&EPA
           United States
           Environmental Protection
           Agency
          Office of Air Quality
          Planning and Standards
          Research Triangle Park NC 27711
EPA-45O/3-83-OO6
March 1984
           Air
Guideline Series

Control of  Volatile
Organic
Compound Leaks
from Synthetic
Organic Chemical
and Polymer
Manufacturing
Equipment

-------

-------
                                 EPA-450/3-83-006
              Guideline Series

 Control of Volatile Organic Compound
Leaks from  Synthetic Organic Chemical
and Polymer Manufacturing Equipment
             Emission Standards and Engineering Division
            U.S. ENVIRONMENTAL PROTECTION AGENCY
                 Office of Air and Radiation
             Office of Air Quality Planning and Standards
            Research Triangle Park, North Carolina 27711

                    March 1984

-------
                                       GUIDELINE SERIES,,,,

The guideline series of reports is issued by the Office of Air Quality Planning and Standards (OAQPS) to
provide information to state and local air pollution control agencies; for example, to provide guidance on the
acquisition and  processing  of air quality data and on  the planning and analysis requisite for the
maintenance of air quality. Reports published in this series will be available—as supplies permit—from the
Library Services Office (MD-35), U.S. Environmental Protection Agency, Research Triangle Park, North
Carolina 27711, or for a nominal fee, from the National Technical Information Service, 5285 Port Royal
Road, Springfield, Virginia 22161.

-------
                        ,,  .   TABLE OF. CONTENTS


                                                                      Page

LIST OF -TABLES		    vi

LIST OF .FIGURES	    viii

Chapter 1.0 - INTRODUCTION	    1-1

Chapter 2.0 - PROCESSES AND POLLUTANT EMISSIONS	    2-1

     2.1  INTRODUCTION .	

     2.2  FACILITIES AND THEIR EMISSIONS	    2-3
          2.2.1  Potential Source Characterization and
                   Description	,	    2-3 -
                 2.2.1.1  Pumps	    2-3
                 2.2.1.2  Compressors. . .-	    2-8
                 2.2.1.3  Process Valves	  .    2-11
                 2.2.1.4  Safety Relief Devices. .........    -2-14
                 2.2.1.5  Agitators. ..-	    2-15
                 2.2.1.6  Open-Ended Lines  ............    2-17
                 2.2.1.7  Sampling Connections 	    2-17
                 2.2.1.8  Flanges	 . . ,	    2-17

     2.3  MODEL UNITS	  .    2-17
          2.3.1  Model Units	    2-17
                 2.3.1.1  Sources of Fugitive Emissions	'  .    2-18
                 2.3.1.2  Model Units Components 	    2-18
                 2.3.1.3  Uncontrolled Fugitive Emission
                            Estimates. . .  . . . .	    2-20

     2.4  REFERENCES	    2-22

Chapter 3.0 - EMISSION CONTROL TECHNIQUES	    3-1

     3.1  PRIMARY CONTROL METHODS. ...-..-	    3-1
          3.1.1  Individual Component Survey for Leak Detection,  .    3-1
          3.1.2  Repair Methods.  .	    3-2
                 3.1.2.1  Pumps.  .	-  .    3-2
                 3.1.2.2  Compressors.	-    3-2
                 3.1.2.3  Safety/Relief Valves	    3-3
                 3.1.2.4  Valves-	 . .••	  .    3-4
                 3.1.2.5  Flanges	    3-4
          3.1.3  Control Effectiveness of Leak Detection and
                   Repair Techniques .... 	    3-4
          3.1.4  Open-Ended Lines	    3-7
                                      n i

-------
                        TABLE OF CONTENTS (CONTINUED)
                                                                      Page

     3.2  OTHER CONTROL STRATEGIES 	 .     3-9
          3,2.1  General	.     3-10
          3.2,2  Allowable Percentage of Valves Leaking. . . ...     3-12
          3.2.3  Alternative Work Practice for Valves	     3-13

     3.3  OTHER CONSIDERATIONS 	 ......     3-16
          3.3.1  Unsafe and Difficult to Reach Components. ....     3-16
          3.3.2  Small Process Unit	     3-16
          3.3.3  Unit Turnarounds	     3-17

     3.4  REFERENCES		     3-20

Chapter 4.0 - ENVIRONMENTAL ANALYSIS OF RACT	•''••.•     4-1

     4.1  REASONABLY AVAILABLE CONTROL TECHNOLOTY. (RACT) .....     4-1

     4.2  AIR POLLUTION	     4-3
          4.2.1  Development of VOC Emission Levels	     4-3
          4.2.2  VOC Emission Reduction	     4-3

     4.3  WATER POLLUTION	 . ;  . ,.     4-3

     4.4  SOLID WASTE DISPOSAL	     4-6

     4.5  ENERGY	     4-6

     4.6  REFERENCES	     4-7

Chapter 5.0 - CONTROL COST ANALYSIS OF RACT	     5-1

     5.1  BASIS FOR CAPITAL COST	 .     5-1
          5.1.1  Cost of Monitoring Instrument ..........     5-1
          5.1.2  Caps on Open-Ended Lines	     5-1
          5.1.3  Initial Leak Repair	     5-4
          5.1.4  Replacement Pump Seals at Initial  Repair, ....     5-4

     5.2  BASIS FOR ANNUALIZED COSTS 	 ......     5-6
          5.2.1  Monitoring Labor	     5-6
          5.2.2  Leak Repair Labor 	  ........     5-6
          5.2.3  Maintenance Charges and Miscellaneous Costs .  . .     5-6
          5.2.4  Administrative Costs.  .	.  , .     5-9
          5.2.5  Capital Charges	     5-9
          5.2.6  Recovery Credits.  .	     5-9

-------
                        TABLE OF CONTENTS (CONTINUED)
                                                                      Paqe
     5.3  EMISSION CONTROL COSTS ..... 	    5-11
          5.3.1  Initial Costs	;	•	    5-11
          5.3.2  Recovery Credits.	    5-11
          5.3.3  Net Annuallzed Cost	    5-11
          5.3.4  Differences in Net Annualized Costs 	    5-13

     5.4'  COST EFFECTIVENESS	    5-13

     5.5  REFERENCES .......... j .  . . .  	 ...    5-17

APPENDIX A - MAJOR COMMENTS RECEIVED ON THE DRAFT CTG	    A-l

     A.I  NEED AND COVERAGE OF THE CTG	 .    A-l

     A.2  ESTIMATES OF EMISSIONS, EMISSION REDUCTIONS,
            AND COSTS	;.....	    A-6

     A.3  RACT SELECTION, PROVISIONS, AND EXEMPTIONS .......    A-16

     A.4  REFERENCES . ,	    A-27

APPENDIX B - COMMENT LETTERS RECEIVED ON DRAFT CONTROL TECHNIQUES
             GUIDELINE DOCUMENT	    B-l

-------
                               LIST OF TABLES
Table

 2-1      EQUIPMENT COUNTS FOR FUGITIVE VOC EMISSION SOURCES
            IN SOCMI MODEL UNITS	     2-19

 2-2      EMISSION FACTORS FOR LEAKS FROM PROCESS EQUIPMENT. .  .  .     2-21

 3-1      INPUT PARAMETERS FOR LDAR MODEL	     3-6

 3-2      EFFECTIVENESS OF QUARTERLY LEAK DETECTION AND REPAIR .  .     3-8

 3-3      ILLUSTRATION OF A SKIP-PERIOD MONITORING PROGRAM ....     3-15

 3-4      COST EFFECTIVENESS FOR QUARTERLY LEAK DETECTION AND
            REPAIR PROGRAMS FOR PROCESS UNITS PROCESSING SMALL
            VOLUMES OF LIGHT LIQUID AND GASEOUS VOC	     3-18

 4-1      ESTIMATED EMISSIONS AND EMISSIONS REDUCTION ON A
            MODEL UNIT BASIS	     4-4

 4-2      EMISSION FACTORS FOR SOURCES CONTROLLED UNDER RACT ...     4-5

 5-1      CAPITAL COST DATA	     5-2

 5-2      CAPITAL COST ESTIMATES FOR IMPLEMENTING RACT (Thousands
            of June 1980 Dollars	     5-3

 5-3      LABOR-HOUR REQUIREMENTS FOR INITIAL LEAK REPAIR UNDER
            RACT	  .  .     5-5

 5-4      BASIS FOR ANNUALIZED COST ESTIMATES	     5-7

 5-5      ANNUAL MONITORING AND LEAK REPAIR LABOR REQUIREMENTS
            FOR RACT	.  .     5-8

 5-6      RECOVERY CREDITS	!  .  .     5-10

 5-7      ANNUALIZED CONTROL COST ESTIMATES FOR MODEL UNITS UNDER
            RACT (Thousands of June 1980 Dollars)	  .     5-12

 5-8      COST EFFECTIVENESS FOR MODEL UNITS UNDER RACT.  .....     5-14

 5-9      COST EFFECTIVENESS FOR COMPONENT TYPES IN MODEL
            UNIT B	  .     5-16

 A-l      ESTIMATES OF VOC EMISSIONS FROM SOCMI	'. .  .     A-3

-------
                         LIST OF TABLES (CONTINUED)


Table                                                                 Page

 A-2      COSTS FOR COMPONENT TYPES IN MODEL UNIT,B.  .......     A-13

 A-3      EQUIPMENT LEAKS OF VOC FROM SYNTHETIC ORGANIC CHEMICAL
            POL'.MER MANUFACTURING:  RACT	     A-15

 A-4.     SUMMARY OF PERCENT OF SOURCES DISTRIBUTION  CURVES AND
            PERCENT OF MASS EMISSIONS CURVES AT VARIOUS ACTION
            LEVELS	:	     A-20

 B-2      LIST OF COMMENTERS AND AFFILIATIONS	     B-2
                                     vn

-------
LIST OF FIGURES
Figure
2-1

2-2
2-3
2-4

2-5

2-6
2-7
2-8
2-9
2-10
2-11
2-12
2-13
2-14
2-15
2-16
3-1



General Schematic of Process Levels that Make Up
the Organic Chemical Industry 	
Diagram of a Simple Packed Seal ..... 	 , , j
Diagram of a Basic Single Mechanical Seal 	
Diagram of a Double Mechanical Seal (Back-to-Back
Arrangement). 	 .....
Diagram of a Double Mechanical Seal (Tandem
Arrangement) 	
Diaphragm Pump 	 	 .
Labyrinth Shaft Seal 	 	
Restrictive-Ring Shaft Seal 	
Mechanical (Contact) Shaft Seal 	 .
Liquid Film Shaft Seal With Cylindrical Bushing . .; . .
Diagram of a Gate Valve 	
Example of Bellows Seals 	 	 .
Diagrams of Valves With Diaphragm Seals ........
Diagram of a Spring-Loaded Relief Valve 	
Diagram of a Hydraulic Seal for Agitators 	
Diagram of Agitator Lip Seal 	 ' . .
Cost Effectiveness of Quarterly Leak Detection and
Repair of Valves With Varying Leak Frequency - SOCMI
Units 	 	 	 - . .
Page

2-2
2-4
2-5'

2-6

2-6
2-7
2-8
2-9
2-10
2-10
2-11
2-12
2-13
2-14
2-16
2-16


3-11
     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 regulations on a
category of stationary sources of volatile organic compounds (VOC) until
after EPA published a control techniques guideline (CTG) for that VOC source
category.  See also 44 FR 53761 (September 17, 1979).  This delay allowed
the states to make more technically sound decisions regarding the applica-
tion 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 operated to produce one or more of the synthetic organic
chemicals listed in Appendix E of the proposed standards of performance for
SOCMI (46 FR 1136, January 5, 1981),'methyl  tert-butyl ether (MTBE),
polyethylene, polypropylene, and polystyrene.  The equipment in process
units in the synthetic organic chemical  manufacturing industry (SOCMI) is
similar to equipment in the polymer 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 plants.
     The SOCMI is a segment of the chemical  industry consisting of some of
the higher volume intermediate and finished  products.  The polymer
manufacturing industries to which the discussion in this document applies
are polyethylene, polypropylene, and polystyrene.  It should be emphasized
that the discussion in this document is  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 mega-
grams (Mg) annually.  Another 12 percent have production capacities in
excess of 500,000 Mg.
                                     2-1

-------
                          RAW MATERIALS
               (CRUDE OIL, CRUDE NATURAL GAS,  ETC.)
                   REFINERIES    |
                                 I
                          CHEMICAL
                          FEEDSTOCK
                           PLANTS
                                                    CHEMICAL
                                                   FEEDSTOCKS
                                                    CHEMICAL
                                                     PLANTS
                                                      o
                                                    CHEMICAL
                                                    PRODUCTS
Figure 2-1.
General  schematic of process levels that make up
the organic chemical industry.            ;
                             2-2

-------
     The polymer 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 polyethylene, polypropylene, and polystyrene.
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, recipro-
cating and rotary action, and special canned and diaphragm pumps, are also
used.  Chemicals transferred by pumps carv 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
                                             2
that flows between the packing and the shaft.   Deterioration of the
packing will result in process liquid leaks.
                                     2-3

-------
                      Fluid
                      End
                                  Stuffing
                                  Box
                                            ^n
    Packing
    Gland
 ^ •""•^    Atmosphere
IZI0    End
                                      ^
                                   Packing
 X Possible
    Leak
    Area
                  Figure 2-2.  Diagram of a simple packed seal.'
     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 sealing device such as packing can be employed.
                                      2-4

-------
                     PUMP
                   STUFFING
                     BOX
                                                     GLAND
                                                      RING
                 FLUID
                  END
                                                         STATIONARY
                                                          ELEMENT

                                                         POSSIBLE
                                                         LEAK AREA
                          SHAFT
                                               ROTATING
                                              SEAL RING
           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-tp-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 than  that  in the stuffing box.  Therefore, any  leakage will  be  from
                                      2-5

-------
POSSIBLE LEAK
 INTO SEALING
    FLUID
       FUIID END
                                          SEAL LIQUID.
                         PRIMARY
                           SEAL
                                v
                                      v
                                          SECONDARY
                                            SEAL
                                                                   — GLAND
                                                                      PLATE
Figure 2-4.
                         Diagram of  a double mechanical seal
                         (back-to-back arrangement)'
            FLUID
             END
                PRIMARY -
                 SEAL
                                   BUFFER LIQUID
                                    OUT   IN
                                   (TOP) (BOTTOM)
                                           I

                     _liJBryy_T ' r>>c*at..>rao j. ll-^rik K*.



                     -\	-/—SHAFT-A
                                          V
                                   SECONDARY
                                     SEAL
                                                            GLAND
                                                            PLATE
                                                           70-17B7.1
           Figure  2-5.  Diagram of  a double mechanical seal
                         (tandem arrangement)8
                                    2-6

-------
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 does not leak from the reservoir, the
                                            g
reservoir can be vented to a control device.
     Another type of pump that has been used is the shaftless 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.
     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.
                      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 unit:; 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 reduce  leakage by  up  to 40 percent  over a  straight pass  type
labyrinth.13
                      PORT MAY BE ADDED
                      FOR SCAVENGING OR
                      INERT-GAS SEAUNG
                      INTERNAL
                      GAS PRESSURE

ATMOSPHERE
                     Figure 2-7.  Labyrinth  shaft  seal.14
                                     2-8

-------
     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.  Rpstrictive  ring seals can achieve lower leak rates
than the labyrinth.    A  restrictive ring seal  is shown in Figure 2-8.

                                        SCAVENGING
                                        PORT MAY Si '
                                      .._ ADDED FOR
                                      •'£ VACUUM
                                        APPLICATION
                                             ATMOSPHERE
                Figure  2-8.   Restrictive-ring shaft seal.
                                                         15
     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
                                            19
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
                           20
released to the atmosphere.     To  eliminate release  of VOC emissions from
the seal oil system, the  reservoir can  be  vented to  a  control device.
                                     2-9

-------
                       INTERNAL
                       GAS PRESSURE
CLEAN OIL IN

   PRESSURE
 /• BREAKDOWN
 I  SLEEVE
                                                ATMOSPHERE
                                   CONTAMINATED
                                   OIL OUT
           Figure 2-9.   Mechanical  (contact) shaft seal.
                                                                17
                                          CLEAN OIL IN
                                                  ATMOSPHERE
                           CONTAMINATED
                           OIL OUT
                                          OIL OUT
Figure 2-10.  Liquid  film  shaft  seal  with  cylindrical  bushing.18
                                     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 th^ 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.  Valves 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'tetrafluoroethylene.  The packing material used
                                                   21
depends on the valve application and configuration.    These conventional
packing glands can be used over a wide range of operating temperatures.  At
                                                                      ??
high pressures these glands must be quite tight to attain a good seal.
                       PACKING
                       GLAND
                      PACKING
                        VALVE
                        STEM
                                                   POSSIBLE
                                                   LEAK AREAS
                 Figure 2-11.  Diagram of a gate valve.
                                                       23
                                    2-11

-------
     Elastomeric 0-rings are also used for sealing process valves.  These
Q-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
         24
material.
Bellows
                   are more effective for preventing process fluid leaks
                                                                        25
than the conventional packing gland or any other gland-seal arrangement.
This type of seal incorporates a formed metal bellows that makes a barrier
between the disk 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
                                        OC
with a leak detector in case of failure.
                   BELLOWS
                                                  BODY
                                                  BONNET
               Figure 2-12.  Example of bellows seals.'
                                                      27
                                     2-12

-------
     A diaphragm may be used to isolate the working parts of the valve and
the environment from the process liquid.  Two 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
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 seals 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
                               28
a source of fugitive emissions.
          DIAPHRAGM
              DISK
                                                              STEM
                                                              DIAPHRAGM
             Figure 2-13.  Diagrams of valves with diaphragm seals.29
                                     2-13

-------
     2.2.1.4  Safety 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, safety 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
                                                                      30
pressure 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 potential causes of leakage from safety 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.
                        Possible
                        Leak Area
                                    Process Side
           Figure  2-14.   Diagram of  a  spring-loaded relief valve.
                                     2-14

-------
     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 to escape from the system
under normal operation.  However, when the disk does rupture, the system
depressurizes ur.til atmospheric conditions are obtained.  This could result
in an excessive loss of product or a corresponding excessive release of
fugitive emissions.
     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
                                                                        32
packing (packed seal), mechanical seals, hydraulic seals, and lip seals.
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, one-
                                        33
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 are
                                                              34
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 fluctua-
tions.  Organic chemicals may contaminate the seal liquid and then be
                                                   35
released into the atmosphere as fugitive emissions.
                                     2-15

-------
               INVERTED CUP
          ANNU .ARCUP
         Figure 2-15.   Diagram of a  hydraulic  seal  for agitators
                                                                 36
     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
been 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
                                37
the pressure limits of the seal.
                Figure 2-16.   Diagram of agitator  lip seal.
                                                          38
                                     2-16

-------
     2.2.1.6  Open-Ended Lines.  Some valves are installed in a system 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.
     2.2.1.8  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
             39
flange faces.
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,
                              40
age, temperature, or pressure.    Therefore, model  units defined for this
analysis represent different levels of process complexity (number of
sources) rather than different unit size.
                                     2-17

-------
     2.3,1,1  Sources of Fugitive Emissions.  Data from petroleum refineries
                                                                    41
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.  Emission rates from agitator seals have not been measured.
Since there are ,10 data from similar 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 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 IT Enviroscience, Inc (formerly
Hydroscience).    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
                                                 44
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.
     Data 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
                     45
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
                                     2-10

-------
       TABLE 2-1.  EQUIPMENT COUNTS FOR FUGITIVE VOC EMISSION SOURCES
                            IN SOCMI MODEL;UNITS*


                                      Number of Components in Model  Unit0

                   .                Model  Unit     Model  Unit     Model  Unit
Equipment Component                     A              B             C

Pump seals    •        d                   '
  Light liquid service
    Single mechanical                    5             19            60
    Double mechanical                    3             10            31
    Sealless                            0              1             1
  Heavy liquid service                    ;
    Single mechanical                    5             24            73
    Packed                              2              6            20

Valves                                    !
  Gas service                          99            402           1232
  Light liquid service                131 '           524           1618
  Heavy liquid service                132            524           1618
Safety/ relief valves
Gas service
Light liquid service
Heavy liquid service
Open-ended lines
Compressor seals
Sampling connections'9
Flanges

11 '
1
1
104 .
1
26
600

42
4
4
415
2
104
2400

130
13
14
1277
8
320
7400
Reference 42.                            i

 Equipment components in VOC service only.
C52 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.

 Light 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  40.
g
 Heavy 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.

 Sample, drain, and purge valves.
9Based on 25 percent open-ended valves.  Reference 1, pg.  IV-3.
                                     2-19

-------
approximately at a vapor pressure of 0.3 kPa at 20°C.   The data collected by
Hydroscience were used to estimate the split between gas/vapor and liquid
                        46
service for each source.    In order to apply emission factors for light
                                              i
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 is 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 open-ended valves are used for sampling
            47
connections.    Emissions which occur through the valve stem, gland, and
open-end are included in the open-ended valve category.  The emission factor
for sampling connection applies only to emissions which result from sample
purging.
     2.3,1.3  Uncontrolled Fugitive Emission Estimates.  The development of
uncontrolled fugitive emission factors for SOCMI is described in
Reference 42.  The resulting emission factors are shown in Table 2-2.
Generally, the method employed the use of leak/no leak emission factors
derived from data in Reference 40 coupled with leak frequencies from
Reference 49 to arrive at average emission factors for equipment in SOCMI.
However, there are three exceptions:  (1) The gas valve emission factor
reported in Reference 50 for SOCMI units had a smaller confidence interval
associated with it, and it was substituted for the emission factor derived
from data in Reference 40; (2) The emission factor for sampling connections
is based on the amount of sampling purge reported for every 1,000 barrels of
refinery throughput   and the average count of sampling connections per
                                              52
1,000 barrels of refinery throughput reported;   (3) The emission factor
for open-ended lines represents valve seat leakage only.  The emissions
attributable to the valve, such as from around the stem and packing are
accounted for in the valve emission factor.
                                     2-20

-------
        TABLE 2-2.   EMISSION FACTORS FOR LEAKS FROM PROCESS EQUIPMENT
          Equipment
Emission Factors
  kg/hr/source
     Pump Seals
          Light Liquid
          Heavy Liquid
     Valves
          Gas
          Light Liquid
          Heavy Liquid
Compressor Seals
Safety Relief Valves - Gas
Flanges
Open-ended Lines
Sampling Connections
     0.0494
     0.0214

     0.0056
     0.0071
     0.00023
     0.228
     0.104
     0.00083
     0.0017
     0.0150
Reference 48.
                                    2-21

-------
2.4  REFERENCES
1.   Erikson, D, G., and V. Kalcevic.  (IT Enviroscience.)  Fugitive
     Emissions.  In:  U. S. Environmental Protection Agency.  Organic
     Chemical Manufacturing Volume 3:  Storage, Fugitive and Secondary
     Sources.  Publication No. EPA-450/3-80-025.  December 1980.  Report 2,
     p. II-2.                                                    !
2.   Reference 1,                                             :
3.   Reference 1, p. II-3.
4.   Ramsden, J. H.  How to Choose and Install Mechanical Seals.  Chem.  E.,
     85(22}:97-102.  1978.
5.   Reference 1, p. I1-3.
6.   Reference 4, p. 99.
7.   Reference 4, p. 99.
8.   Reference 4, p. 99.                                     ,
9.   Reference 4, p. 99.
10.  Perry, R. H., and C. H. Chilton, Chemical Engineers' Handbook, 5th  Ed.
     New York, McGraw-Hill Book Company, 1973.  p. 6-8.
11.  Reference 10, p. 6-13.
12.  Nurken, R. F.  Pump Selection for the Chemical Process Industries,
     Chem. E.» February 18, 1974.  p. 120.
13,'  Nelson, W. E.  Compressor Seal Fundamentals.   Hydrocarbon Processing,
     56(12):91-95.  1977.
14.  American Petroleum Institute, "Centrifugal Compressorsfor General
     Refinery Service", API Standard 617, FourthTaition, November, 1979,
     p. 8.(Figures reprinted by courtesy of the  American Petroleum
     Institute.)
15.  Reference 14, p. 9.
16.  Reference 13.
17.  Reference 14.
18.  Reference 14.
                                    2-22

-------
19.  Reference 13.                         '•                     .
20.  Reference 1, p. 11-7.
21.  Lyons, J. L., and C. L. Ashland, Jr.  Lyons' Enclopedla of Valves.
     New York, Van Nostrand Reinhold Co., 1975.  290p.
22.  Templeton,  I. C.  Valve Installation, Operation and Maintenance.
     Chetn. £.,78(23)141-149, 1971.        '
23,  Reference 1, p. II-5.
24.  Reference 22, p. 147-148.
25.  Reference 22, p. 148.
26.  Reference 22, p. 148.
27.  Reference 22, p. 148.
28.  Pilulik, A.  Manually Operated Valves.  Chem. E., April 3, 1978.
     p. 121.                               ,
29.  Reference 28, p. 121.                 ;  .  .
30.  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.
31.  Reference^1, p. II-7.
32.  Ramsey, W. D. and G. C. Zoller.  How the Design of Shafts, Seals  and
     Impeller Affects Agitator Performance.  Chem. E., 83{18}:1Q1-108.
     1976.
33.  Reference 32, p. 105.                .
34.  Reference 32, p. 105.
35.  Reference 32, p. 105.
36.  Reference 32, p. 106.
37.  Reference 32, p. 106.
38.  Reference 32, p. 106.
                                      2-23

-------
39.  McFarland, I.  Preventing Flange Fires.  Chem. E. Prog., 65(8):59-61.
     1969.

40.  Wetherold, R. S., L. P. Provost, and C. D. Smith (Radian Corporation.)
     Assessment of Atmospheric Emissions from Petroleum Refining,
     Appendix B:  Detailed Results.  (Prepared for U. S. Environmental
     Protection Agency.)  Research Triangle Park, N.C.  Publication No.
     EPA-600/2-83-075C.  April 1980.

41.  Reference 40.

42.  U. S. Environmental Protection Agency.  Fugitive Emission Sources of
     Organic Compounds — Additional Information on Emissions, Emission
     Reductions, and Costs.  Research Triangle Park, N.C.  Publication
     No. EPA-450/3-82-Q10.  April 1982.

43,  Reference 1, pp. IV-1, 2.

44,  Reference 1, p. II-9-13.

45.  Reference 40, pp. 11-23.

46.  Reference 1, p. II-10.

47.  Reference 1, p. IV-8.

48.  Reference 42.

49.  Blacksmith, J. R., et al.  (Radian Corporation.)  Problem Oriented
     Report:  Frequency of Leak Occurrence for Fittings in Synthetic Organic
     Chemical Plant Process Units.  (Prepared for U. S. Environmental
     Protection Agency.)  Research Triangle Park, N.C;  Publication
     No. EPA-600/2-81-003.  September 1980.

50.  Langley, 6. J. and L. P. Provost.  (Radian Corporation.)  Revision of
     Emission Factors for Nonmethane Hydrocarbons from Valves and Pump Seals
     in SOCMI Processes - Technical Note.  (Prepared for the U. S. Environ-
     mental Protection Agency.)  Research Triangle Park, N.C.
     November 1981.

51.  U. S. Environmental Protection Agency.  Compilation of Air Pollutant
     Emission Factors.  Research Triangle Park, N.C.  AP-42.  February 1980.

52.  Powell, D., et al.  (PES, Inc.)  Development of Petroleum Refinery Plot
     Plans.  (Prepared for U. S. Environmental Protection Agency.)  Research
     Triangle Park, N.C.  Publication No. EPA-450/3-78-025.  June 1978.
                                     2-24

-------
                      3.0  EMISSION CONTROL TECHNIQUES

     Sources of process unit equipment leaks of VOC were identified in
Chapter 2.  This chapter discusses the emission control  techniques  which  are
considered representative of reasonably available control  technology (RACT)
for these sources of VOC emissions.  The estimated control  effectiveness  of
each technique is also presented.
3.1  PRIMARY CONTROL METHODS
     Leak detection and repair methods can be 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.  In the  case
of open-ended lines, however, equipment leaks are treated  more effectively
by installation of control equipment.
3.1.1  Individual Component Survey for Leak Detection
     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 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
                                    3-1

-------
displayed on the instrument meter.  The performance criteria for monitoring
instruments and a description of instrument survey methods are given in
Reference Method 21.
     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 of 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  Safety/Relief Valves.  Emissions of VOC from safety/relief
valves, in general, result from leakage of the VOC around the valve seat.
The leakage is most commonly attributable to improper seating of the valve,
initially or after overpressure relieving.  There are basically three means
of eliminating VOC leaks from safety/relief valves:  (1) installation of a
rupture disk in the line prior to the relief valve; (2) connection of the
discharge port of the relief valve to a closed vent system; and (3) use of
soft seat technology such as elastomer "0-rings."
     Used upstream of the safety/relief valve, a rupture disk effectively
seals the process below the set pressure of the disk.  When this set
pressure is exceeded, the rupture disk will break, allowing the safety/
relief valve to relieve the process overpressure.  ASME codes  provide for
such installations and set forth the design constraints for installing
rupture disks in conjunction with relief valves.  ASME codes also provide
design criteria to prevent potential safety hazards from pressure building
between the disk and valve.   For example, a pressure gauge and bleed
valve installed between the disk and relief valve provide an indication of
leakage around the disk and the means to relieve this pressure.
     After an overpressure relief, a new rupture disk would have to be
installed to reseal the system.  For such an arrangement, it may be
necessary to install a 3-way valve with a parallel relief valve.  This would
allow isolation of the rupture disk/relief valve system for disk replace-
ment, while maintaining a backup relief valve in service.  A block valve
upstream of the rupture disk/relief valve system will accomplish the same
purpose where safety codes allow the use of a block valve in relief valve
service.
     The second method that effectively eliminates VOC leaks from safety/
relief valves is connection of the relief valve discharge port to a closed
vent system.  A closed vent system is composed of piping, connections, and,
where necessary, flow-inducing devices (e.g., fans, compressors); the system
transports gas or vapor to a control device such as a flare, incinerator,
boiler, or process heater.  In connecting a safety/relief valve to a closed
vent system, any leakage of VOC through the seat of the valve will  be
destroyed in the control device.
                                     3-3

-------
     The use of soft seat technology in many cases will also eliminate VOC
emissions from safety/relief valves due to improper valve seating.  An
elastomer "0-ring" can be installed so that the valve forms a tight seal
after an overpressure discharge.  Soft seat technology will not, however,
eliminate VOC emissions due to "simmering" (emissions resulting from
operation too close to the relief valve set pressure).
     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
                                   2
petroleum refineries in California.
     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
                               3
emit very small  amounts of VOC.
3.1.3  Control  Effectiveness of Leak Detection and Repair Techniques
     For some sources of fugitive VOC emissions, leak detection and repair
programs provide an effective means of reducing the total  VOC emitted.   A
                                     3-4

-------
control efficiency of a given leak detection and repair program is dependent
upon the program design and several factors associated with the program
design.  These factors include:
     «  monitoring interval;
     •  action level  or leak definition;
     •  the time interval between detection of a leak and repair; and
     d  the emissions associated with leaking sources, nonleaking sources,
         successfully repaired sources, and unsuccessfully repaired sources.
Leak detection and repair programs have been modeled using a set of
recursive equations to describe the behavior of fugitive emissions.  The
                                     4
model is detailed in a technical note  and the development of the model is
summarized in the AID.   Briefly, the leak detection and repair (LDAR)
model examines the distribution of a class (equipment type) of fugitive
emission sources in four categories:
     9  leaking sources (screening above the action level);
     •  non-leaking and successfully repaired sources (screening below
        the action level);
     «  sources that were leaking and were not successfully repaired
        (these sources cannot be repaired on-line and must await a
        turnaround for repair); and
     •  sources that were leaking, repaired, and exhibited early leak
        recurrence.
At each interval, the distribution of sources in these four categories  is
adjusted.  The average emissions rate is then determined for the class  of
sources and is dependent upon this distribution in the categories since each
category is assigned an emissions rate.  The LDAR model  presents the
emissions reduction at each interval for the interval  and as a time-weighted
average over the entire time period since the last turnaround.  The latter
values are used in the analyses presented here.
     The LDAR model computes the distribution considering a number of
parameters.  Table 3-1 lists the parameters on which the simulation results
are based.  Also provided in the table are the input values used in modeling
leak detection and repair programs for pumps in light liquid service and
valves in gas and light liquid services.
                                     3-5

-------
               TABLE 3-1.  INPUT PARAMETERS FOR LDAR MODEL3

Input Parameter
Emission factor, kg/hr/source
Occurrence rate, percent
Initial leak frequency, percent
Fractional emission reduction from:
fa) unsuccessful repair
(b) successful repair
Fraction of sources for which
repair attempts failed
Fraction of repaired sources
exhibiting early leak recurrence
Turnaround frequency, yrs.
Val
Pumps ,
Light Liquid
0.0494
10.2
8.8
0
0.972
0
0
2
ues Selected
Valves,
Gas
0.0056
3.8
11.4
0.626
0.977
0.1
0.14
2

Valves,
Light Liquid
^ 0.007
3.8
6.5
0.626
0.977
0.1
0.14
2
Selection of input parameters discussed in Reference 6.
                                    3-6

-------
     Some of the key input values used in modeling leak detection and repair
programs were the product of an EPA study of the effect of maintenance on
VOC emissions from valves and pumps.  The leak occurrence rates used in the
LDAR model were developed based on data collected in six SOCMI process
units.  For valves, simple on-line maintenance techniques were the basis of
the estimates for emission reduction due to unsuccessful repair, the
emission reduction due to successful repair, and the fraction of maintained
valves exhibiting early leak recurrence.  The 29 percent success rate for
attempted valve repair determined in the Maintenance Study was not used in
examining the effectiveness of leak detection and repair.  This low value
was the result of only simple on-line maintenance techniques, such as
tightening bolts and packing.  Under a rule, maintenance was determined to
be more effective; and, based on documented studies, a 90 percent success
rate of attempted valve repair was selected.
     For examination of leak detection and repair programs for pumps, all
seal repair attempts were assumed to be successful.  Maintenance failures
were assumed to be resultant from the mechanical aspects of the pump; these
problems would be treated under normal maintenance programs.  The emission
reduction associated with successful repair (97.2 percent) is based on the
reduction from the leaking emission factor for pumps to the nonleaking
emission factor for pumps (see Chapter 2 of the AID).
     Using the inputs given in Table 3-1, a quarterly leak detection and
repair plan was examined for valves in gas service and light liquid service;
a quarterly leak detection and repair program was also considered for pumps
in light liquid service.  The effectiveness values for these programs are
given in Table 3-2; also shown in the table are the corresponding emission
reductions for the three model units.
3.1.4  Open-Ended Lines
     Fugitive emissions from open-ended lines result from leakage through
the seat of the valve prior to the open-ended line.  Leakage of VOC to the
atmosphere from open-ended lines is most effectively prevented by installa-
tion of caps, plugs, and double block-and-bleed valves downstream of the
open-end.  Where a double block-and-bleed arrangement is used, the upstream
                                     3-7

-------
      TABLE 3-2.  EFFECTIVENESS OF QUARTERLY LEAK DETECTION AND REPAIR
                                                       Emission Reduction,
Equipment Type               Effectiveness, Percent     Per Source, kg/yr


Pumps (Light Liquid)                   32.5a                   141

Valves
Gas
Light Liquid
Safety/Relief Valves (Gas)
Compressor Seals
63. 9a
43. 9a
44. 2b
32. 9C
31
27
403
657
 Effectiveness estimated using the LDAR model.

 Effectiveness estimated using the ABCD model adjusted with the results of
 the LDAR model for valves in gas service:

          Effectiveness = ABCD$/RV  x ^LDAR gag ya1yes
                                            gas valves/

The effectiveness estimates using the ABCD model are presented in
Reference 6.


 Effectiveness estimated using the ABCD model adjusted with the results of
 the LDAR model for pumps:

          Effectiveness » ABCDcompressors x /LDARpymps


                                             ABCDpumps,

The effectiveness .estimates using the ABCD model are presented in
Reference 6.
                                     3-8

-------
valve must be closed first to ensure product is not trapped between the two
valves; expansion of trapped chemical may leak through valve stems.
     The control efficiency of using these techniques depends upon valve
seat leakage, frequency of valve use, and the amount of material  trapped
between the upstream valve and the closing device (i.e., cap, plug, second
valve, etc.).  To estimate the overall effectiveness of using these
techniques, the annual emissions can be estimated for a leaking open-ended
line that is used about ten times annually and is otherwise closed by a cap,
plug, etc.  A leaking open-ended line results in about 100 kg VOC emitted to
the atmosphere annually.   Assuming that about 0.1 kg VOC is trapped
between the valve and enclosing device, and all of this is lost each time
the open end is operated, about 1 kg VOC would be emitted annually for the
ten times the valve is used.  This relates to 99 percent efficiency; but due
to the conservative nature of this estimate, an efficiency of 100 percent
has been used to estimate the emissions reduction attributed to closing
open-ended lines.
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.  Considera-
tion of alternative control strategies for valves is pertinent because
valves account for such a large percentage of the components to be monitored
(about 90 percent in the model process units).  Furthermore, valve leaks in
general occur slowly with gradual failure of the sealing mechanism.  And the
history of leak behavior for populations of valves indicates how leaks will
occur in a valve population in the future.  Such historical  leak data
permit less frequent monitoring for valve populations with a low probability
of leaking in the near term.  However, alternative control  strategies are
not pertinent for other components (pumps, compressors, safety/relief
valves).  These other equipment types exhibit more unpredictable  failure,
with failure generally being instantaneous.  In addition, these other
components are relatively few in number, a fact which prohibits the
application of statistical sampling plans.
                                    3-D

-------
     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 stra;egies 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.
     On an individual component basis, valves have a lower emission rate
than other equipment components (Table 2-2) and have a percentage leak rate
which is lower than most other components.   As the percent of valves found
leaking decreases, the product recovery credit decreases.  The direct cost
for monitoring, however, remains the same because the number of valves which
must be monitored remains nearly the same.   Therefore, the cost effective-
ness (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.
     Figure 3-1 presents the cost effectiveness of a quarterly leak
detection and repair program for valves as  a function of the initial  percent
                                    3-10

-------
    2,000 ••
t/t

C

I  1,000

4->
U
0)
O)
O
O
          Figure 3-1.
       Leak frequency, percent


Cost effectiveness of quarterly leak detection and
repair of valves with varying leak frequency -
SOCMI units.
                                   3-11

-------
of valves found leaking.  Because it is part of the cost of the overall
implementation of RACT controls and is not specifically part of the valve
control costs, the cost of the monitoring instrument is not included in the
costs of the leak detection and repair programs represented in the figure.
There is no precise breakpoint in the cost effectiveness curve shown in
Figure 3-1.  however, EPA judges that the cost effectiveness of quarterly
leak detection and repair becomes unreasonably high at average leak
frequencies less than one percent.  Based on this judgement, an allowable
percentage of valves leaking was determined that reflects the average of
one percent of valves leaking.
     A process unit 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 process unit averaged one percent
of valves leaking, then t.he 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 process unit
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.
     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 in  light
liquid service and of attempting to repair any valves which are "leaking.

-------
This type of regulation for valves would not affect the monitoring plans  set
for other types of equipment, however.   The percentage of valves found
leaking during the inspection (prior to attempted maintenance)  would be
determined by dividing the number of valves for which a leak was detected by
the number of valves monitored.   Valves that are not monitored  because they
are known to be leaking (e.g., valves that are awaiting shutdown for repair)
are included as leaking valves in the total count of monitored  valves.  If
the results of a performance test showed that the percentage of valves
leaking was greater than the selected performance level of valves leaking
(e.g., two percent), then the process unit would be in violation of the
State regulation.
     Incorporating this type of alternative control strategy in the State
regulation would provide the flexibility of a performance standard.
Compliance with the regulation could be achieved by the method  deemed most
appropriate by the plant 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 its 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 process unit
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 equipment components would not skip monitoring
intervals; they would be subject to their required monitoring intervals.
                                   3-13

-------
     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 i_ consecutive quarters, then jn
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 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 3-3 illustrates how a "skip-period" monitoring
program 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 an
additional three quarters were skipped, all  valves would be monitored again
on the fourth quarter.  This strategy would permit a process unit 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.
     Another strategy would permit monitoring for two consecutive quarters
and skipping to semiannual monitoring.  If in two consecutive quarterly
periods the good performance level of two percent (or less) of valves
leaking is achieved, then a process unit could skip to semiannual monitoring
with 90 percent certainty that the good performance level would be met in
all periods.  Using skip period monitoring,  a process unit could develop and
implement its own leak detection and repair procedures or install valves
with lower probabilities of leaking, thereby optimizing labor and capital
costs required to achieve a good level of performance.

-------
        TABLE 3-3.  ILLUSTRATION OF A SKIP-PERIOD MONITORING PROGRAM9

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 1
Yes 2
Yes 3 '
Yes 4
Yes 5b
- 1
- 2
- 3
No 4C
Yes 1
Yes 2
Yes 3
Yes 4
Yes 5b
- 1
- 2
- 3
Yes 4d
- 1
- 2
- 3
Yes 4d
ai=5, m=3, good performance level  of 2 percent.

 Fifth 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 means1 3 quarters of monitoring  may  be
 skipped.
                                    3-15

-------
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 turnaround.
3.3.1  Unsafe and Difficult to Monitor Components
     Some components might be considered unsafe to monitor because of
process conditions such as 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 valves might be monitored at times
when process conditions are such that the valves are not operating under
extreme temperatures or pressures as would be found in high pressure polymer
reactors.
     Some valves may be difficult to monitor because access to the valve
bonnet is restricted or the valves are located in elevated areas.  These
valves might be 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 would be
monitored quarterly.  However, if the monitoring of certain valves would
require the use of scaffolding or would require the elevation of monitoring
personnel higher than two meters above permanent support surfaces, these
valves might be exempted from quarterly monitoring provided they are
monitored annually.
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 those
process units 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
                                     3-16

-------
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.
     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 3-4 shows the cost effectiveness for quarterly leak detection and
repair of valves in process units processing small quantities of light
liquid and gaseous VOC.  Using the processing rates at the optional cost
effectiveness cutoff levels as a guideline, States may wish to consider
exempting process units designed for processing small volumes of light
liquid and gaseous VOC from regulations requiring control of fugitive VOC
emissions.
3.3.3  Unit Turnarounds
     A State agency might wish to consider a provision in its regulations
which would allow the agency Director to order an early unit shutdown for
repair of leaking components in cases where the percentage of leaking
components awaiting repair at unit turnaround becomes excessive.  Use of
such a provision, however, must be carefully considered in terms of the
emissions reduction achievable and the costs to the process unit in
production down-time and repair cost.
     Alternative methods of treating delay of repair could also be
considered by a State or local agency in reducing the cumulative number of
unrepairable equipment components.  For instance, delays of repair to the
                                    3-17

-------
    TABLE 3-4.  COST EFFECTIVENESS FOR QUARTERLY LEAK DETECTION
         AND REPAIR PROGRAMS FOR PROCESS UNITS PROCESSING
           SMALL VOLUMES OF LIGHT LIQUID AND GASEOUS VOC
                                        Volume Light Liquid
Cost-Effectiveness                        And Gaseous VOC
      ($/Mg)                             Processed (Mg/Yr)

        500                                     3,660

      1,000                                     1,850

      1,500                                     1,210

      2,000                                       890
                             3-18

-------
next scheduled process unit shutdown (or turnaround) could be allowed under
circumstances where it is technically infeasible to repair the component
in-place/on-line (i.e., without a unit shutdown) or where replacement parts
have been depleted from once-sufficient inventory.  By requiring records of
delays and reasons for delays, State enforcement officers would be supplied
with the data necessary to determine compliance.
                                      3-19

-------
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.

2.   Teller, James H.  Advantages Found in On-Line Leak Sealing.   Oil  and Gas
     Journal, 77 (29):54»59, 1979.

3.   Wetherold, R. G., L. P. Provost, and C. D. Smith.  (Radian  Corporation.}
     Assessment of Atmospheric Emissions from Petroleum Refining, Appendix B:
     Detailed Results.  (Prepared for U. S. Environmental Protection Agency.)
     Research Triangle Park, N.C.  Publication No. EPA-600/2-80-Q75c.
     April 1980.

4.   Williamson, H. J., et al.  (Radian Corporation.)  Model for Evaluating
     the Effects of Leak Detection and Repair Programs on Fugitive Emissions.
     Technical Note DCN 81-290-403-06-05-03.  September 1981.

5.   U. S. Environmental Protection Agency.  Fugitive Emission Sources of
     Organic Compounds - Additional Information on Emissions, Emission
     Reductionss and Costs.  Research Triangle Park, N.C.  Publication
     No. EPA-450/3-82-010.  April 1982.

6.   Reference 5.

7.   Reference 5.
                                     3-20

-------
                     4.0  ENVIRONMENTAL ANALYSIS OF RACT
                                                                       s*
     This chapter discusses the environmental impacts that would result from
implementing reasonably available control technology (RACT), which is
presented in Section 4.1.  The primary emphasis is a quantitative assessment
of VOC emissions in the absence of RACT (baseline emissions) and after
implementation of RACT.  The impacts of RACT upon water quality, solid
waste, and energy consumption are also addressed in this chapter.
4.1  REASONABLY AVAILABLE CONTROL TECHNOLOGY (RACT) PROCEDURES
     Reasonably available control technology (RACT) procedures for equipment
leaks of VOC in SOCMI and polymer manufacturing include capping of open-
ended lines and quarterly leak detection and repair of pumps, valves,
compressors, and safety/relief valves.  Routine instrument monitoring of
flanges, connections, and equipment in heavy liquid service is not
necessary.  However, any component that appears to be leaking, on the basis
of sight, smell, or sound, should be repaired.   In addition, difficult-to-
monitor valves may require less frequent monitoring than the quarterly plan
considered as RACT for valves in gas or light liquid service.  Small  process
units (e.g., units processing small quantities  of light liquid and gaseous
VOC) may be exempted from implementing routine  leak detection and repair
programs on the basis of cost effectiveness for these small  units (see
Table 3-4).  Other exemptions might include process units processing  only
heavy liquid VOC or processing only non-VOC and equipment operating under a
vacuum.
     Leak detection should consist of quarterly monitoring the following
components in VOC service with an organic detection instrument:  pumps in
light liquid service, valves in light liquid service, valves in gas service,
compressors, and safety/relief valves in gas service.  However, states may
choose monthly monitoring for pumps instead of  quarterly monitoring,  because
the cost effectiveness ratio associated with monthly monitoring is more
attractive than the cost effectiveness ratio for quarterly monitoring.
Pumps in light liquid service should also be visually inspected weekly for
indications of leaks.  Safety/relief valves should also be monitored  after
each overpressure relief to ensure the valve has properly reseated.

                                     4-1-

-------
     Compressor seals should be monitored quarterly; however, some plant
owners and operators may experience difficulty in reducing concentrations of
organic compounds to less than 10,000 ppmv.  Moreover, repair of compressor
seals often necessitates a potential or complete process unit shutdown
because compressors are generally not spared.  Consequently, plants may find
it preferable to install a compressor vent control system.  Howevejr,
retrofitting existing compressors with these systems may pose a safety
problem.  Because of the problems associated with quarterly monitoring or
with installation of equipment controls in certain cases, RACT for
compressors, therefore, will be determined on a case-by-case basis;.  The
estimates of emission reductions for compressors in this chapter are based
on implementing a quarterly leak detection and repair program.
     The organics detection instrument and the monitoring method employed
should be EPA Reference Method 21 or an equivalent State method.  A source
is considered leaking if monitoring results in an instrument reading of
10,000 ppmv or greater.  A soap solution may be applied to certain equipment
as a preliminary screening technique for leakage.  A soap score equivalent
to 10,000 ppmv is not specified in this guideline document because soap
scoring is not applicable to all source types and because it involves a
subjective evaluation of bubble formation over a specified period of time.
Hov/ever, states may wish to allow plant owners or operators to use the soap
score method based on a correlation between soap scoring and instrument
readings for sources where soap scoring is applicable.  Leaking components
should be repaired within 15 days of the date the leak is detected.  Repair
should be considered as reduction of the measured organics concentration
below 10,000 ppmv.  Leaking components which cannot be repaired without a
unit shutdown should be repaired at the next unit turnaround.
     RACT should be applicable only to components in VOC service.   A
component is 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.  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
                                     4-2

-------
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.
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 in Table 4-1.  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 4-2.
     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.
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 Table 4-1.  The reduction in emissions for the model units
after RACT would be implemented is 37 percent.
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.
                                     4-3

-------
           TABLE 4-1.  ESTIMATED EMISSIONS AND EMISSIONS REDUCTION
                       ON A MODEL UNIT BASIS
Level of
Control
Estimated Emissions
    . (kg/hrj
    Model Unit
                                     Estimated Emissions
                                           (Mg/yr)
                     sm
                     Tur
Model Unit
  A
B
     B
Average Percent
Reduction From
 Uncontrolled
     Level
Uncontrolled

RACT
 4.5   17.2   53.7     39     150    470

 2.8   10.9   34.0     25      96    300
                                          37
                                     4-4

-------
                           TABLE 4-2.  EMISSION FACTORS FOR SOURCES CONTROLLED UNDER RACT
en

Uncontrolled
Emission Source
Pumps
Light Liquid Service
Valves
Gas Service
Light Liquid Service
Safety/Relief Valves
Gas Service
Compressors
Inspection
Interval
Quarterly
Quarterly
Quarterly
Quarterly
Quarterly
Uncontrolled
Emission Factor,
kg/hr
0.0494
0.0056
0.0071
0.104
0.228
Control .
Efficiency
0.33
0.64
0.44
0.44
0.33
Controlled
Emission Factor,
kg/hr
0.0333
0.0020
0.0040
0.0580
0.153
         Trom Table 2-2.

          Control efficiency estimated based on LDAR model  results with inputs detailed in the AID.  For
          compressors and safety/relief valves (gas service), the control  efficiency estimates were made
          using results of the ABCD model adjusted with results of the LDAR model for comparable equipment
          types, as discussed in Section 3.1.3.  References 1, 2, 3.

         'Controlled emission factor = uncontrolled emission factor x [1 - (control efficiency)].

-------
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.   U. S. Environmental Protection Agency.   VOC Fugitive Emissions  in
     Synthetic Organic Chemicals Manufacturing Industry - Background
     Information for Proposed Standards.  Research Triangle Park, N. C.
     Publication No. EPA-450/3-80-033a.   November 1980.

2.   U. S. Environmental Protection Agency.   Fugitive Emission Sources of
     Organic Compounds - Additional Information on Emissions, Emission
     Reductions, and Costs.  Research Triangle Park, N. C.  Publication  No.
    . EPA-450/3-82-010.  April 1982.

3.   Memorandum from Stelling, John, Radian  Corporation, to SOCMI Fugitives
     NSPS File.  March 5, 1982.  1 p.  Estimated effectiveness of leak
     detection and repair programs for pressure relief devices.
                                     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 imple-
mentation 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
     The cost of a VOC monitoring instrument includes the cost of two
instruments.  One instrument is intended to be used as a spare.  The cost
of $4,600 for a portable organic vapor analyzer was obtained from a
             2
manufacturer.
5.1.2  Caps on Open-Ended Lines
     Fugitive emissions from open-ended lines and valves can be controlled
by installing a cap, 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
                                     5-1

-------
                                            TABLE 5-1.  CAPITAL COST DATA
                   Item
       Cost Value
    Used in Analysis
   (June 1980 Dollars)
        Cost Basis
Reference
         Monitoring Instrument        2  x 4600 = 9200/model unit

         Caps for Open-Ended Lines    53/1ine
         Replacement Seal
140/seal
en
i
One instrument used as a spare     2

Based on cost for 1" screw-on      3
type valve.  Cost June 1980 =
$35.  Installation = 1 hour at
$18/hour.

Based on cost of single          4,5,6
mechanical seal, assuming
50 percent credit for old
seal.  June 1980 cost
determined from 1978 cost of
$113/seal and the ratio of
indices (331.8/268.1).

-------
          TABLE 5-2.  CAPITAL COST ESTIMATES FOR IMPLEMENTING RACTa
                      (Thousands of June 1980 Dollars)


1.
2.
3.
4.

Capital Cost Item
Monitoring Instruments .
Caps for Open-ended Lines .
Initial Leak Detection and Repairc'Q'e
Initial Pump Repair Costs (Replacement Seals)
Total
Model Unit Costs
A B C
9.2 9.2 9.2
5.5 22.0 67.9
1.1 4.2 13.0
0.1 0.4 1.1
15.9 35.8 91.2
 Based on cost data presented in Table 5-1.

 Number of open-ended lines from Table 2-1.

 Initial  leak detection and repair costs are treated as capital  costs  since
 they are incurred only once.

 Includes screening and repair labor charges.
A
 Repair costs are industry-averaged per unit and,  therefore,  consider
 fractional  repairs.  Equipment repair was not rounded to whole  component
 repairs.
                                     5-2

-------
line.  The cost of a cap for an open-ended lines is based on a cost of $35
for a one-inch screw-on type globe valve.  This cost was supplied 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.
     Caps, plugs, and blind flanges can be used at much less cost; the
capital cost of installing these enclosures range from about $0.40 per plug
for 1/4-inch hex head plugs to about $26 per 2-inch blind flange.   Costs  for
                                                                   8 9 10
1-inch components range from about $1.20 per plug to $5.20 per cap. ' '
Ninety-two percent of the open-ended lines surveyed in one study were less
than 2-inches in diameter.    Therefore, the cost estimate of $53 per
open-ended line is conservative given the prevalence of small sizes and
alternative enclosing devices.
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 not rounded up to  the
next highest integer, thus resulting in industry-averaged values.   The
repair time for fixing leaks is estimated to be 16 hours for a purnp seal,
40 hours for a compressor seal, and 1.13 hours for a valve.  The repair time
for fixing compressor seals includes the cost of a new seal.  These require-
                                 12
ments 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.
5.1.4  Replacement Pump Seals at Initial Repair
     As with the initial leak detection discussed in the previous section,
the cost of initial seal replacements for pumps in light liquid service is a
one-time cost and is treated as a capital cost.  A replacement seal cost  of
$140 per seal is based on the cost of a single mechanical seal and assumes a
                                     5-4

-------
                      TABLE 5-3.  LABOR-HOUR REQUIREMENTS FOR INITIAL LEAK REPAIR UNDER RACT
en
U1

Number of Components Estimated Number
Per Model Unit Of Initial Leaks*
Source Type
Pumps (Light Liquid)
Valves (In-Line)
Gas
Light Liquid
Safety/Relief Valvesd
(Gas Service)
Compressor Seals
TOTAL
A
8

99
131

11
1

B
29

402
524

42
2

C A B C
91 0.7 2.6 8.0

1232 11.3 45.8 140
1618 8.5 34.1 105

130 0 0 0
8 0.1 0.2 0.7

Repair Time, Labor-Hours
Hours Required
A B
16b 11 41

1.13C 13 52
1.13C 10 38

0 00
40e 4 7
38 138 .

C
128

159
119

0
29
435
        aBased  on  the percent of sources leaking at > 10,000 ppm.  Reference 13.
         Includes  repair time for pump seals replaced in the field and not for retrofitting of packed
         seals  with mechanical seals.  Reference 13.
        Weighted  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.
         References 12, 14.
         These  leaks are corrected by routine maintenance at no additional labor requirements due to
         safety requirements.  Reference 12.
        Includes  labor-hour equivalent cost of new seal.  Reference 15.

-------
50 percent cost credit for the seal being replaced.  The number of initial
leaks per model unit is the percentage of initial leaks multiplied by the
number of pumps (light liquid service) in the model unit.  To present
industry-averaged values for each model units, the fractional repairs
required were not rounded to the next integer.
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.
5.2.1  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 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 man-minute for visual
inspection, 2 man-minutes for valves, 10 man-minutes for pump seals, 16 man-
                                                                      21
minutes for safety/relief valves, and 20 minutes for 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
                      22
of their capital cost.     The annual cost of materials and labor for
maintenance and calibration of monitoring instruments is estimated to be
       23 24 25
$3,000. °>*-H>"  An  additional  miscellaneous charge of four percent of
                                     5-6

-------
               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 RACT

6.   Annualized charge for initial
       leak repairs
7.   Recovery credits
0.163 x capitalc
0.23 x capital
0.05 x,capital
$3,000°
0.04 x capital®
0.04 x capital

$18/hourf

0.40 x (monitoring + repair
  labor)*

E (estimated number of leaking
  components per model unit x
  repair timekx $18/hr  x
  1.4g x 0.163n

$429/Mg VOC1
 Ten year life, ten percent interest.   From Reference 15.

 Six year life, ten percent interest.   From Reference 15.
r
 From Reference 15.

 Includes materials and labor for maintenance and calibration.   Reference  15.
 Cost index » 247.3 * 223.5 (Reference 16 and 17).

eFrom Reference 15.

 Includes wages plus 40 percent for labor-related administrative and  overhead
 costs.  Cost (June 1980} from Reference 1.

"From Reference 15.

.Initial leak repair amortized for ten years at ten percent interest.
References 18, 19, 20.  Producer price index ratio = 327.3/228.8.
                                     5-7

-------
                      TABLE 5-5.   ANNUAL MONITORING  AND LEAK  REPAIR  LABOR REQUIREMENTS FOR RACT
01
CO
Monitoring
Number Of
Components
Per Model Unit
Source Type
Puwps (Light Liquid)

Valves (Gas)
Valves (Light Liquid)
Safety/Relief Valves (Gas)
Compressor Seals
TOTAL
A
8

99
131
11
1

8 C
29 91

402 1232
524 1618
42 130
2 8

Monitoring
T1«te Per
Type Of Source,
Monitoring
Instrument
Visual
Instrument
Instrument
Instrument
Instrument

Fraction
Monitoring
Labor Hours .
Required Annually
Han-Hi n Monitored A -B
10
0.5
2
2
16
20

4
52
3.94
3.94
4
4

5.3
3.5
13.0
17.2
11.7
1.3
52
19
13
53
69
45
2.7
202
C
61
39
162
212
139
10.7
624
Leak Repair
Estimated Repair
Number Of Tine Per
Fraction Leaks Annually Source,
Repaired ABC Han-llr
0.394 3.2 11.4 35.9 16a

0.186 18.4 74.7 229 1.13d
0.186 24.4 97.4 301 1.13d
_e _e _e _e
0.394 0.4 0.8 3.2 40f




Leak Repair
Labor Hours
Required Annual ly"
A
50

21
28
_e
16
115
B
183

84
110
_e
32
409
C
574

259
340
_e
126
1299
Reference 26, 27.
Monitoring labor hours = (Monitoring time) x (Number of components) x (Fraction Monitored); the fraction monitored annually Is an output of the LDAR
 model.
°Fract1onal leaks considered and provided as an output of the LDAR model.
 Leak  repair labor hours - (Repair time) x (Estimated number of leaks).
eThe occurrence of leaks from safety/relief valves  Is dependent upon the frequency of operation of the safety/relief valves.  No estimates of required
 repairs have been presented; however, any leak that 1s detected In the absence of a  RACT requirement would be repaired under normal plant maintenance
 practices.  Reference 27.
^Reference 27.

-------
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 administration and support costs include recordkeeping and
reporting requirement costs.
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
repair 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:

               CRF =  1(1 + if
                     (1 + i)n - 1
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
$429/Mg.28'29'30
                                     5-9

-------
                       TABLE 5-6. -RECOVERY CREDITS

Model
Unit
A
B
C
Uncontrolled
Emissions,
Mg/yr
39
151
470
Emissions
Under RACT,
Mg/yr
25
96
300
Emission
Reduction,
Mg/yr
14
55
170
Recovered3
Product Values
$/yr
6 ,200
24 ,000
74,000
Based on an average price of $429/Mg.
                                   5-10

-------
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.
5.3.1  Initial Costs
     The cost of initially implementing RACT consists of capital costs and
initial leak repair.  The capital cost of $9,200 for two monitoring instru-
ments is the same for all model unit sizes.  Caps for open-ended lines will
cost $5,500 for model unit A, $22,000 for model unit B, and $67,900 for
model unit C.  The one-time initial leak repair cost is $1,200 for model
unit A, $4,600 for model unit B, and $14,100 for model unit C.  The total
initial capital costs for implementing RACT are $15,900 for model  unit A,
$35,800 for model unit B, and $91,200 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 imple-
mentation of RACT will result in saving $6,200 worth of VOC annually in
model unit A, $24,000 worth of VOC in model unit B, and $74,000 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 $3,200 and the
total annual operating costs are $11,500.  Product recovery credits total
$6,200.  The net annualized cost for model unit A is $5,600.
     The annualized capital charges for model unit B are $6,500 and the
total annual operating costs are $28,000.  The recovery credit is  $24,000
per year; thus, the net annualized cost for model unit B is $4,300.
                                     5-11

-------
        TABLE 5-7.  ANNUALIZED CONTROL COST ESTIMATES FOR MODEL UNITS
                    UNDER RACT (Thousands of June 1980 Dollars)

Model Unit
Cost Item
Annual ized Capital Charges
1. Control Equipment
a. Instrument
b. Caps
2. Initial Leak Repair
Subtotal
Operating Costs
1. Maintenance Charges
a. Instrument
b. Caps
2. Miscellaneous (taxes, insurance,
administration)
a. Instrument
b. Caps
3. Replacement seals
4. Labor
a. Monitoring labor b
b. Leak repair labor
c. Plant and payroll overhead
Subtotal
Total Before Credit
Recovery Credits
Net Annual ized Cost
A

2.11
0.90
0.21
3.22

3.00
0.28
0.37
0.22
0.24
0.94
2.06
1.20
8.31
11.53
6.19
5.34
B

2.11
'3.58
0^76
6.45

3.00
1.10
0.37
0.88
0.88
3.62
7.36
4.39
21.60
28.05
23.76
4.29
C

2.11
11.01
JL38
15.50

3.00
3.38
0.37
2.71
2.77
11.23
23.38
13.84
60.68
76.18
73.88
2.30
aSum of labor hours for monitoring in Table 5-5  multiplied  by  $18/hour.
 Sum of labor hours for leak repairs in Table 5-5  multiplied by $18/hour.
°Based on 40 percent of monitoring labor plus leak repair labor costs.
                                     5-12

-------
     Model unit C has annualized capital charges of $16,500 and total
operating expenses of $76,200.  The recovery credit is $74,000 per year.
The net annualized cost for model unit C is $2,300 for 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 of $5,400.  The emission reduction associated with RACT is 14.4 Mg/yr,
resulting in a cost effectiveness of $370/Mg.
     The implementation of RACT in the case of model unit B results in a  net
annualized cost of $4,300.  The emission reduction associated with RACT  is
55.4 Mg/yr and the cost effectiveness is $77/Mg.
     The implementation of RACT in the case of model unit C results in a  net
annualized cost of $2,300.  Th.e emission reduction associated with RACT  is
172 Mg/yr.  Therefore, the cost effectiveness is $13/Mg.
     A comparison of the cost effectiveness of RACT for each model unit
reveals that cost effectiveness improves as model  unit size increases. The
strong influence of recovery credits and the constant charge for monitoring
instruments regardless of model unit are responsible for the increase  in
cost effectiveness.
                                     5-13

-------
TABLE 5-8.  COST EFFECTIVENESS FOR MODEL UNITS UNDER RACT

Model Unit9

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
11.53
6.19
5.34
14.4
370
B
28.05
23.76
4.29
55.4
77
C
76.18
73.88
2.30
172
13
a(XXX) - net credit.
                           5-14

-------
     The cost effectiveness of RACT for each component type in Model  Unit  B
is presented in Table 5-9.  The cost effectiveness of RACT by component for
the other model units is the same since there are no economies of scale
associated with the control techniques and since the cost of the monitoring
instrument is not considered for this individual component analysis.   Thus,
the individual cost effectiveness values by component presented in Table 5-9
for Model Unit B are the same as the by-component cost effectiveness  values
for other model units.  The overall cost effectiveness values for the three
model units differ as a result of the fixed cost for the monitoring
instrument.  The cost of the monitoring instrument cannot be attributed to
any single type of component since all components are monitored by the
instrument.  Therefore, the cost for each component does not include  the
cost of the monitoring instrument.  The cost effectiveness for RACT for
pumps and compressors is higher than other components due to the additional
costs required for leak repair.
                                    5-15

-------
                        TABLE 5-9.  COST EFFECTIVENESS FOR COMPONENT TYPES IN MODEL UNIT B
en
i-1
cr>

Component
Pumps (Light Liquid)
Valves
Gas Service
Light Liquid Service
Safety/Relief Valves
Open-ended Lines
Compressors
TOTAL UNIT (Without
Instrument Cost)
TOTAL UNIT (With
Instrument Cost)
Number Of
Components
29

402
524
42
415
2
1,414
1,414
Annualized
Cost Before
Credit ($)a
6,670

3,670
4,650
1,130
5,560
890
22,570
28,050
Annual
Recovery
Credit ($)
1,750

5,400
6,140
7,260
2,650
560
23,760
23,760
Net
Annualized
Cost ($)a
4,920

(1,730)
(1,490)
(6,130)
2,910
330
(1,190)
4,290
Total VOC Cost
Reduction Effectiveness
(Mg/yr) ($/Hg)
4.08

12.6
14.3
16.9
6.18
1.31
55.4
55.5
1,200

(140)
(100)
(360)
470
250
(21)
77
         Does  not  include cost of monitoring instrument, unless otherwise noted.
         The net cost associated with monthly monitoring of pumps  is  lower than the net  cost  shown  for
         quarterly monitoring of pumps.  The lower net cost associated with monthly monitoring  results
         from  higher emission reductions and, therefore, higher recovery credits.  The following  cost
         figures are applicable to monthly monitoring for  pumps:

            Annualized cost before credit  =  8,439
               (gross cost)  ($)                          .        .              ,
            Emission reduction (Mg/yr)     =  7.6.
            Annual recovery credit ($)     =  3,277
            Net annualized cost ($)        =  5,162
            Cost  effectiveness ($/Mg)      =  680

        "Cost  for  caps on lines only.  Not monitored under RACT.

-------
5.5  REFERENCES

1.   Letter with attachments from Texas Chemical  Council  to Walt Barber,
     U. S. Environmental Protection Agency.   June 30, 1980.

2.   Purchase order from GCA/Technology Division  to Analabs/Foxboro,
     North Haven, Connecticut.   July 3, 1980.

3.   Telecon.  Samuel Duletsky, SCA Corporation with Dave Myer,  Piedmont
     Hub, Greensboro, N.C.  September 25, 1980.  Price of 1" screw-on type
     valve.

4.   U. S. Environmental Protection Agency.   Fugitive Emission Sources of
     Organic Compounds - Additional Information on Emissions, Emission
     Reductions, and Costs.   Research Triangle Park, N.C.  Publication
     No. EPA-450/3-82-010.  April 1982.

5.   Economic Indicators.  Chemical Engineering.   Volume  86, Number 7.
     March 26, 1979.

6.   Economic Indicators.  Chemical Engineering.   Volume  87, Number 21.
     October 20, 1980.

7.   Reference 3.

8.   Richardson Engineering Services, Inc.   Process Plant Construction
     Estimating Standards, Volume 3.  Sol ana Beach, California.   1982.

9.   Economic Indicators.  Chemical Engineering.   Volume  89, Number 21.
     October 18, 1982.

10.  Reference 6.

11.  Memorandum from Hustvedt,  K. C., U.  S.  Environmental Protection  Agency,
     to SOCMI Fugitive NSPS File.  January 5, 1982.  9p.   Summary of  HI
     Survey Data Compiled on 15 July 1980:   Open-ended Valves (Vent,  Drain,
     Sample) by Size.

12.  Letter with attachments from J. M. Johnson,  Exxon Company,  U.S.A., to
     Robert T. Walsh, U. S.  Environmental Protection Agency.  July 28, 1977.

13.  Reference 4.

14.  State of California Air Resources Board.  Emissions  from Leaking
     Valves, Flanges, Pump and  Compressor Seals,  and Other Equipment  at Oil
     Refineries.  April  1978.   p. V-18.
                                    5-17

-------
15.  Environmental Protection Agency.  Control of Volatile Organic Compound
     Leaks from Petroleum Refinery Equipment.  Publication No.
     EPA-450/2-78-036, OAQPS No. 1.2-11.  June 1978.
16.  Economic Indicators.  Chemical Engineering.  Volume 86, Number 2.
     January 15, 1979.
17.  Economic Indicators.  Chemical Engineering.  Volume 87, Number 19.
     September 22, 1980.
18.  Reference 4.
19.  Reference 5.
20.  Reference 6.
21.  Reference 12.
22.  Reference 15.
23.  Reference 15.
24.  Reference 16.
25.  Reference 17.
26.  Reference 4.
27.  Reference 12.
28.  Reference 4.
29.  Reference 5.
30.  Reference 6.
                                     5-18

-------
               APPENDIX A



MAJOR COMMENTS RECEIVED ON THE DRAFT CTG

-------

-------
                                 APPENDIX A
                  MAJOR COMMENTS RECEIVED ON THE DRAFT CTG

     Twenty-two comment letters were received on the August 1981 draft CTG
distributed in December 1981.  Some of the comments received addressed the new
source performance standards (NSPS) for equipment leaks of VOC in SOCMI.
These comments are considered only within the technical content of the CTG;
that is, the technical aspects of the comments are considered, whereas the
regulatory decisions concerning NSPS are not addressed.
     In April 1982, EPA published Fugitive Emission Sources of Organic
Compounds -- Additional Information on Emissions, Emission Reductions, and
Costs (EPA-450/3-82-010), or AID.  The AID represents EPA's current under-
standing of equipment leaks of VOC and contains the methodology for examining
emissions, emission reductions, and costs.  The AID served as the primary
reference in revising cost and emission estimates presented in this document.
     Out of the 22 comment letters, the following major comments were identi-
fied as having appeared several times or as having cited issues that resulted
in revisions to the CTG.  The comment letters are given in their entirety in
Appendix B.  The 20 comments identified in this appendix are addressed speci-
fically with reference to the final CTG.  The following comment areas are
discussed here:
     (1)  Need and coverage of the CTG;
     (2)  Estimates of emissions, emission reductions, and costs; and
     (3)  RACT selection, provisions, and exemptions;

A.I  NEED AND COVERAGE OF THE CTG

Comment:  Some commenters [#3; #9; #151* said that SOCMI fugitive emissions
are minor sources of VOC, and that there is, therefore, no need for the CTG.
The need for the CTG was further mitigated they said, by the fact that
fugitive emissions of VOC from SOCMI were already near the level  EPA hopes to
                                     A-l

-------
achieve through control techniques outlined in Chapter 4.  They based their
argument on a comparison of estimates of controlled emissions based on
petroleum refinery data arid uncontrolled emissions based on SOCMI data.

Response:  Emissions of VOC from SOCMI represent a significant source of VOC
emissions to the atmosphere.  EPA estimates that 540 Gg/yr of VOC (540,000
Mg/yr) of VOC are emitted to the atmosphere from all sources in SOCMI (see
Table A-l).  This estimate of emissions is based on detailed studies of
individual process source types including air oxidation processes, distilla-
tion operations, storage operations, carrier gas processes, equipment leaks,
and secondary sources.  540 Gg/yr of VOC is a significant quantity of VOC to
be emitted as air pollution.  This quantity is large in absolute terms and is
large relative to other VOC source categories.  Fugitive emissions of VOC from
SOCMI are estimated to be approximately 190 Gg/yr, thus contributing a large
proportion of VOC emissions within the SOCMI source category.
     The commenters1 comparison of emission estimates indicates confusion over
the purpose for CTG's and EPA's approach in developing them.  EPA's intent is
not to set a regulatory goal.  Rather, the intent is to provide State and
local air pollution control agencies with information for determining
reasonably available control technology (RACT) for specific stationary
sources.

Comment:  There were some comments [#7; #8; #20] received on the coverage of
equipment leaks by other regulations.  One commenter stated that the control
techniques recommended in the CTG were already in place for vinyl chloride
plants under National Emission Standards for Hazardous Air Pollutants
(NESHAP).  Other commenters discussed the potential overlap in regulations set
forth by EPA and OSHA.  One commenter said that new regulations would be
redundant considering the existence of OSHA regulations.   Another stated that
for some chemical plants OSHA already has standards governing equipment leaks
of VOC.   For example, the commenter said that acrylonitrile plants are subject
to strict workplace exposure limits set by OSHA.
                                     A-2

-------
            TABLE A-l.  ESTIMATES OF-VOC EMISSIONS FROM SOCMI?
Category
Fugitive emissions
Distillation operations
Air oxidation processes
VOL storage operations
Carrier gas processes
Secondary & misc. emissions
Revised total VOC emissions
Gg/yr
189
140
110
47
32
26
544
Percent of Total
35
26
20
8
6
5
100
Estimates for process emission sources estimated using best available
information from current standards development programs (25 October 1982),
Reference 1.

Secondary and miscellaneous emissions estimated as 5 percent of the total
of the other sources.
                                   A-3

-------
Response;  As discussed in Chapter 1, Control Techniques Guidelines (CT6) are
designed to assist states in bringing non-attainment areas into compliance
with national ambient air quality standards by providing them control
technology information.  Their role is distinct from that of OSHA regulations
and national emission standards for hazardous air pollutants (NESHAP).
     NESHAP are developed to control pollutants that are hazardous because
they are carcinogens or the cause of other serious diseases.  Some of the
individual SOCMI chemicals have been identified as hazardous air pollutants
and some SOCMI units may be affected by NESHAP regulations.  However, SOCMI
VOC emissions as a class have not been identified as hazardous pollutants, and
therefore, are not subject to NESHAP.  Therefore, there is still a need for
the CTG.  The CTG is consistent with both NSPS and NESHAP with respect to the
equipment covered.  There is no duplication of efforts required by the CTG;
equipment covered by NSPS or NESHAP is exempt from the CTG since the equipment
is already controlled to a higher degree under these other programs.
     Many of the chemicals covered by the CTG are also listed in Table Z-l,
Toxic and Hazardous Substances, in the general provisions for OSHA (29 CFR
1910.1000), and some of these chemicals are also covered by more specific
health standards under OSHA.  As a consequence, the CTG and the OSHA standards
may affect the same equipment in VOC service.  However, this possibility also
does not negate the need for the CTG.
     Control techniques described in the CTG serve to limit mass emission
rates directly; OSHA standards for toxic chemicals generally do not.  Under
OSHA, control of emission sources may include substitution with less hazardous
materials, process modification, worker rotation, process or worker isolation,
ventilation controls, or modification of work practices.   These controls
reduce occupational exposures, but they do not necessarily reduce the mass
rate of VOC emissions to the atmosphere.   Relying on indirect controls that
may or may not reduce emissions that would degrade air quality would be an
unreasonable approach to reducing emissions of VOC.   However, in some
instances, control of emissions provided by OSHA requirements may be
sufficiently effective to allow an alternative standard (e.g., .percentage of
valves leaking) to be met.  Furthermore,  the need for CTG controls can be
                                     A-4

-------
eliminated for certain sources under specific circumstances.   For example,  the
quarterly monitoring requirement for a pump seal could be eliminated if the
pump is equipped with dual mechanical seals with a non-VOC barrier fluid
system/degassing reservoir connected to a closed vent system.

Comment:  Several [#8; #13; #16; #21; #22] commenters recommended deletion  of
styrene-butadiene latex from the list of processes covered by  the CTG.
Commenters pointed out the fact that styrene-butadiene latex plants consist of
fewer reactors and ancillary equipment than styrene-butadiene  crumb rubber
plants, although the equipment is of a similar type.  Therefore, fugitive
emissions should be similar in magnitude or lower.  Since the  crumb rubber
processes were deleted from the list, it seemed appropriate to the commenters
to delete the latex rubber processes.
     Commenters further pointed out the fact that the production of styrene-
butadiene latex is less than 15 percent of the production of styrene-butadiene
crumb rubber.  This comparatively low production rate was considered further
justification for deleting styrene-butadiene latex from the list, since it  is
a small part of the total styrene-butadiene production.
     Another point made in support of deleting styrene-butadiene latex  was  the
fact that most gas valves in styrene-butadiene plants are in vacuum service,
so they would not be sources of fugitive emissions.
     Some commenters continued their argument that neither polymer nor  resin
manufacturers are similar to the chemical producers in SOCMI.   Specifically, a
number of commenters stated that styrene-butadiene latex plants are not like
the remaining plants in SOCMI.  Furthermore, the commenters said there  were no
styrene-butadiene plants in the SOCMI data base.  The commenters, therefore,
concluded that the SOCMI data base was not representative of their particular
industry segment.

Response:  In the August 1981 draft CTG, four categories of polymers and
resins were included for coverage under this CTG:   polyethylene, polypro-
pylene, polystyrene, and styrene-butadiene latex.   Other polymers and resins
were dropped from consideration under the CTG for equipment leaks of VOC prior
                                     •A-5

-------
to the distribution of the draft CT6.  EPA has also removed styrene-butadiene
latex from the list of affected chemicals.  Thus, this CTG covers only
polyethylene, polypropylene, and polystyrene from the polymers and resins
category.  However, in the future EPA will consider recommending RACT for
equipment leaks of VOC from units producing the polymers and resins not on the
final list.
     The decision to drop styrene-butadiene latex from consideration at this
time is not based on the inapplicability of SOCM! data or RACT to styrene-
butadiene latex units.  The data base presented in the AID (on which the final
CTG is based) is comprised of data on equipment leaks of VOC in SOCMI.  The
data were collected from a variety of SOCMI process types and are considered
representative of VOC emissions from equipment found in chemical and polymer
plants.  The data, therefore, are deemed applicable to those equipment types
found in styrene-butadiene latex units.  However, in order to allow further
consideration of the processing equipment in various polymer and resin
manufacturing units, styrene-butadiene latex (along with several other
polymers and resins) are not included in the scope of this CTG.
     In addition to these considerations made for polymers and resins, EPA has
further evaluated coverage of the CTG since the draft document was released
for comment.  Methyl tert-butyl ether (MTBE) is a relatively new, high-growth
organic chemical that has gained prominence as a gasoline additive, replacing
lead-based additives.  MTBE was not produced in large quantities commercially
when the SOCMI list of organic chemicals was originally composed.  Because
MTBE is a large volume organic chemical with a high growth rate and because it
is produced in the same synthetic organic chemical plants currently covered by
the CTG, MTBE is being added to the list of organic chemicals covered by the
CTG.

A.2  ESTIMATES OF EMISSIONS, EMISSION REDUCTIONS, AND COSTS

Comment:  Commenters [#2; #5; #6; #7; #9; #13; #15; #16; #17; #18;, #19; #20;
#21; #22] objected to the application of fugitive emissions data collected in
petroleum refineries to SOCMI and polymer processes.  The commenters said that
                                     A-6

-------
fugitive emissions from SOCMI are lower in terms of both leak frequency and
mass emission rate.  Several reasons for the differences were given including
smaller unit sizes, lower temperatures and pressures, more expensive products,
more toxic products, and chemicals whose leaks are self-sealing such as
polymers.  They referred to fugitive emissions data generated in SOCMI as
evidence of the lower emissions from SOCMI processes when compared to
petroleum refineries.  Commenters said that the data showed the differences
between the industries and recommended the use of SOCMI data exclusively.
     Commenters [#3; #5; #9; #13; #16; #17; #20; #22] further objected to the
use of refinery data on technical grounds.  They cited differences in
calibration gases and screening instruments used in studies of the industry
and differences in response factors of different chemicals as reasons that
data generated in petroleum refineries should not be considered relevant to
SOCMI.

Response:  EPA's analysis of fugitive emissions data is extensively documented
in Fugitive Emission Sources of Organic Compounds — Additional Information on
Emissions, Emission Reductions, and Costs (EPA-450/3-82-010, April 1982).  As
	
the Additional Information Document (AID)  relates, EPA reviewed all available
fugitive emissions data from SOCMI as well as from petroleum refineries.  EPA
determined that the best studies on which emission estimates for SOCMI
                                                                  3
emission sources could be based were the Refinery Assessment Study  and the
                             4
SOCMI Twenty-four Unit Study.   EPA considered these data sets to show
differences between the SOCMI data and the petroleum refinery data.  The
assessment of differences and similarities between the data sets was not
clearcut.  There were some apparent differences, but they could not be
explained conclusively.  The differences may be due to factors mentioned by
the Commenters.  It is impossible to tell because there are so many variables.
It seemed illogical that on the average, identical equipment handling similar
organic compounds would behave differently.  However, EPA determined that the •
differences, as indicated by the data, were evident.  Because of the
differences, EPA decided that an adjustment of the emission factors used
previously was warranted.
                                     A-7

-------
     After considering alternative approaches, EPA concluded that the best
method of arriving at a complete set of emission factors for equipment leaks
was by using leak frequencies determined in the SOCMI 24-Unit Study to weight
the emission factors determined in the Refinery Assessment Study.  The
resulting emission factors are presented in Chapter 2.
     The technical considerations cited by the commenters refer to the
monitoring instruments, calibrants, and procedures used in the different
studies of fugitive VOC emissions.  These considerations are examined in
detail in the AID.  The differences in measurement methods and response
factors cited by the commenters were considered by EPA and were not found
significant.  The variability seen in repeat sampling of the same source was
           5
23 percent.   This variability is in the same range as the 30 percent
difference seen in response between the TLV-hexane system and the OVA-methane
systems at the 10,000 ppmv action level.   Because the variability in repeat
sampling is so similar to the differences in response at 10,000 ppmv, the data
can be used interchangeably within ±30 percent at the action level.
     Furthermore, laboratory experiments measuring variation in response
factors for a number of organic chemicals indicated that 90 percent of the
                                                  789
chemicals tested had responses between 0.1 and 10. ' '   When considered in
analyzing leak frequencies,   the response factor variation, however, did
not product significant changes in the overall percent leaking estimates
resulting from the SOCMI 24-Unit Study.
     As presented in Section 4.1, RACT requires the use of a VOC detection
instrument and monitoring method in accordance with EPA Reference Method 21 or
an equivalent State method.  An instrument reading of 10,000 ppmv is used as
the definition of a leak.  Soaping is permitted for some sources as a
prescreening tool, but this technique, where applicable, must be supplemented
with instrument screening if leaks are indicated using soaping.  Soaping is an
additional  element of RACT beyond that presented in the draft CTG.

Comment:  Commenters [#9; #14; #17; #19] disagreed with control efficiency
estimates for leak detection and repair programs presented in the draft CTG.
Referring to the ABCD model calculations, commenters said that the occurrence
                                     A-8

-------
and recurrence function used to derive the B-factor was not supportable and
should be revised.  They recommended a time series model  which could be used
in calculating a B-factor.  They also felt that the repair effectiveness
(D-factor) should be lower than the figure used in the draft CTG.

Response:  The control efficiencies presented in the draft CTG for leak
detection and repair programs were estimated using the ABCD model.  These four
factors, when taken together, considered the maximum emissions reduction
potential (A) and accounted for other factors such as delay of repair (C),
emissions reduction to a non-zero emissions ;level  (D) and the occurrence and
recurrence of leaks and the number of non-repairable leaks between monitoring
inspections (B).  Of these factors, the B-factor involved the most subjective
consideration; the selection of the value for the  B-factor was based on the
engineering judgement that the rates for occurrence/recurrence/non-repairable
leaks were non-linear with respect to monitoring interval.
     The commenters are partially correct in stating that occurrence rates
should be linear.  Occurrence rates have been found to be essentially linear
in the studies of fugitive emissions reported in the Maintenance Study.
In this report, the leak occurrence rate is Described by  an exponential
distribution model and the leak recurrence rate is described by a mixed
distribution model, which incorporates an exponential model to describe
long-term leak recurrences.  Both models are non-linear in format.  But, as
applied to the data collected in these studies, the models result in a nearly
linear relationship with time.  In fact, only slightly non-linear leak
occurrence and recurrence rates for valves are noted when considering a
monitoring interval of one year.            [
     Analysis of the results of the Maintenance Study led to the development
of a new model describing leak detection and repair programs.'  This model  is
                                       12
described in detail in a Technical Note   and in the AID.  The Leak
Detection and Repair (LDAR) model is based on a set of recursive equations
describing leaks from equipment in terms of four categories:
                                     A-9

-------
      (1)  Nonleaking equipment (screening < action level),
      (2)  Leaking equipment (screening >^ action level),
      (3)  Leaking equipment which cannot be repaired on-line and are
          awaiting a process unit shutdown for repair, and
      (4)  Repaired equipment that exhibit early leak recurrence.
Three emission rates describe these four categories; a single emission rate is
used  to describe equipment in the last two categories listed.
      In describing these various categories, the LDAR model requires more
information than the ABCD model.  This information includes repair
effectiveness, emissions reduction for successful repair, and emissions
reduction for unsuccessful repair.  These data are available for pumps in
light liquid service and for valves in gas and light liquid service; their
selection is detailed in the AID and summarized in Chapter 3.
     The LDAR model is the preferred predictor of leak detection and repair
effectiveness where the detailed information is available because the
resulting estimates are based on experimental data rather than engineering
judgment alone.  But the ABCD model remains a viable method of estimating the
effectiveness of programs for equipment types for which these data are not
available.  To refine the ABCD estimate for such equipment types, the results
of the LDAR model for comparable equipment types may be applied.  For example,
detailed leak occurrence and repair data have not be generated for safety/
relief valves in gas service.   By comparing LDAR results to ABCD results for
valves in gas services, the ABCD results for safety/relief valves can be
adjusted to yield a refined estimate for leak detection and repair programs
applied to safety/relief valves.  The AID discusses this refinement procedure
for safety/relief valves in additional  detail.   This is the same approach
taken in estimating the effectiveness of leak detection and repair for
compressors.a
Effectiveness, = ABCD1  x  (LDAR Effectivness)9
              1       ~
For safety/relief valves in gas service, comparisons were made with results
for valves in gas service.  Comparisons to pumps were used to adjust ABCD
results for compressors due to similarities in sealing mechanisms.
                                     A-10

-------
Comment:  Two commenters [#9; #14] specifically cited fugitive emission
testwork in supporting their comments on the occurrence rate and recurrence
rate assumed in computing the control efficiency of leak detection and repair
techniques.  One commenter said that for valves the occurrence rate varies
with the leak frequency and that the data in the SOCMI studies are biased to
the high side of the leak frequency spectrum.  Recurrence of leaks was
estimated, according to the commenter, using an extremely sparse data set,
resulting in a recurrence rate of questionable utility.  To support the same
claims, the other commenter cited values determined in a fugitive emission
study in a high density polyethylene plant.

Response:  Occurrence and recurrence of leaks was embodied in the B-factor of
the ABCD model for fugitive emissions.  An improved model, the LDAR model, is
now the basis for estimates of emissions and emission reductions for valves
and pumps operating under leak detection and repair programs.  However, the
LDAR model requires data which are unavailable for some other equipment types.
As discussed in the previous response, the LDAR model is used in conjunction
with the ABCD model for those sources (compressors, safety/relief valves).
     The inputs used for occurrence and recurrence in the LDAR for pumps and
valves were documented and explained in detail  in the AID.  EPA chose the best
values available for these input parameters.  Occurrence rate estimates for
valves were available from two studies.  First, the Maintenance Study had
occurrence rate estimates developed from tests in three SOCMI processes.
Estimates were presented for each type of process (vinyl acetate, cumene, and
ethylene) and by service (gas, light liquid).  An overall estimate for all
                                                        13
units was also developed.  Second, the Allied HOPE Study   presented
occurrence rate estimates for valves in a high density polyethylene unit.  Due
to some inconsistencies noted in this study and due to the broader range of
processes covered by the Maintenance Study, occurrence rates generated in the
Maintenance Study were considered to be the best available estimates of
occurrence rates for valves.  Because the confidence intervals for the
occurrence rates for individual process units showed substantial  overlap
(i.e., the occurrence rates for the process units were not significant
                                    A-ll

-------
different), the overall 30-day occurrence rate of 1.3 percent was selected as
an input to the LDAR model.  The Maintenance Study showed that about 14
percent of all repaired valves started to leak again within 5 days of repair.
The only other recurrence rate data is from the Allied HOPE Study.  However,
that study does not provide information for early failures.  Therefore, early
leak recurrence rate data from the Maintenance Study were used for input to
the model.
     As shown in the AID, the only occurrence rate data available for pumps
are from the Maintenance Study.  This occurrence rate was adjusted to account
for pump seal replacement which normally occurs.  The resulting 30-day
occurrence rate input is 3.4 percent.  Because leaking pump seals are usually
taken off-line and replaced with new seals, a successful repair rate of
100 percent was used, and the early leak recurrence rate was taken as 0.  The
leak recurrence rate equals the leak occurrence rate.

Comment:  Several comments [#3; #4; #6; #9] were received regarding the
estimated costs associated with RACT requirements.  Several comments concerned
increased cost effectiveness estimates resultant from increased costs and
decreased emission reduction estimates.  One commenter stated that the capital
costs estimated for RACT did not include the costs of initial survey
inspections and repair.  Other commenters felt that monitoring time estimates
for valves were underestimated and that this time did not include preparation
time and travel time between sources.  Another commenter cited several causes
of the "unrealistic" cost estimates:  a low interest rate, a low overhead
charge, and an underestimate of valve size.

Response:  The costing methodology for controlling equipment leaks of VOC was
reviewed in detail in the AID.  Costing techniques and cost assumptions were
discussed for equipment control techniques as well as for leak detection and
repair programs.   The methodology presented in the AID has been applied to the
cost estimates presented in Chapter 5.  The revised estimates of costs are
higher than those presented in the draft CTG.  The annualized costs of RACT
for model unit B presented in Table A-2 result from use of these assumptions.
                                    A-12

-------
                               TABLE A-2.   COSTS  FOR  COMPONENT  TYPES IN MODEL UNIT B
*-*
co

Component
Pumps (Light Liquid)
Valves
Gas Service
Light Liquid Service
Safety/Relief Valves
Open-ended Lines
Compressors
TOTAL UNIT (Without
Instrument Cost)
TOTAL UNIT (With
Instrument Cost)
Number Of
Components
29

402
524
42
415
2
1,414
1,414
Annual ized
Cost Before
Credit ($)a
6,670

3,670
4,650
1,130
5,560
890
22,570
28,050
Annual
Recovery
Credit ($)
1,750

5,400
6,140 '
7,260
2,650
560
23,760
. 23,760
Net
Annual ized
Cost ($)a
4,920

(1,730)
(1,490)
(6,130)
2,910
330
(1,190)
4,290
Total VOC
Reduction
(Mg/yr)
4.08

12.6
14.3
16.9
6.18
1.31
55.4
55.5 -
Cost
Effectiveness
($/Mg)
1,200

(140)
(100)
(360)
470
250
(21)
" - 7-7 .
        Does not include cost of monitoring instrument, unless otherwise noted.
        The net cost associated with monthly monitoring of pumps is lower than the net cost shown for
        quarterly monitoring of pumps.  The lower net cost associated with monthly monitoring results
        from higher emission reductions and, therefore, higher recovery credits.  The following cost
        figures are applicable to monthly monitoring for pumps:

            Annualized cost before credit  =  8,439
              (gross cost) ($)
            Emission reduction (Mg/yr)     =  7.6
            Annual recovery credit ($)     =  3,277
            Net annualized cost ($)        =  5,162
            Cost effectiveness ($/Mg)      =  680
        "Cost for caps on lines only.  Not monitored under RACT.

-------
Also, the cost effectiveness, or the ratio of the annualized control cost to
the emissions reduction achieved, is presented in Table A-3, along with the
estimated control efficiency of each selected RACT,  Although these costs are
higher, they are reasonable costs and cost effectiveness of control.
     As in the draft CTG, where leak detection and repair programs have been
selected as RACT, the cost of the initial screening and repair of leaking
equipment components has been capitalized.  These costs are assumed to be
amortized at 10 percent interest over a ten year period (a 2-year period is
used for replacement seals).  A six-year amortization period was used for
capitalizing monitoring instruments.  The 10 percent interest rate used is
conservative in that it represents a real rate of return after taxes and not
merely a typical interest rate.
     The monitoring time estimates for valves were examined in the AID.  The
2 man-minutes per valve monitoring time used was based on information provided
by Exxon Company, USA.    It is a process unit-wide average value and is the
most reasonable estimate available in the absence of data to the contrary.
     The 40 percent overhead rate was found to be low by some commenters who
suggested 100 percent would better reflect an overhead charge.  The labor
charge of $18 per hour includes a 40 percent charge for labor-related adminis-
trative and overhead costs.  An additional 40 percent rate is applied to the
$18 per hour rate to account for the administrative and support costs
associated with implementation of RACT.  These two charges taken together
amount to a cumulative 96 percent total overhead charge rate.
     A one-inch valve size was selected as the basis of the capital costs for
control of open-ended lines.  A survey of the data on which the model units
were based showed that approximately 92 percent of the valves in the process
units surveyed were two-inch diameter or smaller.  Moreover, the one-inch
valve size was used to estimate the cost of controlling open-ended lines only.
The costs of controlling emissions from open-ended lines are based on
installing a second valve.  Plugs, caps, or blind flanges are also expected to
be used to control  open-ended lines; the costs of these materials are similar
or less than the costs of one-inch valves.  In most cases, therefore, the
control cost based on these equipment would be much lower than estimated for
use of a second valve.
                                     A-K

-------
                        TABLE A-3.  EQUIPMENT LEAKS OF VOC FROM SYNTHETIC ORGANIC CHEMICAL AND
                                              POLYMER MANUFACTURING:  RACT
             Equipment
               Type
                              Control Technique
Average Control
Effectiveness,
    Percent
Average Cost
   jctivc
   i cre<
    $/Mg
                                                                                  Effectiveness.
                                                                                 (with  credit).
         Pumps  (light  liquid
               service)
                        Quarterly  leak  detection &  repair
         Valves  (gas and light   Quarterly leak detection & repair
                 liquid service)
      33
                                                                   51
   1,200
                         (H4)(
i
i_»
en
Safety/relief valves
       (gas service)

Open-ended lines
                                Quarterly leak detection & repair
                                Equipment
      44
     100
    (360)c


     470
         Compressors
                        Qua rterly leak  detect ion &  repa i r
      33
     250
          Control effectiveness of leak detection and repair programs is based on results of LDAR model,
          either  solely or in combination with the ABCD model results.

          Costs are 2Q1980 and assume $429/Mg recovery credit for VOC saved or recovered as a result of
          implementing the control technique.

         "Parentheses indicate net savings resulting from the credit value of the saved/recovered VOC.

-------
A.3  RACT SELECTION, PROVISIONS, AND EXEMPTIONS

Comment:  Commenters [#4; #5; #9; #14-, #19] called the 2 percent good
performance level arbitrary and said that it was too low for RACT.  Values
offered as more realistic levels of good performance were 4 percent and 10 to
12 percent.  It was also recommended that flexibility be offered plants in
setting good performance levels.

Response:  As discussed in previous responses and in Chapter 3, estimating the
effectiveness of leak detection and repair in reducing VOC emissions resulting
from equipment leaks has been facilitated for some equipment types by the LDAR
model.  The LDAR model has been used to examine the costs and effectiveness of
leak detection and repair for valves exhibiting varying initial leak frequen-
cies (see Section 3.2.1).  As shown in Figure 3-1, the cost effectiveness of
quarterly leak detection and repair for valves become unreasonable around
1 percent leaking initially in a process unit.  As discussed in Chapter 3, a
performance level of 2 percent leaking would ensure that most units would be
achieving around 1 percent leaking.  Therefore, an alternative to periodic
leak detection and repair for valves could be a performance level  of 2 percent
leaking in a process unit.
     The selection of a performance level based on the percentage of valves
leaking was not a question of technical achievability of such a performance
level.  The selection was based on the high cost effectiveness associated with
routine (quarterly) leak detection and repair of valves in model units
exhibiting low leak frequency.  This type of alternative standard not only
allows low-leak units an exemption from routine monitoring where it is not
cost-effective, but it also provides an incentive to units exhibiting higher
leak frequencies to attain the performance level by means of installing better-
equipment or improving their current maintenance practices.
     In additions skip-period monitoring plans, discussed in Section 3.2.2,
may provide another mechanism for achieving a performance level at minimal
monitoring.  Under such plans, monitoring requirements can be minimized for
individual process units that seek to maintain a given performance level  of
percent of valves leaking.
                                     A-16

-------
Comment:  One commenter [#20] stated that the CTG should relate only to photo-
chemical ly reactive VOC.  And priorities should be established to control  only
those substances determined to be photochemically reactive based on the
documented differences in photodegradation rates and ozone yield of various
organic compounds.  Other comments [#5; #9] referred to the definition of VOC
given in the proposed NSPS.  One commenter stated that the definition included
nonphotochemically reactive organic compounds that do not contribute to ozone
formation.  Another commenter felt that the definition should be clarified and
rely on Reference Method 21 in determining if a compound should be considered
a VOC.  Moreover, he felt that the limitations of the detection instruments
should be accounted for in the definition of VOC.

Response:  Volatile organic compounds (VOC) are any organic compounds which
participate in atmospheric photochemical reactions.  At present, the
Administrator has identified only the following organic chemicals as
nonreactive organic chemicals:
       •  methane
       o  ethane
       •  1,1,1-trichloroethane
       •  methylene chloride
       •  trichlorofluoromethane
       »  dichlorodifluoromethane
       »  chlorodifluoromethane
       t>  trifluoromethane
       •  trichlorotrifluoroethane
       «  dichlorotetrafluoroethane
       •  chloropentafluoroethane
The RACT requirements discussed in this CTG are applicable to equipment that
are "in VOC service," which is defined as containing at least 10 percent VOC
by weight.  In determining whether a piece of equipment is "in VOC service,"
the organic chemicals listed above as nonreactive organic compounds may be
excluded from the total VOC determined by the appropriate reference methods.
                                    A-17

-------
     The definition of "in VOC service" is a means of determining if a piece
of equipment is subject to RACT requirements.  Once the determination is made
that a piece of equipment is "in VOC service," the requirements of RACT stand
alone and do not rely on the definition of VOC.

Comment:  Several commenters [#5; #9; #10; #20] took exception to the
selection of 0.3 kPa as the vapor pressure breakpoint separating light liquid
and heavy liquid services.  They felt this selection was arbitrary and that
other choices would have more relevance to the chemical industry.  The
commenters presented alternative choices based on:

     (1)  the vapor pressure at 20°C corresponding to the concentration
          equivalent of the leak definition (action level);
     (2)  the split between gasoline and kerosene (1.5 psia or 10 percent of
          the ASTM distillation point); and
     (3)  the vapor pressure at operating conditions (with light liquids
          defined below the vapor pressure of 760 mm Hg at operating
          conditions).

Response:  EPA's analysis of fugitive emission rates and vapor pressures has
shown that substances with vapor pressure of 0.3 kPa and higher have
significant emission rates while those with lower vapor pressures are not as
significant.  This vapor pressure (0.3 kPa) represents the split between
kerosene and naphtha and is the criterion used by EPA to distinguish between
light liquid and heavy liquid substances.  The split was made to concentrate
effort in a leak detection and repair program on the sources with the largest
potential to leak.

Comment:  Various comments [#6; #18] dealt with the selection of 10,000 ppmv
as the leak definition.   One commenter felt 10,000 ppmv was a satisfactory
leak definition for all  sources, except valves in gas/vapor service.   He
recommended 100,000 ppmv for gas valves.  Another commenter suggested that
100,000 ppmv should be used for all  sources based on an improved cost

-------
effectiveness.  The commenter stated that the control efficiency of leak
detection and repair techniques would be nearly the same for leak' definitions
of 100,000 and 10,000 ppmv.  Other commenters felt that 10,000 ppmv was too
low a leak definition and that the definition should be based on a mass
emission rate equivalent.                  ;

Response:  One consideration in selecting 10,000 ppmv as the leak definition
for equipment leaks of VOC was the monitoring instrument characteristics.
Data on which the CTG is based were collected using hydrocarbon detectors  that
are readily available.  These instruments provide a direct measurement of
organics concentrations up to 10,000 ppmv; in order to measure higher
concentrations with the instruments most commonly used, additional care and
calibration for devices such as dilution probes are required to obtain
reliable results.  And as a result, additional costs are associated with
measuring concentrations higher than 10,000 ppmv.  Although instruments that
directly measure higher concentrations of organics may be available in the
future, the monitoring requirements are based on the least complicated and
best established portable hydrocarbon detection technique currently available.
     Table A-4 presents a summary of the percent of sources screening above
the action level (leak definition) for various action levels (an indication of
the number of leaks) and percent of mass emissions attributable to these
action levels for valves.  Analysis of the results from the Maintenance Study
demonstrates that a significant quantity of mass emissions would be detected
with an action level of 10,000 ppmv instead of 100,000 ppmv for the SOCMI
sources tested.  In addition, an analysis of leak detection and repair
programs based on 10,000 ppmv and 20,000 ppmv action levels indicate that
improved cost effectiveness and greater emissions reduction is associated  with
                        15
the 10,000 ppmv program.    EPA sees the opportunity to control these leaks
as a significant opportunity for cost-effective emission control.
Comment:  The monitoring requirements of the draft guidelines were said to be
overly restrictive and excessive [#3; #6; #9; #18].  Several  commenters
recommended annual monitoring instead of the quarterly scheme presented in the
                                    A-19

-------
                          TABLE A-4.  SUMMARY OF PERCENT OF SOURCES DISTRIBUTION CURVES AND
                             PERCENT OF MASS EMISSIONS CURVES AT VARIOUS ACTION LEVELS3
ro
o

Valves
Gas
Ethyl ene
Cumene
Vinyl Acetate
Light Liquid
Ethyl ene
Cumene
Vinyl Acetate
Pump Seals
Light Liquid
Ethyl ene
Cumene
Vinyl Acetate
Percent
10,000


15
16
3.7
26
12
0.2
30
14
1.7
of Sources Screening Above
20,000


12
13
2.8
22
9
0.1
24
11
1.0
40,000


10
10
2.0
18
6
0.1
18
8
0.5
100,000


7
6
1.2
13
4
0
12
5
0.2
Percent of Mass Emissions Attributable
to Sources Screening Above '
10,000


94
94
90
89
80
25
96
89
67
20,000


90
89
84
83
71
16
92
83
57
407000


84
83
77
75
61
10
86
75
46
100,000


71
69
62
60
45
4
73
61
31
          Curves  are  based  on  models derived from data collected during 24-unit SOCMI  study.

         ""Screening values  in  ppmv.

         'These values were based  on the original leak rate/screening value correlations presented  in the
          Maintenance Study and  have not been changed to reflect the new correlations  developed in  the
          Technical Note  on the  revision of SOCMI emission factors.  Based on a comparison of empirical
          data, these values are not expected to change significantly.

-------
draft CTG.  They based their recommendation on improved cost effectiveness  of
an annual monitoring plan.  One of these comments suggested that for valves
monitoring frequency could be increased for units where annual  monitoring
failed to achieve an allowable percentage of valves leaking.

Response:  Section 4.1 presents RACT for equipment leaks of VOC in synthetic
organic chemical and polymer manufacturing plants.  RACT procedures include
quarterly leak detection and repair of pumps in light liquid service, valves
in gas service, valves in light liquid service, safety/relief valves in gas
service, and compressors.  However, states may choose to implement monthly
monitoring for pumps because the cost effectiveness is more attractive than
the cost effectiveness for quarterly monitoring.  RACT also includes installa-
tion of plugs, caps, blind flanges, etc. for open-ended lines.   The environ-
mental impacts of RACT are presented for model units in Chapter 4 and the cost
impacts are given for model units in Chapter 5.  The costs, emission
reductions, and cost effectiveness of RACT are reasonable.
     Alternative programs for monitoring valves are also allowed as RACT
requirements.  Under such programs, RACT can be met by meeting a performance
level of 2 percent leaking in a process unit.  This provision allows specific
programs to be tailored to individual process units,.provided an annual
performance test demonstrates 2 percent or less leaking.  Another alternative
program for valves allows implementation of skip-period monitoring techniques;
these programs are also discussed in Chapter. 3.  Either of these approaches
has the .potential to reduce monitoring frequency and cost of valve leak
detection and repair in individual process units.
     Quarterly leak detection and repair has also been retained as the basis
of RACT for safety/relief valves in gas service (see response to comment later
in this appendix) and for compressors.  Since RACT applies to existing
compressors, EPA believes additional provisions should be considered for
compressors.  Leak detection and repair may not always be an effective
technique for compressors.  For instance, leak detection and repair is not
applicable if the compressor seal cannot be repaired below the action level
                                    A-21

-------
(10,000 ppmv) or cannot be repaired on-line (i.e., the seal can only be
repaired during a process unit shutdown).  If leak detection and repair is not
applicable to an existing compressor, equipment should be installed as the
control technique.  An example of the equipment to be used is a mechanical
seal system with a non-VOC barrier fluid and degassing reservoir connected to
a control device (e.g., flare).  Another example of equipment is a seal area
enclosure that is vented to a control device.  In some instances, neither leak
detection and repair nor equipment are feasible due to prohibitive costs or
safety considerations.  Under these circumstances, a waiver from the RACT
requirements could be considered on a case-by-case basis.

Comment:  Referring to comments submitted on the NSPS, commenters [#16; #17;
$21; 122] stated that emissions and emissions reduction potential were lower
than presented because of the current use of flares.  They cited a study of
flares by Siege!  to support their contention that flares can achieve 99+
percent destruction of VOC.

Response:  Flares have not been presented in the CT6 as a control device for
destroying VOC collected from various sources.  The CT6 focuses on the
application of leak detection and repair for reducing emissions rather than
equipment.  As the commenters state, however, flares are effective in
eliminating VOC emissions for certain equipment types.  For example, VOC
emitted through the seat of a safety/relief valve are effectively eliminated
if the discharge of the safety/relief valve is vented to a control device,
such as a flare.   Where flares are used to control VOC emissions from
safety/relief valves, there is greater potential for emissions reduction from
the uncontrolled leak rate than can be achieved through leak detection and
repair techniques.   Flares are also effective in eliminating VOC emissions
from pump seals and compressor seals when used in combination with mechanical
seal/barrier fluid/degassing reservoir systems.  Thus, flares are allowed for
control of equipment leaks from pumps, compressors, and safety/relief valves
in Lieu of instrument monitoring.
                                     A-22

-------
Comment:  Commenters [#7; #9; #10; #15] noted that EPA had made allowances for
inaccessible and unsafe-to-monitor valves.  However, they asked that further
consideration be given to exempting unsafe-to-monitor and inaccessible valves
completely from periodic monitoring requirements.

Response:  EPA's view is that inaccessible and unsafe to monitor valves should
be monitored as often as practicable because of the potential for finding
leaks and reducing emissions.  EPA does not consider annual monitoring or
monitoring at shutdown to be an unreasonable burden for inaccessible and
unsafe to monitor valves.  However, as indicated in Section 3.3.1, the
difficulties of monitoring inaccessible and unsafe to monitor valves should be
considered.  For example, difficult-to-monitor valves might be exempted from
routine quarterly monitoring provided they are monitored annually.  The extent
of the consideration is left to the discretion of the state and local agencies
administering regulations based on leak detection and repair programs.

Comment:  Two commenters [#4; #7] expressed concerns with different aspects of.
safety/relief valves.  One commenter felt that monitoring of safety/relief
valves was unwarranted since serious injury could result if a safety/relief
valve should relieve while being monitored.  The commenter said that such
monitoring presented an undue safety hazard to personnel.  Another commenter
discussed the use of block valves upstream of safety/relief valve.  Acknow-
ledging their existence in the industry, the commenter stressed that such
valves are typically locked open.  Furthermore, in his plant, only authorized
personnel could unlock this kind of block valves and the personnel must remain
with the block valve until it is again locked open. ,

Response:  EPA has examined the monitoring requirements for safety/relief
valves in gas service and does not consider the quarterly monitoring require-
ments to be burdensome or unwarranted.  Safety/relief valves are routinely
inspected as a part of normal safety and maintenance procedures to ensure the
set-pressure is correct.  The quarterly monitoring requirement of RACT may
increase the frequency of this ordinary monitoring practice, but the
                                     A-23

-------
precautions taken during current safety/relief valve inspections are expected
to be used during quarterly monitoring.
     The intent of the RACT selected for safety/relief valves is to eliminate
the large amounts of VOC that can be emitted through the valve if it does
not seat properly after an overpressure release.  Therefore, as part of any
emissions reduction program for safety/relief valves, EPA believes that
monitoring should follow every overpressure relief within 5 days of the
relief.  This is to ensure the valve has reseated properly.
     EPA also considered the existence of current systems in use in the
industry.  For example, many safety/relief valves are already connected to
closed vent systems (e.g., flare headers) for safe disposal of emergency
release gases.  Under such a system, there would be no required monitoring.
Some process units do have block valves installed upstream of safety/relief
valves, as one commenter described.  While this is not recommended practice,
it is an acceptable procedure under engineering standards.  An improvement
over this arrangement is the use of a Y-valve with parallel relief systems.
This arrangement ensures a safety/relief system is in-service at all times and
allows ready repair of one of the safety/relief valves.

Comment;  The 15 day interval allowed for delay of repair was said to be too
short, especially in those cases where repair parts had to be ordered [14;
#9],  One commenter said that 30 to 45 days should be allowed in such cases to
obtain parts.  And in commenting on the NSPS, commenters requested a delay in
repair to the next process unit shutdown for repair where parts had to be
ordered.  This would provide time to obtain repair materials and to schedule
maintenance work.

Response:  Delay of repair for leaking sources can significantly impact
emission reductions achievable under leak detection and repair programs.  EPA
expects most on-line/in-place repairs to'be effected quickly.  The require-
ments of RACT allows a 15-day repair interval to provide time for those
technically feasible repairs to be made; a 15-day interval provides ample time
for such repairs without sacrificing a large amount of emissions reduction.
                                    A-24

-------
For the remaining leaks that are not technically feasible on-line or in-place,
delay of repair is allowed to the next process unit turnaround.

Comment:  Several commenters [#6; #10; #14] objected to the statement that
State enforcement officers might request a unit with an excessive numbers of
leaks to shutdown before their scheduled shutdown.  They said that shutdowns
could cause more emissions than allowing the leaks to continue and that the
shutdowns could result in excessive energy use.

Response:  The intent of requesting a process unit shutdown for repair of an
excessive number of leaks is to promote the use of sophisticated repair
techniques (such as sealant injection) in process units with demonstrated
excessive leaks.  Certainly, any decision to request a process unit shutdown
for repair of an excessive number of leaks prior to a scheduled shutdown must
be carefully considered, taking into account the potential costs of an early
shutdown.  Similar provisions for early shutdown have previously been
presented in a CTG model rule for petroleum refining fugitive VOC emissions
(EPA-450/2-79-004).    In lieu of requesting an unscheduled process unit shut-
down to repair an excessive number of leaks, State and local agencies may
consider including specific provisions for delay of repair of various equip-
ment types. -Under this approach, a delay of repair beyond the repair interval
(15 days) would only be allowed if repair is technically infeasible without a
process unit shutdown and if spare parts for repair have been depleted (after
being sufficiently stocked).  Records of the reasons for delay of repair could
then be used to aid State enforcement officers in determining compliance.

Comment:  Several comments [#5; #6; #9; #12; #14; #18] were received asking
for exemptions.  Exemptions were requested for small production quantities,
for units with few fugitive emissions sources, and for small diameter lines
and valves.

Resp.onse:  State and local control agencies may wish to include exemptions
for plants or process units.  Exemptions would most likely be designed to
                                    A-25

-------
prevent high cost effectiveness ratios for VOC control.  As discussed in
Chapter 3, such exemptions might be based on quantities of light liquid and
gaseous VOC processed.  An exemption based on the number of equipment compo-
nents in a process unit is another possibility, but it is more appropriately
addressed in terms of quantity of light liquid and gaseous VOC processed.
Other suggested exemptions are for equipment in vacuum service and for process
units processing only heavy liquids or non-VOC.  EPA has no documentation  of
fugitive emission rates varying with line size.  Thus, there is no justifica-
tion for an exemption from monitoring requirements based on line size.

Comment:  The recordkeeping and reporting requirements of the draft CTG were
said to be excessive [#3; #9; #16].  Further, the time estimated to handle
these tasks was found to be insufficient.  One commenter felt that tagging and
logging all leaks was unjustified and cost-ineffective, especially where
on-the-spot repairs are successful.  Another commenter, however, stated that
the draft CTG did not provide any discussion of the reporting and record-
keeping requirements.

Response:  EPA sees no way of implementing and administering leak detection
and repair programs without some recordkeeping.  The level of reporting and
recordkeeping the state and local  air pollution agencies will require has  not
been discussed in the CTG.  However, an allowance was made for recordkeeping
and reporting in the cost analysis.
     Tagging and logging equipment that cannot be repaired on-'!ine/in-place is
an effective means of handling those components that must await a process  unit
shutdown for repair.  Such recordkeeping is a necessary tool in establishing
alternative leak detection and repair programs, such as a percent leaking
requirement or a skip-period monitoring plan.  Furthermore, this type of
recordkeeping would be beneficial  to State enforcement officers considering a
request of unscheduled process unit shutdown for repair of an excessive number
of leaks.  For these reasons, records should be maintained of all  leaks.   In
the .case of effective on-the-spot  repair, tagging is not considered productive
for the quarterly leak detection and repair programs selected as RACT; for the
reasons cited above, however, maintaining records is necessary.
                                     A-26

-------
A.4  REFERENCES

1.   Memorandum from Stalling, J. H. E., Radian Corporation, to SOCMI
     Fugitive NSPS file.  November 1, 1982.  11 p.  Estimates of VOC
     emissions from SOCMI.

2.   U. S. Environmental Protection Agency,  fugitive Emission Sources of
     Organic Compounds -- Additional Information on Emissions, Emission
     Reductions, and Costs.  Research Triangle Park, N.C.  Publication No.
     EPA-450/3-82-010.  April 1982.

3.   Wetherold, R. 6., L. P. Provost, and C. D. Smith.  (Radian
     Corporation.)  Assessment of Atmospheric Emissions from Petroleum
     Refining, Appendix B:  Detailed Results.  (Prepared for U. S.  Environ-
     mental Protection Agency.)  Research Triangle Park, N.C.  Publication
     No. EPA-600/2-80-075c.  April 1980.

4.   Blacksmith, J. R., et al.  (Radian Corporation.)  Problem Oriented
     Report:  Frequency of Leak Occurrence for Fittings in Synthetic Organic
     Chemical Plant Process Units.  (Prepared for U. S. Environmental
     Protection Agency.)  Research Triangle Park, N.C.  Publication No.
     EPA-600/2-81-003.  September 1980.

5.   Langley, G. J. and R. 6. Wetherold.  (Radian Corporation.)  Evaluation
     of Maintenance for Fugitive VOC Emissions Control.  (Prepared  for
     U. S. Environmental Protection Agency.)  Research Triangle Park, N.C.
     Publication No. EPA-600/52-81-080.  September 1980.

6.   DuBose, D. A. and G. E. Harris.  (Radian Corporation.)  Response
     Factors of VOC Analyzers to a Meter Reading of 10,000 ppmv for Selected
     Organic.Chemicals.  (Prepared for U. S. Environmental Protection
     Agency.)  Research Triangle Park, N.C.  Publication No.
     EPA-600/2-81-051.  March 1981.

7.   Reference 6.

8.   Reference 6.

9.   DuBose, D. A., G. E. Brown and G. E. Harris.  (Radian Corporation.)
     Response of Portable VOC Analyzers to Chemical Mixtures.  (Prepared  for
     U. S. Environmental Protection Agency.)  Research Triangle Park, N.C.
     Publication No. EPA-600/2-81-110.  June 1981.

10.  Langley, G. J., et al.  (Radian Corporation.)  Analysis of SOCMI VOC
     Fugitive Emissions Data.  (Prepared for U. S. Environmental  Protection
     Agency.)  Research Triangle Park, N.C.  Publication No.
     EPA-600/2-81-111.  June 1981.

11.  Reference 5.
                                      W27

-------
12.  Williamson, H. J., et al.  (Radian Corporation.)  Model for Evaluating
     the Effects of Leak Detection and Repair Programs on Fugitive
     Emissions.  Technical Note DCN 81-290-403-06-05-03.  September 1981.

13.  Harvey, C. M. and A. C. Nelson.  (PEDCo Environmental, Inc.)  VOC
     Emission Data - High Density Polyethylene Process Unit.  (Prepared for
     U. S. Environmental Protection Agency.)  Research Triangle Park., N.C.
     Publication No. EPA-600/2-81-109.  June 1981.

14.  Letter and attachments from Johnson, J. M., Exxon Company, to
     Walsh, R. T., EPA:CPB.  July 28, 1977.  14 p.  Review of "Control of
     Hydrocarbon from Miscellaneous Refinery Sources" draft report.

15.  Memorandum from Stelling, John, Radian Corporation, to SOCMl Fugitives
     NSPS File.  November 1, 1982.  4 p.  Cost effectiveness of leak
     detection and repair programs using different-leak definitions.

16.  Capone, S. V. and M. Petroccia.  (GCA Corporation.)  Guidance to State
     and Local Agencies in Preparing Regulations to Control Volatile Organic
     Compounds from Ten Stationary Source Categories.  (Prepared for
     U. S. Environmental Protection Agency.)  Research Triangle Park, N.C.
     Publication No. EPA-450/2-79-004.  September 1979.
                                     i  or-
                                     1 —£-U

-------
             APPENDIX B

  COMMENT LETTERS RECEIVED ON DRAFT
CONTROL TECHNIQUES GUIDELINE DOCUMENT

-------

-------
                                 APPENDIX B

     This appendix contains letters received commenting on the draft control
techniques guidelines document for fugitive emissions from synthetic organic
chemical, polymer, and resin manufacturing plants.  Twenty-two letters were
received from industry representatives and trade groups.  Table B-l contains
a listing of the commenters and their affiliations.
                                     B-l

-------
              TABLE B-l.  LIST OF COMMENTERS AND AFFILIATIONS
Comment No.                            Commenter and Affiliation

    1*                            Mr. W. M. Reiter, Corporate Director
                                  Corporate Environmental Affairs
                                  Allied Corporation
                                  P.O. Box 2332R
                                  Morristown, New Jersey  07960

                                  *EPA response attached,

    2                             Mr. Henry L. Ramm
                                  Environmental Engineer
                                  Government and Regulatory Affairs Dept.
                                  Rohm and Haas Company
                                  Independence Mall West
                                  Philadelphia, Pennsylvania  19105

    3                             Mr. D. E. Park, Director
                                  Environmental Affairs
                                  Ethyl Corporation
                                  P.O. Box 341
                                  Baton Rouge, Louisiana  70821

    4                             Mr. J. J. Moon, Manager
                                  Environmental and Consumer Protection
                                    Division
                                  Phillips Petroleum Company
                                  Bartlesville, Oklahoma  74004

    5                             Mr. John T. Barr
                                  Air Products & Chemicals, Inc.
                                  Box 538
                                  Allentown, Pennsylvania  18105

    6                             Mr. J. C. Edwards, Manager
                                  Clean Environment Program
                                  Tennessee Eastman Company
                                  Eastman Kodak
                                  Kingsport, Tennessee  37662

    7                             Mr. Allen R. Ellett, Environmental
                                    Specialist
                                  Environmental Affairs and Product Safety
                                  The Standard Oil  Company
                                  Midland Building
                                  Cleveland, Ohio  44115
                                     B-2

-------
        TABLE B-l.  LIST OF COMMENTERS AND AFFILIATIONS (Continued)    •


Comment No.                            Commenter and Affiliation

    8                             Mr. James W. Lewis, Manager
                                  Special Environmental Projects
                                  The BF Goodrich Company
                                  Chemical Group
                                  6100 Oak Tree Boulevard
                                  Cleveland, Ohio  44131

    9                             Mr. A. H. Nickolaus
                                  Texas Chemical Council
                                  1000 Brazos, Suite 200
                                  Austin, Texas  78701

   10                             Mr. F. M. Parker, Environmental  Coordinator
                                  Chevron U.S.A., Inc.
                                  575 Market Street
                                  San Francisco, California  94105

   11                             Mr. H. R. Norsworthy, Manager-Manufacturing
                                  Synpol, Inc.
                                  P.O. Box 667
                                  Port Neeches, Texas  77651

   12                             Mr. R. B. Tabakin, Manager
                                  Environmental Affairs
                                  American Cyanamid Company
                                  One Cyanamid Plaza
                                  Wayne, New Jersey  07470

   13                             Mr. E. J. Burkett, Manager
                                  Corporate Environmental Engineering
                                  The Goodyear Tire & Rubber Company
                                  Akron, Ohio  44316

   14                             Mr. W. F. Blank, Manager
                                  Pollution Control
                                  Corporate Environmental Affairs
                                  Allied Cnemical
                                  P.O. Box 2332R
                                  Morristown, New Jersey  07960

   15                             Mr. Thomas V. Malorzo
                                  Senior Regulations Analyst
                                  Diamond Shamrock Corporation
                                  717 North Harwood Street
                                  Dallas, Texas  75201
                                     B-3

-------
        TABLE B-l.  LIST OF COMMENTERS AND AFFILIATIONS (Continued)


Comment No.                            Conmenter and Affiliation

   16                             Mr. A. H. King, P.E.
                                  Environmental Consultant
                                  The Firestone Tire & Rubber Company
                                  1200 Firestone Parkway
                                  Akron, Ohio  44317

   17                             Mr. R. W. Fourie, Manager
                                  Environmental Programs, Shell Oil Company
                                  One Shell Plaza
                                  P.O. Box 4320
                                  Houston, Texas  77210

   18                             Mr. William P. Gull edge
                                  Manager, Environmental/Scientific Programs
                                  Chemical Manufacturers Association
                                  2501 M Street, N.W.
                                  Washington, D.C.  20037

   19                             Mr. Steven A. Tasher
                                  Legal  Department
                                  E.I. du Pont de Nemours & Company, Inc.
                                  Wilmington, Delaware  19898

   20                             Mr. C. D. Mai loch
                                  Regulatory Management Director
                                  Monsanto Company
                                  800 N. Lindbergh Boulevard
                                  St. Louis, Missouri  63166

   21                             Mr. Bonner L. LaFleur, Chairman
                                  Environmental Impact Committee, Southern
                                    Rubber Group
                                  P.O. Drawer 1361
                                  Lake Charles, Louisiana  70602

   22                             Mr. M. J. Rhoad
                                  Managing Director
                                  International Institute pf Synthetic
                                    Rubber Producers, Inc.
                                  2077 South Gessner Road
                                  Houston, Texas  77063
                                    B-4

-------
                                    Allied Corporation
                                    Corporate Environmental Affairs
                                    P.O.' Box 2332R
                                    Morristowrt, New Jersey 07960
                                             December  23,  1981
Mr. Don R. Goodwin, Director
Emission Standards and Engineering Division  (MD-13)
Office of Air Quality Planning and Standards
U. S. Environmental Protection Agency
Research Triangle Park, North Carolina 27711

Dear Don:

    I have just initiated review of the draft  CTG  entitled,  "Control
of Volatile Organic Compound Fugitive Emissions  from  Synthetic  Organic
Chemical, Polymer, and Resin Manufacturing Equipment."

    I am somewhat disturbed, as it appears that  the contributions  that
were made by NAPCTAC committee members and in  some cases your comments
apparently have been completely disregarded  in the preparation  of  this
document.  I recognize that the CTG is labelled  draft,  however, our
comments were provided many months ago.   Further,  the release of such
a flawed document to the States and EPA Regions  constructs  a foun-
dation for improper and technically unsound  control assessments.  I
recognize that this is not a "final" document, however, the  label
"draft" may be lost in the pressure of permitting.

    Further, the document does not use available SOCMI  and  Polymer
plant data contributed by your contracts  and industry sources  (e.g.
Allied Corporation).  The failure to use  available control  techniques
data is contrary to '§108 of the Clean Air Act.  I  quote from §108(b)(l)

         "Simultaneously with the issuance of  criteria  under
         subsection (a), the Administrator shall,  after con-
         sultation with appropriate advisory committees and
         Federal departments'and agencies, issue to the
         States and appropriate air pollution  control
         agencies information on air pollution control  tech-
         niques, which information shall  include data
         relating to the cost of installation  and  operation,
         energy requirements, emission reduction benefits,
         and environmental impact of the  emission  control
         technology.  Suchinformation shall 1nclude  such
         data as are a'v^Vi'lXble^ oh avai 1 able  technology  and
         alternati ve methods of prevention and control  of
         ai r pol1uti on.  Such i nformati on shall  also
         include data on alternative fuels,  processes^ and _
         operating methods which wil 1 resujt in  elimination
         of significant reduction of emissions".
         (Emphasize added)
                                   B-5

-------
    My specific  concerns are as follows:

    1)   The document does not reflect the constructive comments made by
         NAPCTAC members.

    2)   The document does not reflect the advances made by EPA in their
         study of fugitive emission problems.  Data collected by Radian
         Corporation of numerous SQCMI plants has been entirely neqlected.
         The contribution made by Allied Corporation and other
         industrial firms directly to EPA and via trade organizations
         has not been included.

    3)   Our concern relative to the handling of safetv valves has
         be.en disregarded.  This document again supports the
         installation of a block valve before the safety valve, a stec
         which could eliminate insurance coverage for the facility and
         more significantly, jeapordize the lives of many workers.

    4)   The document can be misinterpreted by the local regulator to the
         point where he might include flanges, agitator seals, and
         emissions from secondary sources such as cooling towers for control.
         This circumstance could arise since there is no clear arid emphasized
         exclusion within the document.   Rather there is a review of
         the losses from such sources with a simple caveat  (difficult
         to find) indicating that these areas may not be covered by
         the CT6 recommendation.

    I feel that  the document does not reflect a professional
evaluation of the fugitive emission problem associated with polymer
and organic chemical plants.  Rather it is an attempt to extrapolate
.from refinery data.  I strongly recommend that you consider
withdrawing the  document, or at least clearly indicating to State and
Federal  regulators who have received the document that the document in
its present state is not to be used in formulating RACT.

    I will provide chapter by chapter comment as rapidly as possible.

    I am taking  the liberty to share these comments with other NAPCTAC
members  and sol licit their comments.

    Have a Merry Christmas and a Happy New Year.

                                       Very truly yours,
                                       W.  M.  Reiter
                                       Corporate Di rector
                                       Pollution Control
/pab
CC:      R.  0.  Blosser                  Dr.  J.  M.  Lents
         R.  J.  Castelli                 R.  A.  Moon,  Jr.
             F.  Dubrowski                   W.  Reilly
         E.  H.  Haskell                  B.  A.  Steiner
         E.  E.  Lemke           0 c
                               D-D

-------
                                        JAN
1952
       Mr. W. M. Reiter
       Corporate Director, Pollution Control
       Corporate Environmental  Affairs
       Allied Corporation
       Post Office Box 2332R
       Morristown, New Jersey 07960

       Dear Bill:

            In response to your letter of December 23, 1931, outlining several
       concerns with the draft control techniques guideline (CTG) document
       entitled "Control of Volatile Organic Compound Fugitive Emissions from
       Synthetic Organic Chemical, Polymer, and Resin Manufacturing Equipment,"
       I would like to draw your attention to the following points.  First, the
       draft CTG document does reflect constructive comments made fay NAPCTAC members.
       For example, the preliminary draft CTG document discussed at the March 1931
       KAPCTAC meeting included a model regulation.  The NAPCTAC recommended that
       model regulations not be included in CTG documents; and, you will note, a
       model regulation is not included in the draft CTG document you received.
       Also, a number of the NAPCTAC members recommended that the CTG document
       accommodate alternative approaches to quarterly inspections for locating  .
       equipment leaks and exemptions for snail process plants.  Again, you will
       note that the draft CTG document you received accommodates skip-Deriod
       monitoring and "an allowable percentage of valves leaking" as alternatives
       to quarterly inspections.  In addition, the document accommodates exemptions
       for sn-.all process plants with less than a hundred valves in gas and/or
       licnt liquid service.

            Second, the draft CTG document includes fugitive emission data from
       the synthetic organic chemical manufacturing industry (SOCMI) to the extent
       these data had been received, comoiled, and'assessed in Hay 1931, when the
       draft rinr-t-^nt- wa«: -FnvwarHor! -f-n i-ha Office of Manaoenent and Budget for
       Preliminary assessment of these data supported extrapolation from the Detroleurn
       refining industry to the SOCMI concerninn fuoitive emissions from process
       equipment.
                                       CONCURRENCES
SYMBOL i! T" r")
M "-" -r.
A'"","-,' ; 	
SURNAMF k *./ \ . ,~
r> -"' C* '" *-'. *
\ ; j - /
~> A T £ &• ;/'-/•-.
' t i -•' - ^_—





















SPA =cm. 12:C-i '17.70)
                           OFFICIAL FILE CC

-------
     Since May 1931, additional SOCMI fugitive emission data have been received
and compiled.  While we have not completed an assessment of these data, it does
appear that some adjustment of various emission factors included in the
draft CTG document may be warranted to reflect differences between the
petroleum refining industry and the SOCMI.  Currently, we plan to complete
our assessment of these data over the next month or two and then publish a
Federal Register notice in the spring surrcnarizing our technical conclusions
regarding fugftrve emissions in the SOCMI.  These conclusions will, of course,
be incorporated in the final CTG document we develop.

     Third, your concern relative to the handling of safety valves has not been
disregarded.  As I mentioned in my letter of May 12, 1981, we find a number of
companies, such as Exxon and Union Oil, who routinely follow the practice
outlined in the draft CTG document and we find that the ASME Boiler and Pressure
Vessel Code permits this practice.  In addition, I should like to point out
that rather than use a block valve, one could use a three-way valve vented
to a second safety relief valve.  In this manner, the process would always
have access to a safety relief valve.

     Fourth, we do not think State or local air pollution control agencies
could misinterpret the draft CTG document and include flanges, agitator seals,
and cooling towers in a leak detection program.  You will note, for
example, that these items are not included in the recon^endation for
reasonably available control technology simarized on page 4-1 of the
draft CTG document which states: "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."

     I have tried to indicate that we have accommodated the suggestions
ir.adc L»y committee members and industry representatives.  We may'have
nissec sone and we will rsview your chanter-by-chapter corments carefully
when they are received.

     !/e appreciate your in-denth review of our technical documents vary
ppch.  1-te do our best to provide solid technical work and comments by
those experienced in the design'and operation of chemical plants are
essential to this effort.

     Best wishes for the N'ew Year.

                                        Sincerely yours,
bcc:  R. 0. Blosser                     Don R. Goodwin
      R. J. Castelli                       Director
      F. Dubrowski                  Emission Standards and
      E. H. Haskeii                  Encnnserina Division
      E. E. Lemke                      •        -
      J. M, Lents
      R. A. Moon, Jr.
      W. Reilly
      B. A. Steiner
                                      B-8

-------
                                                   ' "          COfVlPAN'
January 8,1982                    •        ...      >
Mr. F. Porter
U. S. Environmental Protection Agency
Office of Air Quality  and Planning  & Standards
Research Triangle Park, NC .  27711

•Dear Mr. Porter:

The Rohm and Haas Company  is  a member  of  the  Chemical
Manufacturers Association, and we support  the CMA critique  of
draft documents  issued for the control  of  volatile organic
compound fugitive emission from  synthetic  organic chemical,
polymer, and resin manufacturing equipment.   As  an adjunct  to
their review, the following  comments are  offered.

1.  Model Regulat ions  in Control Techniques Gui_de_l_j_nes  (CTG)

Formerly CTG's  containing model  regulations  tended to  be  more  of
a  regulatory rule than a guidance document because a state
agency, not having  the expertise or technical manpower  available
to your office,  would  in many cases' adopt  the model  regulation as
listed, even if  it was not justified.   By  not including a model
regulation  in the August.1381 draft CTG you  are  bringing  the
document towards its  interided purpose  of  a guidance  document.
The  state and local  agencies  have the  responsibility for  first
line  control and should be encouraged  to  decide  what level  of
control is  necessary.   We  support the  deletion  of the model
regulations section.

2.   ADDendixB,  Tables I and 11

V,"e  agree with  the.listing  of specific  chemicals  in Appendix B,
Table I as  this  explicit  listng  makes  clear  exactly  what
processes  are  referred to.   However, as worded,  some categories
ar'e  not specific.   The terms "acrylic  acid and  esters,"
" ethanolami nes ,t! phenol sul f on i c  acids," "polybutenes , "
"tetra/chloroethanes,"  "toluenesulfonic acids,"  "toluidines,"
"trichlorobenzenes"  could  be construed as  covering broad  classes
of  compounds.   It  is  requested that each  compound to be covered
by these  rules  be listed separately, i.e., "acrylic acid, ethyl
acrylate,  butyl  acrylate,  ethanolami ne, diethanolamine,"  etc.
The  more  general terms should be deleted.
                              B-9

-------
3-  Data Base for CTG Draft

The data for fugitive emissions  from  synthetic organic chemical
plants are significantly different, and  in many cases lower, than
the petroleum refiners  fugitive  emission data base.  The draft
CTG should be revised using'the  SOCMI data base so that the
document is accurate and does mislead the users to develop
unneeded and unproductive emission control regualtions. • •
      fly,
H. L. Rarrm
Environmental Engineer
Government and Regulatory Affairs Dept
cm
01CM21
                                  B-10

-------
                     ETHYL CORPORATION

                           January 11,  1982
                                                                 ADORCSS REPLY
                                                              To: f. O. »ox 3*1
                                                           BATON ROUCC, LA.7OS2I
Mr.  F.  L.  Porter
Emission S'tandards and Engineering Division ,{MD-13)
Environmental  Protection Agency                            .
Research Triangle  Park,  North Carolina  27711

Dear Mr.  Porter;

          Re:     Control  Techniques Guideline; Volatile  Organic  Compound
                 Fugitive  Emissions from Synthetic Organic Chemical,
                 Polymer, and Resin  Manufacturing Equipment

          The  following  is in response to the  request for  comments on  the
Ira'ft document (46 FR 5963Q).

          Ethyl Corporation considers  the  proposed guidelines  overly
estrictive and .excessive  in record keeping, reporting and monitoring
•equirsmsnts.   The control strategies  suggested  are  extremely  difficult
,o implement and  enforce.

          Ethyl's  corporate environmental staff have carefully  reviewed  the
•evised  draft of these guidelines  and  do  not 'find "many significant changes
rom the  previous  document.  Many of the  technical weaknesses  of the
previous  draft  have not been  corrected.

          Fugitive emissions  from the  Synthetic  Organic Chemicals
Manufacturing  Industry are not of large anough magnitude  to warrant such
extensive documentation and  control.    These  emissions ars by nature
Is  minimis.    The  guidelines  propose  over-regulation of these  emissions.

          Regulations to  reduce such  emissions,  if necessary, should
•stabiish clear-cut objectives and specify reduction requirements.   Ths
necHS to achieve  the reduction should be an industry decision.

         - The  cost estimate in the guidelines  is not valid.   It is based  on
i  10 percent interest rate, which is unrealistic  in the  present market.
The  labor cost based on  "wages  plus  40  percent" for  overhead is also  low.
["he  overhead  often is 100% of wages.  The total basis for valve cost  is a
Dns-inch globe valve.  Many different types  and  sizes of valves are used-
                                      B-ll

-------
januery VI, 19SZ
?ac=  2
in SOCMI.   The cost of a valve does  not  increase  linearly  as the  size,
but sxponentially.   Also,  the majority  of valves  in  service "arei larger
shan one  inch.

          We support the  use of the bubble concept  for emission control
strategies.   This  concept  could encompass these de  minimis fugitive emissions
without the excessive control .proposed in the guidelines.

          In considering these comments we  urge  EPA to modify the
guidelines.   As proposed,  the burden of compliance on  industry and en~
forcement is unduly excessive.

                                        Very truly yours,

                                        ETHYL CORPORATION
                                         D.  E,  Park, Corporate Director
                                         Environmental Affairs
DEP:jht
                                   B-12

-------
                   PHILLIPS  PETROLEUM  COMPANY

                       BARTLESVILLE. OKLAHOMA 7AQO4
                                918 661-5356
Corporate engineering

JOHNJ MOON
Manager, Environment and Consumer Protection
January 15, 1982
Mr. Fred L. Porter (2)
Emission Standard & Engineering Division  (MD-13)
Environmental Protection Agency
Research Triangle Park, NC  27711

Dear Mr. Porter:

       Phillips appreciates the opportunity  to  comment on the draft of
"Control of Volatile  Organic Chemical,  Polymer,  and Resin Manufacturing
Equipment" dated August, 1981  transmitted by your letter of December 1,
1981.  After reviewing the draft our  comments are as follows:

       Section 2.1 -  To state  that  the  equipment  in process units in
the synthetic organic chemical manufacturing industry (SOCMI) is similar
to equipment in the polymer and resin manufacturing industry is only
partly true.  The polymer and  resin plants are  basically quite different
from the SOCMI.  A SOCHI facility handles gas and/or light liquid through
most of the facility  while in  a polymer or res-in facility only a small .
part of the facility  handles gas and/or light liquids.

       Section 3.1.2.3 - Most  safety/relief  valves in chemical plants
relieve into a vapor  recovery  system  or into a  flare system.  These
valves should be excluded from the  monitoring requirement for volatile
emissions.                          .          ;                      •

       If a block valve is installed  up stream  of a relief valve the
block valve has to be locked open.  Only  an  authorized person can unlock
the block valve and he has to  stay  with the  block valve until it is
again locked open.

       Section 3.1.3.3 - The allowable  interval before repair of 15 days
does not allow maintenance enough time  if parts have to be ordered.  It
is suggested that 30  to 45 days would be  more appropriate in these cases.

 „     Section 3.1.3.5 - The reduction  efficiency expression A x B x C x
D  is misleading.  The reference number 8 should be placed after the pre-
ceeding sentence that describes the expression.
                                     B-13

-------
 Page 2.
'January  15,  1982
        Section 3.2.2  - The 2% allowable number of valves  leaking appears
 zo be  unrealistic.  Table 3-2 estimates 10% of valves  in  gas  service and
 12S of valves  in lighr liquid service leak.  Table 4-16  "Summary of Valve
 Maintenance Test Results" shows  that  only 56.4% of total  valves  were re-
 paired successfully.   These data would seem to indicate  that  5  to 6%
 would  be  a more appropriate performance level.

        Section 5.1  -  The  capital cost of implementing  Reasonable Avail-
 able Control Technology (RACT)  did not include the cost of' an initial
 survey and inspection of  plant components.   This survey and inspection
 of & large facility can be a mammoth  job and very costly.

        Section 5.3,2  — The amount of  recovery credit for  VOC  saved for
 unit C ($211,100) does not agree with the amount stated in Table 5-7
 ($216,730).

        If you  have  questions on  any of the above,  please  contact A.  C.
 Oliver at (918)  661-5735.

                               Very truly yours,
 JjH/ACO/pkc
                                          B-14

-------
                                                 Air Products and Chemicals, inc.
                                                     Box 538. AHeniown. PA 16105
                                        11 January 1982
Mr. Fred L.  Porter
Emission Standards and Engineering Div. (MD-13)
EPA
Research Triangle Park, NC  27711 .

Dear Mr. Porter:

We 'have reviewed  the  August 1981  draft 'guideline  for Control of Volatile
Organic  Compound  Fugitive  Emissions for  Synthetic  Organic Chemical,  Polymer,
and Resin Manufacturing Equipment,  and offer you  our comments  and  suggestions
on this draft.

This draft  is a useful attempt  to  provide guidance  to regulatory bodies which
are establishing rules on fugitive emissions.  It assumes that these regulations
will be  prepared,  but  it-does not furnish guidance as to which type  of  emissions
deserve  priority treatment;  it  assumes that all emissions  are  equally undesir-
able.   This is  not correct.  The  Agency has  on  several occasions published
documents  illustrating the difference  in the  photodegradation  rates and  ozone
yields  of  various  organic substances.   Therefore,  it would be helpful to  the
agencies considering  the need for  regulations  and the priority to  be  assigned
to various  substances'.to devote a section to  this issue.   You  may  remember our
discussion  on this point at  the RTP hearing in  March of last year.  See Dqcket
A-79-32.

The  primary weakness of this draft is the total  dependence on the similarity
between the refining  and the chemical  industries  for the data  base.   See pages
2-18-20.   In the  above-referenced  docket, we  have  submitted comments on  the
disparity  of the  two industries.   In  brief,  chemical  plants are smaller,,
operate at  lower temperatures and pressures, and process more valuable  streams,.
Certainly,  the  size of a  valve or pump  has  a  direct effect on the  rate  of
leakage to  be expected,  and  thus on its  effect on the environment. We realize
the  problems involved in  obtaining representative  data for such a variety of
processes,  but  believe that the Agency  should not  rely  totally on data from
the  refinery industry.  You have  data  on many chemical operations (see  the
EPA.-450/3-80-G28  series,  for example)  and should  attempt to apply  as  accurate.
data as  possible,  rather than rely on surrogates.

Along  this  line,  we would  suggest that  the alternative  control  strategies
discussed  in Chapter 3 provide  for an  exemption from regulation based on size
as well as  on  the  number  of valves and pumps.   The  value of a substance usually
 increases  as the  equipment  size decreases.   Thus,   the  operator has  a strong
economic incentive not to lose the material.  A cost-effectiveness  analysis'
would  suggest that regulating efforts  be concentrated on the larger potential
 emitters ("page 3-21).
                                         B-15

-------
                                     -2-
Iu  •'=  for  these reasons that we  believe  that the calculations in Appendix D
are cr.erly optimistic and that the projected reductions in emissions will not
be sciieved on an average basis.  The leak rates simply are not representative
of zhe entire  chemical industry, but of the refinery industry.

We support the concept of the skip-test procedure.  We would suggest that the
users of this  guide need further assistance  in selecting the values for the
monitor-skip periods, however,  there should be  some  discussion  of how these
periods are to  be chosen,  and  some  illustrations  of the effects of choosing
other arbitrary values of i and m.

In regard  to  the 2% figure  as  an index of compliance (page 3-17),  it can be
seen  from  table 3-2 that this  value  depends  heavily on the -service in which
the equipment  operates.  Some  flexibility should  be  allowed for this factor.
Similarly, it  would help the regulator to have  some  further illustrations of
the quantitative effect  of alternate repair times (page 3-8).

Another place  where more guidance would  be  useful is in selecting  the  leak
rate  where a  forced turnaround would  be  required.  An inexperienced reader
would conclude  that a  rate  above 2% would justify this action.  More data on
average and excessive leak rates would be helpful here.

There are  several  other underlying factors which  bear heavily on the effec- .
tiveness  of  this control strategy.  One  is  the arbitrary volatility split,
which may  have  some  significance  in the refinery industry, but is not relevant
to the  chemical  industry.   Another is the relative response -to the various
substances by the  detection instrument.   This  wide  range  of  sensitivities,
coupled with  the difference in  photochemical •. reaction rates,  produces  an'
enormous difference  in the actual emissions, and"their  impact on the environ-
ment, from various  'substances.   It is not clear that a casual reader of this
guide will be  aware  of these  facts.

We  hope  that  these comments will be  useful  to you as you revise this draft.
The fact  that the Agency apparently  is  choosing  not to  utilize conventional
rsc'-'latory procedures  in establishing  these regulations makes  it  important
tnai the guidelines be as accurate, effective, and flexible as possible.
            us directly  if you  have  any  questions  regarding these comments and
s.ccssvcns.
                                         Very truly yours,

                                         AIR PRODUCTS AND CHEM^ALS  INC.
                                         Regulatory Response
                                          B-16

-------
                         100-y€ar §tart on torr\grro
                                                  ~
January 15, 1982
Emission Standards  and Engineering
Division (MD-13)
Environmental Protection Agency
Research Triangle Park, North  Carolina  27711

Attention Mr. Fred  Porter

Reference:   Draft (August,  1981)  Control Techniques Guideline Document for
            the Control of  Volatile  Organic Compound Fugitive Emissions from
            Synthetic Organic  Chemical,  Polymer,  and Resin Manufacturing
            Equipment

Eastman Kodak Comany is a manufacturer of photographic products, chemicals,
fibers and  plastics with major manufacturing facilities in seven states.   As
a multistate employer of approximately 100,000 people, we support the
objectives  of the Clean Air Act  and  EPA's efforts to achieve the goals of
this Act.  The proposed Control  Techniques Guideline (CTG) could have a
significant impact  on our operations  since Eastman Kodak Company produces
many of the chemicals, polymers,  and resins listed in Appendix B of the CTG.
Therefore,  we have  the following comments on the  draft Control Techniques
Guideline document.
                                                                 •
We commend  EPA for  making several  adjustments in  the draft CTG in response
to previous public  comments.   Adjustments which we support are:

1.  The exclusion of Chapter Six which contained  a model regulation.  The
    CTG is  as its title implies,  a guideline and  not a regulation.
    Therefore, it is more appropriate to have state specific regulations.

2.  The opportunity to apply alternative control  strategies.

3.  Alternate uonitoring requirements for unsafe  and difficult to reach
    components.

4.  The exemption of process units with  less than 100 valves in gas service
    ?.nd light liquid service from regulations requiring control of fugitive
    VOC omissions.

However, there are  certain  other parts in the draft (August, 1981) CTG that
could be improved.  For example:

1.  The draft CTG uses the  refining  industry data to. estimate Synthetic
    Organic Chemical Manufacturing Industry (SOCMl) uncontrolled fugitive
    emissions, control costs,  and the proposed emission reductions.  The CTG
    assumes fugitive emissions in the refining industry represent fugitive
         EAGT.V.AN KODAK COMPANY • K1NGSPOBT, TENNESSEE o7652 • 615 2^6-2111
                            E,'i'tiT73n r1--—,-:.--ls ij.fis.'on
                                  B-17

-------
Emission Standards and Engineering
Division (KD-13)
Page 2
January 15, 1982  •
    emissions in SOCMI.  However,  in EPA  contracted  studie.s  data have  been
    obtained showing SOCMI uncontrolled fugitive  emissions are much  less
    than the refining  industry  uncontrolled  fugitive emissions.  For
    example, EPA1s Report 600/2-81-111 entitled Analysis of  SOCMI VOC
    Emissions Data which studies  data  from 24 .process units  in the SOCMI,
    reveals SOCMI fugitive emissions are  substantially less  than refinery
    fugitive emissions.  The CTG  should be revised to reflect the SOCMI da
    now available.

    The draft CTG suggests that components which  have a measurable VOC
    concentration of 10,000 ppmv  or greater  should be considered leaking
    conponents and repaired.  However, the Chemical  Manufacturers
    Association (CMA)  in its comments  presented on August  7, 1981, showed
    that the control efficiency for a  100,000  ppmv screening value for
    valves in gas service was more cost effective and had  a  control
    efficiency near the 10,000  ppmv screening  value.

    The draft CTG suggests a quarterly monitoring schedule.  CHA's August
    1981, comments showed that  annual  monitoring  instead of  quarterly
    monitoring only lowered the control efficiency by two  percent  for  valv
    in gas service.  CMA also showed that annual  monitoring  with  a
    100,000 ppmv screening value  for valves  in  gas service could be
    implemented at a more reasonable cost to industry.

    Also, a report on  an EPA 10-month  study  of  fugitive -emissions  at an
    Allied Corporation high-density polyethylene  unit in Baton Rouge,
    Louisiana, indicates that a repetitive monitoring program on a quarter
    or monthly basis would be far less cost  effective and  could  exceed
    ?20,000/Mg VOC.  It was concluded  that a repetitive monthly  or quarter
    program provides no additional benefit and  is not cost effective.'

    The draft CTG on Page 3-21  suggests a state agency might wish  to
    consider a provision in their RACT regulations which would allow the
    ••'fct-rnoy 
-------
?,7\ission Standards snd Engineering
Division (MD-13)
Page 3
January  15,  1982  •
In summary, we suggest the following  changes  in  the  CTG:

1.  An analysis of SOCMI fugitive  emissions should' be made  rather than
    relying on refining industry data.  Then  a decision  of .whether  a CTG  is
    needed for SOCMI could be appropriately made.  If a  CTG is needed, a
    cost effectiveness estimate should  be  based  on SOCMI data.

2.  If a CTG is justifiable, the monitoring of valves in gas  service should
    be performed on an annual basis with:  a screening value  of 100,000 ppray.

3.  The draft CTG should address the  economic  and environmental effects of
    early  shutdowns.

4.  The cost analysis should utilize  a  monitoring time of 3.4 person-minutes
    per source.

Very truly yours,
J. C. Edwards
Manager, Clean Environment Program
Tennessee Eastman Company
Division of Eastman Kodak Company

af
                                    B-19

-------
             THE STANDARD OIL COMPANY
                                                            MIDLAND BUILDING, CLEVELAND, OHIO
   R  E  FA=J?.£U.
     DiRJC-Ga
•NVmONMENTAl AFFAIRS
      AND
  PRC3UC7 SAFETY                               January  15,  1982
      Mr. Fred L. Porter
      Environmental Protection Agency
      Emissions Standards and Engineering
        Division (MD-13)
      Research Triangle Park, NC  27711

      Subject:  Draft CTG - Control of Volatile Organic  Compound  Fugitive  Emissions
                from Synthetic Organic Chermical,  Polymer,  and  Resin  Manufacturing
                Equipment.

      Dear Mr. Porter:

      The Standard Oil Company of Ohio  (Sohio) has reviewed the above cited
      document, and would like to submit the  following comments on it.

         Basis for Regulation

      The guideline appears to be developed based  on  emission data and  rates
      determined from the Radian survey of  13 petroleum  refineries.   On. page 2-20,
      'it says that the operation of SOCMI process  equipment is  not expected to
      differ greatly from refinery operations, so  emissions would be  expected to-be
      sinlar.  We feel that this is not the case.   The data summarized  in  Appendix
      A shows that in most cases the percentage  of leaking  sources in chemical
      plants is much less than the percentage of  leaking sources  in a petroleum
      refinery.  (See table A-l, page A-13.)  It  is our  opinion that  this  table
      identifies enough difference to change  the  economics  significantly.

      Emissions from chemical plants will vary depending on the chemical feedstocks
      and products.  Some plants, such  as acrylonitrile  plants  currently have very
      strict work pla'ce limits under OSHA,  and hence  must be careful  to repair
      leaks to avoid high worker exposure.  Other  plants produce chemicals which
      tend to polymerize, and "self seal" small  leaks.   Based on these
      considerations chemical plants are not  similar  to  petroleum refineries and
      should not be regulated as such.
                                              B-20

-------
                                   - 2 -                     January  15,  1982
   Pump Seal Requirements

Double mechanical seals are not feasible on all pumps  in  a  chemical  plant,  as
the barrier fluid will contaminate the process stream  if  the  seal  leaks.
Tandem seals (double mechanical seals in which the  barrier  fluid  is  not under
pressure) are allowed if the seal oil is degassed and  incinerated.   This  is  a
very costly approach, because it is doubtful  if enough seals  will  be leaking
at any given time to support combustion.  Therefore, the  incinerator would
have to continuously fire alternate fuel.  Significant amounts  of  other
gasseous air pollutants (SO^, NO^, TSP) would be generated  to control a
rather insignificant amount of hydrocarbons.  We feel  that  this requirement
should be eliminated based on cost and these  other  environmental  impacts.

   Relief Valves

The guideline mentions monitoring relief valves.  This poses  a health and
safety hazard to the monitoring team, as if a valve were  to relieve  as a  team
was monitoring it, serious- injury may result.  This may happen  at  any time,
since these valves are designed to automatically vent  during  upset or
overpressure events.  We feel that any attempt to require monitoring safety
valves is unwarranted, and these references should  be  removed from the
guideline.

   Flanges

Flanges are not a source of leaks.  This has been, shown in  both refinery  .and
chemical plant surveys.  Any reference to monitoring these  should  be removed
from the guidelines.
                               i
   Inaccessible Sources

The guideline mentions that some sources in a chemical plant  are
inaccessible, and should not have to be monitored as frequently as accessible
sources.  We feel that it is dangerous to be standing  on  a  ladder  while
monitoring inaccessible valves, and that it would also take appreciably more
time to monitor these sources than the normally accessible  sources.   This
would seriously impact the economics of the requirement to  monitor these
sources at all.  We would recommend removing  the requirement  that  these
sources be monitored.

   Use of Draft Reports

In Chapter 2 references 1, 30, and 40, Chapter 3 reference  3, Chapter 4
reference 2, and Chapter 6 reference 5 are all draft reports.   If  these
reports were never issued as final reports, they should not be  quoted in  a
document that will be used to develop regulations.  The use of  these
documents creates a "house of cards" on which the regulations will be based.
                                     B-21

-------
                                                iv\ r /""\ IN i
                                   -  3  -                    January 15, 1982


Sohio appreciates this opportunity to comment  on this  draft guideline.  If
you have any questions -concerning these comments,  please contact me at
(216) 575-5136.

                                      Sincerely,                     :
                                      Allen R.  Ellett
                                      Environmental Specialist
ASl/dmb
                                       B-22

-------
 Trie SrGoob'ttch Company
 Cr.e.Tiicci Group
 6"i OC Osx Tree Boulevard
 C;« ve.'snd. Ohio 44131
 2 lo-W 7.6000

 January  11,  1982
 Emission Standards and Engineering Division  (MD-13)
 U,  £.  Environmental Protection Agency
 Research Triangle Park, North Carolina  27711

 Attention:   Mr.  Fred Porter

 Dear Mr. Porter:

 Re:   Draft  CTG Document:  Control.of .VQG FugitiveEmissions  from
      Synthetic Organic Chemical, Polymer, and Resin Manufacturing
      EquiPTnent.  Augus t, 1981

 Vie  appreciate the opportunity to comment on  the  subject  document.   We
 respectfully submit these comments pursuant  to your letter of  December
, 1,  1981 and the Federal Register notification of December 7, 1981.

; Our comments are directed to Page B-6 of Appendix B, List of Chemicals
'Defining Synthetic Organic Chemical, Polymer and Resin Manufacturing
• Industries.  We request that the following chemicals be  deleted  from
 this list:

      Table  I:   Synthetic Organic Chemicals  Manufacturing Industry

                OCPDB No.             Chemical "

                  3520       '  '       Vinyl Chloride


      Table II:  Polymer and Resin Manufacturing  Industry

                 Styrene-Butadiene Latex

 "or vinyl chloride, this request  is  based upon  the  fact  that fugitive
. -rls-ier.s frcr,  this jrrr-cess are already regulated by  the National  Zrais-
 sicr. Standard .for Hazardous Air Pollutants - Vinyl  Chloride, 40  CFR 61.65
 •"by.  The fugitive er.issicns frcr. the vinyl  chloride  process are already
 stringently  controlled and all  the  controls  listed  in  the  subject  support
 dccument ara already  in place.  Therefore, no  further  control can be
 expected by  application of the  RACT controls contained in  the subject
 document.  The  potential duplication of reporting would  be  burdensome  and
 and serve no purpose.'

 "or styrene-butaciene  latex, our  request is  based  upon the small potential
 for "OC fugitive  emission  reductions and associated costs  for this snail
 reduction.   We  assume  the  only  reason  styrene-butaciene  latex is still
 listed  in Table II  is,because  the Agency vas developing  a  CTG document for
 this  orccess.   '•:& base this  assumption' on the  fact  that  styrene-butadiene

-------
Pace 2
,'ET.uary 11, 1982
copelyraers were listed in the January, 1981, draft  fugitive guideline
document.

At that- time, both styrene-butadiene  crumb rubber and  styrene-butadiene
latex were included in the CTG document  "Control of Volatile Organic
Corspound Emissions from Manufacture of Styrene-Butadiene  Copolymer".
However, the SBR crumb rubber category was dropped  from the document
pursuant to our testimony at the April 29, 1981 National  Air Pollution
Control Technique Advisory Committee  (NAPCTAC) meeting.   The additional
controls were not cost effective.

Fugitive emissions associated with the styrene-butadiene  latex manufacture
are from styrene unloading/charging pumps, butadiene unloading/charging
pumps, and flanges and valves in the  liquid  lines between the storage
tanks and the reactors'.  These emissions are similar to the emissions
froa the emulsion crumb rubber process.   We  believe, therefore,  that
the Agency should eliminate styrene-butadiene  latex since styrene-
butadiene 'crumb rubber was deleted from  the  category and  the fugitive
emissions from the two processes are  similar.

                                   Sincerely,

                                   THE BFGOODRICH COMPA1SY
                                   CHEMICAL  GROUP
                                    James W.  Lewis
                                    Manager,  Special Environmental
                                    Projects
                                    B-24

-------
                             CHEMICAL CCUWCIL
1000 BRAZOS, SUITE 200, AUSTIN, TEXAS 78701-2476, (512) 477-4465
                                                   January 15,  1982
      'Emission Standards and Engineering Division (MD-13)
      Environmental Protection Agency
      Research Triangle Park, North Carolina  27711
      Attention:  Fred Porter (2)
                                          RE:   Comments  On The  Draft  CTG::
                                               Fugitive  Emissions  From Synthetic
                                               Organic Chemical, Polymer &  Resin
                                               Manufacturing Equipment"
                                               46 FR 59630,  December  7,  1981
      Dear Mr. Porter:
                  Attached are the Texas Chemical Council's  comments  on the
      subject fugitive emission control guideline.


                                          Sincerely Yours,
                                          A. H,  Nickolaus
                                          Chairman,  CTG Subcommittee
      CC:  J. S. Matey - CMA
           ?. J. Sienknecht - Dow
           J. D. Martin - Union Carbide
           J. B. Cox - Exxon
           Roger Wallis - TACB
           Air Policy Committee
           TCC Files
       JEC/rts
      Attachments
                                        B-25

-------
                     COMMENTS BY THE TEXAS CHEMICAL COUNCIL

                                     OS THE

                     DBAFT CONTROL TECHNIQUE GUIDELINE  (CTG)

         FOR THE CONTROL OF VOLATILE ORGANIC COMPOUND FUGITIVE EMISSIONS

FROM SYKTHETIC ORGANIC CHEMICAL, POLDER  & RESIN MFG. EQUIPMENT, DATED AUGUST 1.981
            The Texas Chemical Council  (TCC)  is an association of  85 chemical
companies having more than 67,000 employees in Texas  and representing
approximately 90% of the chemical industry in the state.  Thus the draft CTG
for the control of volatile organic  compound  (VOC) fugitive emissions is of
vital concern to us.


            The draft CTG does not fulfill its stated purpose  (Chapter 1) to
"review existing information and data concerning technology and costs for
fugitive emission control in the Synthetic Organic Chemical Industry (SOCHI)".
It is based on data from petroleum refineries and does not incorporate the
SOdl data (Ref. 1-8) developed specifically  for this purpose.  These SOCHI
data show:
                                                                        •

      •* Leak frequencies and rates  in  the chemical industry are different
         from, and significantly less than, petroleum refining so  the proposed
         control strategies are largely inappropriate.

      •* "Uncontrolled" emission levels are sufficiently close to  the controlled
         levels sought by the EPA so that a CTG may be unnecessary.

      •  The cost of emission control is greatly understated.


            Thus the draft CTG needs extensive revision to make it accurate
and technically sound.


            Most of the deficiencies in this  document have been discussed in
the previous TCC comments listed in  Table 1.  The discussion following
sunsarizes and/or references those that are especially pertinent.  Since
extensive reference is made to our July 27, 1981 comments  (7.ef. 17) on the
SOQ1I studies, a copy of them is attached  (see Attachment 1).  Also,
reference numbers 1 through 16 in this  letter have been kept the same as
those in our July 27th comment to help  avoid  confusion.
                                  B-26

-------
                                     - 2 -
 • I-  Why The CTG Should Be Redone

     The EPA1 s approach to fugitive emission control  is based on certain
key conditions and suppositions.  Several of these are not in agreement with
the facts developed in the SOCMI studies.  The differences and consequences
are discussed in Attachment 1.  Also see page two of  Ref. 10 and pages 25-
29"of Ref. 11 for a further discussion of screening values relative to the
definition of a leak.

     In addition to the comments already made, we continue to be puzzled
by the data shown in. Figure 3-7 _of the Maintenance Study  (Ref. 1) in which
a control sample group o'f 60 valves actually decreased 'in emission rate
over a median 77 day period.  This is contrary to the EPA's theory of leak
occurrence, and we would appreciate an explanation.

     We have also reviewed the High-Density Polyethylene  Plant data
(Ref. 3) and in addition to concluding that more frequent inspection and
monitoring did not reduce the percentage of valves leaking (Ref. 17), we
note the percentage of valves leaking is high also.   These data indicate a
good performance level of ,2% of valves leaking is too stringent for RACT
(See Figure 1).

     We believe the discussion in Attachment 1 gives  ample reason for
redoing this CTG.  We note that although the SOCHI Screening Study (Ref.
2) was published in September 1980, the SOCMI Maintenance"Study (Ref. 1)
in day 19ox, and the-Analysis Study (Ref. 3, 6, 7) in Juse 1931, t:st one
SOCMI data point was used in this draft CTG published in  August 1981 to
estimate SOCMI emissions, to develop a SOCMI control  strategy, or to
estimate SOCMI control costs.  'The latest references  in Cnapters II and
III are 1979.  The CMA and TCC comments in March 1981 on  the preliminary
draft CTG were almost completely ignored.

     If model plants are to be used,, a set more representative of SOCMI
should be developed.  These should give greater consideration to a wide
range of leak frequencies and to the relationships summarized on page 5
of Ref. 3, 6, and 7.  Contrary to the statement in Par. 2.3.1 of the CTG
the model units do not represent different levels of  process complexity.
Each contains the sane components in almost the same  ratios; the only
difference is in Che numbers of each.  Further, in Table  II-3 of the
rererencs they are based on (Ref. 1 in Chapter 2), Model  Units A, 3,
and C are idencified as Small, Medium,, and Large Model Plants.

 II. Is A CTG Heeded?
     In Reference 18 the CMA estimated that SOCMI emissions ba^ed on
SOCMI data are probably less than 30% of EPA's estimate baseo. on petroleum
refinery data.  Their comments on this are quoted on  the  following page,
and their complete submission is attached  (Attachment II) for your review.
                                  B-27

-------
                                       - 3  -
      "Using  the  data  from Table I we calculated a SOCMI emissions
      total of  55 gg/yr.   However, the BID using refinery data •
      estimated SOCMI  fugitive emissions  of 200  gg/yr.   This  further
      confirmed previous  CMA assertions that SOCMI emissions  are
      approximately 30 percent of the refinery emissions.  This is
      true, even  though emissions are calculated using only data
      from the  "high leak" processes  — ethylene,  vinyl acetate and
      cuoene.   The actual SOCMI industry is in large part comprised.
      of  "low"  and "non-leak" processes.   Of the approximately 1,000
      SOCMI plants in  the data base,  the ethylene and cumene  plants
      represent less than 5 percent of the total numb_er. '

      'If emissions were  calculated using a true mix of the plants,
i      including "low"  and "non-leak"  processes,  the SOCMI emissions'
      would be  considerably below 55  gg/yr.  In  fact, the uncontrolled
      SOCMI emissions  might well approach EPA's  proposed regulatory
      goal of 26  gg/yr.   We conclude the present uncontrolled fugitive
      emissions from SOCMI are de mini'mis.  The  data from these reports
      demonstrate no real need for the NSPS or the CTG."

      The TCC agrees with CMA's conclusions, and we urge the EPA to
.seriously reappraise  the need for this CTG.

 Ill-  Control Costs Are Greatly Underestimated

      The analysis of  RACT control costs in Chapter 5 greatly under-
 estimate them.  Using screening times and leak" rates based on 50(241  data
 in Reference 2 we'estimate the minimum cost of  EPA's proposed program
 for Model Unit B to be S980/Mg VOC instead of a $247/Mg credit ~ a
 difference of  $1227/Mg.   Details and basis for  our estimate are shown
 in Table 2.

  IV.  T?rr^er.t On  E?A's RACT Recommendations

      In  Chapters 3 and 4 the CTG makes various  recommendations of  what
 they  consider  to be reasonably available control technology (RACT).   The
,TCC believes that in,many if not most cases existing SOCMI maintenance
 and operating  practices  result in emission levels equal to or better than
 E?Ars proposed program.   For this reason any RACT recommendation the EPA.
 =akes should allow these practices to continue.  The TCC also believes
 that  differences between the SOCMI data in References 1 through 8  and the
 basis used by  the EPA in developing their proposed control strategy  are
 such  chat the  whole strategy should be re-analyzed.  However, we are
 pessimistic  about this being done, so our comments and recommendations
 following are  intended to make the best of what the EPA has recommended,
 not to endorse them.
                                     B-28

-------
                                 - 4 -
A.  Monitoring Frequency

    1.  We recommend the basic monitoring frequency be once, per year
    with those units that do not achieve an "allowable-percentage of
   •valves leaking" increasing their frequency to semi-annually or
    quarterly.

    2.  We also recommend that those"units which meet this good
    performance level for two or three annual inspections be allowed
    to drop the program entirely - provided they continue their same
    general operating and maintenance practices. -

    3.  We also recommend the program start with an initial monitoring
    and that the second monitoring one year later be used to determine
    if a unit must start a more .frequent monitoring period.  Starting
    this way would give a plant time to assess their situation, make
    feasible engineering and operating improvements, or gear up for
    more frequent monitoring.  Since we believe most SOCMI units
    will pass the initial screening test, a great deal of unnecessary
    work will be avoided.  After the second annual monitoring, 'units
    would go from semi-annual or quarterly to annual periods and vice
    versa per the skip-period monitoring plan in Section 3.2.3 or
    some appropriate modification of it.

B.  Allowable Percentage Of Valves Leaking

        The CTG mentions a two percent allowable percentage of valves
    leaking as "a reasonable performance level.  This is the same as
    the NSPS and is too low for RACT.  It should be on the order of
    4% or higher and based on a valid cost-effectiveness analysis,
    SOCMI maintenance effectiveness, the high-density polyethylene
    data, etc.  This number is critical to the reasonableness of RACT
    and should be set"based on mass emissions and the best computer
    analysis of SOCMI data.

C.  Act.ion_ Level (Leak-rDef inition)

        The EPA. has recommended 10,000 ppta or greater observed during
    monitoring as the definition of a. leak.  From the start the ICC
    has argued chat this level is too low and extensive comments and
    reasons have been set forth in Reference 10 (page 2) and
    Reference 11 (pages 25-29).  We recommend the EPA define a. leak
    in terms of a component's' mass emission rate and that they not
    specify a single rate, for the CTG but give the states a choice of
    several.  The concentration level corresponding to the mass
    emission race for the chemicals in question could then be usaa
    for screening purposes.  Giving states a choice of several levels
    would let them tailor a control plan to fit their needs.
                              B-29

-------
                            - 5 -



Capping Of Opfen-Ended Lines

    Okay.  See comments page 24 of Reference 11.

Weekly Pump Inspection

    Okay.  See comments page 18 of Reference 11. ''

Allowable Internal Before Repair

    Generally okay.  See comments on possible delay beyond
scheduled unit shutdown on pages 18-21 of Reference 11.

Definition Of Light Liquid

    Based on petroleum refining splits, a light liquid is
defined as one having a vapor pressure greater than 0.3 kPa
(0.044 psia) at 20°C.  We believe a more rational basis
would be in terms of the vapor pressure at 20°C that equals
the concentration units equivalent to the action level.  For
example, for the EPA's proposed 10,000 ppm level this would
be 0.01 atmospheres or 1.0 kPa (0.147 psia).

Unsafe & Difficult To Reach Components

 •. The discussion of this in Par.1 3.3.1 suggests that for
safety reasons the state may wish to require less frequent
monitoring of certain components in hazardous service.  This
has been added based on TCC/CMA comments but doesn't quite
capture our concern.  Certain processes are carried out at
such extreme conditions that access is not allowed anytime
the unit is in operation.  Thus monitoring while the unit is
in operation is not possible.  This and some alternative
monitoring possibilities are discussed on page 33 of Ref. 11.

Exclusion For Small Valves
    Par. 2.3.1 of the CTG states that fugitive emissions are
not related to capacity, throughput, age, temperature, or
pressure.  We do not find this stated in the reference
document (Ref. 14, pages 11-49).  What we did find was a
statement that source and stream types could be grouped such
that three equations were adequate for predicting leak rates
from screened sources (see pages 11-12).  But this is not
quite the same and in Reports 3,6 and 7 significant effects
were found for pressure and ambient temperature when
analyzing SOCMI data.
                           B-30

-------
                                      - 6 -
              Further we have recent data relevant to valve size' from a
          refinery hydroprocessing unit handling light hydrocarbons•at
          high pressures and temperatures - out of some 5,000 valves
          screened, 200 leaked (>10,000 ppmv) but none of these were  in
          valves 2" or smaller although there were numerous valves 2"
          and smaller in the unit.  We request the EPA to re-analyze
          their data specifically on this point to determine if small
          size valves can't be excluded from the monitoring requirement,
                                                    A. H. Nickolaus
                                                    January 14, 1982
AHN/rtg
                                  B-31

-------
ZRENCZS                                                 c   •

  "Evaluation of Maintenance for Fugitive VOC Emissions Control", EPA.-600/52-81-080,
  May, 1981.

  Blacksmith, Harris, and Langley,  "Frequency of Leak Occurrence for Fittings in
  Synthetic Organic Chemical Plant  Process Units", September, 1980.

, 6.,, & 7.

  "Analysis of SOCHI VOC Fugitive Emissions Data", EPA-600/2-81-111, June, 1981.

  "Response Factors of VOC Analyzers at  a Meter Reading 10,000 ppmv for Selected
  Organic Compounds", EPA-600/2-81/-51,  March, 1981.

  "Response of Portable VOC Analyzers  to Chemical Mixtures", EPA-600./2-81-110,
  June, 1981.

  "VOC Fugitive Emissions Data — High Density Polyethylene Process Unit", EPA-
  600/2-81-109, June, 1981.

  "Flare Sampling and Analysis Methods Development - Nineteenth Monthly Progress
  Report", April 1-30, 1981.

  "Texas Chemical Council Comments  on  EPA's Preliminary Draft CTG For  Control of
  VoTatile Organic Fugitive Emissions  From Synthetic Organic Chemical, Polymer,
  And Resin Manufacturing Equipment" given by A. H. Nickolaus at the March 17-18,
  1981 National Air Pollution Control  Technical Advisory  Committee  (NAPCTAC)
  needing.              •                                                ,

  QtA. & TCC, "Comments of The Chemical Manufacturers Association and The  Texas
  Chemical Council on EPA's Proposed NSPS for VOC Fugitive Emission Source^ Under
  The Clean Air Act,-Docket No..A-79-32", April 6, 1981.

  Letter: A. H. Nickolaus (TCC) to  Don R. Goodwin  (EPA),  "Draft Control Technique
  Guideline Document For Control of VOC  Emissions From  Manufacture of  Polyethylene,
  Polypropylene, and Polystyrene Resins, May 29, 1981,  Page 3.

  "VOC Fugitive Emissions in SOCMI-BID For Proposed Standards", EPA-450/3-80-033a,
  November, 1980.

  "Emission Factors and.Frequency of Leak Occurrence for  Fittings in Refinery
  Process Units", EPA-600/2-79-044, February, 1979. .

  Based on emission rates in Table  4-4 of the BID  (Ref. 14) and the nuaber of
  valves in gas and light liquid service in the Model B plant (Table 6-1  of the
  BID).

  A. H. Nickolaus (TCC) to Central  Docket Section  (A-130) Attn: Docket No.
  A-79-32, "RE: Proposed VOC Fugitive  Emission Regulations For SOCMI Sources,
  Consents on Reports No. 1 thru 8  (46 FR 21789, April  14, 1981)", July 27, 1981.

  J. S. Matey (Chemical Manufacturers  Association) to Central Docket Section
  (A-130) Attn: Docket No, A-79-32, "RE: Proposed  ••• VOC Fugitive Emission
  Sources ••• April 14, 1981 Review of Background Reports", August 7,  1981.

-------
                                     TABLE  I
                      TEXAS CHEMICAL  COUNCIL  (TCC)  COMMENTS
                        TO THE EPA DURING  THE DEVELOPMENT
                OF VOC FUGITIVE EMISSIONS  MONITORING  REGULATIONS
May 17, '1979



February 1, 1980


June 30, 1980


July 28, 1980


July 30, 19.80



March 3, 1981


March 17, 1981


March 23, 1981



April 6, 1981


June 12, 1981



Jtfly 27, 1981
-  Letter From H. H. McClure  (TCC) to David R. Patrick  (EPA).
   Comments on the March 1979 Hydroscience Report on
   Fugitive Loss Control Option.

-  Letter From H. H. McClure  (TCC) to Jack R. Farmer  (EPA).
   Comments on the Draft Background Information Document.

-  Letter From TCC to EPA.  Comments on the Draft BID and
   Recommended SOCMI Standard.

-  Letter From H. H. McClure  (TCC) to Walter Barber  (EPA).
   "Texas Chemical Council Data On Capital  'Creep™.

-  Letter From H. H. McClure  (TCC) to Walter Barber  (EPA).
   "TCC/EPA_Conference on Proposed SOCMI Fugitive Emission
   NSPS".

-  TCC Testimony At the Public Hearing on the SOCMI Fugitive
   Emissions Monitoring NSPS.

-'  TCC Testimony at the March 17-18, 1981 NAFCTAC Meeting
   on the Preliminary Draft- CTG.                     •

-  TCC Clarifying Comments on Questions Raised at the NAPCTAC
   Meeting.  Letter From A. H. Nickolaus  (TCC) to
   Don R. Goodwin (EPA). •

   CMA/TCC Joint -Written Comment  on the Proposed NSPS
   (Docket No. A-79-32). '

-  Letter From A. H. Nickolaus  (TCC) to Don R. Goodwin  (EPA)
   on Questions Raised at the Petroleum Refining NSPS Review
 '  before the NAPCTAC on June 3,  1981.

-  TCC Comments on the SOCMI  Studies to Docket No. A-79-32.
AHN/rtg

1-14-82
                                     B-33

-------
                                          TABLE 2


                            COST EFFECTIVENESS FOR MODEL UNIT B
                                           CTG                   TCC
                                      TABLES 5-6,10         ''   ESTIMATE
3C EMISSIONS, MG/YR
 "ONCOHTROLLED"                            260                    78
 CONTROLLED                                 90.4                  50.9

   REDUCTION                               169.6                  27.1


TOTAL COSTS BEFORE CREDITS

 PUMPS            .                      $9,486                 $9,486

 VALVES

   GAS                                   1,836                 4,168
   LIGHT LIQUID                          1,771                 4,020

 SAFETY/RELIEF VALVES                    1,128    ' •            1,128

 Qg£N-£NPEO VALVES

   GAS SERVICE                             698                 1,584
   LIGHT LIQUID SERVICE                  3,532                 8,018
   HEAVY LIQUID SERVICE  '    '            2,535                 2,535
 COMPRESSORS              •               1,240                 1,240     .
 INSTRUMENT COSTS                        5,494        •         5,494

   TOTAL                      '         $27,720  '              $37,673

SCOVERY CREDIT @ 5410/MG               $69,540                $11,111

ET COST                                ($41,823)               $26,562

OST EFFECTIVENESS, $/HG                    (247)                  980


AS.IS FOR TCC ESTIMATE

 Balssidas;  30% of CTG Based on SOCMI Data - See Attachment  2.

 Monitoring Time;  117 Man-sainutes Per Component - Table 2-2  Ref. 2.

 Labor Efficiency:  75% Based on Experience, Allows Time for  Training, Safety
                    Meetings, Breaks,  Etc.

 Haintanance Efficiency:  71.3% per Ref, 1
                                   B-34

-------