EPA-450/3-83-007
Guideline Series
Control of Volatile Organic
Compound Equipment Leaks
from Natural Gas/Gasoline
Processing Plants
Emission Standards and Engineering Division
U.S. ENVIRONMENTAL PROTECTION AGENCY
Office of Air, Noise, and Radiation
Office of Air Quality Planning and Standards
Research Triangle Park, North Carolina 27711
December 1983
-------
GUIDELINE SERIES
The guideline series of reports is issued by the Office of Air Quality Planning and Standards
(OAQPS) to provide information to state and local air pollution control agencies; for example, to
provide guidance on the acquisition and processing of air quality data and on the planning and
analysis requisite for the maintenance of air quality. Reports published in this series will be
available - as supplies permit - from the Library Services Office (MD-35), U.S. Environmental
Protection Agency, Research Triangle Park, North Carolina 27711, or for a nominal fee, from the
National Technical Information Service, 5285 Port Royal Road, Springfield, Virginia 22161.
-------
TABLE OF CONTENTS
Page
List of Tables .............. ........... v
List of Figures ........................ vii
Metric Conversion Table ............... ..... viii
1.0 Introduction . ........... \ ..... ...... 1_1
2.0 Sources of VOC Emissions ........... ...... 2-1
2.1 General ................ ....... 2-1
2.2 Description of Fugitive Emission Sources. ...... 2-1
2.3 Baseline Fugitive VOC Emissions ....... .... 2-8
2.4 References ....... . ........... .... 2-12
3.0 Emission Control Techniques ................ 3-1
3.1 Leak Detection and Repair Methods ..... ..... 3-1
3.2 Equipment Specifications. . . . . . . ....... . 3-12
3.3 Other Control Strategies. ... ....... .... 3-16
3.4 Other Considerations ...... ........... 3-21
3.5 References. ....... ...... ....... . 3-25
4,0 Environmental Analysis of RACT .............. 4-1
4.1 Reasonably Available Control Techniques (RACT)
Procedures. ...... ........ . ...... 4-1
4.2 Air Pollution ................... . 4-2
4.3 Water Pollution ........ ' ........... 4.3
4.4 Solid Waste ......... V ........... 4-6
4.5 Energy ........ ..... ........... 4_6
4.5 References. ...... ............... 4.9
5.0 Control Cost Analysis of RACT. ... ........... 5-1
5.1 Basis for Capital Costs . . . . . .......... 5-1
5.2 Basis for Annual Costs ........ ' ........ 5-2
5.3 Emission Control Costs of RACT ............ 5-4
5.4 Cost Effectiveness of RACT ...... . ....... 5-5
5.5 Analysis of Compressor Vent Control System
Costs ........ ................ 5-5
5.6 References. ... ................. . 5-20
-------
TABLE OF CONTENTS
(concluded)
Appendix A. Emission Source Test Data
Appendix 8. Model Plants
Appendix C. Public Comments
Appendix D. Summary and Responses to Draft CTG Comments.
A-l
B-l
C-l
D-l
IV
-------
LIST OF TABLES
Table 2-1
Table 2-2
Table 3-1
Table 3-2
Table 3-3
Table 3-4
Table 3-5
Table 4-1
Table 4-2
Table 4-3
Table 4-4
Table 5-1
Table 5-2
Table 5-3
Table 5-4
Table 5-5
Baseline Fugitive Emission Factors for
Gas Plants
Baseline Emissions from Three Model Gas Plants . . .
Percent Emissions from Sources with
Instrument Readings Equal to or Greater
than 10,000 ppmv
Estimated Percent Components Leaking Per Inspection
for Quarterly Monitoring
Average Emission Rates from Components
above 10,000 ppmv and at 1,000 ppmv. .
Controlled Emission Factors for Quarterly
Leak Detection and Repair. .
Illustration of Skip-Period Monitoring
Example Calculation of VOC Fugitive Emissions
From Model Plant B Under RACT. . ,
Annual Emissions on a Component Type Basis for
Each Model Plant
Energy Impacts on a Component Type
Basis For Model Plant B
Energy Impacts on a Model Plant Basis,
Capital Cost Data , , . ,
Labor-Hour Requirements for Initial Leak
Repair Under RACT
Initial Leak Repair Costs. . . . .
Basis for Annualized Cost Estimates,
Annual Leak Detection and Repair Labor
Requirements for RACT
Table 5-6 Annual Leak Detection and Repair Costs
Page
2-9
2-11
3-6
3-9
3-10
3-13
3-21
4-4
4-5
4-7
4-8
5-7
5-10
5-11
5-12
5-13
5-14
-------
LIST OF TABLES
(concluded)
Page
Table 5-7 Example Calculation of Product Recovery
Credits for Model Plant B 5-15
Table 5-8 Model Plant Recovery Cost Credits 5-16
Table 5-9 Annualized Control Costs for Model Units ...... 5-17
Table 5-10 Examples of Cost Effectiveness by Component
Type for Model Plant B .. 5-18
Table 5-11 Costs and Cost Effectiveness of Compressor
Vent Control System for Model Plant B . . . 5-19
Table A-l Gas Plants Tested for Fugitive Emissions A-5
Table A-2 Instrument Screening Data for EPA-Tested Gas
Plants A-6
Table A-3 Soap Screening Data for API-Tested and EPA-Tested
Gas Plants A-7
Table B-l Example Types of Equipment Included and Excluded
in Vessel Inventories for Model Plant
Development B-4
Table B-2 Number of Components in Hydrocarbon Service and
Number of Vessels at Four Gas Plants B-5
Table B-3 Ratios of Numbers of Components to Numbers
of Vessels B-6
Table B-4 Fugitive VOC Emission Sources for Three Model
Gas Processing Plants B-7
VI
-------
LIST OF FIGURES
Figure 2-1 General Schematic of Natural Gas-Gasoline
Processing
Page
2-2
Figure 2-2 Diagram of a Simple Packed Seal 2-3
Figure 2-3 Diagram of a Basic Single Mechanical Seal ..... 2-4
Figure 2-4 Diagram of a Gate Valve 2-7
Figure 2-5 Diagram of Spring-Loaded Relief Valve ....... 2-7
Figure 3-1 Compressor Distance Piece Purge System. . .... . . 3-15
Figure 3-2 Cost-Effectiveness of Quarterly Leak Detection
and Repair of Valves with Varying Initial Leak
Frequency 3-18
Figure 3-3 Cost Effectiveness of Quarterly Leak Detection
and Repair of Valves for Low Production
Volume Units , 3.23
vn
-------
METRIC CONVERSION TABLE
EPA policy is to express all Tieasurements in Agency documents in
•.
metric units. Listed below ara metric units used in this report /vith
conversion factors to obtain equivalent English units. A list of
prefixes to metric units is also presented.
To Convert
Metric Unit
Multiply 3y
Conversion Factor
To Obtain
Englisn Unit
centimeter (cm)
meter (m)
liter (1)
cubic meter (m )
cubic meter (m )
3
cubic meter (m )
kilogram (kg)
megagram (Mg)
gigagram (Gg)
gigagram (Gg)
joule (J)
Prefix
tera
giga
mega
kilo
centi
milli
micro
0.39
3.23
0.25
254.2
6.29
35
2.2
1.1
2.2
1102
9.48 x IO"4
PREFIXES
Symbol
T
G
M
k
c
m
viii
inch (in.)
feet (ft.)
U.S. gallon (gal)
U.S. gallon (gal)
barrel (oil) (bbl)
cubic fee-t (ft3)
pound (Ib)
ton
million pounds (10 Ibs)
ton
British thermal unit (3tu
Multiplication
Factor
10 12
IO9
IO6
IO3
IO"2
io-3
-6
10 °
-------
1.0 INTRODUCTION
The Clean Air Act Amendments of 1977 require each State in which
there are areas in which the national ambient air quality standards
(NAAQS) are exceeded to adopt and submit revised State implementation
plans (SIP's) to EPA. Revised SIP's were required to be submitted to
EPA by January 1, 1979. States which were unable to demonstrate attain-
ment with the NAAQS for ozone by the statutory deadline of December 31,
1982, could request extensions for attainment with the standard. States
granted such an extension were required to submit a further revised SIP
by July 1, 1982. :.
Section 172(a)(2) and (b)(3) of the Clean Air Act require that
nonattainment area SIP's include reasonably available control technology
(RACT) requirements for stationary sources. As explained in the "General
Preamble for Proposed Rulemaking on Approval of State Implementation
Plan Revisions for Nonattainment Areas,ft (44 FR 20372, April 4, 1979)
for ozone SIP's, EPA permitted States to defer the adoption of RACT
regulations on a category of stationary sources of volatile organic
compounds (VOC) until after EPA published a control techniques guideline
(CTG) for that VOC source category. See also 44 FR 53761 (September 17,
1979). This delay allowed the States to make more technically sour.d
decisions regarding the application of RACT.
Although CTG documents review existing information and data concerning
the technology and cost of various control techniques to reduce emissions,
they are, of necessity, general in nature, and do not fully account for
variations within a stationary source category. Consequently, the
purpose of CTG documents is to provide State and local air pollution
control agencies with an initial information base for proceeding with
their own assessment of RACT for specific stationary sources.
1-1
-------
-------
2.0 SOURCES OF VOC EMISSIONS
2.1 GENERAL
Natural gas/gasoline processing plants (gas plants) are a part of
the oil and gas industry. Field gas is first gathered in the field
directly from gas wells or from oil/gas separation equipment (see
Figure 2-1). The gas may be compressed at field stations for the
purpose of transporting it to treating or processing facilities.
Treating is necessary in certain instances for removal of water,
sulfur compounds, or carbon dioxide. Gas gathering, compression, and
treating may or may not occur at a gas plant. For the purposes of
this document, natural gas processing plants are defined as facilities
engaged in the separation of natural gas liquids from field gas and/or
fractionation of the liquids into natural gas products, such as ethane,
propane, butane, and natural gasoline. Excluded from the definition
are compressor stations, dehydration units, sweetening units, field
treatment, underground storage facilities, liquefied natural gas
units, and field gas gathering systems unless these facilities are
located at a gas plant. Types of gas plants are: absorption,
refrigerated absorption, refrigeration, compression, adsorption,
cryogenic — Joule-Thomson, and cryogenic-expander.
2.2 DESCRIPTION OF FUGITIVE EMISSION SOURCES
In this document, fugitive emissions from gas plants are considered
to be those volatile organic compound (VOC) emissions that result when
process fluid (either gaseous or liquid) leaks from plant equipment.
VOC emissions are defined as nonmethane-nonethane hydrocarbon emissions.
There are many potential sources of,fugitive emissions in a gas plant.
The following sources are considered in this chapter: pumps, compressors,
valves, relief valves, open-ended lines, flanges and connections, and
gas-operated control valves. These source types are described below.
2-1
-------
Sulfur
Recovery
Field Gas Gathering Systems
Field Compression
Gas Treating
Sweetening and Dehydration
(H2S, C02, and HO Removal)
Separation of Natural Gas
Liquids from Field Gas
Fractionation of
Natural Gas Liquids
Dry Gas
to Sales
Sales Products
(ethane, propane, iso-butane, butane, natural gasoline, etc.)
Figure 2-1. General Schematic of Natural Gas-Gasoline Processing,
2-2
-------
2.2.1 Pumps
Pumps are used in gas plants for the movement of natural gas liquids,
The centrifugal pump is the most widely used pump. However, other
types, such as the positive-displacement, reciprocating and rotary
action, and special canned and diaphragm pumps, may also be used.
Natural gas liquids transferred by pumps can leak at the point of contact
between the moving shaft and stationary casing. Consequently, all pumps
except the canned-motor and diaphragm type1require a seal at the point
where the shaft penetrates the housing in order to isolate the pump's
interior from the atmosphere.
Two generic types of seals, packed and mechanical, are currently in
use on pumps. Packed seals can be used on both reciprocating and rotary
action types of pumps. As Figure 2-2 shows, a packed seal consists of a
cavity ("stuffing box") in the pump casing filled with special packing
material that is compressed with a packing gland to form a seal around
the shaft. Lubrication is required to prevent the buildup of frictional
heat between the seal and shaft. The necessary lubrication is provided
by a lubricant that flows between the packing and the shaft.2
End C
— i Stuffing
Box
JSSSSS
PCEXIXD
-------
seals. There are many variations to the basic design of mechanical
seals, but all have a lapped seal face between a stationary element and
a rotating seal ring. In a single mechanical seal application (Figure 2-3),
the rotating-seal ring and stationary element faces are lapped to a very
high degree of flatness to maintain contact throughout their entire
mutual surface area. As with a packed seal, the seal faces must be
lubricated to remove frictional heat. However, because of its construction,
much less lubricant is needed.
PUMP
STUFFING
BOX
OLANO
/RING
STATIONARY
ELEMENT
POSSIBLE
LEAK AREA
SHA
\ROTAT1NG
SEAL RING
9
Figure 2-3. Diagram of a basic single mechanical seal.
2.2.2 Compressors
Three types of compressors can be used in the natural gas production
industry: centrifugal, reciprocating, and rotary. The centrifugal
compressor utilizes a rotating element or series of elements containing
curved blades to increase the pressure of a gas by centrifugal force.
Reciprocating and rotary compressors increase pressure by confining the
gas in a cavity and progressively decreasing the volume of the cavity.
Reciprocating compressors usually employ a piston and cylinder arrangement
while rotary compressors utilize rotating elements such as lobed impellers
or sliding vanes.
As with pumps, sealing devices are required to prevent leakage from
compressors. Rotary shaft seals for compressors may be chosen from
several different types: labyrinth, restrictive carbon rings, mechanical
contact, and liquid film. All of these seal types are leak restriction
2-4
-------
devices; none of them completely eliminates leakage. Many compressors
may be equipped with ports in the seal area to evacuate collected gases.
Mechanical contact seals are a common type of seal for rotary
compressor shafts, and are similar to the mechanical seals described for
pumps. In this type of seal the clearance between the rotating and
stationary elements is reduced to zero. Oil or another suitable lubricant
is supplied to the seal faces. Mechanical seals can achieve the lowest
leak rates of the types identified above, but they are not suitable for
all processing conditions.
Packed seals are used for reciprocating compressor shafts. As with
pumps, the packing in the stuffing box is compressed with a gland to
form a seal. Packing used on reciprocating compressor shafts is often
of the "chevron" or nested V type.4 Because of safety considerations,
the area between the compressor seals and the compressor motor (distance
piece) is normally enclosed and vented outside of the compressor building.
If hydrogen sulfide is present in the gas, then the vented vapors are
normally flared.
Reciprocating compressors may employ a metallic packing plate and
nonmetallic partially compressible (i.e., GRAFFOIL,R TEFLONR) material
or oil wiper rings to seal shaft leakage to the distance piece. Nevertheless,
some leakage into the distance piece may occur. .
2.2.3 Process Valves
One of the most common pieces of equipment in gas plants is the
valve. The types of valves commonly used are globe, gate, plug, ball,
butterfly, relief, and check valves. All except the relief valve (to be
discussed below) and check valve are activated through a valve stem,
which may have a rotational or linear motion, depending on the specific
design. This stem requires a seal to isolate the process fluid inside
the valve from the atmosphere as illustrated by the diagram of a gate
valve in Figure 2-4. The possibility of a leak through this seal makes
it a potential source of fugitive emissions. Since a check valve has no
stem or subsequent packing gland, 'it is not considered to be a potential
source of fugitive emissions.
Sealing of the stem to prevent leakage can be achieved by packing
inside a packing gland or 0-ring seals. Valves that require the stem to
move in and out with or without rotation must utilize a packing gland.
2-5
-------
Conventional packing glands are suited for a wide variety of packing
materials. The most common are various types of braided asbestos that
contain lubricants. Other packing materials include graphite, graphite-
impregnated fibers, and tetrafluoroethylene polymer. The packing material
used depends on the valve application and configuraton.5 These conventional
packing glands can be used over a wide range of operating temperatures.
At high pressures these glands must be quite tight to attain a good
seal.
2.2.4 Pressure Relief Devices
Engineering codes require that pressure-rel ieving devices or systems
be used in applications where the process pressure may exceed the maximum
allowable working pressure of the vessel. The most common type of
pressure-relieving device used in process units is the pressure relief
valve (Figure 2-5). Typically, relief valves are spring-loaded and
designed to open when the process pressure exceeds a set pressure,
allowing the release of vapors or liquids until the system pressure is
reduced to its normal operating level. When the normal pressure is
reattained, the valve reseats, and a seal is again formed.8 The seal is
a disk on a seat, and the possibility of a leak through this seal makes
the pressure relief valve a potential source of VOC fugitive emissions.
A seal leak can result from corrosion or from improper reseating of the
p
valve after a relieving operation.
Rupture disks may also be used in process units. These disks are
made of a material that ruptures when a set pressure is exceeded, thus
allowing the system to depressurize. The advantage of a rupture disk is
that the disk seals tightly and does not allow any VOC to escape from
the system under normal operation. However, when the disk does rupture,
the system depressurizes until atmospheric conditions are obtained,
unless the disk is used in series with a pressure relief valve.
2.2.5 Open-Ended Lines
Some valves are installed in a system so that they function with
the downstream line open to the atmosphere. Open-ended lines are used
mainly in intermittent service for sampling and venting. Examples are
purge, drain, and sampling lines. Some open-ended lines are needed to
preserve product purity. These are normally installed between multi-use
product lines to prevent products from collecting in cross-tie lines due
2-6
-------
PACKING
GLAND
POSSIBLE
LEAK AHEAS
PACKING
Figure 2-4. Diagram of a gate valve.
Possible
Leak Area
Process Side
Figure 2-5. Diagram of a spring-loaded relief valve.
2-7
-------
to valve seat leakage. In addition to valve seat leakage, an incompletely
closed valve could result in VOC emissions to the atmosphere.
2.2.6 Flanges and Connections
Flanges are bolted, gasket-sealed junctions used wherever pipe or
other equipment such as vessels, pumps, valves, and heat exchangers may
require isolation or removal. Connections are all other nonwelded
fittings that serve a similar purpose to flanges, that also allow bends
in pipes (ells), joining two pipes (couplings), or joining three or four
pipes (tees or crosses). The connections are typically threaded.
Flanges may become fugitive emission sources when leakage occurs
due to improperly chosen gaskets or poorly assembled flanges. The
primary cause of flange leakage is due to thermal stress that piping or
flanges in some services undergo; this results in the deformation of the
seal between the flange faces.9 Threaded connections may leak if the
threads become damaged or corroded, or if tightened without sufficient
lubrication or torque.
2.2.7 Gas-Operated Control Valves
Pneumatic control valves are used widely in process control at gas
plants. Typically, compressed air is used as the operating medium for
these control valves. In certain instances, however, field gas or flash
gas is used to supply pressure.5 Since gas is either continuously bled
to the atmosphere or is bled each time the valve is activated, this can
potentially be a large source of fugitive emissions. There are also
some instances where highly pressurized field gas is used as the operating
medium for emergency control valves. However, these valves are seldom
activated and, therefore, have a much lower emissions potential than
control valves in routine service.
2.3 BASELINE FUGITIVE VOC EMISSIONS
Baseline fugitive emission data have been obtained at six natural
gas/gasoline processing plants. Two of the plants were tested by Rockwell
International under contract to the American Petroleum Institute,11 and
four plants were tested by Radian Corporation under contract to EPA.12
Baseline fugitive emission factors were developed from these data,12 and
are presented in Table 2-1. The factors represent the average baseline
2-8
-------
Table 2-1. BASELINE FUGITIVE EMISSION FACTORS FOR
GAS PLANTS (kg/day)a
Component
Valves
Relief valves
Open-ended lines
Compressor seals
Pump seals
Flanges and
connections
Emission factor
0.18
0.33
0.34
6.4
1.2
0.011
(0.48)
(4.5)
(0.53)
(18)
(1.5)
(0.026)
95% Confidence interval
0.1 - 0.3
0.007 - 8
\ 0.1 - 0.7
0.5-3
0.006 - 0.02
(0.2 -
(0.1 -
(0.2 -
(0.5 -
(0.01
1)
100)
1)
4)
- 0.05)
xx = VOC emission values.
(xx)= Total hydrocarbon emission values.
Reference 12.
Compressor seal emission factors from Reference 12 were not used
because the data base included dry gas compressors (which are not subject
to RACT). The emission factors shown are a weighted average of wet gas
and natural gas liquids compressor seals; as developed in Reference 13.
2-9
-------
emission rate from each of the components of a specific type in a gas
plant. The compressor seal emission factor represents a weighted
1 O
average of compressor seals in wet gas and natural gas liquids service.
Compressor seal emission factors -are not directly from gas plant
testing because this data included dry gas compressors which are not
subject to RACT.
The total daily and annual emissions from fugitive sources at
each of the three model gas plants (developed in Appendix B) are shown
in Table 2-2. Total daily emissions are calculated by multiplying the
number of pieces of each type of equipment by the corresponding daily
emission factor. The average percent of total emissions attributed to
each component type is also presented in Table 2-2. The average
percent of total emissions attributed to each component type is the
same for each model plant.
2-10
-------
co
r~*
^
_i
a.
GO
C£>
1
UJ
O
o
LU
UJ
o:
T*
(—
3T
O
U_
co
o
h-H
oo
GO
t—t
1^—
^
1 1 1
•UJ
^»
l^_-
1 — (
_1
UJ
oo
CO
•o
•-X
o
•*-.
VI
c
o
M
to
iS
0)
c
'a;
ftj
CO
]
to
4-*
C
O)
0
i
o
u
o
s.
.0
§
21
V]
4J C
C i— O
OJ ro •'-
U 4J co
i" O V)
O) 4J-t-
4-J
cr
ra
"a.
,_ <•>-
ai
"§
S
c
re
CL
00
•g
21
4J
c:
m
'a.
<
r—
ttJ
|
en
c
0) O
° £
s >—
^^
4J V)
C i—
re CU
•— V)
CL v>
03 0)
CU
-0 0
*— '
4-> V»
ro "cu
f— V)
a. vi
< cu
01
•o o
= d
ro
0
(U re--~
•r- io
0) O --,
VI *f— o>
CO 10 •
*^*
LlJ
OJ
a.
>i
4J
CU
c
o
9.
o
C3
CM
O
§
CM
•—•
O
un
*^f
o
m
o
r- 1
"— *
LO
*5f
0
O
n
CM
o
^
0
un
CM
CO*
o
00
I-*.
0
t/)
OJ
ns
3*
CO ^H
CM a^
O o
co to
«— < CM
ro O
T— 1
*3- 0
LT3 CO
O
UD
CO tO
r- • CM
ro
t— ( I--,
^ .
O O
*3- 0
un
CM O
-H IT)
1—4
f
--
V O
un
GT S
<3- O
m ^3-
ro ro
O O
VI
(U
4) »«™
p- -o
CM
CO
Ol
CO
CM
ro
-Q
cn
• *L.
^*"
,
VJ
dj
3
(O
>
e
o
1
5
S
X
x
tn
aj
3
rt3
>
C
O
V)
VI
*E
£
c
o
-G
s_
a
o
o
"2,
r~-
(O
-J-}
o
h-
It
^KK
X
X
1
1
CM
OJ
5
>
S-
0)
O."
v>
s>
•o
IT)
ro
VI
OJ
VI
V)
2-n
-------
2.4 REFERENCES
1. Cantrell, A. Worldwide Gas Processing. Oil and Gas Journal,
July 14, 1980. p. 88.
2. Organic Chemical Manufacturing, Volume 3: Storage, Fugitive, and
Secondary Sources. Report 2, Fugitive Emissions. U.S. Environmental
Protection Agency. Office of Air Quality Planning and Standards.
Emission Standards and Engineering Division. Research, Triangle
Park, North Carolina. EPA-450/3-SO-025. December 1980.
3. Nelson, W.E. Compressor Seal Fundamentals. Hydrocarbon Processing,
56(12):91-95. 1977.
4. Telecon. R.A. McAllister, TRW, to G.H. Holliday, Shell Oil, Houston,
Texas. March 10, 1981. Compressors and seals at gas plants.
5. Letter from Hennings, T.J., TRW to K.C. Hustvedt, EPA. May 13, 1981.
Results of a telephone survey concerning the use of pneumatic
control valves at gas plants.
6. Lyons, J.D., and C.L. Ashland, Jr. Lyons' Encyclopedia of Valves.
New York, Van Nostrand Reinhold Co., 1975. 290 p.
7. Templeton, H.C. Valve Installation, Operation and Maintenance.
Chem. £.,78(23)141-149, 1971.
8. Steigerwald, B.J. Emissions of Hydrocarbons to the Atmosphere from
Seals on Pumps and Compressors. Report No. 6, PB 216-582, Joint
District, Federal and State Project for the Evaluation of Refinery
Emissions. Air Pollution Control District, County of Los Angeles,
California. April 1958. 37 p.
9. McFarland, I. Preventing Flange Fires. Chemical Engineering
Progress, 65_(8):59-61. 1969.
10. Letter from Hennings, T.J., TRW to K.C. Hustvedt, EPA. July 7, 1981.
Results of a telephone survey concerning control of fugitive emissions
from gas plant compressor seals.
11. Eaton, W.S., et al. Fugitive Hydrocarbon- Emissions from Petroleum
Production Operations. API Publication No. 4322. March 1980.
12. DuBose, D.A., J.I. Steinmetz, and G.E. Harris. Radian Corp.,
Austin, TX. Frequency of Leak Occurrence and Emission Factors for
Natural Gas Liquid Plants. Final Report. Prepared for U.S.
Environmental Protection Agency, Research Triangle Park, North
Carolina. EMB Report No. 80-FOL-l. July 1982.
2-12
-------
13. Memorandum, K.C. Hustvedt, EPA, CPB to James F. Durham, EPA, CPB.
Revised Gas Plant Compressor Seal Emission Factor. February 10,
1983.
2-13
-------
-------
3.0 EMISSION CONTROL TECHNIQUES
Sources of fugitive VOC emissions from gas plant equipment were
identified in Chapter 2 of this document. This chapter discusses the
emission control techniques which are considered to be reasonably
available control technology (RACT) for these sources. These techniques
include leak detection and repair programs and equipment specifications.
The estimated control effectiveness of the techniques is also presented.
This chapter (Section 3.3) also presents other control strategies
applicable to control of fugitive emissions from gas plants. However,
the control effectiveness of these alternative strategies has not been
estimated.
3.1 LEAK DETECTION AND REPAIR METHODS
Leak detection and repair methods can be applied to reduce fugitive
emissions from gas plant sources. Leak detection methods are used to
identify equipment components that are emitting significant amounts -of
VOC. Emissions from leaking sources may be reduced by three general
methods: repair, modification, or replacement of the source.
3.1.1 Individual Component Survey
Each fugitive emission source (pump, valve, compressor, etc.) is
checked for VOC leakage in an individual component survey. The source
may be checked for leakage by visual, audible, olfactory, soap solution,
or instrument techniques. Visual methods are good for locating liquid
leaks, especially pump seal failures. High pressure leaks may be
detected by hearing the escaping vapors, and leaks of odorous materials
may be detected by smell. Predominant industry practices are leak
detection by visual, audible, and olfactory methods. However, in many
instances, even very large VOC leaks are not detected by these methods.
Applying a soap solution (soaping) to equipment components is one
individual survey method. If bubbles are seen in the soap solution, a
3-1
-------
potential leak from the component is indicated. The rate of leakage
may be subjectively determined by the observer by determining the
number of bubbles formed over a specified time period. In addition,
soaping may also serve as a preliminary screening technique, in that
the number of equipment components otherwise subject to instrument
monitoring may be reduced to only those components for which bubbles
were detected. Soaping is not appropriate for very hot sources,
although ethylene glycol can be added to the soap solution to extend
the temperature range. This method is also not suited for moving
shafts on pumps or compressors, since the motion of the shaft may
cause entrapment of air in the spap solution and indicate a leak when
none is present. In addition, the method cannot generally be applied
to open sources.such as relief valves or vents without additional
equipment.
The use of portable hydrocarbon detection instruments is the best
individual survey method for identifying leaks of VOC from equipment
components because it is applicable to all types of sources. EPA
Reference Method 21, Determination of Volatile Organic Compound Leaks,
specifies the procedures for instrument monitoring. This method
incorporates the use of a portable analyzer to detect the presence of
volatile organic vapors at the surface of the interface where direct
leakage to atmosphere could occur. This sampling traverse, is called
"monitoring" in subsequent descriptions. A measure of the hydrocarbon
concentration of the sampled air is displayed in the instrument meter.
The approach of this technique assumes that if an organic leak exists,
there will be an increased vapor concentration in the vicinity of the
leak, and that the measured concentration is generally proportional to
the mass emission rate of the organic compound.
3.1.2 Repair Methods
The following descriptions of repair methods include only those
features of each fugitive emission source (pump, valve, etc.) which
need to be considered in assessing the applicability and effectiveness
of each method.
3.1.2.1 Pumps. In many cases, it is possible to operate a spare
pump while the leaking pump is being repaired. Leaks from packed
3-2
-------
seals may be reduced by tightening the packing gland. At some point,
the packing may deteriorate to the point where further tightening
would have no effect or possibly even increase fugitive emissions from
the seal. The packing can be replaced with the pump out of service.
When mechanical seals are utilized, the pump must be dismantled so the
leaking seal can be repaired or replaced. Dismantling pumps may
result in spillage of some process fluid causing emissions of VOC.
The maximum amount of VOC released to atmosphere from these temporary
emissions may be estimated by assuming1 all the.trapped VOC found
between the inlet and outlet block valves are released. The mass
emissions from pump repair were quantified assuming the VOC contained
between the block valves is approximated by 2 m of 10 cm pipe. As
such, a conservative estimate of pump repair emissions is 8.6 kg VOC,
or the equivalent of the emissions from a leaking pump over a three
day period.1 Pumps should be isolated from the process and flushed of
VOC to a closed system as much as possible prior to repacking or seal
replacement to minimize spillage emissions, however, even with spillage,
repair will result in an emission .reduction.
3.1.2.2 Compressor Seals. As discussed in Chapter 2, there are
three types of compressors used in natural gas plants: centrifugal,
rotary, and reciprocating. Centrifugal and rotary compressors are
driven by rotating shafts while reciprocating compressors are driven
by shafts having a linear reciprocating motion. In either case,
fugitive emissions occur at the junction of the moving shafts and the
stationary casing, but the kinds of controls that can be effectively
applied depend on the type of shaft motion involved.
Repair of leaking compressor seals may be accomplished if there
is a spare compressor or spare compressor capacity such that repairs
can be performed on the leaking seal without a unit shutdown. Leaks
from compressor seals may be reduced by the same repair procedure that
was described for pumps (i.e., tightening the packing). Other types of
seals, however, may require that the compressor be taken out of service
for repair.
3.1.2.3 Relief Valves. In general, relief valves which leak
must be removed in order to repair the leak. In some cases of improper
3-3
-------
reseating, manual release of the valve may improve the seat seal. In
order to remove the relief valve without shutting down the process, a
block valve may be required upstream of the relief valve. A spare
relief valve should be attached while the faulty valve is repaired and
tested. As an alternative to the potential hazard introduced by the
chance of a block valve being mistakenly closed when a vessel is over-
pressured, it may be preferable to install a second block valve and
relief valve for use when the first relief valve is under repair. An
even safer alternative is to install a three-way valve with parallel
relief systems so that one of the two relief systems is always open.
Some relief valves may be difficult to monitor. A state or local
agency may wish to require less frequent monitoring for relief valves
that are difficult to access because of location or hazardous operating
conditions.
3.1.2.4 Valves. Most valves have a packing gland whichrcan be
tightened while in service. Although this procedure should decrease
the emissions from the valve, in some cases it may actually increase
the emission rate if the packing is old and brittle or has been over-
tightened. Unbalanced tightening of the packing gland may also cause
the packing material to be positioned improperly in the valve and
allow leakage. Valves which are not often used can build up a "static"
seal of paint or hardened lubricant which could be broken by tightening
the packing gland.
Plug-type valves can be 'lubricated with grease to reduce emissions
around the plug. Some types of valves have no means of in-service
repair and must be isolated from the process and removed for repair or
replacement. Other valves, such as control valves, may be excluded
from in-service repair by operating procedures or safety procedures.
In many cases, valves cannot be isolated from the process for removal.
If a line must be shut down in order to isolate a leaking valve, the
emissions resulting from the shutdown will possibly be greater than
the emissions from the valve if allowed to leak until the next process
change which permits isolation for repair. Depending on site-specific
factors, it may also be possible to repair leaking process valves by
injection of a sealing fluid into the source of the leak.
3-4
-------
3.1.2.5 Flanges and Connections. In some cases, leaks from
flanges can be reduced by replacing the flange gaskets. Leaks from
small threaded connections can be reduced by placing synthetic (e.g.,
Teflon) tape or "pipe dope" on the male threads before the connection
is made. Most flanges and connections cannot be isolated to permit
repair of leaks. Data show that flanges and connections emit relatively
small amounts of VOC (Table 2-1).
3.1.3 Control Effectiveness of Leak Detection and Repair Methods
There are several factors which determine the control effectiveness
of a leak detection and repair program; these include:
• Action level or leak definition,
• Inspection interval or monitoring frequency,
• Achievable emission reduction from maintenance, and
• Interval between detection and repair of the leak.
3.1.3.1 Action Level. .The instrument reading at which maintenance
is required is called the "action level." The RACT action level is
10,000 ppmv. Components which have indicated instrument readings
equal to or higher than this "action level" are marked for repair.
Table 3-1 gives the percent of total mass emissions affected by the
10,000 ppmv action level for a number of component types. Available
data indicate that a 10,000 ppmv action level provides a reasonable
level of confidence that most large leaks will be detected in routine
screening. However, a higher action level (e.g., 20,000 ppmv) will
result in lower maintenance costs because somewhat fewer leaks will be
detected. Higher action levels were considered for RACT, but the
actual savings in maintenance costs are,not likely to be large compared
to the high credits to be realized from product recovery. In addition,
the monitoring instruments presently in.use for fugitive emission
surveys have a maximum meter reading of.10,000 ppm. Add-on dilution
devices are available to extend the range of the meter beyond 10,000 ppm,
but these dilution probes are inaccurate and impractical for fugitive
emissions monitoring surveys. Other considerations for selection of
the action level are component specific.
For valves, the selection of an action level for defining a leak
is a tradeoff between the desire to locate all significant leaks and
3-5
-------
Table 3-1. PERCENT EMISSIONS FROM SOURCES WITH INSTRUMENT READINGS
EQUAL TO OR GREATER THAN 10,000 ppmv
Component
Percent of mass
emissions for 10,000 ppmv
action level
Valves"
Relief valves
Compressor seals
Pump seals
b,d
86 (87)
77 (77)
93 (93)
79 (79)
xx = VOC emission values.
(xx) - Total hydrocarbon values.
aFraction of total emissions from a given source type that is
attributable to sources with instrument readings equal to or greater
than the 10,000 ppmv action level.
Reference 3.
°Reference 4.
dBased on a weighted average of compressor seals in wet gas and natural
gas liquids service. Reference 5.
3-6
-------
to ensure that emission reductions are possible through maintenance.
Although test data show that some valves with meter readings less than
10,000 ppm have significant emission rates, most of the major emitters
have meter readings greater than 10,000 ppm. Information obtained
through EPA in-house testing and' industry testing ' indicates that in
actual fugitive emission surveys, most sources of VOC have meter
readings which are very low or very high. Maintenance programs on
valves have shown that emission reductions are possible through on-line
repair for essentially all valves with non-zero meter readings. There
are, however, cases where on-line repair attempts result in an increased
emission rate. The increased emissions from such a source could be
greater than the emission reduction if maintenance is attempted on Tow
leak valves. These valves, however, should be able to achieve essentially
100-percent emission reduction through off-line repair because the
leaking valves can either be repacked or replaced. The emission rates
from valves with meter readings greater than or equal to 10,000 ppm are
significant enough so that an overall emission reduction will occur
for a leak detection and repair program with a 10,000 ppm action
level. :
For pump and compressor seals, selection of an action level is
different because the cause of leakage is different. Compared to
valves which generally have zero leakage, most seals leak to a certain
extent while operating normally. The routine leakage is generally
low, so these seals would tend to have low instrument meter readings.
With time, however, as the seal begins to wear, the concentration and
emission rate are likely to increase. At any time, catastrophic seal
failure can occur with a large increase in the instrument meter reading
and emission rate. As shown in Table 3-1, over 90 percent of the
emissions from compressor seals and approximately 80 percent of the
emissions from pumps are from sources with instrument meter readings
greater than or equal to 10,000 ppm. Properly designed, installed,
and operated seals have low instrument meter readings, and the bulk of
the pump and compressor seal emissions are from seals that have worn
out or failed such that they have a concentration equal to or greater
than 10,000 ppm.
3-7
-------
3.1.3.2 Inspection Interval. The length of time between inspections
depends on the expected occurrence and recurrence of leaks after a
piece of equipment has been checked or repaired. The choice of the
interval affects the emission reduction achievable since more frequent
inspection will result in leaking sources being found and fixed sooner.
The leak occurrence and recurrence correction factor for quarterly
inspections is estimated to be 90 percent. The estimated percentages
of components found leaking with quarterly inspections are given in
Table 3-2.
3.1.3.3 Allowable Interval Before Repair. If a leak is detected,
the equipment should be repaired within,a certain time period. The
allowable repair time should reflect an interest in eliminating a
source of VOC emissions but should also allow the plant operator
sufficient time to obtain necessary repair parts and maintain some
degree of flexibility in overall plant maintenance scheduling. The
determination of this allowable repair time will affect emission
reductions by influencing the length of time that leaking sources are
allowed to continue to emit pollutants. Some of the components with
instrument readings in excess of the leak definition action level may
not be able to be repaired until the next line shutdown.
The allowable interval before repair for RACT is chosen to be
15 days. The percent of emissions from a component which would be
affected by the 15-day repair interval if all other contributing
' factors were 100 percent efficient is 98 percent. The emissions which
occur between the time the leak is detected and repair is attempted
are increased with longer allowable repair intervals.
3.1.3.4 Achievable Emission Reduction. Repair of leaking components
will not always result in complete emission reduction. To estimate
the' emission reduction from repair of equipment, it was assumed that
leaks are reduced by maintenance to an instrument reading of 1,000 ppmv.
The percent emissions reduction due to repair of leaking valves,
pressure relief valves, and compressor seals is derived from the
average emission rates of these components above 10,000 ppmv and at
1,000 ppmv as shown in Table 3-3.
3-8
-------
Table 3-2. ESTIMATED PERCENT COMPONENTS LEAKING PER INSPECTION
FOR QUARTERLY MONITORING
Component Type
Estimated Percent
of Components
Leaking Initially3
Estimated
Annual Percentage of
Components Found
Leaking with
Quarterly Inspections
Valves
Relief Valves
Compressor Seals
Pump Seals
a
Approximate fraction
equal to or greater
18b
igd
46.7d,f
33b
of components having an i
than 10,000 ppmv prior to
18. 5C
7.66
18. ?e
39. 4C
nstrument reading
repair.
Reference 3.
Reference 8.
Reference 9.
a '
"Annual percent recurrence factors have been applied for quarterly
inspections for relief valves and compressor seals to determine
the percentage of sources maintained. It is assumed that 10 percent
of sources initially detected are found with quarterly monitoring,
therefore, the annual average is calculated as: 0.10 x 4 = 0.4.
The estimated annual percentage of components found leaking at quarterly
inspections is calculated as:
Estimated percentage of
components found leaking
with quarterly inspections
(Percent of\
components \
leaking /
initially /
x 0.4
Reference 5.
3-9
-------
Table 3-3. AVERAGE EMISSION RATES FROM COMPONENTS ABOVE
10,000 ppmv AND AT 1,000 ppmv
Component type
Valves
Compressor Seals
Relief valves
Average emission
rate from sources
above 10,000 ppmv,
kg/day
0.86 (2.3)
18 (39)
1.3 (18.2)
Average emission
rate from sources
at 1,000 ppmv,
kg/day
0.015 (0.017)
0.36 (0.78)
0.141 (0.146)
Percent
reduction
98 (99)
98 (98)
89 (99)
xx = VOC emission values.
(xx) = Total hydrocarbon values.
aEmission factor for leaking sources. Calculated by multiplying,the
baseline emission factor (Table 2-1) times the percent emissions .
with instrument readings greater than 10,000 ppmv (Table 3-1), divided
by the percent components with instrument readings greater than
10,000 ppmv (Table 3-2).
Assumed emission factor for leaking sources that have been successfully
repaired (on the average, repair is not perfect).
clmmediate percent reduction in emissions due to successful on-line
leak repair.
Reference 5.
3-10
-------
3-1'3'5 Development of Controlled Emission Factors. There are
two models available for estimating emission reduction efficiency from
leak detection and repair programs. Controlled emission factors used
in this document are calculated using both models. The first model
(the computer leak detection and repair (LDAR) model.10) is applied to
valves and pumps. It is the preferred model because it incorporates
recently available data on leak occurrence and recurrence and data on
the effectiveness of simple in-line repair. These data are not available
for relief valves and compressor seals. Therefore, a second model
(The ABCD model) is applied to relief valves and compressor seals.
The ABCD model can be expressed mathematically by the following equation:11
Emission reduction efficiency = A x B.x C x-D
Where:
A = Theoretical Maximum Control Efficiency = fraction of total
mass emissions for each source type with instrument readings
greater than the action level (Table 3-1).
B = Leak Occurrence and Recurrence Correction Factor = correction
factor to account for sources which start to leak between
inspections (occurrence) and for sources which are found to
be leaking, are repaired and start to leak again before the
next inspection (recurrence), and for sources not repaired.
C = Noninstantaneous Repair Correction Factor = correction
factor to account for emissions which occur between detection
of a leak and subsequent repair; that is, repair is not
instantaneous.
D = Imperfect Repair Correction Factor = correction factor to
account for the fact that some sources which are repaired
are not reduced to zero emission levels. For computational
purposes, all sources which are repaired are assumed to be
reduced to an emission level equivalent to an instrument
reading of 1,000 ppmv.
The ABCD model control efficiencies for relief valves and compressor
seals, however, have been modified to correct for the accuracy of the
3-11
-------
engineering judgment employed to derive one of the model inputs as
discussed in the AID.10 The accuracy of the judgment was approximated
by the comparison of the LDAR model and the ABCD model control efficiencies
9
for valves, as:
LDAR Control
Effectiveness
/ABCD Model \
Control 1 X
/ Valve LDAR Model
y Control Effectiveness
/Valve ABCD Model
I OVJ HUlVJl I I 1 VA 1 » N- ' t*-' v»* i i www t
^Effectiveness/ ^Control Effectiveness
Emissions reduction efficiencies are presented in Table 3-4, as
are controlled emission factors. The controlled emissions factors are
calculated as:
Controlled
Emission
Factor
= Baseline -
Factor
Baseline
Factor
Emission
x Reduction
Efficiency
using the baseline emission factors in Table 2-2.
3.2 EQUIPMENT SPECIFICATIONS
Fugitive emissions may be reduced by using process equipment
designed to prevent leakage. Equipment specifications are considered
here only for control of emissions from control valves and open-ended
1i nes.
3.2.1 Gas-Operated Control Valves
VOC emissions from pneumatic control valves result when field gas
or flash gas is used as the operating medium. These emissions can be
eliminated by switching to the use of compressed air or nonVOC gas
such as methane. This will require installation of an air compression
system and/or reconnection. of the appropriate pressure supply lines.
3.2.2 Qpen-Ended Lines
Fugitive emissions from open-ended lines are caused by leakage
through the seat of a valve upstream of the open end of the line.
Fugitive emissions from open-ended lines can be controlled by installing
a cap, plug, flange, or second valve to the open end of the line. In
the case of a second valve, the upstream valve should always be closed
first after the use of the valves. Each time the cap, plug, flange,
or second valve is opened, any VOC which has leaked through the first
valve seat will be released. The control efficiency will depend on
3-12
-------
Table 3-4. CONTROLLED EMISSION FACTORS FOR QUARTERLY
LEAK DETECTION AND REPAIR
Baseline
Equipment Emission Factor3
Item (kg/day) '.
Valves
Relief Valves
Compressor Seals
Pump Seals
0.18 (0.48)
0.33 (4.5)
6.4 (18)
1.2 (1.5)_ .
Control
Efficiency
(%)
77 (77)b
63 (69)C
83 (81)d
58 (58)b
Controlled
Emission Factor
(kg/day)
0.041
0.12
1.1
0.50
(0.11)
(1-4)
(3.4)
(0.63)
xx = VOC emissions factor
(xx) = THC emission factor
aFrom Table 2-2.
From LDAR Model. Reference 8.
cFrom ABCD Model, corrected as described in Section 3.1.3.5,
Reference 5.
Reference 9.
3-13
-------
such factors as frequency of valve use, valve seat leakage, and material
that may be trapped in the cap or plug. Annual VOC emissions from a
leaking open-ended valve are approximately 100 kg. Assuming that
open-ended lines are used an average of 10 times per year, that 0.1 kg
of trapped organic material is released when the valve is used, and
that all of the trapped organics released are emitted to atmosphere,
the annual emissions from closed off open-ended lines would be 1 kg.
This would be a 99 percent reduction in emissions. Due to the conservative
nature of these assumptions, a 100 percent control efficiency has been
used to estimate the emission reductions of closing off open-ended
lines.
3.2.3 Compressor Seals
Centrifugal and rotary compressors are both driven by rotating
shafts. Emissions from these types of compressors can be controlled
by the use of mechanical seals with barrier fluid (liquid or gas)
systems or by the use of liquid film seals. In both of these types of
seals, a fluid is injected into the seal at a pressure higher than the
internal pressure of the compressor. In this way, leakage of the
process gas to atmosphere is prevented except when there is a seal
failure. As in the case of pumps, seal fluid degassing vents must be
controlled with a closed vent system to prevent process gas from
escaping from the vent.
Reciprocating compressors involve a piston, cylinder, and drive-shaft
arrangement. Since the shaft motion is linear, a packing gland arrangement
is normally employed to prevent leakage around the moving shaft. This
type of seal can be improved by inserting one or more spacer rings
into the packing and connecting the void area or areas thus produced
to a collection system through vents in the housing. This is referred
to as a "scavenger" system. As with other fugitive emission collection
systems, these vents must be controlled to prevent fugitive emissions
from entering the atmosphere. However, venting the seal does not
eliminate emissions from reciprocating compressors entirely, because
emissions can still occur into the distance piece area. As shown in
Figure 3-1, these leaks can be controlled by enclosing the distance
piece area and installing suitable piping to vent the emissions either
3-14
-------
LU
I—
oo
>-
oo
LU
CD
LU
o
-------
to a flare, a plant process heater, or back into a low pressure point
in the process. For the latter two cases, an auxilliary compressor
may be required to compress the vent stream to a usable pressure.
Purging the distance piece with natural gas could be performed to keep
the enclosure above the upper explosive limit and to ensure a nonexplosive
atmosphere.
Obtaining a good seal at the distance piece door and at the point
where emissions are vented from the distance piece or seal area is
necessary for maintaining a sufficient pressure (e.g., 2 to 5 psig).
Block valves should also be installed in order to close vent lines
during compressor shutdown periods. This will prevent hydrocarbon
vapors from entering the work place and air from entering the vent
system during compressor maintenance. There may be instances where
retrofitting of such a vent control system to a compressor distance
piece may be infeasible for safety reasons. Therefore, the application
of this preventive program as a retrofit will have to be evaluated on
a case-by-case basis.
3.3 OTHER CONTROL STRATEGIES
This section discusses two fugitive emission control strategies
for valves other than the quarterly leak detection and repair procedures
discussed above. These strategies are limited in application to
valves, because the other component types (pumps, compressors, and
relief valves) are relatively few in number. The statistics used in
estimating the effectiveness of the alternative-strategies are inappropriate
for small populations of components. For example, it is difficult to
quantify a "low leak frequency" in reference to a population of six
pumps at a medium-sized gas plant. There are also differences between
valves and other component types in the way that leaks occur. Valves
develop leaks slowly over time with small percent emission increases
over a given time interval. Other component types, however, may leak
at very low levels over a long period of time prior to a sudden equipment
failure that results in a very high emission rate. Therefore, leak
history of individual components other than valves may not be a good
indicator of the likelihood of a leak in the future. This is an
important consideration when selecting an appropriate monitoring
frequency for a particular component type.
3-16
-------
These strategies should be considered alternatives to quarterly
leak detection and repair to allow process units the flexibility to
meet a level of performance using control procedures considered most
appropriate by that process unit. Process units which currently have
relatively few leaking valves because of good design or existing
control procedures would be most likely to benefit from these strategies
if they were included in regulations adopted by a State agency. Thus,
these alternative control strategies might be included in State regulations
as alternative standards to quarterly leak detection and repair.
Before implementing one of these alternative control strategies,
however, an owner or operator should be required to notify the Director
of the State agency.
3.3.1 General
The emission reduction and annualized cost of a quarterly leak
detection and repair program depend in part on the number of valves
found leaking during inspections. Since about 95 percent of the com-
ponents to be monitored in a gas plant are valves, most of the cost of
detecting leaks in a process unit can be attributed to valves. In
general, few leaks mean VOC emissions are low. Consequently, the
amount of VOC emissions that could be reduced through a leak detection
and repair program and the product recovery credit associated with the
program would be small. As a result, the annualized cost of a leak
detection and repair program for a process unit increases as the
number of leaks detected and repaired decreases. As the percent of
valves found leaking decreases the product recovery credit decreases.
The direct cost for monitoring, however, remains the same because the
number of valves which must be monitored remains the same. Therefore,
the cost effectiveness (annualized cost per megagram of emissions
controlled) of a leak detection and repair program varies with the
number of valves (or the percent of valves) which leak within a process
unit. The cost effectiveness for a quarterly leak detection and
repair program for valves appears reasonable for leak percentages of
about one percent or higher as shown in Figure 3-2.14
A process unit averaging about one percent of valves leaking will
sometimes have less than one percent of valves leaking and sometimes have
more than one percent leaking. Statistically, if a process unit averaged
one percent of valves leaking, then the percent of valves found leaking
3-17
-------
c
o
I 4,000
a:
g
ci
3,000
2,000
1,000
4 6
Leak Frequency (Percent)
Figure 3-2. Cost-effectiveness of Quarterly Leak Detection and Repair
of Valves With Varying Initial Leak Frequency
during a random inspection would exceed two percent less than five percent
of the time. Two percent of valves found leaking is a reasonable
criterion to judge the applicability of alternative control strategies
for valves.
3.3.2 Allowable Percentage of Valves Leaking
A State regulation incorporating an alternative control strategy
based on an "allowable percentage of valves leaking" would require a
process unit to limit the number of valves leaking at any time to a certain
3-18
-------
i
percentage of the number of valves to be monitored. As discussed
above, it appears that two percent of valves leaking represents a
reasonable performance level for an allowable percentage of valves
leaking. •
This type of regulation would require the owner or operator to
conduct a performance test at least once a year by the applicable test
method. Additional performance tests could be requested by the State.
A performance test would consist of monitoring all valves. All components
other than valves would be subject to quarterly leak detection and
repair. The percentage of valves leaking would be determined by
dividing the number of valves for which a leak was detected by the
number of valves monitored including known leaks that are awaiting
shutdown repair. If the results of a performance test showed that the
percentage of valves leaking was greater than the performance level of
two percent of valves leaking, then the process unit would be in
violation of the State regulation.
Incorporating this type of alternative control strategy in the
State regulation would provide the flexibility of a performance standard.
Compliance with the regulation could be achieved by the method deemed
most appropriate by the plant. The plant could implement the quarterly
leak detection and repair program for valves to comply with the regulation
or it could implement a program of their choosing for valves to comply
with the performance level in the regulation.
3.3.3 Alternative Work Practice for Valves
A State regulation incorporating an alternative" control strategy
for valves based on "skip-period" monitoring would require that a
plant attain a "good performance level" on a continual basis in terms
of the percentage of leaking valves. As discussed above, it appears
that two percent of valves leaking represents a "good performance
level."
This type of regulation would require the owner or operator to
begin with implementation of a quarterly leak detection and repair
program for valves. If the desired "good performance level" of two
percent of valves leaking was attained for valves for a certain number
of consecutive quarters, then one or more of the subsequent quarterly
leak detection and repair periods for these valves could be skipped.
3-19
-------
This strategy is generally referred to as "skip-period" monitoring.
All other components would be subject to quarterly leak detection and
repair intervals.
If implementation of the quarterly leak detection and repair
program showed that two percent or less of the valves were leaking for
j. consecutive quarters, then in quarterly inspections may be skipped.
If the next inspection period also showed that the "good performance
level" was being achieved, then rn quarterly inspections could be
skipped again. When an inspection period showed the "good performance
level" was riot being achieved, then quarterly inspections of valves
would be reinstituted. If j_ consecutive quarterly inspections then
showed again that the good performance level .was being achieved, then
rn quarterly inspections could be skipped again.
As mentioned above, two percent of valves leaking represents a
good level of performance. Table 3-5 illustrates how "skip-period"
monitoring might be implemented in practice. In this case, the "good
performance level" must be met for five consecutive quarters (i=5)
before three quarters of leak detection could be skipped (m=3). If
the quarterly leak detection and repair program showed that two percent
or less of the valves in a plant were leaking for each of five consecutive
quarters, then three quarters could be skipped following the fifth
quarter in which the percent of these valves leaking was less than the
"good performance level." After three quarters were skipped, all
valves would be monitored again on the fourth quarter. Another possible
skip program would allow semi-annual monitoring following two consecutive
quarters at less than the good performance level.
This strategy would permit a plant that has consistently demonstrated
it is meeting the "good performance level" to monitor valves annually
instead of quarterly. Using this approach, a plant could optimize
labor and capital costs to achieve the good level of performance by
developing and implementing its own leak detection and repair procedures
or installing valves with lower probabilities of leaking. Compared to
a standard based on an "allowable percentage of valves leaking," where
not achieving the good performance level would be a violation of the
State regulation, the penalty under the "alternative work practice"
standard would only be a return to routine quarterly monitoring.
' 3-20
-------
Table 3-5. ILLUSTRATION OF SKIP-PERIOD MONITORING9
Quarterly
leak
detection
period
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
Leak rate
of valves
during
period (%)
3.1
0.8
1.4
1.3
1.9
0.6
-
-
-
3.8
1.7
1.5
0.4
1.0
0.9
-
-
-
0.9
-
-
_
1.9
Quarterly
action
taken
. (monitor vs. skip)
monitor
monitor
monitor
monitor
monitor
monitor
skip
skip
skip
monitor
: monitor
monitor
monitor
monitor
monitor
skip
'. skip
skip
monitor
skip
skip
s ki p
; monitor
Good
performance
1 eve!
achieved?
No
Yes
Yes
Y-es
- Yes
Yes
_
• „
_
No
Yes
Yes
Yes
Yes
Yes
«•
Yes
_
Yes
1
2
3
4.
5b
1
2
3
4C
1
2
3
4,
5b
1
X
2
3^
4d
1
2
3 ,
4d -
i=5, m=3, good performance level of 2 percent.
Fifth consecutive quarter below 2 percent means 3 quarters of monitorinq may
be skipped. • . a J-
Percentage of leaks_above 2 percent means quarterly monitoring reinstituted.
Percentage of leaks below 2 percent means 3 quarters of monitoring may be
skipped.
3-21
-------
3.4 OTHER CONSIDERATIONS
This section identifies and discusses other considerations that a
State agency may wish to address when drafting a regulation. These
considerations include components which are difficult to
monitor, small process units, and unit turnarounds.
3.4.1 Pifficult-to-monitor Components •
Some valves may be difficult to monitor because access to the
valve bonnet is restricted or the valves are located in elevated
areas. These valves might be monitored by the use of a ladder or
scaffolding. Valves which could be monitored by the use of a ladder
or which would not require monitoring personnel to be elevated higher
than two meters should be monitored quarterly. However, valves which
require the use of scaffolding or which require the elevation of
monitoring personnel higher than two meters above permanent support
surfaces might be exempted from quarterly monitoring provided they are
monitored annually.
3.4.2 Small Process Unit
The net annual cost and emission reduction of performing a quarterly
leak detection and repair program is principally related to the number
of equipment pieces in a gas processing plant. In gas plants with
very small throughputs it is reasonable to assume that VOC emissions
would not become a large percentage of the gas processed regardless .of
the number of pieces of equipment involved. Further, small non-complex
gas plants are often manned with a minimum number of operators so that
outside personnel may need to be used to perform the monitoring.
Figure 3-3 shows the cost effectiveness of a quarterly leak detection
and repair program as a function of gas plant throughput based on
these considerations.15 Based on this curve, States may wish to
consider exempting from the RACT requirements non-complex gas plants
(plants that do not fractionate the mixed natural gas liquids) that
have design throughputs of less than 10 million scfd.
3.4.3 Unit Turnarounds
A State agency might wish to consider a provision in its regulations
which would allow the agency Director to order an early unit shutdown
for repair of leaking components in cases where the percentage of
3-22
-------
5,000
c
o
-a
OJ
CO
to
•I—
LU
O
o
C7>
to
a>
-M
O
-------
leaking components awaiting repair at unit turnaround becomes excessive.
Use of such a provision, however, must be carefully considered in
terms of the emissions reduction achievable and the costs to the
process unit in production down-time and repair cost.
Alternative methods of treating delay of repair could also be
considered by a State or local agency in reducing the cumulative
number of unrepairable equipment components. For instance, delays of
repair to the next scheduled process unit shutdown (or turnaround)
could be allowed under circumstances where it is technically infeasible
to repair the component in-place/on-line (i.e., without a unit shutdown)
or where replacement parts have been depleted from once-sufficient
inventory. By requiring records of delays and reasons for delays,
State enforcement officers would be supplied with the data necessary
to determine compliance.
3-24
-------
3.5 REFERENCES
1. Memorandum. Rhoads, T.W., Pacific Environmental Services, Inc.
to K.C. Hustvedt, EPA/CPB. Gas Plants CTG: Pump Repair Emissions
Estimate. December 21, 1982.
2. Teller, James H. Advantages Found in On-Line Leak Sealinq. Oil
and Gas Journal, _77 (29): 54-59, 1979.
3.
D.A., J.I. Steinmetz, and G.E. Harris. Radian Corporation,
lf\ \S ^ £• [TnA.mi.«ii»*A.._fl^_l-/\_ i.— . . _
9 - - - - y ~ — „,_ v- w*. j M.IIV* •*••(_• iiuii i o • rxauiail OUF'UUrdtlOl
Austin, Texas. Frequency of Leak Occurrence and Emission Factors
for Natural Gas Liquid Plants. Final Report. Prepared for U.S.
Environmental Protection Agency. Research Triangle Park, North
Carolina. EMB Report No. 80-FOL-l. July 1982.
Memorandum. Henning, T.J., TRW to 'VOC/Onshore Production Docket.
April 2, 1982. Cumulative Distribution of Mass Emissions and
Percent Sources with Respect to Screening Value for Relief Valves
Memorandum. K.C. Hustvedt, EPAiCPB, to James F. Durham, EPA:CPB,
Revised Gas Plant Compressor Seal Emission Factor. February 10,
1983.
Letter w/enclosures from H. H. McClure, Texas Chemical Council,
to W. Barber, OAQPS, EPA. June 30, 1980.
Lee, Kun-chieh, et al. Union Carbide Corporation. A Fugitive
Emission Study in a Petrochemical Manufacturing Unit. Paper
presented at annual APCA meeting. Montreal,'Quebec. June 22-27,
4.
5.
6.
7.
8. Memorandum. Rhoads, T.W., Pacific Environmental Services, Inc.
to Docket A-80-20-B. Evaluation of the Effects of Leak Detection
and Repair on Fugitive Emissions in the Onshore Natural Gas
Processing Industry using the LDAR Model. November 1, 1982.
9. Memorandum. Rhoads, T.W., Pacific Environmental Services, Inc.
to Docket Number A-80-20-B. Calculation of Controlled Emission'
Factors for Pressure Relief Valves and Compressor Leaks. November
1982. :
1,
10. Fugitive Emission Sources of Organic Compounds. Additional
Information on Emissions, Emission Reduction, and Costs. U. S.
Environmental Protection Agency. EPA-450/3-82-010. April 1982.
11. Tichenor, B.A., K.C. Hustvedt, and R.C. Weber. Controlling
Petroleum Refinery Fugitive Emissions Via Leak Detection and
Repair. Symposium on Atmospheric Emissions from Petroleum Refineries
Austin, Texas. EPA-600/9-80-013. November 6, 1979.
3-25
-------
12. VOC Fugitive Emissions in Petroleum Refining Industry-Background
Information for Proposed Standards. U.S. Environmental Protection
Agency, Research Triangle Park, NC. EPA-450/3-81-015a. November
1982.
13. Memorandum. K.C. Hustvedt,.EPA:CPB, to James F. Durham, EPA:CPB,
Cost Basis for Compressor Vent Control System. February 23, 1983.
14. Memorandum. Dimmick, Fred, EPA/SDB, to K.C. Hustvedt, EPA/CPB.
Natural Gas/Gasoline Processing LDAR Model Results. January 24,
1983.
15. Memorandum. T.W. Rhoads, Pacific Environmental Services, Inc.,
to K.C. Hustvedt, EPA/CPB. Cost Effectiveness of RACT as a
Function of Throughput for Smell Gas Plants. December 20, 1982.
3-26
-------
4.0 ENVIRONMENTAL ANALYSIS OF RACT
This chapter discusses the environmental impacts that would
result from implementing reasonably available control technology
(RACT), which is presented in Section 4.1. The primary emphasis is a
quantitative assessment of VOC emission? in the absence of RACT (baseline
emissions) and after implementation of RACT. The impacts of RACT upon
water pollution, solid waste, and energy consumption are also addressed
in this chapter.
4.1 REASONABLY AVAILABLE CONTROL TECHNOLOGY (RACT) PROCEDURES
RACT procedures include weekly visual inspection of pumps and
quarterly monitoring of pumps, valves, compressors, and relief valves.
Relief valves should be monitored and repaired if necessary after they
have vented to the atmosphere. Routine;instrument monitoring is not
necessary for flanges and connections. Any component that appears to
be leaking on the basis of sight, smell, or sound should be repaired.
In addition, difficult-to-monitor valves may require less frequent
than quarterly monitoring. Except when the open end is in use (e.g.,
relief valves and double block and bleed valves), open-ended lines
should be sealed with a second valve, a blind flange, a cap, or a
plug. In the case of a second valve, the upstream valve should be
closed first after each use.
Compressor seals should be monitored quarterly, however, some
plant owners and operators may experience difficulty in reducing VOC
concentrations to less than 10,000 ppmv. Moreover, repair of com-
pressor seals often necessitates a .potential or complete process unit
shutdown because compressors are generally not spared. Consequently,
plants may find it preferable to install a compressor vent control
system (see Section 3.1.2.2). However, retrofitting existing
compressors with these systems may pose a safety problem. Because
of the problems associated with quarterly monitoring or with
4-1
-------
Installing equipment controls in certain cases, RACT for compressors,
therefore, will be determined on a case-by-case basis.
Quarterly monitoring should be performed according to EPA Reference
Method 21, and a source is considered leaking if monitoring results in
an instrument meter reading equal to or greater than 10,000 ppmv. As
discussed in Section 3.1.1, a soap solution may be applied to certain
equipment as a preliminary screening technique for leakage. A soap
score equivalent to 10,000 ppmv is not specified in this guideline
document because soap scoring is not applicable to all source types
(see Section 3.1.1) and because it involves a subjective evaluation of
bubble formation over a specified period of time. However, states may
wish to allow plant owners or operators to use the soap score method
based on'a correlation between soap scoring and instrument readings
for sources where soap scoring is applicable. Leaking components
should be tagged and repaired within 15 days. In those instances
where a leak cannot be repaired within 15 days because of interference
with plant operations, the leak should be repaired at the next line
shutdown.
RACT should apply only to equipment containing or contacting a
process stream with a VOC concentration of 1.0 percent by weight or
more. The purpose of this cutoff is to exclude equipment in product
natural gas service, which contains much less than 1.0 percent by
weight VOC. Equipment with process streams containing relatively low
percentages of VOC (i.e., between 1.0 and 10 percent) contribute a
significant portion of"total emissions from natural gas plants and,
therefore, are subject to RACT requirements. RACT does not apply to
equipment operating under vacuum and equipment in heavy liquid service.
An equipment component is in heavy liquid service if the percent
evaporated is less than 10 percent at 150°C as determined by ASTM
Method D-86. RACT does not apply to wet gas service reciprocating
compressors in plants that do not have a VOC control device such as a
flare or a continuously burning process heater or boiler. Further,
due to the high cost effectiveness of monitoring in small plants,
plants with less than lOMHcfd capacity that do not fractionate natural
gas liquids are exempt from the RACT monitoring requirements.
4-2
-------
- 4.2 AIR POLLUTION
Implementation of RACT would reduce fugitive emissions of VOC
from gas plants significantly. There are no adverse VOC emission
impacts associated with RACT.
4.2.1 Development of Emission Levels
To estimate the VOC emission level associated with RACT, control
efficiencies and emission factors were determined for each type of
component (e.g., valves, pumps). The baseline emission factors for
process equipment, which represent emissions in the absence of RACT,
were previously presented in Chapter 2 (Table 2-1). Controlled emission
factors were developed for valves, pressure relief valves, pump seals
and compressor seals that would be controlled by the implementation of
a leak detection and repair program. Control efficiencies and controlled
emission factors for pressure relief valves and compressor seals were
derived from the ABCD model correction factors and the leak detection
and repair (LDAR) model as discussed in Chapter 3. Control efficiencies
and controlled emission factors for valves and pump seals were derived
directly from the LDAR model as described in Chapter 3. for RACT
requirements specifying equipment controls (i.e., open-ended lines),
it is assumed that zero emissions result from the controlled source.
The controlled emission factors for each component type are presented
in Table 3-4.
In calculating the total VOC fugitive emissions from model plants
controlled under RACT, the controlled emission factors were multiplied
by the number of pieces of equipment for each model plant given in
Table 2-2. An example calculation for estimating emissions from model
plant B under RACT is shown in Table 4-1. Total annual model plant
emissions for each component type are presented in Table 4-2 for both
baseline and RACT levels of control.
4.2.2 Emission Reduction •
The emission reduction expected from the implementation of RACT
can be determined for each model plant. The emission reduction is the
difference between the fugitive emissions before RACT is implemented
and the fugitive emissions after RACT is implemented. These emissions
4-3
-------
Table 4-1. EXAMPLE CALCULATION OF VOC FUGITIVE EMISSIONS FROM
MODEL PLANT B UNDER RACT
Component
Valves
Relief valves
Open-Ended lines
*
Compressor seals
Pump seals
Flanges and connections
Number of
sources in
model plant
(N)
750
12
150
6
6
3,000
Controlled
emission
factor,
kg/day/source
(E)
0.041 (0.11)b
0.12 (1.4)b
0.0 (0.0)c
1.1 (3.4)M
0.50 (0.63)b
0.011 (0.026)b
Total Emissions
Total emissions,
kg/day
(N x E)
31 (82)
1.4 (17)
0 (0)
6.6 (20)
3.0 (3.8)
33 (78)
75 (200)
xx = VOC emission values.
(xx) = Total hydrocarbon emission values.
aFrom Table 2-2.
DFrom Table 3-4. Controlled emission factors are derived from the baseline
emission factors in Table 2-1.
cAssumes installation of second valve, blind flange, cap, or plug with
100 percent control efficiency.
^Based on leak detection and repair. Installation of a compressor vent
control system would achieve 100 percent control efficiency.
4-4
-------
(O
1 —
2:
1
1
Q.
— 1
f~*\
O
SI
o
UJ
o;
0
u_
oo
oo
rv-i
LU
Q.
^~
1—
i
I —
2:
LU
O
5>"
0
CJ
«c
^-^
o
oo
^**
0
oo
oo
i— i
LLJ
O
O
1
e£
_j
2:
2T
CM
1
•=3-
OJ "~
r—
(t3
1 —
s-
c -~~
3 s
r~r_i
O •
tn o
tn«C
ill
L.
.*
i
SJ3
QJ *
•a r—
C 0
3 S-
in c
c o
o u
vi
m
•*
o
as
"*
0)
a.
*"*
ai
o
ex
§
*O 4J *«O **~
O O o Ln ^ ir>
O CM -— ' CM *— * CTl
""""'
CO O 1^
r-- r-i o co co o
m ^~
a>
^* -o ^* ^ »*-
o <£> o r^ P-I co
m — ' ^-^ — " -^-^ c\j
ir> ^ • ^
^H O O OJ <— < CM
»— 1 i—4
0) *— * H-
*-* ^-> vo -^
0-9 ID *r un
O CVJ O CM O CT\
^^
co r^-
r^ r-i co ro o
fO O O O O ^~
O VO Ufl O «— ' tO
r-i. CO
O ^r csj vo co o
VO <4D ^ ^
t— 4
m
O O O*> CT1 f5 CO
ro csj csj co •»— • CNJ
. -^-^
LT, »
Ol rl Ol «± CM CM
^^ rH r-H *— 1
t£> LO . ^^ LO
. «3- to cn ro «— 4 eft
, .
p~ OO
T CM to CO O
to o to ^- o «^
in
c
0
u
11)
in tn c
0) — C
c « o
V) •»- OJ u
-a
r— -o s- tn c
<0 U O r— 10
> -o vi to
c tn QJ tn
> M- c ex ex c
i — i — a) ~ E fa
re: a> ex o 3 !-=-
> o; o o a. LU
o
CM
0
cr>
ro
f»^
CM
^~
• • — '
O
cn
O
co
• o
o"
IT!
CM
CO
en
*3"
Z-*
CM
ro
,_
a
o
i-
i.
c
o
u
•M
0)
U
> u
en
s: -o
O 3
*-> o .
3
•^. -U
en • in
-^ >> >>
u tn
E C
O 0) r—
S- -I- O
<*- U £-
•f- 4-J '
O M- O
•"- 0) U
e ^ *.
a> o c:
> s- aj
c *-> >
o c
oos-
^ o o
tn
in- +J in
to c o
ro tu s-
• o ex
o s. E
ai o
x ex o
-~~ 0 a
CM O
1 i— I M-
CM O
' ai £ er
<0 ^
— 3 i —
: in "ex "io
c S- tn
0) O C
O * en
a. a. - c
§10 1. •,-
O O ^£
1 o > 10
* flJ Q)
• <4- 0) X i—
. CM O O O
i c 3= -a
CM s. ro e:
0) r— 3
0) .CJ I*- • O
1— -i- ex ••-
in x i— ro > to a)
c • ai 'to c
o • t— i— > o *•>
tn o x
tn a) &- o s- • aj aj
D 4) O tn *O 13
3 c ex S- S- 0) M- a) •r-
rc to in c r— o
• e: u v ^ c o a
vi ^3 tn o tn ro s_
§^3 E 1O i- (O
*— c: tu 4-> ro
10 LO ••— tn 3 • in
oocointt)"- o"o
=» 1— C f- C C S_
in Or— in O 0) -M
§•*- o a> -t— c:
V) i- S -O O O
3 in J-> 3 0) f- 0
- — . in •«- c. tn tn <*-
xx in sov)iot*-o
• — • 10 .O o "O tU t«_
4-5
-------
are presented as "totals" in Table 4-2. The average reduction in
emissions for the model plants after RACT is implemented is 71 percent
for VOC and for total hydrocarbons.
4.3 WATER POLLUTION
Although fugitive emissions of VOC from gas plant equipment
primarily impact atmospheric VOC emissions, they also impact water
quality. In particular-, leaking components handling liquid hydrocarbon
streams increase the waste load entering wastewater treatment systems.
Leaks from equipment can contribute to the waste load by entering
drains via runoff. Implementation of RACT should reduce the waste
load on wastewater treatment systems by -preventing equipment leaks
into the wastewater system; therefore, no adverse water pollution
impact is expected.
4.4 SOLID WASTE DISPOSAL
The quantity of solid waste generated by the implementation of
RACT would be insignificant. The solid waste generated would consist
of used valve packings and components which are replaced.
4.5 ENERGY
Implementation of RACT is expected to require little or no energy
consumption at gas plants. Instead, implementation of RACT will save
energy by reducing emissions to atmosphere of methane, ethane, and
VOC. Table 4-3 shows the amount of energy to be saved on a component
basis from implementation of RACT in terms of joules and in barrels of
crude petroleum. Table 4-4 shows the total energy saved per model
plant.
4-6
-------
CQ
I—
•=t
Q_
LU
Q
O
2
CO
1 1 i)
GO
ef
CQ
LU
O_
^^»
1 —
I—
LU
i
^~
c^j
r-
O
GO
O
0.
t 1
^_
CD
o;
LU
1 1 i
•
ro
i
QJ
X}
1
s_
to
QJ
"~
O
•r-3
cn'-
en c
QJ
C (0
o >
4-> 3
O__
c
3 QJ
•0
O) S-
s- >•
• s"^
O 3
•r- OJ
00 i —
00 0
•i- S_
QJ QJ
Q,
It
O O)
, — i
C S-
. ^
"p.
o S
o -a
^^^
en
S_
QJ
C
1 . 1
I.1 J
03
S-
5
^.
C
O
'F™
+J
rs
•o
OJ
*"
00
to
LU
C
'(O
4^
O
•a
QJ
E
(C
-c
4->
QJ
QJ
C
rt3
4->
JJ
*
u
^J
o
QJ
c-
.E
QJ
QJ
C
to
-C
^QJ
C_5
§
-J
QJ
C
O
Q. QJ
E Q.
-J I—
O O O O LO
^t" CO CO I*** r— t
CO i— 1 CM CM
*— • *— *
O O O O CM
O r^ o o cn
CM r~^ ^- r-»
•> •> 0
LO i— t r-<
O -— ^ .to"
LO C3 O CO CO
— CM cyi i— 4 —
o — ' '
O CM
•* O O O CM
• CM LO O
CO 1 — . LO r-H
1 — 1
O LD O i — 1 - — >
cr> • *=3~ cr> i — i
CM f"-* t-H *- — ,—4
O <£> O C> O:
CD ^^ CTi **O r***
CO 00 LO
A
o
CM co o o o
>J3 i— 1 r-< CM
1 —
cr> LO
CO O 01 CM .— 1
CO ^H ,-H
to oo
QJ r—
E T3 oo «
C to QJ
00 <4- QJ QJ OO
QJ QJ ' 1 i-
> -i- C CL CL
r— i — OJ P= E
rtJ QJ Q. O 3
> o: o o a.
QJ
•r-
1—
QJ
^3 ^3
J3 4->
^^
"« *«
3 3
c o"
C QJ
"3 O i —
QJ 4-> _a
E -Q .0 oo
S- O QJ 3 • -r~
4- 4-> E -^ CM C
3 CQ O QJ
tO T3 00 QJ -r- i_ w
O E'"3OC-i— i —
•>— 3 >— 1 OOJOOCQJ
to to IO«S-OO>
00 oo QJ •!— LO QJ CL-I— •<-
SI cojcnQJOtoo
QJ OO QJ 3 *O£ C_3 3 QJ
•i— S- i — CO 0 Q.
1 — QJ tt3 . •> E to
=C 3QJ OLOQJOi-
cc: i — o: en c
IO C O O rQ H— "
i — > •> -i— co JE: o QJ
CT> QJ 4-> C
3 Cn.O fT3 ' — t -O QJ OO fO
C C f"> QJ | | 4_i ^
fH i ^ ^* .^^ ,*^ *fi v{ ^.
(0 4-> CTl' fd -C
CO QJCQ • t— ) T3 ^T T3
c: 2: <_> QJ 4-> QJ c
4J O=>03Et^
3 •!—) •> ea 4-> 4-J (B (O QJ
oo OSLO i- o -i— S CE
en en 3 > cn4-5
-Q coo cnos_ en. o
QJ i- en aj 4-» t-
•o i- o E ex EC
QJ QJ CO O QJ J3
C CLI S--a-r- S_Of—
•i— CO QJ -f- M- QJ S- *->»
EOjr-~ CX3-I— CLQJ3
"O CT> CT CJ r-i J-i
QJ 3 i— t to •<— flj to CQ
4J S- QJ ^- CL QJ O
QJ Oi_i — GOi — CMO
"O O 3 4-» 3 O
i— "4- O C O T3 CO
O • .QC fO r— f- ( lo'iOCM
•f- aj o en CL en CM
4-5 CL CO -r- -r- QJ -t— QJ
<->>j«34-> CnoOO COCT3
34->r-lo (OC IBC
O) 4-> O -O "=3- S. LO 4J
^- E O r— QJ ai f-i
QJ t)_S_ M_rat)_ 4_ c: , —
cc ocxps_aj o • —
O O 3Qi 4-J 3
•<— CXtOg OO4-> OOC4->
OO E '"— 3 •!— (B « -i— QJ CQ
00 O tO QJ OO C- — > 00 O
E -Cto JDM — Q .aajo
Lu jr S- O -Q CLO
0 QJ 4-5 QJ — QJ „
•QJ 4->CL4->rOQ> 4-JCOCM
CM ^= , —
IS- CQJ E4->3 CQJQJ
>3-o o~o o o ojaja«
4- • 3 0 -r-5 CO
QJ "OS-TC34->itJ-aoO
i — to QJO QJ cna)4->4->aj
1 — ••— ( — O r— 3 f— QJ QJ QJ
to 3 t3 CT«3- 3 S= E S-
EoO O4J (J QJ •— ' Cj'33QJ
S— c fO c «^ >,* . ffi tn /i rti
u_ 3 o 4-> c_j _a — cj ro ro ct:
-Q o "^i
4-7
-------
I/O
oo
^^
ca
f—
o
Q_
SE
I"" |
C3
o;
UJ
z:
UJ
•
^^»
t
«5j-
QJ
s
^
CO
•^
o
•r-j
ov — •
'o> C
O>
C (O
0 >
jj^j 13
O O"
3 CU
•o
S- >,
••^
S= £
0 3
••- CU
CO 1 —
CO O
•r- S_
at cu
QL
H-
O CU
-a
4-> =5
C S-
cu o
4^
C r—
O J3
O JO
*^»*
>fc
C7>
s_
cu
c:
UJ
"(5
[ '
o
cu
•c
2
^_>
cu
1
cu
to
^_>
cu
{ 1
o
fO
S-
>^
^x^.
CD
«\
CO
. o
CO
CO
•r—
F
O)
cu
cu
o
o
O)
o;
cu
c-
ra
^^
4^
cu
cu
c
ro
4J
cu
C_3
0
r— •»->
cu c
O i—
s: Q-
f^^ r*^
cr. 0
•=1- LO
d.
CD O
CD O
1 — 1 «3"
co
^— ^ ^— •*
0 0
0 O
O 0
rt «\
CM ^O
^-** ^— *s
0 O
i~~ tn
i— 1 LO
**-^ >»^^
CD O
O CD
1— < «*
t-H CO
^O CD
CO ^H
.—1
CO CM
CM r--.
*^C ro
o
0
O-l
A
^s
0
0
0
r-4
CO
, — .
o
o
CM
A
CO
^-^
0
0
0
«
0
CM
f^^
o
CD
r^
A
, — 1
o
o
o
1— (
'""'
o
10
CO
CD
CO
CM
C 3
•
CO
0
^-J
ta
S-
O)
c
•r—
o
c:
•r—
S-
o
CO
cu
s_
fO
r—
I+~
o
4-3
CO
c
0
•^
CO
CO
•r"
F^
cu
r-^
rC
a>
CO
0
CO
CO
cu
Q.
r-~
o
o
01
c
•^
4_)
c:
cu
>
^
o
^
•o
cu
S-
CJ
5^
cu
o
o
cu
s-
0
c
CO
•r-*
cu
^
cu
nc^
1—
•
CO
1
^*
JO)
r~
JH
ro
I—
M-
o
•a
•a
£
(0
o
„
o
CO
cu
0
c
4-3
0
0
(4—
cu
cu
CO
£2
4-8
-------
4.6 REFERENCES
1. DOE Monthly Energy Review. January 1981. DOE/EIA-0035 (31/01).
2. Nelson, W.L. Petroleum Refinery Engineering. McGraw-Hill Book
Company, Inc. New York, 1958. p. 32.
3.
Perry, R.H., and C.H. Chilton, eds. Chemical Engineers' Handbook,
Fifth Edition. McGraw-Hill Book Company, New York. 1973. p.
4-9
-------
-------
5.0 CONTROL COST ANALYSIS OF RACT
The costs of implementing reasonably available control technology
(RACT) for controlling fugitive emissions of volatile organic compounds
(VOC) from equipment leaks at gas plants are presented in this chapter.
Capital costs, annualized costs, and the cost effectiveness-of RACT
,are presented. These costs have been developed for the individual
equipment pieces and model plants presented in Chapter 2. To ensure a
common cost basis, Chemical Engineering cost indices are used to
adjust costs to June 1980 dollars.
As, discussed in Section 4.1, RACT for compressor seals is quarterly
leak detection and repair. In many instances, however, a compressor
vent control system would be installed. For the purpose of estimating
RACT cost impacts, Sections 5.1 through 5.4 are based on quarterly
leak detection and repair for compressors. Additionally, compressor
vent control costs are discussed in Section 5.5. Capital costs for
the vent control system are included in Table 5-1.
5.1 BASIS FOR CAPITAL COSTS
Capital costs represent the total cost of starting a leak detection
and repair program in existing gas plants. The capital costs for the
implementation of RACT include the purchase of VOC monitoring instruments,
the purchase and installation of caps for all open-ended lines, the
purchase and installation of a compressor vent control system, and
initial leak repair. The cost for initial leak repair is included as
a capital cost because it is expected to be greater than leak repair
costs in subsequent quarters and is a one-time cost. The basis for
these costs is discussed below and presented in Table 5-1.
5.1.1 Cost of Monitoring Instrument
The cost of a VOC monitoring instrument includes the cost of two
instruments. One instrument is intended to be used as a standby
5-1
-------
spare. The cost of $4,600 for a portable organic vapor analyzer was
obtained from a manufacturer.
5.1.2 Caps for Open-Ended Lines
Fugitive emissions from open-ended lines can be controlled by '
installing a cap, plug, flange, or second valve to the open end. Any
one of these pieces of equipment is included in the definition of a
"cap" for an open-ended line. For the purposes of this analysis, the
cost of a cap for an open-ended line is based on a cost of $43 for a
2
one-inch screw-on type globe valve. Line sizes larger than 2" can be
fitted with a reducer, or as an alternative, can be equipped with a
blind flange at a similar cost. A charge of §18 for one hour of
labor is added to the $43 as the cost for installing one cap. Therefore,
the total capital cost for installing a cap on an open-ended line is
$61,
5.1.3 Initial Leak Repair
The implementation of RACT will begin with an initial inspection
which will result in the detection of leaking components. The number
of initial leaks is expected to be greater than the number found in
subsequent inspections. Because initial leak repair is a one-time
cost, it is treated as a capital cost. The number of initial leaks
was estimated by multiplying the percentage of initial leaks per
component type by the number of components in the model plant under
consideration. The repair time for fixing leaks is estimated to be
16 hours for a pump seal, 1.13 hours for a valve, and 40 hours for a
9 14
compressor seal. ' These requirements are presented in Table 5-2.
The initial repair costs given in Table 5-3 were determined by taking
the product of the number of initial leaks, the repair time, and the
hourly labor cost of $18.
5.2 BASIS FOR ANNUAL COSTS
Annual costs represent the yearly cost of operating a leak detection
and repair program and the cost of recovering the initial capital
investment. This includes credits for product saved as the result of
the control program. The basis for the annual costs is given in Table 5-4.
5-2
-------
5.2.1 Leak Detection Labor ;
The implementation of RACT requires visual and instrument monitoring
of potential sources of fugitive VOC emissions. The monitoring labor-hour
requirements for RACT are presented in Table 5-5. The labor-hour
requirements were calculated by taking the product of the assumed
number of workers to monitor a component (1 for visual, 2 for instrument),
the time required to monitor, the number of components in a model
plant, and the number of times the component is monitored each year.
The monitoring times for the various components are 0.5 minute for
visual inspection, 1 minute for valves, 5 minutes for pump and compressor
seals, and 8 minutes for relief valves.9 Monitoring labor costs
presented in Table 5-6 were calculated based on a charge of $18 per
hour.
5.2.2 Leak Repair Labor
Labor is needed to repair leaks which develop after initial
repair. The estimated number of leaks and the labor-hours required for
repair are given in Table 5-5. The repair time per component is the
same as presented for initfal leak repair. Leak repair costs presented
in Table 5-6 were calculated based on a charge of $18 per hour.
5.2.3 Maintenance Charges and Miscellaneous Costs
The annual maintenance charge for caps, is estimated to be five
percent of their capital cost.1 Annual maintenance costs include pump
seal replacement costs at $140 per pump seal repair.14 The annual
cost of materials and labor for maintenance and calibration of monitoring
instruments is estimated to be $3000.11>12'13 An additional miscellaneous
charge of four percent of capital cost for taxes, insurance, and
capital related associated administrative costs is added for the
monitoring instruments and caps.
5.2.4 Administrative Costs
Administrative and support costs associated with the implementation
of leak detection and repair are estimated to be 40 percent of the sum
of monitoring and leak repair labor costs. The administrative and
support costs include record-keeping and reporting requirement costs.
5-3
-------
5.2.5 Capital Charges
The life of caps for open-ended lines is assumed to be ten years
and the life of monitoring instruments is assumed to be six years.
The cost of repairing initial leaks was amortized over a ten-year
period since it is a one-time cost.
The capital recovery cost is obtained from annualizing the installed
capital cost for control equipment. The installed capital cost is
annualized by using a capital recovery factor (CRF). The CRF is a
function of the interest rate and useful equipment lifetime. The
capital recovery can be estimated by multiplying tfie CRF by the total
installed capital cost for the control equipment. This equation for
the capital recovery factor is:
(1 + i)n - 1
where i = interest rate, expressed, as a decimal
n s economic life of the equipment, years.
The interest rate used was ten percent. The capital recovery factors
and other factors used to derive annualized charges are presented in
Table 5-4.
5.2.6 Recovery Credits
The reduction of VOC fugitive emissions results in saving a
certain amount of VOC which would otherwise be lost. The value of
this VOC is a recovery credit which can be counted against the cost of
a leak detection and repair program. The recovery credits for each
model plant are presented in Table 5-5. The VOC saved is valued in
June 1980 dollars at $192/Mg, using a price of 40^/gallon of LPG and •
a specific gravity of 0.55.16 The methane-ethane saved is valued in
June 1980 dollars at $61/Mg, using a price of $1.46/Mcf and an
assumed composition of 80% methane and 20% ethane at standard temperature
and pressure. An example calculation of product recovery credits is
presented in Table 5-7 based on Model Plant B. Model Plant recovery
credits are summarized in Table 5-8.
5.3 EMISSION CONTROL COSTS OF RACT
This section presents the emission control costs of implementing
RACT for each of the three model plants. Both the initial costs and
the annualized costs are included.
5-4
-------
5.3.1 Initial Costs
The cost of initially .implementing RACT consists of capital costs
and initial leak repair. The cost of $9,200 for two monitoring instruments
is the same for all model, plant sizes. . The capital costs for caps for
open-ended lines are annualized on the basis presented in Table 5-4.
5.3.2 Annualized Costs
The annualized RACT control costs includes the initial leak
detection repair costs, annual leak detection cost, and product recovery
credits, as previously discussed. Table 5-9 presents the annualized
costs for the model plants. The net annualized costs to implement
RACT range from $3,300 for Model Plant A to a cost savings of $17,000
for Model Plant C.
5.4 COST EFFECTIVENESS OF RACT
Cost effectiveness is the annual cost per megagram of VOC controlled
annually. The cost effectiveness of RACT for each model plant is the
net annual cost for implementing RACT divided by the emission reduction
achieved under RACT.
The cost effectiveness of implementing RACT for the model plants
is- presented in Table 5-9. The cost effectiveness for Model Plant A
is $140/Mg VOC reduction, and a cost credit of $28/Mg and $74/Mg for
Model Plants B and C, respectively.
The cost effectiveness for each individual component covered by
RACT is presented in Table 5-10 based on Model Plant B. The cost of a
monitoring instrument cannot be attributed to any single type of
component since all components are monitored by the Instrument,
Therefore, the cost for each component does not include the cost of
the monitoring instrument. The instrument cost is included, however,
in the model plant cost effectiveness.
5.5 ANALYSIS OF COMPRESSOR VENT CONTROL SYSTEM COSTS
The cost to install a compressor vent control system is dependent
upon several factors: (1) the type of compressor (reciprocating,
centrifugal), (2) the presence of an existing VOC control device
(flare, process heater), and (3) the type of process fluid being
compressed (wet gas or natural gas liquids). Product recovery credits
are not included in the compressor vent control system cost analysis
5-5
-------
on the assumption that recovered emissions would be flared. However,
the recovered emissions could be routed to a process heater resulting
in a credit for the captured emissions at their fuel value.
The type compressor is important because reciprocating compressors
may require additional control equipment to ensure the safety of the
vent.system. Also, reciprocating compressors would require instrumentation
for the purge gas system. Hence, the control costs for a reciprocating
compressor are treated separately from a centrifugal compressor.
The vent control system relies upon the venting of captured
emissions to a VOC control device; therefore, in the absence of an
existing control device, additional costs would be incurred. Further,
the individual compressor control costs are dependent upon the number
of compressors per plant due to fixed and variable costs associated
with the vent control system. Capital and annualized costs of the
compressor vent control system are presented in Table 5-11.
The cost effectiveness of the vent control system is dependent
upon the factors previously discussed plus the service a compressor is
in. The compressor emission factor presented in Table 2-1 is based on
compressors in natural gas liquids (NGL) and wet gas service. Individually,
these emission factors are 0.7 Mg/yr for wet gas service compressor
seals and 5.5 Mg/yr for natural gas liquids service compressor seals.
Table 5-11 also presents the cost effectiveness of compressor vent
control systems under the scenarios discussed.
5-6
-------
Table 5-1. CAPITAL COST DATA (June 1980 dollars)
1. Monitoring Instruments
2 instruments (Foxbgro OVA-108)
@ $4,600/instrument
Total cost is-$9,200/plant
2. Caps for Open-Ended Lines
Based on cost for 5.1 cm screw-on gate valve, rated at 17.6 kg/cm2
(250 psi) water, oil, gas (w.o.g.) pressure. June 1981 cost is $46.50b,
June 1980 cost is 8 percent lessc at $43. Retrofit installation =
1 hour at $18/hour . Total cost is $61/1ine.
3. Compressor Seal Vent Control System
A. Centrifugal Compressor Seal Piping6
5m 2.5cm pipe @$2.82/m
5m 5.1cm pipe @ $6.50/m
1 5.1cm x 2.5cm tees @ $8.16
1 2.5gn block valves @ $24.63
1 2.5cm elbows @ $6.22
1 pressure alarm @ 9.90
Total manifold piping
$
14.10
32.50
8.16
24.63
6.22
9.90
$ 96
Labor
10m of pipe
3Om/hr/crew
1.08 crew hrs. x
0.33 hr for installation
0.25 hr for set-up/breakdown
0.5 hr for fabrication
1.08 hours/crew
3 men
crew
x $18-.00/hr =
B.
Reciprocating Compressor Seals
COSTS FOR EACH COMPRESSOR SEAL
Incremental Cost for Double Distance Piece
$ .58
Subtotal
Contingency
Total
Contingency
Total
$ 154
15
$ 170
$2,500
250
$2.750
5-7
-------
Table 5-1. CAPITAL COST DATA (June 1980 dollars)
(continued)
Distance Piece Piping
Material
2.5cm piping -
31m @ $2.82/m • $ 90
2.5cm check valve -
1 6 $80 ' 80
2.5cm block valve -
2 @ $25 50
Misc. Flanges, Fittings,
etc. 160
Labor
380
620
Subtotal
Contingency
Total
$1,000
100
$1.100
FIXED COSTS
Instrumentation for Purge Gas
Material
$ 650
350
Oil Seal Pot
Supply Regulator
2.5cm Block Valve -
2 @ $25
2.5cm Piping -
8m @ $2.82/m
Misc. Flanges, Fittings,
Etc.
Labor
50
20
.160
$1,230
$ 550
Subtotal
Contingency
Total
$1,780
178
$1,960
C. Flare
Cost of Flare
Contingency
Total
$6,670
667
$7,340
5-8
-------
Table 5-1. CAPITAL COST DATA (June 1980 dollars)
(concluded) . '
D. Piping to Flare
Material
Inlet line from
Compressor to Flare -
100m of 5.1cm pipe
@ $6.50/m
Ruptured disk and
holder
Misc. Flares, Fittings
Misc. Costs for
Pipe Supports
Labor
$650
130
370
750
$1,900
$1,450
Subtotal
Contingency
Total
$3,350
335
a - -
One instrument used as a spare.
Reference 2.
$3,700
—
Cost is based on Reference 1,
Referenced
O
Reference 14.
Reference 8.
-.
iminary). References
5-9
-------
t—
o
o-
UJ
O
1
O
CO
— J
*
Cvl
lO
\u
jQ
fl3
-
t- in
•r- t_0
o
c o
U O O Q-
J OLI — Q.
V 0
a. o
4- in t—
0 4-> CO
c: -a -»->
E- CO O C
E a. t- "o
3 E CO
zoo.
{J
o
ca
•a:
CO
o.
•)->
CO
u
J_
3
o
OO
Z> 0 0 0
-1 CO rH
O OO r-H
3 O O CM
n T— i oo
— ( t— 1
-i o oo <— i
r> oo i— i
CO 4— 4—
•0 O O IO
'"i "*
H^
V^3 ^^ VO
o r^* o^ vo
<=t-
oo co o
UO CM CM CM
oo
1—4
10 «a- 10
i — er> 10
UO 0 0 0
•*
co cr> r^. oo
r-l ^-4 «* 00
0000
O •* CM CM
UO
«v
CM
O CM IO IO
JD ^H
O >=*• CM CM
r>
Cvl
01
to
in co
CO OO
i — (— in
to o i—
> Ol "3
Ol CO
01 4- CO OO
CO CO t_
> .r- Q. Q.
r— r— E E
n3 CO O 3
> a: o Q-
•3
o
3
— I
O
i— 1
_l
1 —
O
•
E 0
£ -a o
f~ ) (O
O « IB
O* T- CO
^H
co c
l_ Q. « -f-
CO CL «
CO 3 O CO
£_ 0 r-l _<=
CO 4->
i_ r^ u 3
O T-t C O
0 0 t-
4J CO >>
,— c co
(0 -r- 01 "0
3 £_ CO
cr ••- 4J
CO 3 • O
cr t- co
trt CO T- t-
"c Q. O
co « co o
C CO £-
o ceo
Q. -r- t_ t-
C r— HI fO
0 1 Q.
O C t/)
O (~ J=
e: • T- co
CO CO
X -i- ^^ •!- "3
4J CO i— CO
c s_ s- -a •
co co co cr>
c X g" •» E
0 -U CO 3 CO
o. 01 c m u
o "3 o "— a:
^ t_ -o ••—
CU £- CO O)
ja co a. t- in •
E _Q «i— -4-> O1
3 E LO
C 3 C~~ Q. CO C
u c: c_ CO E
II 0 £- CO
(/) CO •!— t-
•^ 10 -a j«i 3 «i—
« c- at ITS cr 3
co 3 01 co co cr
"~ "f ja "~ t-
i— t- CO >,
to o co .c +-> t-
• .f— _Q C71 4-J CO O
CM -t-5 T3 "3 4— .0
oo c co o in <— •
•r- i- > CTl
Ol -i— to 4-> 4— i—
r— 4- CO c O «3 CO
J2 O Q. T3 CO CO
rtJ CO CO O CO O C
i — t_ C. 4-> C- 01-r- CO
CO JC CO 3 +-> S-
E -Q -^ O1CL rt3'i- CO
0 E t3 -t- 0 T3 4-
c_ rs co coun co~a co
U_ "Z. 1 3CM CQ'O OL
It3 J2 O T3 CO 4-
5-10
-------
Table 5-3. INITIAL LEAK REPAIR COSTS
(June 1980 dollars)
Initial Repair Costs
Initial Annualized Repair
Tnctc -Fnv> Mr,/^«l ii_-j.-D .
For Model Units
Costs for Model Units
1,200 4,200
'-initial Lea. Repair. Cost -
where n = 10 years and 1 » 10 percent. Therefore, the CRF = 0 163
5-11
-------
Table 5-4. BASIS FOR ANNUALIZED COST ESTIMATES
1. Capital recovery factor for capital charges
o Caps on open-ended lines
o Monitoring instruments
2. Annual maintenance charges
o Caps on open-ended lines
o Monitoring instruments
o Replacement pump seals
3. Annual miscellaneous charges
(taxes, insurance, administration)
o Caps on open-ended lines
o Monitoring instruments
4. Labor charges
5. Administrative and support costs
for implementing leak detection
and repair
6. Annualized charge for initial
leak repairs
0.163 x capital3
0.23 x capitalb
0.05 x capital0
$3,000d
$1406
0.04 x capital0
0.04 x capital0
$18/hourf
0.40 x (monitoring +
repair labor)0
(estimated number of
leaking components per
model unit x repair time) x
$18/hrf x 1.4° x 0.1639
eC°VoryNonmethane-nonethane hydrocarbons (VOC) JJ92/M9h
o Methane-ethane $61/Mg1
aTen year life, ten percent interest. Reference 11.
Six year life, ten percent interest. Reference 11.
Reference 11.
Includes materials and labor for maintenance and calibration.
Reference 11. Cost index = 242.7/209.1. References 12 and 13.
Q
Reference 14.
Trora Table 5-2. Includes wages plus 40 percent for labor-related
administrative and overhead costs. Cost (June 1980). Reference 4.
Initial leak repair amortized for ten years at ten percent interest.
hBased on LPG price of 40<{:/gallon for June 1980 and specific gravity of
0.55. References 15 and 16.
1Based on natural gas price of k$1.46/Mcf for June 1980 and assumed
composition of 80% methane and 20% ethane at standard temperature and
pressures. Reference 17.
5-12
-------
i. l/l
•I- U4-
tu 3 -_
C- O
I
LO
I
LO
•
o *•>
;_ —
si
— i_ oo
> irt "fl
i— - «e
CT c
c •—
— • ,-= 1-
-5 a;
P- S
o -=: =£
a. yi *J
3 -*— VI
OJ — '
X1 O — r—
—• -r- VI
5-13
-------
Table 5-6. ANNUAL LEAK DETECTION AND REPAIR COSTS0
(June 1980 dollars)
Leak Detection
Source type
Valves
Relief Valves
Compressor Seals
Pump Seals
TOTAL
For
A -
590
77
23
41
730
Model Uni
B
1,800
230
72
130
2,200
Costs
ts
C
5,900
770
230
390
7,300
Repair Costs
For
A
940
Ob
270
230
1,400
Model
B
2,900
Ob
790
680
4,400
Units
C
9,400
Ob
2,700
2,300
14,000
aFrom Table 5-5, Annual Leak Detection and Repair Labor Requirements for RACT,
Cost = hours x $18.00 per hour.
bBecause of safety requirements safety relief valve leaks are repaired by
routine maintenance at no additional cost. Reference 9.
5-14
-------
ca
H—
^^
«C
^^
Q_
UJ
Q
O
o;
o
u_
I—
Q
UJ
Qi
^— ^
co
o
0
^
UJ
>
o
0
1 1 1
a;
o
o
en
Q.
U_
O
o
t—
•
'*^*N
-b=>
*^»«
OJ
3
O
O
•Q.
•a
O)
5-
O
O
^
^J
O
1 —
03
c
o
•r-
CO
UJ
O)
a>
o
QJ
Q£
O)
(8
03
1
03
£r
-!=
a>
s:
o
^v^
a>
Q.
>^
1—
-M
O)
|
§
CJ O O O O
O CO O 0 ^H
i— I CTi CM LO CO -
n «* «*
i— 1 «^- CO-
t—l
o o o o •=*• '
O CTi —1 CD CM
co r~— <*o CM
oo" r-T
o o o o o
O CT> O O CTl
CO i— I ^O OO CM
r^ co CM
o
^*
CM CO O O O
VO .-t f-H CM
CTI m
OO O CT» CO i— 1
CO ^-« ^l-
i— -a s_ in
(Q 03 O r—
> "O (O
C (/» OJ
to <+- a> a) to
a> 2> i i_
> -i- C Q. Q.
i— r— OSS
its 03 a. 03
=» ce: o «_j a_
o
o
o
CM
o
o
o
o
CO
r— 1
un
o
f— 1
S o>>—
O in JD
S- (O (O
u_ cat—
-------
oo
1—
1— 1
UJ
o;
i—
oo
o
o
^
UJ
••11^
o
O
a:
H-
Z
•a:
i
D_
_J
UJ
a
•^
c
00
1
m
CD
^
ia
I—
"\
»^^
O)
3
"id
-P
o
3
•a
o
Q-
-a
-o
CD
s_
O
O
CD
S-
4-5
ns
O i—
1— f*\
ca
,
.
.0
0)
0
c
4_>
0
o
r-T
tn
CD
r^
fs
(T3
1—
O)
d)
oo
-^3
5-16
-------
Table 5-9. ANNUALIZED CONTROL COSTS FOR MODEL UNITS
(thousands of June 1980 dollars)
Cost Item
Annual ized Capital Costs
A. Control Equipment3
1. Monitoring Instrument
2. Caps for Open-Ended Lines
B. Initial Repairs5
Operating Costs0
A. Maintenance Charges
1. Monitoring Instrument
2. Caps for Open-Ended Lines
3. Replacement Pump Seals
B. Miscellaneous Charges (taxes
insurance, administration)
1. Monitoring Instrument
2. Caps for Open-Ended Lines
C. Labor Charges8
1. Monitoring Labor
2. Leak Repair Labor
3. Administrative and Support •
Total Annual ized Cost Before Credit
Recovery Credits^
Net Annual ized Cost
VOC Emission Reduction (Mg/yr)9
Cost-Effectiveness
($/Mg VOC Emission Reduction)
Model Plant
A
2.1
0.51
0.42
•
3.0
0.16
0.11
0.37
0.12
0.73
1.4
0.85
9.8
(6.5)
3.3
23
140
B
2.1
1.5
1.2
3.0
0.46
0.34
0.37
0.37
2.2
4.4
2.6
18
(20) .
(2)
72
(28)
C
2 1
t- « A
5.1
4.2
3 0
%J • w
1.6
1.1
0.37
1.2
7.3
14
8.5
48
(65)
(17)
230
(74)
(xx) = cost savings
aFrom Tables 5-1 and 5-4.
u •
From Table 5-3.
Basis for cost estimates presented in Table 5-4.
Calculated as: Estimated number of pump seal leaks per year.(Table 5-5)
X *p A TU . .
e '
From Table 5-6.
From Table 5-8.
Table 4-4.
5-17
-------
o
«C
cz.
cz
UJ
o
CO
<£.
Q.
UJ
a
C£
O
UJ
ex.
UJ
o
a.
O T3
>- O
CD CO
CT>
CO r—
CO
UJ OJ
2= C
UJ =3
O
UJ
u_
u_
UJ
CO
o
o
u_
o
CO
UJ
X
UJ
to
(U
-Q
4J a
tn >
o ••-
£
u-
Q)
O -C '
•-£
ai i.
o ~~
ai •—•
u
TJ
3 -O
C U
T3 I- --»
a; u t.
N >>
T- 01 .
i— £. v>-
re o *—•
4J tA
t- u .
Q. tn •<
«a O *
U o
o c
s- c
QJ O
A Q.
O O C O
CO O to r-;
to --< —•• eo
o
o
o
o
co
o
CO
o o
O —i
in m
o
CM
rr>
0
o
0 0
O CO
tn
O
>
Oil—
aj
tn X
sl
§0.
re
•u tn
tn «^-
3
U- S-
O 4->
tn
HI C
tn t-
re
f CU
a; tn
in o
re o
i— !- M
•"-> 3
sg
U (O
-O p-
(D ro
M 4->
••- O
"8
CM t. CO
1 1 •«-
IT) Q. «r-
OJ
fC
U_ Q3
"O OJ '
Q.CD
(0 V*
O
tn
CD tn
§4->
c
L_ CU
tn 5
c s-
tn
ai c
c ••—
'T~
U CU
5-18
-------
o
Q£
z.
O
J_
•z.
uJ
a:
O
to
to
LU
a.
^?
o m
Q; |_
£3
to o_
LU _l
•Z. LU
LU O
Si
go:
u_
LU 2:
1 i ' i
t— 1—
to to
o >-
cj to
O
4J > en
V» 'I- ar
0 4-> ^.
O O **
LU
C C
O O —
V) (J ^>i
i- 3 en
i!S "S-5-
os
c.
*(OT3 "^
3 4JO
C VI O
COO
< 0 «
(^
•M 4JO
•»- v» O
0.0 «
« o •— •
o *•»•
0
V*
VI 0)
(U u
u .—
t£
O 01
o
.Q
O, ai
tu i-
o o a.
o
V) (O
w
^J 4^ r j^ t ij ]
tQ
sz sz Sz »z
V) v)
0) 0 OJ 0
>> = >,=
*
_
i_ c
ns - ^
3 u
*- o
•w .£•
c o
OJ CU
o ae
t^
ia
V
c
10
cn
c
_§
VI
n-
V)
3^
C
1
'c
•o
0)
(Q
£
a
VI
91
U
E
o
: cn
c
•**
. 10
o '
: u-
: CL
O
£
0)
i
•2
a
VI
en
c
1
^t
1
•o
£
<0
o
0) •
i-
o
u
l~
§,
,~*
"C o •
C 4->
0) 0
CD
_^
4-1
O
1
c
QJ
1!
tn
«
•^
^
c.
«3
S-
d)
u
"5
"O
o
u .
4J
O)
US T3
<*- •—
o o
« c
o o
o o
If
V)
c
II
d'a.
"*"**- ^3"
0) 1
O V> LO
> VI 0)
0) O ' r~
•6 o ja
i-^ S ' "
'o t/}
•II 1 s
+J U (T5
(/) 4J OJ
M
OJ E M Wl
^ o o o
3 *J O CJ
jO U -0W
-J
.e
o>
i
o.
p
»-
0
c
c
o
o
o
&
S.1-"
(O
L.
a>
c
J=
tn
s
•a
j^
V)
§
E
i>
»*M-
i
^
CO
E
J!
o
^
a-
~o
"3
•2"
^-
*
•
^
V)
OJ
3
L.
4-
OJ
1
CL
ill.
W
•a
a-
•a
0)
X
•g
Ik-
o
f
v\
c
o
o
i
"O
3
P—
V)
S,
2
3
Z
5-19
-------
5.6 REFERENCES
1. Telephone conversation. Michael Alexander, TRW, with Ms. M. Fecci
of Analabs/Foxboro. March 23, 1982. Price of 'Century Systems
OVA-108 in July 1980.
2. Telephone conversation. Michael Alexander, TRW, with Mr. Harris
of Dillon Supply, Durham, N.C. June 17, 1981. Price of gate
valves.
3. Economic Indicators. Chemical Engineering. Vol. 88 #12. June 15,
1981. p. 7.
4. Letter with attachments from Texas Chemical Council to Walt Barber,
U.S. EPA. June 30, 1980.
5. Telephone conversation. Michael Alexander, TRW, with Danny Keith,
Dillon Supply Co., Raleigh, N.C. June 15, 1981. Costs of valves,
pipes, and fittings.
6. Telephone conversation. Tom Norwood, Pacific Environmental
Services, Inc., with P. Marthinetti, Ingersoll Rand. Distance
Piece Price, December
7. McMahon, Leonard A. ,'1981 Dodge Guide. Annual Edition No. 13,
McGraw-Hill Publishing Co.
8. Memorandum. K.C. Hustvedt, EPA, to J.F. Durham, EPA. Cost basis
for compressor vent control system. February 23, 1983.
9 Letter with attachments from J.M. Johnson, Exxon Company, U.S.A.,
to Robert T. Walsh, U.S. EPA. July 28, 1977.
10. Erikson, D.G., and V. Kalcevic. Hydroscience, Inc. Organic
Chemicals Manufacturing Industry. Volume III, Report No. 2: Fugitive
Emissions EPA 450/3-80-025. September 1980.
11. Environmental Protection Agency. Control of Volatile Organic
Compounds Leaks from Petroleum Refinery Equipment. EPA-450/2-78-036,
OAQPS No. 1.2-111. June 1978.
12. Economic Indicators. Chemical Engineering, Vol. 86 #2. January 15,
1979.
13. Economic Indicators. Chemical Engineering, Vol. 87 #19. September 22,
1980.
14. Fugitive Emission Sources of Organic Compounds - Additional
Information on Emissions, Emission Reductions, and Costs.
EPA 450/3-82-010, April 1982.
5-20
-------
15.
16.
17.
Telephone conversation. T. Mannings, TRW, with Editor, Oilqram
News. February 25, 1981. Price of LPG on June 16, 1980.
Nelson, W.L. Petroleum Refinery Engineering, McGraw-Hill Book
Company, Inc. New York 1958, p. 32.
DOE Monthly Energy Review. January 1981. DOE/EIA-0035(81/01).
p. 88.
5-21
-------
-------
APPENDIX A
EMISSION SOURCE TEST DATA
A-l
-------
APPENDIX A
EMISSION SOURCE TEST DATA
The purpose of Appendix A is to summarize the fugitive emission
test data that have been collected at six natural gas/gasoline processing
plants (see Table A-l) by EPA and industry. Two gas plants were tested
under contract to the American Petroleum Institute (API), and four gas
plants were tested under contract to EPA. All six gas plants were
screened for fugitive emissions using either portable hydrocarbon detection
instruments, soap solution, or both. Instrument screening (using EPA's
proposed Method 21) was performed at all four of the EPA-tested plants
(Plants 3, 4, 5, and 6). The instruments were calibrated with methane.
Soap screening (using the method described in Reference 1) was performed
at the two API-tested plants and at three of the EPA-tested plants.
Selected components were measured for mass emissions at both of the
API-tested plants (Plants 1 and 2) and at two of the EPA-tested plants
(Plants 5 and 6). These mass emission measurements were used in development
of emission factors for gas plant fugitives, which are presented in
Table 2-1. A study of maintenance effectiveness at production field
tank batteries was also performed by API. These data are discussed in
Section A.2.
A.I PLANT DESCRIPTION AND TEST RESULTS
One API-tested gas plant was of the refrigerated absorption type,
and the other was a cryogenic plant. Descriptions and schematics of the
plants are provided in Reference 1. Of the four EPA-tested plants, the
first tested was a solid bed adsorption type (Reference 2). Natural gas
liquids are removed by adsorption onto silica gel, then stripped from
the bed with hot regeneration gas and condensed out for sales. There
were three adsorption units, of which only one was operating. This unit
had a capacity of 60 MMSCFD (million standard cubic feet per day), and
A-2
-------
was operating between 33 and 55 MMSCFD during the testing period. The
second unit was shut down and depressurized, and therefore not tested.
The third unit was also not operating, but it was under natural gas
pressure and was tested.
The second EPA-tested plant was of the cryogenic type (Reference 3).
Feed gas to the plant is compressed and then chilled. Natural gas
liquids are condensed out and split into two streams: ethane/propane
and butane-plus. The cryogenic plant was operating at its rated capacity
of 30 MMSCFD.
The third EPA-tested plant was of the refrigerated absorption type
(Reference 4). There were three absorption systems for removal of
natural gas liquids. The liquids were combined and sent to a single
fractionation train. The fractionation train separated the liquids into
ethane, propane, iso-butane, butane, and debutanized natural gasoline.
Testing was performed on the fractionation train and on the largest
absorption system. The absorption system that was tested was operating
at 450 MMSCFD, near its capacity of 500 MMSCFD.
The fourth EPA-tested plant was also of the refrigerated absorption
type (Reference 5). There were two parallel absorption trains, and one
fractionation train. Natural gas liquids were fractionated into ethane/
propane, propane, iso-butane, butane, and debutanized natural gasoline
streams. The plant was operating at approximately 450 MMSCFD, about
half of its rated capacity of 800 MMSCFD.
A summary of the instrument screening data collected at the four
EPA-tested plants is presented in Table A-2. A summary of the soap
screening data collected at the two API-tested plants and at all of the
EPA-tested plants is presented in Table A-3. (Only a very small amount
of soap screening data were collected at Plant 6). The instrument
screening data are tabulated for each plant, showing the number of' each
type of component tested and the percent emitting. The soap screening
data are not tabulated for each plant but are instead summarized by soap
score. A complete tabulation of the soap screening data by plant and by
soap score is provided in Reference 6.
A.2 INDUSTRY VALVE MAINTENANCE STUDY
The API study that developed the gas plant data presented in Section A.I
also included a study of maintenance. Gate valves in gas and condensate-
A-3
-------
service in oil and gas production field tank batteries were studied.
The sources were monitored with soap scoring at intervals over a 9-month
period. The results of an analysis of this data show that monthly leak
occurrence was 1.3 percent, monthly leak recurrence was 1.6 percent, and
leak repair effectiveness was 100 percent.7 These results compare
favorably with the 1.3 percent monthly leak occurrence and recurrence
and 90 percent repair effectiveness used to analyze leak detection and
repair control effectiveness. Maintenance was performed on a portion of
the valves studied. The industry study results were not specifically
used here, however, because (1) the data were gathered in tank batteries
which, based on API data, appear to have different leak characteristics,
(2) very few valves were studied (25 total data points), and (3) a
soap score value of 3 was used to define a leak rather than a meter
reading of 10,000 ppm.
A-4
-------
Table A-l. GAS PLANTS TESTED FOR FUGITIVE EMISSIONS9
Plant
No.
Data
collection
sponsor
Plant process
type :
Principal screening
method(s) used
1
2
3
4
5
6
API
API
EPA
EPA
EPA
EPA
Refrigerated Absorption
Cryogenic
Adsorpti.on
Cryogenic
Refrigerated Absorption
Refrigerated Absorption
Soaping
Soaping
Instrument, Soaping
Instrument, Soaping
Instrument, Soaping
Instrument
Reference 6.
Less than 50 components were soap screened at plant #6.
A-5
-------
1-1 10 O •-<
cj n cJ| n"
*-< co
CO VO
CO
CO
3
88
tlo
•a- n CM CM
to
tO
d:
o.
Ul
cc.
o
u.
z
UJ
CO
UJ
to
OU
0)
(0
I
c
o
<§•
.•s
S3
§c
ss
-si
8
• OJ
5 £§
e
ce
1 o o
i i"*"!
S j C- O
p2 ' ~*
-* *-t CSJ
A-6
-------
to
oo
T3
f- W O
»— I—
CD
LU
(X.
oo
a.
CO
1
9}
J y_ 1
ui { O fl
CL 1.
"3 O
O U
l/l OO
A-7
-------
A.3 REFERENCES FOR APPENDIX A
1 Eaton, W. S., et al. Rockwell Corporation. Fugitive Hydrocarbon
Emissions from Petroleum Production Operations. API Publication
No. 4322. March 1980.
2 Harris, G. E., Radian Corporation. Fugitive VOC Testing at Houston
* Oil and Minerals Smith Point Plant. U.S. EPA, ESED/EMB Report
No. 80-OSP-l. October 1981.
3. Harris, 6. E., Radian Corporation. Fugitive VOC Testing at the
Amoco Hastings Gas Plant. U.S. EPA, ESED/EMB Report No. 80-OSP-2.
July 1981.
4 Harris, G. E., Radian Corporation. Fugitive VOC Testing at the
Texaco Paradis Gas Plant, Volume I and II. U.S. EPA, ESED/EMB
Report No. 81-OSP-7. July 1981.
5. Harris, G. E., Radian Corporation. Fugitive Test Report at the
Gulf Venice Gas Plant, Volume I and II. U.S. EPA, ESED/EMB Report
No. 80-OSP-8. September 1981.
6. DuBose, D. A., J. I. Steinmetz, and G. E. Harris, Radian Corporation,
Data Analysis Report: Emission Factors and Leak Frequencies for
Fittings in Gas Plant. Final Report. U.S. EPA, ESED/EMB Report
No. 80-FOL-l. July 1982.
7. Memorandum, Hustvedt, K.C., EPA to Durham, J.F., EPA "API/Rockwell
Maintenance Data". December 9, 1982.
A-8
-------
APPENDIX B
MODEL PLANTS
B-l
-------
APPENDIX B
MODEL PLANTS
The purpose of this appendix is to present model plants. The model
plants were selected to represent the range of processing complexity in
the industry. They provide a basis for determining environmental and
cost impacts of reasonably available control technology (RACT).
B.I DEVELOPMENT OF MODEL PLANTS
There are a number of different process methods used at gas plants:
absorption, refrigerated absorption, refrigeration, compression, adsorption,
cryogenic - Joule-Thompson, and cryogenic-expander.1 Process conditions
are expected to vary widely between plants using these different methods.
However, available data show that fugitive emissions are proportional to
the number of potential sources, and are not related to capacity, throughput,
age, temperature, or pressure.2 Therefore, model plants defined for
this analysis represent different levels of process complexity (number
of fugitive emission sources), rather-than different process methods.
In order to estimate emissions, control costs, and environmental
impacts on a plant specific basis, three model plants were developed.
The number of components for each model plant is derived from actual
component inventories performed at four gas plants. Two of the plants
were inventoried during EPA testing, and two were inventoried during
testing by Rockwell International under contract to the American Petroleum
Institute. The model plants are based on four rather than on all six of
'the plants presented in Appendix A because two of the plant visits did
not obtain information on vessel or equipment inventories. Nevertheless,
the four plants for which vessel and equipment inventory data were
obtained are representative of the range of plant complexity found in
the natural gas processing industry.
B-2
-------
Complexity of gas plants can be indexed by means of calculating
ratios of component populations to a more easily counted population.
For gas plants, number of vessels appears to be best suited to this
need. Example types of equipment included and excluded in vessel
inventories are listed in Table B-l. The vessel inventories for the
industry-tested gas plants are taken from the site diagrams and des-
criptions provided in the API/Rockwell report,5 and the vessel inventories
from the EPA-tested plants were performed during the testing. These
vessel inventories and the component inventories are shown in Table 8-2.
Table B-3 shows the ratios of numbers of components to numbers of vessels
at the four gas plants. The mean and standard deviation of the,four
ratios is also shown in Table B-3.
Three model plants have been developed using the average ratios of
components to vessels. The number of vessels in the model gas plants
are 10, 30, and 100. This range in number of vessels is based on the
vessel inventories shown in Table B-2. The low end of the range, 10
vessels, is approximately equivalent to the number of vessels that are
accounted for in one of the three process trains at the EPA-tested plant
A. It is assumed that there are existing gas plants with a similar
configuration to the EPA-tested plant A, that have only one process
train. The high end of the range, 100 vessels, is slightly larger than
the number of vessels at the industry-tested plant A. Since this was s
the largest of the plants tested, it appears reasonable to use this as a
guide in calculating the number of components at the largest model
plant. The middle model plant has 30 vessels. This is approximately
the same number of vessels as at three of the four plants tested, and
appears that it may be representative of a common gas plant size. The
three model plants and their respective number of components are shown
in Table B-4.
B-3
-------
1^
j"^ | f 1
O 2EZ
rD a.
_J 0
X UJ
HI ^*
UJ
O O
^C 1
;jp»
o «t
UJ —1
Q Ou
i |
O UJ
2S a
as o;
UJ tO
*5T" 1 1_
•£• I-1—
a.
I— I CO
•=> UJ
cr>-i
UJ Q£
O
u- tr
02=
CO >
UJ as
Q. )— 1
1 - 1
^UJ
UJ CO
_J CO
Q. UJ
x as
UJ >-H
,
4 -
I
CO
0)
3
ro
to
"O e^t
cu E
3 Q-
p— •
^i **
x to
UJ t_
o
to
to
cu
c»
<_>
4
to
•r-
c
=>
o
o.
Included
rption/Desor
Absorbers
o
^Q TC
•P C
cu cu
E
• o.
C 3
o cr
•e- CU
tO O i—
t= CU O
CU C (-
4-5 C 4->
to O c
>» o o
to o
A
t_ CO "
cu cu to
^3 en Q}
ro c en
CU (O 3
to .E i — (O
E ^^ *^~ en
CU TO
[ ^ ^~ ^ A
to O to to
>>t>_ CU «-
CO -r- > CU
C r— 4->
cn rei ITS cu
5.
«r— • • •
Q- fO J3 O
CM
to
c_ c_
to o cu
C- +J N 4-
cu ns -i- cu
.0 5- r— Q.
.a -a -i- o.
3 >,-Q -t-
t_ J= (0 C-
O CU 4-» 4->
co a co to
....
.a o T3 cu
cu
cn
(O
c_
o
to
c.
cu
•4^
C
c:
o
•r—
a.
o
to
-a
=£
.
evj
cn
c:
c:
cu
4-)
CU
cu
to
c"
4-> CO
•r- CU
•3. to
to
-o cu
cu o
4J 0
("Q i_
•i- a.
u
0 >,
to t-
to cu
(0 >
O
4-> 0
C CU
CU C-
E
("^ £..
-i- 13
23 *+—
a; 3
(/>
>n
c t-
.,- ^^ M- tO
4_> C C S-
O ITS Z3 O)
ITS t- £- 1 — en
c_ cu C- cu ~^» en c
U-NCUN E C lO to
-^ .,_ |sj .r- 3 -t- -C tO S-
CC-r-Ct-t- i — tOOt-CU
OiOCfOCUQ OtO^-XCUiotO
.,-_c:4-J^ZO4->-Cr— C_3CU^- -I— CUCU
n3CU4-3£-T-tOr— — 4-)r— 4->OT3i—
i — cu cu cu o-i — 4-> ccujrcucuoo
to rei
•51-
co
,-
C-
. o
C- 4->
O * 4->
tO 4-> i — 3
-*J IT5 3 O
c c_ cu E -^
(O
-------
C£
LU
CO
y
jf
-*^7-
Q
^£
LU
O
l— l
a:
LU
to
^y ^O
CD 1—
03 Z
£c2
O Q.
O
ee: co
C£
z ID
>-» O
U_
tO
1— H-
•:=.<£.
LU
^ tO
CD —I
Q_ LU
s to
CD to
CO LU
^»
u_
CD U_
CD
OS
LU
CO
*^
— ^
^^
•
CM
1
CO
CO
1—
(Q
Era
to
. ' O-
-O
CO
1—
^
s_
4-5 <
3
-o
E
1— i
(O 03
CO
to
CL.
-»
CO
CO
CO
«=c«=c
Q.
LU '
LO ' CM r-~ CO CD CO O
CM ^o r~- oo
r-^ ^n o
CO
O O 0 CT^ LO CM O
CT> CO CM to CO CO (~^.
CO l£3 00
** «
CO LO
l-H
co 5 in **" °° ^ °
LO """ ^ •" 5
A
1—t
J^ § "fe s . ° ^- us
LO LO
A
r—
CO
E
O
!|j
O
-o
i— -a s- co E
"3 CO O i— to
> "O co to
to E co CO co
•— co 4- uj dj ^ ^
<1> CO CO 1 S_ 05
to > •"- E Q. Q. E
co 1o "cu Q. § § ,2
> > OH CD C_5 Q- Li_
a
00
CO
o
E
CO
CO
<4-
co
Qi
03
0
J=
CO
E "~
•i- T3
4-> CO
CO 4J
LU E
3
O
• 0
CO CO
•i— 4J
S- E •
O CO 4->
4-J E -r-
E O E
CO 0. 3
^* F^
E O S-
••- o o;
T3 4- i —
E O (O
(0 E
5- CO
-0 CO
CO JD CO
4-> E -E
CO 3 4->
CO E
4-> E
CO •!-
CO J=
S- 4J TD
CO CD
S M- 4J
O E
4-» 3
£= E 0
ta 3 u
i — to
CL co
CO 4->
CO JE 'E
-E 4-> CO
4J E
E O
4-5 0 CL
to E
-a o
to CO O
4-5 to
•i- OS <4-
E jQ 0 •
3
CO J-
s= s_ cu
O O J2
•i- M- S
4-> CO 3
Q. S- E
i- co
O JE O)
to 4-> J=
•a 4->
E Q.
• Q
O -r-
CO M- E
0 O O 3
E 3
CO 4-5 S_ S-
5- CO CO
CO >,-Q a»
4- r- E S-
CO E 3 to
CU O E i—
-Q cj
B-5
-------
1 1
ro
UJ
CO
CO
UJ
>
u_
o
CO
OS
UJ
m
n
o
H-
co
1—
UJ
— ^
1
u_
0
CO
OS
UJ
CD
5
u_
co
o
5
•
CO
1
CD
CU
(O
l_
•a
ro •
(O
CO C
0)
en
co
s_
OJ
— 1
1) 4—
3 0
0
4J
OS
CO
*£.
CO
cc
en i-i o CM csi r-»
en o co o o cr>
in «=i- . vo -cvj cvi «*
if) o «3" o o r^
CNJ . cr»
in co o i — <>J
o o ^o o o r—
CO CM
^™
0 CVJ * * * CO
r*. o r~ o o o
CO t^~
O ""d" i— CO «S" O
CO CO CVJ O O CO
r~
«=j- in o o o "*
tO O CNl , O O CT»
^" "Ssl
IO 00
CU i —
C ro
00 -i- 0)
ai i co
> T3 OO
, — T3 t- OO C C
rO CU Or— ro O
> "O OO ro T-
C 00 CD oo -M
to 't- UJ CU CO CD O
ai cu i . s_ cn cu
> •!- £1 CL CL != E
r- r- OJ S S rO S=
ro CU Q. O 3 i — O
>• o; to co o_ u_ o
CNJ
1
ca
cu
r^l
ro
^
•r"
"O
O)
cu
00
CD
a.
rO
+-^
ro
-a
c
o
-a
cu
to
rO
CO
rO
B-6
-------
Table B-4. FUGITIVE VOC EMISSION SOURCES FOR
THREE MODEL GAS PROCESSING PLANTS
*" Number of components
Component type3
Valves
Relief Valves
Open-Ended Lines
Compressor Seals
Pump Seals
Flanges and Connections
Model plant
(1.0 vessels)
250,
4;
, • . 50 . .
2
2
1,000
Model plant
B .
(30 vessels)
. 750
12
,.150
6
6
3,000
Model plant
C
(100 vessels)
2,500,
40
500
20
20
10,000
Number of Components based on average ratios presented in Table B-3.
B-7
-------
B.2 REFERENCES
1. Cantrell, A. Worldwide Gas Processing. Oil and Gas Journal,
July 14, 1980. p, 88.
2. Assessment of Atmospheric Emissions from Petroleum Refining,
Volume 3, Appendix B. EPA 600/2-80-075c, April 1980. Pages
266 and 280.
3. Hustvedt, K.C., memo to James F. Durham, Chief, Petroleum Section,
OAQPS, U.S. EPA. Preliminary Test Data Summaries of EPA Testing
at Houston Oil and Minerals Smith Point Gas Plant and Amoco Production
Hastings Gas Plant. March 19, 1981. <,
4. Eaton, W.S., Rockwell International, letter to D. Markwordt, OAQPS,
U.S. EPA. Component Inventory Data from Two API-Tested Gas Plants.
September 11, 1980.
5. Eaton, W.'S., et al. Fugitive Hydrocarbon Emissions from Petroleum
Production Operations. API Publication No. 4322. March 1980.
B-8
-------
APPENDIX C
PUBLIC COMMENTS
C-l
-------
This appendix presents the public comments received by the EPA
on the draft CT6. Table C-l summarizes the respondents and lists
their corresponding number used to identify commenters in the
Summary of Public Comments and Responses presented in Appendix D.
Table C-l. LIST OF COMMENTERS ON THE
DRAFT CTG FOR NATURAL GAS/GASOLINE
PROCESSING PLANTS
Comment
Number Company
1 ARCO Oil & Gas Company
2 Cities Service Company
3 Columbia Gas System
Service Corporation
4 Southern California
Gas Company
5 Michigan Consolidated
Gas Company
6 Texas Oil & Gas Company
7 Texas Air Control
Board
8 American Petroleum
Institute
9 Michigan Wisconsin
Pipe Line Company
10 Chevron
11 Amoco Production
Company
12 Flour Engineers and
Constructors, Inc.
Commenter
L.E. Bartlett
D.V. Trew
M.J. Atherton
S.E. Kurmas
J.W. Boley
R.R. Wallis
C.T. Sawyer
J.V. Mehta
Date
of Comment
March 9, 1982
March 9, 1982
March 9, 1982
G.M. Gardetta March 9, 1982
March 10, 1982
March 10, 1982
March 10, 1982,
March 11, 1982
March 11, 1982
R.W. Kreutzen March 12, 1982
R.E. Mahaffey March 18, 1982
S.J. Thomson
March 22, 1982
C-2
-------
ARCO Oil and Gas Company
Engineering Department
Post Office Sox 2819
Dallas. Texas 75221
Telephone 214 351 5151
Luther E. Bartiett
Manager
Operations and
March 9, 1982
Mr. Fred. Porter
Emission Standards and Engineering Division (MD-13)
U. S. Environmental Protection Agency
Research Triangle Park, North Carolina 27711
Re: Draft Control Techniques Guideline (CTG) for
'• Le£TL» l°rat±^ °Tnl° ComP°un
-------
Mr. Fred Porter
Emission Standards and Engineering Division (MD-13)
U. S. Environmental Protection Agency
March 9, 1982
Page Two
analysis supporting this guideline suggests control of the fugitive VOC
emissions will have a direct economic benefit to the operations of ARCO
Oil and Gas Company, thus justifying the proposed RACT. We do not
agree with many of the assumptions and, therefore, do not feel the RACT
has been justified. For example, the CTG economic analysis equates
"front-end costs" with "capital costs" (comment 13). Capital cost has
a specific economic definition and does not include operating costs.
In addition, the cost analysis for the labor associated with leak
detection severely underestimates the actual cost (comment 18). A
complete cost estimate must include the front-end set-up costs,
depreciation on the equipment, additional - and otherwise unnecessary -
platforms for each inaccessible source, and maintenance on the VOC
analyzer. Although the conclusion of the draft CTG's economic analysis
is that the oil and gas industry has lost significant revenue from not
controlling fugitive VOC emissions (excepting the smallest plants), we
feel the costs of implementing and maintaining the recommended
practices are much greater than estimated with little if any
improvement in the air quality. Consequently, we believe a net
long-term loss will result from the use of the proposed RACT. This is
of specific concern since the draft CTG, although published as only a
guidance to the states, will serve as the basis for many of the state
regulations.
We appreciate your, consideration of our concern. If it would be
helpful, we would welcome an opportunity to further discuss our
concerns associated with implementing the proposed RACT to control VOC
emissions from natural gas/gasoline processing plants.
Sincerely,
Luther E. Bartlett
C-4
-------
CITIES SERVICE COMPANY
BOX 300
TULSA, OKLAHOMA 74102
March 9,, 1982
Emission Standards & Engineering
Division (MD-13)
Environmental Protection Agency
Research Triangle Park, NC 27711
Attention Mr. Fred Porter
Dear Mr. Porter:
SSSS
able control device and we recommend that the use of this
gas plant compressors
are b.sed
guideline a more meaningful and workable document? ^
.Sincerely,
C-5 "'
DVT - (
D. V..Trew
Manager, Environmental Services
-------
GAS SYSTEM SERVICE CORPORATIOM'
2C MC'.~C ——^INI =OAC
.•-.G-C-.. z=:_A.-.i=E -sac-
March 9 , 1982
Emissions Standards and Engineering Division (MD-13)
Environmental Protection Agency
Research Triangle Park, North Carolina 27711
Dear Sirs:
Re: .Control Techniques Guideline Document;
Equipment Leaks From Natural Gas/Gasoline
Processing Plants "_ "'
Columbia Gas System Service Corporation on behalf of
Columbia Gas System (Columbia) herewith submits comments on the
above draft control techniques guideline (CTG), the release of
which was announced in the Fe'deral Register of January 25, 1982.
Columbia is one' of the largest natural gas systems in
the United States and is composed of The Columbia Gas System,
Inc. , a registered public utility holdling company, a service
company and eighteen operating subsidiaries. The operating
subsidiaries are primarily engaged in the production, purchase,
storage, transmission and distribution of natural gas at wholesale
and retail. Columbia supplies directly through its retail operations,
or indirectly, through other utilities, the gas requirements of
about 4,200,000 customers in an area having a population of
approximately 18,000,000. Columbia's service area includes large
parts of the states of Ohio, Pennsylvania, Kentucky, New York,
Virginia, West Virginia, Maryland and the District of Columbia.
Columbia serves at retail 1,890,000 customers residing in communi-
ties witfi a total population of 7,400,000.
As an owner of natural gas processing plants, Columbia
has a direct interest in this CTG. Therefore, Columbia is sub-
mitting the following comments.
C-6
-------
Emissions Standards and Engineering Division (MD-13)
Page 2 ..,-...
March 9, 1982
Model Plant A Considerations
While the stated purpose.of the GTG is to provide
information to state and local air pollution agencies it will
actually be used by these agencies to''develop and adopt new
regulatory programs for compliance with Sections 172(a)(2) and
(b)(3) of the Clean Air Act. Thus, Columbia believes the
fo3?X5?mentS °f Presidential Executive Order 12291 of February 17
1981 (45 FR 13193; February 19, 1981) must be considered in ^he" '
development of RACT for natural gas/gasoline plants. Section 2 of
this Presidential Executive Order requires that potential benefits
must outweigh costs and be maximized. In this respect, EPA makes
no recommendations in the CTG concerning the limits of applica-
bility of RACT (Reasonably Available Control Technology)"Vo
natural gas/gasoline processing plants. . '. *'
However, the CTG states that plants of the size of the
model plant A will incur a net annual cost for RACT of $2300
with a cost effectiveness of $115/Mg. For these plants requiring
conversion of pneumatic control valves from gas to air, the costs
?XiiH r ^?r (Section 5-1'5^ 'The larger plants, model plants -
B and C, will receive a net savings ($4,200 and $24 500
respectively) and have a cost effectiveness of -$70/Mz ind -$1207
Ms, respectively (Sections 5.3.2 and 5.3.3). Given the low annual
emissions of volatile organic compounds (VOC) 29 Mg (Table 2-2)
of the model plant A, the annual costs and cost effectiveness for
these plants dictate that these plants'should not be subject to
the requirements of RACT.
Secondly, an annual emission rate of 29 Mg is equivalent
u u2™°nS/7ear VOC' This level of emissions is less than
that whzch. EPA has defined as "significant" or as a de minimis
value in several regulations (for example, 40 CFR Part 51
Appendix 5, II.A.10; 40 CFR 52.24(f)(13); and 45 FR 52705-52710
August 7, 1980). The "significant" or- de minimis value applicable
to ozone is 40 tons per year of volatile organic compounds.
Finally, the smaller plants will have fewer employees
£™ larger Plants- Thus, it may be anticipated the proposed
RACT measures will more severely strain their limited staff.
Based upon the above three considerations of costs and
benefits, emissions below "significant" or de minimis levels and
potential manpower limitations, Columbia recommends that, in the
C-7
-------
Emissions Standards and Engineering Division (MD-13)
Page 3
March 9, 1982
CTG, planes of the size of model plant A (10 vessels as defined
in Appendix B) or smaller be excluded from the requirements of
RACT. This exclusion, similar to the following, should be added
as the last paragraph of Section 1.0 Introduction:
Natural gas/gasoline processing plants
equal to and less than the size of a model
plant A (10 vessels as defined in Appendix)
should be excluded from the requirements for
RACT. This recommendation is based upon an
analysis of costs and the low level of emissions
of volatile organic compounds from such plants.
Further, the CTG should point out in Section 2.3 that VOC emissions
from the model plant A are less than those considered as
"significant" or de minimis . Thus, the imposition of RACT, with
its attendant costs (Section 5.0), is not needed.
Monitoring Instrumentation
One of the major costs for implementing RACT is that
for the purchase and maintenance of monitoring equipment. Natural
gas facilities already have portable monitoring instruments for
detection, of leaks of combustible hydrocarbons as part of their
safety programs. Further, this type of instrument is considerably
less expensive than the monitoring instruments described in the
CTG. Leak detection of combustible hydrocarbons with these
instruments and knowledge of the processes before and -after various
components could provide an estimate of VOC emissions. Use of
this instrument and approach would provide the operator with a
method to estimate VOC emissions, to determine those components
requiring repair and to measure the effectiveness of repair, but
at much -less cost than the recommended type of monitoring
instruments .
Columbia recommends that EPA include this type of
instrument as an alternative to the purchase of expensive, new
instrumentation for leak detection and implementing RACT.
Columbia appreciates the opportunity to comment on the
CTG and trusts the above comments will be evaluated and of value
to EPA.
Sincerely,
C-8
Michael J. Atherton, Ph.D.
Environmental Affairs
MJA/ljh
-------
u
SOUTH RN CALIFORNIA | g(3S
COMPANY
G M GAROETTA
Environmental Affairs
Aarmmstrator
810 SOUTH FLOWER S"EET • LOS ANGELES. CALIFORNIA 90017
Mailing Aaart-s SOX 3249 TERMINAL ANNEX. LOS ANGELES. CALIFORNIA 90051
March 9, 1982
Mr. Fred Porter
Emission Standards and Engineering Division (MD-13)
Research Triangle Park
North..Carolina 27711
Dear Mr. Porter:
Southern California Gas Company (SoCal) appreciates the
opportunity to submit the following consents on the Envh on-
mental Protection Agency's (EPA) draft control techniques
guideline (CTG) document entitled "Control of Volatile Organic
Compound Equipment Leaks from Natural Gas/Gasoline Processing
Plants", for review and consideration.
SoCal is the nation's largest natural gas distribution
company. Accordingly, it has a serious concern regarding the
applicability of the proposed CTG to underground gas storage
facilities. It is not clear from the language of the proposed
document whether or not the definition of a natural gas/gasoline
storage operation excludes underground gas storage fields.
SoCal strongly feels that a gas storage operation should not
be compared with conventional oil/gas production facilities and
gas processing plants. It is important to recognize that among
other factors, the magnitude of fugitive emissions will be
dependent on the complexity and number of component processes.
The liquid and gas processes performed at an underground gas
storage facility are relatively few and simple when compared
to those at a conventional gas processing- plant and oil/qas
production operations.
In order to demonstrate the basis for its concern, SoCal
has provided the following summary of the facilities which could
be impacted and has compared these to traditional gas processing
plants and oil/gas productfon operations. SoCal operates six
underground storage fields located in Honor Rancho, Aliso Canyon,
Playa del Rey, Montebello, East Whittier and Go!eta. The gas
withdrawal capacity ranges from 1.5 billion cubic feet per day
to 72 million cubic feet per day. The larger gas storage fields
operate to meet peak winter load demand while smaller fields are
C-9
-------
Ltr. to F. Porter
dated 3/9/83
Page two
usually used to meet daily peak load demands. Therefore, operations_at
SoCal's larger storage fields are seasonal compared to conventional oil/
fields where production is usually continuous throughout
gas
the
production
year.
SoCal's gas storage fields are depleted oil or dry gas fields, and
any oil production is obtained primarily due to repressurization of the
field to store gas. Coincident oil production from these underground gas
storage fields is not significant. The gas to oil ratios in SoCal's opera-
tions range from 90,000 to 766,000 cubic feet per barrel of oil produced.
This is significantly higher than the reported gas to oil ratio of 1 ,000
cubic feet per barrel of oil produced from conventional oil/gas production
operations. A. high gas to oil ratio clearly implies a smaller scale of
oil treatment''operations and'consequently results in significantly lower
fugitive emissions.
Figure 1 (attached) represents a simplified flow sheet of the gas
withdrawal process in a typical underground gas storage field. In general,
gas injection, withdrawal and dehydration operations are similar at all
SoCal's storage fields with the exception of Playa del Rey where there is
no dehydration. Oil treatment (stabilization) and oil/condensate storage
are other operations where additional HC gas is generated and the methods
of processing or handling this gas varies from field to field. At Montebello
and Playa del Rey, this gas is directly delivered into the low pressure
distribution or transmission pipeline system. At Honor Rancho, it is
"delivered to an oil company gasoTine plant for further processing. At Aliso
Canyon, the gases are compressed and then the liquid fractions (gasoline)
are removed in a Hydrocarbon Recovery Unit (HRU). It is important to note
that the volume of HC gases generated at oil treatment and storage operations
represent only a small fraction of the total natural gas processed.
To study the composition of the aforementioned gas streams, one should
refer to Table 4 (attached). The compositions of non-methane and non-methane
plus non-ethane were obtained from actual field test data. The table also
compares SoCal's data with the average gas analysis reported in the emission
factor table of the API/Rockwell report - "Fugitive Hydrocarbon Emissions
from"Petroleum Production Operations", March 1980. The non-methane and non-
methane plus non-ethane hydrocarbons present in SoCal's wet gas range from
5.26 to 8.47 percent by volume and 1.54 to 2.21 percent by volume respect-
ively. Conversely, an average composition of similar wet gas reported in
the API study contains 22.93 percent by volume non-methane hydrocarbon and
18.4 percent by volume non-methane plus non-ethane HC. This difference
clearly indicates that fugitive emissions of reactive hydrocarbons from
an underground gas storage operation are significantly lower than a con-
ventional oil/gas production facility. The low concentration of non-methane
or non-methane plus non-ethane hydrocarbons in SoCal's wet gas is not mere
coincidence.
C-10
-------
Itr. to F. Porter
dtd. 3/9/82
Page three
imnr^T 1 <-°-1/9aS Product1on operation the gas withdrawn
unprocessed, and contains significant amounts of higher fractions of
P6e ainin9 ethane> Propane, butane'and natural gasoline.
aawh.Vh ,:S-°Perat^nS the gas injected is commercial natural
ma am°U^S °f ^^..hydrocarbons and conse-
hydr°Carbon Actions ™ Picked up
plant operations on the other hand,
-JT j.i ; ~' "• ..-.•»., wniuiie, propane and butane from thp hala
of the species present in an unprocessed natural gas. The remaining
liquid .is..P1ped to refineries or chemical processing plan?. 9
Two types of gas plants are primarily
the field
°f
he Si9n]'f1cantly fewer processing steps used
in
CONCLUSIONS AND RECOMMENDATIONS
Fugitive emissions from underground storage fields ooeratinn< ^0
s-
Sincerely,
GMG:avs
Attachments
c-n
-------
1
H«i | r
• To ProceMincI
To Hetering I
s
I±j
:tl <
£
IW
S-o-i
s' r^j
8
U
S
. U S
2 *
= o
• «• s
to o
C-12
-------
c
c
*«•
J0
n
«
0
e
•*
h
J
«
U
E
J
£M
a
S
u
;s
z
z
••
u
s
s
z
^
u
.*"
«*
e e
S °
o -^
«. ^
a! S
•a
I 2
£ 5 I
g 5 *
^ 03 C 4(
!•»••* W
. 8 , *
3 ^ * °
5 § S °
s
bl
hi
s
g^
§
i
to o
§ :
wi
S «B
01
^•1 «p4
— «
2
• a
41
£
u
a
tiB
Z
**
u
1
•
r
^
e
o
•4
i
CO
'
VO
^*
•H
VO
»<•
^
^
ea
ta
m
n
in
en
N
ea
in
in
"*
e
en
£
g
2
u
I
CM .at i*
n •* CM
rt N «s
^* ^4 Ol
• a
• -4
s§s
222
*M IM iw
a a a
n
a
o
b
i
u
V
3
H
^
o
^^
a
o
u
CO b
fif"
^0
. ffl
at
\a IN
41
b
0
b
a>
•o
0)
Sb-<
0)2
Scu ua
< '
«•
NOTES t
J3 0-0
01 W
ncomin
4«4
a
IB
(9
at
•S
ii
b
ii
s «
e g
t) 4)
X S
II
t i
Ed
a 3
II
C-13
-------
MICHIGAN CONSOLIDATED GAS COMPANY
March 10, 1982
Emissions Standards and Engineering Division
Environmental Protection Agency
Research Triangle Park,
North Carolina 27711
Attention: Mr. Fred Porter
Dear Sir: .
Michigan Consolidated Gas Company is a natural gas utility currently
servicing in excess of one million residential, industrial and com-
mercial customers. Although Michigan Consolidated does not operate any
gas processing plants, as defined by your department, the operation of
such plants does have an indirect impact on the cost of natural gas to
our customers. Therefore, Michigan Consolidated believes that the Draft
Guideline Series for the Control of Volati1e Organic Compounds Equipment
Leaks from Natural Gas/Gasoline Processing. PUnts. imposes excessive
finaricTifrburden without realizing significant air quality gains.
Michigan Consolidated disputes the premise made in the guidelines that
Volatile Organic Compound (VOC) leaks are not being adequately detected
by industry (Section 3,1.1.)- F9r" economic reasons and, more importantly,
safety reasons extensive precautions are taken to detect and repair all
but the most insignificant sources. Although visual, audible, and
olfactory methods are the primary safeguards, it is not uncommon to
supplement these methods with oxygen and hydrogen sulfide monitors.
Since many of the major sources (pumps, compressors, etc.) as identified
in the guidelines are frequently located inside buildings where emissions
are confined, early leak detection becomes .even more important. There-
fore, Michigan Consolidated believes that fugitive VOC emissions from
natural gas processing plan-t-s are adequately controlled and further
regulation would prove overburdensome.
Michigan Consolidated also disagrees with-the methods used in estimating
current emissions from processing plants. Fugitive sources vary signif-
icantly depending on a variety ofvconditions. These include: system
pressure, equipment age, climate, past maintenance, gas composition (as
it affects corrosivity) and a multitude of other factors. To assume
that these emissions are simply a.function of the number of valves,
pumps and flanges at a facility i-s a gross oversimplification.
Although, the guidlines mention the fact that repairing most fugitive
sources will require venting the gas to the atmosphere, they do not
include this as a source in either their emission estimates or in the
computation of recovery cost credits. Since repairing many insignif-
icant leaks will require blow down of potentially large portions of the
system, significant emission reduction benefits and recovery cost credits
are questionable.
C-14 i
Our natural gas is your most economical form of energy . . . please help conserve :t.
-------
March 10, 1982
Emissions Standards and Engineering Div
Attention: Mr. Fred Porter
Page Two
In conclusion, Michigan Consolidated believes that regulations resulting
from the implementation of these draft guidelines will be economically
burdensome to our natural gas customers while resulting in insignificant
improvements to air quality. Therefore, we request that the cost/benefit
aspect of these regulations be carefully reconsidered, and that the
deadline for comments be extended allowing industry additional time to
further analyze these complex regulations.
Sincerely,
Steven E. Kurmas
Senior Environmental Engineer
SEK/sl
C-15
-------
TEXAS OIL & GAS CORP.
~ ' O i L I T Y UNION ~OW=:^
DALLAS, TEXAS 7S2OI
JACK W. BOLEY
MANAQtR
I CNVtHONMCNTAt. Arr
March 10Y~T982
Emission Standards and Engineering Division (MD--3)
Environmental Protection Agency
Research Triangle Park, North Carolina 27711
Attn: Mr. Fred Porter
RE:
Control of Volatile
Organic Compound
Equipment Leaks from
Natural Gas/Gasoline
.Processing Plants
Dear Mr. Porter:
Texas Oil & Gas Corp. is an independent energy company that operates 18 natural
gas processing plants and is processing approximately 811 million cubic feet
per day (MMCFD) of natural gas is responding to the referenced document. The
processing is accomplished in various types of plants, such as, turbo-expander
cryogenic, refrigerated lean oil, and propane refrigeration. We believe we
are quite experienced in the processing of natural gas. It is the opinion of
Texas Oil & Gas Corp. that prudent operations eliminate any fugitive emissions-
associated with the plant processing of natural gas.
The information presented in the document indicate various emission factors
for various segments of a natural gas processing plant. It is apparent that
much of the information was obtained from a Draft Final report by Radian
Corp; September 1981. Since this report is not available to Texas Oil & Gas
Corp. there is no way to dispute or support the data based on the reference.
In 1979, the Environmental Protection Agency (EPA) published "Guidence for
Lowest Achievable Emission Rate from 18 Major Stationary Sources of Particulates,
Nitrogen Oxides, Sulfur Dioxide, or Volatile Organic Compounds," EPA-450/3-
79-024. No where in that report were natural gas/gasoline plants mentioned as
a major source. Texas Oil & Gas Corp. believes the regulation of fugitive
emissions from natural gas plants will not provide a significant benefit to
the environment.
Texas Oil & Gas Corp. believes that there are a number of flaws in the document.
The flaws as perceived by Texas Oil & Gas Corp. will be discussed section by
section.
C-16
-------
Mr. Fred Porter
March 10, 1982
Page 2
Section 2.2.4" Pressure Relief Devices
This section discusses the possible emissions from relief valve
seais. No where in this, section1, or in is document is the
consideration of a closed system discussed. In many ^nstanc-s '
pressure relief valves are vented to a oiant flare whe^lS?
voiauiie organic compounds (VOC) are combusted and thus "no VOC
would oe detected around these' devices. -From operational standpoint
seais or more properly pressure relief valve seats cannot ^
allowed to leak.
Section 2.2.7 Gas Operated ' Control Valves '
The instrument gas used for process control -is oredcnnnately
1^h^.an?.stha5e- /either methane nor ethane'.are included
ohe definition of volatile organic compounds, consequently
bl
Section 2.3 BASELINE FUGITIVE VOC EMISSIONS '
This section makes an assumption that all natural gas plants
Sn /f^1^06 ?e S3Me type °f TOC sessions. The document
. fails to differ between the type of plant, i.e. cryogenic^ lean
oil or refrigeration, size of plant (for both gS^STllquiS)
and the gas composition, i.e. the amount of light hydrocarbons '
ScLr^h ? Sisnificant ^act ^ fugitive emilsioS SStSe?'
' the SJ? ^-n0t ^fussed ^ fche total mix of fluids at
the plant. Various hydrocarbons,, amines, glycols, slop and lube
oils and condensates can be encountered, at any or «n natural
gas plants. Plants that handle the fractionated
Section 3.1.1 Individual Component Survey
The monitoring for this study was accomplished
using of a portable hydrocarbon analyzer. The data
obtained using the portable hydrocarbon analyzer was used to
measuring an actual flow rate. In order to accept the emission
' obtatains the
C-17
-------
Mr. Fred Porter
March 10, 1982
Page 3
Section 3.1.2.2 Compressors
The venting of fugitive emissions by enclosing the compressor
seals is not cost effective and is also an operational hazard.
At the present time these emissions are allowed to disipate in
the atmosphere, by enclosing and venting these emissions, the
VOC air mixture would reach explosive limits, thus creating a
serious explosion potential. Texas Oil & Gas Corp. questions
the ability of a system, as described in this section, to
achieve a 15 to 20 psig pressure. It seems apparent that a
sufficient exit rate would be required to obtain this pressure.
These operations may reach an operational equiiibruim where the .
vent system will never reach the 15 to 20 psig release pressure,
thus trapping the VOC in the vent system.
This section mentions a "combustion device" to handle these
emissions. At these low pressures, 15 to 20 osig, a flare is
not safe for similar reasons as mentioned above. We are -unaware
of any type of combustion equipment that would facilitate the
removal of these VOC from the atmosphere. Texas Oil & Gas
suggests that the EPA throughly reevaluate this type of control
systenu
Secton 3.1.3.3 Allowable Interval Before Repair
The previous Section 3.1.3-2 Inspection Interval indicates a
quarterly inspection'Interval. If there were a number of
components in excess of the 10,000 ppmv level, the 15 day repair
internal may not be met. Also, if there are sufficient fugitive
emissions, the only way the 15-day repair internal could be
met is by shutting down the plant. The natural gas industry
does not believe the economic implications of plant shutdowns
have been anticipated by the EPA, Texas Oil & Gas Corp. suggests
that the 15-day repair interval be eliminated in favor of, repairing
leaking components during the next regular maintainence period.
Texas Oil & Gas Corp. strongly believes that the regulation of fugitive VOC
emissions from natural gas/gasoline plants will not have a significant positive
environmental impact. We suggest that the data based to evaluate these fugitive
emissions be expanded to insure a. representative cross section of plants, not
six plants in the entire United States, is used to obtain a more accurate
emission profile. The EPA must consider the economic impacts in light of the
current economic picture of the United States. The increased regulation of
the already overregulated Oil and Gas industry will not enhace the future
economic stature of the country.
Very truly yours,
>r.
Jack W. Boley f
Manager,
Safety & Environmental Affairs
JWB/mkc
C-18
-------
AJLR-CUIMI'JKUJL .JOAKU
JOHN L BLAIR
Chairman
CHARLESR.JAYNES
Vice Chairman
BILL STEWART, P. E.
Executive Director
6330 HWY. 290 EAST
AUSTIN, TEXAS 78723
512/451-5711
WILLIAM N.ALLAN
VITTORIO K.ARGENTO.P. E.
FREDHARTMAN
D.JACKK!L1AN,M. D.
OTTO R. KUNZE.Ph. D..P.-E.
FRANK H.LEWIS
WILLIAM 0. PARISH
March 10, 1982
Mr. Fred Porter
Environmental Protection Agency
Emission Standards and Engineering
Division (MD-13) °
Research Triangle Park,.. North Carolina
Dear Mr. Porter:
27711
We offer the following comments on the January 25 1982
Federal Register notice concerning a.-draft control techniaues
guideline (CTG) for control of volatile organic comoound"'
emissions from equipment leaks from natural gas/gasoline
processing plants.
From this notice, it is our understanding that the Environmental
Protection Agency now plans to use CTG documents to provide
technical and cost comparative data to state and local air
pollution control agencies to assist in analysis of reasonably
available control technology (RACT) for various industrial
processes. We understand that the CTG's are not to be
regulatory and will not impose any new requirements.
We fully endorse the need for the Environmental Protection
Agency to prepare and distribute to state and local »overn-
m o n t* c 't-^/^'V'iv-in/-*''^! •iw^.—.w-i™.^ •*-.:—. ______ • .. _ ^
cost and avail-
ments technical information concerning the
ability of control technology. Such activity effectively
supports state and local regulatory efforts without r«=stric^-
ing the opportunity for state and local governments to taylor
regulations to meet specific local conditions. We encourage
/™ ^ r"1 ^ T V\ 1 1 j~\ s3 m*tW14.>..«4.-__ _^ /^ m f* * - ^
continued publication of CTG's
tional and not regulatory.
so long as they are informa-
C-19
-------
Mr. Fred Porter
-2-
March 10, 1982
Thank you for the opportunity to comment. .If you desire
additional information, please call.
Sincerely,
Roger R. Wallis, Deputy Director
Standards and Regulations Program
cc: Mr. Dick Whittington, P.E;, Regional Administrator,
U.S. Environmental Protection Agency, Dallas
C-20
-------
American Petroleum Institute
2101 L Street Northwest
Washington, D.C. 20037
202-457-7330
C. T. Sawyer
Vice President
March 11, 1982
Mr. D. R. Goodwin
Emission Standards and Engineering
Division
Environmental Protection Agency (MD
Research Triangle Parx, NC 27711
Dear Mr. Goodwin: '
13).
,ho ?i ? ? ^r°ieUm InStltUte (API) herewith submits comment
on the Control Techniques Guideline; Control of Volatile Organic
Compound Equipment Leaks From Natural Gas/Gasoline Processing
Plants listed at 47FR 3403 (January 25, 1982). "*>mg
API maintains that this Control Technique Guideline (CTG) is
unwarranted since EPA has not shown the .need for such guidelines
Furtnermore, EPA has failed to demonstrate the effectiveness SI
the control measures proposed, and has misrepresented the costs
29 ?SS2 I*** tiveness. Nevertheless, in response to the January
comments! letter ot Mr' J* R' Farmer, API offers the attached
n
plants
based.on extensive first hand experience with gas
comprehensive experience with fugitive hydrocarbon
n5f^SP°nSOred the mOSt si9nific^t fugitive emissions
nr™ ^Ctl^eqUlpment (Eaton' et al* 1980) which has been
performed. The study included two gas plants where a large
number ot components were tested. All of the gas -plant data were
provided EPA tor use in formulating this CTG.
In addition to sponsoring the fugitive emissions study, API
tro??!LTeHal ^meS with'EPA' as «« CTG was being developed,
to offer technical advice and the benefit of operating experience
Further, API presented a statement oefore the National ^P.erience-
-Pollution Control Techniques Advisory Committee (NAPCTAC) when a
preliminary draft of the CTG was reviewed (Woodruff, 1981)
filed by API '
n* EPA haS accePted the API technical advice
and responded to our other comments. Nevertheless, a number of
?!.?3006"18 remain- Primarily, these concerns have to do
leak measurement method, transf erability of information
C-21
-------
Page Two
from tne petroleum refining and caemical industries, safety ana
economic analysis in support of tnis guideline. Our concerns are
detailed in the attached comments.
If tnere are questions on these comments, contact Mr. E. P.
Crocxett, 202/457-7084.
Sincerely,
uju\
C. T. Sawyer
Attachments
C-22
-------
AMERICAN PETROLEUM INSTITUTE -
Comments on Draft
Control of Volatile Organic Compound Equipment Leaks
from Natural Gas/Gasoline Processing Plants
(47FR 3403, January 25, 1982)
• . .- Chapter 2 — Sources of VQC Emissions-
1" . General — The process streams to which this CTG applies
should be clearly stated. The requirements of the CTG should
pertain, to plant components in contact with fluids containing;10%
or more volatile organic compounds (VOC) by weight since leaks
from fluids containing less than 10% VOC represent deminimus
losses. .Excluded, therefore, from the provisions of this guide-
line would be (1) compressors handling only methane-ethane and
(2) other equipment in hydrocarbon service where the VOC content
is low. . , . \
Additionally, components installed on lines operating at
negative gauge pressure should be exempted'-from., this .CTG, since
leakage from the component is impossible. Otherwise valuable
time and resources will be expended monitoring components having
no actual potential to leak, to the atmosphere.
r
Chapter 3 — Emission Control Techniques
2. General — Chapter 3 is lifted almost entirely from the CTG
for fugitive VOC emissions from chemical manufacturing [EPA 1981]
which is based "on studies of chemical plant and refinery processes
There is no technical basis for the transf er'abil ity of .chemical
C-23
-------
plant or refinery VOC emissions data to natural gas/gasoline
processing plants because (1) the processes are different, (2)
no description is provided of the statistical analysis of the
Emission Source Test Data (CTG Appendix A)to demonstrate leaks
were from gas-liquids handling components/ (3) no derivation is
provided of confidence limits to establish the correctness of the
predictions and, (4) ho comparison is supplied of gas plant data
with the other study data. Following are some of the critical
issues which must be addressed by the Agency in the CTG to
demonstrate their recommendations have technical merit:
(1) What is the repeatability of the instrument reading?
(2) What is the accuracy of the instruement reading?
(3) What is the source of the estimate that after repair, 10
percent of the original number of sources''develop leaks
each quarter?
(4) What is the basis for the recommendation to isolate and
purge a pump before repair? Is this feasible?
(5) Why is quarterly monitoring required for all valves,
pumps and relief valves when existing data [Eaton, at
al. 1980] dispute this monitoring frequency?
(6) Why is a 15 day (98%) repair interval recommended as
opposed to a 5 day (99%) or a 30 day (96%)?
(7) What is the basis for the assumption that leak repair
reduce emissions to 1000 ppm?
The above API concerns are not addressed by EPA in the CTG.
Specifically, EPA is silent on the accuracy and repeatability
of the instrument readings. Recent field testing shows the
average error made by 15 people screening 28 leaks using a
.hydrocarbon detection instrument was 65,000 ppm. This represents
C-24
-------
6.5 times the "action level" which triggers repair of a component
and clearly demonstrates the poor repeatability of the instrument
(see comment-4 below).
Additionally, the Agency has not supported- the assumption-
"10 percent" of the original components will develop leaks
quarterly. Eaton, e_t al. (1980) indicated. a- much lower rate of
leak recurrence' for components, in production service. Further,
the assumption repaired components emissions will be 1000 ppm is
wholely unsupported by any documentation offered by EPA. There-
fore, API-concludes, these, parameters were selected arbitrarily.
3- Page 3-1, first paragraph"— It is stated that the CTG is
based, in part, on the transfer of technology from other indus-
tries because of similarity in types of equipment'used by these
industries. Work practice/performance type control techniques
may be applicable to gas plants from related industries. However,
it is erroneous to imply other aspects such as specific sources,
emission rates, monitoring .techniques and maintenance schedules
are directly transferable to gas plants from other industries
since known differences exist in operating temperature, .operating
pressures, vibrational problems and product compositions. Radian
(1980) and Eaton e_t _al. (1980) document fugitive emission rates
are independent of pressure and temperature within chemical
plants, refineries and production facilities. However, there is
no documentation to show differences do not exist between
facilities within these related industries as the result of
differences in temperatures and pressure. For example, most
C-25
-------
- 4 -
modern gas plants are cryogenic plants. On the other hand,
refineries and chemical plants operated at elevated temperatures.
The feed to a refinery is different than the feed to a gas
plant. In fact the gas plant product is gas liquids (natural
gasoline) which is one- of the inlet streams to .a refinery. The
inlet to the gas plant is natural gas and its entrained (vapo-
rized) liquids. Further, the vaporized hydrocarbon liquids in a
refinery process are different than in a gas plant since the
mixture is more complex. The mix is more complex because of the
presence of heavier hydrocarbons with greater VOC emissions
resulting from the higher temperature and pressures inherent with
a refinery. EPA has not addressed these differences in the CTG.
Thus, without supporting data, technology transfer is question-
able at best.
4. Page 3-2, second full paragraph — The CTG states without
support the portable hydrocarbon detection instrument is the best
survey method. The soap score method is discussed in the CTG in
a negative and superficial manner. There is no discussion in the
CTG of the limitations of the VOC analyzer. Some of the limita-
tions of the detection instrument which must be dealt with
are:
o Extreme delicacy of the instrument;
o Sensitivity to correct calibration;
o Weight and inconvenience of the instrument?
o Poor repeatability?
o Lack of demonstrated accuracy
C-26 „ -^
-------
- 5 -
o Time delay in achieving a reading; and
o Difficulty in receiving timely repairs.
The hydrocarbon detector is cumbersome and difficult to
handle at gas plants. Many gas plant components ."are at elevated
locations not accessible from platforms. For example, relief
valves are located on top of fractionating columns. The need to
access these gas plant valves is infrequent. 'Thus, platforms are
not installed on the columns. Also, the columns are small in
diameter, i.e., two to four feet at the top. Thus, space '"for" a'
platform, piping and valves-is limited. Monitoring of elevated'
components with a hydrocarbon _detector at a gas plant requires
the inspector to climb the column with the detector over his or
her shoulder. At • the top of the column the detector must be '
brought to the front of the operator, the probe moved about' the
surface of the potentially leaking, component, and readings taken
while standing on a ladder 50 - 80 feet above the ground. This
acrobatic challenge is not impossible to perform,' but it is dif-
ficult, dangerous and time' consuming. The documented difficulty
that a woman operator experienced in'handling the instrument in a
training program^, along with her lack of confidence in the
method, is related in Attachment A. 'During the training program,
15 participants measured 28 leaks (four times .-each) . TO assess
variation in readings, calculations were made for (1)
m?n^ nv? readinq' and (2> the absolute difference between
minimum OVA reading
the maximum and mijiimim readings. The results averaged overall
the leaks are 240 for the "tic and 65000 ppm as hexane for the
-------
difference. The OVA's were all calibrated immediately prior to
the field work. This demonstrates the lack of
repeatability of the instrument. It also brings into serious'
question the accuracy of instrument readings.
API has consistently advocated use of the soap score method
as an effective, simple, economic screening method as demon-
strated by Saton, j_t _al. (1980.) The only place where soap
scoring is not applicable is on rotating and reciprocating
equipment. Rotating equipment refers to liquid pumps in gas
plants. Eaton, _et _al. (1980) shows liquid leaks are small. EPA
data does not demonstrate this_ fact since their study did not
differentiate between liquid and gaseous leaks. Additionally,
EPA indicates leaks from reciprocating compressors need not be
monitored, since most modern compressors are provided with closed
distance pieces which will be vented. Thus, the alleged advan-
tages of the instrument monitoring technique are not utilized in
practice. We urge EPA to adopt the soap score method as the
principal leak detection technique since the instrument technique
is known to be non-repeatable.
5. Page 3-3, 3.1.2.1, last sentence — Isolating the pump and
flushing it of VOC prior to pumps repacking or seal replacement
is vague and difficult to understand. Additionally, the flushing
fluid must be disposed. It cannot be returned to the process
stream. Thus, flushing does not appear practical, and the
sentence should be deleted.
C-28
-------
6- ' Page 3-3, 3.1.2.2 — Collection and combustion of emissions
from the compressor seal area is impractical, unsafe, and not
cos-t effective. Reducing the compressor seal concentration level
below 10,000 ppm by repair is very difficult and- often impracti-
cal based on-experience during the API maintenance study.
Although compressor seal areas are frequently enclosed and vented
outside of the compressor building (S2.2.2), these vents are
unrestricted in most Natural Gas/Gasoline Plants unless hydrogen
sulfide is present.
In some cases it is also dangerous, if not impossible, to
enclose the distance piece to-_hold gas pressure from' a lea-king
packing. For example, one compressors model will not hold over 5
psi pressure without blowing the metal cover off the distance
piece according to the manufacturer.' The potential also exists,
in some cases, for pressurized hydrocarbons to pass through the
engine crank case seal and enter the crank case thus creating an
explosion hazard. .
The fact that 80% of the gas content is non-reactive methane-
ethane makes these systems different from the typical refinery
compressor systems used as the reference case. In gas plants
this emission source represents only 2.6% of the VOC emissions
(Table 4-2) and the cost of control, in a safe manner,' is
unreasonably .high. Additional block valves, a pressure control
valve, and a pressure relief valve between the compressor and
the first block valve are all required for a safe connection of
this vent line to a flare line. The assumed materials and labor
C-29
-------
- 8 -
in Table 5-1 are deficient and both are about one-half of the
real cost when allowances are made for pressure relief valves,
pressure control valves, unions' and the actual installation
labor. The result'is a capital cost of about $11,000 in a Model
"C" plant to collect 7-.3 megagrams of VOC annually. This amounts
to 19,5% of the RACT capital cost to eliminate only 2.5% of the
VOC emissions from gas/gasoline plan.ts. The VOC emission elimi-
nation cost of. $1,507 per megagram is unreasonably high and this
control technique should not be considered RACT. Compressors
should be exempt from control, or the action level raised to a
reasonably attainable level. •_
<
7. Page 3-8, 3.1.3.3 — The CTG specifies 15 days for required
repair without explanation. This could represent as few as
nine scheduled work days at plants where repair crews may only
be' available two or three days per week. The repair of all leaks
detected within fifteen days may be impossible. A 30 day repair
period would be more appropriate.
8. Page 3-16 — There is- a numbering error on pages 3-16 and
3-17. Paragraphs 3.2.1, 3.2.2, and 3.2.3 should be 3.3.1,' 3.3.2
and 3.3.3. Our comments are based on the numbers shown in the
CTG.
Chapter 4 — Environmental Analysis of RACT
9. Page 4-1, second paragraph — The soap score method should
be adopted as the principal leak detection method as discussed in
comment 4 above.. C-30
-------
- 9 -
1°* Page 4-2, 4.0, last paragraph — This paragraph is misleading
in casting doubt on the viability of soap score monitoring.
Further, it is not likely an acceptable correlation between soap
scores and instrument readings will be established since the
instrument readings are not repeatable. Accordingly, this
paragraph should be deleted from the CTG.
11' page 4-9, Table 4-6 -- This table is in error. Recovered
energy for open-ended lines should be corrected to 250• bbi crude
?etroleum/yr equivalent.
tt
Chapter 5 — Control Cost Analysis of RACT
12' Page 5-1, 5.1 — The economic analysis has equated "front'
end costs" with "capital costs,." Capital cost has an extremely
limiting economic definition which does not allow the inclusion
of operating cost. Only the VOC analyzer and the piping of the
compressor seal emission should be defined as capital costs. The
remaining costs•incurred are classified as expenses. Expense and
capital costs cannot be combined without using an amortization
schedule for the life of the capital purchases.
13' Page 5-1, 5.1.1 — The estimated cost for capping open flew
lines is based on the price of a one-inch screw-on type globe
valve. EPA assumed that any larger line size can be reduced -o
one inch. Normally, gas lines are specific sizes for a reason
(e.g., the hose size to be attached):, and therefore can not be
reduced arbitrarily. According to CTG Appendix A 'Table A-2),
721 open-ended lines were t,es±ed. Those data cculd be used —
w ™ O I s"- " *""'
-------
- 10 -
develop a distribution of line sizes and choose an appropriate
average valve cost rather than make questionable assumptions.
In the economic analysis, the double valving of an open-
ended line has been assumed to be a capital cost. Installed
equipment costs of less than S250 are rarely considered to be -a
capital cost. In fact, both the valves and labor to install the
double valves are expense costs. In addition, the estimated cost
of adding the valves to open-ended lines was underestimated due
.-!*
to the omission of miscellaneous costs, e.g., record keeping*,
vehicle use, source identification and tagging.
14. Page 5-4, 5.1.4 — There is a discrepancy between 43 hours
for a pump seal repair and Table 5-2 which shows 11 hours for
repair for the same seal.
15. Page 5-4, 5.1.4 — The statement:
"Because initial leak repair is a one-time cost, it is
treated as a capital cost."
is not correct because a one time expenditure is not a correct
criteria to define a capital cost. The initial leak repair is a
one time operating cost and must be included in the initial cost
of emissions reduction, not distributed over a reasonable time
period as proposed. For example, reducing the financial the
first year's repair cost is distributed over ten years forcing a
reduced impact. The emissions reduction realized during the
first year will not, however, be saved each year as assumed.
Each year's cost must be compared against each year's corre-
sponding benefits to determine the cost effectiveness of RACT.
C-32
-------
16' Page 5-4, 5.1.5 — RACT for the process controls-and actua-
tors has been defined as conversion to compressed air. With no
estimate of VOC emissions .saved and no economic analysis, there
is no support for this RACT. The expectation that most gas .
plants -already use compressed air does not constitute justifi- .
cation for RACT.
17- Page 5-6, 5.2.1 _— The cost analysis for the labor associ-
ated with leak detection severely underestimates the actual cost.
Whether it requires one minute for;VOC sampling of a valve or 5
minutes is a function of such factors as the plant configuration,
the monitoring method, the personnel, the weather, and the
location of the component. While there is no current evaluation
of this activity for gas plants, the reference cited (letter from
J. M. Johnson, Exxon, 1977) is not appropriate for gas plants and
out of date. The information contained in that letter was
determined for refineries for an entirely different purpose. •
An API member company has obtained recent and realistic cost
information from independent firms performing VOC monitoring.
•These firms have identified the cost per source tested. Although
contracting the monitoring to a third party may be slightly
higher (assume 15 percent profit) than .performing the monitoring
program internally, these costs include all of the associated
hidden costs and overhead. Typical bids from contractors who
regularly perform this service vary from $ 1.80/source/sampling
for an unsophisticated system with 100,000 to 200,000 fugitive
sources to $5.60/source/sampling for. a computerized data system,
C-33
-------
- 12 -
with the majority in the $2 to $4.range. A S4/source/sampling is
a reasonable contractor cost, assuming 15 percent is profit.
Deducting contractor profit, it is possible for an operator to
monitor his own VOC sources for approximately $3.50/source (1981
dollars). This estimate does not include leak repair? resampling
after the repair; or initial design, acquisition or implementa-
tion of the monitoring network. The development and implementa-
tion of a new monitoring program will be about equal to the
first year sampling 'cost for smaller gas plants (such as Model
Plant A) and 60 percent of the first year's cost for larger gas
plants - plus the cost of the'.instrumentation.
In brief, the estimated coat of solely maintaining a moni-
toring program is 5 to 6 times the estimate cited in the CTG,
assuming $3.50/source/sample and referring to Table 5-4. First
the approximations for man time are extremely low. In addition,
front-end set-up costs; depreciation on the equipment; and
maintenance on the VOC analyzer must be included. Furthermore,
the cost analysis must include the cost of the otherwise unneces-
sary platforms for each inaccessible source." If other options
are implemented, such as mobile platforms, then the time and cost
required to sample each valve must be included in the economic
analysis.
18. Page 5-6, 5.2.2 — The estimate for leak repair costs
consider only the labor of actually repairing the valve or pump.
Omitted is a number of associated hidden costs such as record
keeping, use of a vehicle, provisions for inaccessibility, cost
C-34
-------
- 13 -
of repair parts, loss of production, and overtime factor. If the
gas planfis keeping sufficient records to take'-advantage of' the
statistical relief of Section 3.3; the' costs are'further increased.
Although actual cost of repair is facility dependent, a'realistic
estimate of maintenance costs are ?120/repaired valve and S1000/
repaired pump. Relief valves are included in the"number'of •'"
'valves; Even if'the valve only requires reseating', qualified
maintenance personnel must perform the function ana the choice of
personnel is not optional, especially under union contracts.'
>
19. Page 5-10, 5.2.6 — ..Using 1981 dollars for recovery credits,
realistic estimates are nearly"twice what is quoted in the CTG in
value per gallon and one and a half times that per MCF. The
error in these approximations is partly due to the assumption
that all the VOC is propane. The recovered VOC is 6 Ibs/gal
rather than 4 Ibs/gal if the correct product density for propane
is used. Therefore, the value of the recovery credits per Mg is
$146 not $210. . :
20- Page 5-10, 5.3 — m section 5.3 the "cost" of RACT has been
estimated. The conclusion, based on erroneous assumptions
discussed above, is (with the exception of the smallest plants)
the petroleum industry has lost significant revenue due to the
lack of controlling the VOC,emissions. In reality the coses of
implementing and maintaining the-recommended control of VOC
emissions from a gas plant are greater than estimated and a
net long-term, loss will result from the use of -his RACT based or.
the above cost information. -
-------
- 14 -
Appendix 3 — Model Plants
21. Page B-1, 8.1 — The three model plants were developed from
questionable data. All four should be included in the analysis
or the reason for including only two of four EPA tested plants in
the vessel- and component inventories should be explained. Final
selection by API of two plants was based on maximizing the number
of components at sites for emission measurements. The resulting
API component inventories at the tested plants are unusually
large and thus of questionable value in developing model plant
configurations.
C-36
-------
- 15 -
References . ..
Eaton, W. S.f e_t _al. [1980]. Fugitive Hydrocarbon Emissions
From Petroleum Production Operations. API Publication 4322,
March, 1980. - . . .
Woodruff, W. J., [1981]. Phillips Petroleum Company, Statement
on Behalf of the American Petroleum-Institute before the
National Air Pollution Control -Techniques .Advisory Committee,
Raleigh, NC, April 29, 1981.'
Sawyer, C. T._, [1981]. Letter to The National Air Pollution
.Control Techniques Advisory- Committee, May 1 5,. -.1-9.8.1. -
EPA [1981], Guideline Series, Control of Volatile Organic Coumound
Fugitive Emissions from Synthetic Organic-Chemical, Polymer/
and Resin Manufacturing Equipment. ,
Radian [1980], Assessment of Atmospheric Emissions
Refining. . . .... -.-•••
from Petroleum
C-37
-------
r
(RgV. 5-78)
Shell Oil Company
Interoffice Memorandum
FEBRUARY 5, 1982
FROM: SENIOR ENVIRONMENTAL ANALYST
PACIFIC DIVISION, VENTURA
T0: SENIOR''STAFF ENVIRONMENTAL SPECIALIS'
WESTERN E&P OPERATIONS
On January 25, 1982, I attended a Rockwell International
school to learn how to operate an OVA. I would _i.
-------
In short, I found that the OVA is awkward to -ase and
completly unreliable. On a positive note, some models are
explosion-proof.
S.S. Walker
'SSW: jk
C-39
-------
MICHIGAN WISCONSIN PIPE LINE COMPANY
MSV15£= Z' "-5 iME5.CAi\i ">jATwSAL RESOURCES SYSTEM
3 = " = C." MICHIGAN 4322S
March 11, 1982
Emission Standards and Engineering Division (MD-13)
Environmental Protection Agency
Research Triangle Park, North Carolina 27711
Attention: Mr. Jack R. Farmer
Re: Control Techniques Guideline Document Equipment
Leaks from Natural Gas/Gasoline Processing Plants
Dear Mr. Farmer:
The following comments, on the referenced document, are being sub-
mitted by Michigan Wisconsin Pipe Line Company (Michigan Wisconsin) which
owns and operates an extensive interstate natural gas transmission and
underground storage system. Michigan Wisconsin transports gas from pro-
ducing fields in the Oklahoma-Texas Panhandle area, Louisiana and offshore
in the Gulf of Mexico and through-connecting pipelines from Canada to
52 distributing utilities in nine states. Michigan Wisconsin and its
'subsidiary ANR Production Company -are also engaged in exploration for
gas and oil in the major gas prone areas of the South, Southwest, Rocky
Mountains, Michigan, Offshore in the Gulf of Mexico and in Western Canada.
We have carefully reviewed the above referenced report, which is
clearly well researched and we are in agreement with many of the proposed
control techniques. However, there are certain requirements which con-
cern us and we have concentrated our comments on these requirements.
Emission factors presented in this document are based on natural
gas liquid processing plants, gasoline plants and natural gas liquid
fractionation plants. For natural gas gathering plants and natural gas
dehydrating plants, emission factors for non-methane/nonethane hydrocarbons
(VOC's) will be much smaller because the ratio of methane and ethane to
total hydrocarbons is much larger for these facilities. The costs of
implementing Reasonably Available Control Technology (RACT) for controlling
fugitive emissions of Volatile Organic Compounds (VOC) cannot be justified,
for such facilities because the quantity of non-methane/non-ethane hydro-
carbons is very small for natural gas gathering plants.
There are leak prevention and control procedures in place at most
natural gas plants in compliance with minimum federal safety standards,
The proposed technology in this document will be repetition in most in-
stances and increase regulatory burden.
C-40
-------
- 2 -
Costs projected for converting from natural gas to compressed air
actuated control valves -should also include costs for the .air compressor
Compressed air is generally not available at most small natural -/as com-
pressor stations, natural gas gathering and dehydration plants.. The cost
cannot be justified for remote locations, which are not sources- of, 1 ar^e
/quantities, of VOC's-. ' • ' '•'•' °
C'os,t.-of a portable VOG analyzer is reported to be $4,600 by the
report. "The annual cost of materials and '-.labor for maintenance and
'.calibration of monitoring instruments is estimated to be S3,aoO." --The ;; '
., combined cost of $7,600 per year is very large compared to' .currently •''
practiced technology of leak detection, which employs visual,' olfactory '"
or audible means for the purpose.
Based on our past experience in controlling leaks and successful
operation of natural gas processing/compression facilities, a Table com-
paring our current- practice and proposed techniques is attached for com-
parison. The table will show that in most instance's, proposed control
techniques are already in practice, at our facilities/- - " ••• -
.- Michigan Wisconsin would like to thank you for providing this oppor-
tunity to comment on this document. We believe that the present techniques
employed by natural gas liquid removal, dehydration, and compression faci-
lities are adequate and the majority of the emission factors developed in
this document should not be used to judge performance of these "facilities.
" Sincerely yours,
Jitendra V. Mehta,
Environmental Engineer
Attachment
cc: Messrs.
Mrs.
J. P. Cencer
V. "D. Lajiness
R. J. Lecznar
P. B. Thompson
M. L. Webster
File
C-41
-------
u
ta
en
a.
t-H
a,
S^
I-H
tn
WISCON
MIC
tn
s
o
u
o
o
•H
4->
U
ca
o.
tH
ca
H
4>
o
a
® /->
ca BO
i-H O
•'ca o
> c
< JS,
o
a I-H
0 N
tn +J
a e
o o
as u
v_/
tn
o
o
t o
r-l (U J3
3 fn a
O O O 4-1
i-« 3 s a
A O
ca fn JH O
4-1 O O O
Q_ (sj flj
« xo
U r-l O 4-1
o ca • o
a cos
a o
H 0 E-
0 U *CJ
4-> i a
4-1 U i-H O
xj fl} 5(J [|5
O P< 3
0
O 'fH
<-l
•H X
4-) •-<
tn s
3 O
.2
s -a
o o •
O tn s~\
3 90
_a tn cn x o C
C -H fH
fH -H > 0
o *->
X w> o
O i-H C Cfi
a a
' rH a ®
3 I-H a
in 4->
•H X M
> ^H O f-
X O t-
i-H H -iH
r\ Q3 «^
a N rH
4J X 0
M r-< 3
O td
a. c o
v_/ ca 4->
•
4
)
]
,-/ 4-> bo X
f-
^?
^ f-i C •-
^
CD a -H a
o 3 tn c
4
o: s cy 3 a
V)
w >
tj __j
4J
S
0
c
o
a
S
o
CJ
^ w~ I
o a
tn >
O
tn o <4
tn o JH (
Q, > o.-i
H
u tn
H CL,
S PH S rH S
3 a o o 3
a, > cj ca a.
0
CO
a
,
C
<;
tn
M S
o o
•H
tn •(->
o o
90 O
s s
a s
r-4 O
a. cj
tn
a
00
tn
C -3
o o
a FH
o o
O M
i-H 4-1
X O
s o
£ a
IS
US
3 «
0«
o a
o i* tj
tw s a
O -H S
PH y js.
a o 4->
3 O<'H
tn tn 3
•H S
> -H tn
^
s o a
O tn O
•H f—I
o o
4-1 a o
tn
a
o o
S -3 +J
O i-H O
CO O S
<4H 3
a PH
C -H
S
S 0 S
0 = 2
'II '2 >
tin^
o « S
o
-------
Chevron
Chevron U.SA Inc.
555 Market Street, San Francisco, California . Phone (415) 894-3041
Maii Address: P.O. Sox 3069. San Francisco. CA 94119
R. W. Kreutzen
General .Manager
Environmental Affairs
March 12, 1982
Draft CTG for Natural Gas/Gasoline
Processing Plants
Mr. Jack R. Farmer, Chief
Chemical's and Petroleum Branch
Emission Standards and Engineering Division (MD-13)
Office of Air Quality Planning and
Standards
U.S. Environmental Protection Agency
Research Triangle Park, North Carolina 27711
Dear Mr. Farmer:
*'
Chevron is happy to comment on the draft CTG for natural gas/gasoline processing plants.
We note considerable similarity between this document and the draft NSPS for VOC emis-
sions, upon which we commented last November 16. There are enough differences
between the two drafts, however, to warrant our offering a new set of comments.
Section 2.2.7 Gas-Operated Control Valves
No actual emission factors are given for these valves, nor any indication of how many use
compressed air vs. "field gas." If the "field gas" is natural gas, very little nonmethane/-
nonethane hydrocarbon will be present.
Table 2-2 Baseline Emission from Model Plants
The baseline emissions are based upon the emission factors in Table 2-1. Although the
9596 confidence interval for these factors is extremely large (generally a factor of five or
more), there are no error estimates for the total baseline emissions. What are the confi-
dence figures on these numbers?
Section 2.3 Baseline Fugitive VOC Emissions
Estimating components by ratioing to major equipment has merit. However, ratioing
components to all vessels pooled together (columns, heat exchangers, and drums/tanks) is
an oversimplification, as demonstrated by Table B-3. Comoaring the average ratio of
flanges and connections to vessels (97.4) and the standard deviation of the ratio (59.7)
shows the correlation is sometimes poor. This is also confirmed by the disparate ratios
reported - about 50 for the EPA-tested plants and about 150 for the industrv-tested"
plants. .
-------
Mr. 3ack R. Farmer, Chief
Page 2
March 12, 1982
Section 3.1.1 Individual Component Survey
This section fails to' mention the considerable problems associated with maintaining and
operating a portable hydrocarbon analyzer. They are temperamental precision instruments
sensitive to heat, humidity, and the type of gas being sampled. Reliable use requires
thorough personnel training and careful maintenance and calibration procedures. For
example, the inspection and maintenance program in Chevron's El Segundo refinery requires
one full-time person to maintain and service the detectors. Even if the same type of
instruments are calibrated and used side by side, reproducible results can be elusive. This
has been Chevron's experience at our Ventura County, California production lacilities.
The point is that portable hydrocarbon analyzers are very tricky, and their required use
could be a considerable burden on smaller operators.
Monitoring requirements for unsafe and difficult to reach components should receive spe-
cial consideration. At the least, this CTG should include the attached statement taken
from the draft CTG on "Control of Volatile Organic Compound Fugitive Emissions from
Synthetic Organic Chemical, Polymer, and Resin Manufacturing Equipment" (August,
1981).
Section .3.1.2.1 Pumps
It is stated that a pump should be isolated from the process and flushed of VOC-as much
as possible prior to repacking or seal replacement. The reason for doing this would be to
assure that the temporary VOC emissions from repair do not exceed the emissions from
the original leak. We believe that in practice this would be nearly impossible to do. EPA
should consider the prospect that the emissions resulting from the repair of a pump leak
could offset any long-term benefits, depending on the extent of the original leak.
Section 3.1.2.2 Compressors
We must take strong exception to the control strategy discussed here for reciprocating
compressors. There are serious physical and safety considerations associated with enclosing
compressor seals in the manner suggested. Rather than detail these issues here, let me
refer you to K. C. Hustvedt of your RTP facility, who is very familiar with the issues
raised by industry when a similar strategy was proposed for refineries.
Table 2-2 assigns compressors only 3% of the total gas plant emissions. In view of the
serious problems associated with controlling these emissions, we strongly urge that this
strategy be dropped.
Section 3.1.3.2 Inspection Interval
The draft CTG apparently considers only quarterly inspection intervals. EPA should seri-
ously consider annual inspections, since quarterly inspections are practical only for a rela-
tively small number of major components (like compressors). Even in Los Angeles, which ^
generally has the toughest hydrocarbon control regulations in the nation, inspection with a
C-44
-------
Mr. Jack R. Farmer, Chief
Page 3
March 12, 1982
detector is required annually for pumps, valves, and flanges, and quarterly for compressors.
More frequent inspections are not judged cost effective. -
Section 3. T.3.3 Allowable Interval Before Repair
The report suggests that 15 days is reasonable to allow a plant operator enough time to
obtain repair parts. While often true for readily available parts, this time is much too
short for difficult-to-get parts. Allowing sixty days in such special cases is more reason-
able.
Section 3.3 Other Control Strategies
It is suggested that if less than 2% of the valves are found to be leaking, then the operator
may skip inspections. We feel this a good concept, but it should be extended. The CTG
still requires that inspections be carried out on a yearly basis. We believe that if data
indicate a longer time interval would maintain a leak rate of less than 2%, then this inspec-
tion interval should be used. In addition, we can see no reason not to apply this concept to
other fittings as well as valves.
The subsections in section 3.3 are misnumbered.
Chapter 4.0 Environmental Analysis of RACT
This chapter could be improved considerably by looking at other inspection intervals. We
thought the corresponding information in the NSPS was much more comprehensive and
generally pretty accurate (with certain exceptions noted in our earlier letter). We would
encourage you to add some of this information to the CTG.
It is interesting to compare the values given in Tables 4-2 and 4,-3 with Chevron's Los
Angeles area refinery experience. The inspection and maintenance program for valves at
our El Segundo refinery (10,000 ppm cutoff, annual inspection) yields a reduction
efficiency of about 65%, not too far from the entry in Table 4-3 for quarterly inspections.
On the other hand,' our I<3aM program for compressors (quarterly inspections) yields about a
35% reduction. The 100% reduction noted in Table 4-2 is not realistic (see our comment
on section 3.1.2.2).
Chapter 5.0 Control Cost Analysis of RACT
We have some concerns about the costs that were used in estimating the cost effectiveness
of the CTG.
a. We currently would estimate the cost of labor to be $23.00 per hour rather than $13.00
. per hour (28% higher).
C-45
-------
Vlr. Jack R. Farmer, Chief
Page 4
March 12, 1982
b. We do not believe a venting system could be installed on compressor seals .for $700.00
(page 5.10). National Air Oil Company estimates the cost of their smallest flare to
be $8,000. This would not include the cost of piping to each seal.
Cost effectiveness numbers are given for a RACT program involving quarterly inspections.
While these numbers are somewhat open to questions, for the reasons given earlier, the
lack of incremental cost effectiveness calculations is a serious omission. What is the
incremental cost effectiveness for annual or monthly inspections versus quarterly, for
example?
******
We hope these comments are' helpful. For your information, we are attaching a copy of a
technical paper given last November, which describes Chevron's experience with I&M
programs at our El Segundo, California refinery. If you wish to discuss our comments
further, please contact Michael Foster of my staff at (415) 894-6107.
Very truly yours,
MSFrrdg
Attachments
C-46
-------
ATTACHMENT
The following statement is from the Draft CTG:, "Control of Volatile Organic
Compound Fugitive Emissions from Synthetic Organic Chemical, Polymer, and
Resin Manufacturing Equipment"4August, 1981, page 3-20).
Unsafe and Difficult to Reach Components
Some components might be considered unsafe to monitor because process condi-
tions include extreme temperatures or pressures. A State agency may wish to
require J^ess Jrequent monitoring intervals for these components because of the
potential danger which may be presented to monitoring personnel! For example,
some pumps might be monitored at times when process conditions are such that
the pumps are not operating under extreme temperatures or pressures.
Some valves may be difficult to reatch because access to the valve bonnet is
restricted or the valves are located in elevated areas. These valves might be
reached by the use of a ladder or scaffolding. Valves which could be reached by
the use of a ladder or which would not require monitoring personnel to be
elevated higher than two meters might be monitored quarterly. However, valves
which require the use of scaffolding or which require the elevation of monitoring
personnel higher than two meters above permanent support surfaces might be
monitored annually, for example.
C-47
-------
CONTROL OF FUGITIVE HYDROCARBON EMISSICNS
IN PETROLEUM REFINERIES
Charles W. Aarni
Ohevrcn U.S.A.
El Segundo, California
Clayton R. Freeberg
Chevron Research Company
Richmond, California •
C-48
For presentation at the Annual AIChE Meeting in New Orleans or
November 8-12, 1981.
-------
ABSTRACT
This paper discusses, the effect of two inspection and .maintenance
(I & M) programs for reducing fugitive hydrocarbon emissions at . •
Chevron's refinery in El Segundo, California. 'Two I •'& M regulations,
one covering, valves and flanges and the other covering pumps 'and com-
pressors, have been imposed on this refinery: by the South Coast Air -
Quality Management District. First-hand experience in meeting these
regulations is presented along with estimates of hydrocarbon, emission
reductions and estimates of the cost effectiveness of. the regula-
tions. The paper also explains how. to estimate fugitive, .hydrocarbon
emissions for
tion permits.
new facilities, which is necessary to obtain construc-
C-49
-------
SUMMARY
Ozone levels in the South Coast Air Basin, which include parts of four
counties in the Los Angeles Area, currently exceed the national
ambient air quality standard for ozone. This has resulted in the
South Coast Air Quality Management District (SCAQMD)' establishing
several regulations to reduce emissions of hydrocarbon; an ozone pre-
cursor. Consequently, Chevron's El Segundo Refinery has had several
years of experience with two SCAQMD regulations requiring inspection
and maintenance (I & M) programs—one for valves and flanges and
another for pumps and compressors. Based on this experience and the
use of the emission factors from a recent Radian report,1 results
indicate that the valve I & M program currently achieves a net
economic return. The effect of the flange I & M program has not yet
been completely evaluated, but it currently appears to be less cost
effective than the valve program. The pump and compressor I i M
program has a significant net cost ($i-$2/lb hydrocarbon controlled)
because of the higher.maintenance cost involved'in replacing pump and
compressor seals. Proposed future regulations are estimated to be
much more costly (perhaps $5/lb); therefore, industry must continue to
provide input to regulatory agencies to ensure that the most cost
effective controls are implemented.
Chevron's experience with I & M programs has proven to be an asset in
obtaining construction permits from regulatory agencies. An accurate
prediction of the fugitive emission control programs has provided
significant emission information which can now be used to develop more
valid estimates of emissions from proposed new facilities.
INTRODUCTION
Under the Clean Air Act, .the South. Coast Air Basin must meet the ozone
standard. This will require further reduction of hydrocarbon emis-
sions. In addition, many states are adopting New Source Performance
Standards which include fugitive emission controls and which will
affect almost all new facilities or major modifications. Thus, opera-
tors in ozone nonattainment areas are being confronted with the need
to participate in the development of new regulations to ensure that
the most cost-effective controls are used first.
At Chevron's El Segundo Refinery, which is located in the South Coast
Air Basin, controls on hydrocarbon emissions have been coming into
effect over the past 25 years. During this time, many major sources
of hydrocarbon emissions have been controlled. These include tanks,
oil/water separators, valves, flanges, pumps, and compressors. • Yet,
the Basin is still nonattainment for ozone. While mobile sources
C-50
-------
-2-
represent more than half of the hydrocarbon emissions in the basin
Chevron expects additional controls to be imposed on stationary
be more expensive and difficult
imcle-
spurces. Such measures will
ment than those in the past.
OVERVIEW OF ONE REFINERY'S f . ••-
HYDROCARBON EMISSIONS _
A summary of the estimated- -hydrocarbon ' emissions for Chev-o-'s
El S-egundo Refinery is shown on Table l7 "These emjss
-------
Table I
'Refinery Hydrocarbon
Emission Summary
Emission Source
Combustion Sources!
Solvents /Organic s2
Tanks1
Bulk Loading2
Fugitive Emissions3
Total
Lb/Day_
5CO
400
2,300
600
23,200
27,000
% of
Total
2
1
9
2
86
100
•'•Based on EPA's "Compliance of Air
Pollutant Emission Factors," AP-^
effective April 1981.
2Based on South Coast Air Quality
Management District emission
factors with current controls in
effect.
Table II for emission basis.
C-52
-------
— 4-
Table II
Fugitive Emission Summary1?
Fugitive Emission Source
Flanges^.
Valves ' •••-''
Gas and Light Liquid3,5
Heavy Liquid^
Compressors"^ . • '
Pumps ~ •
Light Liquid^ > 5-
Heavy Liquid^
Relief Valves _;
Separators?
Cooling Tower?
Drains
Total
Lb /Day
1,600.
13,600
' 300 *
700-
2,000
700
200*
900
1V500
1,700
23,200
•%; of To tar-'
.Fugitive
Em is s: ion's •
7 :
••••-59 •• -•'
:;. .- i •-_• :.-
' •' • • 3. •'. : '•
'
- , 9
3
1 ..
''k ,.
• 6
1 .'
100
%' of .Total ;;
Re'finery : •
Hydro-carbon
Emissions
Y-'ir " '
•- 50— ^
• ' "": i •-'•' ••
'::- "^ '-•-. '.:.
v. ,.;7 ;v;'
• >. 3 "I, ;
'- , :' ,1 : ' .' ,
' ' :3'J "' "
"•' ' 6" . '."
6 •• •"•
86.' ::
1Unless otherwise noted, calculations are based on '"'Radian
e,mission factors for nonmethane hydrocarbons
No reduction credit is assumed for' the flange I & M program
since the program is not yet completely evaluated.
•>A 65% reduction is applied to hydrocarbon gas and light
f»i to,Jccount for th* I & M. program. The method
for calculating the percent reduction is outlined in
Reference 2. ' •
5TA, 3H fjduction is applied to''account for .the I & M- program
^Light liquid is any liquid with Reid vapor pressure >1.55
gpsia. —
Heavy liquid is any compound with ReJ.d vapor pressure '<1.55
7COctober ision facfcors
EPA Publication AP-42
C-53
-------
PHILOSOPHY OF THE RULES
During the development of the valve and flange rule and the pump and
compressor rule, two significantly different approaches were pro-
posed. The first, which was'put forth by the regulatory agencies,
stressed enforcement of the rule. The basic concern was how to ensure
that companies were complying with the rule. So the agencies proposed
that any leak found by an agency inspector would be a violation.
However, this concept could create major problems for inductry since
it is impossible to stop all equipment leaks. At any given time,
there will be some fraction "of the valves, flanges, pumps, and com-
pressors which leak. In addition, 'leak rates are Variable—a piece of
equipment may not be leaking .one day but be leaking the next day.
These facts must be recognize.d in the rule development process.
Therefore, the industry approach was that- the rules should require
repair of all leaking equipment within a certain time period. A leak •
found by an agency inspector is not a violation, but it must be
repaired. A rule of this sort achieves the desired emission reduction
through inspection and directed maintenance without penalizing an
operator for expected occurrences beyond his control. This is the
approach which finally prevailed.
VALVE AND FLANGE RULE
The valve and flange rule requirements are shown in Table III. The
1,55 psia RVP limit makes a split between naphtha and kerosene which
is the same split made by Radian Corporation in their studies of fugi-
tive emissions.^ Valves and'flanges in heavy liquid service have very
low fugitive hydrocarbon emissions, and so they are exempt. Ethane
and methane are exempt because they do not contribute significantly to.
photochemical smog. Also, any-stream containing more than 80* hydro-
gen is exempt. •
Currently, Wery valve and flange subject to the rule must be
inspected annually with a portable hydrocarbon detector. This
requires a full-time three-man team in the field plus additional sup-
port people. The three-man team inspects and makes the first attempt
to repair, every leak found. The team consists of:
1. Operator
- Identifies applicable valves and flanges and records
the data. Assists in securing a valve for repacking
or replacement.
2. Technician - Operates and services the hydrocarbon analyzer.
C-54
-------
-6-
Table ,111
SCAQMD
Valve and Flange
Rule Summary
Applicability
- Applies to hydrocarbon gas and liquid streams with Re^d vaoor o
sure >1.55 psia, except methane and ethane. " ' "
Inspection Requirements
- Two complete valve inspections during the first year.
- One complete valve inspection each year thereafter.
- One -complete flange inspection each year effective May 1980.
- Reinspect each leaking valve three months after reoai rs a--
completed . -
Leak Definition
- Liquid leakage at rate >3 drops per minute.
- Gaseous hydrocarbon concentration 2.10,000 ppm at the source.
Repair Requirements
- Repair to nonleaker status «10,000 ppm) within two -working days
to
Recordkeeping Requirements .
- Maintain records of valve inspections for one year.
- Make records available to the District upon request.
- No records required for flange inspections.
Exemptions
- Natural gas valves and flanges.
- Hydrogen valves and flanges (>80% H2 ) •
-Inaccessible valves.., and 'flanges .
C-55
-------
Mechan
•> r.
- Performs necessary .maintenance.
The -earn is equipped with a Century OVA-108 hydrocarbon analyzer.
This analyzer "satisfies the instrument performance standards imposed
oy SCAQsMD.
In the field, the technician measures the hydrocarbon concentration at
the source with the analyzer. The operator records the data. All
valves tested are counted, but only thosewith emission concentrations
greater than 10,000 ppm have detailed data recorded to identify them
{e.g., plant, size, type, service, leak concentration). A bright
orange numbered tag is attached to each of these valves. These tags
help the inspection teams relocate the leakers during the quarterly
reinsoections.
The mechanic on the team attempts to repair a leaker as soon as it is
found. Since the inspection team is still inthe area, the repaired
valves are reinspected immediately. Considerable followup maintenance
tir.e is reduced by streamlining the repair and reinspection program in
this way.
Some valves and flanges are inaccessible. These are the valves ,and
flanges which cannot be inspected or repaired without excessive cost
and effort. Based on this criterion, less than ^% of the valves and
flanges are considered inaccessible.
There are two additional full-time technicians involved with this
program, one who performs the reinspections with a hydrocarbon ana-
lyser and another who handles the recordkeeping duties. All of the
records required by this rule are kept on a computer. The computer
greatly reduces the labor spent in data compilation and data handl-
ing. In addition, the computer provides a tickler file which flags
any special action, such as reinspection of a repaired leak.
The results of the I & M programs to date are shown in Table IV. The
first complete inspection showed that 4.3# of the valves subject to
the rule were leaking in excess of 10,000 ppm. The leak rate during
the second inspection six months later was 2.2%. The third and fourth
inspections, which were done at 12-month intervals, showed an average
of 2.8% leakers. Based on this average leak rate and the emission
factors from the Radian report!, the calculated emission reduction
from valves is currently about 19,000 Ib/day. This results in a net
economic return assuming a hydrocarbon value of $0.10/lb. If Chevron
were to use the emission factors from EPA Publication AP-42 (October
1977), which are significantly lower than the newer Radian1 factors,
the estimated hydrocarbon savings would not offset the cost of the
program.
C-56
-------
CD
cd
£-1
co
-u
rH
3
• CO
CD
OS
C
o
•H
.p
O
CD
a
CO
C
CD
bO
C
03
•a
C
03
CD
CO
CD ifl
bO C
C -H
03 ^
rH Oj
fr« CD.
CO
CD
bO bfl
C C
Oj -H
rH .*
fe crj
O
CO
CD t3
bO CD
C -P
crj o
rH' CD
&4 a
CO
• C
O M
CO
CD bC
3 C
rH «H
03 Js!
> 03
CD
CO
CD
> bO
rH C
03 «H
> x
OS
. CD
O >-3
CO
CD *O
> CD
rH .P
03 0
> CD
a
• CO
O C
z: M
T3
O
•H
CD
cu
C
o
1-
4_>
o
CD
a
CO
c
rH
VO
in
•
o
ON
•sr
•
o
CO t"-
rH rH
CM
0
rH f— 1 CO
t^^ ^^ pr^
•>
\o
CM
rO CM rH
• « •
.3- CM rn
CM -ST rH
^T VO ON
O t~» CT\
t— 1
CM ^" C—
co ON CO
JT PO JT
•> •> »
.=r in rH
CM rn on
ON
ON O-
C»— ON
ON rH
^H
SL,
CD CD
C J3
3 S
r-3 CD
1 0
>> 4)
^ a
03 1
3 >j O
C rH CO
OS 3 ON
oo
•=r
o\
•>
rH •
OO
- VD
*
CM
—
CM
CM
CO
VO
ON
O
t
rH
CO
rH
CO
r-l
*
0
co
ON
OS
S
•a
CD
•O
•D
03
CO
JJ
C
CD
S
CD
s*
3
'o-
CD
C
O
JJ
o
CD
a.
w
c
CD
bO
C
03
fe
rH
O
CO
rH
03
S
CD
4-4
as
TJ
CD
CJ
CD
Ou
CO
C
•H
CO
•H
C
3
CO
CO
CD
O
O
lj
a,
>5
rH
C
o
CD
TD
3
0
C
•^
CO
4-i
rH
2
CO
CD
CD
bO
C
cd
fe
CM
C-57
-------
-9-
Some new regulations currently being, considered by SCAQM.D, such as
including heavy liquid valves in the I & M program and increasing the
inspection frequency, would be much less cost effective than the pre-
sent program. For example, the incremental cost effectiveness of
adding the heavy liquid valves to the program would be about 35/1° of
hydrocarbon controlled, which is currently not considered economically
justified by most regulatory agencies and industry.
PUMP AND COMPRESSOR RULE • .
The current rule requirements for pumps and compressors are shown in
Table V. This rule applies to pumps in light liquidservice and com-
pressors in hydrocarbon gas service. The affected refinery stocks are
the- same stocks-as those covered by the valve and flange rule.
In the case of the pump and compressor rule,' there are two kinds of
leaks—visible leaks and leaks detectable only with a hydrocarbon
analyzer. Visible leaks are defined as a visible mist or three
drops/minute of liquid leakage. The early version of this rule, which
has been in place for several years> required only that all pumps and
compressors subject to the rule be inspected once a shift for visible
leaks and that all leaks must be repaired. Any visible leak found by
a District inspector is an immediate rule violation.
Since the rule was first passed in 1976, there have been many changes
in the area of fugitive emission control. There has been an improve-
ment in technology available, to control and quantify the emissions as
well as a vast improvement in the understanding of fugitive emis-
sions. Fugitive emissions are known to be much more significant than
was originally thought.
The need for further hydrocarbon reduction led.to the development of
another inspection requirement for pumps and compressors which became
effective July 1, 1981. In addition to the inspections for visible
leaks, each pump must be inspected annually and each compressor must
be inspected quarterly with a portable hydrocarbon detector. Any pump
or compressor with a concentration greater than 10,000 ppm at the seal
must be repaired. If a leaking pump or compressor has a spare, it
must be shut down within two days; if it is not spared, repairs may be
deferred until the next unit shutdown. A leak found by District per-
sonnel with a hydrocarbon detector must be repaired, but it is not a
violation.
C-58
-------
-JLU-
Table V
• SCAQMD
Pump and Compressor Rule Summary
Applicability
- Pumps/compressors handling hydrocarbon,gas or liquid with Reid vaoor
pressure >1.55 psia, except methane and ethane.
Requirements
Visible Leaks
- Maintain pumps/compressors so there is no visible vapor leakae-* or
visible liquid leakage >3 drops/minute.
- Any visible leak greater than the above limits ^ound HY Distr-'c4-
personnel is a, violation. - . - -
- Inspect pumps/compressors for visible leaks once/shift.
Invisible Leaks
- Inspect each pump annually and- each compressor quarterly with oor-
table detector. Repair any leak >10,000 ppm, measured 1 cm from
S 63.J. *
minimize leakage within one day and repair
- For spared equipment, take it out of service within H8 hours and put
spare in service. If the spare leaks, one pump must be repaired
witnin 15 days.
- Repair requirements:
1. Repair to <10,000 ppm, if possible.
? Tf
& » j. u. «
TnivQ ,te!:ls >75,000 ppm (10,000 -ppm after
July 1, 1982), then leak must be vented to pollution control
device or variance obtained. Must be repaired or reolaced at
next shutdown.
Exemptions
Pumps under 1 brake horsepower.
Pumps/compressors with applicable hydrocarbon content <20%.
Pumps with double seals. .
C-59
-------
r
-11-
The rule covering the annual inspections with a hydrocarbon detector
uses a phased approach to-allow industry time to attain compliance.
Currently, the goal is to repair all. leaks to below the 10,000 ppm
chreshold. Current mechanical seal repair practices are not sophisti-
cated enough to ensure that all seals can be made to satisfy the
10,000 ppm limit. Thus, during the first year the rule is in place,
leakage must only be reduced to 75,000 ppm. If -the leakage exceeds
75,000 ppm, a variance must be obtained or the emissions vented to an
air pollution control device. After the first year, the -limit becomes
10,000 ppm instead of 75,000 ppm. The purpose of this two-step •
approach is to allow industry one year to gather data on seal reli-
ability and repairabillty. If the data show that enough pumps cannot
meet the 10,000 ppm leakage limit, then the rule may be modified.
Chevron is currently gathering data on the leak, rates of the 513 pumps
and 29 compressors subject to the current rule. These pumps and com-
pressors are the second largest source of our refinery's fugitive
emissions. Preliminary data indicate that about 20% of the pumps do
not meet the 10,000 ppm limit. We are unable to predict at this time
how many of these pumps can be made to satisfy the 10,000 ppm limit by
repairing or replacing their seals.
The emission reductions due to the elimination of visible leaks have
not yet been quantified. The annual, inspection program .with a leak
detector is in 'its infancy, and changes are still being made to
improve the effectiveness of the- program. The limiting factor on the
rate of which pumps and compressors can be inspected is -the rate at
which they can be repaired in the Machine Shop. Preliminary data
indicate that about 100 of the affected- pumps can be expected to leak,
so an average of eight 'to nine pumps must be repaired every month..
This could be as much as a 25% increase in the number of pump repairs
previously required and is a significant increase in maintenance
requirements. Whenever possible, the inspection program is scheduled
so that any seal repairs can be done when a pump is sent to the shop
for other maintenance.
A major concern is how to control emissions from old reciprocating
compressors, many of which have been in service for more than
iJO years. If the leakage cannot be reduced to less than 10,000, ppm,
then the distance piece must be enclosed and vented to a closed systen
or the compressor must be replaced. Ultimately, several old compres-
sors may have to be replaced since venting to a closed system is some-
times impractical.
C-60
-------
-12-
One other .feature of. the refinery sea*l maintenance program is worth
mentioning. There is one full-time mechanical seal technician whose
job is to inspect and test every seal before it is installed. The
technician replaces any defective parts•or repairs a leaking new
seal. Once a seal is installed, the technician retests the seal for
leakage before the. pump, is reinstalled in the field. This approach
greatly improves the quality control for mechanical seal repairs and
replacements.
So far it has not been possible to calculate precisely the cost
effectiveness for the seal I & M program. A very rough estimate of
the cost effectiveness- of the rule indicates that this rule costs.
$l-$2/lb of hydrocarbon emission reduced. This includes a credit of
$0.10/lb for the recovered hydrocarbon. The pump and compressor rule
is less cost effective than the valve and flange•rule due to the high
ccsx of replacing seals.
EFFECT ON MAJOR NEW PROJECTS
Fugitive hydrocarbon emissions are playing .an increasingly significant
role in obtaining construction permits for major new projects, espe-
cially in nonattainment areas for ozone. For some large construction
projects, as many as a hundred s.eparate permits may be required.
However, the most critical permits are usually air permits.
In order to obtain an air permit to construct a new. facility or. modify
an existing plant, the applicant first must supply a valid estimate of
the emission rate of each pollutant from his new .project. Since it
can sometimes take several years from the inception of a project to
get final permit approval, the permit application with emission esti-
mates must be submitted as early as possible to avoid costly construc-
tion delays. The applicant usually must apply for permits long before
detailed process designs are available, which puts a severe strain on
the engineering staff to come up with valid equipment counts of fugi-
tive sources (e.g., valves and pumps) before any detailed designs are
complete. Here is where the applicant can draw from his past experi-
ence with I 4 M programs for similar plants to estimate the number of
fugitive sources within the new project.
A quick guess of fugitive emissions
taken to develop an accurate
is not acceptable. Care must be
estimate of emissions from a project
overestimating or underestimating can severely impact the
ng of a project. if the applicant underestimates "'
sions from his project, this will reduce
quality impact; however, when the
plant, he may find
because
permitting of
the emis-
the project's predicted air
- - operator wants to start up his new
that the permitting agency will only allow him to
C-61
-------
-13-
operate a. fraction of the pumps in the plant. Then the operator
either has to delay startup until he can renegotiate a new permit or
operate only a portion of the plant. If the applicant overestimates
the emissions, this will overstate the project's impact; permit
approval will be more difficult to obtain'. For example, in nonattain-
ment areas, the applicant would have to develop more emission offsets
than necessary, since the emission increases from his project have
been overestimated.
In permitting of new ..projects, the need for an early valid estimate of
emissions is now obvious. Estimating pollutant emissions from point
sources, such as furnace stacks, is relatively simple since this
involves a straightforward engineering calculation. But predicting
fugitive emissions is somewhat less accurate; however, the procedure
is becoming more standardized as more data becomes available. The
procedure for estimating fugitive emissions generally involves four
basic steps as outlined below.
1. Obtain Design Data
This includes equipment counts (e.g., valves and pumps), cooling tower
rate, waste water effluent rate, product loading rates, and tankage
information. Equipment counts ..are usually the most difficult to'pre-
dict. Since permitting requires such large lead times, final piping
and instrumentation diagrams are usually not available for developing
accurate equipment counts. The applicant has to estimate the equip-
ment counts based upon actual equipment counts for similar existing
process units or based upon the Radian report1 which quotes average
equipment-counts for many typical process units from 13 U.S. refiner-
ies. It is advisable for the applicant to have his engineering staff
review these equipment counts for reasonableness before the informa-
tion is submitted to the permitting agency. Some agencies may require
an adjustment in the permit after the new plant starts up based upon
the I & M program for that new plant. Therefore, to avoid any sur- .
prises, such as being required to supply extra emission offsets after
startup, the emission estimate in a permit application should be as
accurate as possible.
2. Select Emission Factors
Emission factors are average measured emission rates per equipment
unit. For example, the Radian report states the emission factor for
light liquid valves is 0.024 Ib/hr valve. Developing valid emission
factors usually requires sampling very large populations of similar
sources'. It appears that the most widely accepted fugitive factors
currently are those quoted in the Radian reportl.
C-62
-------
-11-'
3- Agree on Control Efficiency
Emission factors are usually quoted on an uncontrolled oas's Th?
applicant and the permitting agency then have to agr« on ^V cort-o^
efficiencies to assume the specific types of control Mechanisms
later when the application is reviewed
* • Calculate the Emission Rates •
By the time the applicant gets to Step
'
about-
;
Emission Rate -(Equipment Units ) (Emission Factor) ( 1-Eff iciency )
- (380 Valves) (0.024 Lb/Hr. Valve) ( l-o . 65) :
= 3.2 Lb/Hr Hydrocarbon . • ' . :
project i..u.uall, preliminary and subject to change • ng
determine exact equipment counts for a proposed pro^e-t ihii'f s-
the design phase is extremely difficult. But for ?he sake of -
ting, specific numbers must be supolied- this e
programs for existing fugitive
CONCLUSIONS
C-63
-------
-15-
1. The data from the valve I Sc M program currently show a net eco-
2.
3.
1|,
nomi
return due to reduced stock losses when che Radian- emission
factors are used. Decreasing the emission factors or increasing
the inspection frequency would adversely affect the cost effec-
tiveness. The flange I & M program has not yet been completely
evaluated.
The I & M program for pumps and compressors is a relatively cost
effective program from a regulatory agency's viewpoint. It is an
expensive program to operate, but the cost effectiveness is better
than most other hydrocarbon reduction strategies currently under
consideration for petroleum refineries.
Future emission reduction rules will be more expensive to comply
with, and the emission reductions- will be smaller. Industry
should participate in the regulatory development to make sure that
the most cost-effective controls are used first. All sources/
mobile and stationary, should be evalua
ed.
Estimating fugitive emissions is a critical part of obtaining'
construction permits for most new projects. Current I 4 M pro
grams provide a .valuable data base which helps to expedite the
permitting process.
References
and
I. Mesich, Prank C., Radian Corporation, "Results of Measurement
Characterization of Atmospheric Emissions from Petroleum
' Refineries," presented at Symposium on Atmospheric Emissions from
Petroleum Refineries (November 1979, Austin, Texas).
2.
Tichenor, B. A., Hustvedt, K. C., Weber, R. C., U.S. Environmental
Protection Agency, "Controlling Petroleum Refinery Fugitive
Emissions Via Leak Detection and Repair," presented at Symposium
on Atmospheric Emissions from Petroleum Refineries (November 1979,
Austin, Texas).
:lkf, smm
C-64
-------
Amoco Production Company
. 16825 Nonhchase Drive
?cst cff'ce 3cx -L3S1
Houston. ~exas 210
Robert E. Mahaffey -.''-..
Manager. Plant Engineering ana Construction (USA! ' "-- • , '
March 18, 1982
Emission Standards and Engineering Division (MD-13)
Environmental Protection Agency
Research Triangle Park, North Carolina 27711
Attention: Fred Porter
Re: Control Techniques Guideline Document
Equipment Leaks from Natural Gas/Gasoline Plants
File: LY-46-986.622 . .
In accordance with the notice contained in the January 25, 1982 Federal
Register, Amoco Production Company (USA) welcomes this opportunity to
comment on the referenced document. Amoco Production Company (USA) is an
oil and gas exploration and production company which operates 40 aas
processing plants in the U.S.
The natural gas/gasoline plant 'industry has an interest in keeping fugitive
emissions to a minimum. The economic value of the hydrocarbons and the
conservation of a valuable natural resource as well as protection of the
environment are all important considerations.
We feel the gas processing plants can maintain a low level of volatile
organic emissions without the necessity of the detailed monitoring and
record keeping proposed in the CTG document. A much more cost effective
procedure could be based on ambient concentration monitoring at or near the
plant boundary. Such a system would provide more continuous data of any
Sl!SS Jf SSCape fr°m the plant- If an afanonnal concentration is
detected, then a two step operation should be set in motion. First, the
plant maintenance force would be notified and they would seek to reduce the
concentration to normal levels. Second, on those occasionfwheS Se
immed1atel* "ccessful . a detailed monitoring
.-nn K° *6 n° standard concerning the sizes and configuration of the
sampling probe for the monitoring instrument. The probe tip must of
necessity be quite small in order to- reach the less accessible points
around small valves and flanges. This will likely make sampling less than
precise and subject to varying dilution effects aampnng less man
C-65
-------
Page 2
Comnents for the specific paragraphs of the Guideline Document are attached
hereto. If we can be of further assistance, please fee free to call on
Dr. Lyman Yarborough at 713-931-2943.
R. E. Mahaffey
LEP/pdh
221/Y
C-66
-------
AMOCO PRODUCTION-. COMPANY (USA). "''./.-
Comments Re: Control Techniques Guideline Document; Equipment Leaks from
Natural Gas/Gasoline Processing Plant . --'- '•, ••
The sketch shown in Fig. 2-1 indicates "methane*to sales"; It is felt that
this is not meant to be compositionally specific since most sales streams
also contain ethane and: frequently smaller quantities of .the heavier
molecular weight hydrocarbon as well as some inerts; i.e.,•nitrogen and
carbon dioxide. •'-.:•- ' '. ^.:.'.-.,..".
Re: 3.1.2.2
Page 3-3
Compressor
Installation of the additional valves (checks and blocks) will be
expensive and potentially'require downtime, for .installation.
Qpe.ration.of the vent space at a pressure ofAS-to 20 psig will
not be possible without rather extensive modification of many
machines. The distance piece.enclosure- of-many-machines will not
stand 15.to,20 psig. The door seal may not:be-suitable for this
type service. The pressure may force volumes.of hydrocarbons
into_the compressor-crankcase, ruining the lubricating oil,
causing engine damage and significantly increasing the danger of
a crankcase explosion. Compressor manufacturers could provide
details about the requirements of the specific machines.
To permit operation of pressured distance.pieces-for most
compressors would require reconstruction of the cylinder, new
compressor rods repiping all the process gas side of the
cylinder, and possibly modification of,,the compressor building
floor and walls. This would be prohibitively,expensive and would
require significant downtime on each machine. -
If vent lines are installed from compressors, the sizing should
be increased to 1 1/2 or 2", at least for the headers, to reduce
pressure buildup potential. -• • '.."-
Re: 3.1.2.3
Page 3-4
Relief Valves
Many vessels have been installed .without block valves to permit
relief valve removal and removal of such valves may require large
hydrocarbon emissions while depressing the vessel and/or its
operating system and require expenditure of hundred of dollars
per valve.
Many relief valves are installed in-rather inaccessible
locations. It would not be uncommon to require a cr-ane be
brought to the plant site to facilitate relief valve removal
these cases, costs of thousands of dollars "per valve can be
expected.
In
C-67
-1-
-------
Relief valves thus located will be difficult to monitor and
checking at 1 to 3 years intervals (depending upon the service)
is suggested.
The construction of relief valves (metal to metal seats) makes
zero emission difficult, especially after the valve has operated
one time. Testing a relief valve for leaks by use of a
hydrocarbon detector may be unrealistic since even a minor leak
into-the relief valve stack may, over a period of time, displace
all or part of the air and result in a high hydrocarbon
indication, particularly if the hydrocarbon vapor is. heavier than
air.
3.1.3.3 Allowable Interval Before Repair
Page 3-8
The time interval for repair of a leak after its discovery is
currently proposed to be 15 days'. We suggest that the operator
be permitted more flexibility in scheduling this work. As
accepted practice, most operators will repair significant leaks
as soon as practical after they are' observed, without waiting for
a specified monitoring period. The operator seeks to minimize
leakage and prevent further damage to his equipment. However,
there can be a need to order and receive maintenance supplies
before proceeding. For minor problems, the 15 day figure can be
reasonable but times of 60 days or even much longer times should
be made acceptable, "there are some repairs that cannot be made
without a plant shutdown.
The calculation of 98% efficiency for a 15 day repair period
seems to use 365 days (1 year) as a basis. This seems to imply
one leak per year per piece of equipment, which is unrealistic.
Static equipment, (flanges, valves) may be in operation for 10,
20, even 30 years without a failure or leakage of any kind.
3.2.2 - Open-Ended Lines -
Page 3-13
The CT6 advocates plugging or capping flanging or valving all
open ended lines. Many open-ended lines are vents installed for
safety purposes. Capping or plugging those lines will result in
added danger to personnel and equipment. Some of those lines
might conceivably be routed to a flare system but others must be
left free to prevent cross contamination or back pressure against
a piece of equipment. The CTG recognizes that the caps or plugs
cannot eliminate the emissions from the .first valve, only that
it's release is controlled. The technique then becomes of
questionable value.
3.3 Other Control Strategies
Page 3-15
Section 3.3 recognizes that valves will have a much lower leak
frequency than compressors, pumps, and relief valves and suggests
C-68
-2-
-------
that quarterly inspections may not be necessary. This rationale
is even more•applicable to flanges and connections and the
extension of time between inspections is most appropriate. Due
to the low leak frequency of valves and connections, it is
suggested that these items be removed from the monitoring
program.
Tables 4-5 and 4-6
Page 4-8 and 4-9
Energy Recovery Credits
These tables allow recovery energy credits for. all estimated
emissions (100 per cent reduction)' from open ended lines. This
seems .to be an error since Section 3.2.2 had recognized that much
of this emission could not be stopped. In many other instances,
the open ended vents would be routed to flare and energy recovery
credits would not be applicable. This error is also reflected in
the cost analysis of RACT (Section 5).
Section 5.0
Page 5-1
Control Cost Analysis of RACT
Many of the cost figures shown in this section appear understated
and, in addition, the costs do not seem to allow anything other
than ideal work conditions and new materials, i.e., no charges
are shown for pipe support material. Costs at the ^nfller
plants, especially where attendance is minimal, areTikely to be
much higher per unit of emission reduction. Such plants will be
forced to call upon outside assistance.
Table 5.2 and 5.4
Page'5-5 and 5-8
Labor Requirements
These tables show zerp labor time and cost for relief valve
repair on the condition that these repairs would"be done by
routine maintenance. The repair of compressor and pump seals
would also be done routinely as needed. It seems inconsistent to
charge the emission reduction program with the cost associated
with one repair and not another.
The monitoring times shown for valves seem inordinately low
Only 1 minute per valve is estimated. The instrument response
time alone may be as long as 30 seconds. Sampling procedure
specifies moving the Probe slowly along the interface periphery
while observing the instrument readout. At the point of maximum
readout, the Probe is held stationary for at least twice the
instrument response time. -Once this has been noted (record
rnn?]3IJS re?¥1n!d) the operator procedes to the next valve.
SnnJoSnil all the.calibration time, instrument warmup, care and
maintenance of the instrument and associated gas supplies as
well as the testing procedure, the time required could easily be
5 minutes or more per test.
The tables of monitoring times and cost do not show sampling
times for flanges and connections. There are generally a large
number of these devices in the plants and by the very nature of
C-69
-3-
-------
r
their construction, the test time per unit will be substantially
greater than for a valve.
L. E. Petty
Room*579, Ext. 2941
GP III
LEP/pdh
221/Y
C-7|
-4-
-------
FLUOR ENGINEERS AND CONSTRUCTORS, INC.
SOUTHERN CALIFORNIA DIVISION
3333 MICHELSON DRIVE
IRVINE. CALIFORNIA 92730
TELEPHONE: (714) 975-2000
TELEX: 69-2485
March 22, 1982
Mr. Jack R. Farmer
Chief Chemicals and Petroleum Branch
Emissions Standards and Engineering Division
U.S. Environmental Protection Agency
Research Triangle Park, North Carolina 27711
Dear Sir:
We have reviewed the draft document "Control of Volatile Organic Compound
Equipment Leaks From Natural Gas/Gasoline Processing Plants." We suggest
that if EPA conducts periodic inspections of plants, only those facifities
found not to be in reasonable compliance with guidelines be required to
compile reports and be subjected to quarterly inspection. This practice
will significantly reduce the burden of paperwork and costs to both
industry and EPA and still achieve the same overall goal.
Thank you very much for the opportunity to comment on the draft guidelines.
Very truly yours,
Hilliam M. Hathaway
Vice President, Process Engineering
iney J. Thomson
Senior Manager, Environmental Engineering
WMH/SJT:nr
C-71
-------
-------
APPENDIX D
SUMMARY AND RESPONSES TO DRAFT ,CT6 COMMENTS
D-l
-------
APPENDIX 0
SUMMARY AND RESPONSES TO DRAFT CTG COMMENTS
On January 25, 1982, the Environmental Protection Agency (EPA)
announced the release of the draft control techniques guideline (CTG)
document for control of volatile organic compound (VOC) emissions from
equipment leaks in natural gas/gasoline processing plants (gas plants)
in the Federal Register (47 FR 3403). Public comments were requested
on the draft CTG and comments were received by industry representatives
as listed in Table D-l. The comments that were submitted, along with
response to these comments, are summarized in this appendix. This
summary of comments and responses serves as the basis for revisions
made to the draft CTG.
D.I GENERAL
Comment;
One commenter (5) requested that the comment period be extended
to give industry more time to further analyze these complex regulations.
Response:
The CTG incorporates public comments from the preliminary draft
CTG (March 1981) presented to the National Air Pollution Control
Techniques Advisory Committee (NAPCTAC) in April 1981 and comments
received on the draft CTG (December 1981) that was announced in the
Federal Register January 25, 1982 (47£R_3403). Comments were also
received at the NAPCTAC meeting (July 1982) for the gas plants NSPS
and these have also been addressed in the CTG as applicable. EPA has
been working with industry since the inception of the gas plants
NSPS/CTG projects in December of 1979. Therefore, there has been
ample time for public comment.
D-2
-------
D.2 NEED FOR CTG
Comment:
One commenter .(8) questions in general, the need for the standard
stating that EPA has failed to.demonstrate the effectiveness of the
control measures proposed and has misrepresented the costs and cost
effectiveness. . .;
Response-':
National Ambient Air Quality Standards (NAAQS) (Section 109 of
the Clean Air Act) set a ceiling for public exposure to criteria
pollutants by establishing an ambient concentration level that must
not be exceeded anywhere in the United States. This control techniques
guidelines document will provide guidance to States and air pollution
control agencies for RACT-based provisions applicable to gas plant
facilities to reduce significantly volatile organic compound emissions
to achieve and maintain NAAQS for ozone. The CTG environmental.and
cost impacts are based upon actual field studies in gas plants and on
comments received on the draft CTG and on similar fugitive VOC control
projects. Specific comments on the controls, costs, and cost effectiveness
of RACT are addressed in the following sections.
Comment:
Another commenter (9) wrote that leak prevention and control
procedures are already in place at most natural gas plants in compliance
with minimum federal safety standards. .Proposed requirements are
repetitious and burdensome.
Response:
The commenter is apparently referring to occupational safety
requirements which have different purposes and may result in different
environmental benefits. Present industry practices (e.g., enclosing
compressor distance pieces and venting emissions outside of a compressor
house) may reduce occupational exposures, but they do not necessarily
reduce the mass emissions to the atmosphere. The data base upon which
these recommendations are made is from plants in compliance, with
existing rules. These data show gas plants have significant emissions
and that cost-effective controls will reduce these emissions.
D-3
-------
D.3 APPLICABILITY
Comment:
One commenter (3) argued that small plants (10 vessels or fewer)
should be exempt from RACT based on cost effectiveness (benefits don't
outweigh costs per Executive Order 12291).
Response:
In Section 4.1 a small plant cutoff is recommended based on cost
effectiveness of RACT. Small plants may need to rely upon outside
personnel to perform the leak detection and repair program and, hence,
may incur higher costs per unit of emission reduction to implement
RACT than large gas plants. Section 3.4.2 provides the basis for the
small plant cut-off.
Comment:
Another commenter (8) wrote that components with less than 10 percent
VOC by weight should be excluded.
Response:
Based on API testing, sources with less than 10 percent VOC have
significant emissions, therefore, EPA has not exempted these sources.
However, it seems reasonable that at some low percentage of VOC,
sources would have very limited VOC emission reduction potential.
Therefore, dry gas equipment (less than 1 weight percent VOC) are exempt
from RACT as described in Section 4.1
Comment:
A commenter (4) noted that the liquid and gas processes performed
at underground gas storage facilities are few and simple and, therefore,
should be excluded from the definition of a natural gas/gasoline
storage operation. Since there is no potential for leakage from
components operating at negative pressure, another commenter (8)
requested that these components be exempt from the CTG.
Response:
EPA concurs with the comments, therefore, the description of RACT
in Section 4.1 exempts equipment at underground storage facilities and
equipment which operate under vacuum service.
D-4
-------
D.4 CONTROL TECHNOLOGY
'D.4.1 Compressors
Comment:
One commenter (8) wrote that reducing VOC emissions from compressors
below 10,000 ppmv is difficult and impractical.
Response:
EPA recognizes that compressor repair to achieve organics
concentrations below 10,000 ppmv may be difficult and impractical for
reciprocating compressors with packed seals. Therefore, alternative
RACT impacts are based on compressor seal vent control systems.
Nevertheless, leak detection and repair would be required (unless a
vent control system-is installed) in those instances where repair can
achieve VOC emission concentrations below 10,000 ppmv. Centrifugal
compressors may operate in tandem, one in service while the other
serves as a spare. In such instances seal repair may be performed
without need for a process unit shutdown.
Comment:
Several commenters (2,6,8,10,11) remarked that many compressors
are not designed to operate with back pressure against the distance
piece; enclosing and venting emissions from compressor seals.and the
distance piece poses mechanical and safety problems. The enclosed VOC
air mixture could reach explosive limits. Enclosing compressors would
require extensive modification which would be expensive and require
significant downtime.
Response:
Many compressors are equipped with enclosed distance pieces.
Enclosed distance piece emissions are generally vented outside of
compressor houses; however, these emissions .can be safely vented to a
VOC control device (e.g. flare). EPA has reconsidered the safety and
cost aspects of venting compressor distance piece emissions. The
Chapter 5 cost analysis has been revised to include necessary safety
equipment for a compressor distance piece purge system. However, if
plant owners/operators can demonstrate that enclosing and venting
emissions from distance pieces and seal packing vents is either unsafe
or requires unreasonable cost such as replacement of the compressor,
D-5
-------
these compressors may be exempt from RACT. Most compressor seal
packing vents can be vented to a VOC control device.
Comment:
Another commenter (10) stated that since compressor emissions
represent only 3 percent of total gas plant emissions, they should not
be covered.
Response:
The gas plant compressor seal emission factors used in the draft
CT6 are based on emission measurements from open reciprocating compressor
distance pieces and does not include seal vent emissions or measurement
of emissions from distance pieces that are enclosed and vented outside
of a compressor house to atmosphere. The data base also includes dry
gas compressors, which are exempt from RACT requirements. The gas
plants compressor seal emission factors have, therefore, been revised.1
Using the revised emission factors in the model plants (see Table 2.2),
compressor seals contribute approximately 14 percent of total emissions.
For the actual equipment counts found during API and EPA testing,
compressor VOC emissions ranged from 0-42 percent and averaged 13 percent.2
Therefore, compressor emissions are significant and emission control
is considered.
D.4.2 Leak Detection and Repair Methods
Comment:
Two commenters (8 and 10) wrote that isolating a pump and purging
before repair (repacking or seal replacement) is not practical.
Flushing fluid disposal is a problem. Another commenter (9) further
questioned whether emissions resulting from pump repair might offset
long-term benefits, depending on the extent of the original leak.
Another commenter (5) noted that the venting of gas during repair of
fugitive emission sources is not -included in emissions estimates or in
computing recovery credits.
Response:
Process industry pumps are now routinely isolated and purged.
prior to repair. Flushing fluid is routed to the oily storm sewer for
treatment and disposal. This fluid is expected to be a small percentage
of total plant waste. RACT does not mandate purging pumps prior to
repair although a pump would normally be emptied prior to repair.
D-6
-------
Even if the pump were not purged and all process fluid in a pump were
allowed to evaporate to atmosphere.as a result of pump repair, these
emissions are approximately equivalent to the mass emissions released
to atmosphere by a leaking pump over a 3-day period, 9 kg. This would
not offset the long-term benefits of'RACT.
The final comment is based on RACT requiring shutdown and purge
for repair of leaking equipment. The draft CT6 included a provision
that required repa-ir of all leaks within one year of detection. RACT
has been revised such that repairs requiring a unit shutdown may be
delayed until the next scheduled shutdown. As such, RACT no longer
requires yearly turnaround for repair of .these equipment leaks.
However, as discussed in Section 3.4.3, a State agency might wish to
consider a provision in its regulation which would allow the Agency
director to.order an early unit shutdown for repair of leaking components
in cases where the percentage of leaking components awaiting repair at
unit turnaround becomes excessive.
Comment:
Commenters (3 and 9) stated that facilities already have portable
monitoring instruments (for safety purposes) that are effective and
less costly than the recommended monitors. Facilities should be
allowed to use their own monitors. The recommended monitors are
temperamental, sensitive to heat, humidity, and type of gas sampled,
and their required use can place a financial burden on small facilities.
Response:
Facilities may use any instrument as long as it satisfies the
requirements specified in Reference Method 21. EPA recognizes that
monitoring instruments will,require periodic maintenance and has
accounted for instrument maintenance in the annual'cost of implementing
the leak detection and repair program, including the cost of a spare
instrument. Nevertheless, EPA agrees that small facilities (plants
with few equipment pieces) may incur higher costs per emission reduction
and, therefore, as noted in Section 4.1 and D.3, small plants are
exempted from RACT.
Comment:
One commenter (8) wrote that soap scoring should be allowed as a
VOC detection method.
D-7
-------
Response:
Soaping is permitted as a preliminary screening technique on
certain equipment pieces as discussed in Section 3.1.1.
Comment:
Several comments (6, 8, 10 and 11) were received stating that EPA
should extend the 15-day repair interval. One commenter suggested
that repair should be completed during the next regular maintenance
period, while others suggested repair within 30 days and 60 days for
repairs that require hard to get parts.
Response:
The 15-day repair interval was selected for RACT because it
allows operators sufficient time to accomplish repairs while achieving
effective emission reduction. Most repairs can be completed quickly,
while a few may take up to 15 days. Repair intervals beyond 15 days
reduce the effectiveness of emission reductions and do not substantially
improve the efficiency in handling complex repair tasks. If repair is
not technically feasible without shutting down the process unit,
repair may be delayed until the equipment can be isolated for repair
or during the next scheduled process unit turnaround.
Comment:
Another commenter (10) stated that unsafe and difficult-to-monitor
components should be considered.
Response:
Guidelines are included in the CT6 for less frequent monitoring
of equipment pieces that are difficult-to-monitor. Guidelines, however,
do not address unsafe components because such equipment components are
not found in gas plants.
Comment;
Two commenters (8 and 10) wrote that EPA should consider annual
inspections because existing data (Eaton, 1980) dispute quarterly
monitoring for all valves, pumps, and relief valves. Quarterly inspections
are practical only for a relatively small number of major components
(e.g. compressors).
Response:
EPA data and models presented in the EPA report, "Fugitive Emission
Sources of Organic Compounds -- Additional Information on Emissions,
D-8
-------
Emission Reductions, and Costs", EPA 450/3-82-010, April 1982, (AID),
show quarterly monitoring to be reasonable as discussed in Chapter 5.
The Eaton data, as discussed in Appendix A, support the data used in
the .analyses.3
Comment:
For the "skip-period" monitoring alternative work practice for
valves, one commenter (10) suggested that EPA should allow less frequent
than annual inspections .for valves if data indicate that a longer
interval would keep the leak rate less than 2 percent. This, concept
should apply to other fittings as well as valves.
Response:
In developing skip-period monitoring, EPA did not consider inspection
intervals longer than one year. In skip-lot sampling theory it is
assumed that failures do not accumulate with time. For skip period
monitoring, it is likely that leaks .that occur will not be detected
and will accumulate. EPA does not feel it is reasonable to allow
leaks to accumulate for greater than one year.4 Facilities with very
low leak percentages may., however, elect to comply with the allowable
percent 'leaking alternative.
"Skip" monitoring is not allowed for other sources because there
are not enough other sources present for the statistics of skip monitoring
to apply. In addition, leaks from these other sources are not as
predictable as leaks from valves. Valves develop leaks"slowly over
time with small percent increases over a given time interval, whereas
other sources might operate with low leak rates for long periods of
time and then fail instantaneously with sudden increases in leak
rates. Consequently, no matter how many consecutive successful inspections
are performed, there is little assurance that a low leak rate would
continue if skipping were allowed.
Comment:
The same commenter (10) also questioned emissions reduction
estimates stating that 100 percent emission reduction (Table 4-2) for
compressor controls is not realistic. Chevron's leak detection and
repair program of quarterly inspections reduces emissions by 35 percent.
D-9
-------
Response:
The assumption that compressor controls reduce emissions by
100 percent is based on enclosing compressors and venting emissions to
a control device rather than on leak detection and repair programs.
EPA has recalculated the impacts of RACT based on quarterly leak
detection and repair (assuming gas plants can use leak detection and
repair) and determined an emission reduction of 83 percent. EPA
maintains that enclosing and venting seal emissions to a control
device will result in essentially 100 percent control.
Comment:
Another commenter (11) remarked that achieving zero emissions
from relief valves is difficult due to the metal-to-metal seat construc-
tion. Testing relief valves for leaks by using hydrocarbon detectors
may be unrealistic because even minor leakage into the relief valve
stack might give a high hydrocarbon reading, particularly if the
hydrocarbon vapor is heavier than air.
Response:
RACT for safety relief valves does not require zero emissions as
the commenter implied. Rather, quarterly leak detection and repair is
required which results in approximately a 63 percent VOC (69 percent
THC) emission reduction efficiency. Also, relief valves may be designed
to utilize an elastomeric 0-ring seat as a backup to the conventional
metal-to-metal seat while any leakage is controlled by the elastomeric
0-ring seat. Relief valves with 0-ring seats have been tested and
found to be bubble tight up to over 95 percent of set pressure and to
reseat to this condition through several cycles. Finally, Method 21
specifies that relief valves be monitored at (and not in) the relief
valve opening (horn), so minor leakage should not be detected.
Comment:
One commenter (6) noted that pressure relief valves can be vented
to a plant flare where VOC would be combusted.
Response;
A flare or other VOC control device (i.e., process heater, carbon
adsorption unit, refrigeration unit, gas recovery compressors) can be
used to effectively control relief valve leakage and are allowed under
RACT.
D-10
-------
Comment:
One commenter (11) expressed concern that the removal of pressure
relief valves may result in large emissions from depressurizing vessels
at a cost of hundreds of dollars per valve (thousands of dollars per
inaccessible valve).
Response:
Three-way valves or block valves may already be in place to
isolate pressure relief valves for repair on-line without depressuring
the unit. If so, repair within 15 days should be accomplished.
However, if pressure relief devices cannot be isolated for repair on
line, repair can be delayed until the next process unit turnaround.
D.4.3 Technology Transfer
Comment:
One commenter (7) wrote that there is no technical basis for the
transferability of chemical plant or refinery VOC emissions data to
natural gas plants because the processes, feedstocks, operating temperature,
operating pressures, vibrational problems and product compositions are
different; EPA should address these differences and give supporting
data for technology transfer.
Response:
EPA recognizes that differences exist between chemical and refinery
process units and gas plants; however, these differences do not preclude
the transfer of control technology to the gas processing industry. In
testing conducted in ethylene plants, process conditions approximate
that of equipment pieces in cryogenic units of gas plants. In addition,
only a small proportion of gas plant equipment are subject to conditions
which are unlike that of chemical or refinery process units. Finally,
the parameters addressed by the commenter Either do not affect the
frequency of emissions :or are.unquantifiable. In API testing of the
natural gas production industry, the process type, operating temperature
and pressure, and line size were determined to be unrelated to equipment
leaks early in the testing program; therefore, the .recording of these
data was discontinued. Feedstocks and product chemical types have
been found to be important only in terms of vapor pressure; as a
result, heavy liquids have been exempt from routine monitoring.
Vibrational problems and other site-specific differences are
unquantifiable.
D-ll
-------
D.4.4 Gas Operated Control Valves
Comment:
One commenter (10) stated that there are no emission factors
given for gas operated control valves nor indication of how many are
used. There are very little nonmethane/nonethane hydrocarbons present
in gas operated control valves. Similarly, another commenter (9)
argued that because many remote natural gas gathering stations use gas
operated control valves, compressed air is not acceptable RACT.
Response:
Gas operated control valves normally use air. In those instances
in which gas is used to operate control valves, dry gas (methane and
ethane) is normally used. Since RACT exempts dry gas service equipment,
the recommendation for controlling gas operated control valves is
deleted from RACT. Further, the RACT recommendations are for gas
plants and not for natural gas gathering stations.
D.5 MODEL PLANTS
Comment:
Two commenters (8 and 10) questioned the model plants. One
remarked that EPA should have included all four of the EPA-tested gas
plants, or EPA should explain the reason for including only two of the
plants in the vessel and component inventories. In addition, the
component inventories at the two API-tested plants are unusually large
and of questionable value in developing model plant configurations.
Another wrote that the method of ratioing components to all vessels
combined (columns, heat exchangers, drum/tanks) is an'oversimplification,
as indicated by Table B-3 of the draft CT6.
Response:
The model plants are based on four rather than on six plant
visits because the last two EPA-tested plants were visited after the
model plants were derived. Furthermore, as stated in Appendix B, the
latter two plant visits did not obtain information on vessel or equipment
inventories. The purpose of model plants is to characterize the range
of processing complexity. The diversity within the gas production
industry is represented in the four gas plants were examined and is shown
in the three model plants selected. In addition, the cost effectiveness of
D-12
-------
RACT controls are independent of the number of pieces of equipment
because there are.no economies of scale for leak detection and repair
programs.
D.6 ENVIRONMENTAL IMPACT
Comment:
Two commenters (2 and 6), maintained that the test data base on
which emission reduction estimates are based is limited (only 6 plants).
They contended that the test data are not statistically sound and
should be expanded to obtain a more representative sample.
Response:
Emissions test data are used to estimate the magnitude of fugitive
emissions and the magnitude of potential emission reductions through
the application of reasonably available control techniques. For this
purpose, the emissions test data obtained from gas plants indicate
that significant emissions are released to atmosphere from leaking
equipment and that implementation of the ,RACT requirements will reduce
these emissions.-
i .
Comment:
Two commenters (5 and 6) noted that other factors should be .
considered (besides the number of components) in estimating, emissions
(e.g., system pressure, equipment age, climate, past performance, gas
composition, differences in plant type, size, total fluid mix, etc.)
Response: -
EPA has conducted numerous equipment emissions studies at petroleum
refineries, synthetic organic chemical manufacturing plants , gas
plants, coke oven by-product plants, etc., as discussed in detail in
"Fugitive Emission Sources of Organic Compounds—Additional Information
on Emissions, Emission Reductions, and Costs." U.S. Environmental
Protection Agency, Research Triangle Park. EPA-450/3-82-010. April
1982. The major conclusions drawn from these studies are that the
only equipment or process variable found to correlate with fugitive
emissions was the volatility and/or phase of the process stream.
Consistent with to this finding, RACT for gas plants exempts equipment
that contact or contain heavy liquid VOC. Other variables such as
line temperature and pressure indicated much lower degrees of correlation.
D-13
-------
Comments:
Another commenter (6) wrote that the method for obtaining emission
rates needs to be described in more detail. The commenter specifically
questioned how EPA derives emission factors from concentrations measured
by a hydrocarbon analyzer without measuring actual flow rate.
Response:
Equipment emission .rates were determined by enclosing the emission
sources and measuring mass emissions. The screening values, simul-
taneously measured, were correlated to the measured leak rate. The
derivation of emission factors is presented in "On-shore Production of
Crude Oil and Natural Gas-Fugitive Volatile Organic Compound Emission
Sources-Data Analysis Report Frequency5of Leak Occurrence and Emission
Factors for Natural Gas Liquid Plants," U.S. Environmental Protection
Agency, Research Triangle Park, N.C. EMB Report No. 80-FOL-l. July
1982. The emission factor development methodology was reviewed by
industry representatives and they determined that EPA's methods for
emission factor development were appropriate.5
Comment:
One comment was received (11) stating that it is wrong to assume
that 100 percent of emissions would be reduced by controls on open-ended
lines. In many-cases, open-ended vents would be routed to a flare, and
energy recovery credits would not be applicable.
Response:
Capping an open-ended line limits emissions to .the amount of VOC
trapped between the valve and cap (whether it be a plug, second valve,
etc.) which is released when the line is again opened. However, by
closing the first valve prior to capping or closing the second valve,
the safety of the technique is ensured and emissions minimized. A
conservative estimate of the amount of VOC trapped in the line and the
frequency of open-ended line use, nevertheless, results in almost
100 percent VOC control (as discussed in Section 3.2.2). RACT for
open-ended lines effectively controls emissions between each time the
line is used.
D-14
-------
It appears as though the commenter has incorrectly assumed that
process vents would need to be capped. Any open-ended line that is in
use, a pressure relief valve, a double block and bleed line, or a
process vent, would not be capped. Process vents could be routed to a
flare, with no recovery credit as the commenter states, but this is
not part of th-e requirements of RACT.
D.7 COST IMPACTS
Comment:
*» ' - " - , -
Commenters (1 and 8) wrote that front end costs of RACT should
not be combined with capital costs. For example, double valving
open-ended lines and initial leak repair are front end expenses that
should be considered as operating costs, and not capital costs to be
amortized (and thus minimizing the impact of their expense in the year
when they are incurred).
Response:
Although the control cost of open-ended lines and.initial leak
repair could be treated as operating expenses because they are one-time
start-up costs, for the purposes of this CTG they are treated as
though they were capital costs and amortized. This assumes capital
would be borrowed to pay these initial costs.
Comment:
Two commenters (8 and 10) remarked that estimates for monitoring
times do not apply to gas plants and the costs are outdated. Monitoring
labor charges of $4/source for contractor labor and $3.50/source for
plant personnel (not including leak repair, resampling after repair,
or initial design, acquisition, or implementation of the monitoring
network), as well as the current labor rate of $23/hr (as opposed to
EPA estimate of $18/hr) were offered. It was also argued that labor
costs for leak detection should include: front end set-up cost,
equipment depreciation, and instrument maintenance. In addition, the
costs apply to ideal work conditions and new materials, and the costs
at smaller plants will be much higher per emission reduction because
they would have minimal attendance and would be forced to call upon
outside assistance in implementing RACT.
D-15
-------
Response:
The monitoring time estimates for plant equipment are based on
the results of refinery inspections and have been corroborated in
chemical plant testing. The EPA labor rate ($18/hr) is based on June
1980 dollars. Updating the EPA estimate to present (June 1982) dollars
results in a labor rate that exceeds the rate suggested ($18 x June 1982
Cost Index 295.9/June 1980 Cost Index 210.5 = $26. Reference: Chemical
Engineering 87(20):7 and 89(19):7). Set-up costs, equipment depreciation,
and instrument maintenance costs are included in the cost-analysis.
Leak detection costs account for field labor time only. Administrative,
support, and instrument costs to implement RACT are itemized separately.
The leak detection and repair costs are based on field monitoring
under all weather conditions. For model plant B, EPA's estimated
costs fall within the range of costs the commenter quotes. With
750 valves maintained at 2-man-minutes per inspection and one fourth
the annual instrument cost of $5,500, the cost per valve inspection is
$2.67. Using the above cost indices this would update to $3.75 per
source.
The EPA agrees, however, that small plants may incur higher costs
per emission reduction if outside personnel are relied upon to conduct
the leak detection and repair program. Chapter 4, therefore, recommends
a small plant exemption from RACT (Section 4.1).
Comment:
One commenter (8) wrote that the costs for adding double valves
on open-ended lines are underestimated because these costs should
include: recordkeeping, vehicle use, and source identification and
tagging. In addition the commenter wrote that the cost estimate for
capping open-ended lines is based on the price of a one-inch screw-on
type globe valve and the incorrect assumption that any lines larger
than one inch can be reduced to one inch. The commenter suggested
that EPA should review the 721 open-ended lines tested as reported in
the CTG-Appendix A and base the costs on a distribution of line sizes.
Response:
Double valving an open-ended line does not require additional
recordkeeping or tagging. The second valve is not subject to the •
D-16
-------
valve leak detection and repair requirements, but is considered as
RACT. Complying With RACT, however, does not necessitate a second
valve. Open-ended lines may be capped ,or plugged. The basis for the
cost estimate is the price of a one-inch screw-on type globe valve
which reflects the maximum cost likely incurred for open-ended line
emissions control. Larger lines would likely have a blind flange
installed at a similar cost, and smaller lines would be capped at a
much lower cost.
Comment:
Similarly, the same commenter (8) stated that leak repair costs
are too low. Costs should include: recordkeeping, vehicle use, provisions
for inaccessibility, repair parts, loss of production and overtime.
Cost of $120/repaired valve and $1000/repaired pump are realistic
repair costs.
Response;
In Chapter 5, a $140/seal replacement cost is included in the
cost of pump repair. The cost analysis also includes annual miscellaneou:
(0.04 x capital cost) and maintenance (0.05 x capital cost) costs plus
an annual calibration and maintenance cost for the monitoring equipment
of $3000 (1980 dollars). Administrative and support costs to implement
RACT (0.40 x monitoring labor + maintenance labor) are also included.
Very few valves will require repacking. Chapter 4 includes provisions
for less frequent monitoring of difficult to monitor valves and repairs
that cannot be completed on-line. These repairs may be delayed until
the next scheduled shutdown.
Comment;
One respondent (11) remarked that monitoring time and costs are
not included for flanges and connections.
Response:
Flanges and connections need not be monitored routinely under
RACT. Therefore, there are no monitoring time or costs associated
with it.
Comment;
One commenter (10) was concerned that RACT compressor control-
costs are too high in consideration of their small proportion to total
emissions. The venting system would require extra valves for safety.
D-17
-------
The entire costs for the control system will exceed $700, and the cost
of a small flare is about $8000.
Response:
Section 5.5 presents revised control costs for controlling compressors
based on public comments received on the enclosed compressor vent
system. The cost effectiveness for the enclosed compressor vent
control system also-reflects revised emission factors for compressor
seals. The revised compressor and emission factors are based on
wet gas and natural gas liquids compressors.1 Data from dry gas service
compressors was excluded because dry gas service components are exempt
from RACT. Based on the revised cost effectiveness of enclosing
compressors, wet gas reciprocating compressors at facilities that do
not have a VOC control device are exempt from RACT.
Comment:
Similarly, another commenter (9) expressed concern that the cost
of a VOC monitoring instrument, its maintenance, and calibration are
high compared to current practices of leak detection.
Response:
The draft CT6 has included the cost of two monitoring instruments,
instrument maintenance calibration time, and two-man monitoring teams
to obtain maximum cost impacts from implementation of RACT. In Chapter 5,
the costs of RACT including the maximum instrument costs are shown to
be reasonable. Actual plant costs incurred may be much less because
one man monitoring teams may be employed and less expensive monitoring
instruments may be purchased. Also, many equipment pieces will not
require instrument monitoring if soaping is used as a preliminary
screening technique.
Comment:
One commenter (8) pointed out that the value of recovery credits
for VOC ($210/Mg) is incorrectly based on the assumption that all the
VOC is propane. Also, the value of the recovery credits .is $146 if
the correct product density for propane is used.
Response:
Recovery credits have been revised. Nonmethane/nonethane hydrocarbons
are valued at $192/Mg based'on LPG price of 40
-------
and a specific gravity of 0.55 (the original Incorrect, .credit of
$210/Mg was/based on a specific gravity .-of 0.50), Methane and ethane
are valued at $61/Mg based on $1.46/Mcf, of natural gas for June 1980,
-assuming-a weight equivalent composition of 80 percent methane and
20 percent ethane at standard temperature and pressure.
Comment: •-•,..;.-... ..-.:•-••
One commenter (10) stated that more frequent inspections than
annual are not cost effective. • ;
Response:
EPA has determined that quarterly leak detection, and repair .is
cost effective and represents reasonably'available control technology.
In Chapter 5, Table 5-10 presents the cost effectiveness of quarterly
leak detection and repair for valves, pumps, relief valves, and compressors,
Comment:
One commenter (10) wrote that incremental cost-effectiveness
figures should be calculated for different inspection intervals.
Response:
The purpose of control technique guidelines (CTG) is to inform
air pollution control agencies responsible for achieving and maintaining
national ambient air quality standards of reasonably available control
technology (RACT). These agencies may formulate their own regulations
based upon the CTG; however, the CTG itself is not intended to evaluate
alternative control strategies and the incremental cost effectiveness
between the alternatives.
D-19
-------
D.8 References
1. Memorandum. K.C. Hustvedt EPA:CPB to James F. Durham, Revised
Gas Plant Compressor Seal Emission Factor. February 10, 1983.
2. Memorandum, Kent C. Hustvedt CPB:EPA to James F. Durham, CPB:EPA,
Estimated Compressor Seal VOC Emissions Contribution for API and
EPA Tested Gas Plants. March 22, 1983.
3. Memorandum, K.C. Hustvedt, EPA:CPB, to J.F. Durham EPA:CPB,
API/Rockwell Maintenance Data. December 9, 1982.
4. Memorandum, Kent C. Hustvedt CPBrEPA to James F. Durham CPB:EPA
Skip Monitoring for Fugitive Emission Sources. December 14,
1981.
5. Letter with attachment from D.A. DuBose, Radian Corporation, to
W.E. Kelly, EMB:EPA, July 22, 1982 - attachments are Radian and,
TRW report, of the January 28, 1982 meeting with API on gas plant
emission factors.
D-20
-------
TECHNICAL REPORT DATA
(Please read Instructions on the reverse before completing)
1. REPORT NO. • „ _
EBA-450/ 3-83-007
2.
3. RECIPIENT'S ACCESSION NO.
4. TITLE AND SUBTITLE
Control of Volatile Organic Compound Equipment Leaks
from Natural Gas/Gasoline Processing Plants-
Guideline Series
5. REPORT DATE
December 1983
6. PERFORMING ORGANIZATION CODE
7. AUTHOR(S)
8. PERFORMING ORGANIZATION REPORT NO.
9. PERFORMING ORGANIZATION NAME AND ADDRESS
Office of Air Quality Planning and Standards
Environmental Protection Agency
Research Triangle Park, North Carolina 27711
10. PROGRAM ELEMENT NO.
11. CONTRACT/GRANT NO.
68-02-3511
12. SPONSORING AGENCY NAME AND ADDRESS
Director of Air Quality Planning and Standards
Office of Air Quality Planning and Standards
U.S. Environmental "rotection Agency
Research Triangle Park, North Carolina 27711
13. TYPE OF REPORT AND PERIOD COVERED
14. SPONSORING AGENCY CODE
EPA/200/04
15. SUPPLEMENTARY NOTSS
16. ABSTRACT
Control Technique Guidelines (CTG) are issued for volatile organic compound
(VOC) equipment leaks from natural gas/gasoline processing plants
to inform Regional, State, and local air pollution control agencies of reasonably
available control technology (RACT) for development of regulations necessary to
attain the national ambient air quality standard for ozone. This document contains
information on RACT environmental and cost impacts.
17.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
b.lOENTlFIERS/OPEN ENDED TERMS
c. COSATI Field/Group
Air Pollution
Gas/gasoline Processing Plants
Pollution Control
Reasonably Available Control Technology
Volatile Organic Compounds (VOC)
Air Pollution Control
13b
18. DISTRIBUTION STATEMENT
Unlimited
19. SECURITY CLASS (This Report)
Unclassified
21. NO. OF PAGES
173
20. SECURITY CLASS (Thispage)
Unclassified
22. PRICE
SPA Form 2220—1 (Rev. 4—77) -PREVIOUS EDITION is OBSOLETE
-------
------- |