Lessons
 Learned
NaturalGas
EPA POLLUTION PREVENTER
SBl
       N
\
 CD
From  Natural  Gas STAR  Partners
DIRECTED INSPECTION AND MAINTENANCE AT

COMPRESSOR STATIONS

Executive Summary

The U.S. natural gas transmission network contains more than 279,000 pipeline miles. Along this network, com-
pressor stations are one of the largest sources of fugitive emissions, producing an estimated 50.7 billion cubic
feet (Bcf) of methane emissions annually from leaking compressors and other equipment components such as
valves, flanges, connections, and open-ended lines. Data collected from Natural Gas STAR partners demon-
strates that 95 percent of these methane emissions are from 20 percent of the leaky components at compressor
stations.

Implementing a directed inspection and maintenance (DI&M) program is a proven, cost-effective way to detect,
measure, prioritize, and repair equipment leaks to reduce methane emissions. A DI&M program begins with a
baseline survey to identify and quantify leaks. Repairs that are cost-effective to fix are then made to the leaking
components. Subsequent surveys are based on data from previous surveys, allowing operators to concentrate
on the components that are most likely to leak and are profitable to repair. Baseline surveys of Natural Gas STAR
partners' transmission compressor stations found that the majority of fugitive methane emissions are from a rela-
tively small number of leaking components.

Natural Gas STAR transmission partners have reported significant savings and methane emissions reductions by
implementing DI&M. One 1999 study that looked at 13 compressor stations demonstrated that the average value
of gas that could be saved by instituting  a DI&M program at a compressor station is $88,239 per year, at an
average cost of $26,248 per station.
Leak
Source
Compressor
Station
Components
Potential
Average Gas
Savings
(Mcf/yr)
29,413 per com-
pressor station
Method for
Emissions
Reduction
Identify and
measure leaks.
Make cost-
effective repairs.
Value of Gas
Saved
($/yr)1
$88,239 per com-
pressor station
Average Initial
Implementation
Cost2
$26,248 per com-
pressor station
Potential
Average First
Year Savings
$61,991 per com-
pressor station
  1Gas valued at $3.00 per Mcf. Total cost for initial baseline survey and leak repairs.
This is one of a series of Lessons Learned Summaries developed by EPA in cooperation with the natural gas industry on superior
applications of Natural Gas STAR Program Best Management Practices (BMPs) and Partner Reported Opportunities (PROs).

-------
Introduction
Technology
Background
Transmission compressor stations boost pressure at various points along
natural gas transmission pipelines to overcome the pressure losses that
occur along a long distance pipeline. The more than 279,000 miles of natu-
ral gas transmission pipeline are supported by approximately 1,790 com-
pressor stations. Most compressor stations are equipped with either gas-
fired reciprocating compressors or centrifugal compressors (turbines). These
compressors and associated components, such as pipelines and valves, are
subjected to substantial mechanical and thermal stresses, and as a result
are prone to leaks.

A DI&M program at compressor stations can reduce methane emissions
and yield significant savings by locating leaking components and focusing
maintenance efforts on the largest leaks that are profitable to repair.
Subsequent emissions surveys are directed towards the site components
that are most likely to leak, as well as cost-effective to find and fix.

DI&M  programs begin with a comprehensive baseline survey of all equip-
ment components at the compressor stations in the transmission system.
Operators first identify leaking components and then measure the emissions
rate for each leak. The  repair cost for each leak is evaluated with respect to
the expected gas savings and other economic criteria such as payback peri-
od. The initial leak survey results and equipment  repairs are then  used to
direct  subsequent inspection and maintenance efforts.

Leak Screening Techniques

Leak screening in a DI&M program may include all components in a com-
prehensive baseline survey, or may be focused only on the components that
are likely to develop significant leaks. Several leak screening techniques can
be used:

*   Soap Bubble Screening is a fast, easy, and very low-cost  method to
    screen for leaks. This technique involves spraying a soap solution on
    small, accessible components such as threaded connections. Soaping
    is effective for locating loose fittings and connections, which can  be
    tightened on the spot to fix the leak, and for quickly checking the tight-
    ness of a repair.  Operators can screen about  100 components per
    hour by soaping.
*   Electronic Screening using small hand-held gas detectors or "sniff-
    ing" devices provides another fast and convenient way to detect
    accessible leaks. Electronic gas detectors are equipped with catalytic
    oxidation and thermal conductivity sensors  designed to detect the
    presence of specific gases. Electronic gas detectors can be used on

-------
    larger openings that cannot be screened by soaping. Electronic
    screening is not as fast as soap screening (averaging 50 components
    per hour), and pinpointing leaks can be difficult in areas with high
    ambient concentrations of hydrocarbon gases.
*  Organic Vapor Analyzers (OVAs) and Toxic Vapor Analyzers (TVAs)
    are portable hydrocarbon detectors that can also be used to identify
    leaks. An OVA is a flame ionization  detector (FID), which measures the
    concentration of organic vapors over a range of 9 to 10,000 parts per
    million (ppm). A TVA combines both an  FID and a photoionization
    detector (PID) and can measure organic vapors at concentrations
    exceeding  10,000 ppm. TVAs and OVAs measure the concentration of
    methane in the area around a leak.
*  Acoustic Leak Detection uses portable acoustic screening devices
    designed to detect the acoustic signal that results when pressurized
    gas escapes through an orifice. As gas  moves from a high-pressure to
    a low-pressure environment across a leak opening, the turbulent flow
    produces an acoustic signal, which is detected by a handheld sensor
    or  probe, and read as intensity increments on a meter.  Although
    acoustic detectors do not measure leak rates, they provide a relative
    indication of leak size—a high intensity or "loud" signal corresponds to
    a greater leak rate. Acoustic screening devices are designed to detect
    either high  frequency or low frequency signals.
    High  Frequency Acoustic Detection is best applied in noisy environ-
    ments where the leaking corn-
                                  Exhibit 1: Acoustic Leak Detection
                                          iij
ponents are accessible to a
hand-held sensor. As shown
in Exhibit 1, an acoustic sen-
sor is placed directly on the
equipment orifice to detect
the signal. Alternatively,
Ultrasound Leak Detection is
an acoustic screening
method that detects airborne
ultrasonic signals in the fre-
quency range of 20 kHz to
100 kHz. Ultrasound detec-
tors are equipped with a hand-
held acoustic probe or scanner
that is aimed at a potential leak source from a distance up to 100 feet.
Leaks are pinpointed by listening for an increase in sound intensity
through headphones. Ultrasound detectors can be sensitive to back-
                                 Source: Physical Acoustics Corp.

-------
    ground noise, although most detectors typically provide frequency tun-
    ing capabilities so that the probe can be tuned to a specific leak in a
    noisy environment.


Leak Measurement Techniques

An important component of a DI&M program is measurement of the mass
emissions rate or leak volume of identified leaks, so that manpower and
resources are allocated only to the significant leaks that are cost-effective to
repair. Four leak measurement techniques can be used:

*  Toxic Vapor Analyzers (TVAs) can be used to estimate mass leak
    rate. The TVA-measured concentration in ppm  is converted to a mass
    emissions rate by using a correlation equation.  A major drawback to
    TVAs for methane leak measurement is that the correlation equations
    are typically not site-specific. The mass leak rates predicted by general
    TVA correlation equations have been shown to deviate from actual
    leak rates by as much as three or four orders of magnitude. Similarly, a
    study conducted jointly by Natural Gas STAR partners, EPA, the Gas
    Research Institute (GRI-now GTI, the Gas Technology Institute), and
    the American Gas Association (AGA) found that TVA concentration
    thresholds, or "cut-off" values, such as 10,000  ppm or 100,000 ppm,
    are ineffective for determining which methane leaks at compressor sta-
    tions are cost-effective to fix. Because the use of general TVA correla-
    tion equations can increase measurement inaccuracy, the develop-
    ment and use of site-specific correlations will be more effective in
    determining actual leak rates.
*  Bagging Techniques are commonly used to measure mass emissions
    from equipment leaks. The leaking component  or leak opening is
    enclosed in a "bag" or tent.  An inert carrier gas such as nitrogen is
    conveyed through the bag at a known flow rate. Once the carrier gas
    attains equilibrium, a gas sample is collected from the bag and the
    methane concentration of the sample is measured. The mass emis-
    sions rate is calculated from the measured methane concentration of
    the bag sample and the flow rate of the carrier gas. Leak rate meas-
    urement using bagging techniques is a fairly accurate (within ± 10 to
    15 percent), but slow process (only two or three samples per  hour).
    Although bagging techniques are useful for direct measurement of
    larger leaks, bagging may not be  possible for equipment components
    that are very large, inaccessible, and unusually  shaped.
*  High Volume Samplers capture all of the emissions from a leaking
    component to accurately quantify leak emissions rates. Exhibit 2

-------
 Exhibit 2: Leak Measurement Using a
 Might Volume Sampler
Source: Oil & Gas Journal, May 21, 2001
   shows leak measurement using
   a high volume sampler. Leak
   emissions, plus a large volume
   sample of the air around the
   leaking component, are pulled
   into the instrument through a
   vacuum sampling hose. High
   volume samplers are equipped
   with dual hydrocarbon detec-
   tors that measure the concen-
   tration of hydrocarbon gas in
   the captured  sample, as well as
   the ambient hydrocarbon gas concentration. Sample measurements
   are corrected for the ambient hydrocarbon concentration, and a mass
   leak rate is calculated by multiplying the flow rate of the measured
   sample by the difference between the ambient gas concentration and
   the gas concentration in  the measured sample. Methane emissions are
   obtained by calibrating the hydrocarbon detectors to a range of con-
   centrations of methane-in-air.
   High volume samplers are equipped with special attachments
   designed to ensure complete emissions capture and to prevent inter-
   ference from other nearby emissions sources. High volume samplers
   measure leak rates up to 8 standard cubic feet per minute (scfm), a
   rate equivalent to 11.5 thousand cubic feet per day (Mcfd). Leak rates
   greater than 8 scfm must be measured using bagging techniques or
   flow meters. Two operators can measure thirty components per hour
   using a high volume sampler, compared with two to three measure-
   ments per hour using bagging techniques.
* Rotameters and other flow meters are used to measure extremely
   large leaks that  would overwhelm other instruments. Flow meters typi-
   cally channel  gas flow from a leak source through a calibrated tube.
   The flow lifts a "float bob" within the tube, indicating the leak rate.
   Because rotameters are  bulky, these instruments work best for open-
   ended lines and similar components, where the entire flow can be
   channeled through the meter. Rotameters and other flow metering
   devices can supplement measurements made using bagging or high
   volume samplers.
Exhibit 3 summarizes the application and usage, effectiveness, and approxi-
mate cost of the leak screening and measurement techniques described
above.

-------
Exhibit 3: Screening and Measurement Techniques
Instrument/Technique
Soap Solution
Electronic Gas
Detectors
Acoustic Detectors/
Ultrasound Detectors
TVA (flame ionization
detector)
Bagging
High Volume
Sampler
Rotameter
Application and Usage
Small point sources,
such as connectors.
Flanges, vents, large gaps,
and open-ended lines.
All components. Larger
leaks, pressured gas, and
inaccessible components.
All components.
Most accessible
components.
Most accessible
components (leak rate
<11.5Mcfd).
Very large leaks.
Effectiveness
Screening only.
Screening only.
Screening only.
Best for
screening only.
Measurement
requires site-
specific leak
size correlations.
Measurement
only. Time-
consuming.
Screening and
measurement.
Measurement
only.
Approximate
Capital Cost
$100-$500
(depends on cost of
facility)
Under $1,000
$1,000-$20,000
(depends on
instrument
sensitivity, size,
associated
equipment)
Under $10,000
(depends on
instrument
sensitivity/size)
Under $10,000
(depends on sample
analysis cost)
> $10,000
Under $1,000

-------
Decision
Process
A DI&M program is implemented in four steps: (1) conduct a baseline sur-
vey; (2) record the results and identify candidates for cost-effective repair; (3)
analyze the data, make the repairs, and estimate methane savings; and (4)
develop a survey plan for future inspections and follow-up monitoring of
leak-prone equipment.

Step 1: Conduct Baseline Survey. A DI&M program typically begins with
baseline screening to identify leaking components. As the leaking compo-
nents are located, accurate leak rate measurements are obtained using bag-
ging techniques, a high volume sampler, or TVAs that have site-specific con-
centration correlations.
Partners have found that leak
measurement using a high
volume sampler is cost-effec-
tive, fast,  and accurate.
                                                           1.
                                                           2.

                                                           3.
                                                           4.
   Decision Steps for DI&M

Conduct baseline survey.
Record results and identify candi-
dates for repair.
Analyze data and estimate savings.
Develop a survey plan for future
DI&M.
The cost of the baseline sur-
vey to find and measure leaks
at the 13 compressor sta-
tions included in the 1999
EPA/GRI/PRCI study was
approximately $6,900 per
compressor station or about $2.55 per component. A baseline survey that
focuses only on leak screening is substantially less expensive. However, leak
screening alone does  not provide the information needed to make cost-
effective repair decisions. Partners have found that follow-up surveys in an
ongoing DI&M program cost 25 percent to 40 percent less than the initial
survey because subsequent surveys focus only on the components that are
likely to leak and are economic to repair. For some equipment components,
leak screening and measurement can be accomplished most efficiently dur-
ing a regularly scheduled DI&M survey program. For other components, sim-
ple and rapid leak screening can be incorporated into ongoing operation and
maintenance procedures. Some operators train maintenance staff to con-
duct leak surveys, others hire outside consultants to conduct the baseline
survey.

Step 2: Record Results and Identify Candidates for Repair. Leak meas-
urements collected in  Step 1 must be evaluated to pinpoint the leaking com-
ponents that are cost-effective to repair. Leaks are prioritized by comparing
the value of the natural gas  lost with the estimated cost in parts, labor, and
equipment downtime to fix the leak. Some leaks can be fixed on the spot by
simply tightening a connection.  Other repairs are more complicated and
require equipment downtime or new parts. For these repairs, operators may

-------
choose to attach identification markers, so that the leaks can be fixed later if
the repair costs are warranted. Repair costs for components such as valves,
flanges, connections, and open-ended lines are likely to be determined by the
size of the component, with repairs to large components costing more than
repairs to small components. Some large leaks may be found on equipment
normally scheduled for routine maintenance, in which case the maintenance
schedule may be advanced to repair the leak at no additional cost.

As leaks are identified and measured, operators should record the baseline
leak data so that future surveys can focus on the most significant leaking
components. The results of the DI&M survey can be tracked using any con-
venient method or format.  The information that operators may choose to
collect include:

* An identifier for each leaking component.
* The component type (for example, blowdown OEL).
* The measured leak rate.
* The survey date.
* The estimated annual gas loss.
* The estimated repair cost.

This information will direct  subsequent emissions surveys, prioritize future
repairs, and track the  methane savings and cost-effectiveness of the DI&M
program.

Understanding of fugitive methane emissions from leaking equipment at
compressor stations has evolved since the mid-1990s as the result of a
series of field studies sponsored by EPA, GRI, and AGA's Pipeline Research
Committee International (PRCI). A study published in 1996 reported on
emissions factors from emissions measurements at six compressor stations
in 1994. An extension of this study published by Indaco Air Quality Services
in 1995 reported on the results of emissions surveys of 27,212 components
at 17 compressor stations. The third study published in  1999 by EPA, GRI,
and the PRCI  is the most comprehensive to date, and surveyed fugitive
emissions from 34,400 components at 13 compressor stations.

The compressor stations surveyed in the 1999 EPA/GRI/PRCI study range in
size from stations with 15 reciprocating compressors to  stations with only
two reciprocating compressors. Three of the compressor stations surveyed
contain two centrifugal compressors (turbines) each, and no reciprocating

-------
compressors. Two stations contain both reciprocating compressors and tur-
bines. The compressor stations equipped with reciprocating compressors
contain an average of seven reciprocating compressors per station.
Compressor stations with turbines contain an average of two turbines per
station. The compressors are typically installed in parallel so that individual
compressors can be on- or off-line as needed, and each compressor can be
isolated and  depressurized as needed for maintenance. The inlet pressure at
the compressor stations typically ranges from 500 psig to 700 psig, while
the outlet pressure ranges from 700 psig to 1,000 psig.

On average,  the number of components surveyed per compressor station
was 2,707, and 5 percent of these components were found to be leaking.
The total leak rates at the 13 compressor stations ranged from 385 Mcf per
year to 200,000 Mcf per year. The average total station  leak rate was 41,000
Mcf per year. The largest 10 percent of leaks were found to contribute more
than 90 percent of emissions. Exhibit 4 summarizes average emissions fac-
tors for the compressor station components.

At the site emitting 200,00 Mcf per year, a single source accounted for
142,000 Mcf per year of emissions—a vent from the gas system used to
control compressor unloaders. This was not a significant source of gas
emissions at the other sites. The compressor station with the extraordinary
emissions was otherwise quite average, containing only seven reciprocating
compressors. The experience of this station  underscores the value of DI&M
for detecting huge and costly gas leaks at compressor stations of all sizes.

Exhibit 5 illustrates the average leak repair costs for the  13 compressor sta-
tions included in the 1999 EPA/GRI/PRCI study. The repair costs include the
fully loaded cost of labor as well as parts and materials.

-------
Exhibit 4: Average Fugitive Emissions Factors For Equipment Leaks
From Compressor Station Components
COMPONENTS UNDER MAIN LINE PRESSURE1
Component
Description
Ball/Plug Valve
Blowdown Valve
Compressor Cylinder Joint
Packing Seal- Running
Packing Seal - Idle
Compressor Valve
Control Valve
Flange
Gate Valve
Loader Valve
Open-Ended Line (DEL)
Pressure Relief Valve (PRV)
Regulator
Starter Gas Vent
Connector -Threaded
Centrifugal Seal - Dry
Centrifugal Seal - Wet
Unit Valve3
ON COMPRESSOR
Natural Gas
Emissions Factor2
(Mcf/Yr/Comp.)
0.64 (±1.04)

9.9 (±11.1)
865 (± 247)
1,266 (±552)
4.1 (± 3.8)

0.81 (± 0.89)

17.2 (±5.6)


—

0.74 (± 0.46)

—

Total No.
Components
Measured
189

148
178
42
2,324
—
864
—
940
—

—

1,625

—

OFF COMPRESSOR
Natural Gas
Emissions Factor2
(Mcf/Yr/Comp.)
5.33 (±3.71)
207.5 (±171.4)
—

—

4.26 (±7.13)
0.32 (±0.21)
0.61 (±0.43)

81. 8 (±79.6)
57.5 (±63.2)
0.2 (± 0.2)
40.8 (± 43.3)
0.6 (± 0.3)
62.7 (± 66.3)
278
3,566
Total No.
Components
Measured
2,406
57
—

—

33
2,727
1,476

168
117
171
5
10,338
14
2
12
COMPONENTS UNDER FUEL GAS PRESSURE4

Ball/Plug Valve
Control Valve
Flange
Fuel Valve
Gate Valve
Open-Ended Line
Pneumatic Vent
Regulator
Connector — Threaded
ON COMPRESSOR
0.1 (±0.1)

—
27.6 (± 13.5)
—

—

1.21 (±1.66)
414

—
479
—

—

2,511
OFF COMPRESSOR
0.51 (± 0.37)
2.46 (±3.89)
0.2 (±0.2)

0.43 (± 0.36)
2.53 (±2.1 9)
76.6 (±118.1)
4.03 (±3.98)
0.32 (±0.1 6)
654
69
1,650

640
42
14
103
3,654
10
                       1Main line pressure range from 500 psig to 1,000 psig.
                       Emission factors with associated 95% confidence intervals.
                       3Unit valve leakage is  measured on depressurized compressors. Most of the compressors surveyed remained pressurized when taken off-line.
                       fuel gas pressure is typically 70 psig to 100 psig. The components on the compressor are located at the top of pistons on reciprocating compres-
                       sors and are subjected to substantial vibration and heat. These components only leak when the compressor is running.
                       Source: Indaco Air Quality Services, Inc., 1999, Cost Effective Leak Mitigation at Natural Gas Transmission Compressor Stations, Report No. PRC-
                       246-9526.

-------
Exhibit 5. Average Repair Cost and Payback Period For Equipment
Leaks At Compressor Stations
Component
Description
Ball Valves -1"
Bull Plug on Valve
Compressor Blow Down
Compressor Blow Down
Compressor Valve Cap
Flange -30"
Flange - 6"
Fuel Valve
Gate Valve
Grease Port
Head End of
Compressor
Loader Valve Flange
Loader Valve Stem
Needle Valve
DEL on Valve
Pig Receiver Door
Pipe Thread Fitting
Plug Valves
Pressure Relief
Valve -1"
PRV Flange
Rod Packing
Rod Packing
Rod Packing
Station Blow Down
Tubing
Union
Unit Valve
Unit Valve -10" Plug
Type of
Repair
Replace
Add Teflon Tape & Tighten
Replace
Rebuild
Replace Gasket
Change Gasket
Change Gasket
Replace
Teflon Repack
Replace

Pull & Change Gaskets
Replace Gasket
Rebuild
Replace
Grease
Tighten
Tighten, Add Teflon Tape
Grease

Replace
Tighten
Change Packing Rings
Without Removing Rods
Pull Packing Case and
Rods to Change Rings,
Rework Packing Case
Pull Packing Case
and Rods to Change Rings,
Rework Packing Case &
Replace Rod
Reverse Plug
Tighten
Tighten
Clean & Inject Sealant
Replace
Average
Cost
$120
$15
$600
$200
$60
$1,250
$300
$200
$40
$80

$450
$80
$300
$100
$45
$120
$30
$40

$1,000
$40
$750
$2,600
$ 5,600
$720
$10
$10
$70
$2,960
Source: Indaco Air Quality Services, Inc., 1999, Cost Effective Leak Mitigation at Natural Gas Transmission
Compressor Stations, Report No. PRC-246-9526.
11

-------
                                           Step 3: Analyze Data and Estimate Savings. Cost-effective repair is a criti-
                                           cal part of successful DI&M programs because the greatest savings are
                                           achieved by targeting only those leaks that are profitable to repair. In all
                                           cases, the value of the gas saved must exceed the cost to find and fix the
                                           leak. Partners have found that an effective way to analyze baseline survey
                                           results is to create a table listing all leaks, with their associated repair cost,
                                           expected gas savings, and expected life of the repair. Using this information,
                                           economic criteria such as net present value or payback period can be easily
                                           calculated for each leak repair. Partners can then decide which leaking com-
                                           ponents are economic to repair.

                                           Exhibit 6 shows the total potential savings at the 13 compressor stations
                                           included in the 1999 EPA/GRI/PRCI study, based on fixing only the leaks
                                           with an estimated payback of less than one year. Repair life is assumed to
                                           be two years. For most sites the initial expense of the baseline survey and
                                           repair costs were quickly recovered in gas savings. For two sites, (station 11
                                           and station  12) the baseline survey and repair costs never payback within
                                           the two-year repair period because the total leakage at these compressor
                                           stations  is low.

                                           This example illustrates that a comprehensive DI&M baseline survey, which
                                           includes all of a partner's transmission compressor stations, may uncover a
                                           few individual stations where the baseline DI&M survey may not be prof-
                                           itable. If  DI&M program is profitable for the transmission system as a whole,
                                           the information gained from the few unprofitable stations is still useful. At the
                                           very least, the unprofitable compressor stations for DI&M are  identified and
                                           managed separately in future surveys. Such stations may  be excluded from
                                           future DI&M surveys, surveyed less frequently, or screened with more highly
                                           focused and cost-effective techniques to reduce costs.
12

-------
   si g.
V)  3
co  a
CO  ^-



aS  CD
$  B-
        ffl
        "S
        3
        "2.

        "
        BO
>


(Q
w
O)
"CD
o
ro
             CD
03

TVS
             O)

             £
             03
             01
             -°
             ivs
p

8
                   §

                   ro

                   w
                   a
       CD

       Ol
             ^

             2    8
       Ol
      "os
       ixs
       o

                   s
p

8
                          CO
                                                   CD
                                                                                        CO
                               O1
                               cn

                               03
                               ^

                               o
                                     o
                                     Q
                                     o
      O1
      ro

      en
      Q

      o
                               en
                               Q
                               o
o
Q
o
                                      CD
                                            CJI
                                                   CD
                                     ffi
                                     CD
                                     "en
            £»
            CO

            CO
                                                                                  Q
                                                                                                    -^I


                          CD


                          O
                                      CD


                                      O

                                                                                              00
                                                         en
                                                         ro
                                                         o
                         03
                         o
                                     oo

                                     "0
      oo

      "00
                                                        go
                         -  |

                         Cn    «
                         og
-

 g]
 g
                         a!
                         g
                         "-vl
                               H
                               "
I
03
"-vl
                                                                                 I
                               ±
                               as


                                           CO

                                           O

                                                                                        55     §
                                                                                              w

                         -^    0

                         00    0
                                     CD

                                     W

      03

      00

      CD

      W


                                      PP
                                                                                          JS
                                                                                          fi>
                                                                                          =r.

                                                                                                              » »
                                                                                                                 e
                                              III
                                                                                               i|
                                                                                                 30
                                                                                                                 s?
                                                                                                            S» o>  SL

                                                                                                            tsli
                                                                                               !T3!
                                                                                               a>  5
                                                                                           S
                                                                                                            1. sr 1
                                                   CD CD =

                                                   |-I.i
                                                                                                                        m
                                                                                                                        x
                    en

                    V)

                    i

                    &>

                                                                                                    o
                                                                                                    o
                                                                                                    1
                                                                                                                        V)
                                                               o
                                                               o

                                                               •o

                                                               n
                                                                                                                        v>
                                                                                                                              13

-------
                                          Step 4: Develop a Survey Plan for Future DI&M. The final step in a DI&M
                                          program is to develop a survey plan that uses the results of the initial base-
                                          line survey to direct future inspection and maintenance practices. The DI&M
                                          program should be tailored to the needs and existing maintenance practices
                                          of the facility. An effective DI&M survey plan should include the following ele-
                                          ments:

                                          *  A list of components to be screened and tested, as well as the equip-
                                             ment components to be excluded from the survey.
                                          *  Leak screening and measurement tools and procedures for collecting,
                                             recording, and accessing DI&M data.
                                          *  A schedule for leak screening and measurement.
                                          *  Economic guidelines for leak repair.
                                          *  Results and analysis of previous inspection and maintenance efforts,
                                             which will direct the next DI&M survey.
                                          Operators should develop a DI&M survey schedule that achieves maximum
                                          cost-effective methane savings yet also suits the unique characteristics of a
                                          facility (e.g., the age of the compressors, the number and size of reciprocat-
                                          ing and centrifugal compressors in service, the line pressure and the fuel gas
                                          pressure). Some partners schedule DI&M surveys based on the anticipated
                                          life of repairs made during the previous survey. Other partners base the fre-
                                          quency of follow up surveys on maintenance cycles or the availability of
                                          resources.  Since a DI&M program is flexible, if subsequent surveys show
                                          numerous large or recurring leaks, the operator can increase the frequency
                                          of the DI&M follow-up surveys.  Follow-up surveys may focus on  compo-
                                          nents repaired during previous surveys, or on the classes of components
                                          identified as most likely to leak. Over time, operators can continue to fine-
                                          tune the scope and frequency of surveys as leak patterns emerge.
              Estimated
              Savings
The potential gas savings from implementing DI&M programs at compressor
stations will vary depending on the size, age, equipment, and operating
characteristics of the compressor stations. Natural Gas STAR partners have
found that the initial expense of a baseline survey is quickly recovered in gas
savings.

Exhibit 7 presents three partners' experience in implementing DI&M pro-
grams. Note that the benefit/cost ratio is positive in each case, but varies
widely from 1.7:1 to 95:1.
14

-------
       Exhibit 7: Natural Gas STAR Transmission Partners' Experience

Company A: Fifteen compressor stations were surveyed annually. Total costs
for the DI&M survey and repairs were $350 per station. Leaks were most
commonly found at unit valves. Gas savings totaled 166,010 Mcf, averaging
11,067 Mcf per station.
           Total Gas Savings
           Total Cost of Survey and Repairs
       $498,030
       $5,250
            Net Savings
            Year One Benefit/Cost Ratio
       $492,780
       95:1
Company B: Two compressor stations were surveyed quarterly. Survey costs
averaged $200 per station. Leaks were most commonly found at valve stem
packings, shaft seals, and flange leaks. Of 24 leaks detected, 23 were
repaired at an average cost of $50. Gas savings totaled 17,080 Mcf, averag-
ing 8,540 Mcf per station.
              Total Gas Savings
              Total Survey Costs
              Total Cost of Repairs
$51,240
$1,600
$1,150
              Net Savings                 $48,490
              Year One Benefit/Cost Ratio    19:1
Company C: Sixty-seven compressor stations were surveyed (survey sched-
ule included both quarterly and annual surveys, depending on the station).
Leaks were most commonly found at gaskets and loose fittings, as well as at
compressor valves and packing. Close to 1,150 repairs were made. Gas sav-
ings totaled 132,585 Mcf, averaging 1,978 Mcf per station.
              Total Gas Savings
              Total Survey Costs
              Total Cost of Repairs
$397,755
$176,175
$57,180
              Net Savings                 $164,400
              Year One Benefit/Cost Ratio    1.7:1
Assumes gas price of $3/Mcf.
                                                                                   15

-------
             Lessons
             Learned
DI&M programs can reduce survey costs and enhance profitable leak repair.
Targeting problem stations and components saves time and money needed
for future surveys and helps identify priorities for a leak repair schedule. The
principal lessons learned from Natural Gas STAR partners are:

* A relatively small number of large leaks contribute most of a compres-
   sor station's fugitive emissions.
* Screening concentrations do not accurately identify the largest leaks,
   nor do they provide the information needed to identify which leaks are
   cost-effective to  repair. Effective leak measurement techniques must
   be used to obtain accurate leak rate data.
* A cost-effective DI&M program will target the components that are
   most likely to leak and are economic to repair.
* Natural Gas STAR partners have also found that some compressor
   stations are more leak-prone than others. Tracking of DI&M results
   may show that some compressor stations  may need more frequent
   follow-up surveys than other stations.
* Partners have found it useful  to look for trends, asking questions such
   as "Do gate valves leak more than ball valves?" and "Does one station
   leak more than another?"
* Re-screen  leaking components after repairs are made confirms the
   effectiveness of the  repair. A quick way to check the effectiveness of a
   repair is to use the soap screening method.
* Institute a "quick fix" step that involves making simple repairs to simple
   problems (e.g., loose nut, valve not fully closed) during the survey
   process.
* Develop a system for repairing the most severe leaks first, incorporat-
   ing repair of minor leaks into  regular O&M practices.
* Focus future surveys on stations and components that leak most.
* Record methane emissions reductions at each compressor station and
   include annualized reductions in Natural Gas STAR Program reports.
          References
Bascom-Turner Instruments, personal communication.

Foxboro Environmental Products, personal communication.

Gas Technology Institute (formerly the Gas Research Institute), personal
communication.

Henderson, Carolyn, U.S. EPA Natural Gas STAR Program, personal com-
munication.
16

-------
Howard, louche, Indaco Air Quality Services, personal communication.

Indaco Air Quality Services, Inc., 1995, A High Flow Rate Sampling System
for Measuring Leak Rates at Natural Gas Facilities. Report No. GRI-
94/0257.38. Gas Technology Institute, Chicago, Illinois.

Indaco Air Quality Services, Inc., 1995, Leak Rate Measurements at U.S.
Natural Gas Transmission Compressor Stations. Report No. GRI-
94/0257.37. Gas Technology Institute, Chicago, Illinois.

Indaco Air Quality Services, Inc., 1999, Cost Effective Leak Mitigation at
Natural Gas Transmission Compressor Stations, Report No. PRC-246-9526.
PRC International (report available from the American Gas Association,
Arlington, Virginia).

King Instrument Company, personal communication.

Omega Engineering, personal communication.

Physical Acoustics Corporation, personal communication.

Radian International, 1996,  Methane Emissions from the Natural Gas
Industry, Volume 2,  Technical Report, Report No. GRI-94/0257.1. Gas
Technology Institute, Chicago, Illinois.

Radian International, 1996,  Methane Emissions from the Natural Gas
Industry, Volume 8,  Equipment  Leaks, Report No. GRI-94/0257.1. Gas
Technology Institute, Chicago, Illinois.

Thermo Environmental Instruments  Inc., personal communication.

Tingley, Kevin, U.S.  EPA Natural Gas STAR Program, personal
communication.

UE Systems Inc., personal communication.

U.S. Environmental  Protection Agency, 1994 - 2001, Natural Gas STAR
Program, Partner Annual Reports.

U.S. Environmental  Protection Agency, 1995, Natural Gas STAR Program
Summary and Implementation Guide for Transmission and Distribution
Partners.
                                                                    17

-------
&EPA
    United States
    Environmental Protection Agency
    Air and Radiation (6202J)
    1200 Pennsylvania Ave., NW
    Washington, DC 20460
    EPA430-B-03-008
    October 2003

-------