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

-------