it may be possible to define the relationship between
horsepower and methane emission rate for the 241 models of compressors in this
database. With this information, it would be possible to extrapolate from the
test conditions to actual operating conditions at lower horsepower levels.
Because of the limitations of this study, an analysis of this type was not
performed.
Because the Test Database contains only emission data and the other
databases contain only operating data, a fourth database was developed that
contains data for compressors for which information was found in both the
Operating Database and Test Database. This fourth database was called the
"Emissions Database." A total of 775 reciprocating engines and 86 gas
turbines (out of 1,515 units in the Operating Database) fit the criteria and
were included.
Table 1 describes the contents of each of the four databases.
Model-Matching Hierarchy
A model-matching schema was designed to maximize the amount of
correlation between the Test Database data and the Industry and Operating
Databases. Originally, when the compressors were matched according to exact
model names, the Test Database accounted for only 38 percent of the units
6

-------
TABLE 1. DESCRIPTION OF THE DATABASES
Industry Database
Operating Database
Test
Database

Emissions Database
Data
Engines
Turbines
Total
Engines
Turbines
Total
Engines
Turbines
Total
Engines
Turbines
Total
Total Units
7,489
793
8,282
1,385
130
1,515
NA
NA
NA
775
86
861
Manuf/Model"
922
144
1,066
318
31
349
229
12
241
120
7
127
Total hp (MM)
11.2
5.0
16.2
2.4
0.8
CM
NA
NA
NA
1.3
0.4
1.7
Total Hours (MM)
NAb
NA
NA
4.6
0.3
4.9
NA
NA
NA
0.2
0.2
3.1
"Number of unique compressor models in the database.
bNot applicable.

-------
92-142.05
in the Operating Database and only 7 percent of	the Industry Database.
Through consultations with experts in the field	of compressor emissions, a
three-step hierarchy was designed to match more	data in the Industry and
Operating Databases to the Test Database.
The first step, therefore, in the hierarchy was based on an exact match.
For instance, an Ajax DPC-360 engine in the Industry or Operating Database was
identified with data from the Test Database for an Ajax DPC-360 in the first
step of the hierarchy.
The second step in the matching procedure matched compressors by
substring name, where the horsepower of both compressors was within
± 20 percent. An example of the second step in the hierarchy was a match of a
Clark BA8T engine with a Clark HBA8T engine (common substring of BA8T), where
the horsepower for the Clark HBA8T engine in the Emissions Database was 1,911,
and was 2,050 for the Clark BA8T engine in the Industry Database. The third
step in the hierarchy consisted of matching the rated horsepower per cylinder
(± 20 percent) in compressors with similar initial names (manufacturer) but
with varying substrings. An example of the third step is a match of a
Clark BA5 engine, with 248 horsepower per cylinder (total horsepower - 1,242)
and a Clark BA8 engine, with 211 horsepower per cylinder (total horsepower =
1,000).
Following the execution of the three steps in the model-matching
hierarchy, 37 percent of the models and 57 percent of the units in the
Operating Database (and 23 and 55 percent, respectively, in the Industry
Database) were matched to emissions data. The execution of the three-step
hierarchy also increased the amount of data in the Emissions Database to a
total of 1.7 million horsepower, from the previous total of 1.3 million
horsepower after the first step only, and accounting for over half of the
horsepower in the Operating Database and Industry Database.
METHOD TO ESTIMATE METHANE EMISSIONS
The Emissions Database was used to estimate methane emissions from
compressors in the natural gas industry. The method used to calculate
emissions is described below.
Equations
An average emission rate was obtained for each model of compressor
engine and turbine in the Emissions Database from the average of all methane
emission tests in the Test Database for that model. Since the time period in
which each test was conducted was not given, the emission rates, in units of
grams of methane per hour, were calculated from the reported methane emissions
per unit of fuel (in grams per m3) and the reported fuel use rate (in m3 per
hour) for each compressor:
8

-------
92-142.05
Emission Rate. Average Emission Rate. FUR.
=	,	x „	\->)
(g/hr)	(g/m3)	(m /hr)
where FURi was the average fuel use rate for compressor^
The total annual methane emissions for each compressor were then calculated as
in Equation (6):
Emissions.
1 = ER. x Annual Operating Hours.
(grams)
where ERj^ was the methane emission rate of compressor calculated with
Equation (5) , and Annual Operating Hoursj^ were obtained from the data for
compressor in the Emissions Database.
The total methane emissions for the compressors in the Emissions
Database were determined using:
Annual Emissions ^	n	.	/7\
- £ [ ER. x Annual Operating Hours. I	\>)
(grams)	i=1 1	v	iJ
where the emissions from each unit were calculated as in Equation (6), and N
was the total number of compressors in the database.
Estimate of National Emissions
If data for all compressors in the industry were available, a national
estimate could be calculated using Equation (7). Since the compressors in the
Emissions Database were only a subset of the compressors in the natural gas
industry, a procedure was necessary to relate the methane emissions calculated
using Equation (7) to a national estimate.
The ratio of the total horsepower from compressors in the industry
(16.7 million horsepower) to the total horsepower of the compressors in the
Emissions Database (1.7 million horsepower) was used to scale up the methane
emissions calculated in Equation (7) by a factor of (16.7/1.7), or 9.8, to
estimate national emissions. This relationship is shown by:
National Methane Methane Emissions for
Emissions	Emissions Database
Scaling Factor
Hpi	(8)
Hp
ED
where Hpj is the total horsepower in the industry3 and HpED is the total
horsepower in the Emissions Database, which produce a scaling factor of 9.8.
9

-------
92-142.05
RESULTS
National Emission Estimate
An estimate of national methane emissions from compressors using the
approach discussed above was 0.22 teragrams (Tg) of methane. Over 99 percent
of the emissions were estimated to be from reciprocating engines and less than
1 percent from gas turbines. This estimate was based on th§ assumption that
the compressors in the Emission Database represent compressors used in the
natural gas industry, on the average. If the industry compressors are
operated at horsepower levels much less than the rated horsepower, the methane
emissions estimated here could underestimate the industry emissions, since
methane emissions are thought to vary inversely with horsepower."
Field, Plant, and Pipeline Compressor Emissions
Compressors are used in field and plant operations as well as in
transmission activities. The emission estimate above does not apportion
methane emissions among the segments of the natural gas industry that use
compressors. The U.S. Department of Energy (DOE) provides estimates of the
amount of fuel used in field (lease), plant, and pipeline applications, and
these estimates were used to apportion methane emission estimates among the
sources.4 The methane emission estimates were based on the assumption that
all the fuel reported for field and plant purposes was used by compressors.
This assumption is likely to be an overestimate, because fuel is known to be
used for field and plant purposes by equipment other than compressors.
Although the portion of lease and plant fuel used for other purposes was not
known, estimates of the fuel used by these sources served as rough estimates
of fuel used by compressors in field, plant, and pipeline operations.
The result of using DOE estimates of fuel use was that 39 percent of
total compressor fuel was attributed to field compressors, 26 percent to plant
compressors, and 35 percent to pipeline. It is likely, however, that a higher
percentage of total compressor fuel is used for pipeline, and that lower
portions are used for field and plant. The AGA estimated that 84 percent of
total compressor horsepower was used for the pipeline3; if fuel use can be
assumed to be proportional to horsepower, the percent of fuel used by pipeline
compressors could be over twice as high as the estimate based on the DOE
information.
Since methane emissions were assumed to be proportional to fuel use, the
DOE breakdown in fuel between the three sources of compressor emissions in the
natural gas industry was used to apportion the national estimate of 0.22 Tg of
methane emissions between the three sources. The results were that, of the
total 220 megagrams (Mg) of annual methane emissions estimated for compressors
in the industry, 86 Mg of methane was estimated to be emitted from field
compressors, 57 Mg from plant compressors, and 77 Mg from pipeline
compressors. As discussed above for apportioning fuel use between these three
sources, the emissions from pipeline compressors were probably underestimated,
10

-------
92-142.05
and the portions for the other two sources were overestimated. Another
possible error in the breakdown of emissions was the assumption that the
emission rate and operating characteristics of compressors in field, plant,
and pipeline applications were similar.
UNCERTAINTY IN EMISSION ESTIMATES
Since the true value of methane emissions from compressors in the
natural gas industry is not known, the accuracy of the estimate of 0.22 Tg of
methane emissions cannot be assessed. However, the data can be used to
produce an approximation of the uncertainty in the emission estimate. Like
confidence intervals, the uncertainty estimated for a value can describe, with
a high degree of confidence, the range in which the true value lies.
The target uncertainty for the total estimate of methane emissions from
all sources in the natural gas industry was close to 3 billion m3 (100 billion
scf), Because many segments in the natural gas industry emit methane that are
part of the overall methane emission estimate in the GRI/EPA project, the goal
for each segment was to produce a methane estimate with as little uncertainty
as possible. The target uncertainty for compressors was approximately 800
million m3 (30 billion scf).
The following explains the procedure used to estimate the uncertainty in
the annual methane emission estimate from compressors and the results of this
estimate.
Basis of Uncertainty in the Compressor Emission Estimate
The only sources of uncertainty in the approach described above that
could be estimated were the analytical error associated with the test data
used to calculate the methane emission rates and the scale-up to national
emissions based on the ratio of the horsepower in the Emissions Database to
the estimate of compressor horsepower in the industry.
The uncertainty in the hydrocarbon analysis was estimated to be
± 10 percent, based on the expected gas chromatograph (with a flame ionization
detector) capabilities. Likewise, the uncertainty associated with fuel flow
measurements was estimated to be ± 2,5 percent. The uncertainty associated
with the scale-up between the Emissions Database and total industry horsepower
was more difficult to quantify. Two indirect assessments of this uncertainty
are discussed below, using comparisons between the Emissions Database and the
Industry Database, which was taken to be a fairly good representation of the
industry as a whole. The uncertainty for the scale-up to nationwide emissions
was estimated to be approximately 1 percent, based on estimated significant
figures in total horsepower. For an estimated industry horsepower of
16.7 million, the uncertainty was estimated to be 0.17 million horsepower; for
the Emissions Database, with 1.7 million horsepower, the uncertainty was
estimated to be 0.017 million horsepower.
11

-------
92-142.05
Procedure to Calculate Uncertainty
The software "@Risk"5 was used to estimate the uncertainty in the
estimate of methane emissions from compressors. The variables of uncertainty
were entered into the program along with the values discussed above. The
"@Risk" uncertainty estimating procedure recalculated the total methane
emission estimate using Latin Hypercube sampling at normally distributed
intervals along the range of each variable, defined by the expected value and
the (estimated) standard deviation of that value. In the case where the
standard deviation was estimated from significant figures, a uniform interval
(as opposed to a normally distributed interval) was used for the Latin
Hypercube sampling.
In the "@Risk" program, the recalculation was performed 500 times, each
time using a different value in the interval of each variable in the emission
equations. The results of the 500 calculations, excluding values below zero,
were analyzed for variation about the mean. The uncertainty was reported as
the coefficient of variation, or the standard deviation divided by the mean,
as a percent.
Estimates of Uncertainty
The uncertainty for the estimate of methane emissions from compressor
engines and turbines, at 0.22 Tg (11.5 billion scf), was estimated to be
approximately 1 percent, or 4.3 million m3 (0.15 billion scf). This value is
well below the 800 million m3 target for compressor engines and turbines.
Comparison of the Database
One component of uncertainty in the methane emission estimate that could
not be quantified for the uncertainty analysis was the degree to which the
compressors in the Emissions Database represented the compressors in the
industry. The Emissions Database was compared with the Industry Database in
two procedures that were designed to assess the representativeness of the
Emissions Database for compressors in the industry and, indirectly, the
uncertainty associated with the scaling factor.
Figures 2 and 3 are plots for engines and turbines, respectively,
showing the total horsepower for each model in the Industry Database versus
the total horsepower of that model in the Emissions Database. The line formed
from the relationship between the total horsepower for each database was
plotted on each graph. For engines, a line through 11.2 million and
1.3 million, the total horsepower for engines in the Industry and Emissions
Databases, respectively, constituted the ideal relationship between the total
horsepower for each type of engine in the two databases. Similarly, for
turbines, a line through 5 million and 0.4 million produced the ideal
relationship for the total horsepower for each type of turbine in the two
databases. The figures show that the horsepower contribution of the
compressors in the Emissions Database was higher than the horsepower
12

-------
90
80 -
70 -
(Thousands)
Horsepower Contribution In the Industry Database
K>
Figure 2. Relative horsepower contribution of each engine model in the
Emissions and Industry Databases

-------
220 j—
210 -
200 -
190 -
180 -
170 -
160 -
150 -
140 -
130 -
120 -
110 -
100 -
90 -
80 -
70 -
60 -
50 -
40 -
30 -
20 -
10 -
0 -
Ideal
0
400
600
(Thousands)
HP Contribution - Industry Database
Figure 3.
Relative horsepower contribution of each turbine model in the
Emissions and Industry Databases

-------
92-142.05
contribution in the Industry Database. However, the trend in general is in a
positive direction for both databases, showing that compressors that
contribute a large amount of horsepower to the total in one database also
contribute significantly to the other. Conversely, compressors that are small
contributors in one database are of similar ranking
in the other. This comparison presents a qualitative assessment of the
representativeness of the Emissions Database for the industry's compressors.
A more quantitative assessment of how well the Emissions Database
represents industry compressors was performed using a comparison between the
emission factors derived from the Emissions Database and the Industry
Database, with the Industry Database used as a close approximation of
compressors in the natural gas industry. To determine the emission factors
for the Emissions Database, an overall average emission factor was calculated
for engines and turbines, separately, after Step 1 in the hierarchy, where
only the data that were exact matches with the compressors in the Test
Database were included. Overall average emission factors were calculated for
engines and turbines, weighting the model-specific emission factors for each
engine and turbine, respectively, by the horsepower contribution of that model
in the Emissions Database.
To determine the emission factors for the Industry Database, model-
specific emission factors were developed after Step III in the hierarchy, and
overall average emission factors were calculated for engines and turbines,
weighted by the horsepower contribution of the engine and turbine models,
respectively, in the Industry Database. The results were that the overall
average emission factors calculated for the Emissions Database were 18 grams
of methane emitted per m3 of fuel for engines, and 0.16 grams of methane
emitted per m3 of fuel for turbines. The overall average emission factors
calculated for the Industry Database were 20 grams of methane emitted per m3
of fuel for engines, and 0.17 grams of methane emitted per m3 of fuel for
turbines. This analysis showed quantitatively that the Emissions Database was
a fairly good representation of the Industry Database and, therefore, the
industry.
CONCLUSIONS AND RECOMMENDATIONS
The annual methane emissions estimate of 0.22 Tg for compressors in the
natural gas industry was less than the previous estimate of 2.2 Tg by a power
of 10. Although compressors are still a significant source of methane
emissions, this estimate reduces the importance of compressors in the
assessment of the natural gas industry's contribution to global warming.
The uncertainty for the estimate, at 4.3 million m3 of methane
(0.15 billion scf), was much less than the portion of total industry methane
emissions contributed by compressors, and was 0.15 percent of the target
uncertainty (100 billion scf) for the industry.
Future work in this area could include an assessment of the effect of
compressor operation, in terms of horsepower, on the methane emission rate.
15

-------
92-142.05
On the basis of the data on this subject, the annual methane emissions could
be higher than the current estimate by a factor of 2, if compressor operation
is at horsepower levels significantly lower than compressors' rated
horsepower.
REFERENCES
1.	Blizzard, D. and A. D. Gillette. Two Cycle Clean Burn (Integral) Gas
Engines. Presented at the Small Prime Movers N0X Control Workshop, San
Diego, CA. June 1991.
2.	TRANSDAT: Compressor Module. Tom Joyce and Associates, Washington,
D.C., for the Gas Research Institute, Chicago, IL. August 1991.
3.	1990 Gas Facts. The American Gas Association, Arlington, VA. 1990.
Page 65.
4.	Natural Gas Annual 1989. Department of Energy, Energy Information
Administration, Washington, D.C. Volume I. 1989.
5.	@Risk, Risk Analysis and Simulation Add-in for Lotus 1-2-3.
Version 1.5. Palisade Corporation, Newfield, NY. March 1989.
ACKNOWLEDGMENTS
This project was cofunded by the Gas Research Institute, Chicago, IL,
and the Air and Energy Engineering Research Laboratory,* U.S. Environmental
Protection Agency, Research Triangle Park, NC. Radian gratefully acknowledges
the contribution made by the sponsors.
(*) Later redesignated the Air Pollution Prevention and Control Division
of EPA's National Risk Management Research Laboratory.
16

-------
A„™T t-, 1100 TECHNICAL REPORT DATA
J\ hi -tiiJo, 1_/- r~ llzo (Please read Instructions on the reverse before completing)
—
1. REPORT NO. 2.
EPA/600/A-96/131
3. RECM
4. TITLE AND SUBTITLE
National Estimate of Methane Emissions from
Compressors in the U. S. Natural Gas Industry
5. REPORT DATE
6. PERFORMING ORGANIZATION CODE
7. AUTHOR(S)
D.L.Jones, L.M.Campbell, C. E. Burklin, M. Gun-
dappa (Radian); and R. Lott (GRI)*
8. PERFORMING ORGANIZATION REPORT NO.
9. PERFORMING ORGANIZATION NAME AND ADDRESS
Radian Corporation
P.O. Box 13000
Research Triangle Park, North Carolina 27709
10. PROGRAM ELEMENT NO.
11. CONTRACT/GRANT NO.
68-D1-0031
12. SPONSORING AGENCY NAME AND ADDRESS
EPA, Office of Research and Development
Air Pollution Prevention and Control Division
Research Triangle Park, NC 27711
13. TYPE OF REPORT AND PERIOD COVERED
Published paper; 4/92-9/93
14. SPONSORING AGENCY CODE
EPA/600/13
is. supplementary notesproject officer is David A. Kirchgessner, Mail Drop 63,
919/541-4021. Presented at AWMA Conference, Kansas City, MO, 6/22—26/92.
(*) Gas Research Institute, Chicago, IL 60631. (Paper No. 92-142.05.)
i6. abstractpaper discusses a cofunded Gas Research Institute/EPA program to
evaluate methane emissions from the natural gas industry in the U. S. The program
consists of an emission testing program and an engineering assessment program
for the major methane emission sources within the natural gas industry. One me-
thane emission source is reciprocating engines and turbines that are used to drive
compressors in the natural gas industry. In the study, the evaluation of methane
emissions from the natural gas industry had a target unce rtainty of 2, 841 million
standard cubic meters (100 billion standard cubic feet). The methane emissions
estimate of 0.22 Tg for compressors in the natural gas industry is much less than
the previous estimate of 2.2 Tg, by a power of 10. Although compressors are still
a significant source of methane emissions, this estimate reduces the importance
of compressors in the assessment of the natural gas industry's contribution to glo-
bal warming. The uncertainty for the estimate, at 0.003 Tg (0.15 billion scf), is
much less than the portion of the total industry methane emissions contributed by
compressors, and is 0.15% of the target uncertainty (100 billion scf) for the indus-
try.
17. KEY WORDS AND DOCUMENT ANALYSIS
a. DESCRIPTORS
b.IDENTIFIERS/OPEN ENDED TERMS
c. COSATI Field/Group
Pollution Distribution
Methane
Emission
Compressors
Natural Gas
Production
Pollution Control
Stationary Sources
Global Warming
13 B
07 C
14G
13 G
2 ID
13. DISTRIBUTION STATEMENT
Release to Public
19. SECURITY CLASS (This Report)
Unclassified
21. NO. OF PAGES
20. SECURITY CLASS (This page)
Unclassified
22. PRICE
EPA Form 2220-1 (9-73)

-------