United States
Environmental Protection
Agency
Office of Air Quality
Planning and Standards
Research Triangle Park NC 27711
August 1981
Air
IvvEPA Guideline Series
Control of Volatile Organic
Compound Fugitive
Emissions from Synthetic
Organic Chemical, Polymer,
and Resin Manufacturing
Equipment
Draft
-------
NOTICE
This document has not been formally released by EPA and should not now be construed to represent
Agency policy. It is being circulated for comment on its technical accuracy and policy implications.
Control of Volatile Organic Compound
Fugitive Emissions from Synthetic
Organic Chemical, Polymer,
^
Resin Manufacturing Equipment
Emission Standards and Engineering Division
Contract No. 68-02-3168
Prepared for
U.S. ENVIRONMENTAL PROTECTION AGENCY
Office of Air, Noise, and Radiation
Office of Air Quality Planning and Standards
Research Triangle Park, North Carolina 27711
August 1981 •
-------
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.
11
-------
TABLE OF CONTENTS
Page
List of Tables v
List of Figures viii
Chapter 1.0 Introduction 1-1
Chapter 2.0 Processes and Pollutant Emissions 2-1
2.1 Introduction 2-1
2.2 Facilities and Their Emissions 2-3
2.3 Model Units 2-17
Chapter 3.0 Emission Control Techniques. .' 3-1
3.1 Leak Detection and Repair Methods 3-1
3.2 Other Control Strategies 3-14
3.3 Other Considerations 3-18
Chapter 4.0 Environmental Analysis of RACT 4-1
4.1 Introduction 4-1
4.2 Air Pollution 4-2
4.3 Water Pollution 4-6
4.4. Solid Waste Disposal 4-6
4.5 Energy 4-6
Chapter 5.0 Control Cost Analysis of RACT 5-1
5.1 Basis for Capital Costs 5-1
5.2 Basis for Annualized Costs 5-3
5.3 Emission Control Costs 5-8
5.4 Cost Effectiveness 5-12
m
-------
TABLE OF CONTENTS (continued)
Page
Appendix A. Emission Source Test Data A-l
Appendix B. List of Chemicals Defining Synthetic Organic
Chemical, Polymer, and Resin Manufacturing
Industries B-l
Appendix C. Method 21. Determination of Volatile Organic
Compound Leaks C-l
Appendix D. Example Calculations for Determining Reduction
in Emissions from Implementation of RACT D-l
iv
-------
LIST OF TABLES
Page
Table 2-1 Fugitive Emission Sources for Three Model Units . . 2-19
Table 2-2 Uncontrolled Fugitive Emission Factors in Process
Unit Equipment 2-21
Table 2-3 Estimated Total Fugitive Emissions from Model
Units 2-22
Table 2-4 Average Percent of Total Fugitive Emissions Attributed
to Specific Component Types 2-23
Table 3-1 Percentage of Emissions as a Function of Action
Level 3-6
Table 3-2 Estimated Occurrence and Recurrence Rate of Leaks
for a Quarterly Monitoring Interval 3-7
Table 3-3 Average Emission Rates from Sources Above
10,000 ppmv and at 1000 ppmv 3-9
Table 3-4 Impact of Monitoring Interval on Correction Factor
Accounting for Leak Occurrence/Recurrence (For Example
Calculation) 3-12
Table 3-5 Example of Control Efficiency Calculation 3-13
Table 3-6 Cost Effectiveness Versus Initial Percent of
Valves Leaking in Model Units 3-16
Table 3-7 Illustration of Skip-Period Monitoring 3-19
Table 3-8 Cost Effectiveness of Quarterly Leak Detection
and Repair for Typical Process Units 3-22
Table 4-1 Estimated Hourly Emissions and Emissions Reduction
on a Model Unit Basis 4-3
Table 4-2 Estimated Annual Emissions and Emissions Reduction
on a Model Unit Basis 4-3
Table 4-3 Emission Factors for Sources Controlled Under RACT . 4-4
Table 4-4 Example Calculation of VOC Fugitive Emissions from
Model Unit A Under RACT 4-5
-------
Table 5-1
Table 5-2
Table 5-3
Table 5-4
Table 5-5
Table 5-6
Table 5-7
Table 5-8
Table 5-9
Table 5-10
Table 5-11
Table A-l
Table A- 2
Table A-3
Table A-4
Table A- 5
Table A-6
Table A-7
Table A-8
Capital Cost Data
Capital Cost Estimates for Implementing RACT . . .
Labor-Hour Requirements for Initial Leak Repair
Under RACT
Basis for Annual i zed Cost Estimates
Annual Monitoring and Leak Repair Labor Requirements
for RACT
Recovery credits
Annuali zed Control Cost Estimates for Model Units
Under RACT
Cost Effectiveness for Model Units under RACT . . .
Cost Effectiveness for Component Types in
Model Unit A
Cost Effectiveness for Component Types in
Model Unit B
Cost Effectiveness for Component Types in
Model Unit C
Frequency of Leaks from Fugitive Emission Sources in
Synthetic Organic Chemical Units
Twenty-four Chemical Process Units Screened for
Fugitive Emissions
Summary of SOCMI Process Units Fugitive Emissions .
Average Fugitive Emission Source Screening Rates. .
Sampled Process Units from Nine Refineries During
Refinery Study
Leak Frequencies and Emission Factors from Fugitive
Emission Sources in Petroleum Refineries
Comparison of Leak Frequencies for Fugitive Emission
Sources in SOCMI Units and Petroleum Refineries . .
Frequency of Leaks from Fugitive Emission Sources
in Two DuPont Plants
Page
5-2
5-2
5-4
5-5
5-7
5-9
5-11
5-13
5-14
5-15
5-16
A-4
A-5
A-8
A-9
A-10
A-l 2
A-l 3
A-l 4
VI
-------
Page
Table A-9 Frequency of Leaks from Fugitive Emission Sources
in Exxon's Cyclohexane Unit A-16
Table A-10 Summary of Maintenance Study Results from the Union
Oil Co. Refinery in Rodeo, California A-18
Table A-ll Summary of Maintenance Study Results from the Shell
Oil Company Refinery in Martinez, California. . . . A-20
Table A-12 Summary of EPA Refinery Maintenance Study Results . A-21
Table A-13 Maintenance Effectiveness Unit D Ethylene Unit
Block Valves A-22
Table A-14 Occurrence Rate Estimates for Valves and Pumps
by Process in EPA-ORD Study A-24
Table A-15 Valve Leak Recurrence Rate Estimates A-25
Table A-16 Summary of Valve Maintenance Test Results A-26
Table D-l Uncontrolled Fugitive Emission Factors in Process
Unit Equipment D-2
Table D-2 Controlled Emission Factors for Equipment Affected
by RACT D-2
Table D-3 Example Calculation of Uncontrolled Emissions from
an Illustrative Process Unit D-3
Table D-4 Uncontrolled Emissions from Components Affected
by RACT D-4
Table D-5 Controlled Emissions from Components Affected by
RACT D-4
Table D-6 Emission Reduction Expected from RACT D-5
vn
-------
LIST OF FIGURES
Page
Figure 2-1 General schematic of process levels that make up
the organic chemical industry 2-2
Figure 2-2 Diagram of a simple packed seal 2-4
Figure 2-3 Diagram of a basic single mechanical seal 2-5
Figure 2-4 Diagram of a double mechanical seal 2-6
Figure 2-5 Diagram of a double mechanical seal 2-6
Figure 2-6 Diaphragm pump 2-7
Figure 2-7 Labyrinth shaft seal 2-9
Figure 2-8 Restrictive-ring shaft seal 2-9
Figure 2-9 Mechanical (contact) shaft seal 2-10
Figure 2-10 Liquid film shaft seal with cylindrical bushing . . . 2-10
Figure 2-11 Diagram of a gate valve 2-11
Figure 2-12 Example of bellows seals 2-12
Figure 2-13 Diagrams of valves with diaphragm seals 2-13
Figure 2-14 Diagram of a spring-loaded relief valve 2-14
Figure 2-15 Diagram of hydraulic seal for agitators 2-16
Figure 2-16 Diagram of agitator lip seal 2-16
Figure 3-1 Cumulative distribution of total emissions by screening
values - valves in light liquid service 3-10
Figure 3-2 Cumulative distribution of sources by screening
values - valves in light liquid service 3-10
Figure C-l Calibration Precision Determination C-8
Figure C-2 Response Time Determination C-9
vm
-------
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 attainment 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 are 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," (44 FR 20372, April 4, 1979) for ozone SIP's, EPA permitted States
to defer the adoption of RACT regul-ations 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
sound 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
unique 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
analysis of RACT for specific stationary sources.
1-1
-------
2.0 PROCESSES AND POLLUTANT EMISSIONS
2.1 INTRODUCTION
The discussion presented in this document applies to equipment in
process units which manufacture synthetic organic chemicals and polymers and
resins. The equipment in process units in the synthetic organic chemical
manufacturing industry (SOCMI) is similar to equipment in the polymer and
resin manufacturing industry. Both industries process volatile organic
compounds. Therefore, the information and discussion presented in this
chapter and subsequent chapters applies equally to SOCMI plants and polymer
and resin plants.
The SOCMI is a segment of the chemical industry consisting of some of
the higher volume intermediate and finished products. A list of these
chemicals is presented in Appendix B, Table I. The polymer and resin
manufacturing industries to which the discussion in this document applies are
presented in a list in Appendix B, Table II. It should be emphasized that
the discussion in this document are intended to apply to equipment in process
units which manufacture these chemicals.
Most of the SOCMI chemicals produced in the United States are derived
from crude petroleum or natural gas. The ten principal feedstocks used in
the manufacture of organic chemicals are produced primarily in petroleum
refineries. After chemical feedstocks are manufactured from petroleum,
natural gas, and other raw materials, they are processed into chemical
intermediates and end-use chemicals (see Figure 2-1). Approximately 12 percent
of the plants in the United States produce less than 5,000 megagrams (Mg)
annually. Another 12 percent have production capacities in excess of
500,000 Mg.
2-1
-------
RAW MATERIALS
(CRUDE OIL, CRUDE NATURAL GAS, ETC.)
CHEMICAL
REFINERIES | FEEDSTOCK
| PLANTS
CHEMICAL
FEEDSTOCKS
CHEMICAL
PLANTS
CHEMICAL
PRODUCTS
Fi gure 2-1.
General schematic of process levels that make up
the organic chemical industry.
2-2
-------
The polymer and resin manufacturing industry includes operations which
convert monomer or chemical intermediate materials obtained from the basic
petrochemical industry and the SOCMI into polymer products. Such products
include plastic materials, synthetic resins, and synthetic rubbers.
2.2 FACILITIES AND THEIR EMISSIONS
2.2.1 Potential Source Characterization and Description
In this document, fugitive emissions from process units are considered
to be those volatile organic compound (VOC) emissions that result when
process fluid (either gaseous or liquid) leaks from plant equipment. There
are many potential sources of fugitive emissions in a typical process unit.
The following sources will be considered in this chapter: pumps, compressors,
in-line process valves, pressure relief devices, open-ended valves, sampling
connections, flanges, agitators and cooling towers. These potential sources
are described below.
2.2.1.1 Pumps. Pumps are used extensively in process units for the
movement of organic 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, are also used.
Chemicals transferred by pumps can leak at the point of contact between the
moving shaft and stationary casing. Consequently, all pumps except the
shaftless type (canned-motor and diaphragm) require 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 Deterioration of the packing will result
in process liquid leaks.
2-3
-------
End C
— i Stuffing
Box
L * nn
U
CxixxixMxExl
Packing
/ Gland
rf Fnrf
1XIXDOXIXIXIX1 ,
/ 1
r / UJd
Packing
\
/ Possible
Leak
Area
Figure 2-2. Diagram of a simple packed seal.3
Mechanical seals are limited in application to pumps with rotating
shafts and can be further categorized as single and double mechanical 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.
A mechanical seal is not a leak-proof device. Depending on the condition
and flatness of the seal faces, the leakage rate can be quite low (as small
as a drop per minute) and the flow is often not visually detectable. In
order to minimize fugitive emissions due to seal leakage, an auxiliary
4
sealing device such as packing can be employed.
2-4
-------
PUMP
STUFFING
BOX
GLAND
•RING
FLUID
END
SHAFT
\ROTATING
SEAL RING
STATIONARY
ELEMENT
POSSIBLE
LEAK AREA
Figure 2-3. Diagram of a basic single mechanical seal.
In a dual mechanical seal application, two seals can be arranged
back-to-back or in tandem. In the back-to-back arrangement (Figure 2-4), the
two seals provide a closed cavity between them. A seal liquid, such as water
or seal oil, is circulated through the cavity. Because the seal liquid
surrounds the double seal and lubricates both sets of seal faces in this
arrangement, the heat transfer and seal life characteristics are much better
than those of the single seal. In order for the seal to function, the seal
liquid must be at a pressure greater than the operating pressure of the
stuffing box. As a result some seal liquid will leak across the seal faces.
Liquid leaking across the inboard face will enter the stuffing box and mix
with the process liquid. Seal liquid going across the outboard face will
exit to the atmosphere. '
In a tandem dual mechanical seal arrangement (Figure 2-5), the seals
face the same direction. The secondary seal provides a backup for the
primary seal. A seal flush is used in the stuffing box to remove the heat
generated by friction. The cavity between the two seals is filled with a
buffer or barrier liquid. However, the barrier liquid is at a pressure lower
2-5
-------
SEAL LIQUID.
POSSIBLE LEAK
INTO SEALING
FLUID
FLUID END
w \/ o
PRIMARY — V
SEAL
\
*
— SECONDARY
SEAL
GLAND
PLATE
Figure 2-4. Diagram of a double mechanical seal
(back-to-back arrangement)7
PRIMARY
SEAL
SUFFER LIQUID
OUT IN
(TOP) (BOTTOM)
V
SECONDARY
SEAL
GLAND
PLATE
70-1787-1
Figure 2-5. Diagram of a double mechanical seal
(tandem arrangement)8
2-6
-------
than that in the stuffing box. Therefore, any leakage will be from the
stuffing box into the seal cavity containing the barrier liquid. Since this
liquid is routed to a closed reservoir, process liquid that has leaked into
the seal cavity will also be transferred to the reservoir. At the reservoir,
the process liquid could vaporize and be emitted to the atmosphere. To
ensure that VOC's do not leak from the reservoir, the reservoir can be
vented to a control device.
Another type of pump that has been used is the shaft!ess pump which
includes canned-motor and diaphragm pumps. In canned-motor pumps the cavity
housing the motor rotor and the pump casing are interconnected. As a result,
the motor bearings run in the process liquid and all seals are eliminated.
Because the process liquid is the bearing lubricant, abrasive solids cannot
be tolerated. Canned-motor pumps are being widely used for handling organic
solvents, organic heat transfer liquids, light oils, as well as many toxic or
hazardous liquids, or where leakage is an economic problem.10
Diaphragm pumps (see Figure 2-6) perform similarly to piston and plunger
pumps. However, the driving member is a flexible diaphragm fabricated of
metal, rubber, or plastic. The primary advantage of this arrangement is the
elimination of all packing and seals exposed to the process liquid. This is
an important asset when hazardous or toxic liquids are handled.11
DISCHARGE
CHECK VALVE
INLET
CHECK VALVE
DIAPHRAGM
PISTON
Figure 2-6. Diaphragm pump.
12
2-7
-------
2.2.1.2 Compressors. Gas compressors used in process units are
similar to pumps in that they can be driven by rotary or reciprocating
shafts. They are also similar to pumps in their need for shaft seals to
isolate the process gas from the atmosphere. As with pumps, these seals
are likely to be the source of fugitive emissions from compressors.
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 devices; none of them
completely eliminate leakage. Many compressors may be equipped with ports
in the seal area to evacuate gases collecting there.
The labyrinth type of compressor seal is composed of a series of close
tolerance, interlocking "teeth" which restrict the flow of gas along the
shaft. A straight pass labyrinth compressor seal is shown in Figure 2-7.
Many variations in "tooth" design and materials of construction are available.
Although labyrinth type seals have the largest leak potential of the different
types, properly applied variations in "tooth" configuration and shape can
13
reduce leakage by up to 40 percent over a straight pass type labyrinth.
Restrictive carbon ring seals consist of multiple stationary carbon
rings with close shaft clearances. This type of seal may be operated dry
or with a sealing fluid. Restrictive ring seals can achieve lower leak
rates than the labyrinth type. A restrictive ring seal is shown in
Figure 2-8.
Mechanical contact seals (shown in Figure 2-9) are similar to the
mechanical seals described for pumps. In this type of seal 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 described here, but they are
not suitable for all processing conditions.
Centrifugal compressors also can be equipped with liquid film seals. A
diagram of a liquid film seal is shown in Figure 2-10. The seal is formed by
a film of oil between the rotating shaft and stationary gland. When the
circulating oil is returned to the oil reservoir, process gas can be released
to the atmosphere.18 To eliminate release of VOC emissions from the seal oil
system, the reservoir can be vented to a control device.
2-8
-------
PORT MAY BE ADDED
FOR SCAVENGING OR
INERT-GAS SEALING
INTERNAL
GAS PRESSURE
'//'///'/.'///'/) ATMOSPHERE
Figure 2-7. Labyrinth shaft seal.
14
SCAVENGING
PORT MAY BE
AOOEO FOR
VACUUM
APPLICATION
ATMOSPHERE
Figure 2-8. Restrictive-ring shaft seal.
15
2-9
-------
INTERNAL
GAS PRESSURE
CLEAN OIL IN
PRESSURE
/• BREAKDOWN
.' SLEEVE
STATIONARY SEAT
CARSON RING
ATMOSPHERE
CONTAMINATED
OIL OUT
Figure 2-9. Mechanical (contact) shaft seal.
^- CLEAN OIL IN
INNER BUSHING OUTER BUSHING
ATMOSPHERE
CONTAMINATED
OIL OUT
OIL OUT
Figure 2-10. Liquid film shaft seal with cylindrical bushing.
2-.10
-------
2.2.1.3 Process Valves. One of the most common pieces of equipment in
organic chemical plants is the valve. The types of valves commonly used are
control, globe, gate, plug, ball, relief, and check valves. All except the
relief valve (to be discussed further below) and check valve are activated by
a valve stem, which may have either 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 2-11. 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. Va'lves that require the stem to move
in and out with or without rotation must utilize a packing gland. 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 tetrafluoroethyler
tetrafluoroethylene. The packing material used depends on the valve application
19
and configuration. These conventional packing glands can be used over a
wide range of operating temperatures. At high pressures these glands must be
20
quite tight to attain a good seal.
PACKING
GLAND
PACKING
VALVE
STEM
POSSIBLE
LEAK AREAS
Figure 2-11. Diagram of a gate valve.
21
2-11
-------
Elastomeric 0-rings are also used for sealing process valves. These
0-rings provide good sealing but are not suitable where there is sliding
motion through the packing gland. Those seals are rarely used in high pressure
service and operating temperatures are limited by the seal material.22
Bellows seals are more effective for preventing process fluid leaks than
the conventional packing gland or any other gland-seal arrangement.23 This
type of seal incorporates a formed metal bellows that makes a barrier between
the disc and body bonnet joint. An example of this seal is presented in
Figure 2-12. The bellows is the weak point of the system and service life
can be quite variable. Consequently, this type of seal is normally backed up
with a conventional packing gland and is often fitted with a leak detector in
24
case of failure.
BELLOWS
BODY
BONNET
Figure 2-12. Example of bellows seals.
25
A diaphragm may be used to isolate the working parts of the valve and
the environment from the process liquid. Tw.o types of valves which utilize
diaphragms are illustrated in Figures 2-13(a) and (b). As Figure 2-13(b)
shows, the diaphragm may also be used to control the flow of the process
fluid. In this design, a compressor component pushes the diaphragm toward
the valve bottom, throttling the flow. The diaphragm and compressor are
2-12
-------
connected in a manner so that it is impossible for them to be separated under
normal working conditions. When the diaphragm reaches the valve bottom, it
seats firmly against the bottom, forming a leak-proof seal. This configuration
is recommended for fluids containing solid particles and for medium-pressure
service. Depending on the diaphragm material, this type of valve can be used
at temperatures up to 205°C and in severe acid solutions. If failure of the
seal occurs, a valve employing a diaphragm seal can become a source of fugitive
emissions.
DIAPHRAGM
DISK
STEM
DIAPHRAGM
Figure 2-13. Diagrams of valves with diaphragm seals.
27
2.2.1.4 Pressure Relief Devices. Engineering codes require that
pressure-relieving 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 2-14). 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
28
is reattained, the valve reseats, and a seal is again formed. 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. Two
2-13
-------
potential causes of leakage from relief valves are: "simmering or popping",
a condition due to the system pressure being close to the set pressure of the
valve, and improper reseating of the valve after a relieving operation.29
Possible
Leak Area
Process Side
Figure 2-14. Diagram of a spring-loaded relief valve.
Rupture disks are also common 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's to escape from the system
under normal operation. However, when the disk does rupture, the system
depressurizes until atmospheric conditions are obtained. This could result
in an excessive loss of product or a corresponding excessive release of
fugitive emissions.
2-14
-------
2.2.1.5 Agitators. Agitators are commonly used to stir or blend
chemicals. Like pumps and compressors, agitators may leak organic chemicals
at the point where the shaft penetrates the casing. Consequently, seals
are required to minimize fugitive emissions from agitators. Four seal
arrangements are commonly used with agitators. These are compression
packing (packed seal), mechanical seals, hydraulic seals, and lip seals.31
Packed seals for agitators are very similar in design and application to
the packed seals for pumps (Section 2.2.1.1).
Although mechanical seals are more costly than the other three seal
arrangements, they offer a greatly reduced leakage rate to offset their
higher cost. The maintenance frequency of mechanical seals is, also,
32
one-half to one-fourth that of packed seals. In fact, at pressures
greater than 1135.8 kPa (150 psig), the leakage rate and maintenance frequency
' 33
are so superior that the use of packed seals on agitators is rare. As
with packed seals, the mechanical seals for agitators are similar to the
design and application of mechanical seals for pumps (Section 2.2.1.1.)
The hydraulic seal (Figure 2-15) is the simplest and least used agitator
shaft seal. In this -type of seal, an annular cup attached to the process
vessel contains a liquid that is in contact with an inverted cup attached
to the rotating agitator shaft. The primary advantage of this seal is that
it is a non-contact seal. However, this seal is limited to low temperatures
and pressures and can only handle very small pressure fluctuations. Organic
cherricals may contaminate the seal liquid and then be released into the
34
atmosphere as fugitive emissions.
INVERTED CUP
ANNULAR CUP
Figure 2-15. Diagram of a hydraulic seal for agitators.
2-15
35
-------
A lip seal (Figure 2-16) can be used on a top-entering agitator as a
dust or vapor seal. The sealing element is a spring-loaded elastomer. Lip
seals are relatively inexpensive and easy to install. Once the seal has
beer installed the agitator shaft rotates in continuous contact with the
lip seal. Pressure limits of the seal are 2 to 3 psi because it operates
without lubrication. Operating temperatures are limited by characteristics
of the elastomer. Fugitive VOC emissions could be released through this
seal when this seal wears excessively or the operating pressure surpasses
36
the pressure limits of the seal.
•*— —*•
37
Figure 2-16. Diagram of agitator lip seal.
2.2.1.6 Open-Ended Valves or Lines. Some valves are installed in a
systeir so that they function with the downstream line open to the atmosphere.
Examples are purge valves, drain valves, and vent valves. A faulty valve
seat or incompletely closed valve would result in leakage through the valve
and fugitive VOC emissions to the atmosphere.
2.2.1.7 Sampling Connections. The operation of a process unit is
checked periodically by routine analyses of feedstocks and products. To
obtain representative samples for these analyses, sampling lines must first
be purged prior to sampling. The purged, liquid or vapor is sometimes drained
onto the ground or into a sewer drain, where it can evaporate and release VOC
emissions to the atmosphere.
2-16
-------
2.2.1.8 Flanges. 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. Normally, flanges are employed
for pipe diameters for 50 mm or greater and are classified by pressure and
face type.
Flanges may become fugitive emission sources when leakage occurs due to
improperly chosen gaskets or a poorly assembled flange. 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.38
2.3 MODEL UNITS
This section presents model process unit parameters. The model units
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).
2.3.1 Model Units
Available data show that fugitive emissions are proportional to the
number of potential sources, but are not related to capacity, throughput,
39
age, temperature, or pressure. Therefore, model units defined for this
analysis represent different levels of process complexity (number of sources)
rather than different unit sizes.
2.3.1.1 Sources of Fugitive Emissions. Data from petroleum refineries
40
indicate that cooling towers are very small sources of VOC emission.
Differences in operating procedures, such as recirculation of process water,
might result in cooling tower VOC emissions, but no data are available to
verify this. The number of agitator seals in the industry is not known.
Furthermore, the emission rate from agitator seals has not been measured.
Since there are no data from similiar sources in other industries, no estimates
of emission rate can be made. Because of these uncertainties, cooling towers
and agitator seals are not included in the Model Units.
2.3.1.2 Model Units Components. In order to estimate emissions, control
costs, and environmental impacts for process units on a unit specific basis,
three model units were developed. The equipment components comprising the
2-17
-------
model units are shown in Table 2-1. These three model units represent the
range of emission source populations that may exist in SOCMI process units.
The number of equipment components for each model unit was developed from a
data base compiled by Hydroscience, Inc. The data base included equipment
source counts from 62 SOCMI plants which produce 35 different chemicals.
These plant sites represent approximately 5 percent of the total existing
SOCMI plants and include large and small capacities, batch and continuous
production methods, and varying levels of process complexity. The source
counts for the 35 chemicals include pumps, valves, and compressors. These
counts were used in combination with the number of sites which produce each
chemical in order to determine the average number of sources per site.
Hydroscience estimates that 52 percent of existing SOCMI plants are similar
to the Model Unit A, 33 percent are similar to B, and 15 percent are similar
to C.
Dcta from petroleum refineries indicate that emission rates of sources
decrease as the vapor pressure (volatility) of the process fluid decreases.
Three classes of volatility have been established based on the petroleum
refinery data. These include gas/vapor service, light liquid service, and
heavy liquid service. The split between light and heavy liquids for the
refinery data is between naphtha and kerosene. Since similar stream names
may have different vapor pressures, depending on site specific factors, it is
difficult to quantify the light-heavy split. The break point is approximately
at a vapor pressure of 0.3 kPa at 20°C. The data collected by Hydroscierce
were used to estimate the split between gas/vapor and liquid service for each
44
source. In order to apply emission factors for light and heavy liquid
service, it is assumed that one-half of SOCMI liquid service sources are in
light liquid service. There are no data available on the actual distribution
of sources in volatility ranges. It is assumed that all packed seal pumps
are in heavy liquid service. This assumption is reasonable, since more
volatile liquid are more suitable for mechanical seal applications, and newer
process units tend to use fewer packed seals. Sampling connections are a
subset, of the open-ended valve category. Approximately 25 percent of
45
open-ended valves are used for sampling connections. Emissions which occur
through the valve stem, gland, and open-end are included in the cpen-ended
valve category. The emission factor for sampling connection applies only to
emissions which result from sample purging.
2-18
-------
TABLE 2-1. FUGITIVE EMISSION SOURCES FOR THREE MODEL UNITS
Equipment component3
Pump seals
Light liquid service0
Single mechanical
Double mechanical
Sea 11 ess
Heavy liquid service0
Single mechanical
Packed
Valves
Gas service
Light liquid service
Heavy liquid service
Safety/relief valves
Gas service
Light liquid service
Heavy liquid service
Open-ended valves and lines6
Gas service
Light liquid service
Heavy liquid service
Compressor seals
Sampling connections
Flanges
Cooling towers
Number
Model unit
A
5
3
0
5
2
90
84
84
11
1
1
9
47
48
1
26
600
_-9
of components in
Model unit
B
19
10
1
24
6
365
335
335
42
4
4
37
189
189
2
104
2400
..9
model unit
Model unit
C
60
31
1
73
20
1117
1037
1037
130
13
14
115
581
581
8
320
7400
..9
aEquipment components in VOC service only.
b52 percent of existing SOCMI units are similar to model unit A.
33 percent of existing SOCMI units are similar to model unit B.
15 percent of existing SOCMI units are similar to model unit C.
cLight liquid is defined as a fluid with vapor pressure greater than 0.3 kPa
at 20°C. This vapor pressure represents the split between kerosene and
naphtha and is based on data presented in reference 39.
dHeavy liquid is defined as a fluid with vapor pressure less than 0.3 kPa
at 20°C. This vapor pressure represents the split between kerosene and
naphtha and is based on data presented in reference 39.
eSample, drain, and purge valves.
fBased on 25 percent of open-ended valves. Reference 1, pg. IV-3.
90ata not available.
2-19
-------
2.3.1.3 Uncontrolled Fugitive Emission Estimates. Data characterizing
the uncontrolled levels of fugitive emissions in the SOCMI are presently
unavailable. However, data on this type have been obtained for the refining
industry. These data are presented in Table 2-2, and represent the average
uncontrolled emission rate from each of the components of a specific type in
the process unit. Because the operation of the various process equipment in
the SOCMI is not expected to differ greatly from the operation of the same
equipment in the refining industry, the refinery fugitive emission data can
be used to approximate the levels of fugitive emissions in SOCMI.
The total amount of VOC emitted from fugitive sources can be estimated
for each Model Unit. Total hourly emissions can be calculated by multiplying
the number of pieces of each type of equipment (Table 2-1) by the corresponding
hourly emission factor (Table 2-2). The total annual emissions have been
calculated by multiplying the total hourly emissions for each Model Unit by
the number of hours in a yeer (8,760 hours/year). These estimated annual
emission rates appear in Table 2-3.
The average percent of total VOC emissions attributed to each component
type is presented in Table 2-4. The percent attributed to each component
type is the same for each model unit.
2-20
-------
TABLE 2-2. UNCONTROLLED FUGITIVE EMISSION FACTORS
IN PROCESS UNIT EQUIPMENT
Uncontrolled emission
Fugitive emission source factor,9 kg/hr
Pumps .
Light liquids
With packed seals 0.12
Wtih single mechanical seals 0.12
With double mechanical seals 0.12
With no seals 0.0
Heavy liquids
With packed seals 0.020
With single mechanical seals 0.020
With double mechanical seals 0.020
With no seals 0.0
t
Valves (in-line)
Gas . 0.021
Light liquid^ 0.010
Heavy liquid0 0.0003
Safety/relief valves
Gas . 0.16
Light liquid° 0.006
Heavy liquid0 0.009
Open-ended valves
Gas . 0.025
Light liquid^ 0.014
Heavy liquid0 0.003
Flanges 0.0003
Sampling connections 0.015
Compressors 0-44 e
Cooling towers 13.6-1107
Agitators NAT
aThese uncontrolled emission levels are based upon the refinery data presented
in reference 39-
bLight liquid is defined as a fluid with vapor pressure greater than 0.3 kPa
at 20°C. This vapor pressure represents the split between kerosene and naphtha
and is based on data presented in reference 39.
cAssumes the inner seal leaks at the same rate as single seal and that the VOC
is emitted from the seal oil degassing vent.
dHeavy liquid is defined as a fluid with vapor pressure less than 0.3 kPa at 20°C.
This vapor pressure represents the split between kerosene and naphtha and is
based on data presented in reference 40.
eThese levels are based on cooling tower circulation rates that range from
0.05-3.66 ms/sec (714-58,000 GPM). Reference 46-
NA no data available.
2-21
-------
TABLE 2-3. ESTIMATED TOTAL FUGITIVE EMISSIONS FROM MODEL UNITSa
Model unit Model unit Model unit
ABC
Estimated total
emissions (Mg/yr) 67 260 800
aBased upon equipment component counts in Table 2-1, the uncontrolled
emission factors in Table 2-2, and 8,760 hours/yr.
2-22
-------
TABLE 2-4. AVERAGE PERCENT OF TOTAL FUGITIVE EMISSIONS ATTRIBUTED
TO SPECIFIC COMPONENT TYPES
Component
Percent of
total uncontrolled
emissions attributed to
to component type
for model units A,B,C
Pump seals
Light liquid service
Heavy liquid service
In-line valves
Gas service
Light liquid service
Heavy liquid service
Safety/relief valves
Gas service
Light liquid service
Heavy liquid service
Open-ended valves
Gas service
Light liquid service
Heavy liquid service
Compressor seals
Sampling connections
Flanges
12
2
26
11
23
3
9
2
4
5
2
Less than one percent.
2-23
-------
2.4 REFERENCES
1. Erikson, D.G., and V. Kalcevic. Emissions Control Options for the
Synthetic Organic Chemicals Manufacturing Industry, Fugitive Emissions
Report, Draft Final. Hydroscience, Inc., 1979. p. II-2.
2. Ref. 1.
3. Ref. 1, p. II-3.
4. Ramsden, J.H. How to Choose and Install Mechanical Seals. Chem. E.,
85(22):97-102. 1978.
5. Ref. 1, p. II-3.
6. Ref. 4, p. 99.
7. Ref. 4, p. 99.
8. Ref. 4, p. 99.
9. Ref. 4, p. 99.
10. Perry, R.H., and C.H. Chi 1 ton, Chemical Engineers' Handbook, 5th Ed.
New York, McGraw-Hill Book Company, 1973. p. 6-8.
11. Ref. 10, p. 6-13.
12. Nurken, R.F. Pump Selection for the Chemical Process Industries, Chem.
E., Feb. 18, 1974. p. 120.
13. Nelson, W.E. Compressor Seal Fundamentals. Hydrocarbon Processing,
56_( 12): 91-95. 1977.
14. American Petroleum Institute, "Centrifugal Compressors for General Refinery
Service", API Standard 617, Fourth Edition, November, 1979, p. 8.
Reprinted by courtesy of the American Petroleum Institute.
15. Reference 14, p. 9.
16. Ref. 13.
17. Ref. 13.
18. Ref. 1, p. 11-7.
19. Lyons, J.L., and C.L. Ashland, Jr. Lyons' Encyclopedia of Valves. New
York, Van Nostrand Reinhold Co., 1975. 290p.
20. Templeton, H.C. Valve Installation, Operation and Maintenance. Chem.
E., 78(23)141-149, 1971.
2-24
-------
References (continued)
21. Ref. 1, p. II-5.
22. Ref. 18, p. 147-148.
23. Ref. 18, p. 148.
24. Ref. 18, p. 148.
25. Ref. 18, p. 148.
26. Pikulik, A. Manually Operated Valves. Chem. E., April 3, 1978.
p. 121.
27. Ref. 24, p. 121.
28. 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.
29. Ref. 1, p. II-7.
30. Cooling Tower Fundamentals and Application Principles. Kansas City,
Missouri, The Marley Company, 1969. p. 4.
31. Ramsey, W.D. and G.C. Zoller. How the Design of Shafts, Seals and
Impeller Affects Agitator Performance. Chem. E., 83_(18): 101-108.
1976.
32. Ref. 29, p. 105.
33. Ref. 29, p. 105.
34. Ref. 29, p. 105.
35. Ref. 29, p. 106.
36. Ref. 29, p. 106.
37. Ref. 29, p. 106.
38. McFarland, I. Preventing Flange Fires. Chem. E. Prog., 65_(8): 59-61.
1969.
39. Wetherold, R.G., et al. Emission Factors and Frequency of Leak Occurrence
for Fittings in Refinery Process Units, interim report, EPA Contract
No. 68-02-2665. Austin, Texas, Radian Corporation, February 1979.
pp. 11-49.
2-25
-------
References (continued)
40. Radian Corporation. The Assessment of Environmental Emissions From Oil
Refining. Draft Report, Appendix B. EPA Contract No. 68-02-2147,
Exhibit B. Austin, Texas. August, 1979. pp. 3-5 through 3-16.
41. Ref. 1, pp. IV-1, 2.
42. Ref. 1, p. II-9-13.
43. Ref. 37,p p. 11-23.
44. Ref. 1, p. 11-10.
45. Ref. 1, p. IV-8.
46. Letter with Attachments from J.M. Johnson, Exxon Company, U.S.A., to
Robert T. Walsh, U.S. EPA. July 28, 1977.
2-26
-------
3.0 EMISSION CONTROL TECHNIQUES
Sources of fugitive VOC emissions from process unit equipment were
identified in Chapter 2. This chapter discusses the emission control
technique which is considered representative of reasonably available control
technology (RACT) for these sources. The estimated control effectiveness of
the technique is also presented.
3.1 LEAK DETECTION AND REPAIR METHODS
Leak detection and repair methods can b£ applied in order to reduce
fugitive emissions from process unit 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, 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.
Portable hydrocarbon detection instruments are the best method for
identifying leaks of VOC from equipment components. The instrument is used
to sample and analyze the air in close proximity to the potential leak surface
by traversing the sampling probe tip over the entire area where leaks may
occur. This sampling traverse is called "monitoring" in subsequent descriptions.
The VOC concentration of the sampled air is displayed on the instrument
meter. The performance criteria for monitoring instruments and a description
of instrument survey methods are included in Appendix C.
3-1
-------
The VOC concentration at which maintenance is required is called the
"action level". An action level of 10,000 ppmv is considered representative
of RACT. Components which have indicated concentrations higher than this
"action level" are marked for repair. Emission data indicate that large
variations in mass emission rate may occur over short time periods for an
individual equipment component.
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. They are
not intended to be complete repair procedures.
3.1.2.1 Pumps. Many process units have spare pumps which can be
operated while the leaking pump is being repaired. Leaks from packed 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. These temporary emissions could be greater than the continued
leak from the seal. Therefore the pump should be flushed of VOC as much as
possible before opening for seal replacement.
3.1.2.2 Compressors. Leaks from packed seals may be reduced by the
same repair procedure that was described for pumps. Other types of seals
require that the compressor be out of service for repair. Since most compressors
do not normally have spares, repair or replacement of the seal would require
a shutdown gf the process. If the leak is small, temporary emissions resulting
from a shutdown could be greater than the emissions from the leaking seal.
3-2
-------
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 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
installed upstream of the relief valve. A block valve is required upstream
of the safety/relief valve in order to permit in-service replacement of the
valve if it cannot be repaired. In some chemical plants, installation of a
block valve upstream of a pressure relief device may be a common practice.
Although allowed by ASME codes , this practice may be forbidden by operating
or safety procedures of a particular company. A spare relief valve should be
attached while the faulty valve is repaired and tested. After a relief valve
has been repaired and replaced, it is possible that the next over-pressure
relief will result in another leak.
t
3.1.2.4 Valves. Most valves have a packing gland which can 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 overtightened. 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. Most control valves have a manual bypass loop which allows them
to be isolated easily, although temporary changes in process operation may
allow isolation in some cases. If a process unit must be shut down in order
to isolate a leaking valve, the emissions resulting from the shutdown might
be greater than the emissions from the valve if allowed to leak until the
next scheduled unit turnaround which permits isolation for repair.
Depending on site specific factors, it may be possible to repair process
valves by injection of a sealing fluid into the source. Injection of sealing
fluid has been successfully used to repair leaks from valves in petroleum
2
refineries in California.
3-3
-------
Fugitive emissions from open-ended valves are the result of leakage
through the seat of the valve. Approximately 28 percent of valves (excluding
safety/relief and .check valves) in VOC service are open-ended. They include
drain, purge, sample, and vent valves. Fugitive emissions from open-ended
valves can be controlled by installing a cap, plug, flange, or second valve
to the open end of the valve. In the case of a second valve, the upstream
valve should always be closed first after use of the valves. Each time the
cap, plug, flange, or second valves is opened, any VOC which has leaked
through the first valves seat will be released. These emissions have not
been quantified. The control efficiency will be dependent on the frequency
of removal of the cap or plug. Caps, plugs, etc. for open-ended valves do
not affect emissions which may occur during use of the valve. These emissions
may be caused by line purging for sampling, draining or venting through the
open-ended valve.
3.1.2.5 Flanges. In some cases, leaks from flanges can be reduced by
replacing the flange gaskets. Most flanges cannot be isolated to permit
replacement of the gasket. Data from petroleum refineries show that flanges
emit very small amounts of VOC.
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:
t Action level (leak definition),
• Inspection interval (monitoring frequency),
• Achievable emission reduction of maintenance, and
t Interval between detection and repair of the leak.
Some of these factors can be estimated by using data collected from petroleum
refineries.
3.1.3.1 Action Level. The action level is the VOC concentration
observed during monitoring which defines a leaking component which requires
repair. The choice of the action level for defining a leak is influenced by
a number of important considerations. First, the percent of total mass
emissions which can potentially be controlled by the monitoring and repair
3-4
-------
program can be affected by varying the leak definition, or action level.
Table 3-1 gives the percent of total mass emissions affected by the 10,000 ppmv
action level for a number of equipment types. The choice of an appropriate
leak definition is most importantly limited by the ability to repair leaking
components. The ability to repair leaking equipment from above 10,000 ppmv
to below 10,000 ppmv has been demonstrated in field testing. (Table A-16,
Appendix A). This repair ability has not been as well demonstrated for a
1,000 ppm leak defintion, however. Some available data do not support the
conclusion that repairing leaks in the 1,000 to 10,000 ppmv range would result
in an overall reduction in emissions.
The nature of repair techniques for pipeline valves, for instance, is
such that attempts to repair leaks below a certain level by tightening the
packing gland may result in an increase in e'missions. In practice, valve
packing material can become hard and brittle after extended use. As the
packing loses its resiliency, the valve packing gland must be tightened to
prevent loss of product to atmosphere. Excessive tightening, however, may
cause cracks in the packing, thus increasing the leak rate.
3.1.3.2 Inspection Interval. The length of time between inspections
should depend 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 repaired sooner. In order to
evaluate the effectiveness of the quarterly monitoring interval which is
considered representative of RACT, it is necessary to estimate the rate at
which new leaks will occur and repaired leaks will recur. The estimates
which have been used to evaluate quarterly monitoring are shown 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
3-5
-------
TABLE 3-1. PERCENTAGE OF EMISSIONS AS A FUNCTION OF ACTION LEVEL3
Fraction of mass emissions
from sources with leak rates
above the 10,000 ppmv
action level (as %)
Source type
Pump seals
Light liquid service 87
Heavy liquid service 21
Valves
Gas service 98
Light liquid service 84
Heavy liquid service 0
Safety/relief valves 69
Compressor seals 84
Flanqes 0
3-6
-------
TABLE 3-2. ESTIMATED OCCURRENCE AND RECURRENCE RATE OF LEAKS FOR A
QUARTERLY MONITORING INTERVAL
Component
type
Estimated Percent of sources
percent of leaking at quarterly
components inspection from leak
leaking occurrence, recurrence,
initially and leaks not repaired^
Pump seals
Light liquid service
Heavy liquid service
23
2
2.3
0.2
Valves
Gas service
Light liquid service
Heavy liquid service
10
12
0
1.0
1.2
0.0
Safety/relief valves
Compressor seals
8
33
0.8
3.3
Flanges
0
0.0
Approximate fraction of components with a concentration greater than or equal
to 10,000 ppmv prior to repair.
^Estimated that 10 percent of the initial leaks represent subsequent occurence
and recurrence rate for quarterly inspections. This estimate is based on
engineering judgement.
3-7
-------
affect emission reductions by influencing the length of time that leaking
sources are allowed to continue to emit pollutants. Some of the components
with concentrations in excess of the leak action level may not be able to be
repaired until the next scheduled unit shutdown.
The allowable interval before repair considered representative of RACT
is fifteen days. The percent of emissions from a component which would be
affected by the repair interval if all other contributing factors were
100 percent efficient is 97.9 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 repair to a level equivalent to a concentration reading of 1,000 ppmv.
The average emission rates of components above 10,000 ppmv and at 1,000 ppmv
are shown in Table 3-3.
3.1.3.5 Development of Controlled Emission Factors. The uncontrolled
emission levels for the emission sources that are typically found in the
model plants were previously presented in Chapter 2 (Table 2-2). Controlled
VOC emission levels can be calculated by a "controlled emission" factor.
This factor can be developed for each type of emission source by using the
general expression:
Controlled emission factor = Uncontrolled factor - [uncontrolled
factor x emission reduction efficiency]
The reduction efficiency can be developed by the following expressions and
correction factors:
p
Reduction efficiency = AxBxCxD
Where:
A = Theoretical Maximum Control Efficiency = fraction of total mass
emissions for each source type with VOC concentrations greater than
the action level (Table 3-1, Figure 3-1).
3-8
-------
TABLE 3-3. AVERAGE EMISSION RATES FROM SOURCES ABOVE
10,000 PPMV and at 1000 PPMV9
Source type
Pump seals
Light liquid service
Heavy liquid service
In-line valves
Gas service
Light liquid service
Heavy liquid service
Safety/relief valves
Compressor seals
Fl anges
(Y)
Emission rate
from sources above
10,000 ppmv
(kg/hr)
0.45
0.21
0.21
0.07
0.005
1.4
1.1
0.003
(X)
Emission rate
from sources at
1000 ppmvb
(kg/hr)
0.035
0.035
0.001
0.004
0.004
0.035
0.035
0.002
Y-X
Percentage
reduction
92.2
83.3
99.5
94.3
20.0
97.5
96.8
33.3
ng
ng
^Average emission rate of all sources, within a source
values above 10,000 ppmv.
^Emission rate of all sources, within a source type, having screening values
of 1000 ppmv.
3-9
-------
100
CO
2 90
CO
to
80
to 70
3
_i 60
<
i-
2 50
u.
o
uj
(C
UJ
40
20
10
0
UPPER LIMIT OF 90%
^ /CONFIDENCE INTERVAL
ESTIMATED PERCENT OF
TOTAL MASS EMISSIONS
LOWER LIMIT OF 90%
CONFIDENCE INTERVAL
10 I02 I03 I04 I05 I06
SCREENING VALUE (ppmvH LOG,0 SCALE)
PERCENT OF TOTAL MASS EMISSIONS - PERCENT OF TOTAL
EMISSIONS ATTRIBUTABLE TO SOURCES WITH SCREENING
VALUES GREATER THAN THE SELECTED VALUE.
Figure 3-1. Cumulative distribution of total emissions by screening
values - valves in light liquid service.9
UPPER LIMIT OF 9S%
CONFIDENCE INTERVAL
IMATED PERCENT OF SOURCES
LOWER LIMIT OF 9S%
CONFIDENCE INTERVAL
10 10*103 I04 10s
SCREENING VALUE (ppmv) (LOG|QSCALE)
PERCENT OF SOURCES - PERCENT OF SOURCES WITH
SCREENING VALUES GREATER THAN THE SELECTED SOURCE.
Figure 3-2. Cumulative distribution of sources by screening
values - valves in light liquid service.10
3-10
-------
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)
(Table 3-2, 3-4).
C = Non-Instantaneous 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 a 1000 ppmv emission
level equivalent to a concentration of 1000 ppmv (Table 3-3).
t
These correction factors can, in turn, be determined from the following
expressions:
\
(DB-l-
(3) D 1-f
'Where:
n = Total number of leaks occurring and recurring over the monitoring
m interval.
N - Total number of sources at or above the action level (Figure 3-2).
t = Average time before repairs are made (with a 15-day repair limit,
7.5 is the average used).
f = Average emission factor for sources at the average screening value
achieved by repair.
F Average emission factor for all sources at or above the action level
An example of a control effectiveness calculation is presented in Table 3-5.
Support data for this calculation are presented in Tables 3-1, 3-2, 3-3,
and 3-4, as well as in Figures 3-1 and 3-2.
3-11
-------
TABLE 3-4. IMPACT OF MONITORING INTERVAL ON CORRECTION FACTOR ACCOUNTING
FOR LEAK OCCURRENCE/RECURRENCE (FOR EXAMPLE CALCULATION)
Monitoring a Bb
interval m
3 months 0.2NC 0.90
a n = Total number of leaks which occur, recur, and remain between
monitoring intervals.
B = Correction factor accounting for leak occurrence and recurrence.
c N = Total number of components at or above the action level.
3-12
-------
TABLE 3-5. EXAMPLE OF CONTROL EFFICIENCY CALCULATION
Assume:
1) A leak detection and repair program to reduce emissions from
valves in light liquid service.
2) Action level = 10,000 ppmv.
3) Average screening value after directed repair = 1,000 ppmv.
4) Leak detection monitoring interval = 3 months.
5) Allowable repair interval = 15 days.
6) Number of valves having new or recurring leaks between repair
intervals, r\m 0.2N (see Table 3-4).
Calculations:
A = 0.84 (from Figure 3-1 for a screening value of 10,000 ppmv)
B = 0.9 (from Table 3-4)
C = 0.979 (for 15-day interval)
where:
F = A(Avg. uncontrolled emission factor)9 .
Fraction of sources screening 2l 10,000 ppmv
= (0.84)(0.010 kg/hr)/0.11 = 0.076 kg/hr
f = Emission factor at 1,000 ppmva
= 0.004 kg/hr
. 0.947
Overall percentage reduction =AxBxCxD
(0.84) x (0.90) x (0.979) x (0.947)
= 70 percent
Therefore:
Emission factor after control = 0.010 kg/hr - (0.70) (0.010 kg/hr)
= 0.003 kg/hr
Reference 3.
bFrom Figure 3-2.
3-13
-------
3.2 OTHER CONTROL STRATEGIES
This section discusses two fugitive emission control strategies for
valves in gas service and valves in light liquid service other than the
quarterly leak detection and repair procedures discussed above. Consideration
of alternative control strategies for valves is pertinent because they account
for such a large percentage of the components to be monitored (about 90 percent
in the model process units). However, alternative control strategies are not
pertinent for other components (pumps, compressors, safety/relief valves)
because these other components are relatively few in number.
These strategies should be considered alternatives to quarterly leak
detection and repair to allow plants the flexibility to meet a level of
performance using control procedures considered most appropriate by that
plant. Plants 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.
3.2.1 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 90 percent of the components to be monitored
in a process unit 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.
3-14
-------
On an individual component basis valves have a lower emission rate than
the other components (Table 2-2) and have a percentage leak rate which is
lower than most other components (Table 3-2). As the percent of valves found
leaking decreases the product recovery credit decreases. The direct cost for
monitoring, however, remain 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.
Table 3-6 presents the cost effectiveness of a quarterly leak detection
and repair program for the model process units as a function of the initial
percent of valves found leaking. As shown in Table 3-6, the cost effectiveness
for a quarterly leak detection and repair program for valves appears reasonable
for leak percentages of one percent or higher.
A plant averaging 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 plant averaged one percent of valves leaking,
then the percent of valves found leaking during a random annual inspection
should exceed two percent less than five percent of the time. In other
words, if a random annual inspection indicated that no more than two percent
of valves are leaking, the probability is greater than ninety-five percent
that an average of one percent of valves leaking is actually being achieved
in practice. Therefore, two percent of valves found leaking is a reasonable
criterion to judge the applicability of alternative control strategies for
valves.
3.2.2 Allowable Percentage Of Valves Leaking
A State regulation incorporating an alternative control strategy based
on an "allowable percentage of valves leaking" would require a plant to limit
the number of valves leaking at any time to a certain 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.
3-15
-------
TABLE 3-6. COST EFFECTIVENESS VERSUS INITIAL PERCENT OF VALVES
LEAKING IN MODEL UNITS3
Initial percent
of valves leaking
0.1
0.5
1
2
5
10
20
Cost
A
23,500
4,400
2,000
850
100
(150)
(270)
effectiveness in model
B
8,500
1,400
460
35
(220)
(300)
(350)
unit ($/Mg)b
C
4,500
630
140
(HO)
(300)
(350)
(370)
aFor quarterly leak detection and repair.
bAssumes value of VOC of $410/Mg.
( ) = indicates credit
3-16
-------
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 in gas service and valves in light liquid
service, and attempting to repair any valves which are leaking. All other
components 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. 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
t
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 for each process unit. 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.
3.2.3 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 in gas service and light liquid service 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. This
strategy is generally referred to as "skip-period" monitoring. All other
components would be subject to quarterly leak detection and repair intervals.
3-17
-------
If Implementation of the quarterly leak detection and repair program
showed that two percent or less of the valves in gas service and valves in
light liquid service were leaking for j_ consecutive quarters, then m quarterly
inspections may be skipped. If the next inspection period also showed that
the "good performance level" was being achieved, then m quarterly inspections
could be skipped again. When an inspection period showed the "good performance
level" was not 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 m_ quarterly inspections
could be skipped again.
As mentioned above, two percent of valves leaking represents a good
level of performance. Table 3-7 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 gas service and
valves in light liquid service in a process unit 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.
This strategy would permit a plant that has consistently demonstrated it
is meeting the "good performance level" to monitor valves in gas service and
valves in light liquid service 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.
3.3 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 unsafe or difficult to reach, small process units,
and unit turnarounds.
3-18
-------
TABLE 3-7. ILLUSTRATION OF SKIP-PERIOD MONITORING0
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
skip
monitor
Good
performance
level
achieved?
No
Yes
Yes
Yes
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
2
3
4d
1
2
3
4d
ai=5, m=3, good performance level of 2 percent.
bFifth consecutive quarter below 2 percent means 3 quarters of monitoring may be
skipped.
Percentage of leaks above 2 percent means quarterly monitoring reinstituted.
Percentage of leaks below 2 percent means 3 quarters of monitoring may be
skipped.
3-19
-------
3.3.1 Unsafe and Difficult to Reach Components
Some components might be considered unsafe to monitor because process
conditions include extreme temperatures or pressures. A State agency may wish
to require less frequent 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 reach 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.
3.3.2 Small Process Unit
Some process units have so few components to be monitored that the cost
effectiveness of a quarterly leak detection and repair program for that
process unit would be high. A State agency may wish to consider such process
units ''small" and exempt them from compliance with a regulation.
The total cost of a leak detection and repair program would consist of the
capital cost of VOC detection instruments and the cost of labor for leak
detection and repair. The cost of VOC detection instruments would be the same
for all sizes of process units, but the cost of labor for leak detection and
repair would depend on the number of components to be monitored. As the
number of components to be monitored decreases, both the labor cost and the
recovery credit associated with VOC emission reduction decrease. This results
in a lower total cost. However, since the cost of the VOC detection instruments
is fixed, a leak detection and repair program becomes less cost effective as
the number of components subject to monitoring decreases.
3-20
-------
Valves in light liquid service and valves in gas service are the greatest
percentage (about 90 percent) of the components which would be subject to
monitoring in a typical process unit. In addition, the number of valves in
gas service and light liquid service can be used as a crude indicator of the
total number of components in a process unit which would be subject to monitoring,
Table 3-8 shows the emission reduction, net annualized cost, and cost
effectiveness for quarterly leak detection and repair in process units with
different numbers of valves. The magnitude of emission reduction and cost
effectiveness of emission control suggest that implementation of a leak
detection and repair program for units which have more than 100 valves in gas
service and valves in light liquid service appears reasonable. Thus, States
may wish to consider exempting process units with less than 100 valves in gas
service and light liquid service from regulations requiring control of fugitive
VOC emissions.
3.3.3 Unit Turnarounds
A State agency might wish to consider a provision in their regulations
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.
3-21
-------
TABLE 3-8. COST EFFECTIVENESS OF QUARTERLY LEAK DETECTION AND REPAIR FOR
TYPICAL PROCESS UNITS
ro
ro
Number of
valves in
process
unit
10
50
100
200
Uncontrolled
emissions
(Mg/yr)*
1.3
6.3
12.6
25.2
Potential
emission
reduction
(Mg/yr)
1.1
5.3
10.6
21.1
Total
annual ized
cost ($)
5,590
6,050
6,620
7,740
Net
annual ized
cost ($)b
5,140
3,880
2,270
(910)
Cost
effectiveness
($/Mg)
4,670
730
210
(40)
Cost
effectiveness
per valve
($-valve/Mg)
467
15
2.1
(0.2)
Based on model unit proportion of valves in gas service and valves in light liquid service and
operating 365 days per year.
bBased on VOC value of $410/Mg.
( ) = net credit.
-------
3.4 REFERENCES
1. Part UG - General Requirements (Section VIII, Division I.) In: ASME
Boiler and Pressure Vessel Code, An American National Standard. New
York, The American Society of Mechanical Engineers, 1977. p. 449.
7.
8.
Teller, James H. Advantages Found in On-Line Leak Sealing.
Journal, 77 (29):54-59, 1979.
Oil and Gas
Wetherold, R.G., and L.P. Provost. Emission Factors and Frequency of
Leak Occurance for Fittings in Refinery Process Units. Interim Report.
EPA/600/2-79-044. Radian Corporation. February 1979. p. 2.
from F.R. Bottomley, Union Oil Company.
Feldstein, Bay Area Quality Management
Ref. 3.
Valve Repair Summary and Memo
Rodeo, California. To Milton
District, April 10, 1979.
Ref. 3.
Ref. 3.
Tichenor, B.A., K.C. Hustvedt, and R.C. Weber. Controlling Petroleum
Refinery Fugitive Emissions Via Leak Detection and Repairs in Proceedings:
Symposium on Atmospheric Emissions from Petroleum Refineries (November 1979,
Austin, Texas). EPA-600/9-80-013. Radian Corporation. March 1980.
pp. 421-440.
9. Ref. 3
10. Ref. 3.
3-23
-------
4.0 ENVIRONMENTAL ANALYSIS OF RACT
4.1 INTRODUCTION
The environmental impacts resulting from implementation of reasonably
available control technology (RACT) are examined in this chapter. Implementing
a quarterly leak detection and repair program and capping of open-ended lines
with a second valve, cap, plug, or blind flange is considered representative
of RACT for control of fugitive VOC emissions from equipment components in
the SOCMI and in the polymer and resin manufacturing industry.
Leak detection should consist of quarterly monitoring the following
components in VOC service with a VOC detection instrument: pumps in light
liquid service, valves in light liquid service, valves in gas service, compressors,
and safety/relief valves in gas service. Pumps in light liquid service
should be visually inspected weekly for indications of leaks. The VOC detection
instrument and the monitoring method employed should be EPA Reference Method 21
(Appendix C) or an equivalent State method. A component should be considered
in VOC service if it contains ten percent or greater VOC by weight. A VOC
is any organic compound which participates in atmospheric photochemical
reactions and is measured by the applicable test methods described in EPA
Reference Method 21 or equivalent State method. For the purpose of this
document, a light liquid is defined as a fluid with a vapor pressure greater
than 0.3 kPa at 20°C. A component should be considered in light liquid
service if it contacts a fluid containing greater than ten percent by weight
light liquid. A component should be considered in gas service if it contains
process fluid that is in the gaseous state at operating conditions.
Components which have a measureable VOC concentration of 10,000 ppmv or
greater should be considered leaking components. Leaking components should be
repaired within 15 days of the date the leak is detected. Repair should be
considered as reduction of measureable VOC concentration below 10,000 ppmv.
Leaking components which cannot be repaired without a unit shutdown should be
repaired at the next unit turnaround.
4-1
-------
4.2 AIR POLLUTION
Implementation of RACT would reduce VOC fugitive emissions from process
units. A significant beneficial impact on air pollution emissions would
result. The hourly and annual emissions from each model unit before and
after control by RACT are presented (Tables 4-1 and 4-2). There would be no
adverse air pollution impacts associated with RACT.
4.2.1 Development of VOC Emission Levels
The uncontrolled emission factors for process unit equipment were
previously presented in Chapter 2 (Table 2-2). Emission factors were developed
for those sources that would be controlled by the implementation of RACT.
These controlled fugitive emission levels were calculated by multiplying the
uncontrolled emissions from this equipment by a control efficiency. The
control efficiency is determined by several factors which are described and
presented in Chapter 3. The controlled VOC emission factors for each source
are presented in Table 4-3.
In calculating the total fugitive emissions from model units controlled
under RACT, the uncontrolled and controlled emission factors were used.
These emission factors were multiplied by the equipment source inventories
for each model unit. An example calculation for estimating emissions from
model unit A under RACT is shown in Table 4-4.
4.2.2 VOC Emission Reduction
The emission reduction expected from the implementation of RACT can be
determined for each model unit. The emission reduction is the difference
between the amount of fugitive emissions before RACT is implemented and the
amount of fugitive emissions after RACT is implemented. These amounts are
presented in Tables 4-1 and 4-2. The reduction in emissions for the model
units after RACT would be implemented is 66 percent.
4-2
-------
TABLE 4-1. ESTIMATED HOURLY EMISSIONS AND EMISSIONS REDUCTION
ON A MODEL UNIT BASIS.
Estimated emissions
(kg/hr) Average percent
Level of Mode! unit reduction from
control ABC uncontrolled level
Uncontrolled 7.7 29.3 91.2
RACT 2.6 10.2 31.6 66
TABLE 4-2. ESTIMATED ANNUAL EMISSIONS AND EMISSIONS REDUCTION
ON A MODEL UNIT BASIS.
Estimated emissions
(Mq/yr) Average percent
Level of Model unit ~ reduction from
control ABC uncontrolled level
Uncontrolled 67 260 800
RACT 23 89 277 66
4-3
-------
TABLE 4-3. EMISSION FACTORS FOR SOURCES CONTROLLED UNDER RACT
Uncontrolled
emission source
Pumps
Light liquid service
Valves
Gas service
Light liquid service
Safety/relief valves
Gas service
Compressors
Uncontrolled
emission
Inspection factor,
interval kg/hr
Quarterly
Quarterly
Quarterly
Quarterly
Quarterly
0.120
0.021
0.010
0.160
0.440
Ab
0.87
0.98
0.84
0.69
0.84
Correction
factors
Bc
0.90
0.90
0.90
0.90
0.90
Cd
0.98
0.98
0.98
0.98
0.98
De
0.92
0.99
0.94
0.97
0.97
Control
ef f i ci ency
(AxBxCxD)
0.71
0.86
0.70
0.59
0.72
Controlled
emission
factor,
kg/hr
0.035
0.003
0.003
0.066
0.123
aFrom Table 2-2.
theoretical maximum control efficiency. Reference 1.
cLeak occurrence and recurrence correction factor. Reference 2.
Non-instantaneous repair correction factor - for 15-day maximum allowable repair time, the correction
factor is [365 - (15/2)] * 365. Reference 2.
elmperfect repair correction factor, calculated as 1 - (f f F), where f = average emission rate for sources
at 1000 ppmv and F = average emission rate for sources greater than 10,000 ppmv. References 1, 2.
Controlled emission factor = uncontrolled emission factor x [1 - (AxBxCxD)].
-------
TABLE 4-4. EXAMPLE CALCULATION OF VOC FUGITIVE EMISSIONS FROM
MODEL UNIT A UNDER RACT
Number of
sources in.
model unit0
(N)
Emission.
factor,
kg/hr-source
(E)
Emissions
from sources,
kg/hr
(N x E)
Emission Source:
Pumps .
Light liquid single
mechanical jseal
Light liquid double
mechanical seal
Heavy liquid single
mechanical seal
Heavy liquid packed seal
In-line valves
Gas service .
Light liquid service
Heavy liquid service
Safety/relief valves
Gas service j
Light liquid service
Heavy liquid service
Open-ended valves
Gas service ,
Light liquid service
Heavy liquid service
Compressors
Sampling connections
Flanges
5
3
5
2
t
90
84
84
11
1
1
9
47
48
1
26
600
Total
0.035
0.035
0.020
0.020
0.003
0.003
0.0003
0.066
0.006
0.009
0.003
0.003
0.0003
0.123
0.015
0.0003
emissions
0.175
0.105
0.100
0.040
0.270
0.252
0.025
0.726
0.006
0.009
0.027
0.141
0.014
0.123
0.390
0.180
2.583
aFrom Table 2-1.
RACT emission factors include uncontrolled factors from Table 2-2 and controlled
factors from Table 4-3.
°Sources in VOC service.
Light liquid is defined as having a vapor pressure equal to or greater than
0.3 kPa at 20°C. A component is in liquid liquid service if it contains
greater than 10 percent by weight light liquid.
6Heavy liquid is defined as having a vapor pressure less than 0.3 kPa at 20°C.
Open-ended valve factor is equivalent to the in-line valve factor because of
capping the open end.
4-5
-------
4.3 WATER POLLUTION
Implementation of RACT would result in no adverse water pollution
impacts because no wastewater is involved in monitoring and leak repair.
Some liquid chemicals may already be leaking and entering the wastewater
system as runoff. A beneficial impact on wastewater would result from
implementation of RACT since liquid leaks are found and repaired. This
impact, however, cannot be quantified because no applicable data on liquid
leaks are available.
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
The implementation of RACT calls for an emission control technique that
requires no additional energy consumption for any of the model unit sizes.
A beneficial impact would be experienced by saving VOC which has been heated,
compressed, or pumped.
4-6
-------
4.6 REFERENCES
1. Wetherold, R. and L. Provost, Emission Factors and Frequency of Leak
Occurrence for Fittings in Refinery Process Units. EPA-600/2-79-044,
February 1979.
2. Tichenor, B.A., K.C. Hustvedt, and R.C. Weber. Controlling Petroleum
Refinery Fugitive Emissions Via Leak Detection and Repair, Draft.
Symposium on Atmospheric Emissions from Petroleum Refineries, Austin,
Texas.
4-7
-------
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
process units are presented in this chapter. Capital costs, annualized costs,
and the cost effectiveness of RACT are presented. These costs have been
developed for the model units presented in Chapter 2. All costs presented
in this chapter have been updated to second quarter 1980 dollars.
5.1 BASIS FOR CAPITAL COSTS
Capital costs represent the total cost of starting a leak detection and
repair program in existing process units. The capital costs for the implemen-
tation of RACT include the purchase of VOC monitoring instruments, the
purchase and installation of caps for all open-ended lines, 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.
Capital cost estimates for model units under RACT are presented in Table 5-2.
Labor costs were computed using a charge of $18 per labor-hour. This rate
includes wages plus 40 percent for related administrative and overhead
costs .
5.1.1 Cost of Monitoring Instrument
i
The cost of a VOC monitoring instrument includes the cost of two
instruments. One instrument is intended to be used as a standby spare. The
cost of $4600 for a portable organic vapor analyzer was obtained from a
2
manufacturer.
5-1
-------
TABLE 5-1. CAPITAL COST DATA
Cost value
used in analysis
Item (June 1980 dollars) Cost basis Reference
Monitoring instrument 2 x 4600 9200/model unit One instrument used as a spare 2
Caps for open-ended lines 53/1ine Based on cost for 1" screw-on 3
type valve. Cost June 1980 = $35.
Installation = 1 hour at $18/hour.
TABLE 5-2. CAPITAL COST ESTIMATES FOR IMPLEMENTING RACT
(thousands of June 1980 dollars)
Capital cost item
Model Unit A
1. Monitoring instruments .
2. Caps for open-ended lines (104 caps)
3. Initial leak detection and repair cost
Total
Model Unit B
1 . Moni tori ng i nstruments .
2. Caps for open-ended lines (415 caps)
3. Initial leak detection and repair cost
Total
Model Unit C
1. Monitoring instruments .
2. Caps for open-ended lines (1277 caps)
3. Initial leak detection and repair cost
Total
Level
Uncontrolled
0.0
0.0
0.0
0.0
0.0
0.0
0^0
0.0
0.0
0.0
0.0
0.0
of control
RACT
9.2
5.5
4.48
19.18
9.2
22.0
14.35
45.55
9.2
67.7
40.34
117.24
aBased on capital cost data presented in Table 5-1.
bFrom Table 2-1.
""Initial leak detection and repair are treated as capital costs because they are one-time cost.
5-2
-------
5.1.2 Caps for Open-Ended Lines
Fugitive emissions from open-ended lines and valves can be controlled by
installing a cap, plug, flange, or second valve to the open end. These
pieces of equipment are all included in the definition of a cap for an
open-ended line. The cost of a cap for an open-ended line is based on a
cost of $35 for a one-inch screw-on type globe valve. This cost was supplied
3
by a large distributor. A charge of $18 for one hour of labor is added to
$35 as the cost for installing one cap. Therefore, the total capital cost
for installing a cap on an open-ended line is $53.
5.1.3 Initial Leak Repair
The implementation of RACT will begin with an initial inspection which
will result in the discovery of leaking components. The number of initial
leaks is expected to be greater than the number found in subsequent inspec-
tions. 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 unit under consideration. Fractions were rounded up to the
next highest integer. The repair time for fixing leaks is estimated to be
80 hours for a pump seal, 40 hours for a compressor seal, and 1.13 hours for
a valve. The repair time for fixing pump seals and compressor seals includes
the cost of a new seal. These requirements are presented in Table 5-3.
The initial repair cost was 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 ANNUALIZED COSTS
Annualized 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 annualized costs is presented in Table 5-4.
5-3
-------
TABLE 5-3. LABOR-HOUR REQUIREMENTS FOR INITIAL LEAK
REPAIR UNDER RACT
Source type
Pumps (light liquid)
Single mechanical seal
Double mechanical seal
Number of
components
per model
unit
ABC
5 19 60
3 10 31
Estimated
number of
initial
leaks3
ABC
2 5 14
1 3 8
Repair
time,
hrs
80b
80b
Labor-hours
requi red
ABC
160 400 1120
80 240 640
Valves (in-line)
Gas
Light liquid
Safety/relief valvesd
(gas service)
90 365 1117
84 335 1037
9 37 112
11 41 125
1.13 10 42 127
1.13C 12 46 141
11 42
130
Valves on open-ended lines
Gas
Light liquid
Compressor seals
9
47
1
37
189
2
115
581
8
le
6e
1
4e
23e
1
12e
70e
3
1.13C
1.13C
40b
1
7
40
5
26
40
14
79
120
TOTAL
310 799 2241
aBased on the percent of sources leaking at _> 10,000 pom. From Table 3-2.
blncludes labor-hour equivalent cost of new seal. Reference 6.
Steighted average based on 75 percent of the leaks repaired on-line, requiring 0.17 hours per repair,
and on 25 percent of the leaks repaired off-line, requiring 4 hours per repair. Ref. 5, p. B-12.
dlt is assumed that-these leaks are corrected by routine maintenance at no additional labor
requirements. Ref. 6.
^The estimated number of initial leaks for open-ended valves is based on the same percentage of
sources used for in-line valves. This represents leaks occurring through the stem and gland of the
open-ended valve. Leaks through the valve seat are eliminated by adding caps.
5-4
-------
TABLE 5-4. BASIS FOR ANNUALIZED COST ESTIMATES
1. Capital recovery factor for
capital charges
• Caps on open-ended lines
• Monitoring instruments
2. Annual maintenance charges
• Caps on open-ended lines
• Monitoring instruments
3. Annual miscellaneous charges
(taxes, insurance, administration)
• Caps on open-ended lines
• Monitoring instruments
4. Labor charges
5. Administrative and support costs
for implementing PACT
6. Annualized charge for initial
leak repairs
7. Recovery credits
0.163 x capital0
0.23 x capital13
0.05 x,capital0
$3,000°
0.04 x capital p
0.04 x capital
$18/hourf
0.40 x (monitoring + repair
labor)9
E (estimated number of leaking
components per model unit x
repair time) x $18/hrf x 1.49
x 0.163n
$410/Mg
aTen year life, ten percent interest. From Ref. 5, pp. IV-3,4.
bSix year life, ten percent interest. From Ref. 5, pp. IV-9,10.
cFrom Ref. 5, pp. IV-3,4.
Includes materials and labor for maintenance and calibration. Reference 6.
Cost index = 242.7 * 209.1 (Reference 7 and 8).
eFrom Ref. 5, pp. IV-3,4,9,10.
Includes wages plus 40 percent for labor-related administrative and overhead
costs. Cost (June 1980) from Ref. 1.
gFrom Ref. 5, pp. IV-9,10.
Initial leak repair amortized for ten years at ten percent interest.
•^References 9,10,11.
5-5
-------
5.2.1 Monitoring 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 require-
ments were calculated by taking the product of the number of workers needed
to monitor a component (1 for visual, 2 for instrument), the time required
to monitor, the number of components in a model unit, and the number of
times the component is monitored each year. The monitoring times for the
various components are presented in Table 5-5. They are 0.5 minute for
visual inspection, 1 minute for in-line valves and open-ended valves,
5 minutes for pump seals, 8 minutes for safety valves, and 10 minutes for
4
compressor seals. Monitoring labor costs 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 for each component is the same as
presented for initial leak repair. Leak repair costs 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. The annual cost of materials and labor for maintenance
678
and calibration of monitoring instruments is estimated to be $3000. ' ' An
additional miscellaneous charge of four percent of capital cost for taxes,
insurance, and 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
RACT 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.
5-6
-------
TABLE 5-5. ANNUAL MONITORING AND LEAK REPAIR LABOR REQUIREMENTS FOR RACT
01
Monitoring
Number of
components per
model unit
Source type
Pumps (light liquid)
Single mechanical
seals
Double mechanical
seals
Valves (In-line)
Gas
Light liquid
Safety/relief valves
(gas service)
Valves on open-ended
lines"
Gas
Light liquid
Compressors seals
TOTAL
A
5
3
90
84
11
9
47
1
B C
19 60
10 31
365 1117
335 1037
42 130
37 115
189 5B1
2 8
Type of*
monitoring
Instrument
Visual
Instrument
Visual
Instrument
Instrument
Instrument
Instrument
Instrument
Instrument
Monitoring
tlme.b
m1n
5
0.5
5
0.5
1
1
8
1
1
10
Times
monl tored
per year
4
52
4
52
4
4
4
4
4
4
Estimated
Monitoring labor- number of d
hours requ1redc leaks per year
A
3.3
2.2
2.0
1.3
12.0
11.2
11.7
1.2
6.4
1.3
52.6
B
12.8
8.2
6.8
4.3
49.0
44.8
44. »
4.9
25.2
2.7
203.5
C A B C
40.0 1 2 5
26.0
20.8 1 1 3
13.4
149.0 4 15 45
138.4 4 16 50
139.0
15.3 1 2 5
77.6 2 9 28
10.7 1 1 2
630.2
Leak repair
Repair Leak repair labor-
time, hours required6
hrs A
80b 80
80b 80
1.13f 4.5
1.13f 4.5
O9 0
1.13e 1.1
1.13e 2.3
40b 40
212.4
B
160
80
17.0
18.0
0
2.3
10.2
40
327.5
C
400
240
50.9
56.5
0
5.7
31.6
80
864.7
'Two workers for Instrument monitoring, one for visual.
Reference 4.
Ttonltorlng labor-hours «• number of workers x number of components x time to monitor (total 1s minimum of 1 hour) x number of times monitored per year.
dFrom Table 3-2.
eLeak repair labor-hours • number of leaks x repair time.
Weighted average based on 75 percent of the leaks repaired on-line, requiring 0.17 .hour per repair, and on 25 percent of the leaks, repaired off-line,
requiring 4 hours per repair. Ref. 5, p. B-12.
It Is assumed that these leaks are corrected by routine maintenance at no additional labor requirements. Ref. 4
estimated number of leaks per year for open-ended valves 1s based on the same percent of sources used for In-Hne valves. This represents leaks
occurring through the stem and gland of the open-ended valve. Leaks through the seat of the valve are eliminated by adding caps.
-------
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 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 the CRF by the total installed capital cost for the
control equipment. This equation for the capital recovery factor is:
1(1 + i)n
CRF= (l+i)n-l
where i = interest rate, expressed as a decimal
n = economic life of the equipment, years.
The interest rate used was ten percent (June 1980). 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 unit are presented in
Table 5-6. The VOC saved is valued in June 1980 dollars at $410/Mg.9'10'n
5.3 EMISSION CONTROL COSTS
This section will present and discuss the emission control costs
of implementing RACT for each of the three model units. Both the initial
costs and the annualized costs are included.
5-8
-------
TABLE 5-6. RECOVERY CREDITS
Model
unit
A
B
C
Uncontrolled
emissions,
Mg/yr
67
260
800
Emissions
under RACT,
Mg/yr
23
89
277
Emission
reduction,
Mg/yr
44
171
523
Recovered
product
value,
$/yr
18,040
70,100
214,400
Based on an average price of $410/Mg. References 9,10,11.
5-9
-------
5.3.1 Initial Costs
The cost of initially implementing RACT consists of capital costs and
initial leak repair. The 'capital cost of $9200 for two monitoring instruments
is the same for all model unit sizes. Caps for open-ended lines will cost
$5500 for model unit A, $22,000 for model unit B, and $67,700 for model
unit C. The one-time initial leak repair cost is $7812 for model unit A,
$20,135 for model unit B, and $56,470 for model unit C. The total initial
capital costs for implementing RACT are $19,180 for model unit A, $45,550
for model unit B, and $117,240 for model unit C.
5.3.2. Recovery Credits
The value of VOC saved each year as a result of implementing RACT is
included as an annual credit against the net annualized costs. The implemen-
tation of RACT will result in saving $18,260 worth of VOC annually in model
unit A, $69,540 worth of VOC in model unit B, and $211,100 worth of VOC in
model unit.C.
5.3.3 Net Annualized Cost
The net annual cost for controlling emissions is the difference between
the total annualized cost and the annual recovery credit for each model
unit. Net annualized control cost estimates for model units under RACT are
presented in Table 5-7. Capital cost data were previously presented in
Table 5-1.
For model unit A, the annualized capital charges are $4280 and the
total annual operating costs are $10,550. Product recovery credits total
$18,260. The net annualized cost for model unit A is a negative $3436,
which means that $3436 is actually gained every year by preventing loss of
VOC.
The annualized capital charges for model unit B are $8990 and the total
annual operating costs are $18,730. The recovery credit is $69,540 per
year. The net annualized cost for model unit B is a negative $41,823, which
means that $41,823 is saved every year by controlling VOC emissions.
5-10
-------
TABLE 5-7. ANNUALIZED CONTROL COST ESTIMATES FOR MODEL UNITS
UNDER RACT (thousands of June 1980 dollars)
Cost
item
A
Model unit
B
C
Annualized capital charges
1. Control equipment
a. Instrument 2.12 2.12 2.12
b. Caps 0.89 3.59 11.04
2. Initial leak repair 1.27 3.28 9.2
Subtotal 4.28 8.99 22.36
Operating costs
1. Maintenance charges
a. Instrument
b. Caps
2. Miscellaneous (taxes, insurance,
administration)
a. Instrument
b. Caps
3. Labor
a. Monitoring labor .
b. Leak repair labor
c. Plant and payroll overhead
Subtotal
Total before credit
Recovery credits
Net annual ized cost
3.0
0.275
0.37
0.22
0.95
3.82
1.91
10.545
14.825
18.26
(3.44)
3.0
1.1
0.37
0.88
3.66
5.9
3.82
18.73
27.72
69.54
(41.82)
3.0
3.39
0.37
2.71
11.34
15.56
10.76
47.13
69.49
216.73
(147.24)
aSum of labor hours for monitoring in Table 5-5 multiplied by $18/hour.
bSum of labor hours for leak repairs in Table 5-5 multiplied by $18/hour.
cBased on 40 percent of monitoring labor plus leak repair labor costs.
dThese costs are credits. (XXX) = net credit.
5-11
-------
Model unit C has annualized capital charges of $22,360 and total
operating expenses of $47,130. The recovery credit is $216,730 per year.
The net annualized cost for model unit C is a negative $147,240, which is an
annual savings as a consequence of controlling fugitive VOC emissions.
5.3.4 Differences in Net Annualized Costs
The cost for RACT is different for each model unit. The cost for caps
for open-ended lines varies because the number of open-ended lines is
different for each model unit. Because the larger model units have more
components, more labor-hours are needed for monitoring and leak repair. For
this reason, labor costs will increase as model unit size increases.
5.4. COST EFFECTIVENESS
Cost effectiveness is the annualized cost per megagram of VOC controlled
annually. The cost effectiveness of RACT for each model unit is the net
annualized cost for implementing RACT divided by the emission reduction
gained under RACT. The cost effectiveness of RACT is summarized in Table 5-8.
The implementation of RACT on model unit A results in a net annualized
cost which is a credit of $3436. The emission reduction associated with
RACT is 44.5 Mg/yr. The cost effectiveness is -$77/Mg.
The implementation of RACT in the case of model unit B results in a net
annualized cost which is a credit of $41,820. The emission reduction associated
with RACT is 169.6 Mg/yr. The cost effectless is -$247/Mg.
The implementation of RACT in the case of model unit C results in a net
annualized cost which is a credit of $147,240. The emission reduction
associated with RACT is 528.6 Mg/yr. The cost effectiveness is -$279/Mg.
A comparison of the cost effectiveness of RACT for each model unit
reveals that cost effectiveness increases as model unit size increases. The
strong influence of recovery credits is responsible for the increase in cost
effectiveness.
5-12
-------
TABLE 5-8. COST EFFECTIVENESS FOR MODEL UNITS UNDER RACT
Annual ized cost before credit ($1000)
Annual recovery credit ($1000)
Net annual ized cost ($1000)
Total VOC reduction (Mg/yr)
Cost effectiveness ($/Mg VOC)
A
14.82
18.26
(3.44)
44.54
(77.0)
Model unita
B
27.72
69.54
(41.82)
169.6
(247)
C
69.49
216.73
(147.24)
528.6
(279)
'(XXX) net credit.
The cost effectiveness of RACT for each component type is presented in
Tables 5-9, 5-10, and 5-11 for model units A, B and C, respectively. The
cost of the monitoring instrument cannot be attributed to any single type of
component since all the components are monitored by the instrument. Therefore,
the cost for each component does not include the cost of the monitoring
instrument. The cost effectiveness for RACT for pumps and compressors is
higher than other components due to the additional time required for leak
repair.
5-13
-------
TABLE 5-9. COST EFFECTIVENESS FOR COMPONENT TYPES IN MODEL UNIT A
on
Component
Pumps (light liquid)
Valves
Gas service
Light liquid service
Safety/relief valves
Open-ended valves
Gas service
Light liquid service
Heavy liquid service
Compressors
TOTAL UNIT (without
instrument cost)
TOTAL UNIT (with
instrument cost)
Number
of
components
8
90
84
11
9
47
48
1
250C
250C
Annuali zed
cost before
credit ($)a
5,239
456
442
295
183
877
643
1,204
9,339
14,825
Annual
recovery
credit ($)
2,444
5,822
2,111
3,715
709
1,857
463
1,140
18,261
18,261
Net
annual ized
cost ($)a
2,795
(5,366)
(1,669)
(3,420)
(526)
(880)
180
64
(8,822)
(3,436)
Total VOC
reduction
(Mg/yr)
5.96
14.20
5.15
9.06
1.73
4.53
1.13
2.78
44.54
44.54
Cost
effectiveness
($/Mg)
469
(378)
(324)
(377)
(304)
(194)
159
23
(198)
(77)
Does not include cost of monitoring instrument, unless otherwise noted.
3Cost for caps on lines only. Not monitored under RACT.
"Total does not include open-ended lines in heavy liquid service.
(XXX) = net credit.
-------
TABLE 5-10. COST EFFECTIVENESS FOR COMPONENT TYPES IN MODEL UNIT B
en
i—•
en
Component
Pumps (light liquid)
Valves
Gas service
Light liquid service
Safety/relief valves
Open-ended valves
Gas service
Light liquid service
Heavy liquid service
Compressors
TOTAL UNIT (without
instrument cost)
TOTAL UNIT (with
instrument cost)
Number
of
components
29
365
335
42
37
189
189
2
999C
999C
Annualized
cost before
credit ($)a
9,486
1,836
1,771
1,128
698
3,532
2,535
1,240
22,226
27,720
Annual
recovery
credit ($)
8,852
23,595
8,421
14,178
2,923
7,466
1,833
2,275
69,543
69,543
Net
annual i zed
cost ($)a
634
(21,759)
(6,650)
(13,050)
(2,225)
(3,934)
702
(1,035)
(49,989)
(41,823)
Total VOC
reduction
(Mg/yr)
21.59
57.55
20.54
34.58
7.13
18.21
4.47
5.55
169.62
169.62
Cost
effectiveness
($/Mg)
29
(378)
(324)
(377)
(312)
(216)
157
(186)
(295)
(247)
Does not include cost of monitoring instrument, unless otherwise noted.
Cost for caps on lines only. Not monitored under RACT.
Total does not include open-ended lines in heavy liquid service.
(XXX) = net credit.
-------
TABLE 5-11. COST EFFECTIVENESS FOR COMPONENT TYPES IN MODEL UNIT C
en
i-«
cn
Component
Pumps (light liquid)
Valves
Gas service
Light liquid service
Safety/relief valves
Open-ended valves
Gas service
' Light liquid service
Heavy liquid service
Compressors
TOTAL UNIT (without
instrument cost)
TOTAL UNIT (with
instrument cost)
Number
of
components
91
1,117
1,037
130
115
581
581
8
3,079C
3,079C
Annualized
cost before
credit ($)a
25,882
5,560
5,491
3,503
2,130
10,868
7,790
2,779
64,001
69,490
Annual
recovery
credit ($)
27,782
72,213
26,072
43,890
9,086
22,952
5,633
9,106
216,734
216,734
Net
annual i zed
cost ($)a
(1,900)
(66,653)
(20,581)
(40,387)
(6,956)
(9,010)
2,157
(6,327)
(149,657)
(147,240)
Total VOC
reduction
(Mg/yr)
67.76
176.13
63.59
107.05
22.16
55.98
13.74
22.21
528.62
528.62
Cost
effectiveness
($/Mg)
(28)
(378)
(324)
(377)
(314)
(161)
157
(285)
(283)
(279)
Does not include cost of monitoring instrument, unless otherwise noted.
Cost for caps on lines only. Not monitored under RACT.
otal does not include open-ended lines in heavy liquid service.
(XXX) = net credit.
-------
5.5. REFERENCES
1. Letter with attachments from Texas Chemical Council to Walt Barber,
U.S. EPA. June 30, 1980.
2. Purchase order from GCA/Technology Division to Analabs/Foxboro, North
Haven, Connecticut. July 3, 1980.
3. Telecon. Samuel Duletsky, GCA Corporation with Dave Myer, Piedmont
Hub, Greensboro, N.C. September 25, 1980. Price of 1" screw-on
type valve.
4. Letter with attachments from J.M. Johnson, Exxon Company, U.S.A., to
Robert T. Walsh, U.S. EPA. July 28, 1977.
5. Erikson, D.G., and V. Kalcevic. Emission Control Options for the
Synthetic Organic Chemicals Manufacturing Industry, Fugitive Emissions
Report, Draft Final. Hydroscience, Inc. 1979. p. IV-9.
*
6. 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.
7. Economic Indicators. Chem. Eng. Vol. 86 #2. January 15, 1979.
8. Economic Indicators. Chem. Eng. Vol. 87 #19. September 22, 1980.
9. Letter from Vincent Smith, Research Triangle Institute, to Russell
Honerkamp, Radian Corporation. November 30, 1979.
10. Reference 8.
11. Economic Indicators. Chem. Eng. Vol. 86 # 1. January 14, 1980.
5-17
-------
APPENDIX A. EMISSION SOURCE TEST DATA
-------
APPENDIX A
EMISSION SOURCE TEST DATA
The purpose of Appendix A is to describe testing results used in the
development of the Control Techniques Guideline (CTG) document for VOC
fugitive emissions from the Synthetic Organic Chemicals Manufacturing Industry
(SOCMI) and the polymer and resin manufacturing industry. The information
in this appendix consists of a description of the tested facilities, and the
sampling procedures and test results of VOC fugitive emissions studies in
SOCMI and the petroleum refining industry.
Fugitive emission sources of VOC in SOCMI and in the petroleum refining
industry are similar. Considerable data exist concerning both the incidence
and magnitude of fugitive emissions from petroleum refineries. Studies of
fugitive emissions in SOCMI have been undertaken by EPA to support the use
of emission factors generated during studies of emissions in petroleum
refineries for similar sources in the Synthetic Organic Chemicals Manufacturing
Industry. The results of the EPA SOCMI studies, EPA data from a study of
fugitive emissions from petroleum refineries, and some industry studies of
fugitive emissions are discussed in Section A.I.
Section A.2 consists of the results of three studies on the effects of
maintenance on reducing fugitive VOC emissions from valves in petroleum
refineries and two studies on maintenance of valves in SOCMI process units.
These results are included as an indication of the reduction in emissions
which could be expected as a function of the designated action level, and by
applying routine on-line maintenance procedures.
A.I FUGITIVE EMISSIONS TEST PROGRAMS
Three SOCMI test programs have been conducted by EPA. One was a study
performed by Monsanto Research Corporation of a small number of fugitive
emission sources in four SOCMI units. More intensive screening was performed
at six SOCMI units in another study. The third EPA study of SOCMI fugitive
A-l
-------
emissions was a screening and sampling program conducted at twenty-four
SOCMI units. The results of these studies are presented in this section.
Similar types of studies have been performed by industry. This section also
contains the results of an Exxon study of fugitive emissions in cyclohexane
unit and a DuPont study of fugitive emissions in unidentified process units.
The results of a study on fugitive emissions from petroleum refineries
are also presented in this section. Data on fugitive emissions were obtained
from 64 units in thirteen refineries located in major refining areas throughout
the country. Data on the effects of maintenance were obtained at the last
four of these refineries. These results are presented later in Section A.2
of this Appendix.
A.1.1 Study of Fugitive Emissions at Four SOCMI Units
Monsanto Research Corporation (MRC) conducted EPA Industrial Environmental
Research Laboratory (EPA-IERL) sponsored study of fugitive emissions at four
SOCMI units. The process units were monochlorobenzene, butadiene, ethylene
oxide/glycol, and dimethyl terephthalate. Due to the small number of
plants/processes sampled and the experimental design of this study, the
results were not considered to be comparable with the results of other
studies. Since the data generated by the MRC study could not be considered
representative of the SOCMI and valid conclusions could not be drawn
concerning the relative magnitude of fugitive emissions in the SOCMI, the
results of the study were not used in the development of standards for
fugitive emissions control. This study demonstrated the need for more
intensive sampling and screening which was undertaken by EPA.
A.1.2 Description and Results of EPA Study of Six SOCMI Units2'3'4'5
The objective of this test program was to gather data on the percentage
of sources which leak (as defined by a VOC concentration at the leak interface
of _> 10,000 ppmv calibrated with methane). To achieve this objective, an
attempt was made to screen all potential leak sources (generally excluding
flanges) on an individual component basis with a portable organic vapor
analyzer. The test crews relied on plant personnel to identify equipment
handling organics. Normally, all pumps and compressor seals were examined,
and the percentage of valves carrying VOC which were screened ranged from 33
A-2
-------
to 85 percent. All tests were performed with a Century Systems Corporation
Organic Vapor Analyzer (OVA), Model 108, with the probe placed as close to
the source as possible. The results of this study are shown in Table A-l.
Six chemical process units were screened. Unit A is a chlorinated
methanes production facility in the Gulf Coast area which uses methanol as
feedstock material. The individual component testing was conducted during
September 1978. Unit B is a relatively small ethylene production facility
on the West Coast which uses an ethane/propane feedstock. Testing was
conducted during October 1978. Unit C is a chlorinated methanes production
facility in the Midwest. This plant also uses methanol as the basic organic
feedstock. OveV the last few years, several pieces of equipment have been
replaced with equipment the company feels is more reliable. In particular,
the company has installed certain types of valves which they have found do
not leak "as much" as other valves. The individual component testing was
conducted during January 1979. Unit D is an ethylene production facility on
the Gulf Coast, using an ethane/propane feed. The facility is associated
with a major refinery, and testing was conducted during March 1979. Units E
and F are part of an intermediate size integrated petroleum refinery located
in the North Central United States. Testing was conducted during November
1978. Unit E is an aromatics extraction unit that produces benzene, toluene,
and xylene by extraction from refined petroleum feedstocks. Unit E is a new
unit and special attention was paid during the design and startup to minimize
equipment leaks. All valves were repacked before startup (adding 2 to 3
times the original packing) and all pumps in benzene service had double
mechanical seals with a barrier fluid. Unit F produces benzene by hydro-
dealkylation of toluene. Unit F was originally designed to produce a different
chemical and was redesigned to produce benzene.
A.1.3 Description and Results of an EPA Study of 24 SOCMI Units
The U.S. EPA Industrial Environmental Research Laboratory coordinated
a study to develop information about fugitive emissions in the SOCMI. A
total of 24 chemical process units were selected for this purpose. The
process units were selected to represent a cross section of the population
of the SOCMI. Factors considered during process unit selections included
annual production volume, number of producers, volatility, toxicity, and
value of the final products. Table A-2 shows the process unit types selected
for screening.
A-3
-------
TABLE A-l. FREQUENCY OF LEAKS FROM FUGITIVE EMISSION SOURCES IN
SYNTHETIC ORGANIC CHEMICAL UNITS (Six Unit Study)
Unit A*
Chloromethanes
Equipment
type
Valves
Open-ended lines
Pimp seals
Compressor seals
Control valves
Pressure relief
valves
Flanges
Drains
Number
of
sources
tested
600
52
47
_e
52
7
30
_e
Percent with
screening
values
MO, 000 ppmv
1
2
15
6
0
3
Unit Ba
Ethyl ene
Number Percent with
of
screening
sources values
tested
2,301
386
51
42
126
e
_e
_e
MO, 000 ppmv
19
11
21
59
20
Unit Cb
Chloromethanes
Number
of
sources
tested
65B
_e
39
3
25
_e
_e
_e
Percent with
screening
values
MO ,000 ppmv
0.1
3
33
0
Unit Dc
Ethyl ene
Number
of
sources
tested
862
90
63
17
25
_e
_e
39
Percent with
screen! ng
values
MO, 000 ppmv
14
13
33
6
44
10
Unit Ed
BTX Recovery
Number Percent with
of screening
sources values
tested MO ,000 ppmv
715 1.1
33 0.0
33f 3.0
_e
S3 4.0
_e
_e
_e
Unit Fd
Toluene H)A
hunber
of
sources
tested
427
28
30
_e
44
e
_e
e
Percent with
screening
values
MO ,000 ppmv
7.0
11.0
10.0
11.0
'Source: Reference 6.
Source: Reference 7.
cSource: Reference 8.
Source: Reference 9.
*No data.
Pump seals In benzene service have double mechanical seals.
-------
TABLE A-2. TWENTY-FOUR CHEMICAL PROCESS UNITS SCREENED FOR FUGITIVE
EMISSIONS
Unit Type
1. Vinyl Acetate
2. Ethylene
3. Vinyl Acetate
4. Ethylene
5. Cumene
6. Cumene
7. Ethylene
8. Acetone/Phenol
9. Ethylene Dichloride
10. Vinyl Chloride Monomer
11. Formaldehyde
12. Ethylene Dichloride
13. Vinyl Chloride Monomer
14. Methyl Ethyl Ketone
15. Methyl Ethyl Ketone
16. Acetaldehyde
17. Methyl Methacrylate
18. Adi pic Acid
19. Tri chloroethylene/Perchloroethylene
20. 1,1,1-Trichloroethane
21. Ethylene Dichloride
22. Adi pic Acid
23. Acrylonitrile
24. Acrylonitrile
Source: Reference 11
A-5
-------
The screening work began with the definition of the process unit
boundaries. All feed streams, reaction/separation facilities, and product
and by-product delivery lines were identified on process flow diagrams and
in the process unit. Process data, including stream composition, line
temperature, and line pressure, were obtained for all flow streams. Each
process stream to be screened was identified and process data were obtained
with the assistance of plant personnel, in most cases. Sources were screened
by a two-person team (one person handling the hydrocarbon detector and one
person recording data).
The Century Systems Models OVA-108 and OVA-128 hydrocarbon detectors
were used for screening. The HNU Systems, Inc., Model PI 101 Photoionization
Analyzer was also used to screen sources at the formaldehyde process unit.
The detector probe of the instrument was placed directly on those areas of
the sources where leakage would typically occur. For example, gate valves
were screened along the circumference of the annular area around the valve
stem where the stem exits the packing gland and at the packing gland/valve
bonnet interface. All process valves, pump seals, compressor seals,
agitator seals, relief valves, process drains, and open-ended lines were
screened. From five to twenty percent of all flanges were randomly selected
and screened. For the purpose of this program "flange" referred to any
pipe-to-pipe or tubing-to-tubing connection, excluding welded joints.
Each screening instrument was calibrated at least daily. The model
OVA-108 instruments, with a logarithmic scale reading from 1 ppmv to 10,000
ppmv, were calibrated with high (8,000 ppmv) and low (500 ppmv) concentration
methane-in-air standards to ensure accurate operation at both ends of the
instrument's range. The model OVA-128 instruments, with a linear readout
ranging from 0 ppmv to 1,000 ppmv, were also calibrated with high and low
concentration standards. A pre-calibrated dilution probe was required with
the OVA-128 when calibrating with the 8,000 ppmv standard.
The HNU Photoionization instrument, used to screen the formaldehyde
process unit, was calibrated with isobutylene which has an ionization potential
close to that of formaldehyde.
A-6
-------
Results of the screening program at the 24 process units are summarized
in Table A-3.
The fugitive emission sources in the study were screened at an average
rate of 1.7 minutes per source for a two-person team (or 3.4 person-minutes
per source). This average screening rate includes time spent for instrument
calibration and repair. Table A-4 presents screening time data on a unit-
by-unit basis. These time requirements are somewhat higher than would be
expected for routine monitoring because of the extensive record keeping
associated with the screening project.
A.1.4 Description and Results of Refinery Fugitive Emissions Study12
Data concerning the leak frequencies and emission factors for various
fugitive sources were obtained primarily at nine refineries. More complete
information for compressors' and relief valves' emissions was obtained by
sampling at four additional refineries. 'Refineries were selected to provide
a range of sizes and ages and all of the major petroleum refinery processing
units were studied. The type of process units and the number of each studied
in the first nine refineries are listed in Table A-5.
In each refinery, sources in six to nine process units were selected
for study. The approximate number of sources selected for study and testing
in each refinery is listed below:
Valves 250-300
Flanges 100-750
Pump seals 100-125
Compressor seals 10-20
Drains 20-40
Relief Valves 20-40
There were normally 500-600 sources selected in each refinery.
The distribution of sources among the process units was determined
before the selection and testing of individual sources was begun. Individual
sources were selected from piping and instrumentation diagrams or process
flow diagrams before a refinery processing area was entered. Only those
preselected sources were screened. In this way, bias based on observation
of individual sources was theoretically eliminated.
A-7
-------
TABLE A-3. SUMMARY OF SOCMI PROCESS UNITS FUGITIVE EMISSIONS
(Twenty-four unit study)
>
oo
Source Type
Flanges
Process Drains
Open Ended Lines
Agitator Seals
Relief Valves
Valves
Pumps
Compressors
Other8
Service
Gas
Light Liquid
Heavy Liquid
Gas
Light Liquid
Heavy Liquid
Gas
Light Liquid
Heavy Liquid
Gas
Light Liquid
Heavy Liquid
Gas
Light Liquid
Heavy Liquid
Gas
Light Liquid
Heavy Liquid
Light Liquid
Heavy Liquid
Gas
Gas
Light Liquid
Heavy Liquid
(1)
Number
Screened
1.443
2,897
607
83
527
28
923
3.603
477
7
8
1
85
69
3
9.668
18.294
3.632
647
97
29
19
33
2
(2)
X Not
Screened
4.6
2.6
2.4
23.1
1.9
0.0
17.5
10.4
21.5
46.1
11.1
66.7
72.7
40.5
66.7
17.5
12.2
9.9
4.3
40.5
9.4
9.5
19.5
33.3
(3) (4)
% of Screened Sources 95% Confidence Interval
with Screening Values for Percentage of Sources
•s. 10.000 ppmv 2 10.000 pprav
4.6 (3.6. 5.8
1.2 (0.9, 1.8
0.0 (0.0. 0.6
2.4
3.8
7.1
5.8
3.9
1.3
14.3
0.0
0.0
3.5
2.9
0.0
11.4 (]
6.4
0.4
8.8
2.1
0.3. 8.4
2.3, 5.8
0.9, 23.5)
4.4, 7.5
3.3, 4.6
0.5, 2.8
0.4, 57.9
0.0, 36.9
0.0, 97.5
0.7, 10.0
0.3. 10.1
0.0. 70.8
10.8. 12.1)
6.1. 6.8
0.2. 0.7
6.6, 11.1)
0.3, 7.3)
6.9 (0.9. 22.8)
21.0
6.1
0.0
6.0, 45.6
0.7, 20.2
0.0. 84.2
Includes filters, vacuum breakers, expansion joints, rupture disks, slight glass seals, etc. Source: Ref. 13
-------
TABLE A-4. AVERAGE FUGITIVE EMISSION SOURCE SCREENING RATES
(Twenty-four Unit Study)
Average Screening
Number of Time Per
Process Unit Type Screened Sources Source, Minutes
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
20.
21.
22.
23.
24.
Vinyl Acetate
Ethyl ene
Vinyl Acetate
Ethyl ene
Cumene
Cumene
Ethyl ene
Acetone/Phenol
Ethylene Dichloride
Vinyl Chloride Monomer
Formaldehyde
Ethylene Dichloride
Vinyl Chloride Monomer
Methyl Ethyl Ketone
Methyl Ethyl Ketone
Acetaldehyde
Methyl Methacrylate
Adi pic Acid
Tri chol oroethyl ene/Perchl oroethyl ene
1 ,1 ,1-Trichloroethane
Ethylene Dichloride
Adi pic Acid
Acrylonitrile
Acrylonitrile
1,391
5,078
2,780
5,278
1,025
1,573
3,685
3,207
1,430
868
230
744
2,619
585
679
1,148
2,019
1,577
2,720
570
42
664
1,406
1,864
2.0
1.3
0.9
1.5
0.9
1.0
1.9
3.2
2.6
1.8
1.6
1.6
2.2
1.2
0.9
0.7
1.6
1.9
2.5
1.9
Total 43,182 1.7
aAverage source screening time was determined for a two-person team,
one person screening with a portable hydrocarbon detector and one
person recording data. Average screening time includes time spent
for instrument calibration, maintenance, and repair.
Source: Ref. 14
A-9
-------
TABLE A-5. SAMPLED PROCESS UNITS FROM NINE REFINERIES
DURING REFINERY STUDY
Refinery Number of
process unit sampled units
Atmospheric distillation 7
Vacuum distillation 4
Thermal operations (coking) 2
Catalytic cracking 5
Catalytic reforming 6
Catalytic hydrocracking 2
Catalytic hydrorefining 2
Catalytic hydrotreating 7
Alkylation 6
Aromatics/isomerization 3
Lube oil manufacture 2
Asphalt manufacture 1
Fuel gas/light-ends processing 11
LPG 2
Sulfur recovery 1
Other 3
Source: Ref. 15
A-10
-------
The screening of sources was accomplished with portable organic vapor
detectors. The principal device used in this study was the J. W. Bacharach
Instrument Co. "TLV Sniffer" calibrated with hexane. The components were
tested on an individual basis and only those components with VOC concentrations
in excess of 200 ppmv were considered for further study.
A substantial portion of these leaking sources was enclosed and sampled
to determine both the methane and nonmethane emission rates. An important
result of this program was the development of a correlation between the
maximum observed screening value (VOC concentration) and the measured nonmethane
leak rate.
Emission factors and leak frequency information generated during this
study are given in Table A-6.
A.1.5 Comparison of Fugitive Emissions Test Data from Refineries and
SOCMI Units
The results of the SOCMI studies and those of the refinery emissions
study are compared in Table A-7.
A.1.6 Description and Results of the DuPont Study
DuPont conducted a program of VOC fugitive emission measurement from
pumps and valves at two of their plants. The processes of the 5 and 10 year
old plants were not revealed. The OVA-108 was used for screening (leak
identification) and for leak rate determination (analysis of collected leak
vapors). The leak rate was determined by taking Tedlar bags partially
filled with air and enclosing the leaking valve. The hydrocarbon concentration
in the bags was recorded as a function of time. Visual estimates of the
initial bag volume were assumed to be ±5 percent. Dupont did not have a
dilution probe and, therefore, measurements above 10,000 ppmv were not
made. Analysis of the data collected indicates that no significant difference
in leak rates exists between manual and automatic control valves. Significant
trends were observed with changes in product vapor pressure. It also seemed
that full open or closed valve seat positions resulted in lower leak rates
than intermediate positions. The results of the DuPont study are shown in
Table A-8.
24 25
A.1.7 Description and Results of the Exxon Study '
A fugitive emissions study was conducted by Exxon Chemical Company at
the cyclohexane unit at their Baytown plant. The total number of valves,
pumps and compressor seals, and safety valves was determined. For all
A-ll
-------
TABLE A-6. LEAK FREQUENCIES AND EMISSION FACTORS FROM FUGITIVE
..EMISSION SOURCES .IN PETROLEUM REFINERIES
Percent of
sources having
screening values
Equipment 2. 10,000 ppmv
type TLV-Hexane
Valves
Gas service
Light liquid service
Heavy liquid service
Pump seals
Light liquid service
Heavy liquid service
Compressor seals (hydrocarbon
service)
Pressure relief valves
Gas service
Light liquid service
Heavy liquid service
Flanges
Open-ended lines
Gas service
Light liquid service
Heavy liquid service
NA
10
12
0
NA
23
2
33
8
0
NA
Estimated emission
factor for
refinery sources,
kg/hr-source
NA
0.021
0.010
0.0003
NA
0.12
0.02
0.44
0.086
0.16
0.006
0.009
0.0003
NA
0.025
0.014
0.003
Source: Ref. 17
A-12
-------
TABLE A-7. COMPARISON OF LEAK FREQUENCIES FOR FUGITIVE EMISSION
SOURCES IN SOCMI UNITS AND PETROLEUM REFINERIES
Equipment Type
Valves (all)
Gas
Light Liquid
Heavy Liquid
Open-ended lines (all)
Gas
Light Liquid
Heavy Liquid
Pumps (all)
Light Liquid
Heavy Liquid
Percent of SOCHI Sources
Having Screening Values
S 10.000 ppmv, OVA-108
Methane (six unit study)
11
10
17
Percent of SOCHI Sources
Having Screening Values
£ 10,000 ppmv, OVA-108 b
Methane (24 unit study)
11.4
6.4
0.4
5.8
3.9
1.3
8.8
2.1
Percent of Petroleum
Refinery Sources Having
Screening Values
i 10, 000 ppmv, TLV-Hexanec
10
12
0
N/A
23
2
Compressors (Gas)
Pressure Relief Valves (all)
43
0
6.9
33
Gas
Light Liquid
Heavy Liquid
Flanges (all) 3
Gas
Light Liquid
Heavy Liquid
Process Drains (all) N/A
Gas
Light Liquid
Heavy Liquid
Agitator Seals (all) N/A
Gas
Light Liquid
Heavy Liquid
Other* N/A
3.5
2.9
0.0
4.6
1.2
0.0
2.4
3.8
7.1
14.3
0.0
0.0
N/A
N/A
N/A
0
N/A •
0.00
7.00
3.74
N/A
N/A
"Source: Ref. 18, 19, 20, 21
bSource: Ref. 22
cSource: Ref. 23. Screening with OVA-108 (methane) at 10,000 pprav is equivalent to screening with TLV (hexane)
at 4,121 ppmv.
Includes filters, vacuum breakers, expension joints, rupture disks, sight glass seals, etc.
A-13
-------
TABLE A-8. FREQUENCY OF LEAKS3 FROM FUGITIVE EMISSION
SOURCES IN TWO DuPONT PLANTS.
Equipment
type
Valves
Gas
Light liquid
Heavy liquid
Pumps
Light liquid
Heavy liquid
No. of
leakers
47
35
11
1
1
1
0
No. of
non-leakers
741
120
143
478
35
6
29
Percent
leakers
6.1
23.1
7.1
0.2
2.7
14.3
0
aLeak defined as 10 ppmv or greater.
Source: Ref. 26
A-14
-------
sources, except valves, all of the fugitive emission sources were sampled.
For valves, a soap solution was used to determine leaking components. All
leaking valves were counted and identified as either small, medium or large
leaks. From the set of valves found to be leaking, specific valves were
selected for sampling so that each class of leaking valves was in approximately
the same proportion as it occured in the cyclohexane unit.
Heat resistant mylar bags or sheets were taped around the equipment to
be sampled to provide an enclosed volume. Clean metered air from the
filter apparatus was blown into the enclosed volume. The sampling train was
allowed to run until a steady state flow was obtained (usually about 15
minutes). A bomb sample was then taken for laboratory analysis (mass
spectrometry). Table A-9 presents the results of the Exxon study.
A.2 MAINTENANCE TEST PROGRAMS
The results of four studies on the effects of maintenance on fugitive
emissions from valves are discussed in this section. The first two studies
were conducted by refinery personnel at the Union Oil Co. refinery in Rodeo,
California, and the Shell Oil Co. refinery in Martinez, California. These
programs consisted of maintenance on leaking valves containing fluids with
Reid vapor pressures greater than 1.5. The third study was conducted by
EPA. Valves were selected and maintained at four refineries. The fourth
study was conducted by EPA at Unit D (ethylene unit). The study results and
a description of each test program are given in the following sections.
29
A.2.1 Description and Results of the Union Maintenance Study
The Union valve maintenance study consisted of performing undirected
maintenance on valves selected from 12 different process units. Maintenance
procedures consisted of adjusting the packing gland while the valve was in
service. Undirected maintenance consists of performing valve repairs
without simultaneous measurement of the effect of repair on the VOC concentration
detected. This is in contrast to directed maintenance where emissions are
monitored during the repair procedure. With directed maintenance, repair
procedures are continued until the VOC concentration detected drops to a
specified level or further reduction in the emission level is not possible.
Also, maintenance may be curtailed if increasing VOC concentrations result.
A-15
-------
TABLE A-9. FREQUENCY OF LEAKS3 FROM FUGITIVE EMISSION
SOURCES IN EXXON'S CYCLOHEXANE UNIT
Equipment
Source
Valves
Gas
Light
liquid
Safety
valves
Pump .
seals
Compressor
sealsb
Total
in Unit
136
201
15
8
N/A
Screened and
Sampl ed
136
100
15
8
N/A
Percent Emission
Leaking factor(kg/hr)
32
15
87
83
100
0.017
0.008
0.064
0.255
0.264
99.8% Confidence
Interval (kg/hr)
0.008 - 0.035
0.003 - 0.007
0.013 - 0.5
0.082 - 0.818
0.068 - 1.045
N/A - Not available
aLeak defined as 200 ppmv or greater.
Double mechanical seal pumps and compressors were found to have negligible
leaks.
Source: Reference 27,28
A-16
-------
The Union data were obtained with a Century Systems Corporation Organic
Vapor Analyzer, OVA-108. All measurements were taken at a distance of 1 cm
from the seal. Correlations developed by EPA have been used to convert the
data from OVA readings taken at one centimeter to equivalent TLV readings at
the leak interface (TLV-0).30 This facilitates comparison of data from
different studies and allows the estimation of emission rates based on
screening values-leak rate correlations.
The results of the Union study are given in Table A-10. Two sets of
results are provided; the first includes all repaired valves with before
maintenance screening values greater than or equal to 5,300 ppmv (OVA-108),
and the second includes valves with before maintenance screening values
below 5,300 ppmv (OVA-108). A screening value of 5,300 ppmv, obtained with
OVA at 1 cm from the leak interface, is equivalent to a screening value of
10,000 ppmv measured by a Bacharach Instrument Co. "TLV Sniffer" directly at
the leak interface. The OVA-1 cm readings have been converted to equivalent
TLV-0 cm readings because:
1) EPA correlations which estimate leak rates from screening values
were developed from TLV-0 cm data.
2) Additional maintenance study data exists in the TLV-0 cm format.
3) EPA Reference Method 21 specifies 0 cm screening procedures.
The results of this study indicate that maintenance on valves with
initial screening values above 10,000 ppmv (OVA-108) is much more effective
than maintenance on valves leaking at lower rates. In fact, this study
indicates that emissions from valves are reduced by an average of 51.8
percent for valves initially over 5,300 ppmv while valves with lower initial
screening values experienced an increase of 30.5 percent.
A.2.2 Description and Results of the Shell Maintenance Study
The Shell maintenance program consisted of two parts. First, valve
repairs were performed on 171 leaking valves. In the second part of the
program, 162 of these valves were rechecked and additional maintenance was
performed. Maintenance consisted of adjusting the packing gland while the
valve was in service. The second part of the program was conducted approximately
one month after the initial maintenance period. It was not determined
whether the maintenance procedures were directed or undirected, based on the
information reported by Shell.
A-17
-------
TABLE A-10. SUMMARY OF MAINTENANCE STUDY RESULTS FROM THE UNION OIL CO.
REFINERY IN RODEO, CALIFORNIA9
00
All valves
with initial
screening values
^5300 ppmvb
Source: Ref. 33.
All valves
with initial
screening values
<5300 ppmv
Number of repairs attempted
Estimated emissions before maintenance, kg/hr
Estimated emissions after maintenance, kg/hr
Number of successful repairs (<5300 ppmv after maintenance)
Number of valves with decreased emissions
Number of valves with increased emissions
Percent reduction in emissions
Percent successful repairs
Percent of valves with decreased emissions
Percent of valves with increased emissions
133
9.72
4.69
67
124
9
51.8
50.4
93.2
6.8
21
0.323
0.422
--
13
8
-30.5
—
61.9
38.1
value of 5300 ppmv, taken with the OVA-108 at 1 cm, generally corresponds to a value of
10,000 ppmv taken with a "TLV Sniffer" at 0 cm.
-------
Fugitive VOC emissions were measured one centimeter from the source
using the OVA-108. These data have been transformed to TLV-0 cm values as
were the Union data. The same methods of data analysis described in
Section A.2.1 have been applied to the Shell data. The results of the
Shell maintenance study are given in Table A-ll.
A.2.3 Description and Results of the EPA Maintenance Study32
Repair data were collected on valves located in four refineries. The
effects of both directed and undirected maintenance were evaluated. Maintenance
consisted of routine operations, such as tightening the packing gland or
adding grease. Other data, including valve size and type and the processes'
fluid characteristics, were obtained. Screening data were obtained with the
Bacharach Instrument Company "TLV Sniffer" and readings were taken as close
to the source as possible.
Unlike the Shell and Union studies, emission rates were not based on
the screening value correlations. Rather, each valve was sampled to determine
emission rates before and after maintenance using techniques developed by
EPA during the refinery emission factor study. These values were used to
evaluate emissions reduction.
The results of this study are given in Table A-12. Of interest here is
a comparison of the emissions reduction for directed and undirected maintenance.
The results indicate that directed maintenance is more effective in reducing
emissions than is undirected maintenance, particularly for valves with lower
initial leak rates. The results showed an increase in total emissions of
32.6% for valves with initial screening values less than 10,000 ppmv which
were subjected to undirected maintenance. However, this increase is due to
a large increase in the emission rate of only one valve.
A.2.4 Description and Results of Unit D (Ethylene Unit) Maintenance Study
Maintenance was performed by Unit D personnel. Concentration measurements
of VOC were made using the OVA-108, and readings were obtained at the
closest distance possible to the source. The results of this study are
shown in Table A-13. Directed and undirected maintenance procedures were
used. The results show that directed maintenance results in more repairs
being successfully completed than when undirected maintenance is used.
A-19
-------
TABLE A-11.
SUMMARY OF MAINTENANCE STUDY RESULTS FROM THE SHELL OIL COMPANY
REFINERY IN MARTINEZ, CALIFORNIA3
I
ro
O
Number of repairs attempted
Estimated emissions before
maintenance, kg/hr
Estimated emissions after
maintenance, kg/hr
Number of successful repairs
(<5300 ppmv after maintenance)
Number of valves with decreased
emissions
Number of valves with Increased
emissions
Percent reduction In emissions
Percent successful repairs
Percent of valves with decreased
emissions
Percent of valves with Increased
emissions
March
All repaired valves
with Initial screening
values >5300 ppmvb
161
11.08
2.66
105
161
0
76.0
65.2
100.00
0.0
maintenance
All repaired valves
with Initial screening
values <5300 ppmv
11
0.159
0.0
—
11
0
100.0
"
100.0
0.0
April
All repaired valves
with Initial
(March) screening
values _>5300 ppmv
152d
2.95
0.421
45
151
1
B5.7
83.3
99.3
0.7
maintenance
All repaired valves
with initial
(March) screening
values <5300 ppmv
lle
0.060
0.0
—
11
0
100.0
—
100.0
0.0
aSource: Reference 34.
bThe value of 5300 ppmv. taken with the OVA-108 at 1 on. generally corresponds to a value of 10,000 ppmv taken with a "TLV Sniffer" at 0 cm.
""Shell reported the screening value of a]l valves which measured <3000 ppmv (<1500 ppmv-TLV at 0 cm) as non-leakers. Emissions estimates obtained from
emission factors. Reference 14.
Initial value of 90 of these valves was <1SOO ppm-TLV at 0 cm. 54 valves screened j>5300 (note nine valves from Initial data set not rechecked In April).
eln1t1al value of 10 of these valves was <1500 ppm-TLV at 0 cm.
-------
TABLE A-12. SUMMARY OF EPA REFINERY MAINTENANCE STUDY RESULTS
I
ro
Repaired values with initial
screening values >10,000 ppmv
Number of valves repaired
Measured emissions before
maintenance, kg/hr
Measured emissions after
maintenance, kg/hr
Number of successful repairs
(<10,000 ppmv after maintenance)
Number of valves with decreased
emissions
Number of valves with increased
emissions
Percent reduction in emissions
Percent successful repairs
Percent of valves with decreased
emissions
Percent of valves with increased
emissions
Di rected
maintenance
9
0.107
0.0139
8
9
0
87.0
88.9
100.0
0.0
Undirected
maintenance
23
1.809
0.318
13
21
2
82.4
56.5
91.3
8.7
Repaired values with initial
screening values <10,000 ppmv
Directed
maintenance
10
0.0332
0.0049
--
6
4
85.2
--
60.0
40.0
Undirected
maintenance
16
0.120
0.159
--
15
1
-32.6
--
93.8
6.3
Source: Ref. 36
-------
TABLE A-13. MAINTENANCE EFFECTIVENESS
UNIT D ETHYLENE UNIT BLOCK VALVES
1. Total number of valves with VOC > 10,000 ppm
from unit survey 121
2. Total number of valves tested for
maintenance effectiveness 46
% Tested 38%
UNDIRECTED MAINTENANCE
3. Total number subjected to repair attempts 37
4. Successful repairs (VOC<10,000 ppm) 22
% Repaired 59%
Fo-llowup
DIRECTED MAINTENANCE
5. Number of valves unrepaired by undirected 14
maintenance subjected to directed maintenance
6. Number repaired by followup directed maintenance 5
% of unsuccessful repaired by
directed maintenance 36%
7. Total number repaired based on undirected 27
maintenance subset (3) above
% Repaired 73%
8. Total number of repairs including leaks not 29
found before initial maintenance
Total % repaired 63%
Total % not repaired 37%
Source: Reference 37
A-22
-------
38
A.2.5 Description and Results of EPA-ORD Valve Maintenance Study
A study was undertaken by the EPA Office of Research and Development
(ORD) in order to determine the effectiveness of routine (on-line) maintenance
in the reduction of fugitive VOC emissions from in-line valves. The overall
effectiveness of a leak detection and repair program was examined by studying
the immediate emission reduction due to maintenance, the propagation of the
leaks after maintenance, and the rate at which new leaks occur for pumps and
valves. Testing was conducted at six chemical plants, two for each of three
chemical processes (ethylene, cumene, and vinyl acetate production).
It was found that an estimated 71.3 percent (95 percent confidence
limits of 54 percent to 88 percent) reduction in fugitive emissions from all
valves leaking at various concentrations resulted immediately following
maintenance (lasting up to six months). The 30-day rates of occurrence for
valves and pumps initially screened at less than 10,000 ppm were 1.3 percent
(95 percent confidence interval of 0.7 percent to 2.1 percent) and 5.5 percent
(95 percent confidence interval of 2.2 percent to 10 percent), respectively,
as shown in Table A-14. In Table A-15, 30-day, 90-day, and 180-day recurrence
rate estimates are given along with approximate 95 percent confidence limits.
Maintenance of valves in the study averaged about 10 minutes per valve.
A.2.6 Comparison of Maintenance Study Results
A summary of the results of the maintenance programs described in the
preceding sections is presented in Table A-16. Generally speaking, the
results of these maintenance programs would tend to support the following
conclusions:
• A reduction in emissions may be obtained by performing maintenance
on valves with screening values above 10,000 ppmv (measured at the
source).
• The reduction in emissions due to maintenance of valves with screening
values below 10,000 ppmv is not as dramatic and may result in
increased emissions.
A-23
-------
TABLE A-14. OCCURRENCE RATE ESTIMATES FOR VALVES AND PUMPS BY PROCESS IN EPA-ORD STUDY
a,b
30-day
estimate
VALVES
Cumene units
Ethyl ene units
Vinyl acetate units
All units
PUMPS
Cumene units
i\> Ethyl ene units
Vinyl acetate units
All units
1
2
0
1
5
18
2
5
.9
.0
.3
.3
.8
.4
.8
.5
95 percent
confidence
interval
(0.2,
(0.9,
(0.0,
(0.7,
(0.7,
(2.8,
(0.8,
(2.2,
5.9)
3.6)
0.6)
2.1)
20)
42)
6.2)
10)
90- day
estimate
5
6
0
3
16
45
8
15
.6
.0
.8
.8
.3
.7
.1
.7
95 percent
confidence
interval
(0.6,
(2.7,
(0.1,
(2.0,
(2.1,
(8.2,
(2.2,
(6.6,
17)
10)
1.9)
6.0)
49)
80)
17)
27)
180-day
estimate
10.8
11.6
1.5
7.4
30.0
70.5
15.6
29.0
95 percent
confidence
interval
(1.3,
(5.3,
(0.3,
(4.0,
(4.2,
(16,
(4.4,
(12,
30)
20)
3.8)
12)
74)
96)
32)
47)
Reference 38.
JA leak from a source is defined as having occurred if it initially screened <10,000 ppmv and at some
later date screened >10,000 ppmv.
-------
TABLE A-15. VALVE LEAK RECURRENCE RATE ESTIMATES3'b
Recurrence rate 95 percent confidence
estimate limits on the recurrence
(percent) rate estimate
30-day
60-day
180-day
17.2
23.9
32.9
(5,
(7,
(10,
37)
48)
61)
Reference 38.
Data from 28 maintained valves were examined. Only those valves that
screened greater than or equal to 10,000 ppmv immediately before maintenance
and screened less than 10,000 ppmv immediately after maintenance were
considered having a potential to recur.
A-25
-------
TABLE A-16. SUMMARY OF VALVE MAINTENANCE TEST RESULTS
Maintenance
test
Number of
valve repairs
attempted
Number of
successful
repairs
Percent
repaired
Union3 133 67 50.4
Shell3
March 1979 161 105 65.2
April 1979 54 45 83.3
EPA-4 refineries
Directed0 . 9 8 88.9
Undirected0 23 13 56.5
Unit D (Ethylene)b
Directed and undirected 46 29 63.0
EPA-ORDb
Directed 97 28
TOTAL 523 295
3Ihitial screening value of >5,300 ppmv at 1 cm was used to define the
population subject to repair. Repair was successful when a valve screened
<5,300 ppmv at 1 cm.
Before maintenance screening value of_>10,000 ppmv at 0 cm was used to
define the population subject to repair. Repair was successful when a
valve screened <10,000 ppmv at 0 cm.
C0irected maintenance refers to a valve maintenance procedure whereby the
hydrocarbon detector is utilized during maintenance. The leak is monitored
with the instrument until no further reduction of leak is observed or the
valve stem rotation is restricted.
Undirected maintenance refers to action by plant personnel in which an
assigned worker tightens the valve packing gland with a wrench to further
compress the packing material around the valve stem and seat.
A-26
-------
The information presented in the tables of Appendix A has been compiled with
the objective of placing the data on as consistent a basis as possible. However,
some differences were unavoidable and others may have gone unrecognized, due
to the limited amount of information concerning the details of methods used
in each study. Therefore, care should be exercised before attempting to draw
specific quantitative conclusions based on direct comparison of the results of
these studies.
A-27
-------
A.3 REFERENCES
1. Tlchenor, Bruce A., memo to K.C. Hustvedt, U.S. Environmental Protection
Agency, Office of Air Quality Planning and Standards. Research Triangle
Park, N.C. October 27, 1980.
2. Muller, Christopher, memo to files. U.S. Environmental Protection
Agency, Emission Standards and Engineering Division. Research Triangle
Park, N.C. January 18, 1979 (Plants A & B).
3. Muller, Christopher, memo to file. U.S. Environmental Protection
Agency, Emission Standards and Engineering Division, Research Triangle
Park, N.C. March 19, 1979 {Plant C).
4. Muller, Christopher, memo to file. U.S. Environmental Protection
Agency, Emission Standards and Engineering Division. Research Triangle
Park, N.C. March 19, 1979. (Plant D).
5. Hustvedt, K.C., trip report to James F. Durham, Chief, Petroleum
Section. U.S. Environmental Protection Agency. January 5, 1979
(Plants E & F).
6. Reference 2.
7. Reference 3.
8. Reference 4.
9. Reference 5.
10. Blacksmith, J.R., G.E. Harris, and G.J. Langley, Frequency of Leak
Occurence for Fittings in Synthetic Organic Chemical Plant Process
Units. EPA Contract Numbers 68-02-3176-01,06/68-02-3173-02,11/68-02-
3171-01/68-02-3174-04. Radian Corporation, Austin, Texas. September
1980.
11. Reference 10.
12. Wetherold, R.G., and L.P. Provost, Emission Factors and Frequency of
Leak Occurence for Fittings in Refinery Process Units. EPA-600/2-79-
044. Radian Corporation, Austin, Texas. February 1979.
13. Reference 10.
14. Reference 10.
15. Reference 12.
16. Meeting Report. Meeting held between U.S. Environmental Protection
Agency and DuPont at Durham, N.C. June 12, 1979.
A-28
-------
17. Reference 12.
18. Reference 2.
19. Reference 3.
20. Reference 4.
21. Reference 5.
22. Reference 10.
23. Reference 12.
24. Fugitive loss study summary and memo from James B. Cox, Exxon Chemical
Company, Baytown, Texas, to J.W. Blackburn, Hydroscience Incorporated.
February 21, 1978.
25. Letter from James B. Cox, Exxon Chemical Company, Baytown, Texas, to
R.T. Walsh, U.S. Environmental Protection Agency, Chemical and Petroleum
Branch. March 21, 1979.
26. Reference 16.
27. Reference 24.
28. Reference 25.
29. Valve Repair Summary and Memo from F.R. Bottomley, Union Oil Company,
Rodeo, Calfornia, to Milton Feldstein, Bay Area Quality Management
District. April 10, 1979.
30. Reference 12.
31. Valve Repair Summary and Memo from R.M. Thompson, Shell Oil Company,
Martinez Manufacturing Complex, Martinez, California. To Milton
Feldstein, Bay Area Quality Management District. April 26, 1979.
32. Radian Corporation. Assessment of Atmospheric Emissions from Petroleum
Refining, Final Report, Appendix B, Detailed Results. EPA Report No.
600/2-80-075C, Exhibit B. Austin, Texas. April, 1980.
33. Reference 29.
34. Reference 31.
A-29
-------
35. Air Pollution Emission test at Phillips Petroleum Company, Sweeney,
Texas. EMB Report No. 78-DCM-12E, December 1979.
36. Reference 32.
37. Reference 35.
38. Langley, G.J. and R.G. Wetherold. Evaluation of Maintenance for Fugitive
VOC Emissions Control. Radian Corporation. Industrial Environmental
Research Laboratory. Cincinnati, OH. EPA-600/S2-81-080. May 1981.
A-30
-------
APPENDIX B. LIST DEFINING SYNTHETIC ORGANIC CHEMICAL, POLYMER,
AND RESIN MANUFACTURING INDUSTRIES
-------
APPENDIX B
LIST OF CHEMICALS DEFINING SYNTHETIC
ORGANIC CHEMICAL, POLYMER, AND RESIN
MANUFACTURING INDUSTRIES
TABLE I: Synthetic Organic Chemicals Manufacturing Industry
OCPDB No.*_ Chemical
20 Acetal
30 Acetaldehyde
40 Acetaldol
50 Acetamlde
65 AcetanlHde
70 Acetic acid
80 Acetic anhydride
90 Acetone
100 Acetone cyanohydrin
110 Acetonltrile
120 Acetophenone
125 Acetyl chloride
130 Acetylene
140 Acrolein
150 Acrylamlde
160 Acrylic add and esters
170 Acrylonltrlle
180 Adlpic acid
185 Adipon1tr11e
190 Alkyl naphthalenes
200 Ally! alchohol
210 Allyl chloride
220 Am1nobenzo1c add
OCPDB No. Chemical
230 Aminoethylethanolamlne
235 p-amlnophenol
240 Arayl acetates
250 Amy! alcohols
260 Amyl amine
270 Arayl chloride
280 Aroyl mercaptans
290 Arayl phenol
300 Aniline
310 Aniline hydrochloride
320 Anisidine
330 Anisole
340 Anthranilic acid
350 Anthraquinone
360 Benzaldehyde
370 Benzamide
380 Benzene
390 Benzenedisulfonic add
400 Benzenesulfonic acid
410 Benzil
420 BenziHc acid
430 Benzole acid
440 Benzoin
*The OCPDB Numbers are reference indices assigned to the various chemicals
in the Organic Chemical Producers Data Base developed by EPA.
B-l
-------
OCPD8 NO.
Chemical
OCPDB No.
Chemical
450 BenzonltMle
460 Benzophenone
480 Benotr-1 chloride
490 Benzoyl chloride
500 Benzyl alcohol
510 Benzyl aralne
520 Benzyl benzoate
530 Benzyl chloride
540 Benzyl d1chloride
550 B-lphenyl
560 Blsphenol A
570 Bromobenzene
580 Brononaphthalene
590 Butadiene
592 1-butene
600 n-butyl acetate
630 n-butyl acrylate
640 n-butyl alcohol
650 s-butyl alcohol
660 t-butyl alcohol
670 n-butyl amine
680 s-butyl anine
690 t-butyl anrlne
700 p-tert-butyl benzole add
710 1,3-butylene glycol
750 n-butyraldehyde
760 Butyric acid
770 Butyric anhydride
780 Butyronltrile
785 Caprolactam
790 Carbon disulfide
800 Carbon tetrabromide
810 Carbon tetrachloride
820 Cellulose acetate
840 Chloroacetic add
850 ra-chloroani1i ne
860 o-chloroanillne
870 p-chloroanH1ne
880 Chlorobenzaldehyde
890 Chlorobenzene
900 Chlorobenzoic add
905 Chlorobenzotrichlorlde
910 Chlorobenzoyl chloride
920 Chlorodlfluoroethane
921 Chlorodlfluoromethane
930 Chloroform
940 Chloronaphthalene
950 o-chloronitrobenzene
951 p-chloronitrobenzene
960 Chlorophenols
964 Chloroprene
965 Chlorosulfonlc add
970 ra-chloroto1uene
980 o-chlorotoluene
990 p-chlorotoluene
992 ChlorotHfluororaethane
1000 m-cresol
1010 o-cresol
1020 p-cresol
1021 Mixed cresols
1030 Cresyllc add
1040 Crotonaldehyde
1050 Crotonic acid
1060 Cumene
1070 Cumene hydroperoxide
1080 Cyanoacetic acid
1090 Cyanogen chloride
1100 Cyanuric add
1110 Cyanuric chloride
1120 Cyclohexane
1130 Cyclohexanol
1140 Cyclonexanone
B-2
-------
OCPDB No. Chemical
1150 Cyclohexene
1160 Cyclohexylamine
1170 Cyclooctadlene
1180 Decanol
1190 D1acetone alcohol
1200 D1am1nobenzo1c acid
1210 Dichloroaniline
1215 m-dichlorobenzene
1216 o-dichlorobenzene
1220 p-d1chlorobenzene
1221 Dlchlorodifluoromethane
1244 l,2-d1chloroethane (EDC)
1240 Dichloroethyl ether
1250 Dichlorohydrin
1270 Dichloropropene
1280 D1Cyclohexylamine
1290 D1ethyl amine
1300 Diethylene glycol
1304 Diethylene glycol dlethyl ether
1305 Diethylene glycol dimethyl ether
1310 Diethylene glycol monobutyl ether
1320 Diethylene glycol monbutyl ether acetate
1330 Diethylene glycol monoethyl ether
1340 Diethylene glycol nonoethyl ether acetate
1360 Diethylene glycol nonomethyl ether
1420 Dlethyl sulfate
1430 Dif1uoroethane
1440 DHsobutylene
1442 DUsodecyl phthalate
1444 OUsooctyl phthalate
1450 Oiketene
1460 Dimethylamine
1470 N,N-d1methylan1line
1480 N,N-d1methyl ether
1490 N,N-d1methylformam1de
OCPDB No. Chemical
1495 Dimethylhydrazine
1500 Dimethyl sulfate
1510 Dimethyl sulflde
1520 Dimethyl sulfoxide
1530 Dimethyl terephthalate
1540 3,5-d1n1trobenzo1c acid
1545 Dinitrophenol
1550 Dinltrotoluene
1560 Dioxane
1570 Dloxolane
1580 D1 pheny1 amine
1590 Dlphenyl oxide
1600 Dlphenyl thiourea
1610 Dipropylene glycol
1620 Dodecene
1630 Dodecylaniline
1640 Dodecylphenol
1650 Epichlorohydrin
1660 Ethanol
1661 Ethanolamines
1670 Ethyl acetate
1680 Ethyl acetoacetate
1690 Ethyl acrylate
1700 Ethylamine
1710 Ethyl benzene
1720 Ethyl bromide
1730 Ethylcellulose
1740 Ethyl chloride
1750 Ethyl chloroacetate
1760 Ethylcyanoacetate
1770 Ethylene
1780 Ethylene carbonate
1790 Ethylene chlorohydrin
1800 Ethylenediamine
1810 Ethylene dlbromide
1830 Ethylene glycol
B-3
-------
OCPD8 No. Chemical
1840 Ethylene glycol diacetate
1870 Ethylene glycol dimethyl ether
1890 Ethylene glycol monobutyl ether
1900 Ethylene glycol monobutyl ether acetate
1910 Ethylene glycol monoethyl ether
1920 Ethylene glycol monoethyl ether acetate
1930 Ethylene glycol monomethyl ether
1940 Ethylene glycol monomethyl ether acetate
1960 Ethylene glycol monophenyl ether
1970 Ethylene glyco! monopropyl ether
1980 Ethylene oxide
1990 Ethyl ether
2000 2-ethylhexanol
2010 Ethyl orthoformate
2020 Ethyl oxalate
2030 Ethyl sodium oxalacetate
2040 Formaldehyde
2050 Formamlde
2060 Formic add
2070 Fumarlc add
2073 Furfural
2090 Glycerol (Synthetic)
2091 Glycerol dlchlorohydrln
2100 Glycerol trlether
2110 Glydne
2120 Glyoxal
2145 Hexachlorobenzene
2150 Hexachloroethane
2160 Hexadecyl alcohol
2165 Hexamethylened1am1ne
2170 Hexanethylene glycol
2180 Hexamethylenetetramlne
2190 Hydrogen cyanide
2200 Hydroqulnane
2210 p-hydroxybenzole add
2240 Isoamylene
2250 Isobutanol
2260 Isobutyl acetate
2261 Isobutylene
2270 Isobutyraldehyde
2280 Isobutyrlc add n_*
OCPDB No. Chemical
2300 Isodecanol
2320 Isooctyl alcohol
2321 Isopentane
2330 Isophorone
2340 Isophthallc add
2350 Isoprene
2360 Isopropanol
2370 Isopropyl acetate
2380 Isopropylamine
2390 Isopropyl chloride
2400 Isopropylphenol
2410 Ketene
2414 Linear alkyl sulfonate
2417 Linear alkylbenzene
2420 Haleic add
2430 Haleic anhydride
2440 Halle add
2450 Mesltyl oxide
2455 HetanlUc add
2460 Hethacrylic add
2490 Methallyl chloride
2500 Methanol
2510 Methyl acetate
2520 Methyl acetoacetate
2530 Hethylamlne
2540 n-methylannine
2545 Methyl bromide
2550 Methyl butynol
2560 Methyl chloride
2570 Methyl cyclohexane
2590 Methyl cyclohexanone
2620 Methylene chloride
2530 Methylene d1an111ne
2635 Methylene dlphenyl d11socyanate
2640 Methyl ethyl ketone
2644 Methyl formate
2650 Methyl Isobutyl carblnol
2660 Methyl Isobutyl ketone
2665 Methyl methacrylate
2670 Methyl pentynol
2690 a -methylstyrene
-------
OCPDB No. Chemical
2700 Morpholine
2710 a-naphthalene sulfonlc add
2720 a-n»phthalene sulfonic add
2730 a-naphthol
2740 0-naphthol
2750 Neopentanolc add
2756 o-nitroaniline
2757 p-nitroaniline
2760 o-nitroanisole
2762 p-n1troan1sole
2770 Nitrobenzene
2780 N1trobenzo1c add (o, m, and p)
2790 NUroethane
2791 Nltromethane
2792 Nltrophenol
2795 Nltropropane
2800 Nitrotoleune
2810 Nonene
2820 Nonyl phenol
2830 Octyl phenol
2840 Paraldehyde
2850 Pentaerythrltol
2851 n-pentane
2855 1-pentene
2860 Perchloroethylene
2882 Perchloromethyl mercaptan
2890 o-phenetidine
2900 p-phenet1dine
2910 Phenol
2920 Phenolsulfonic acids
2930 Phenyl anthranilic acid
2940 Phenylenedlamine
2950 Phosgene
2960 Phthalic anhydride
2970 Phthalimide
2973 s-picoline
2976 Piperazine
OCPDB No. Chemical
3000 Polybutenes
3010 Polyethylene glycol
3025 Polypropylene glycol
3063 Proplonaldehyde
3066 Prop1on1c add
3070 n-propyl alcohol
3075 Propylamine
3080 Propyl chloride
3090 Propylene
3100 Propylene chlorohydrln
3110 Propylene dichlorlde
3111 Propylene glycol
3120 Propylene oxide
3130 Pyrldine
3140 Quinone
3150 Resorcinol
3160 Resorcylic acid
3170 Salicylic acid
3180 Sodium acetate
3181 Sodium benzoate
3190 Sodium carboxymethyl cellulose
3191 Sodium chloroacetate
3200 Sodium formate
3210 Sodium phenate
3220 Sorbic acid
3230 Styrene
3240 Succinic add
3250 Sucdnltrile
3251 Sulfanilic acid
3260 Sulfolane
3270 Tannic acid
3280 Terephthalic acid
3290 & 3291 Tetrachloroethanes
3300 Tetrachlorophthalic anhydride
3310 Tetraethyllead
3320 Tetrahydronaphthalene
3330 Tetrahydrophthalic anhydride
3335 Tetramethyllead
B-5
-------
OCPDB No. Chemical
3340 Tetrainethylened1am1ne
3341 Tetramethylethylenedlamine Table II: Polymer and
3349 Toluene Manufacturing Industry
3350 Toluene-2,4-d1am1ne
3354 Toluene-2,4-d11socyanate Polyethylene
3355 Toluene dUsocyanates (mixture) Polypropylene
3360 Toluene sulfonoride Polystyrene
3370 Toluene sulfonlc adds Styrene-butadlene latex
3380 Toluene sulfonyl chloride
3381 Toluldlnes
3390, 3391 Trlchlorobenzenes
& 3393
3395 1,1,1-tHchloroethane
3400 1,1,2-trichloroethane
3410 Trichloroethylene
3411 Trlchlorofluoromethane
3420 1,2,3-trlchloropropane
3430 I,l,2-tr1chloro-l,2,2-tr1f1uoroethane
3450 Tr1ethylamine
3460 THethylene glycol
3470 THethylene gylcol dimethyl ether
3480 TrUsobutylene
3490 THmethylamlne
3510 Vinyl acetate
3520 Vinyl chloride
3530 V1nyl1dene chloride
3540 Vinyl toluene
3541 Xylene (mixed)
3560 o-xylene
3570 p-xylene
3580 Xylenol
3590 Xyl1d1ne
B-6
-------
APPENDIX C. METHOD 21. DETERMINATION OF VOLATILE ORGANIC
COMPOUND LEAKS
-------
APPENDIX C
METHOD 21. DETERMINATION OF VOLATILE
ORGANIC COMPOUND LEAKS
1. Applicability and Principle
1.1 Applicability. This method applies to the determination of volatile
organic compound (VOC) leaks from organic process equipment. These sources
include, but are not limited to, valves, flanges and other connections,
pumps and compressors, pressure relief devices, process drains, open-ended
valves, pump and compressor seal system degassing vents, accumulator vessel
vents, and access door seals.
1.2 Principle. A portable instrument is used to detect VOC leaks from
individual sources. The instrument detector is not specified, but it must
meet the specifications and performance criteria contained in paragraph 2.1.
2. Apparatus
2.1 Monitoring Instrument. The monitoring instrument shall be as follows:
2.1.1 Specifications.
a. The VOC instrument detector shall respond to the organic compounds
being processed. Detectors which may meet this requirement include, but
are not limited to, catalytic oxidation, flame ionization, infrared absorption,
and photoionization.
b. The instrument shall be intrinsically safe for operation in explosive
atmospheres as defined by the applicable U.S.A. Standards (e.g., National
Electrical Code by the National Fire Prevention Association).
C-l
-------
c. The instrument shall be able to measure the leak definition
concentration specified in the regulation.
d. The instrument shall be equipped with a pump so that a continuous
sample is provided to the detector. The nominal sample flow rate shall be
1-3 liters per minute.
e. The scale of the instrument meter shall be readable to ±5 percent
of the specified leak definition concentration.
2.1.2 Performance Criteria. The instrument must meet the following
performance criteria. The definitions and evaluation procedures for each
parameter are given in Section 4.
2.1.2.1 Response Time. The instrument response time must be 30 seconds
or less. The response time must be determined for the instrument system
configuration to be used during testing, including dilution equipment. The
use of a system with a shorter response time than that specified will reduce
the time required for field component surveys.
2.1.2.2 Calibration Precision. The calibration precision must be less
than or equal to 10 percent of the calibration gas value.
2.1.2.3 Quality Assurance. The instrument shall be subjected to the
response time and calibration precision tests prior to being placed in
service. The calibration precision test shall be repeated every 6 months
thereafter. If any modification or replacement of the instrument detector
is required, the instrument shall be retested and a new 6-month quality
assurance test schedule will apply. The response time test shall be repeated
if any modifications to the sample pumping system or flow configuration is
made that would change the response time.
C-2
-------
2.3 Calibration Gases. The monitoring instrument is calibrated in terms of
parts per million by volume (ppmv) of the compound specified in the applicable
regulation. The calibration gases required for monitoring and instrument
performance evaluation are a zero gas (air, <3 ppmv VOC) and a calibration
gas in air mixture approximately equal to the leak definition specified in
the regulation. If cylinder calibration gas mixtures are used, they must be
analyzed and certified by the manufacturer to be within ±2 percent accuracy.
Calibration gases may be prepared by the user according to any accepted gaseous
standards preparation procedure that will yield a mixture accurate to within
±2 percent. Alternative calibration gas species may be used in place of the
calibration compound if a relative response factor for each instrument is
determined so that calibrations with the alternative species may be expressed
as calibration compound equivalents on the meter readout.
3. Procedures
3.1 Calibration. Assemble and start up the VOC analyzer and recorder
according to the manufacturer's instructions. After the appropriate warm-up
period and zero or internal calibration procedure, introduce the calibration
gas into the instrument sample probe. Adjust the instrument meter readout
to correspond to the calibration gas value. If a dilution apparatus is
used, calibration must include the instrument and dilution apparatus assembly.
The nominal dilution factor may be used to establish a scale factor for
converting to an undiluted basis. For example if a nominal 10:1 dilution
apparatus is used, the meter reading for a 10,000 ppm calibration compound would
be set at 1000. During field surveys, the scale factor of 10 would be used
to convert measurements to an undiluted basis.
C-3
-------
3.2 Individual Source Surveys.
3.2.1 Case I - Leak Definition Based on Concentration Value. Place the
probe inlet at the surface of the component interface where leakage could
occur. Move the probe along the interface periphery while observing the
instrument readout. If an increased meter reading is observed, slowly probe
the interface where leakage is indicated until the maximum meter reading is
obtained. Leave the probe inlet at this maximum reading location for
approximately two times the instrument response time. If the maximum observed
meter reading is greater than the leak definition in the applicable regulation,
record and report the results as specified in the regulation reporting
requirements. Examples of the application of this general technique to
specific equipment types are:
a. Valves—The most common source of leaks from valves is at the seal
between the stem and housing. Place the probe at the interface where the
stem exits the packing gland and sample the stem circumference. Also, place
the probe at the interface of the packing gland take-up flange seat and sample
the periphery. In addition, survey valve housings of multipart assembly at
the surface of all interfaces where leaks can occur.
b. Flanges and Other Connections—For welded flanges, place the probe
at the outer edge of the flange-gasket interface and sample around the
circumference of the flange. Sample other types of nonpermanent joints (such
as threaded connections) with a similar traverse.
c. Pumps and Compressors—Conduct a circumferential traverse at the
outer surface of the pump or compressor shaft and seal interface. If the
source is a rotating shaft, position the probe inlet within one centimeter
of the shaft seal interface for the survey. If the housing configuration
C-4
-------
prevents a complete traverse of the shaft periphery, sample all accessible
portions. Sample all other joints on the pump or compressor housing where
leakage can occur.
d. Pressure Relief Devices—The configuration of most pressure relief
devices prevents sampling at the sealing seat interface. For those devices
equipped with an enclosed extension, or horn, place the probe inlet at
approximately the center of the exhaust area to the atmosphere for sampling.
e. Process Drains—For open drains, place the probe inlet at
approximately the center of the area open to the atmosphere for sampling. For
covered drains, place the probe at the surface of the cover interface and
conduct a peripheral traverse.
f. Open-Ended Lines or Valves—Place the probe inlet at approximately
the center of the opening to the atmosphere for sampling.
g. Seal System Degassing Vents and Accumulator Vents—Place the probe
inlet at approximately the center of the opening to the atmosphere for
sampling.
h. Access Door Seals—Place the probe inlet at the surface of the
door seal interface and conduct a peripheral traverse.
3.2.2 Case II-Leak Definition Based on "No Detectable Emission".
a. Determine the local ambient concentration around the source by
moving the probe inlet randomly upwind and downwind at distance of one to
two meters from the source. If an interference exists with this determination
due to a nearby emission or leak, the local ambient concentration may be
determined at distances closer to the source, but in no case shall the distance
be less than 25 centimeters. Note the ambient concentration and then move
the probe inlet to the surface of the source and conduct a survey as described
C-5
-------
in 3.2.1. If a concentration increase greater than 2 percent of the
concentration-based leak definition is obtained, record and report the results
as specified by the regulation.
b. For those cases where the regulation requires a specific device
installation, or that specified vents be ducted or piped to a control device,
the existence of these conditions shall be visually confirmed. When the
regulation also requires that no detectable emissions exist, visual observations
and sampling surveys are required. Examples of this technique are:
i. Pump or Compressor Seals—If applicable, determine the type of
shaft seal. Perform a survey of the local area ambient VOC concentration
and determine if detectable emissions exist as described in 3.2.2.a.
ii. Seal System Degassing Vents, Accumulator Vessel Vents, Pressure
Relief Devices—If applicable, observe whether or not the applicable ducting
or piping exists. Also, determine if any sources exist in the ducting or
piping where emissions could occur prior to the control device. If the
required ducting or piping exists and there are no sources where the emissions
could be vented to the atmosphere prior to the control device, then it is
presumed that no detectable emissions are present.
4. Instrument Performance Evaluation Procedures
4.1 Definitions.
4.1.1 Calibration Precision. The difference between the average VOC
concentration indicated by the meter readout for consecutive calibration
repetitions and the known concentration of a test gas mixture.
4.1.2 Response Time. The time interval from a step change in VOC
concentration at the input of the sampling system to the time at which
90 percent of the corresponding final value is reached as displayed on the
instrument readout meter.
C-6
-------
4.2 Evaluation Procedures. At the beginning of the instrument performance
evaluation test, assemble and start up the instrument according to the
manufacturer's instructions for recommended warmup period and preliminary
adjustments. If a dilution apparatus is used during field surveys, the
evaluation procedure must be performed on the instrument-dilution system
combination.
4.2.1 Calibration Precision Test. Make a total of nine measurements
by alternately using zero gas and the specified calibration gas. Record the
meter readings (example data sheet shown in Figure 21-1).
4.2.2 Response Time Test Procedure. Introduce zero gas into the
instrument sample probe. When the meter reading has estabilized, switch
quickly to the specified calibration gas. Measure the time from concentration
switching to 90 percent of final stable reading. Perform this test sequence
three times and record the results (example data sheet given in Figure 21-2).
4.3 Calculations. All results are expressed as mean values, calculated
by:
7 = n 1?1 *1
where:
x. = Value of thetndividual measurements
= Sum of the individual values.
x" = Mean value.
n = Number of measurements.
C-7
-------
Instrument ID
Calibration Gas Data
Calibration = ppmv
Run Instrument Meter Difference^ '
No. Reading, ppm ppm
1.
2.
3.
4.
5.
6.
7.
8.
9.
Mean Difference
Calibration Precision = * 10°
(1)Calibration Gas Concentration - Instrument Reading
Figure C-l. Calibration Precision Determination
C-8
-------
Instrument ID
Calibration Gas Concentration
90% Response Time:
1. Seconds
2. Seconds
3. Seconds
Mean Response Time Seconds
Figure C-2. Response Time Determination
C-9
-------
APPENDIX D. EXAMPLE CALCULATIONS FOR DETERMINING REDUCTION IN
EMISSIONS FROM IMPLEMENTATION OF RACT
-------
APPENDIX D
Example Calculations for Determining Reduction in Emissions from
Implementation of RACT
The purpose of this appendix is to provide emission factors and procedures
for computing the emission reduction associated with reasonably available
control technology (RACT) for control of fugitive emissions in the synthetic
organic chemical, polymer, and resin manufacturing industries. The emission
factors and procedures allow the reader to compute total process unit uncontrolled
emissions, uncontrolled emissions from equipment affected by RACT, controlled
emissions from equipment affected by RACT, and the emission reduction from
implementing RACT. The emission factors and computation procedures in this
appendix should be used only where specific plant fugitive emission data are
not available.
Uncontrolled emission factors for each type of component are presented
in Table D-l. The uncontrolled emissions from a process unit can be computed
by multiplying the uncontrolled emission factor for each component type by
the number of components of the given type, then summing the emissions from
all component types.
The average emissions from components controlled under RACT (pumps in
light liquid service, compressors, valves in gas service and light liquid
service, and pressure relief valves in gas service) following implementation
of RACT are presented in Table D-2. The basis for these factors has already
been discussed in Chapters 3 and 4.
The number of components in a process unit may be estimated (at the
Director's discretion) or derived from actual count. A sample calculation
of uncontrolled emissions from an illustrative process unit is presented
below in Table D-3.
D-l
-------
TABLE D-l. UNCONTROLLED FUGITIVE EMISSION FACTORS
IN PROCESS UNIT EOUIPMENTa
Uncontrolled
Fugitive emission factor,
emission source kg/hr
Pumps
Light liquids 0.12
Heavy liquids 0.020
Valves (in-line)
Gas service 0.021
Light liquid service 0.010
Heavy liquid service 0.0003
Pressure relief valves
Gas service 0.16
Light liquid service 0.006
Heavy liquid service 0.009
Open-ended valves
Gas service 0.025
Light liquid service 0.014
Heavy liquid service 0.003
.Flanges 0.0003
Sampling connections 0.015
Compressors 0.44
aFrom Table 2-2.
TABLE D-2. CONTROLLED EMISSION FACTORS FOR EQUIPMENT
AFFECTED BY RACT
Controlled
Fugitive emission factor,
emission source kg/hr
Pumps in light liquid service 0.036
Valve in gas service 0.003
Valve in light liquid service 0.003
Gas pressure relief valves 0.066
Compressors . 0.123
aFrom Table 4-3.
D-2
-------
TABLE D-3. EXAMPLE CALCULATION OF UNCONTROLLED EMISSIONS FROM
AN ILLUSTRATIVE PROCESS UNIT
Fugitive emission source
Pumps
Light liquid service0
Heavy liquid service
In-line valves
Gas service
Light liquid service
Heavy liquid service
Pressure relief valves
Gas service
Light liquid service
Heavy liquid service
Open-ended valves
Gas service
Light liquid service
Heavy liquid service
Compressors0
Sampling connections
Flanges
Number
of sources
(N)
16
10
•
15°H
120d
100d
17
2,
3
12
19
17
2
34
750d
Emission
factor ,
kg/hr-source
(E)
0.12
0.020
0.021
0.010
0.0003
0.16
0.006
0.009
0.025
0.014
0.003
0.44
0.015
0.0003
Total emissions
Emissions from
source, kg/hr
(N X E)
1.92
0.2
3.15
1.2
0.03
2.72
0.012
0.027
0.3
0.266
0.051
0.88
0.51
0.225
= 11.49 kg/hre
Determined by actual count or by estimate.
bFrom Table D-l.
Inspected under RACT.
Because of the large number of components, an estimate may be appropriate.
eThe expected annual emissions are 100.65 Mg/yr (8,760 hr/yr).
D-3
-------
The overall emission reduction associated with RACT can be computed by
finding the emission reduction for those process units maintained under
RACT. The emission reduction is the difference between the emissions expected
from components in the unrepaired (uncontrolled) condition and in the repaired
(controlled) condition. The uncontrolled emissions are presented in
Table D-4:
TABLE D-4. UNCONTROLLED EMISSIONS FROM COMPONENTS AFFECTED BY RACT.
Emission source
Light liquid pump
Gas service valve
Light liquid service valve
Gas pressure relief valve
Compressors
Number of
sources
(N)
16
150
120
17
2
Emission
factor
kg/hr-source
(E)
0.12
0.021
0.010
0.16
0.44
Emissions from
sources, kg/hr
(N X E)
1.92
3.15
1.2
2.72
0.88
9.87
The emissions expected after implementing RACT are also presented in
Table D-5:
TABLE D-5. CONTROLLED EMISSIONS FROM COMPONENTS AFFECTED BY RACT.
Emission source
Light liquid pump
Gas service valve
Light liquid service valve
Gas pressure relief valve
Compressors
Number
of sources
(N)
16
150
120
17
2
Emission
factor
kg/hr-source
(E)
0.036
0.003
0.003
0.066
0.123
Emissions, from
sources, kg/hr
(N X E)
0.576
0.45
0.36
1.12
0.246
2.75
From Table D-2.
D-4
-------
The difference between emissions before and after the implementation of
RACT can be expressed as an annual emission reduction. This emission
reduction is presented in Table D-6:
TABLE D-6. EMISSION REDUCTION EXPECTED FROM RACT
1. Uncontrolled emissions from process unit 100.65 Mg/yr
(11.49 kg/hr)
2. Uncontrolled emissions from components affected 86.46 Mg/yr
by RACT (9.87 kg/hr)
3. Controlled emissions from components 24.09 Mg/yr
affected by RACT (2.75 kg/hr)
4. Total emission reduction:
86.46 Mg/yr - 24.09 Mg/yr = 62.35 Mg/yr (7.12 kg/hr)
5. Percent emission reduction:
62.35 Mg/yr inn _
100.65 Mg/yr x 10° ~
D-5
------- |