United States      Office of Air Quality       EPA-450/3-82-024b
           Environmental Protection  Planning and Standards      May 1985
           Agency        Research Triangle Park NC 27711

           Air
vvEPA      Equipment Leaks    EIS
           of VOC in  Natural
           Gas Production
           Industry —
           Background
           Information for
           Promulgated
           Standards

-------
                              EPA-450/3-82-024b
       Equipment Leaks of VOC
in Natural Gas Production Industry -
       Background  Information
            for Promulgated
                Standards
           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 2771 1

                   May 1985

-------
This report has been reviewed by the Emission Standards and Engineering Division of the Off ice of Air Quality Planning
and Standards, EPA, and approved for publication. Mention of trade names or commercial products is not intended to
constitute endorsement or recommendation  for use. Copies of this report are available through the Library Services
Office (MD-35), U.S. Environmental Protection Agency, Research Triangle Park, North Carolina 27711;or, fora fee, from
the National Technical Information Services, 5285 Port Royal Road. Springfield, Virginia 22161.

-------
                     ENVIRONMENTAL  PROTECTION AGENCY
                         Background  Information
                and  Final Environmental  Impact Statement
  for  Equipment  Leaks of VOC from Onshore Natural Gas  Processing Plants
Jack R.  Fanner
Director,  Emission Standards and Engineering Division
U.S. Environmental Protection Agency
Research Triangle Park, North Carolina  27711

1.  The  promulgated standards of performance will limit equipment leaks
    of VOC from new, modified, and reconstructed gas plant process
    units and compressors.  Section 111 of the Clean Air Act (42 U.S.C.
    7411), as amended, directs the Administrator to establish standards
    of performance for any category of new stationary source of air
    pollution that ". . . causes or contributes significantly to air
    pollution which may reasonably be anticipated to endanger public
    health or we! fare."

2.  Copies of this document have been sent to the following Federal
    Departments:  Labor, Health and Human Services, Defense, Transportation,
    Agriculture, Commerce, Interior, and Energy; the National Science
    Foundation; the Council on Environmental  Quality; State and
    Territorial Pollution Program Administrators; EPA Regional
    Administrators; Local Air Pollution Control Officials; Office of
    Management and Budget; and other interested parties,

3.  For additional  information contact:

    Ms. Dianne Byrne or Mr. Gilbert Wood
    Standards Development Branch (MD-13)
    U.S.  Environmental  Protection Agency
    Research Triangle Park, NC  27711
    Telephone:  (919) 541-5578

4.  Copies  of this  document may be obtained from:

    U.S.  EPA Library (MD-35)
    Research Triangle Park, NC  27711
    Telephone:  (919)  541-2777

    National  Technical  Information Service
    5285  Port Royal  Road
    Springfield,  VA  22161
                                   11

-------
                                TABLE OF CONTENTS

Section        Title                                              Page

               TABLE OF CONTENTS	  v

               LIST OF TABLES	  viii

               LIST OF FIGURES	  ix


  1.0          SUMMARY 	  1-1

               1.1  Introduction 	  1-1

               1.2  Summary of Changes Since Proposal  	  1-1

               1.3  Summary of Impacts of Promulgated  Action ...  1-4

               1.4  Summary of Public Comments 	  1-5


  2.0          NEED FOR STANDARDS  	  2-1

               2.1  Significance of Emissions  and Public Health
                    Impact 	  2-1

               2.2  Adequacy of Current Controls  and
                    Regulations  	  2-3

               2.3  Industry Growth,  Industry  Mobility,
                    and Seasonal  Considerations   	  2-5


  3.0          BASIS FOR STANDARDS   	  3-1

               3.1  Introduction  	  3-1

               3.2  Valves  	  3-1

               3.3  Pumps   	  3-1

               3.4  Compressors  	  3-2

               3.5  Open-Ended  Lines   	  3-2

               3.6  Pressure Relief Devices  	  3-2

               3.7  Sampling Connections     	  3-4

-------
                       TABLE  OF  CONTENTS  (Continued)



Section        Title                                             pa ge



  4.0          FORMAT AND  REQUIREMENTS  OF  STANDARDS  	  4-1



               4.1  Format of Standards 	  4-1



               4.2  Valves  	  4-1



               4.3  Pumps    	  4-4



               4.4  Compressors  	  4-5



               4.5  Open-Ended Lines   	  4-7



               4.6  Pressure  Relief Devices  	  4-9



               4.7  Control Devices   	  4-13



               4.8  Leak Detection and  Repair  	  4-18






  5.0          APPLICABILITY  OF  STANDARDS	  5-1



               5.1  Definitions   	  5-1



               5.2  Selection of Sources  	  5-7



               5.3  Selection of Affected  Facilities   	  5-10



               5.4  Applicability Date  	  5-12



               5.5  Alaskan North Slope 	  5-13



               5.6  Small  Plants 	  5-15





  6.0          ENVIRONMENTAL  IMPACTS  	  6-1



               6.1  Emission  Reductions  	  6-1



               6.2  Leak Frequencies   	  6-4



               6.3  Emission  Factors   	  6-5



               6.4  Model  Plants   	  6-7



               6.5  Nonair Environmental  Impacts   	  6-7

-------
                           TABLE OF CONTENTS (Concluded)



Section        Title
                                                                  Page
  7.0          COST OF CONTROL	   7_1




               7.1  General    	   7_1




               7.2  Compressors   	   7.3




               7.3  Pressure Relief Devices 	   7-6




               7.4  Control  Devices 	   7.7




               7.5  Leak  Detection  and  Repair   	   7_g






  8.0          ECONOMIC  IMPACTS   	   8-1



               8.1  Prices   	   8_1




            •   8.2  Industry Impacts   	   8-3



               8.3  Small  Businesses	   8-4






  9.0          MODIFICATION  AND  RECONSTRUCTION	   9-1



               9.1   Capital  Expenditure   	   g_l




               9.2   Modification of Existing Sources  	   9-2



               9.3   Miscellaneous   	   9_3






10.0           TEST  METHODS   	   10_l




               10.1  Leak Detection Methods 	   10-1



               10.2  Leak Definition  	   10-4






11.0          RECORDKEEPING AND REPORTING  	   11-1



              11.1  General   	   ll-l



              11.2  Reporting  	   11-4




              11.3  Recordkeoping   	   11-5






12.0          MISCELLANEOUS   	   12_!
                                VI 1

-------
                              LIST OF TABLES


Number       Title                                                        Paqe
1-1          Projected VOC Equipment Leak Emissions
             from Uncontrolled, Baseline, and NSPS
             Control Levels	   1-6

1-2          List of Cornrnenters on Proposed Standards •
             of Performance for Equipment Leafcs of VOC
             from Onshore Natural  Gas Processing Plants	   1-8

3-1          Control Costs per Megagram of VOC
             Reduced 	   3-3

7-1          Cost Effectiveness of Compressor Vent
             Control Systems for Model Plant B  	  7-5

7-2          Summary of Total Annual Costs to Perform
             Leak Detection and Repair Program at a
             500-Componen.t Gas Plant 	   7-12
                                            VII 1

-------
                               LIST OF FIGURES

Number         Title                                                    Page
5-1            Cost Effectiveness versus Plant
               Size for Small  Plants 	   5-16
                                      i x

-------
                              1.0  SUMMARY

1.1  INTRODUCTION

     On January 20, 1984, the Environmental  Protection Agency (EPA)
proposed standards of performance for equipment leaks of volatile
organic compounds (VOC) emissions from onshore natural  gas processing
plants (49 FR 2636) under the authority of Section 111 of the Clean  Air
Act.  Public comments were requested on the proposal  in the Federal
Register.  EPA received 37 comment letters on the proposed standards
and statements from five speakers at the public hearing on March 7,  1984.
Industry representatives submitted most of the comments.  Also commenting
was a vendor of safety equipment.  The comments that  were submitted,
along with responses to these comments, are summarized in this document.
This summary of comments and responses serves as the  basis for the
revisions made to the applicability and requirements  of the standards
between proposal  and promulgation.

1.2  SUMMARY OF CHANGES SINCE PROPOSAL

     The proposed standards were revised as a result  of reviewing public
comments.  Changes were made in the following areas:

     •  Exemption for reciprocating compressors in wet gas service

     •  Revisions to flare requirements

     •  Definition of "in VOC service"

     •  Monitoring requirements for pressure relief devices at certain
        pi ants

     t  Provision for difficult-to-monitor valves in  new units

     t  Leak detection and repair requirements for natural gas processing
        plants located in the North Slope of Alaska

     •  Alternative for determining a "capital expenditure"

1.2.1  Exemption for All Reciprocating Compressors in Wet Gas Service

     At proposal, EPA exempted  reciprocating compressors in wet gas
service only if they were located at a gas plant that did not have an
existing control  device.  The cost effectiveness of controlling such
compressors was high due to the cost of installing and operating a
control device.  However, the cost effectiveness of controlling wet
gas reciprocating compressors at plants with an existing control
device ($1700 per megagram of VOC reduced) was considered reasonable,
given that the average cost effectiveness (combining  cost-effectiveness
numbers for centrifugal and reciprocating compressors) was estimated
to be much lower ($460 per megagram).  However, since proposal, several
industry representatives commented that many gas plants, especially
small  ones, will  use reciprocating compressors almost exclusively.
For such plants, the compressor control cost effectiveness would be
essentially the same as the cost effectiveness for controlling only

                                 1-1

-------
wet gas reciprocating compressors  at plants with  an  existing  control
device (i.e., $1700 per megagram).   This  cost  effectiveness,  when
considered representative of the overall  compressor  control costs  for
small plants, is judged to be unreasonably high.   For  this  reason,
the promulgated standards have been revised to exempt  all wet gas
reciprocating compressors.  Reciprocating compressors  used  in natural
gas liquids (NGL) service and all  centrifugal  compressors  in  wet gas or
NGL service are still required to  be equipped  with closed  vent systems,
however, because they can be controlled  at a reasonable  cost  effectiveness.

1.2.2  Revisions to Flare Requirements

     The velocity and heating value requirements  for flares have been
changed since proposal  of the standards  to allow  flares  burning gas
streams with high heating values to use  high velocities.   The final
standards present equations for calculating the maximum  permitted
velocity for flares to provide for velocities  up  to  122  meters per
second (m/sec) (400 feet per second (ft/sec))  depending  on  the gas heat
content.  The purpose of the equations  is to allow streams  with heat
contents greater than 11.2 megajoules per standard cubic meter (MJ/scm)
(300 British thermal  units per standard  cubic  foot (Btu/scf)) to be
flared at higher velocities, while ensuring a  VOC destruction efficiency
that reflects best demonstrated technology (BDT).  If  the  net heating
value of the gas being combusted is greater than  37.3  MJ/scm  (1,000
Btu/scf), then a flare exit velocity of  122 m/sec (400 ft/sec) will  be
accepted.  The basis  for these changes  in the  flare  requirements  is
discussed in Section  4.7.

1.2.3  Definition of  "In VOC Service"
     The EPA received numerous requests from commenters for raising  the
VOC concentration for "in VOC service" from 1 weight percent to 10
weight percent VOC.   The commenters stated that a 1 weight percent
VOC limit would cause many dry (residue) gas streams with low VOC content
to be regulated, and, as a result, poor cost effectiveness.

     The EPA intended to exempt dry gas in the proposed standards due
to the fact that dry gas has very low (generally less than 1.0 weight
percent) VOC content and, therefore, is generally not cost effective to
control.  The EPA did, however, intend for wet gas (inlet or field gas)
components to be regulated, since components in wet gas service have
a high VOC content and are cost effective to control.  At proposal,
EPA selected 1.0 weight percent VOC as the VOC service definition to
exclude dry gas, while assuring that wet gas streams would be regulated.

     Commenters on the proposed standards provided data indicating that
1 weight percent was not an appropriate limit, since dry gas streams
can contain more than 1 weight percent VOC.  The EPA agrees that it  is
difficult to define a range of VOC concentrations for either wet gas
or dry gas streams and, therefore, has decided to include equipment  in
wet jas service by covering them as a class.

     Considering that gas plant streams, other than wet gas streams,
containing less than 10 weight percent VOC (dry gas or VOC/non-VOC

                                   1-2

-------
mixtures) would not be cost effective to control, a VOC concentration
limit of 10 weight percent was selected as a representative limit for
the "in VOC service" definition.  Therefore, the definition of "in VOC
service" has been changed in the promulgated standards to refer to a
10 we-ight percent VOC content.

1.2.4  Monitoring Requirements for Pressure Relief Devices

     Since proposal, EPA has reviewed comments concerning the
monitoring of pressure relief devices following pressure release.
Commenters representing small nonfractionating gas plants claimed
that the requirement to monitor pressure relief devices within 5
days of a pressure release was burdensome because of the unavailability
of monitoring instruments at the plant and problems in scheduling
contractor assistance in such a short time.  Based on these comments,
EPA decided to allow owners and operators of nonfractionating
plants which are monitored only by non-plant personnel to monitor
pressure relief devices the next time the monitoring personnel are on-
site or within 30 days of a pressure release, whichever cornes first,
instead of within 5 days.

1.2.5  Provision for Difficult-to-monitor Valves in New Units

     At proposal, EPA allowed annual  monitoring of difficult-to-monitor
valves in units covered by the modification or reconstruction
provisions, but there was no similar provision for difficult-to-monitor
valves in new units.  Commenters stated that all new units could not
be designed to eliminate difficult-to-monitor valves.  Upon reviewing
these comments, EPA decided to allow plant owners or operators to
designate up to 3 percent of the total number of valves in a new unit
as difficult to monitor.  Based on existing units, about 3 percent of
the total  number of valves may be difficult to design to be accessible
in new units without significant additional costs.

1.2.6  Leak Detection and Repair Requirements for Natural Gas
       Processing Plants Located in the North Slope of Alaska

     The EPA reviewed comments concerning natural gas plant operations
in the North Slope of Alaska and detennined that the costs to comply
with certain aspects of the proposed standards can be unreasonable.
Leak detection and repair programs incur higher labor, administrative,
and support costs at plants that are located at great distances from
major population centers and particularly those that experience extremely
low temperatures as in the arctic.  Thus, EPA decided to exempt
plants located in the North Slope of Alaska from the routine leak
detection and repair requirements.  The EPA excluded these plants only
from the routine leak detection and repair requirements because the
costs of the other requirements are reasonable.

1.2.7  Alternative for Determining a  "Capital  Expenditure"
     The annual  asset guidel i ne repair allowance (AAGRA]~and the
original  cost basis are used to define capital  expenditures (see 40 CFR
60.2).   The definition of AAGRA is specified by the Internal  Revenue
Service (IRS), and its use has not changed despite tax law revisions  in

                                  1-3

-------
1982.  In response to comments concerning the basis for determining d
capital  expenditure, EPA has amended the definition of a capital
expenditure to provide an alternative procedure for determining the
original  cost basis.  This procedure will allow plant personnel to use
a percentage of replacement costs rather than original  costs to determine
capital  expenditure; however, this procedure does not affect the  coverage
of the standards.  Details of the alternative are discussed in Chapter 9.

1.3  SUMMARY OF IMPACTS OF PROMULGATED ACTION

1.3.1  Alternatives to Promulgated Action

     The regulatory alternatives are discussed in Chapter 6 of the
background information document (BID) for the proposed standards
EPA-450/3-82-024a.  These regulatory alternatives reflect the different
levels of emission control that were used in the selection of best
demonstrated technology (BDT), considering costs, nonair quality health,
and environmental and economic impacts for sources of equipment leaks
of VOC in the natural gas processing industry.  No changes have been
made in the regulatory alternatives since proposal.

1.3.2  Environmental Impacts of Promulgated Action

     Environmental  impacts of the standards are described in Chapter
7 of the BID for the proposed standards.  Changes made  in the  standards
since proposal affect the estimated VOC  emission reduction.  These
changes, which will reduce the industry-wide emission reduction of the
standards, include a change in the estimated compressor service split,
the exemption for reciprocating compressors in wet gas  service, and  the
effects of the Control Techniques Guideline (CTG) requirements
(EPA-450/3-83-007), which have been published since  proposal of the
standards.  The  EPA has revised the estimate of VOC  emissions  that
will be reduced  by  the standards.

     The promulgated standards will reduce equipment leaks of  VOC  from
newly constructed,  modified, and  reconstructed facilities by  16,100
megagrams  (Mg) in the fifth year  of implementation of  the standards,  a
reduction  of emissions from 22,000 Mg of VOC per year  (Mg/yr)  to  5,900
Mg/yr.  This reduction represents a 73  percent decrease in emissions
from the current  industry baseline level  of emissions.  The basis  of
the  revised estimate of VOC emissions that will  be reduced  by  the
standards  is discussed in Docket  Item IV-B-9.  The water  quality  and
solid waste impacts  have  not changed  significantly as  a result of  the
revised emission  estimates.

     With  the changes noted above, the  analysis  of the  environmental
 impacts in the BID  for the  proposed standards  now  becomes  the  final
 Environmental  Impact Statement  for  the  promulgated standards.

 1.3.3   Energy and  Economic  Impacts of Promulgated  Action

      Section  7.5 of the  BID for  the  proposed  standards  describes  the
 energy  impacts of the standards.   The revisions  made in the  standards
 do  not  change  significantly the  energy  impacts  of  the  standards

                                   1-4

-------
 because the energy value and crude oil  equivalent of VOC emission
 reductions are about the same as they were at proposal.

      Chapters 8 and 9 of the BID for the proposed standards  describe
 the cost and economic impacts of the proposed standards.   Since
 proposal, the cost analysis of the standards  has been revised  to  reflect
 the changes discussed above that have been made  since proposal.   The
 industry-wide cost of the standards is  lower  than the estimate at
 proposal  because of the effects of CTG  requirements  for  onshore natural
 gas processing plants (published since  proposal)  in  the  fifth  year  of
 the standards and because of the exemption of reciprocating  compressors
 in wet gas service from the standards since proposal.   The promulgated
 standards will  require an industry-wide capital  investment and net
 annualized cost of approximately $6.2 million and $1.5 million,
 respectively, for newly constructed, modified, and reconstructed  facili-
 ties  in the fifth year of implementation of the  standards.   The basis
 of the revised industry-wide costs is discussed  in Docket  Item IV-B-9.
 As discussed in Chapter 8 of this  document, the  economic  impact of  the
 promulgated standards remains  reasonable.

 1.3.4   Other Considerations

      1-3.4.1  Irreversible and  Irretrievable  Commitment.   Section 7.6.1
 of the BID for  the proposed  standards concludes  that  the standards
 will  not  result in any irreversible or  irretrievable  commitment of
 resources.   It  was also  concluded  that  the  standards  should  help  to
 save  resources  due to the energy savings  associated with the reduction
 in emissions.   These  conclusions remain  unchanged since  proposal.

      1.3.4.2  Environmental  Impacts  of  Delayed Standards.  Table  1-1
 of this document  summarizes  the  environmental   impacts associated
 with delaying  promulgation of  the  standards.   The  air impacts  have
 been revised since proposal  as  explained  in Table  1-1.  Uncontrolled
 and controlled  emissions  would  occur  at  the rates  shown for  each  of
 the 5  years.   Fifth-year  emission  reductions  from  the uncontrolled
 level  and  baseline level  are presented.

 1.4  SUMMARY OF PUBLIC  COMMENTS

     EPA received  37  comment letters  and statements from five  speakers
 at  the public hearing on  the proposed standards and the BID  for the
 proposed standards.   Comments from the public  hearing on the proposed
 standards were  recorded, and a transcript of  the  public hearing was
 placed in the project docket along with all public comment letters.
 At the request  of  some of  the commenters, the  comment period was
 extended 60  days to allow more time  for review and comment.  A list
 of commenters,  their affiliations, and the EPA docket item number
assigned to  their correspondence is given in Table 1-2.
                                 1-5

-------
           TABLE 1-1.   PROJECTED  VOC  EQUIPMENT  LEAK EMISSIONS
        FROM UNCONTROLLED, BASELINE,  AND NSPS  CONTROL  LEVELS
(Proposal Basis) (Promulgation Basis)
YEAR Uncontrolled*
1983 5.1
1984 10.2
1985 15.3
1986 19.5
1987 23.6
5th year (1987) • —
Emission Reduc-
tion from
Uncontrol led
5th year (1987) —
Emission Reduc-
tion from Base-
line
Cost
effectiveness
($/Mg)
Proposed
NSPSb Uncontrol ledc
1.2 5.1
2.4 10.1
3.6 15.2
4.5 19.3
5.5 23.5
18.1
— ~ ~~
140
Proposed
Basel 1ned NSPSe
4.7 1.2
9.5 2.5
14.2 3.7
18.1 4.7
22.0 5.7
1.5 17.8
16.3
150
Final
NSPSf
1.3
2.6
3.9
4.9
5.9
17.6
16.1
93
 From  Docket  Item IV-B-9.  The  uncontrolled emissions  assumed at proposal  that all
 natural  gas  processing plants  are operated with no regulations for control  of
 equipment  leaks of VOC.
b
 Emission levels under the proposed standards are based  on  Docket Item II-B-38,
 which  presents the controlled  emission levels for each  model plant, and Table 7-5
 of the BID for the proposed standards, which presents the  anticipated number of
 model  plants affected.  The memo did not take Into account, however, that one-half
 of all  reciprocating compressors in wet gas service would  be exempt, which  was
 assumed  at proposal.  Also assumes that 34 percent of all  compressors are in NGL
 service  and  that 66 percent are  in wet gas service.

 Since  proposal, as explained in  Docket Item IV-B-9. the uncontrolled emission estimates
 have  been  revised due to a change In the assumed distribution of compressor service.
 This  results In a change In emissions of 100 Mg/yr in 1987 from the uncontrolled level
 reported at  proposal as there  are now assumed to be more compressors In wet gas
 service  and  less in NGL service.
                                         1-6

-------
                                       TABLE 1-1

                         PROJECTED VOC EQUIPMENT LEAK EMISSIONS
            FROM UNCONTROLLED, BASELINE, AND NSPS CONTROL LEVELS (Concluded)

d
 The baseline level  of control assumes that 10 percent of new gas plants will be
 built in ozone nonattainment areas and will be subject to requirements similar to
 those in the CTG for natural gas processing plants (Docket Item II-J-1).  Calculated as
 described in Docket Item IV-B-9.
e
 Emission levels under the proposed standards reflect the fact that wet gas
 reciprocating compressors in plants with an existing control  device were required
 to be controlled.  The final NSPS does not require controls on these compressors.
 However, emissions'under the proposed standards are higher than assumed at
 proposal because of a change in the ratio of compressors in NGL and wet gas
 service (i.e., 50% of all compressors are exempt wet gas reciprocating units,
 and the remaining 50% are split equally between wet gas centrifugal units and NGL
 units of both types).  Under these conditions, 25 percent of all compressors
 would be exempt from the proposed standards, since they would be wet gas
 reciprocating units at plants without a control device.

 Revised emission levels from implementing the final standards are based on
 Docket Item IV-B-9 and are slightly higher than the emission levels under the
 proposed standards as a result of the exemption of all  wet gas reciprocating
 compressors.  This exemption decreases the emission reduction from the proposal
 level by 200 Mg/yr in 1987, resulting in a total  emission reduction of 16,100
 Mg/yr from the baseline level of control.
                                              1-7

-------
   TABLE 1-2.  LIST OF COMMENTERS ON PROPOSED STANDARDS OF  PERFORMANCE
                    FOR EQUIPMENT LEAKS OF VOC FROM ONSHORE
                         NATURAL GAS PROCESSING PLANTS
Commenter and Affiliation                                   Docket  Item  No.
1.  Mr. J.D. Reed                                                 IV-D-1
    Standard Oil  Company (Indiana)
    200 East Randolf Drive
    Chicago, IL  60601

2.  Mr. Louis R.  Harris                                           IV-D-2
    BS & B Safety Systems, Inc.
    7455 East 46th Street
    P.O. Box 470590
    Tulsa, OK  74147-0590

3.  Mr. Rodney D. Long                                            IV-D-3
    Minerals, Inc.
    P.O. Box 1320
    Hobbs, NM  88240

4.  Mr. Greg Lewis                                                IV-D-4
    Liquid Energy Corporation                                     IV-D-21
    2001 Timberlock Place
    P.O. Box 4000
    The Woodlands, TX  77380

5.  Mr. J.D. Geiger                                               IV-D-5
    Aminoil  , USA                                                  IV-D-23
    P.O. Box 94193
    Houston, TX  77292

6.  Mr. R.E. Cannon                                               IV-D-6
    Gas Processors Association                                    IV-D-26
    1812 First Place
    Tulsa, OK  74103

7.  Mr. John J. Moon                                              IV-D-7
    Phillips Petroleum Company                                    IV-D-31
    Bartlesville, OK  74004

8.  Mrs. Laura G. Daniel                                          IV-D-8
    Conoco, Inc.                                                   IV-D-36
    P.O. Box 2197, Suite 450 RT
    Houston, TX  77252
                                    1-8

-------
    TABLE  1-2.   LIST OF COMMENTERS ON PROPOSED STANDARDS OF PERFORMANCE
                  FOR EQUIPMENT LEAKS OF VOC FROM ONSHORE
                       NATURAL  GAS PROCESSING PLANTS
                                (conti nued)
 Commenter  and  Affiliation
Docket Item No.
 9.   Mr.  H.K.  Holland,  Jr.
     Mobil  Oil  Corporation
     150  East  42nd  Street
     New  York,  NY   10017

 10.  Mr.  A.G.  Smith
     Shell  Oil  Company
     One  Shell  Plaza
     P.O. Box  4320
     Houston,  TX  77210

 11.  Mr.  Robert H.  Lovell
     Mountain  Fuel  Supply Company
     180  East  First South
     P.O. Box  11368
     Salt Lake  City, UT  84139

 12.  Mr.  A.E.  Middents
     Western Slope  Gas Company
     One  Park  Central  - 1515 Arapahoe Street
     P.O. Box 840
     Denver, CO  80201

 13.  Mr.  Peter W. McCallum
     The  Standard Oil  Company (Ohio)
     Midland Building
     Cleveland , OH  44115-1098

 14.  Mr. B.L. Walters, Jr.
     Marathon Oil  Company
     Findlay, OH  45840

15. Mr. H.B. Barton
     Exxon Company,  U.S.A.
    P.O.  Box 2180
    Houston, TX  77001

16. Mr. W.W. Cofield
    Transco Energy  Company
    2700  Post Oak  Boulevard
    P.O.  Box 1396
    Houston, TX  77251
     IV-D-9
     IV-D-10
     IV-D-11
     IV-D-lla
     IV-D-12
     IV-D-13
     IV-D-14
     IV-D-15
     IV-D-16
                                    1-9

-------
   TABLE 1-2.  LIST OF COMMENTERS ON PROPOSED STANDARDS OF PERFORMANCE
                 FOR EQUIPMENT LEAKS OF VOC FROM ONSHORE
                      NATURAL GAS PROCESSING PLANTS
                               (continued)
Commenter and Affiliation                                   Docket Item No.
17. Mr. Steven E.  Kurmas, P.E.                                    IV-D-17
    Michigan Consolidated Gas Company
    500 Grinswold  Street
    Detroit, MI  48226

18. Mr. P.K. Smith, Jr.                                           IV-D-18
    Kerr-McGee Corporation
    Kerr-McGee Center
    Oklahoma City, OK  73125

19. Mr. H.  Schuyten                                              IV-D-19
    Chevron U.S.A., Inc.
    575 Market Street
    P.O. Box 7643
    San Francisco, CA  94120-7643

20. Mr. J.P. Keehan                                              IV-D-20
    Mobil  Oil  Corporation
    150 East 42nd  Street
    New York, NY  10017

21. Mr. L.  James Anderson   '                                     IV-D-22
    Union Oil  of California
    Union Oil  Center
    Box 7600
    Los Angeles, CA  90051

22. Mr. J.  Donald  Annett                                         IV-D-24
    Texaco  U.S.A.
    1050 17th Street, N.W.
    Suite 500
    Washington, D.C.  20036

23. Mr. George H.  Lawrence                                       IV-D-25
    The American Gas Association
    1515 Wilson Boulevard
    Arlington, VA  22209

24. Mr. Jack Swenson                                             IV-D-27
    Rocky Mountain Oil and Gas
      Association, Inc.
    345 Petroleum  Building
    Denver, CO  80202
                                   1-10

-------
   TABLE 1-2.  LIST OF COMMENTERS ON PROPOSED STANDARDS OF PERFORMANCE
                 FOR EQUIPMENT LEAKS OF VOC FROM ONSHORE
                      NATURAL GAS PROCESSING PLANTS
                               (conti nued)
Commenter and Affiliation
25. Mr. Stuart C. Mut
    ARCO Oil  and Gas Company
    P.O. Box  2819
    Dallas, TX  75221

26. Mr. Frank J. Duffy
    Northern  Gas Products Company
    2223 Dodge Street
    Omaha, NE  68102

27. Mr. R.J.  Cinq-Mars
    Cities Service Oil and Gas Corporation
    Box 300
    Tulsa, OK  74102

28. Dr. Howard Reiquam, Ph.D.
    El  Paso Natural  Gas Company
    P.O. Box  1492
    El  Paso,  TX  79978

29. Mr. P.J.  Early
    Amoco Production Company
    200 East  Randolf Drive
    P.O. Box  5340A
    Chicago,  IL  60680

30. Mr. Lawrence J.  Ogden
    Interstate Natural Gas Association
      of America
    1660 L Street, N.W.
    Washington, D.C.  20036-5611

31. Mr. John  G. Blackburn, Jr.
    American  Petroleum Institute
    1220 L Street, N.W.
    Washington, D.C.  20005

32. Mr. L.T.  Reed
    Warren Petroleum Company
    P.O. Box  1589
    Tulsa, OK  74102
Docket Item No.
      IV-D-28
      IV-D-29
      IV-D-30
      IV-D-32
      IV-D-33
      IV-D-34
      IV-D-35
      IV-D-37
                                   1-11

-------
   TABLE 1-2.   LIST OF COMMENTERS ON PROPOSED STANDARDS OF PERFORMANCE
                 FOR EQUIPMENT LEAKS OF VOC FROM ONSHORE
                      NATURAL GAS PROCESSING PLANTS
                               (concluded)
Commenter and Affiliation                                   Docket Item No.
  Public Hearing

33.  Mr.  Jim Anderson                                              IV-F-la
    Gas  Processors Association
    1812 First Place
    Tulsa, OK  74103

34.  Mr.  William S. Taylor                                         IV-F-lb
    Union Texas Petroleum Corporation
    P.O. Box 2120
    Houston, TX  77252

35.  Mr.  Gary Reed                                                 IV-F-lc
    (Representing the Independent Petroleum
     Association of America)
    Texas Oil & Gas Corp.
    First City Center
    Lock Box No. 10
    Dallas, TX  75201

36.  Mr.  Scott Ronzio                                              iV-F-ld
    ARCO Alaska, Inc.
    Anchorage, Alaska  99510

37.  Mr.  W.J. Woodruff                                             IV-F-le*
    (Representing American Petroleum Institute)
    Phillips Petroleum Company
    Bartlesville, OK  74004
*At the public hearing, API requested that its oral and written comments to
 EPA on the proposed standard and the related CTG be incorporated in the
 record.  These API comments relevant to the basis of the promulgated standards
 have been included in this document.  The docket items are listed in
 parentheses as "II-B-xx" or "II-D-xx" in the appropriate comment summaries.
                                   1-12

-------
     The comments have been categorized under the following
topics:
          Need for Standards (Section 2)
          Basis for Standards (Section 3)
          Format and Requirements of Standards (Section 4)
          Applicability of Standards (Section 5)
          Environmental Impacts (Section 6)
          Control  Costs (Section 7)
          Economic Impacts (Section  8)
          Modification and Reconstruction  (Section 9)
          Test Methods (Section 10)
          Recordkeeping and Reporting (Section 11)
          Miscellaneous (Section 12)

     The comments, the issues they address, and  EPA's responses are
discussed in the following sections  of this document.
                                 1-13

-------
                     2.0  NELL) FOR STANDARDS

     This chapter summarizes public comments and responses to comments
on the need for the standards.   Included are comments on the significance
of emissions and public health benefit of the standards, the adequacy
of current controls and regulations, industry growth and mobility, and
effects of seasonal variations in ozone formation.

2.1  SIGNIFICANCE OF EMISSIONS AND PUBLIC HEALTH IMPACT

     Comment:  Many commenters (IV-D-3; IV-D-20; IV-D-29;  IV-D-31;
IV-F-4) questioned the significance of VOC emissions from  yas plants.
The commenters questioned EPA's contention (or judgment) that VOC
emissions from natural  gas plants contribute to air pollution signifi-
cantly enough to reasonably be expected to endanger the public health
and welfare and to violate the National Ambient Air Quality Standards
(NAAQS) for ozone.  Commenter  IV-D-3 questioned whether propane and
heavier gases break down in sunlight and air or whether they are
dispersed sufficiently so as to be of no significance.  Commenter IV-F-4
claimed that natural  gas processing plants are not large emitters of
VOC, particularly with the exclusion of methane and ethane.  Commenter
IV-D-27 specifically felt that the oil  and gas industry in the Rocky
Mountain region is not a significant contributor of VOC emissions.

     Commenter IV-D-31 added that VOC as a class of compounds are not
criteria pollutants.   Hydrocarbons as a class of pollutants under
the NAAQS were determined by EPA not to pose a direct threat to the
public health or welfare at typical ambient levels, and the NAAQS for
hydrocarbons were revoked on January 5, 1983.

     Response: In response to the commenter's (IV-D-3) question about
the fate of propane and heavier gases when exposed to sunlight and air,
these gases do break down in the presence of sunlight and air; however,
they remain in the atmosphere long enough to participate in photochemical
reactions.  Consequently, propane and heavier gases pose a significant
threat to the environment.

As reported in Table 1-1 of this document, an estimated 20,100 meyagrams
of VOC per year (Mg/yr) would be emitted from about 220 new, modified,
or reconstructed onshore natural  gas processing plants in  1987 if no
additional  controls were implemented by the industry.   On an individual
plant basis, a typical  medium-sized gas plant having nonfractionating
refrigeration units or fractionating cryogenic units is estimated to emit
about 100 Mg/yr assuming no additional  controls.   EPA considers both
nationwide and individual  plant emissions from natural  gas processing
plants to be significant based on the above estimates  and on the listing
of crude oil  and natural  gas production on the EPA Priority List
(40 CFR 60.16, amended  at 47 FR 951, January 8, 1982).

     In consideration of the significant quantity of VOC emissions from
natural  gas processing  plants, EPA examined the "public health and
welfare" effects associated with  these  emissions.   The Administrator
clearly documented  the  need to regulate VOC to protect public
health and welfare  in the  EPA publication "Air Quality Criteria for

                                  2-1

-------
Ozone and Other Photochemical  Oxidants"  (EPA-600/8-78-004,  April  1978).
VOC emissions are precursors to ozone formation,  and  ozone  has  been
determined to be harmful  to human health and has  been shown to  have
environmental effects on  vegetation and  materials as  well  as  other
damage to the ecosystem.   Some of the health and  welfare effects
associated with ozone exposure are the following:

     Human health effects - Ozone exposure has been shown  to  cause
increased rates of respiratory symptoms, such as  coughing,  wheezing,
sneezing, and shortness of breath; increased rates of headache, eye
irritation, and throat irritation; and physiological  damage to  red
blood cells.  Experimental data link ozone exposure to human  cell damages
known as chromosomal  aberrations.

     Vegetation effects - Reduced crop yields as  a result  of  damages  to
leaves and/or plants  are  documented for  several crops, including  citrus,
grapes, and cotton.  The  reduction in crop yields is  shown  to be  linked
to the level and duration of ozone exposure.

     Materials effects -  Ozone exposure  is shown  to accelerate  the
deterioration of organic  materials, such as plastics  and rubber
(elastomers), textile dyes, fibers, and  certain  paints and  coatings.

     Ecosystem effects -  Continued ozone exposure is  linked to  structural
changes of forests, such  as the disappearance of  certain tree species
(Ponderosa and Jeffrey pines)  and death  of some  types of vegetation.
Hence, ozone causes stress to  the ecosystem.

     Commenter IV-D-31 correctly states  that the  NAAQS for hydrocarbons
were revoked.  The revocation  notice (48 FR 628), however,  clearly
points out that "—  nonetheless, hydrocarbons should continue  to be
controlled or restricted  because of their contribution to  the formation
of ozone and the resultant health and welfare effects of this pollutant
and other photochemical oxidant products."  In addition, the  notice
states that specific  hydrocarbons (including VOC) which are shown to
cause adverse effects can be regulated separately.  The notice  does  not
restrict EPA or State authority in regulating emissions of hydrocarbon
as a class, particularly hydrocarbon compounds or any other VOC that
may be found to pose  a threat  to public  health or welfare.

     Comment:  One commenter (IV-D-3) questioned  if the judgment  of  the
Administrator was a valid reason for implementing standards and asked
what facts back up this judgment.

     Response: The judgment of the Administrator  is a valid reason  for
implementing standards.  The Clean Air Act mandates that the
Administrator list a  source category "... if in his judgment it [the
source category] causes,  or contributes significantly to, air pollution
which may reasonably be anticipated to endanger public health or  welfare"
[Section lll(b)(1)(A)].  The facts considered by  the Administrator  in
making this judgment  include,  but are not limited to, the quantity  of
pollutant(s) emitted  and the health and welfare effects associated
with the pollutant.  The quantity of VOC emitted  by the natural  gas
                                  2-2

-------
 processing  industry and  the health and welfare effects associated with
 ozone  exposure are discussed  in the response to the previous comment.

 2.2  ADEQUACY OF  CURRENT CONTROLS AND REGULATIONS

     Comment:  A  number  of commenters questioned the need for the
 standards because, in their opinions, equipment leaks of VOC in this
 industry may be adequately controlled by other regulations or approaches.
 Several commenters (IV-D-12;  IV-D-20; IV-D-31) felt that the proposed
 standards are not necessary in areas where ozone ambient air standards
 and  Prevention of Significant Deterioration (PSD) requirements are
 being  met.  Commenters IV-D-12 and IV-D-20 maintained that, while
 large, urban, non-attainment areas might need to control VOC emissions
 strictly, most natural gas plants (particularly in the West) are located
 in remote,  rural  areas far from major population centers or ozone
 nonattainment areas.  Similarly, another commenter (IV-F-la) claimed
 that the effects  of the  proposed standards would be regional, citing a
 statement from the Oil and Gas Journal (July 18, 1983) that 90 percent
 of allgas  plants and industry capacity is located in five States.

     Commenter IV-D-24 maintained that a State and local approach to
 regulating  gas plants would better take into account the many differ-
 ences  among gas plants and result in cheaper and more effective pollution
 controls.   The commenter noted that Ventura County, California, has a
 fugitive emissions program covering gas plants and that the South Coast
 Air Quality Management District (Los Angeles, Orange, Riverside, and
 San Bernardino Counties) and Kern County are in the process of establishing
 similar type regulations.  The commenter was also concerned that the
 proposed standards could remove a potential source of hydrocarbon
 offsets in  counties currently or predicted to become attainment for
 ozone  (i.e., Santa Barbara and San Luis Obispo).   Commenter IV-D-31
 added  that  there  is ample EPA technical  guidance (e.g., CTG and BID
 documents for VOC equipment leaks from natural  gas processing plants)
 available to enable States and local  agencies to make reasoned determi-
 nations of  best available control  technology (BACT) under PSD regulations
 if located  in an attainment area and lowest achievable emission rate
 (LAER) if located in a nonattainment area.

     Response:  In setting new source performance standards under
 Section 111 of the Clean Air Act, location of the industry in attainment
 or nonattainment areas is not relevant.   Location of an industry in an
 attainment or nonattainment area is  relevant to achieving the National
 Ambient Air Quality Standards (NAAQS)  under Sections 109 and 110 of the
 Clean Air Act.  The NSPS complements  the ambient  air quality-based
 rules as a means of achieving and maintaining the NAAQS and PSD
 requirements, while on a  broader basis it prevents new sources from
making air pollution problems worse,  regardless of the existing  quality
 of ambient air.   Similarly, in most  cases State and local  regulations
complement the NSPS.   The intent of  Congress in establishing NSPS was
 to establish a uniform level  of stringency nationwide, thereby preventing
States with lenient air pollution requirements  from attracting new
 industry construction.  The standards  will  limit  VOC emissions from
newly constructed, modified,  and reconstructed  facilities  in natural
gas  processing plants  and will  result  in  emission  reductions well  into

                                  2-3

-------
the future.  Even though these reductions may not bear directly now on
attainment or nonattainment of NAAQS for ozone, they will  make  room for
future industrial growth while attainment and nonattainment areas  would
benefit from these standards.

     Comment:  One commenter (IV-D-3)  stated that, in his  opinion, the
proposed standards represent a prime example of government intervention
in an industry that is  already overtaxed and overburdened  with  rules
and regulations.   The industry has proven it can regulate  itself and
has done a commendable  job of self-regulation.

     Response:  The commenter's assertions regarding overtaxation,
overburdening, and self-regulation are unclear, and he provided no
supporting information  about his claims.  In the absence of the standards,
the natural gas processing industry will emit about 20,100 Mg/yr of VOC
emissions from equipment leaks, which  reflects  normal  existing  gas
plant operations  with self-regulation  plus State regulations (Table 1-1).
Uith the control  requirements that will  be implemented by  the standards,
gas plants can achieve  significant emission reductions at  reasonable
costs.  A total of 14,600 Mg/yr of VOC emissions are expected to be
reduced nationwide as a result of implementation of the control tech-
niques required by the  standards.  This  represents an emission  reduction
of 74 percent from existing levels, which already reflect  self-regulation
plus State regulations.

     Comment:  One comrnenter (IV-D-12) pointed  out that the industry
has several incentives  to control VOC  emissions.  The entire justifi-
cation of the processing plant, according to the commenter, is  to
extract liquid and heavy gaseous hydrocarbons for sale.  Given  the
current value of  these  hydrocarbons, the commenter maintained that the
industry has a large incentive to capture and sell as large a quantity
as reasonably possible.  The commenter also noted the incentive of
preventing safety hazards associated with leak  control.

     Another commenter  (IV-D-11) indicated that his company had spent
considerable time, effort, and money to  establish leak prevention  and
control procedures to comply with current regulations, and thus felt
that natural gas  plant  VOC leaks were  adequately controlled already.

     Similarly, another commenter (IV-D-22) noted that plant personnel
will detect most  significant leaks visually or  by sound during  routine
daily inspections, and  that it is standard operating procedure  to  repair
leaking equipment as soon as possible.  The commenter added that
individual component surveys using a hydrocarbon detector or any other
type of instrument would be extremely time consuming and in many cases
impractical.

     Response:   EPA concurs with the first commenter on the industry
incentives to control emissions.  These incentives form the basis  for
the recovery credits used by EPA in calculating the net costs  for
compliance with  the standards.

     The commenters imply that the standards are unnecessary due to
the industry's own safety and economic  incentives to eliminate  leaks

                                  2-4

-------
 or due  to  industry  practices  that  have  been  implemented  to  comply  with
 current regulations.   However,  the  data  base  used  in  development of  the
 standards  shows  that  significant emissions do occur  from  equipment
 leaks at natural  gas  processing plants.   As  given  in  Table  7-2  of  the
 BID for proposed  standards, adjusted  for the  revised  compressor seal
 emission factor,  the  baseline emissions  for a typical  new gas  plant
 (Model  Plant  B)  would be  274  kg VOC/day  (100  Mg/yr).   The data  were  based on
 field testing  of  currently operating  gas plants and  reflect  the emission
 levels  remaining  after industry leak  prevention measures  have  been
 taken.   The leak  detection and  repair programs and equipment requirements
 of the  standards  are  aimed at reducing as many of  the  leaks  as  possible
 in a cost-effective manner and  are  expected to reduce  VOC emissions  by
 about 74 percent.

     The commenter's  questions  on the practical aspects of  leak monitoring
 are addressed  in  Chapters 4,  7, and 10.

 2.3 INDUSTRY  GROWTH,  INDUSTRY  MOBILITY,  AND  SEASONAL  CONSIDERATIONS

     Comment:  Two commenters (IV-D-3;  IV-D-29) questioned the  need  to
 implement  VOC  regulations based on  projections of growth  in  the industry.
 One commenter  (IV-D-3)  reasoned that  if  natural gas  production  is
 expected to decline as  stated in Chapter 9 of the BID  for the  proposed
 standards, air quality  would  improve without  regulations.

     The other commenter  (IV-D-29)  indicated  that although a few additional
 plants  may be  built,  liquid gas production is on the decline.   In  the
 commenter's opinion growth projections should be based on liquids  pro-
 duction, which are the  source of VOC leaks, and not the number  of  plants.

     Response:  The commenters1  opinions are  that the  standards are  not
 necessary because emissions will decrease without standards due to an
 expected decline  in natural  gas production and that growth projections
 should  be based on production instead of number of plants.  The decrease
 in  natural  gas production is not relevant to  the need  for new source
 performance standards.  The necessity of the  standards is based on
 projected growth within the industry and the  potential for long-term
 improvement in air quality.   Even though older gas plants may be shut-
 down, new gas  plants will  be built, as the second commenter correctly
 pointed  out.  Contrary  to the commenter's opinion, the projected
 number  of new plants  is the appropriate measure for the impacts of the
 standards, as emissions reductions are achievable as  new  facilities
 become  regulated.  The  general  goal  of NSPS is to require new facilities
 to  incorporate best demonstrated technology as they are being constructed.
 EPA has  estimated that  180 newly constructed and 40 modified and
 reconstructed gas plants will  be affected by the standards through
 1987.   These plants can achieve  significant emission  reductions at
 reasonable costs.

     Comment:   Several commenters  (IV-D-26, IV-D-29;  IV-D-31; IV-D-35)
contended that national standards  for natural  gas processing are not
mandated because, among other reasons, natural gas plants are not  geo-
graphically mobile and must  locate  near  the source of gas.  As  such,


                                  2-5

-------
States cannot compete for gas processing plants based on relaxed air
quality standards.

     Response: The considerations upon which the standards are based,
as required by Section 111, include the quantity of air pollutant emissions
from the source category, the extent to which the pollutant may reasonably
be anticipated to endanger public health or welfare, and the availability
of a demonstrated system of continuous emission reduction considering
costs, nonair quality health and environmental  impacts, and energy
requirements.  Therefore, mobility and competitive nature of the industry
are not the only criteria that were considered.  The first two factors,
quantity of emissions and public health and welfare impacts, support
the listing of this source category as discussed in Section 2.1.

     Comment:  One commenter (IV-D-31) stated that ozone formation is
accepted by the scientific community as being a seasonal phenomenon,
and that a review of the information pertaining to the formation of
ozone and the demonstrated lack of adverse public health or welfare
effects due to VOC emissions per se would lead  to the conclusion that
VOC emissions cannot be considered to endanger  the public health or
welfare during the months of October through April.

     Since one of the underlying premises of the NSPS program as
embodied in Section 111 of the Clean Air Act is that the source category
being regulated emits a* pollutant which can reasonably be anticipated
to endanger the public health or welfare, the commenter recommended
that, at a minimum, the scope of the NSPS be limited to the months of
May through September, during which time there  would at least be some
potential for environmental benefit.

     Response:  Sections lll(a)(l) and 302(1) of the Clean Air Act
require that NSPS reflect systems of continuous emission reduction.
As explained elsewhere, these standards reflect such systems, which are
operable throughout the year.
                                  2-6

-------
                      3.0   BASIS  FOR  STANDARDS

 3.1   INTRODUCTION

      Section  111 of  the Clean  Air  Act,  as  amended,  requires  that
 standards  of  performance  be  based  on  the best  system  of  continuous
 emission reduction that has  been adequately  demonstrated, considering
 costs,  nonair quality health and environmental  impacts,  and  energy
 requirements.   The following sections  present  the  basis  for  the promul-
 gated standards  for  each  type  of equipment and  any  changes in  the basis
 since proposal.  The selection of  the  basis  for the promulgated standards
 was  based  on  an  analysis  of  the  VOC  emission reductions, control costs,
 the  cost effectiveness, and  the  incremental  cost effectiveness  for  the
 control  techniques considered.   These  impacts  are  summarized in Table 3-1.

 3.2   VALVES

      In  selecting the basis  of the standards for valves, EPA considered
 quarterly  and  monthly monitoring periods.  Each of  these intervals  was
 compared in terms of the  emission  reduction  achievable and cost
 effectiveness  of the leak detection  and repair  program as presented  in
 Appendix H of the BID for the  proposed  standards.   At proposal , monthly
 monitoring was selected as the basis  of the  standards for valves
 because  it achieves  the largest  emission reduction, 40.4 Mg  per year
 for  Model  Plant  B.   The EPA  also judged that monthly  monitoring has a
 reasonable cost  effectiveness, $7  per  Mg.  However, EPA  also recognized
•at  proposal that some valves have  lower leak occurrence  rates  than
 others.   Monthly monitoring  of valves  that do  not  leak for 2 consecutive
 months  was judged to be unreasonable  when compared  to the additional
 emission reduction achieved  by monthly monitoring  over quarterly
 monitoring.   Therefore, although EPA proposed  that  leak  detection and
 repair  programs  include monthly monitoring for  valves, the proposed
 standards  allowed quarterly  monitoring  for valves  that have  been found
 not  to  leak for  2 successive months.   The annual emission reduction
 achieved by this monthly/quarterly implementation would  be 37.7 Mg  for
 Model  Plant B.   The  cost  effectiveness  and incremental cost  effectiveness
 would be a credit of $100/Mg and a cost of $240/Mg, respectively.   The
 promulgated standards are based on monthly/quarterly  implementation,
 which was  allowed at proposal.

 3.3   PUMPS

      In  selecting the basis  for  pump  requirements,  EPA considered
 quarterly  leak detection  and repair, monthly leak detection and repair,
 and  the  installation of dual mechanical seals.   As  with  valves, the
 three alternative control  levels were  examined  to determine the
 achievable emission  reduction, the net cost of  the  control technique,
 and  the  resulting cost effectiveness.  Also examined was the incremental
 cost effectiveness resulting from  dividing the  increased cost of the
 next more  stringent  control technique  by the resulting emission reduction
 i ncrease.
                                  3-1

-------
     Although dual mechanical  seals provided the greatest emission
reduction (2.6 Mg/yr for Model  Plant B)  for pumps,  the cost effectiveness
($4,900/Mg) and incremental  cost effectiveness ($31,000/Mg) were considered
unreasonable.  Monthly monitoring, however, resulted in emission reductions
of 2.3 Mg/yr, at a cost effectiveness of $610/Mg, which EPA considered
reasonable.  The incremental  cost effectiveness  of monthly monitoring
over quarterly monitoring, $800/Mg, was  also considered reasonable.   As
such, monthly monitoring was  selected as the basis  for the proposed
standards, and remains the basis for the promulgated standards.

3.4  COMPRESSORS

     The requirements for compressors have been  changed since proposal
to exempt all reciprocating  compressors  in wet gas  (field gas)  service.
In preparing the basis for the  proposed  standards,  EPA averaged  the
cost effectiveness for reciprocating and centrifugal  compressors (based
on one half of all compressors  being centrifugal),  and wet gas  and
natural  gas liquids (NGL) service (based on 34 percent of all  compressors
being in NGL service and 66  percent of all compressors being in  wet  gas
service).  Since the resulting  cost effectiveness ($460/Mg) was  considered
reasonable, EPA required the  installation of closed vent systems on  all
compressors except wet gas reciprocating compressors in plants  without
a control device.

     Table 3-1 shows the emission reductions and cost effectiveness  for
a closed vent system for wet  gas and NGL reciprocating and centrifugal
compressors.  The EPA considers the average cost effectiveness  ($1950/Mg)
of control systems on all wet gas reciprocating  compressors at  plants
with and without a control device to be  unreasonable and, therefore,
has exempted these compressors  from the  promulgated standards.
Reciprocating compressors in  NGL service and centrifugal compressors in
wet gas or NGL service can be equipped with closed  vent and control
systems at a reasonable cost  effectiveness ($250/Mg, $500/Mg, and
$64/Mg, respectively) and remain subject to the  requirements of  the
promulgated standards.

3.5  OPEN-ENDED LINES

     The basis for the promulgated standards for open-ended lines
remains the same as the basis at proposal.  The  EPA estimated that
capping or adding a second valve to all  open-ended lines will reduce
emissions from a typical (Model Plant B) plant by 19 Mg/yr, while
saving the industry $103/Mg.

3.6  PRESSURE RELIEF DEVICES

     The basis for the standards for pressure relief devices is  a
routine leak detection and repair program with quarterly monitoring.
In selecting standards for pressure relief devices, EPA considered
three regulatory alternatives,  including the selected quarterly leak
detection and repair program, a monthly  leak detection and repair
program, and the installation of rupture disks upstream of relief
devices to prevent emissions.
                                  3-2

-------
              TABLE  3-1.    CONTROL COSTS PER MEGAGRAM  OF  VOC  REDUCED0
Fugitive Emission Source ••
Pressure relief devices




Compressors'1
reclproca ting-wet gas
rec1procating-NGL
centrifugal -wet gas
centrifugal -NGL
Open-ended valves and lines
Sampling connections
Valves




Pumps



Control technique1"
Quarterly leak detection and
repair'
Monthly leak detection and
repair
Rupture disks

Closed vent and seal system
Closed vent and seal systemf
Closed vent and seal system^
tiosed vent and seal systemr
Caps on open endsf
Closed purge sampling1
Quarterly leak detection and
repair
Monthly/Quarterly leakf
detection and repair
Monthly leak detection
and repair
Quarterly leak detection
and repair
Monthly leak detect1onf
and repair
Dual mechanical seals
Emission
Reduct1onc,
Mg/yr

0.95

1.0
1.5

4.2
33
4.2
33
19
0.22

7.3
7.7

0.4

2.0
2.3
2.6
Average'',
$/Mg

(610)9

0
6,800

1.950J
250J
500J
64J
(103)
7,000

(104)
(100)

7

590
610
4,900
Incremental e,
S/Mg

(610)

5 800
22,000

1,950
250
500
64
(103)
7,000

(104)
240

1,450

590
800
31,000
 aThe  control costs per VOC emission  reduction are considered typical  of control techniques for
  equipment leaks 1n gas processing plants and are used 1n selecting  the level of control required
  by the standards.                                                                       H


 further discussion of control  techniques used can be found In Chapter 4.


 ^Emission reductions are for Model Plant B:  12 pressure relief valves. 6 compressor seals
  150  open-ended valves or lines,  6 liquid service and 6 gas service  sampling connections  '
  750  valves, and 6 pump seals.


 <*Average dollars per megagram -  (net annual  cost of the control  technique - emission reduction
  of the control technique).


 elncremental  dollars per megagram = (net annual  cost of the control  technique - net annual  cost
  of the next  ess restrictive control technique) (annual  emission reduction of control  technique  -
  annual emission reduction of the next less restrictive control  technique).

 fControl  techniques selected  as basis for the final  standards  are underlined.

 9Parentheses  denote cost savings.


 "Emission reduction and costs for compressors are from  the  BID  for the proposed standards
 Appendix H,  Table  3.   Costs  based on half of all  compressors  requiring a dedicated control  device.

 iCosts and emission reduction for closed purge sampling  represent both Inlet gas
 sampling and product  liquids sampling.                                      S


JThese mjmbers  are  based  on an average of the cost-effectiveness numbers  for compressors
 with  and without a control device (BID  for  proposed  standard. Appendix G,  Table 1).
                                                     3-3

-------
     Although monthly leak detection and repair provided  greater  emis-
sions reduction (1.0 Mg/yr) than quarterly leak detection and  repair
(0.95 Mg/yr), and the average cost effectiveness was  reasonable  ($0/Mg),
the incremental cost effectiveness was  not reasonable.   The  incremental
cost effectiveness of $5,800/Mg between monthly monitoring and quarterly
monitoring indicates that, since 0.95 Mg/yr VOC emissions can  be  reduced
with quarterly monitoring at an industry savings of $610/Mg, the
remaining 0.05 Mg/yr attributable to monthly monitoring  requires  an
industry expenditure of $5,800/Mg.

     The EPA also considered the installation of upstream rupture disks
as a regulatory alternative.  Rupture disks would result  in  essentially
100 percent emissions reduction, or 1.5 Mg/yr for a typical  plant.
However, the average ($6,800/Mg) and incremental ($22,000/Mg)  cost
effectiveness values were not considered reasonable.

     Based on all of these data, EPA selected quarterly  leak detection,
and repair as the basis for the promulgated standards for pressure
relief devices.

3.7  SAMPLING CONNECTIONS

     The EPA considered a single regulatory alternative,  closed  purge
sampling, to control VOC emissions during purging of  sampling  systems.
However, examination of the costs and emissions reductions for closed
purge sampling indicated that the cost effectiveness  ($7,000/Mg)  was
unreasonable.  Therefore, there are no requirements in  the standards
for sampling connections.
                                  3-4

-------
                4.0   FORMAT  AND  REQUIREMENTS  OF  STANDARDS

 4.1   FORMAT  OF  STANDARDS

      Comment:   One  coinmenter  (IV-D-12) stated that  the approach  taken
 by F.PA  in  setting standards  for  various pieces  of equipment  to control
 VOC  emissions is unnecessary, noting that  in other  industries EPA has
 allowed a  bubble concept  for  the aggregation of total emissions  from
 large facilities.

      The cornmenter  indicated  that  if EPA can demonstrate that total VOC
 emissions  from  a specific plant  need to be controlled due to an  ambient
 ozone violation in  the area,  the Agency should  provide a total VOC
 emissions  limit for  the plant and  let the  plant determine the most
 appropriate means of meeting  the emissions limit.

      The coinmenter  specifically  suggested  that  EPA  revise the proposed
 standards  to set a  total  emission  standard for  a plant of approximately
 150  pounds per  day  of VOC from valves and  process equipment.  In using
 this  limit, the commenter stated that a plant operator could choose
 technology and  operating  procedures tailored for the specific plant
 design  and could decide on  necessary recordkeeping  and maintenance
 scheduling as appropriate to meet  the limit.  The commenter  believed
 that  several other  industries are  regulated  in  this way under 40 CFR 60.

      Response:  A combination of  equipment, work practice, design, and
 operational standards was selected as the  format of the standards for
 process units and compressors (i.e., the affected facilities).   Different
 formats are required for different types of leaking equipment because
 characteristics of  the equipment,  the available emission control
 techniques, and the applicability  of the measurement method used for
 equipment leaks differ.

     Setting a  single emission limit for a given plant would be
 impracticable because the cost associated with measuring emissions from
 each potentially leaking piece of  equipment would be unreasonable for
 most plants.  In addition, the emissions limit (if one were set  for the
 entire  plant) would likely be nearly zero because most pieces of equip-
 ment do not leak.    (The emissions  from a few pieces of equipment that
 may leak averaged with emissions from the majority of equipment that may
 not leak would be close to zero.)

     The EPA is developing a policy for approving NSPS bubbles on a
 case-by-case basis.   This policy will  be proposed in the Federal  Register
 in the near future.   Applications  for NSPS bubbles will  be considired
 in accordance with this policy.

 4.2  VALVES

 4.2.1  Difficult-to-Monitor  Valves
     CommentlOne commenter (IV-D-30)  stated that difficult-to-monitor
 valves can  not be  eliminated from new facilities, and  recommended that
difficult-to-monitor valves  in new  facilities be subject to  the same
annual inspection  requirements as those in modified  or reconstructed

                                  4-1

-------
facilities.  The comrnenter noted that block valves under relief valves,
open-ended valves for testing of relief devices, and certain open-ended
valves for purging equipment at start-up must be placed in high locations.
Another commenter (IV-D-37) suggested that up to 10 percent of valves
in general should be considered difficult to monitor and should be
exempted from monitoring requirements except annual  visual  inspections.

     Response:  Upon reviewing the comments on the proposed standards,
EPA accepts that eliminating all difficult-to-monitor valves from new
facilities may substantially increase the costs of constructing new
facilities, for example, due to the necessity for additional fixed
ladders and platforms.  Estimation of these additional  costs is not
possible due to the wide variability of factors such as the height of
the valves and the ability to co-locate di ff icul t-to-rnonitor valves.

     Commenter IV-D-30 did not provide any data to indicate how many
valves in new gas processing plants would be difficult  to monitor.
Although commenter IV-D-37 suggested that 10 percent of valves in
general should be considered difficult to monitor, no basis for this
number was given.  The EPA expects the proportion of diff icul t-to-rnoni tor
valves in gas plants to be similar to that of refineries.  A refinery
maintenance study (Docket Item II-A-11) found that about 3 percent of
over 8,000 total  valves investigated could not be reached without
extraordinary aids such as scaffolding or cherry pickers.  Based o-n
this study and on telephone conversations with refinery process design
engineers (Docket Item IV-B-8), the Agency believes the 3 percent
figure is accurate representation for gas plants.  Therefore,  the
promulgated standards allow the owner or operator of a  newly constructed
facility to designate up to 3 percent of its valves as  difficult to
monitor.  The standards require annual monitoring of these valves.
Limiting the percentage of allowable valves that may be difficult to
monitor provides the incentive to minimize the number of such valves in
new units, while ensuring that an owner .or operator would not incur
unreasonable costs by attempting to eliminate all difficult-to-monitor
valves in new units.

     Comment:  One commenter (II-D-10) noted difficulties in performing
monitoring of valves located atop gas plant towers since they are not
equipped with platforms.  In addition, the commenter remarked that not
all valves are accessible, and many valves cannot be reached or monitored
in 1 minute.  The commenter also noted that the economic analysis in
the draft NAPCTAC BID did not consider relocation of components to
eliminate "difficult-to-monitor" valves.

     Response:  Provisions for difficult-to-monitor valves have been
added to the standards since this comment was written in May 1981.
Less frequent monitoring (annual) for difficult-to-monitor valves was
allowed for existing plants in  the proposed standards,  while annual
monitoring for all di ff icul t-to-inoni tor valves  is being allowed
in the final standards, since monitoring of these valves on a monthly
basis is not cost effective.  The EPA used 1 minute per valve as an
average monitoring time based on actual process unit testing.  Since
the 1 minute value represents an average monitoring time for all valves,
some valves may take longer, while others will  take less time.  The

                                  4-2

-------
cost of monitoring valves is based on the 1-minute average; therefore,
1 minute is considered to be a reasonable monitoring time for all  valves.

4.2.2 A1ternative Standards

     Comment:  One commenter (IV-D-20) commended EPA for developing
alternative standards for valves in VOC service.  However, the cornmenter
stated that 2.0 percent allowable leakers may not be an attainable
standard.  The commenter stated that his company operates several  gas
plants with less than 20 valves in VOC service, and that a leak in only
one valve would therefore result in a violation.  The commenter recommended
that the standard be changed to 10 percent allowable leakers or at
least one valve leaking over the inspection interval  at each facility.

     The commenter recommended that the second alternative standard for
valves (skip-period monitoring) be retained except for changing the
allowed 2 percent leakers to 10 percent.

     Response: Alternative standards for valves have been provided in
the standards because EPA judged that the emission reduction and the
annual cost relationship is unreasonably high for process units that
have fewer than 1.0 percent of valves leaking over an extended period.
Due to the variability in valve leak detection, process units that
average less than 1.0 percent of valves leaking will  have, at times,
more than 1.0 percent of valves leaking.  Therefore, to provide for the
variability in leak detection, EPA set the allowable percentage of valves
leaking for any point in time at 2.0 percent.  (A complete description
of the methodology used to determine the allowable percentage of valves
leaking is presented in Docket Item II-B-43.)

     The allowable percentage of valves leaking applies to alternative
standards for valves, hence, owners or operators are not required  to meet
a percentage of valves leaking unless they elect to do so.  Also,  under
the skip monitoring alternative standards, if greater than 2.0 percent
valves are detected leaking, a violation has not necessarily occurred.
Owners or operators would simply be required to return to monthly/quarterly
leak detection and repair.  If, as the commenter claims, a gas plant has
less than 20 valves in VOC service, the owner or operator may elect to
follow an alternative standard for valves.  However,  one valve leaking
would exceed the allowable percentage (2.0)  of valves leaking.  Since
20 valves would require very little time to  monitor,  alternative
standards are really unnecessary.   The EPA does not expect many gas plants
processing more than 10 MMscfd with 20 valves or less.

     Comment:  One cornmenter (II-B-23) requested that a "reference leak
detection program" be defined in Section 60.633-2 of  the regulation.

     Response:  A "reference leak  detection  program"  is simply the
requirements for valves in Section 60.632-6, namely,  a  monthly/quarterly
leak detection and repair program.   The required standards serve as the
basis for the skip period leak detection and repair alternative standards.
After an owner or operator has demonstrated  that he can maintain a
performance level  of 2 percent or  less for either 2 or  5 consecutive
quarters using the required  standards as the "reference leak detection

                                  4-3

-------
program, " then the owner or operator Cdn reduce the monitoring f re
-------
     Response:  Visual  inspection of pumps is necessary to check for
liquids dripping from the seal.   As the seal  wears over time,  VOC
leakage is likely to increase, so it is important to monitor pumps on a
regular basis.  As stated in Table 8-3 of the BID for the proposed
standards, weekly visual  inspections would take about one-half minute
per pump or about 10 minutes per week (9 hours per year) for one person
at a large plant with about 20 pumps.  The annual  cost of these visual
inspections would be about $160 (June 1980 dollars). The cost of weekly
visual  inspection is reasonable.  The EPA has concluded that this
situation is not likely to occur.

4.4  COMPRESSORS

     Comment:  Several  commenters (II-B-23; II-D-10; II-D-30; IV-D-13;
IV-D-15; IV-D-19; IV-D-21; IV-D-23; IV-D-24;  IV-D-26; IV-D-27; IV-D-29;
IV-D-30; IV-D-31; IV-D-35; IV-D-33; IV-D-34;  IV-D-36; IV-F-la; IV-F-lc;
IV-F-le) expressed concerns about potential safety hazards associated
with the installation of closed vent systems  on reciprocating compressors.

     The commenters stated that compressors must be equipped with
foolproof systems to prevent pressurizing the distance piece and must
absolutely prevent the intrusion of air and backpressure from the flare
header.  Additionally, flow of the compressed gas into the compressor
driver crankcase must be prevented.

     Other commenters stated that, where existing control devices do
exist, they are often used to control sour gas.  Since the gas-contacting
surfaces of the compressors are not metallurgically compatible with
sour gas, tie-in to the existing control device would not be possible
due to potential corrosion problems.

     Similarly, other commenters (IV-D-33; IV-D-36) said their experience
shows the only times they would consider it a safe practice of tying a
distance piece into a closed vent system to a control device is when a
low pressure flare system is available.

     The commenter further noted that some gas plants use flare systems
solely for high pressure process vessel relief, such as straddle type
plants, plants processing naturally occurring high pressure field gas,
or plants where compression is done in the field prior to plant
processing.  The commenter stated that their  experience shows that d
check valve is not sufficient to prevent backpressure on the flare
header from pressurizing the distance piece.   The commenter recommended,
at the very least, that the requirement for tying the distance piece
into a closed vent system be evaluated on a case-by-case basis.  In
addition, the commenter had no objection from a safety viewpoint to
tying the packing vent into a closed vent system and maintained that
the majority of compressor VOC emissions would be controlled with
just the packing vents tied to a closed vent  system.  The commenter
cited a letter (Docket No. II-D-44) previously submitted that supports
their safety concern for pressurizing single  distance pieces.

     Another commenter (II-D-10) stated that  tying seal  emissions back
into the compressor inlet was not viable because of pressure considerations,

                                  4-5

-------
     All of these commenters recommended that EPA exempt reciprocauny
compressors from the standards.

     Response:  The commenters1  primary concerns are pressurization of
the compressor distance piece and flow of the compressed product into
the driver crankcase.   The EPA cost analysis was based on equipping the
compressors with two-compartment distance pieces, which are designed  to
collect hazardous or toxic gases.  The two-compartment distance pieces
allow for the compressor side compartment to be sealed and vented to  a
control device, while the crankcase side compartment may remain open  to
the atmosphere.  By allowing the crankcase side to remain open, the
"break" between the compressor and the driver crankcase is maintained,
preventing the flow of compressed gases into the crankcase.  Double
distance pieces have been recognized by the industry as a safe method
of seal gas collection for at least 10 years, as evidenced by the 1974
API Reciprocating Compressor Standard.1

     Given that the flow of gas  into the compressor crankcase may
be prevented using double distance pieces, the remaining concern of the
commenters was that the distance pieces might become pressurized from
the flare header or develop explosive atmospheres due to air intrusion.

     In order to ensure that the distance pieces did not become over-
pressurized, EPA recommended several safety features, as follows:

     •   Distance pieces should  be routed to the flare through a separate
         flare line, and not the plant's main flare header.  The EPA
         included the cost of-100 meters (300 feet) of 2-inch pipe
         for this purpose.

     •   Each distance piece would be equipped with a check valve to
         prevent backflow, and the flare line would be equipped with a
         low-burst pressure rupture disk to relieve any catastrophic
         failures.

     •   The system should be pressurized.  A water-sealed trap was
         included in the flare line to prevent air intrusion.

Although no system can be made entirely failsafe, EPA has estimated the
costs for a control system for compressors which, although simple in
design, was designed with safety as the primary consideration.

     Several commenters indicated that, where existing flares were in
sour gas service or high-pressure service, a dedicated control device
could be required for the compressor vent control system.  The EPA
recognized at  proposal that, in many cases, yas plants would not have a
control device present, and exempted wet gas reciprocating compressors
in plants where a control device was not already present.  However,
based on the comments received since proposal, and the unreasonable
cost effectiveness of wet gas reciprocating compressor controls used at
more plants than EPA estimated at proposal, EPA has decided to exempt
1
 API Standard 618, 2nd Edition, July 1974.  Docket Item No. II-I-35.

                                  4-6

-------
 all  wet gas reciprocating compressors from the promulgated standards
 (see Section 7.2).   However,  since  installation of  a  control  device  is
 cost effective for  wet gas centrifugal  compressors  and all compressors
 in natural  gas liquids service,  control  of these compressors  is  still
 required by the promulgated standards (Docket Item  IV-B-7).

      Numerous  control  techniques  are  available for  compressor seal
 emissions.   Some plants may not  be  able  to pipe emissions into a
 compressor  inlet, as noted by the commenter.   The EPA also recognizes
 that some plants do not have  suitable flares  available,  as use of a
 high pressure  flare would require an  auxiliary compressor.  Hence, EPA
 included the cost of a suitable  low pressure  flare, which would  not
 require an  auxiliary compressor,  in the  cost  analysis presented  in the
 BID  for the proposed standards  (Appendix G).

      Comment:   One  commenter  (IV-F-la)  stated that  compressors should
 be exempt from the  standards  because  the control  of emissions from them
 requires the greatest capital  investments while providing the lowest
 cost effectiveness.

      Similarly,  another commenter (IV-D-24) noted that Table  3-2 of  the
 BID  for the proposed standards shows  that compressors represent only
 2  percent of the EPA's total  estimated VOC  leakage.   He  concluded that
 the  increased  risk  and cost of controlling  compressor emissions is not
 justified.

      Response:   Compressor emissions, as shown  in the BID for the pro-
 posed standards,  Appendix H,  are  15.8 percent of  all  gas  plant emissions.
 When  examining  the  selection  of best  demonstrated technology  (BDT)
 considering costs for  the  standards,  EPA compared available demonstrated
 controls for each type of source.   For compressor seals,  closed vent
 and  seal  systems were  considered  as the  only  available demonstrated
 technology.  Since  the use of closed  vent and seal systems can be
 accomplished at  reasonable  cost and cost effectiveness,  the controls
 are  required.  The  EPA does,  however, consider  the costs  of controlling
 reciprocating  compressors  in wet  gas  service  to be unreasonable and  has
 exempted those compressors  from the promulgated  standards.

 4.5   OPEN-ENDED  LINES

      Comment:  Two  commenters (IV-D-15 and  IV-D-24)  opposed the require-
 ments  in  Section 60.632-5  to plug open-ended  valves or lines  because they
 present  a potential   safety  hazard.  The  open  end  is frequently some
 distance  downstream  from  the valve,  and  there could be significant
 safety  hazards if the  operator fails  to  remove  the covering prior to
 opening  the  drain valve or  purging valve.  One  commenter  (IV-D-15)
 said  there  was a need  to  include an  alternative standard  for  frequently
 used  valves, such as to sample the outlet quarterly.  One commenter
 (IV-D-24) questioned the accessibility of the open ends noting that a
 vent  line may be on  a  riser pipe well  above the work area or a drain
 line may  dump into a sump  tank.   He  further questioned the benefit of
 the open-ended line  requirements  because many lines  must be routinely
opened to vent,  purge, sample or  drain some system,  and any fugitive
emissions would be released at that  time.

                                  4-7

-------
     In contrast, one commenter (IV-D-20) agreed with the use of caps,
plugs, and second valves to control leakage from open-ended lines.  The
commenter stated that the practice was already being done in his
company's gas plants.

     Response:  The standards for open-ended valves or lines provide
operators the flexibility to add either a cap, blind flange, plug, or a
second valve, depending upon the individual application.  The EPA
acknowledges that plugging open-ended valves or lines may present
a potential safety hazard in certain situations.  Accordingly, if a
second valve is used, the standards require that the upstream valve be
closed before closing the downstream valve.  This operational requirement
is merely sound practice that plant operators currently follow to
prevent process fluid from being trapped between the valves.  If hot
(or cold) product is trapped between the two valves, as it contracts
(expands) from cooling (heating) to ambient temperature, it could cause
the pipe, the valve stem, or the valve seat to fail.  Should the inner
valve leak through the valve seat, however, the product will eventually
fill the piping between the valves with ambient temperature fluid
without stressing the valve seat.

     The standards provide sufficient flexibility for the operator's
concern with plugging frequently used valves.  Adding a second
valve avoids the risk of premature failure of pipe connections caused
by the frequent removal of a cap or plug.

     Leak detection and repair for the control of VOC emissions from
open-ended valves or lines is inappropriate because it would achieve
less emission reduction and may cost more to implement than the equipment
and operational standards for open-ended valves because of repeated
inspections of nonleaking sources.

     Comment:  One commenter (II-D-10) questioned EPA's recommendation
to use teflon tape to seal threaded connections.  The commenter noted
that teflon tape works well for small sizes (2-inch nominal diameter
and smaller); however, the sealing ability on larger sizes has not
proved satisfactory.

     Response:  Joints larger than 2 inches in diameter are usually
either flanged or welded, and not threaded.  The EPA's point discussed
in the BID for the proposed standards is reiterated by the commenter,
namely, that leaks from small threaded connections can be reduced by
using teflon tape on the threads before the connection is made.  The EPA
makes no claim that the tape is effective on larger connections.

     Most open-ended lines subject to the standards would be 2 inches
or smaller in diameter.  These lines could be sealed with teflon tape
on the appropriate cap or plug.  Although the standards would require
the closure of larger open-ended lines, which would require more costly
control, the cost of plugging open-ended lines is based on the addition
of a second valve instead of a cap or plug.  In most cases, small
open-ended lines can be capped or plugged at significantly lower cost
than with the second valve assumed by EPA.  Large blind flanges cost
about the same as small valves and, therefore, are reasonable.

                                  4-8

-------
 4.6  PRESSURE RELIEF DEVICES

 4.6.1  Annual Certification

      Comment:  One  commenter  (IV-D-15)  offered  an  alternative  method  of
 achieving at least  equivalent emission  control  of  pressure  relief  devices
 because he thought  the  testing,  reporting,  and  recordkeeping of  the
 hundreds of relief  valves  that would  typically  be  installed in a new
 plant would be an immense  burden.   Also,  the additional  investment to
 tie all  relief valves into a  closed system  to gain exemption from
 monitoring would be large.  The  commenter preferred  to  follow  certain
 design criteria and test practices  to verify relief  set pressures.
 The commenter proposed  that an operator annually certify that  these
 conditions are met.  The commenter  also stated  that  relief  valves  should
 be  exempt when burst-heads and test inserts are installed.  The commenter
 further wrote that  if such installations  are made  for operator design
 considerations, then control  effectiveness  should  be acknowledged, and
 the relief valve should  be exempt from  monitoring, compliance  testing,
 and reporting.

      Response:  The EPA  considered  alternative  control  techniques,
 including rupture disk  (burst-heat) installations, leak  detection  and
 repair,  and no control  (baseline) for pressure  relief devices  prior to
 proposing the standards.   The  routine leak  detection and repair require-
 ments selected for  inclusion  in  the proposed standards  were estimated
 (BID for proposed standards,  page H-4)  to reduce VOC emissions from
 relief devices by 0.076  Mg/yr  per valve,  with a savings  to  industry of
 $46/device-year.  As such, the quarterly  leak detection  and repair
 program  required by the  standards will  result in a savings  to  industry
 while reducing VOC  emissions.

      The  commenter  offered an  alternative control technique for relief
 devices  based  on verification  of set pressure and certain relief device
 design  criteria, apparently based on the  assumption that relief device
 leaks are  the  result of  valve  openings  during operation.  However, most
 relief  device  leaks are  from  "closed" valves and are the  result of
 corrosion  and/or improper  reseating after a pressure release.   As  such,
 set-pressure  verification would not ensure that a pressure relief
 device would  not leak.   The standards,  therefore, remain  unchanged.

      The  commenter  also  requested a blanket exclusion for relief devices
 equipped with  rupture disks and test inserts.   The EPA  considers rupture
 disk  systems  to be  effective control devices for emissions from
 pressure  relief devices.  Relief devices equipped with rupture disks
 may  be designated for "no detectable emissions"  and require only an
 initial acceptance  test  and annual  compliance  tests to ensure  proper
 installation and operation.  Since  rupture disks may burst or  develop
 leaks through  corrosion, annual performance tests,  as well as  rapid
 (5-day) replacement after they burst,  are  necessary;  therefore, the
 standards remain unchanged.

4.6.2  Rupture Disks

     Comment:  One  commenter (IV-D-20) stated  that control of  VOC
emissions from pressure  relief devices by  installation of a rupture disk
upstream or downstream creates potential safety  problems.  Furthermore,

                                  4-9

-------
the commenter claimed that pressure relief devices  are already checked
and maintained in as leak-free condition  as possible  since  leaks  would
result in the loss of valuable hydrocarbon products.

     Response:  The standards  for pressure relief devices are  based  on
quarterly leak detection and repair; however,  the standards  allow
operators to use control techniques that  achieve  emission reductions
equivalent to quarterly leak detection and repair.   The EPA  has determined
that pressure relief devices that comply  with  the "no detectable  emissions"
limit and pressure relief devices that are controlled through  a closed
vent system to a control device would achieve  equivalent emission
reductions.

     Rupture disks eliminate equipment leaks of VOC through  the relief
device unless an overpressure  occurs.  After an overpressure release,
replacement of the rupture disk once again eliminates equipment leaks.
Pressure relief devices are required to operate with  no detectable
emissions as indicated by an instrument reading of less than 500  ppm
above background and are required to return to this condition  within
5 days following a pressure release.

     While it is possible that operators  would not use rupture disks
because of safety concerns, rupture disks are  allowed in the standards
and they are presently used in gas plants.  If a  rupture disk  is  used,
a pressure sensor could be installed to warn operators if a  pressure
increase has occurred between  the disk and relief valve.

4.6.3  Accessibility Rechecks

     Comment:  Several commenters (IV-D-19, IV-D-21,  IV-D-23,  IV-D-26,
IV-D-30, IV-D-36, and IV-F-4)  requested that EPA revise the  monitoring
requirements for pressure relief devices.  The commenters  stated  that
relief devices cannot be placed in accessible  locations for  safety
reasons.  Commenters (IV-D-23 and IV-D-30) contended  that elevating
relief valves (i.e., atop towers and columns)  provides for  better
atmospheric dispersion of any released hydrocarbons to minimize explosion
hazards.  Therefore, relief devices are accessible only by  adding
ladders and platforms or by crane, and these costs were not  included in
considering the economic costs of this rule.  One commenter  (IV-D-30)
also pointed out that elevated relief valves will have fewer emissions
than relief valves at lower locations in  equipment processing  both  gas
and liquid phases because a smaller mass  of gas would be released to
provide the same pressure drop.

     One commenter (IV-D-23) thought that there was no benefit to the
requirement that pressure relief devices  be monitored within 5 days  of
relieving because leaking pressure relief valves cannot be  repaired  until
the unit is shutdown.  Also, at an unmanned or partially manned plant,  a
relief valve could relieve without anyone's knowledge.  Other commenters
(IV-D-36 and  IV-D-26) suggested retesting relief valves after over-
pressures at the next scheduled monitoring of valves and pumps.  These
commenters pointed out that many facilities will  be utilizing contract
services to handle their leak detection and repair, and a special trip
                                 4-10

-------
 (costing  in  excess  of  $1,000) would  be required to recheck a  pressure
 relief device  after an  overpressure.  Another commenter  (IV-F-4) requested
 that  EPA  allow a  14-day period  for retesting of relief devices after
 overpressures.  One commenter (IV-D-21) added that inspection of pressure
 relief valves  following each pressure release is difficult to arrange
 since the economics of  small to mid-size plants will not allow either
 the necessary  instruments or the  instrument technicians  to be available
 at all  times.

      One  commenter  (IV-D-19) recommended monitoring pressure  relief
 devices 1 week after each pressure relief and disregard  quarterly
 monitoring because  relief devices will leak within the first week after
 relieving, or  not at all.  This commenter argued that quarterly monitoring
 would be  an  unnecessary duplication  of effort.  Another  commenter (IV-D-30)
 suggested that  relief  valves be subject to annual  leak detection and
 repair as are  difficult-to-monitor valves.

      Response:  As  presented in the  background information document for
 the proposed standards,  emissions from pressure relief devices in natural
 gas plants can  be reduced by 0.076 Mg/year per pressure  relief device
 by implementing quarterly leak detection and repair.  This emission
 reduction value, which  was based on  emission measurements at gas
 processing plants,  resulted in the product savings ($73/device-year)
 being  almost triple  the  costs ($27/device-year)  of monitoring.  The
 statement made  by commenter IV-D-30  is not relevant since the elevation
 of a  pressure  relief device for gas  phase service does not affect the
 amount  of gas  released.  An elevated relief device in liquid phase
 service would  affect the mass of emissions.   However, very few, if any,
 pressure  relief devices at gas plants would  be in  liquid service, and
 the standards  apply  only to gas service relief devices.   The EPA based
 the emission factor  for  pressure relief devices  on gas phase releases
 only'  because few relief devices are  used in  liquid phase service.

      Commenter  IV-D-19  is incorrect  in stating that relief devices  will
 leak  within 1 week or not at all.   While it  may  be true  that significant
 leakage due to improper seating or unseating of  the relief valve occurs
 relatively soon after a release, minor leaks may develop over longer
 periods due to corrosion or valve "chatter".   Since safety regulations
 already require most pressure relief devices to  be inspected annually,
 EPA requires only three additional inspections  per year.   Pressure
 relief devices could be monitored  at each  quarterly inspection using
 the same  equipment that is used  during the current annual inspections.
 Quarterly inspections could improve plant  safety through  both leak
 detection and detection of closed  block valves.

     One commenter contended  that  relief valves  cannot  be repaired
 without a  shutdown.   While this  is true  in some  cases, most  relief
 valves are equipped  with upstream  block  valves to  allow  testing  or
 replacement of the relief valve  while the  unit is  in  operation.   These
 relief valves may  be repaired  with the unit  in operation.

     Several  commenters expressed  concern  about  performing rechecks  on
pressure relief devices within  5 days of a pressure  release,  since


                                  4-11

-------
monitoring personnel and instruments may not be available.   The EPA
recognizes that many small plants will use contractors or central
office personnel to perform routine monitoring and that in  these cases
the monitoring team is not available except during scheduled test  periods.
In these situations, EPA believes that the cost of obtaining the
monitoring services for a single relief valve would be unreasonably
high.  Therefore, the final standards allow small  plants that are  not
monitored by on-site personnel  to delay rechecks up to 30 days after a
pressure release.

     Comment:  One commenter (II-D-10) questioned  the benefit of monitoring
safety valves/rupture disks considering the inherent danger in monitoring
them.  In support of his concern, the commenter added that  a person who
is monitoring when a high pressure release occurs  could be  injured or
killed, particularly if the person is holding onto a ladder near the
top of the vessel and is handling all the paraphernalia required for
monitoring.  The commenter noted that gas plants normally have relief
valves and/or rupture disks installed on top of vessels or  columns
without access platforms.

     Response:  The standards for gas service pressure relief devices
in natural  gas processing plants are work practices consisting of  a
quarterly leak detection and repair program.  The  pressure  relief
devices are also required to be monitored within 5 days after each
overpressure to determine if a  leak has occurred as a result of the
overpressure.  Monitoring of these devices should  be performed by
personnel  who understand the precautions and practices recommended by
industry and ASME codes.  If a  pressure relief device is likely to
relieve when monitoring occurs, then special precautions, such as
monitoring of process conditions (temperature and  pressures), should be
taken by the owner or operator.  Based on EPA's experience  in collect-
ing data for pressure relief devices, these devices can be  monitored
safely.  As mentioned earlier,  most plants already perform  annual  relief
valve tests, further indicating that monitoring can be performed safely.

     Comment:  One commenter (II-D-10) stated that EPA's requirement
for capping all open-ended lines suggests capping  relief valve discharge
lines that must remain open to  provide a path for  discharging the
fluids from the valve in case of an emergency.   Another commenter
(IV-D-34)  requested that a new  paragraph be added  to the standards for
open-ended valves or lines:

    (c)  Open-ended valves which serve as automatic or manually
         actuated emergency shutdown systems for gas processing plants
         are exempt from the provisions of the section and  shall comply
         with the provisions of §60.632-6.

The commenter noted that emergency shutdown (ESD)  systems are required
to protect people and equipment.  ESD systems dump gas to atmosphere by
automatic or manual actuation through a series of  open-ended valves.
The commenter maintained that as a safer alternative open-ended valves
or lines in ESD systems should  be subject to the valve standards rather
than to the open-ended valve or line provisions of Section  60.632-5 of


                               4-12

-------
 the proposed standards.   The commenter added  that leakage  in  a  pressured
 ESD system is self-monitoring,  since  a drop in  system  pressure  actuates
 the ESD system.

      Response:   As stated in the definition of  "open-ended  line"
 (Section 60.631  of the  proposed standards), open-ended  lines  do not
 include pressure relief  valves.   Relief valve discharge  lines are
 exempt from the  capping  requirement because relief valve horns must
 always remain open for the  relief valve to  function.   Assuming  ESD
 systems are relief valves,  they would  be  exempt  from the standards  for
 open-ended lines because they are excluded  from  the definition of
 open-ended lines.

 4.7  CONTROL DEVICES

 4.7.1  Efficiency  Requirements

      Comment:  One commenter (II-B-23)  stated that  vapor recovery
 compressors are  more  representative of gas  plant  applications than
 condensers or absorbers  and  should be  the appropriate standard for gas
 plants.

      Response:   Control  devices  are not limited  to  flares,  incinerators,
 condensers, or absorbers.  They  are given as examples of acceptable
 systems.   Control  devices must  have a  VOC control  efficiency of at
 least 95  percent.   If a  recovery compressor were  designed and operated
 with  greater than  95  percent  reduction  efficiency,  it would be an
 acceptable control  device for natural   gas processing plants.

      Comment:  Several commenters (II-B-23, IV-D-10, IV-D-13  IV-D-15
 IV-D-20,  IV-D-24,  IV-D-31, IV-D-33, IV-D-35, IV-F-lc, IV-F-le) took  '
 issue  with the velocity  and  heat  content limits  for flares, and requested
 they  be deleted  from the standards.  Many of these commenters maintained
 that  existing emergency  flares are capable of greater than 98 percent
 combustion  efficiency if properly operated.

      Some  of the commenters specifically questioned limiting heat content
 values  provided  by the standards.  One commenter  (IV-D-33)  stated that
 gases  with  heat  contents of about 5.6  MJ/scm (150 Btu/scf)  could be
 flared successfully, while others (IV-D-24 and IV-D-35) noted that
 produced (field)  gas often has low heat values due to  N2 and CO?
 entrainment, especially in the latter  stages of enhanced recovery.

     Several commenters indicated that studies available to  EPA  showed
 that  flares operating over a wide range of gas heat content  values  and
 flare exit velocities have destruction efficiencies in  excess  of
 98 percent.  One  of the commenters (IV-D-10) maintained that EPA based
 the flare exit velocity limit on test  data used  for the SOCMI  NSPS,
with the testing  limited  to  velocities of less than 18.3 rn/sec (60 ft/sec)
The commenter did not agree  with EPA's conclusion from  these data  that
higher velocities would result in lower efficiencies, stating  that  EPA
should require that flame stability be maintained rather than  limiting
the heating value and  velocity of the  gas  stream.  Many  commenters


                                  4-13

-------
(IV-D-21, IV-D-23, IV-D-26, IV-D-33,  IV-D-36,  IV-F-la)  were  concerned
that API Recommended Practice (RP)  521  was  ignored,  noting that  RP  521
recommends a flare tip velocity of  0.2  Mach  for continuous flaring  and
0.5 Mach for short-term flaring for sizing  the flare.   Several
commenters indicated that to design a flare  for maximum exit velocities
of 18.3 m/sec (60 ft/sec) under emergency conditions would result  in
very large flare tips, increasing the chance of air  migration and  stack
explosions during normal  operation.

     Four commenters (II-B-23,  IV-D-27, IV-D-29, IV-D-35)  favored  a
20 percent opacity level  (except for  5  minutes in  any 2 consecutive
hours) and contended that the "no visible emissions" requirement was
too stringent.

     Response:   As noted  above, EPA selected control devices that
achieve at least 95 percent reduction efficiency as  the "best demon-
strated technology" (BDT) for control devices  used in complying  with
the standards.   In reflecting this  selection of BDT, EPA proposed
specific design and operational requirements to ensure that  flares  used
to comply with  the standards achieve  at least  95 percent reduction
efficiency.  The proposed design and  operational requirements were, in
EPA's judgment, the only  set of limits  that  ensure at least  95 percent
reduction efficiency.  The EPA considered setting  requirements that
better reflect  95 percent reduction efficiency but rejected  this because
data do not support setting design  and  operational requirements  that
would distinguish between 95 and 98 percent  reduction efficiency.   The
EPA agrees with the commenters who  stated that existing flares in  this
industry are capable of greater than  98 percent reduction  efficiency  if
properly operated.  However, the only way EPA knows  to ensure this
judgment is to  set limits on design and operation  of the flares  that
reflect BDT.

     Since proposal, EPA further examined previously available and
new flare studies cited by the commenters (Palmer (1972);  Siegel (1980);
Lee and Whipple (1981); Howes (1981); McDaniel  (1983)), and  concluded
that flares with exit velocities up to 122  m/sec (400 ft/sec) can
achieve 98 percent destruction efficiency,  as suggested by one commenter,
if the gas heat content  is sufficiently high.    Therefore,  the final
standards provide an additional equation for determining velocities up
to 122 m/sec (400 ft/sec) depending on the  gas  heat content, as follows:


                     Logio  (Vmax) =  (HT + 28.8)/31.7

     where Vmax = maximum permitted velocity, m/sec
     and   Hj   = the net heating value of the  gas, MJ/scm


     This equation was based on  flare efficiency data available to EPA,
primarily from a study conducted by  Energy and  Environmental Research
Corporation, entitled "Evaluation of the Efficiency of  Industrial
Flares-Test Results," "  Pohl , et. al. (1984), Docket Item IV-A-1.   This
equation will allow  streams  with net heating  values greater  than 11.2
MJ/scm  (300 Btu/scf) to  be  flared at higher velocities, while ensuring

                                   4-14

-------
an effective reduction efficiency that reflects BDT.  These limits do
not allow flares to be used that burn streams with a net heating value
of 5.6 MJ/ scm  (150 Btu/scf), as suggested by one of the commenters.
In EPA's judgment, flares burning streams with this net heating value
can not be considered to be effective control devices.  The available
data do not support revising the lower net heating value limit (i.e.,
300 Btu/scf).

     An underlying concern of the commenters appears to result from a
misunderstanding that flares must never exceed the maximum permitted
velocity (e.g., an 18 in/sec (60 ft/sec) tip velocity) for steam-
assisted and nonassisted flares.  As can be determined by reviewing
40 CFR 60.8, emergency releases to the flare are not considered normal
operation and,  therefore, the velocity limits do not apply in such
circumstances.  The specifications described in API RP 521 ("...a
velocity of up  to 0.5 Mach for a peak, short-term, infrequent flow...")
are intended to be used in emergency situations, whereas the maximum
velocity requirements for flares specified in the EPA standards are
intended to be  used during normal flare operation.

     With respect to the "no visible emissions" limit placed on flares,
EPA considers a requirement for no visible emissions, with the exception
of 5 minutes in any 2-hour period, to be reasonable.  Smokeless
flares have been regulated in States with limits similar to the proposed
opacity limits.  The commenters offered no appropriate basis for changing
the limits to another level.  In addition, most streams in natural  gas
processing plants are highly volatile gases or light liquids, and are
readily burned  in smokeless flares.   Some commenters indicated that
continuous smokeless operation would require either automatic or manual
control of the  flare on a continuous basis.  However, EPA believes that
if this type of control  is needed to ensure the use of BDT, it is
reasonable, and that most gas plants would have little difficulty in
obtaining smokeless combustion.

     Comment:  One commenter (IV-D-31) thought that the establishment
of operating criteria is not permissible under Section 111 of the Clean
Air Act.  Section 111 specifically precludes the Administrator from
adopting design or work practice requirements unless it is not feasible
to prescribe or enforce a performance standard.  According to the
commenter, a performance standard is readily available in the case of
controlling VOC emissions by combustion and is included in the proposed
standards, namely, an opacity limit.

     Response:  As indicated by the  commenter, the "no visible emissions"
requirement for flares does constitute a performance type standard.   However.
this portion of the flare requirements only partially reflects BDT.
All  of the requirements  as promulgated are needed to reflect BUT.
Since unburned hydrocarbons can be colorless and can be emitted without
being visible, the opacity requirement does not provide a complete
measure of VOC destruction efficiency.  The EPA could set a performance
standard at a specific destruction efficiency level.   However, available
data indicate that, unless flares are operated within the velocity and
heating value requirements of the standards, they would likely not meet
performance specifications.   More importantly, performance testing  for

                                  4-15

-------
flares is extremely complex and costly,  so  that a  standard  requiring
demonstration of a given efficiency level  is,  in fact,  not  practicable.

     Comment:  One commenter (IV-D-24) claimed that the proposed
standards will  force extensive design  changes  to existing flare  systems
at a cost that would adversely affect  the  economics of  modification
projects.  The commenter stated that his company did not have  any gas
plants equipped with flares that meet  the  provisions of Section  60.632-9
of the proposed standards.

     Response:   If an existing flare does  not  achieve the required
limits of the promulgated standards, then  EPA  would not recommend using
the flare as a control  device because  such  a flare is not likely  to be
effective in burning VOC emissions.  However,  flares are not the  only
control devices suitable for use in controlling gas plant emissions.
The EPA believes that,  in most cases,  existing flares will  meet  the
requirements of the standards where installed.  In addition, the
velocity limits for flares  have been expanded  (allowing more flares to
comply with the standards)  and are not required to be maintained  during
emergency releases.

     For plants that do not have suitable  flares,  EPA has determined
that it is cost effective to install a new  flare solely for VOC  control
purposes if the plant decides that another  control device is not
available (Docket Item  IV-B-7).

     Comment:  One commenter (IV-D-33) noted that API Publication 931,
Chapter 14, states that residence times  of  0.2 to 0.7 seconds  at  1150
to 1400°F are sufficient to obtain satisfactory combustion  and
decomposition in most applications.  The commenters1 experience  confirms
the API parameters as sufficient to incinerate waste gases.  The  commenter
favored deletion of the residence time and  temperature  requirements
from the standards in favor of a 95 percent efficiency  requirement.

     Another commenter (IV-D-15) recognized that the residence time and
temperature parameters  were provided in  the standards as alternatives
to demonstration of 95  percent control  efficiency, but  was  concerned
that the parameters could be treated as  requirements for some new
processes where they would  not be necessary for 95 percent  destruction
efficiency.  The commenter also noted  that  the control  device require-
ments should only apply during normal  design conditions.

     Response:  As correctly noted by commenter IV-D-15, the temperature
and residence time specifications presented in the standards were provided
to allow demonstration of 95 percent efficiency through accepted  design
parameters rather than emissions testing.   Incineration systems  based
on other design parameters may be used subject to demonstration  of
control efficiency by source emissions testing or suitable  engineering
calculations.  Although the EPA parameters  of 0.75 seconds  and 816°C
(1500°F) are somewhat higher than those  stated in API Publication 931,
the EPA parameters are based on known 95 percent destruction efficiencies
for all cases, rather than "most applications" so that the use of the
EPA parameters can be accepted as proof  of 95 percent efficiency.  The
EPA realizes that, in some cases, shorter  residence times and/or lower

                                  4-16

-------
temperatures can provide 95 percent efficiency.   An owner/operator,
therefore, can provide design information  such as emission  rates  and
gas stream component analyses,  and incinerator design  specifications,
for approval as an acceptable control  device.

 4.7.2  Flares

     Comment:  Commenters IV-F-la; IV-D-29;  and  IV-D-23 stated that most
companies will not include a smokeless flare in  their  new plant designs.
Commenters IV-F-la and IV-D-23 also noted, however, that in plants that
have a flare system, the flare design  would  not  adequately  accommodate
the seal  leakage and/or would not meet the requirements of  the proposed
standards.

     Response:  Flares are one of several  VOC  control  devices that
might be used to comply with the standards.   Flares are not required
under the standards, but they are included because they can be used  to
reduce emissions of VOC at conditions  that reflect BDT.  Flares operated
in accordance with the operating conditions  provided in the regulation
are acceptable alternatives to other control devices used to comply
with the standards.  Although EPA allowed  for  plants that did not
include flares as a part of their initial  design in setting the compressor
standards, many plants are expected to include flares  because gas plants
are major sources of VOC; and in selecting best  available control
technology (BACT) and lowest achievable emissions reduction (LAER),
smokeless flares are likely alternatives for the control of VOC emissions
from emergency venting.

    Comment:  One commenter (IV-F-lc)  indicated  that the cost of
equipment to monitor flare pilot lights is high, and the equipment is
unreliable.  Another commenter (IV-F-le) noted that EPA provided no
cost estimates for flare monitoring equipment.

     Response:  The presence of a flare pilot  may be monitored with  a
thermocouple and appropriate transmitter.   Contrary to the  commenter's
suggestion, thermocouples are highly reliable  and are used  throughout
the petroleum industry.

     The costs associated with the flare requirements are discussed  in
Chapter 7.

4.7.3  Continuous Operation

     Comment:  Two comrnenters (IV-D-15 and  IV-D-20) questioned the
requirement in Section 60.632-9(d)(g)  of the proposed standards
for full-time operation of vent systems and  control devices.  The
commenters wrote that the operation of vent  systems and control devices
should not apply during shutdown and start-up  operations.  One commenter
(IV-D-20) aryued this is necessary because during such times the gas
composition may be highly variable.  Consequently, it would be difficult
to design a flare to handle such a wide range  of flows and  compositions.

     Another commenter (IV-D-15) requested that the requirements for  a
continuous burning pilot in Section 60.632-9(d)(2) , include "... or  an

                                  4-17

-------
acceptable reignition device."  The commenter stated that automatic
ignition devices should be allowed as an alternative to monitoring the
presence of a flare pilot.  Other commenters (IV-D-29, IV-F-la)  stated
that the fuel wasted to keep the flare burning would cost more than
could ever be saved from implementing the standards.

     Response:  Start-up and shutdown operations are described in the
General  Provisions.  Section 60.8(c) states that "Operations during
periods  of start-up, shutdown, and malfunctions shall  not constitute
representative conditions for the purpose of a performance test nor
shall emissions in excess of the level  of the applicable emission limit
during periods of start-up, shutdown, and malfunction be considered a
violation of the applicable emission limit unless otherwise specified
in the applicable standard."

     The EPA concurs that there is no basis for operating the flare when
no leakage is emitted.   The proposed standards (Section 60.632-9(g))
required closed-vent systems and control devices (including flares)
to be operated at all  times when emissions are vented J,p_ these devices.
Therefore, the commenter's suggestion is not~ necessaryT" This provi-
sion will  allow the use of a flare header flow detector and automatic
ignition system to prevent waste of pilot fuel.  However, when a flare
is operating, the pilot light must be burning at all times to ensure
the destruction of VOC emissions that may be released.

4.8  LEAK DETECTION AND REPAIR

4.8.1  Monitoring Frequency

     Comment:  Several  commenters (II-B-23; IV-F-lc ;IV-F-le; IV-D-14;
IV-D-15; IV-D-19; IV-D-27) stated that monthly monitoring of valves in
light liquid and gas/vapor service is too frequent.

     Three of the commenters (IV-F-le; IV-D-14; IV-D-15) urged EPA to
adopt quarterly monitoring for valve leaks, indicating that the small
additional emissions decrease does not justify increased cost for
monthly monitoring.  Another (IV-D-24) urged EPA to adopt quarterly
monitoring at most, yet recommend annual monitoring due to the mammoth
paperwork burden the standards create.

     One commenter (II-B-23) requested that EPA present an evaluation
of more extended inspection frequencies as supported by the API-Rockwell
study.  Another commenter (IV-F-lc) requested that the monitoring
period for valves be changed to annual, stating that the cost
effectiveness of monthly monitoring is questionable due to the large
number of valves and the manhours involved, while still other commenters
(IV-D-13 and IV-D-23) said that, even though the standards allow
quarterly monitoring, the initial frequency should be reduced to
(IV-D-23) noted that it will not always be possible to reach facilities
in North Dakota for monthly monitoring, whereas quarterly testing could
be scheduled around the weather.  In addition, quarterly monitoring
would reduce yearly travel expenses by two-thirds; in North Dakota this
is a $16,000-per-year savings.
                                  4-18

-------
     Response:  As stated in the preamble to the proposed standards,
 EPA considered monthly, monthly/quarterly, and quarterly monitoring
 intervals  for valves.  Each of these alternatives was compared in terms
 of the emission reduction achievable and the cost effectiveness of the
 leak detection and repair programs.  At proposal , the standards were
 based on monthly monitoring because it provided the greatest emission
 reduction  at reasonable costs.

     The EPA recognized, however, that some valves have lower leak
 occurrence rates than others.  Monthly monitoring of valves that do not
 leak for 2 consecutive months was judged to be unreasonable when
 compared to the additional emission reduction achieved by monthly
 monitoring over quarterly monitoring.  Therefore, EPA proposed to allow
 a monthly/quarterly implementation program, whereby valves that do not
 leak for 2 successive months may be monitored quarterly until  a leak is
 found.  Also, under one of the alternative standards for valves, an
 owner or operator can skip from the monthly/quarterly program to less
 frequent leak detection (semiannually or annually) if a performance
 test level of 2.0 percent is achieved on a continuous basis (for either
 2 or 5 consecutive quarters).  The EPA expects that most gas plants
 would implement a monthly/quarterly leak detection and repair program
 for valves.  In addition, the incremental  cost effectiveness between
 monthly/quarterly and quarterly leak detection and repair is reasonable
 (see Table 3-1).  Therefore, the standards still allow a monthly/
 quarterly  program for valves, which has not changed since proposal.

     One of the commenters suggested that valves be monitored quarterly
 from the onset of the program due to the remoteness of gas plants.   As
 shown in the BID for the proposed standards, Appendix G, EPA carefully
 analyzed the cost effectiveness of monitoring small  remote plants using
 central  office or contract personnel.  These analyses for small  plants
 indicated that, for plants smaller than 10 MMscfd, monthly monitoring
 was not cost effective.

     Comment:  One commenter (II-B-23)  questioned the validity of the
 leak detection and repair (LDAR) model  upon which the economics  of con-
 trolling emissions from valves and pump seals is based.  The commenter
 asserted that the model was not developed  using any gas plant data
 based on information presented in the AID (Docket Item II-A-25),
 Table 4-10.

     Another commenter (IV-D-19) wrote  that EPA's use of the leak
 occurrence rate function in the LDAR model  is faulty.   The commenter
 disagreed with the use of the 1.3 percent  per month leak occurrence
 rate based on limited tests in chemical  plants.  The commenter noted
 that his company's refinery experience  indicates a leak occurrence
 rate of 0.3 percent leakers per month.   The commenter acknowledged  that
 gas plant valves may be expected to leak slightly more than refinery
 valves,  but not by a factor of four.

     The commenter also disagreed with  the LDAR model's assumption  that
 the number of leakers  increases linearly with time.   As an example, the
commenter said  that the model  estimates  the percentage of leakers for an
annual  program  to  be at least 16 percent (1.3% x 12  months, plus minimum

                                  4-19

-------
level).  The commenter stated that this high  rate is  why EPA does  nut
believe it is reasonable to (allow leakers  to)  accumulate for greater
than 1 year.

     Response:   In response to the first commenter's  remarks, the  LDAR
model  does not  contain any data; it is merely an  algorithm to "model"  a
leak detection  and repair program.  To use  the model, monitoring data
must be provided as input.  The LDAR model  as presented  in the AID is
the basis for the cost analysis of controlling valve  and pump seal
emissions.  However, the use of the model  for natural  gas processing
plants was based on the best leak data available, including gas plant
emission factors.  These model results are  presented  in  Table E-l  of
the BID for the proposed standards and in Docket  Item II-B-18.

     The second commenter contends that the rate  of leak occurrence for
valves used by  EPA as inputs to the LDAR model  is high;  and the model
assumes that the number of leakers increases  linearly with time.   The
1.3 percent rate of leak occurrence for valves is based  largely on data
from SOCMI plants and has been corroborated by API valve maintenance
data from petroleum production operations (II-I-20, II-I-21).  The
EPA analyzed the API valve maintenance data and concluded that the API
data support the estimates of leak occurrence and recurrence used  by
EPA for gas plants (II-A-9).

     The LDAR model- assumes a constant valve  leak occurrence rate  as
mentioned by the commenter.  However, this  rate is applied by the  model
on only the non-leaking sources after each  month.  For example, the
1.3 percent occurrence rate mentioned by the  commenter would result in
13 leaks per month during the first 6 months (for a plant with 1,000
valves), 12 leaks per month during the next 6 months  and so forth  until
all 1,000 valves were leaking, which would  require over 40 years at a
1.3 percent/month rate.  Although another description of the rate  at
which  leaks develop might be  slightly more  accurate,  the method used  in
the LDAR model  is based on reasonable assumptions necessary for modeling
an average case.  For example, valves within a plant may have different
leak occurrence rates due to  different designs, services, manufacturers,
or frequencies of use.  Modeling of such factors would require extremely
complex information over long periods of time for every  plant analyzed.
Consequently, the LDAR model  attempts to simplify the leak occurrence
phenomena to represent an average case for all plants based on test
data  from typical processing  plants.

     The second commenter indicated that his company's refinery
experience  indicated a leak occurrence rate of 0.3 percent leakers per
month, and  that gas plants may be expected to have slightly  higher
occurrence  rates.  However, data  provided by another commenter (IV-D-26)
showed a 3.74 percent monthly occurrence rate at  a gas plant  recently
tested, three times the occurrence rate used by  EPA  for  the  LDAR model
input.  Therefore, EPA considers  these comments  to confirm the reason-
ing used  by  EPA that, although occurrence rates  will  be  widely variable,
average valve leak occurrence rates can be used  in developing model
programs  for cost-effectiveness  analyses.
                                  4-20

-------
     Comment:  One commenter (II-B-23) stated that the LDAR model
 generates negative control efficiencies for some inspection frequencies.

     Response:  The commenter is pointing out an incorrect use of the
 LDAR model.A negative control  efficiency theoretically is generated
 when the occurrence rate  for a given inspection interval is greater
 than the initial leak frequency.  This may occur at a plant with an
 overall occurrence rate of 1.3 percent, in which the leak rate
 increases after leak detection and repair.

 4.8.2  Delay of Repair

     Comment:  Two commenters (IV-D-15; IV-D-35) requested that the
 delay of repair provisions in Section 60.632-8(a) be revised to read,
 in part, "Repair ... shall occur, however, at the first scheduled
 process unit shutdown."   The commenter held that during unplanned
 shutdowns, operating personnel  are trying to make repairs due to upsets
 or equipment malfunctions.

     Response:  The EPA agrees with the commenter that, during brief
 emergency shutdowns, operating personnel  may not be available for
 repair of leaking components due to the importance of restarting the
 process.  To make allowances for delay of repairs beyond an unscheduled
 shutdown in the case of shutdowns of too short a duration for operating
 personnel to make repairs, a process unit shutdown has been defined as
 longer than 24 hours.   Since proposal, the definition of "process unit
 shutdown" has been revised to exclude from the definition any unscheduled
 work practices or operational procedures that stop production from a
 process unit or part of a process unit for less than 24 hours.  This
 definition is in Subpart VV, standards for equipment leaks of VOC in
 SOCMI, and is considered reasonable for gas plants as well.

     Comment:  One commenter (IV-F-4) stated that Section 60.632-8
 of the proposed standards should have provisions to extend the repair
 interval beyond 15 days in the event that replacement parts, equipment,
 or personnel  are unavailable.  The commenter held that the delay of
 repair requirements should be revised because these provisions unfairly
 penalize the small  gas processor.  The commenter claimed that large
 inventories of spare parts cannot be maintained at small plants due to
 the high capital  costs, tax considerations, and space requirements
 i nvolved.

     Response:  The provisions  of Section 60.632-8 of the proposed
 standards allow for delay of repair beyond a process unit shutdown in
 the event sufficient supplies were stocked and then depleted.   As
 recognized by the commenter, this "parts  depletion" provision  does not
 apply to the normal  15-day repair period  for sources not requiring a
 process unit shutdown.   Many smaller gas  plants are remotely located
and may maintain little,  if any, stock of spare parts.   However, many
of these plants would  have an existing source of readily available
 parts to prevent unnecessary plant downtime.  Most parts should be
available on  an immediate-delivery basis, either from a  central
location from the company owning the plant, or from parts distributors.
                                  4-21

-------
     Even though it is conceivable that  a  plant  could  have  a monitoring
period in which a greater than  normal  number  of  repairs  requiring spare
parts was necessary and depleted  the  inventory of  spare  parts,  such
situations should be rare.   In  any event,  provisions  for delay  of
repair beyond the normal  15-day period should not  be  necessary.  Since
most repairs requiring parts  replacement would require that the  leaking
component be removed from service, the component requiring  the  parts
would either be one which can be  isolated  from the process  for  repair
or one which could not be repaired until the  next  shutdown.   In  the
latter case, the delay of repair  provisions allow  for  delay beyond
shutdown if sufficient stock  of spare parts  is depleted.

     For components isolated  from the process for  repair, the  leaking
component may remain isolated until  parts  are acquired.   Repair of
these components legally could  be delayed  indefinitely  (provided there
are no VOC in the line), and  no provision  for spare parts,  other than
the one provided in Section 60.632-8(b), is required.

4.8.3  Sealless Equipment

     Comment:  One commenter  (IV-D-20) noted  that  EPA discusses seal less
pumps as not having a potential leak  area  and then requires initial
and annual performance tests.  The commenter  questioned  why leakless
equipment requires any monitoring and suggested  that  the requirements
for monitoring leakless equipment be  deleted.

     Response:  The standards require an initial performance  test to
verify that a piece of leakless equipment  meets  the "no  detectable
emissions" limit and annual rechecks  to ensure continued operation  with
"no detectable emissions."  The EPA believes  that these  requirements  are
necessary to verify the integrity of  the equipment at installation  and
throughout its operation.  Leakless equipment may fail,  for example,
as a result of corrosion or wear.  For these  reasons, EPA has determined
that initial and annual performance tests  are necessary.

4.8.4  Repair Period

     Comment:  Several commenters (II-B-23,   II-D-10,  II-D-30, IV-D-19,
IV-D-20,  IV-D-27,  IV-D-30, IV-D-34, and IV-D-35) objected to  the repair
period requirements.  Commenters  II-B-23,  IV-D-27, IV-D-34, and IV-D-35
requested an extension of the  first attempted repair period from 5
calendar days to 7 and 15 days.   Another commenter (IV-D-19)  requested
that the final repair requirements be changed to at least 30 days.
Reasons offered for these changes include:

     1.  Replacement parts are not available within 15  days,  therefore,
         gas plants would have to maintain an expensive  inventory
         with no significant air  quality benefit  (IV-D-19),

     2.   Remoteness of plants, inclement weather, and manpower
         availability at small plants justify extending  the repair time
          (IV-D-27),
                                  4-22

-------
     3.  An extension would allow a regularly scheduled maintenance
         crew to handle repairs, whereas retention of the 5-day limit
         would require contractor maintenance personnel to be paid  for
         expensive overtime and holiday work and create scheduling
         problems (IV-D-34 and IV-D-35), and

     4.  At most gas processing plants personnel are not available  to
         repair all  leaks within a 5-day period.  Most new plants  are
         designed to run unattended for most of the day.

     Another commenter (IV-D-20) asked that the 5-and 15-day repair
time limits not apply to equipment for critical  valves, pumps, and
compressors used to keep the process operational.  There is no reason
to specify a repair period for equipment that can be taken out of
service.

     Response:  The standards require that a first attempt at repairing
a leaking equipment should be accomplished as soon as practicable  but
no later than 5 days after detection of a leak.   Attempting a first repair
of the leak within 5 days will help maintenance personnel identify  the
leaks which can be repaired without shutdown of the process unit.
Equipment that continue to leak after simple field repair attempts  must
be repaired within 15 days following initial leak detection.  This
interval provides time for properly isolating equipment that require
more than simple field repair.  The 15 days provides sufficient time to
schedule and effect on-line repairs that a shorter period might not
allow.  Provisions have been made for delaying repair of valves in
critical service that cannot be bypassed.  The two repair period require-
ments provide efficient reduction of emissions and allow sufficient
time for flexibility in scheduling repairs of leaking equipment.  A
15-day period for initial repair would simply permit delays in repairs
that could otherwise be accomplished quickly.

     Most valve repairs can be done quickly and involve simple mainte-
nance procedures, such as packing gland tightening and grease injection.
This is evident from compliance experience of refineries with the  South
Coast Air Quality Management District Rule (Rule 466.1) for valves
which require repair within 2 working days.  A 5-day period for initial
attempts provides sufficient time to schedule field repair.

     Valves that require off-line repair (25 percent is estimated  in
the BID for the proposed standards, Chapter 8) may require new packing
or valve parts (i.e., gland flange or nut); however, these spare parts
and others would be stocked on-site for routine maintenance.  Therefore,
stocking of these parts does not represent a burden imposed by the
standards.

     The EPA realizes that many gas processing plants would be located in
remote areas, have limited manpower availability, and subject workers
to inclement weather.  Consequently, EPA has exempted small nonfraction-
ating plants from the leak detection and repair requirements based  on
the cost effectiveness to employ corporate office personnel or contractors
to perform the leak detection and repair program.  The EPA believes that
the leak detection and repair requirements, including the repair periods,

                                  4-23

-------
are reasonable for all  fractionating plants and  nonfractionating  plants
with a capacity of 10 MMscfd  or more.

4.8.5  Process Unit Shutdown

     Comment:   One commenter  (II-D-10)  remarked  that some  sources of
emissions in gas plants cannot be  repaired  until  there  is  complete
plant shutdown, adding  that a mandated  routine shutdown  for  leak  repair
would result in greater emissions  than  would be  saved  by repair.   The
commenter noted that the majority  of gas  plants  do  not  have  periodic
turnarounds as refineries do  because gas  plants  process  gas  as  it is
produced from  field wells; otherwise gas  products are  lost.

     Response:  The standards do not require routine shutdown  for repair
of leaks for the reason cited by the commenter that a mandated  routine
shutdown could result in greater emissions  than  would  be reduced  by
repair.  The EPA does allow owners or operators  to  delay repair of
leaking valves, pumps,  and relief  valves  beyond  15  days  after  leak
detection if the component cannot  be repaired without  a  shutdown.  For
valves, EPA allows a delay of repair beyond a facility  shutdown if  the
entire valve assembly is required  to be replaced, provided the  plant
owner or operator can demonstrate  that  sufficient stock  of spare  valve
assemblies was maintained before depletion  of the stock  of spare  parts.

     Comment:   One commenter  (IV-D-29)  requested that  the  word
"production" in the definition of  "process  unit  shutdown"  be amended  so
that it is clear it means production of natural  gas and/or liquids.

     Response:  The definition of  "process  unit  shutdown"  clearly states
that a shutdown is a termination of "production  from a  process  unit  or
part of a process unit."  Since a  "process  unit" is also defined  as
equipment assembled for the extration and fractionation  of natural  gas
liquids, the word "production" in  the definition clearly means  the
extraction or  fractionation of natural  gas  products.

     It should be noted that  natural gas  production from the field  can
continue while the processing plant is  shut down.
                                  4-24

-------
                       5.0  APPLICABILITY OF STANDARDS
 5.1  DEFINITIONS

 5.1.1  Natural  Gas Processing Plant

      Comment:  Eleven commenters (IV-D-10; IV-D-11; IV-D-14;  IV-D-15;
 IV-D-16; IV-D-17; IV-D-19; IV-D-34; IV-D-35;  IV-F-lc;  and IV-F-le)
 requested that  the definitions of "natural gas processing plant"  and
 "processing unit" in Section  60.631 of the proposed standards  be  clarified
 to exclude units that EPA intended to  exempt,  such as  field  separators
 and well site production  facilities.   Most of  these commenters  suggested
 that changing the word "separation" in the definition  of natural  gas
 processing plant to "extraction" would imply  units operating with a
 forced process,  while "separation" implies simple  removal  by gravity or
 natural  condensation.

      Another cornmenter (IV-D-11) asked that EPA confirm  in the  promul-
 gated standards  that the  regulations do not apply  to facilities such as
 sour gas treatment facilities that separate impurities other than
 natural  gas liquids from  the  field gas.   Other commenters  (IV-D-14;
 IV-D-16; IV-D-34; and  IV-F-le)  asked that facilities not intended to be
 covered  by the regulation  be  clearly excluded.   One commenter  (IV-D-19)
 asked that the exemption  given  in Section 60.630(e) clarify that  "not
 located  at"  means not  within  the boundaries of the gas plant.

      One commenter (IV-D-15)  requested deleting  subparagraphs  (1), (2),
 and (3)  of Section 60.630(a), and stating in paragraph (a)  that  "The
 provisions  of this subpart apply to specific process units within
 natural  gas  processing  plants."   The cornmenter  favored this change
 because  it  eliminates  the  need  for the field facility exclusion in
 paragraph  (e) of Section  60.630  and because, he  argued,  compressors
 should not  be designated  as affected facilities, but should be
 specifically  excluded.

      Response:   In  the  proposed  standards,  a "natural gas processing
 plant" (gas  plant)  is defined as  "any  processing site engaged in
 the  separation of  natural gas liquids  from  field gas, fractionation of
 mixed natural gas  liquids to  natural gas  products,  or both."  The
 definition was intended to exclude  facilities  that  remove liquids from
 field gas by means  other than a  forced  process  (e.g., gravity or
 natural  condensation).  Therefore,  EPA  has  revised  the definition of
 "natural  gas  processing plant" in  the  promulgated  standards to read
 "...  any processing site that extracts natural  gas  liquids from  field
 gas." The definition of "process  unit"  has  also  been revised by substituting
 the word "extraction" for the word  "separation."

     The EPA considers sour gas treatment  facilities to  be affected
 facilities (i.e., subject to   the  standards) unless  they are not located
at an onshore natural gas processing plant.  The EPA knows of no  reason
 (and the commenters did not provide any reasons) why sour gas treatment
facilities should be exempted  from  the  standards.   If these facilities
have equipment in VOC service, then they  should be covered by the standards.

                                  b-1

-------
     Subparagraphs (1), (2), and (3) of Section 60.630(a) clearly
define the affected facilities for natural gas processing plants.
Compressors are affected facilities, as explained in Section 60.630;
therefore, subparagraph (2) is necessary.  Paragraph (e) is also necessary,
regardless of how paragraph (a) is worded, to make it clear which process
units and compressors are exempt from the requirements of the standards.

5.1.2  Liquified Petroleum Gas and Natural Gas Liquids

     Comment:  One commenter (IY-D-10) suggested adding the following
definitions to the standards:

     •  "Liquified petroleum gases" mean ethane, propane, normal
         butane, and/or iso-butanes or mixtures of these liquified
         gases.

     •  "Natural gasoline" means a liquid composed of pentanes and
         heavier hydrocarbons  or mixtures of pentanes,  heavier hydrocarbons,
         and/or lighter hydrocarbons.

     Another commenter (IV-F-1; No. 5) suggested changing the term
"natural gas products" to "liquified petroleum gases."

     Response:  As noted by the first commenter, liquified petroleum
gases (LPG) generally consist primarily of propane and butane, while
"natural gasoline" consists of pentanes and/or heavier hydrocarbons.
Both of these streams are considered VOC by EPA and are, therefore,
subject to the requirements of the standards.  The EPA used the terms
"field gas" and "natural gas liquids" to define the feedstock and
product streams in natural gas processing plants, and the term "natural
gas products" to define the individual products (such as LPG or natural
gasoline) resulting from the fractionation of natural gas liquids. As
such, both liquified petroleum gases and natural gasoline as defined  by
the commenter would be included in the term "natural gas products,"
which is contained in the definition of "natural gas processing plant."
Since "LPG" and "natural gasoline" could have specific meanings, the
use of these terms could result in exclusion of other natural gas
products that were intended to be included.  Therefore, EPA will continue
to use the terms "natural  gas  liquids" and "natural gas products" as
generic terms for natural  gas  processing plant product streams.

5.1.3  VOC

     Comment:  One commenter (IV-D-20) requested that ethane handling
equipment be exempt from the requirements of the standards.  The
commenter stated that the preamble indicated that methane and ethane  have
negligible photochemical reactivity, but the definition for natural  gas
liquids does not exclude ethane.  The commenter indicated that this  lack
of exclusion would result in requiring inspection of ethane handling
components.

     Another commenter (IV-D-25) requested that EPA revise the definition
of VOC in Section 60.2 of the  General Provisions to specifically exclude
methane and ethane.   The commenter thought a revised definition was
necessary because Reference Method 21 detects any organic compound and

                                  5-2

-------
 nonreactive organic compounds (such as methane and  ethane  which  are  not
 considered VOC by EPA).   The commenter noted that methane  is  used  as  a
 calibration gas in Reference Method 21.

      Response:  The first commenter is correct in stating  that the
 definition for natural  gas liquids  (NGL)  does not exclude  ethane.
 However, a piece of equipment is covered  by the  standards  if  the
 stream is in VOC service, not NGL service.   The  standards  allow  the
 exclusion of substances  not considered photochemically  reactive  by EPA
 when determining the percent VOC in the process  fluid (i.e.,  determining
 whether a piece of equipment is  in  VOC service).  The VOC  content  is
 determined by the ASTM methods  E-260,  E-168, and  E-169, not by Reference
 Method 21.  Therefore, the fact  that ethane is included as an NGL  does
 not necessarily mean that ethane-handling components are covered by the
 standards.  Thus, the exclusion  of  ethane from the  definition of NGL  or
 a revision to the definition of  VOC in the  General  Provisions is
 unnecessary.

 5.1.4   Control  Device

      Comment:   One commenter (IV-D-36)  recommended  that EPA revise the
 definition of a  control  device  to read that a  control device  "means an
 enclosed combustion device,  vapor recovery  system,  or flare,  designed in
 accordance with  the provisions of Section 60.632-9."

     Response:   The EPA  does  not  believe that  it  is necessary to amend
 the definition  of control  device as recommended by  the commenter.
 Section 60.631 of the proposed standards provides a definition for a
 "control  device,"  and control device requirements are given in Section
 60.632-9,  which  has  been  revised as discussed  in  Section 4.7.
 5.1.5   Heavy  and  Light Liquid Service

     Comment:   One  commenter  (IV-D-35)  recommended  changes in the
 definitions of  "heavy liquid  service"  and "light  liquid service."  The
 commenter  indicated  that  all  liquid service  streams  are currently excluded
 from the  "heavy  liquid service" definition.    For  light liquid service,
 the commenter suggested using the ASTM test methods  incorporated in
 Section  60.635(d)(2)  instead of the stream  VOC content as used in
 Section  60.635(e).

     Response:   The commenter has correctly noted two errors  in  the
 regulation, which have been corrected  in the promulgated standards.
 One error, the definition  for "in heavy liquid service," was  corrected
 to  read  "... means  that a  piece of equipment is not  in gas/vapor service
or  in light liquid service."  The word  "light" was inadvertently omitted
 from the definition.

     The second error noted by the commenter is in the definition for
 "in light  liquid service."  The EPA incorrectly cited paragraph  (e) of
Section 60.635 instead of paragraph  (d)(2)  in referring  to  the determination
of equipment in light liquid service.   This  error has also  been  corrected
in the promulgated standards by incorporating the definition  in  Subpart
VV, as  amended.
                                   5-3

-------
5.1.6  Quarter

     Comment:  Two commenters  (IV-D-35 and  IV-D-36)  requested  that  EPA
change the definition of "quarter."  One commenter  (IV-D-36)  recommended
that the definition read that  quarter "...  means  a 3-month period,
commencing on January 1, April  1,  July 1, and October 1  of each year"
because the gas processing industry  generally uses  this  definition  for
all other internal, State, and  Federal  reporting requirements  and,
consequently, standardization  of the NSPS reporting  requirements with
industry practices would facilitate  compliance with  these requirements.

     The second commenter (IV-D-35)  recommended that NSPS monitoring
start with the first quarter after the first  180 days have elapsed.
The commenter suggested changing the definition of  "quarter" in the
standards from "The first quarter  concludes on the  last  day  ..." to
"The first quarter commences the next day after the  last day of the
last full month during the 180  days  following initial  startup."  The
commenter wrote that ending the first quarter during the 180-day startup
period essentially reduces the  startup period to 90  days while gas  plants
typically require about 180 days to  startup,  debug,  balance  and establish
routine operations.

     Response:  The EPA does not wish to force industry  into specific
calendar dates for monitoring  activity.  Therefore,  calendar dates  are
not used for defining quarters; however, a  plant owner or operator  may
choose to use the calendar quarters  suggested by the commenter provided
the first date of the calendar  quarter is no  later  than  the  day after
the last day of the last full  month  during  the 180  days  following
initial startup.

     The regulation specifies  that the plant  be in  compliance  with  the
standards within 180 days of initial startup, not  after  the  180-day
period.  The EPA has been given no data or  documented information  in
support of the second commenter's  claim that  gas  plants  typically
require the full  180 days to prepare to be  in compliance with  the
standards.  Startup times for  gas  plants are  not  expected  to be any
different from startup times for refineries  or chemical  plants.

5.1.7  In VOC Service

     Comment:  Numerous commenters (II-B-23;  II-D-30; IV-D-13; IV-D-14;
IV-D-15; IV-D-16;  IV-D-17; IV-D-19;  IV-D-21;  IV-D-23; IV-D-24; IV-D-25;
IV-D-26; IV-D-27;  IV-D-29; IV-D-30;  IV-D-31;  IV-D-33; IV-D-34; IV-D-35;
IV-D-36; IV-D-37;  IV-F-la; IV-F-lb;  IV-F-le)  requested that  EPA raise
the VOC concentration limit for "VOC service" from  1 weight  percent
VOC to either 10 volume percent VOC  or 10 weight  percent VOC.   Several
reasons were presented by the  commenters,  including:

     •   Use of a  10 weight percent  VOC limit would  make the standards
         consistent with other equipment leak standards.

     •   Raising the limit on  VOC concentration would concentrate
         attention on equipment that is more  likely to exhibit VOC
         leaks and would result in more cost  effective control.

                                  5-4

-------
     •   Under the current limit of 1 weight percent, the commenters
         indicated that many "residue gas" streams, which EPA intended
         to exempt, would be covered.  Other commenters (IV-F-la;  IV-D-
         26; IV-D-36) stated that an industry-wide survey indicated
         that a 10 weight percent cutoff would exclude dry gas service
         components at most of the plants surveyed, and that EPA's data
         base for establishing a 1 weight percent limit was inadequate.

     •   One commenter provided plant size and residue gas VOC content
         data for over 100 plants to illustrate his point.  Similarly,
         another commenter (IV-D-21) wrote that only 13 out of 36  of
         his company's operating plants produce residue gas that is
         less than 1.0 weight percent VOC, while 29 of these same
         36 plants have residue gases that are less than 10 weight percent
         VOC.

     •   Another of the commenters (IV-F-le) said that a 1 weight
         percent limit implied an uncommonly high degree of liquid removal.

     •   Two commenters (IV-D-15, IV-D-35) wrote that the API/Rockwell
         Study shows primarily methane streams (residual or sales  gas),
         a non-VOC, would be exempt by this 10 percent criterion.

     •   One commenter (IV-D-17) indicated that "pipeline quality" gas
         often contains more than 1 weight percent VOC and indicated
         that EPA had told him that pipeline gas was intended to be
         excluded.

     •   Other commenters (IV-D-31; IV-D-35) suggested, as an alternative
         to or in addition to changing the definition of VOC service,
         that residue gas be classed a priori  not in VOC service.

     •   One commenter (IV-D-19) stated that a refinery valve handling
         a 10 weight percent VOC stream has roughly the same health
         effects and control  costs as does a valve in a gas plant
         handling a 10 weight percent VOC stream.  The commenter added
         that EPA's argument that 1 to 10 weight percent streams
         contribute a significant portion of gas plant emissions is  not
         sufficient justification to control that portion, especially
         when the total source category emits  very little.

     •   One commenter (IV-D-24) remarked that many automatic control
         valves are operated using "natural gas" with VOC content
         greater than 1.0 percent by weight.  Consequently, these  valves
         have to be replaced or a new instrument air compressor must
         be purchased.  The commenter also stated that product natural
         gas from the average yas plant has a  VOC content in the range
         of 3 to 8 percent by weight.  The commenter thought the VOC
         content should be raised to 4 mole percent or 8 weight percent.

     Response:   In setting the VOC concentration limit, EPA took into
consideration the VOC content below which equipment leak controls  may
not be cost effective.  In the case of synthetic organic chemical
plants and petroleum refineries, the costs of  controlling the small

                                  5-5

-------
number of streams containing less than 10 weight percent VOC  appeared
to be unreasonable in light of the emission reduction  potential.
Therfore, EPA considered 10 weight percent VOC to be an  appropriate  VOC
concentration limit for those standards.   In contrast,  gas  processing
plants can have a large number of components in streams  containing
between 1 weight percent and 10 weight percent VOC, and  the cost
effectiveness of controlling emissions from these components  is
reasonable.  Thus, the lower VOC concentration limit appeared to  be
warranted as a way to cover these streams.

     The underlying concern of the commenters was that,  while EPA
stated in the preamble that product natural gas (residue ("dry")  gas)
was to be excluded, the 1.0 weight percent VOC limit was so low  that
many residue gas streams would be covered by the standards.  To  support
this contention, the Gas Processors Association (GPA)  provided data  for
throughput, non-organic content, and VOC  content for  the residue  gas
streams from over 100 plants.  An analysis of these data is presented
in Docket Item IV-B-10.  Some of the plants listed in  the GPA survey had
residue gas VOC contents below 1.0 weight percent VOC,  with many less
than 0.1 weight percent VOC.  Most of the plants had  residue  gas  streams
containing between 1.0 percent and 10 weight percent  VOC, and a  few
plants had residue gas streams exceeding  10.0 weight  percent  VOC.
The GPA data illustrated to EPA that 1 weight percent  VOC was not
an appropriate limit to exclude residue gas streams,  since  residue  gas
streams can often contain more than 1 weight percent  VOC.  Therefore,
the Administrator decided to raise the limit.  Considering  that  the
costs of controlling most gas plant streams, other than  wet gas  streams,
containing less than 10 weight percent VOC would be  too  high  for the
emission reduction that would be achieved, a VOC concentration limit of
10 weight percent was selected as a representative limit for  the "in
VOC service" definition.  Therefore, the  definition  of "in  VOC service"
has been changed in the promulgated standards to refer to a 10 weight
percent VOC content.

     However, the original intent in utilizing a 1.0  weight percent  VOC
limit was to ensure that inlet (wet) gas  streams were  subject to NSPS
control, since emissions can be reduced at reasonable  costs from inlet
gases.  Therefore, the promulgated standards require  that all equipment
in wet gas service be controlled (except  wet gas reciprocating compressors),
regardless of VOC content.  Since wet gas streams may  vary around 10 weight
percent VOC, this requirement will prevent repeated  testing to determine
VOC service and will clarify the intent of the standards for  plant
owners, operators and enforcement personnel.

       Comment:  One commenter (IV-D-37)  recommended  that  Section
60.635(e)(l) be revised to require that a detennination that  a piece of
equipment is not in VOC service should only apply during normal  and
ordinary operations.  According to the commenter, this change is
necessary to eliminate brief periods during upset conditions  when the
line may contain heavier hydrocarbons than nonndl.

     Response:  The General Provisions, Section 60.8(c) of the Clean
Air Act, cover operations during upset conditions.  Specifically,
Section 60.8(c) states that:

                                  b-6

-------
          Operations  during  periods  of  start-up, shutdown, and malfunctions
          shall  not constitute  representative conditions  for the purpose
          of  a  performance test nor  shall emissions in excess of the level
          of  the applicable  emission  limit during periods of startup, shut-
          down,  and malfunction be considered a violation of the applicable
          emission  limit  unless otherwise specified in the applicable standard.


          Therefore,  the  commenter's  suggested revision is unnecessary.

 5.1.8   Vacuum  Service  Components

     Comment:   One commenter (IV-D-20) agreed with the exclusion for
 vacuum  equipment.  The commenter requested that common English equivalents
 of measurement  be  included  in  parenthesis where SI units are given.

     Response:   The  EPA  has not found  it necessary to revise the definition
 for  "in vacuum  service"  by  including the appropriate English equivalent
 measure to the  SI  units  given.  The  conversion from kilopascals (kPa)
 to the  appropriate English  equivalent, pounds per square inch (psi), is
 6.895 kPa =  1  psi.   Therefore, the definition for in vacuum service
 expressed in English units  would mean  equipment operating at an internal
 pressure that  is at  least 0.725 psi  below ambient pressure.

 5.1.9   Connectors

     Comment:   One commenter (IV-D-20) stated that requirements to
 repair  leaking  pumps and valves in heavy liquid service, relief valves
 in light and heavy liquid service, and flanges and other connectors are
 generally reasonable.  The commenter did not agree that it is reasonable
 to define a "connector"  as a welded  joint.  The commenter argued that
 such connections do not  leak and should not be subject to these
 requi rements.

     Response:   The EPA maintains that it is reasonable to include a

 welded  joint in  the definition of a  "connector".   Welded joints may
 leak due to an  improper  seal or fitting, or leaks may develop over time
 due to  corrosion or deterioration of the weld.   Therefore, welded
 joints,  like other connections, are  subject to the standards as specified
 in Section 60.632-7 of the proposed standards.   If evidence of a potential
 leak is  found by visual, audible, olfactory, or other detection method,
 the connection must be monitored within 5 calendar days (or assumed to
 be a leaker), a  first attempt  at repair must be made  within 5 days of
 detection; and repair must be completed as soon as practicable, but no
 later than 15 calendar days  after the leak is detected.

 5.2  SELECTION 01 SOURCES

 b.2.1  .VdJ_v_e_s jin_d jDpen-Ended Lines

     Comment:  One commenter (IV-D-3) noted  that  valves and open-ended
 lines account for 79  percent of the  VOC emissions  from gas  plants  and
questioned why the standards are not limited  to  these two sources.

                                      5-7

-------
     Response:   Based on the  use of  the revised  compressor  seal  emission
factors (BID for proposed standards,  Appendix  G),  valves  and  open-ended
lines represent approximately 70 percent (191  Mg/yr  for valves  and
open-ended lines and 82.8 Mg/yr for  other sources  combined  from Model
Plant B) of VOC emissions from gas  plants.

     The EPA agrees that valves and  open-ended lines  represent  the
largest sources of gas plant  emissions.  However,  the baseline  emission
factors in the  BID for the proposed  standards, Table  7-2, as  corrected
in the appendices, show that  pressure relief devices, compressor seals,
and pump seals  are also significant  emission sources.  Together these
sources account for 49.2 Mg/yr of VOC emissions  from  a typical  plant.
The emissions from these sources are  reasonable  to control  as shown  in
Table 3-1.

     In selecting best demonstrated  technology (BDT), EPA selected
standards for each type of component  in gas  plants with demonstrated,
cost-effective  control techniques.   Since the  general purpose of NSPS
is to require new facilities  to incorporate  best demonstrated technology
as they are being constructed, and  cost-effective  controls  are  available
for compressors, pumps, and pressure  relief  devices,  these  sources  are
also included.

5.2.2  Heavy Liquids                                     '

     Comment:  One commenter  (IV-D-15) wrote that  the standards should
delete the monitoring requirements  for components  "in heavy liquid
service." The commenter noted that  these components  have  a  low  potential
for VOC emissions.

     Response:  The proposed standards (Section 60.632-7)  do not require
routine monitoring of components in  heavy liquid service.  Components  in
heavy liquid service are not  as likely to develop  leaks  as  are  gas
service or light liquid service components.   However, heavy liquid
service components that do "leak have emission rates  comparable  to
components in gas or light liquid service.   Therefore, the  standards do
require that leaks detected by visual, audible,  or olfactory means  be
measured to determine if the  leak results in an  instrument  reading  of
10,000 ppm or greater and, if so, that the leaks be  repaired as soon as
possible within 15 days, with a first attempt  at repair within  5 days.
The operator could repair any leaks  found in heavy liquid service
without prior instrument measurement, if desired.

5.2.3  Small Valves or Lines

     Comment:  Several commenters (IV-D-15,  IV-D-20,  IV-D-22,  IV-D-23,
IV-D-26,  IV-D-36, and  IV-D-37) wrote that small  valves or lines should
be exempt from the standards.  Commenters requested  exemptions   for
valves ranging from 2  inches  or less to  1/2-inch or  less.  The commenters
contended that eliminating small valves  from the routine monitoring
requirements would greatly reduce recordkeeping, identification, and
reporting requirements.  The  commenters claimed that  small  valves
tend to leak less, both  in frequency and amount of leakage,  and
contribute a relatively  small percentage of total  fugitive VOC  emissions

                                       b-8

-------
 at  gas  processing  plants.   The  commenters  also  argued  that  exempting
 these valves  from  monitoring  requirements  would  reduce overall  plant
 monitoring  costs.

     Response:   In developing similar  standards  for  petroleum  refineries,
 EPA analyzed  both  the costs of  small valve  repair as compared  to the
 BID average repair cost  estimates and  the  emission factor dependency
 on  valve  size (Docket  Item  IV-B-11).   These analyses indicate  that the
 emission  rates  from  leaking valves  are independent of valve size and
 that repair of  small valves is  no more difficult or costly  than repair
 of  larger valves.   The EPA concluded that,  since both the environmental
 benefits  and  cost  of control  for small  valves are the same  as  larger
 valves, small  valves are cost effective to  control.

     The  emission  source test data  presented in  Appendix C  of  the BID
 for the proposed standards were collected at six natural gas/gasoline
 processing  plants  by EPA and  industry.  Valves of all sizes were
 monitored in  these tests, and valve size was determined not to  be a
 factor.   The  emission factors developed from these test data are based
 on  all  valves  in gas processing plants.  The basis of the standards for
 valves  included all  valve sizes and was determined to be cost  effective.
 Therefore, an exclusion  for small lines and valves is not 'warranted and
 is  not  included in the promulgated  standards.

 5.2.4   Small  Compressors

     Comment:   One commenter  (IV-D-21) was concerned that the  standards
 for compressors apply to all sizes  of  compressors including those used
 to  recompress vapor  from storage tanks, refrigerate gas, boost  regen-
 eration gas, and even collect emissions from a larger compressor.   The
 commenter complained that without a horsepower cut-off in the  designation
 of  a compressor as an affected facility, the replacement or modification
 of  a small compressor at an existing plant could require installation
 of  a complete vent control  system for  all  the compressors at a  plant.

     Response:  As with other fugitive emission sources, EPA has found
 no  relationship between component size and emissions.  Emission test
 data were collected  from all sizes of components and  compressors,  so
 that the emissions and cost effectiveness  of controlling compressors
 are based on compressors of all  sizes.   Consequently, EPA knows of no
 reason to exempt small  compressors based on horsepower.

 5.2.5  Difficult-to-monitor Valves
     Comment:   A number of  commenters  (IV-D-23; IV-D-26;  IV-D-30;
 IV-D-33 and  IV-D-36) recommended that  EPA  exempt insulated  valves'  from
 the monitoring requirements of the proposed standards.   The  commenters
 stated that  some valves in  natural  gas  plants  require insulation because
 they are in  very hot or very cold service.   Monitoring  of these valves
 would require  removal and replacement of the insulation,  often by  an
 insulation contractor,  which would be both  time consuming and  costly
 Two commenters (IV-D-30;  IV-D-33)  noted problems with icing  on some
cryogenic  temperature valves,  delaying  replacement  of the insulation
until  a  shutdown so that  components  may warm up.
                                       5-9

-------
     Response:  The EPA recognizes that additional  time  would  be  required
to remove and replace valve insulation to monitor or repair  these
components.  There may be problems associated with removing  insulation
(e.g., ice formation).  Nevertheless, insulated  valves can  be  monitored
as close to the potential leak source as possible without removing  the
insulation.  If a leak is detected, repairs can  be scheduled within
15 days without having to delay repair until  the next turnaround.

     The EPA determined that it is cost effective to repair  insulated
valves that leak rather than delay repair until  the next unit  turnaround
(Docket Item IV-B-12).  Therefore, the standards for valves, including
insulated valves, require that these components  be monitored as close
to the source as practicable without requiring removal of the  insulation.
If a leak is detected, as with other valves,  repair of  insulated  valves
must be attempted as soon as possible within  15  days with a  first
attempt at repair within 5 days.

5.3  SELECTION OF AFFECTED FACILITIES

     Comment:  One commenter (II-B-23) suggested that the list of  VOC
fugitive emission source categories that are  excluded from  coverage by
the gas plants standards should also include  other VOC sources such as
VOL storage and benzene storage tanks.

     Response:  The purpose of the list was to exclude  sources from the
gas plants standards if they were also subject to other  standards  for
equipment leaks of VOC.  These standards include promulgated standards
for equipment leaks from petroleum refineries (Subpart  GGG), SOCHI
plants (Subpart VV), and benzene equipment leaks (Part  61,  Subpart J).
The volatile organic liquid (VOL) storage standard is based  on emissions
from process vents rather than equipment leaks.   As a result,  a  VOL
storage tank (such as an NGL receiver tank) at a natural gas processing
plant would be covered by Subpart K (VOL storage) for vented emissions
and by Subpart KKK for equipment leaks from valves, pumps,  compressors,
open-ended lines, and pressure relief devices.

     Comment:  One commenter (IV-D-20) took issue with  EPA's presumption
that a narrower designation of the affected facility was proper in deter-
mining regulatory applicability.  The commenter  claimed  that a narrow
definition may be proper at gas plants located in ozone  nonattainment
areas and subject to New Source Review, but is improper  when applied to
all gas plants.  The commenter stated that the application of  NSPS by
way of I3ACT requirements associated with PSD permitting  is  subject to
the PSD source definition and, therefore, requested that the narrower
designation of affected  facilities be removed from the  NSPS.  The
commenter specifically requested that each compressor train rather than
each individual compressor be defined as the affected facility, because
the enclosing of distance pieces would create potential  explosion
problems on certain affected units.  The commenter suggested that each
compressor train be considered as modified only if aggregate emissions
from that train increase.

-------
      Response:  The EPA uses the term "affected facility" in NSPS to
 designate either the individual piece of equipment or groups of
 equipment within a plant chosen as the "source" affected by a given
 standard.  The term "affected facility" is used in NSPS to effect the
 greatest emission reduction at reasonable costs and cost effectiveness.
 This  may result in a designation different from that used in PSD.  The
 designation of affected facilities is based on EPA's interpretation of
 Section 111 of the Clean Air Act and on the judicial construction of
 its meaning [ASARCO, Inc., v. EPA, 578 F. 2d 319 (D.C. Cir. 1978)].
 The designation of the affected facilities is fully explained in the
 preamble to the proposed standards (49 FR 2637-38).  Furthermore, since
 proposal, EPA has excluded certain compressors from coverage under the
 standards (e.g., wet gas reciprocating compressors) that are generally
 located in trains or groups.  The compressors that are still  covered by
 the promulgated standards (e.g., NGL compressors) are generally isolated
 and are not located in trains.  Therefore, EPA does not agree that a
 broader designation of compressors as an affected facility is appropriate.
 Instead, EPA has concluded that it is appropriate to designate each
 individual  compressor as an affected facility.

      Comment:  One commenter requested a clarification to Section
 60.632-l(f) of the proposed standards to protect owners of existing
 plants which have certain equipment designed for purposes other than
 control of emissions from arbitrary subjection to high emission control
 costs.  The commenter offered the following revision: (f)  Reciprocating
 compressors in wet gas service that are located at an onshore natural
 gas processing plant that does not have a control  device designed for_
 VOC emission control  present at the plant site are exempt from the
 compressor seal  control  requirements of Section 60.632-3.

      Response:  As provided in the promulgated standards, EPA has
 exempted all  reciprocating compressors in wet gas service.   Therefore,
 the presence or design of a control  device for these compressors is no
 longer relevant.

      Comment:   One commenter (IV-F-6) requested that the exemption for
 compressors that cannot  be economically or technologically retrofitted
 with  emission controls,  as discussed in the preamble to the proposed
 standards,  be incorporated as a specific paragraph in the standards.

      Response:  Because  compressors  are a separate affected facility,
 they would  be  exempt  from reconstruction if controls are technically or
economically infeasible  as specified  by the General  Provisions  (see
 Section 60.15(e)).   Therefore,  a  specific paragraph exempting  these
compressors is not necessary.   The determination of whether a
reconstruction or  replacement of  a  compressor is  technologically or
economically infeasible  will  be made  by the Administrator on  a
case-by-case  basis.   Unlike  the reconstruction provisions,  modification
does not consider  technical  or  economic  feasibility.
                                  5-11

-------
5.4 -APPLICABILITY DATE

5.4.1  Applicability Date of Standards

     Comment:  One commenter (IV-D-35) recommended  that  the  Administrator
publish in the Federal  Register notice of the retraction of  the  January  20,
1984, applicability date for purposes of the definition  of  a "new
source."  He stated that the history of other standards  which have  been
promulgated indicates that careful  review of comments  by the Agency can
result in substantial delay beyond  the 90-day period  imposed for promul-
gation by the Clean Air Act.  The commenter further wrote that such
action by the Administrator would be well within  his  inherent power and
in keeping with the rationale of the Act, Section 111, while failure to
take such action would  impose severe and unnecessary  burdens on  industry,
which the Clean Air Act intended to be mitigated.

     Another commenter  (IV-D-20) suggested that the promulgated  standards
be applicable to plants starting construction 180 days after promulgation
rather than on the date of proposal.  The commenter indicated that  plants
with existing construction permits, but not yet constructed, could  require
redesign or re-permitting.  The commenter indicated that EPA should
consider the potential  economic impacts of the earlier applicability date.

     Response: Proposal of the standards is legal notification that an
owner or operator of a  source will  be subject to  a  standard.  The  Clean
Air Act (Section lll(a)(2)) clearly states the Congressional intent
that the proposal  date  will be the  applicability  date.   Section  lll(a)(2)
defines the new sources subject to  an NSPS as those sources  built  or
modified after proposal, not after  promulgation or  some  later event.

     The proposal  date  is the applicability date  unless  the  promulgated
(or revised) standard is not based  on, and achievable  by, the same
technology specified at proposal.  This decision  is consistent with
EPA's practice of applying NSPS's to all sources  built after proposal
where the promulgated (or revised)  standards are  based on and achievable
by the same technology  as the proposed standards.

     Section lll(a)(2)  implements the basic Congressional objective of
preventing new pollution problems and improving air quality  as industry
changes by requiring all  new sources (built or modified  after proposal)
to use best demonstrated technology (BDT) in achieving  emission  reductions.
Congress recognized the existence of some uncertainty as to  the  final
standards that this technology must meet (S. Rep. No.  91-1196, 91st Cong.
2d Sess.  (Senate Bill, §113(b)(2)).  The passage of  time between
identification of BDT as the basis  of a standard  and  the final specifi-
cation of the performance required  of that technology,  therefore,  is not
relevant by itself.  Under Sections lll(b)(l)(B)  and  307(d)(10), an
NSPS should be promulgated within 6 months of proposal.   However,
Congress did not intend that if promulgation took longer, the category
of new sources should change.  Commonwealth of Pennsylvania  v. EPA, 618
F. 2d 991, 1000 (3rd Cir. 1980); 45 FR 8210, 8232 (February 6, 1980);
United States v. City of Painesville, 644 F. 2d 1186 (6th Cir. 1931),
cert. den. 102 S. Ct. 392 (1981); 49 FR 18076 (April  26, 1984).
                                  5-12

-------
5.4.2  Existing Sources

     Comment:  One commenter (IV-D-3) noted that the statement in
Section 2.3 of the BIO for the proposed standards indicates that
"Standards of performance must... (3) be applicable to existing sources
..." and asked if the proposed standards might be modified at a later
date to include existing plants.

     Response:  Section 2.3 of the BID for the proposed standards makes
it clear that the standards are applicable to existing sources that are
modified or reconstructed.  The terms "modification" and "reconstruction"
are defined in the General Provisions (see Sections 60.14 and 60.15,
respectively).  The definitions and their applicability to the standards
are discussed in Chapter 5 of the BID for the proposed standards.

     Comment:  One cornmenter (IV-D-20) proposed that all compressors
manufactured before the effective date of the standards be exempt
from the vapor recovery system requirements of the standards.  The
commenter noted that EPA stated in the preamble that compressors are
supplied with enclosed distance pieces and vented seals.  The commenter
disagreed with EPA's assumption that all  new compression capacity results
from installation of new units.  The cornmenter pointed out that many
newly installed compressors are existing units that have been rebuilt.
Hence, retrofitting controls onto these rebuilt compressors may result
in considerable delay and expense.

     Response:  The effective date of the standards is the date of
promulgation; however, all newly installed compressors at a plant,
rebuilt compressors and factory new compressors, are subject to the
standards if they are constructed after the applicability (proposal)
date of the standards (January 20, 1984), unless they are specifically
exempted (e.g., wet gas reciprocating compressors, or not "in VOC service"),
The EPA made no attempt to correlate the number of new compressors with
the amount of new gas to be produced.  New compressors will be well
controlled, and rebuilt compressors will  be controlled as well.
As stated in a previous response to comment, compressors would be
exempt from the reconstruction provisions of the General  Provisions if
controls are technically or economically infeasible (Section 60.15(e)
of the General Provisions).

     Furthermore, since proposal, EPA has decided to exempt all recipro-
cating  compressors in wet gas service.  Therefore, the only reciprocating
compressors covered by the standards are those in NGL service.

5.5  ALASKAN NORTH SLOPE

     Comment:   One commenter (iV-F-ld; IV-D-2H),  in both written
comments and public hearing testimony, requested  that natural  gas
processing plants located north of the Arctic  Circle be exempted from
the proposed standards.   The commenter based his  request on higher
control  costs  per unit of emission reduction,  as  follows:
                                  5-13

-------
     1.  The commenter cited the higher cost of operation in arctic
         regions.  Due to the extreme environment, processing facilities
         must be completely enclosed.  Since plants are enclosed;
         space is limited, and safety is of paramount importance.
         Consequently, any design changes or retrofitting in arctic
         facilities is extremely expensive.  Another important factor
         that increases the costs of operation in the Arctic is the
         high cost of labor.  The commenter indicated that a technician
         in the Arctic costs approximately $124 per hour in 1980 (based
         on $33.50 per hour wage plus 370 percent administrative and
         overhead) due to the requirement for extensive life support
         systems and area labor costs.

     2.  The commenter thought that the basic rationale for uniform
         national VOC standards, to prevent the creation of local
         pollution shelters, is inapplicable in the Arctic.  The degree
         of low solar insolation, the low concentration of photochemical
         precursors, and the cold ambient temperatures lead to a lack
         of photochemical ozone formation.   Additionally, since processing
         facilities are enclosed, sophisticated gas detection systems
         are utilized to ensure safety which would detect leaks of VOC
         from processing equipment.  Based  on these factors, the
         commenter concluded that the standards would provide no
         emission reduction.

     Response:  The presence of an in-place hydrocarbon gas detection
system does not necessarily ensure emission reductions.   Several  rnega-
grams of VOC could be released to the atmosphere annually without  the
use of specific control  techniques like those required by the standards.
The commenter did not demonstrate that their system resulted in at .
least equivalent emission reductions as the standards.  Based on EPA's
experience, gas detection systems alone are ineffective for reducing
equipment leaks of VOC.   Thus, EPA has  not  exempted process units  using
these systems from the standards.  The promulgated standards do, however,
allow an existing control program to be continued if EPA determines
that the program is at least equivalent to  the requirements of the
standards.  40 CFR 60.634

     The EPA has studied the commenter's concerns and acknowledges that
there are several unique aspects to the operation of natural gas processing
plants north of the Arctic Circle.  Because of the unique aspects  of
natural gas processing plants north of the  Arctic Circle, the increased
costs to perform routine leak detection and repair may result in an
unreasonable cost effectiveness.  These operations incur higher labor,
administrative, and support costs associated with leak detection and
repair programs because (1) they are located at great distances from
major population centers, (2) they must necessarily deal with the
long-term, extremely low temperatures of the Arctic, and, consequently,
(3) they must provide extraordinary services for plant personnel.
Therefore, EPA has decided that natural gas processing plants in the
North Slope of Alaska are exempt from the routine leak detection and
repair requirements of the standards.  This exemption does not include
the equipment requirements in the standards because installation of
equipment controls is common practice in the region.

                                  b-14

-------
 5.6   SMALL  PLANTS

 Comment:  Several commenters (iV-F-la, IV-D-15, IV-D-20;  IV-D-21, IV-D-23,
 IV-D-26,  IV-D-29, IV-D-30,  IV-D-31,  IV-D-36) requested that the small
 plant exemption size limit  be revised.  One commenter (IV-D-15) agreed
 with  the  small plant size limit of 10 MMscfd, but he, along with commenter
 IV-D-20,  stated that the plant size  exemption should be based on throughput
 instead of  capacity.  The commenter  indicated that the same staff and
 economic  considerations used by EPA  to determine the appropriate cutoff
 apply to  a  throughput-based cutoff.  Other commenters requested that the
 plant size  limit be raised.  Two commenters (IV-F-la and  IV-D-29)
 requested that the 10 MMscfd or less plant exemption include fractionating
 plants, and that the exemption be extended to somewhere between 35 to 50
 MMscfd for  nonfractionating plants.

      Other  commenters argued that the small plant size limit should be
 raised to 35 MMscfd.  Several  commenters (IV-D-21; IV-D-26; IV-D-36)
 provided  cost and emission  reduction data from a recently completed leak
 detection and repair program in support of a 35 MMscfd cutoff.  One
 commenter (IV-D-30) offered several  reasons for raising the small plant
 size  limit  to 35 MMscfd, including that EPA had overstated the potential
 emissions reduction, tha-t the cost to implement the standards was
 underestimated by EPA due to the ignoring of the substantial clerical
 requirements, and EPA ignored the remoteness and lack of  personnel  at
 small facilities.  Another commenter (IV-D-23) based his  request
 to raise  the small plant cutoff to 35 MMscfd (including fractionation)
 on EPA misconceptions of the average gas processing plant size.  The
 average plant size, he contends, is  biased by a few extremely large gas
 plants processing billions of standard cubic feet per day of natural  gas.

     Commenter IV-D-33 requested that EPA consider the extra travel
 time in getting to remote plant sites in setting the cutoff and the
 effect of extreme climatic conditions on travel  schedule.  In addition,
 the commenter stated that many gas plants with capacities of 10 MMscfd or
 less, whether they fractionate or not, are only j>artial ly attended  or
 unattended.   The commenter cited as examples two types of gas plants,
 straddle  plants and casinghead plants, that differ in their staffing
 requirements and profitability because of their different numbers of
 pieces of equipment and feed rates.  The commenter further stated that if
 his company hires an outside contractor, the recordkeeping and reporting
 requirements alone would be sufficient to warrant hiring a person at a
cost of about $36,000 per year for each central  office.  The commenter
concluded that there would be  an obvious negative impact on plant efforts
to improve manpower productivity.

     Response:   As presented in Appendix F of the BID for the proposed
standards, EPA carefully analyzed  the cost effectiveness of routine
leak detection and repair programs for small,  nonfractionating plants.
The cost  effectiveness  values  derived were presented  graphically on
page F-7  and are reproduced in Figure 5-1  on the next page.   The graph
in Figure 5-1 shows  that the cost  per megagrarn VOC reduced begins to
increase  rapidly as the plant  throughput falls below 10 MMscfd.   There-
fore, 10  MMscfd represents  an  appropriate  cutoff,  and nonfractionating
plants processing more  than 10 MMscfd are  subject  to  the monitoring
requirements of the  standards.

                                  5-15

-------
   10-
   B —
8 6-\
o
3  4-J
>
8
   0 —
                 Appendix F of BID for proposed standards
                                r        r

                       10        IS       20

                        PLANT  SI2E (MHscfd)  -
I

25
30
35
45
                                           so
           Figure  5-1.   Cost Effectiveness versus  Plant  Size for Small  Plants
                                               5-16

-------
      Figure 5-1 also shows that the basis  for the siilall plant leak
 detection and repair programs has been changed slightly since proposal.
 The  primary revision to the cost basis was  in changing the instrument
 cost.  The analysis presented in Appendix  F of the BID for the proposed
 standards was based on the assumption that  each plant would maintain
 two  monitoring instruments.  However, if the monitoring is to be performed
 by contractors or central office personnel, it is likely that the
 instrument cost would be distributed among  several plants.  The revised
 cost is based on the assumption that the contractor or central office
 personnel would maintain two instruments but would use these instruments
 among five plants.  Complete details of the revised costs are available
 in Docket Item IV-B-6.

      The cost effectiveness values presented in Figure 5-1 are based on
 the  cost of performing the required leak detection and repair program
 using either contractor or central office  personnel.  These costs
 include travel costs, as well as additional rnanhours for travel  time.
 Since fractionating plants are more complex than nonfractionating
 plants, both physically and operationally,  plants having fractionation
 trains are likely to have full-time operating and maintenance personnel
 available to perform the monitoring program "in house."  The costs of
 such  an "in house" program would be greatly reduced from a program
 requiring expenses for outside personnel  and travel; consequently, an
 "in-house" program would result in a much  better cost effectiveness.
 Fractionating plants are subject, therefore, to the requirements for
 routine monitoring regardless of capacity.

      Some commenters requested that the small  plant exclusion be
 based on throughput rather than capacity.  Although the economics
 (as  well as the emissions) of plant operation is highly dependent on
 the  process throughput, capacity is the appropriate defining factor for
 excluding small plants.  Throughput may vary from day to day; therefore,
 a small  plant exclusion based on throughput would be unmanageable.  An
 affected facility within the plant may be covered by the standards
 while operating under one capacity, but as the throughput decreases,
 the  facility may be exempt.   This system would require extensive
 recordkeeping and would be difficult to enforce.   Since the plant
 capacity remains  constant, and the plant throughput is limited by the
 capacity, the exclusion is based on plant capacity.

     Comment:  One commenter (IV-D-35) recommended that all  intermit-
 tently attended gas plants be exempted from the proposed standards.  The
 commenter claimed that in both casually attended and partially attended
 plant situations, the proposed standards  could force premature
 abandonment of many remote or marginal  wells,  while plants manned 24
 hours per day would, in all  likelihood, have sufficient personnel  to
 perform the monitoring, recordkeeping, and reporting required by the
 standards.

     Response:   The standards have an  insignificant effect on the
 abandonment  of wells because  the costs  of compliance with  the standards
are very small  compared to the operating  costs of a plant  and the
 revenues generated  by a plant.   The LPA agrees with the commenter that
there are  some  relatively small  plants  that operate without  technically
trained  personnel  being present  because  of the type of process  that is

                                  5-17

-------
performed there.   The EPA was aware of such plants and  considered  the
burden imposed upon small  plants  subject to the standards.   At proposal,
EPA exempted small  nonfractionating plants (capacity of 10  MMscfd  or
less) from the standards.   Assuming that nonfractionating plants  would
hire contractors  or use central  office personnel  to perform the leak
detection and repair program, EPA determined that it would  not be  cost
effective to require these plants with 10 MMscfd  or less capacity  to
perform the routine leak detection and repair requirements.  The  LPA
believes, however,  that fractionating plants require the presence  of
technically trained personnel  having the ability  to carry out responsibly
a leak detection  and repair program.
                                  5-18

-------
                        6.0  ENVIRONMENTAL IMPACTS
6.1  EMISSION REDUCTIONS

     Comment:  One commenter (IV-D-20) did not agree with the EPA
estimates of the environmental  benefit from implementation of the
standards.  The commenter wrote that EPA's assumptions regarding typical
plant size, inventories of valves, and the number of plants to be built
in the future overestimate emission reductions.  The commenter requested
that EPA re-evaluate the estimates.

     The commenter noted that the emission reduction estimates are based
on Model Plant 13 and argued that Model Plant B is not reflective of
future gas plant sizes and designs.  The commenter stated that most new
gas plants will be cryogenic plants due to operational and economic
reasons, and cryogenic plants have at least 50 percent fewer valves than
EPA estimated for a "typical" Model Plant (3.

     The commenter also took issue with EPA's average emission factor
of 6.3 kg/day for compressors,  stating that since few new plants will
have compressors in NGL service, the emission factor of 1.9 kg/day
given by EPA for wet gas compressors should be used for all compressor
calculations.  Another commenter (IV-D-35) indicated that the Rockwell
study (Docket Item II-I-20) showed the maximum compressor VOC emission
factor is 1.5 Ib/day (0.68 kg/day).

     The commenter recalculated the emission reductions and estimated
that the standards would control only 5,159 Mg/yr based on his assumptions
that (1) Model  Plant B contains 250 valves, (2) 1983 plant construction
is not included, (3) compressor emission factor for Model Plants A, B, and
C is 1.90 kg/day, and (4) emission reductions from capping open-ended lines
are not included.  The commenter also presented valve counts and plant
capacity for four gas processing plants to support his argument that
new gas plants  would have fewer valves than EPA estimates.  Based on data
from 1982-1983  Hydrocarbon Processing Construction, the commenter stated
that more than  60 percent of the new plants constructed were 25 MMscfd
or less in contrast to EPA's estimated average capacity of 90 MMscfd.

     Response:   The EPA provides nationwide environmental impacts
to illustrate the potential emission reduction associated with the
standards on emission levels anticipated in the future.  Promulgation
of the standards, however, is based on the availability of demonstrated
cost-effective  control  techniques.  Even if the nationwide impacts
presented in the preamble to the proposed standards were overstated,
the point is that significant emissions are attributable to natural
gas processing  plants and that  growth is predicted in the industry.
In addition, cost-effective controls are available to reduce these
emissions significantly. Therefore, new source performance standards
should be promulgated for the industry.

     The EPA noted several  misconceptions in the conlinenter's analysis of
nationwide impacts.  Primarily, these faults are as follows:
                                  6-1

-------
     •  The commenter suggested  using  a  compressor  emission  factor  of
        1.9 kg/day for all  compressors,  based  on  emissions  from  wet gas
        compressors,  because  few plants  would  have  compressors  in NGL
        service.   Test data used by  EPA  in  developing  the  proposed
        standards  (as given in  Appendix  G of the  BID  for the  proposed
        standards) showed  that  34 percent of all  compressors  were in
        NGL service.   Thus, the  basis  for using 34  percent  NGL  service
        compressors was actual  test  data rather than  assumption.

     •  The commenter claimed that emission reductions  from  open-ended
        lines should  not be considered a result of  the  NSPS,  since  lines
        in his plants are  capped already.   As  evidenced  by  EPA  plant
        visits and other comments pointing  out difficulties  in  capping
        open-ended lines,  most  gas plants do not  cap  or  plug  open-ended
        lines.  Therefore,  it is unrealistic to exclude  open-ended
        line emissions from the  environmental  impacts  of the  standards.

     •  The commenter based his  emission reduction  impacts  on 250 valves
        in Model  Plant B.   The  EPA recognizes  that  different  designs for
        the same  plant process  will  utilize widely  variable  numbers
        of components due  to  the level of control desired  and the
        economics  of  construction.  The  valve  counts  used  in  the BID
        for the proposed standards were  based  on  actual  operating
        plants and represent average values.   Since Model  Plant  B is
        intended  to represent a  relatively  complex  plant,  such  as a
        fractionating cryogenic  plant  or a  nonfractionating  refrigerated
        absorption plant,  750 valves is  an  appropriate  count.   Test
        data supplied by the Gas Processors Association  for  the  Conoco
        Cashion plant showed 540 components (mostly valves)  for  a
        30 MMscfd  nonfractionating cryogenic plant, which  would  be  a
        Model Plant A.

     •  The commenter disagreed  with the inclusion  of 1983  emission
        reductions.  The emission reductions are  presented  to show  a
        fifth-year impact, rather than the  impacts  by given  calendar
        years.  Although 1983 emission reductions will  not  occur, a new
        5-year impact analysis  would include 1987 reductions  with
        the same  results.   The  projected emissions  and  emission
        reductions from baseline levels  of  control  through  1987  are
        presented  in  Table 1-1.

     Comment:  Another commenter (IV-D-22)  believed that the emission
reduction estimates were overstated.  The commenter based  his argument
on the premise that most of the VOC  emission reduction resulting from
the standards will be achieved  during  the  initial testing  and that
subsequent monthly tests are not justified. The  commenter noted that
initial  testing at a  gas plant  found many leaking valves (over
100,000 ppm).  The second test  at this plant  found  19 leaks out of  508
components.  Of these, more than 50  percent had  readings less than
50,000 ppm arid only 1 greater than 100,000  ppm.   Similarly, initial
testing at another plant found  many  leaks  with a  soap score over 4,
while a second test 19 months later  found  only two  leaks with a soap
score of 4.
                                  6-2

-------
     The commenter presented a comparison between soap solution test
results and results obtained using a Foxboro OVA-108 monitoring
instrument as his method of establishing a relationship between measured
VOC concentrations and leak rate.   The commenter estimated that a
monthly program would achieve an incremental reduction of 1,709 kg/yr
more than an annual monitoring program.  Therefore, the commenter
concluded that annual monitoring of all components would serve both
EPA's goal of establishing standards that will  improve and maintain the
environment and industry's goal  of supplying the public with competitively
priced products.

     Response:  The commenter contends that, although the leak frequencies
used by EPA may be correct, new leaks occurring or recurring during the
period between tests will be smaller than leaks found during initial
testing.  The assumption that more recent leaks will  be smaller is
true, as evidenced by the data presented by the commenter.  However,
the leak detection and repair programs are not  intended as a means of
repairing massive leaks, but rather as a means  of preventing such
leaks.  The leaks discovered in each followup inspection are likely to
be smaller than those found in the initial inspection which have developed
over an extended period.  This fact is the reason for shortening the
monitoring period to monthly, as significant emissions can be prevented
from occurring.  If the leak detection and repair program were not
performed, the leaks left undiscovered and unrepaired would eventually
return to the magnitude found during the initial inspection.  The
commenter is partially correct in stating that  most of the emission
reductions occur in the initial  monitoring period.  In fact, the ongoing
program provides the work practice to prevent these emissions from
occurring again.

     In response to the commenter's claims of low incremental emission
reduction between monthly and annual monitoring, the standards allow
special provisions to prevent the unwarranted monitoring of nonleaking
sources.  These provisions include monthly/quarterly monitoring for all
plants and skip-period monitoring, or alternative standards, for low
leak rate plants, which are discussed in detail  in Section 6.2.

     Comment:  One commenter (II-D-10) noted that there is no scientific
basis for assuming a "significant emissions reduction," since there is
no definitive quantitative study demonstrating  fugitive emission
reduction from an inspection and maintenance program for refineries or
gas plants.  Furthermore, the commenter added that EPA had no technical
basis for the transfer of chemical  plant or refinery VOC emissions
data to gas plants.

     Response:  This comment was made in Hay 1981 by API and is included
at their request (IV-F-le).  Since that time, several  leak detection
and repair programs have been conducted by both EPA and industry.
To assess the potential  effects  of leak detection and repair programs,
EPA conducted emissions testing  at gas plants and related facilities.
The results of the EPA gas plant emissions tests, combined with the
results of the API/Rockwell study, formed the basis for the gas plant
emission factors (Docket Item II-A-23) and subsequent emission estimates.
To assess the control technique  (leak detection and repair) for valve,

                                   6-3

-------
pump, and relief device emissions,  EPA has performed  a  maintenance
effectiveness study (Docket Item II-A-11)  and has  examined  the  results
of on-yoing State and local  equipment leak rules.   These studies  all
support the effectiveness of leak detection and  repair  for  reducing
emissions.  Although most of the studies  were performed for refineries,
the components regulated in gas  plants are identical  or very similar,
and the same results are expected.   Therefore, transfer of  refinery/
chemical  data to gas plants is reasonable.

6.2  LEAK FREQUENCIES

     Comment:  One commenter (IV-F-2) submitted  in-plant leak screening
test data for a 20 MMscfd fractionating plant, a plant  the  commenter
considered as typical of current industry  design and  construction.   The
first test was conducted on 4,195 components, including flanged and
screwed connections.  Using the  API Soap  Score Method and recording
scores of 3 or greater as a leak, 61 components  or 1.45 percent were
found leaking.  A second test was performed 18 months later on  722
components covered by the proposed  standards (omitting  flanged  and
screwed connections) using API soap and instrument methods.  Each test
method found 5 leaking components,  all of  which  were  valves corre-
sponding  to a 0.68 percent leak  rate.  The commenter  also provided  the
total time taken to perform monitoring in  each of  these tests.

     Based on these data, the commenter claimed  EPA had both over-
estimated the emission reductions for the  standards and underestimated
the control cost.

     Response:  The EPA recognizes  that some plants will have a low
rate of leaks, either due to good maintenance practices or  other  factors,
and also  recognizes that frequent leak monitoring  for such  low  leak
facilities is not cost effective.  Consequently, EPA provided standards
for valves in which nonleaking valves may  be monitored  less frequently.
For example, in the commenter's  case, the  second test resulted  in only
5 leaks out of 722 components.  If a test the next month gave the same
results,  the nonleaking valves,  which are  likely to be  most of  the  717
nonleaking sources, would then be monitored quarterly.   The plant would
also be a likely candidate for the "allowable percentage of valves
leaking"  alternative provided in Section  60.633-1  of the proposed
standards.  After the initial performance  test,  the facility could
elect to  use annual perfonnance tests to  demonstrate that less  than
2 percent of all valves were leaking.  Otherwise,  the plant could use
the skip-period monitoring program, as specified in Section 60.633-2,
to reduce the monitoring frequency for valves to semiannually or
annually.  The standards provide enough flexibility to  increase cost
effectiveness and avoid the unnecessary monitoring of nonleakers.

     The commenter  indicated that the total monitoring  time for 722
components using the hydrocarbon monitoring instrument  was  14.4 hours.
Assuming  a two-man crew performed the monitoring,  this  equates  to
approximately 2.4 man-minutes per source.   As shown in  Table 8-3  of  the
BID for the proposed standards, Model Plant B has 774 components, which
were estimated to require 121 labor hours per year on a quarterly
basis, or 30.25 man-hours per monitoring  cycle.   This results in  an

                                  6-4

-------
 average monitoring time of 2.34 man-minutes per component.  These two
 values are nearly identical, so that the commenter's monitoring time
 supports EPA's estimated monitoring time.

 6.3  EMISSION FACTORS

      Comment:  API (II-D-10) stated that EPA's assertion that the
 Rockwell compressor seal factor "probably underestimates emissions" is
 speculation and is untrue, according to the Rockwell research team
 leader.

      Response:  The Rockwell compressor emission factor is based on EPA
 and API testing of emissions into the distance piece area from only
 open frame compressors.  The factor does not include emissions into the
 seal packing vent or into enclosed distance pieces.   Therefore, the
 Rockwell emission factor is probably understated substantially.  A
 revised emission factor, discussed in the BID  for the proposed standards
 (Appendix G), was necessary because a review of the  data used  in developing
 the 1  kg/day emission factor showed that a large portion of  the data
 were from residue (dry) gas compressors, which EPA intended  to exempt.
 Therefore, the overall  emission factor of 1 kg/day was  not representative
 of the population of compressors  to be regulated.  The  revised emission
 factor (6.4 kg/day)  is  based on a weighted average of emission factors
 for compressors in both wet gas and NGL service because there  is a
 large  difference in  process stream VOC concentration between  wet gas
 and NGL service compressors.

     Comment:   Two comrnenters  (IV-D-13 and IV-D-19)  stated that EPA
 had overestimated the number of centrifugal  compressors in assuming
 that 50 percent of all  compressors would be centrifugal.   One  commenter
 wrote  that  about  10  percent are centrifugal  and 90 percent are
 reciprocating.   Commenter IV-D-13 stated  that,  because  of  this  assumption,
 EPA overestimated compressor VOC  emissions,  which  are minor compared to
 total  VOC  emissions.  The commenter suggested  that the  compressor
 requirements  be  eliminated.

     Response:   In all  analyses of compressor  controls,  EPA used an
 average  emission  factor  for  all types  of  compressors, including
 centrifugal  and  reciprocating ones.   Emissions  are related to  the
 compressor  service (NGL  and  wet gas),  but  not  to the  type  or operation
 of  compressors  (centrifugal  and reciprocating).  Therefore, the  fraction
 of  compressors assumed  to  be centrifugal  has no  bearing on the compressor
 emissions  estimate.   At  proposal,  EPA  assumed  that 50 percent of all
 compressors would  be  centrifugal  only  in  order  to  determine an average
 cost for compressor control systems.

     Comment:  One commenter (IV-F-lb) stated that nonfractionating
 plants will not use compressors in  NGL service  (unless  a propane
 refrigerant system is required  to  optimize plant recoveries).   Therefore,
 the commenter stated  it would be more appropriate to base compressor
 seal leakage on wet gas  compressors only.

     Response:  As used  in the HID for the proposed standards, as well
as the  proposed standards, NGL service streams  are considered  to be

                                  6-b

-------
a1!  hydrocarbon streams that contain greater than 50 weight percent
VOC.  Since the purpose of the natural  gas plant is to extract hydro-
carbons from the methane residue gas, all  natural  gas plants have NGL
streams.  Based on the content and physical  characteristics of the NGL
produced, the NGL stream may be moved by either pumps (for low vapor
pressure products) or compressors (for high  vapor pressure products).
For nonfractionating plants, the NGL stream  produced is typically
either ethane and heavier hydrocarbons or propane and heavier hydro-
carbons.  Either of these mixtures would likely be high vapor pressure
streams and would be transported from the process with compressors.
Therefore, EPA does not necessarily agree that nonfractionating plants
will not employ NGL compressors.  Since an exemption for all wet gas
reciprocating compressors has been added to  the promulgated standards,
NGL compressors represent an even larger portion of the affected
compressors than assumed at proposal.

     Comment:  Three commenters (IV-F-le; IV-D-31; IV-D-34) urged
that compressors be exempt from the standards because of the small
amount of reactive hydrocarbons emitted from compressors.   Comrnenter
IV-F-le noted that compressors constitute less than 4 percent of
the total gas plant emission inventory.

     Response:  Compressor seal emissions were estimated in Appendix G
of the BID for the proposed standards to be  2.3 Mg/yr per seal, the
highest emission factor of all VOC emission  sources in natural gas
plants.  This average emission value was based on wet gas compressors
having emissions of 0.7 Mg/yr, and NGL compressors having emissions of
5.5 Mg/yr.

     As shown in Table H-l of the BID for the proposed standards,
compressors represent 17.8 percent of the emission reduction from
implementing the standards for a typical new plant.  Contrary to  the
commenter's statement that compressors represent 4 percent of the  gas
plant  emission inventory, the Model  Plant C compressor emissions  (37.8
Kg/day) are 13.8 percent of the total uncontrolled Model Plant B
emissions.

     Comment:  One cornmenter  (IV-D-33) saw no justification  for the
baseline emission  factors developed  in Radian's July  1982 report  (Docket
Item II-A-23).  The commenter claimed, as an example, that  the VOC
emission factor for valves  (0.18 kg/day) was  too high and should  be
0.09 kg/day based  on  Table  1-2 of the Radian  report and on  the 95  percent
upper  bound curve.  The commenter determined that  leak occurrence  rates
of  18  percent at  100,000 ppmv  (0.3b  kg/day) and 82 percent  at 500  ppmv
(0.03  kg/day) resulted  in an  emission  factor of 0.09  ky/day.

     Response:  The commenter contends  that,  based on the data  presented
in  the EPA  report  "Frequency  of  Leak  Occurrence and  Emission  Factors  for
Natural  Gas  Liquid  Plants"  (80-FOL-l),  the  average emission  factor for
valves should  be  0.09 kg/day,  rather than 0.18  kg/day used  by EPA.
 It  is  true  that an  average  leak  rate  calculated  from  the commenter's
values of  500  ppmv  and  100,000  ppmv  would be  0.09  ky/day.   However,  the
arithmetic  average  of these  values  does not  represent a true  average
emission  factor  for  all  valves  due  to  the log  normal  distribution of

                                   6-6

-------
leak rates.  Chapter 4 and Appendix C of the referenced document describe
in detail the statistical techniques used by EPA in developing the
final emission factor for valves (0.18 kg/day), which is presented in
Table C-2 of the document.

6.4  MODEL PLANTS

     Comment:  One cornmenter (II-D-10, II-D-30) questioned the basis for
the model plant analysis.  The cornmenter questioned the use of components
per vessel in developing model plants since different vessels have
different functions and require different controls.  In addition, the
cornmenter said the use of industry-tested plants to determine the
ratios of numbers of vessels is unrealistic because the API/Rockwell
study did not attempt to assess emissions from typical  gas plants.  The
Rockwell study, according to the cornmenter, was designed to assess
emissions from typical components in gas plants; and Rockwell determined
the number and variety of components after a preliminary site visit to
the facilities.

     Response:  Since this comment was submitted in Hay 1981, EPA has
incorporated data from test programs at two additional  plants into the
analysis.  The model plants used in the analysis represent different
levels of process complexity based on numbers of components in the
process rather than on the function of the vessel.   Vessels do vary in
terms of function and control  required, but the function or process
method (e.g., refrigerated absorption, cryogenic) is not as appropriate
in describing plant complexity as are numbers of components.  Appendix B
of the CTG for the natural gas production industry discusses the basis
for the analysis, including a  description of the ratio  calculations and
derivation of the component counts.

     It should be noted that the cost and emission reductions on a per
piece of equipment basis for equipment leak prevention  programs, with
the exception of compressor controls, is relatively independent of
component count.  The regulatory alternatives presented in Chapters 7
through 9 of the BID for the proposed standards were used primarily to
illustrate the impacts of various available control strategies on typical
plants.  The standards do not  correspond with any one of the available
"regulatory alternatives," but are based on individually selected
control techniques for each component class.  For this  reason, prior to
proposing the standards, EPA performed a second cost analysis (Appendix H
of the BID for the proposed standards) based on different control techniques
for each component type rather than for model plants.  This single component
control cost analysis served as the basis for the standards rather than the
model plants, which are used solely to estimate nationwide impacts.

6.5  NONAIR ENVIRONMENTAL IMPACTS

     Comment:  Two cornmenters  (IV-F-lb and IV-D-15) disagreed with
EPA's claim of a positive impact on water quality from  the standards.
The first cornmenter stated that most plant operators are required to
formulate and implement spill  prevention control  systems and measures.
He also noted that any VOC spills prevented would be likely to evaporate
before causing an impact on water quality.   The second  commenter indicated

                                  6-7

-------
that liquid leaks that may affect water treatment systems  are not  the
principal  focus of the proposed standards.

     Response:   The EPA realizes that most  gas plant streams  consist of
high vapor pressure components that, as indicated by the commenter,
will evaporate  before entering drainage systems.   However, some  yas
plants do  produce light liquid products.   Leaks  from the sources
handling light  liquid products could be reasonably expected to enter
plant drainage  systems.  Although water quality  impacts  are probably
very minor, the impacts of the promulgated  standards are positive.

     The Clean  Air Act directs EPA to consider the environmental impacts
that result from promulgation of new source performance  standards.

     Section 111 states that:

     ... a standard of performance shall  reflect  the degree of emission
     limitation and the percentage reduction achievable  through  application
     of the best technological system of continuous emission  reduction which
     (taking into consideration the cost of achieving such emission  energy
     reduction, any nonair quality health and environmental impact  and
     energy requirements) the Administrator determines has been  adequately
     demonstrated.

Hence, it  is appropriate for EPA to have considered the  water quality
impacts of the  standards.
                                  6-8

-------
                           7.0  COST OF CONTROL

7.1  GENERAL

     Comment:  One commenter (IV-D-3) claimed that all  of the capital
costs presented in Table 8-1 of the BID for the proposed standards
are low.  As an example, the commenter presented costs  for compressor
distance piece flare piping.  The commenter argued that very few com-
pressors in gas processing plants have only one cylinder and that most
compressors have four cylinders.  The commenter reasoned, therefore,
that the capital cost of compressor distance piece piping ($2,451)
should be increased to $9,804 per compressor.  The commenter maintained
that the same reasoning applies to the seal piping, gas supply system,
and the rest of the cost data.

     Response:  The capital  costs for control equipment presented in
Table 8-1 of the BID for the proposed standards were all obtained from
vendors or industry sources.  In developing the capital costs, EPA
attempted to generate cost estimates for control systems that would be
representative of the "worst-case" cost situations (for example, offset
mounting costs were included in the rupture disk installation cost
estimate).  Consequently, other commenters believed the average actual
control costs should not exceed those- in the BID for the proposed
standards.

     The commenter indicated that the capital costs for control of
compressor distance piece piping would be low by a factor of four if
the plant had four-cylinder compressors.  The commenter is correct in
realizing that, since the items included in the "distance piece piping"
cost must be installed on each distance piece, a four-cylinder
"compressor" would actually cost four times as much.  However, the
capital costs in the BID for the proposed standards are presented
on an "emission source" basis, which, for the case of compressors, is
each compressor seal.  Since the four-cylinder compressors would have
four controlled seals, the emission reductions would also be quadrupled,
with no change in the control cost effectiveness.

     Comment:  One commenter (IV-D-3) indicated that the energy savings
or recovery credits for hydrocarbons are very misleading.  The commenter
indicated that, according to Table 7-5 of the BID for the proposed
standards, the equivalent of 750,000 barrels of oil  over 5 years would
would be recovered under Regulatory Alternative II.   The commenter
stated that this recovery would barely cover the cost of the necessary
additional compression and operational  costs to recover the hydrocarbons.

     Response:  The data presented in Table 7-5 of the  BID for the
proposed standards indicate  the energy impacts of the NSPS.   These data
are not intended to illustrate the economic impacts of  the energy
supplies recovered, but are  intended to show the impacts of the standards
on the national  energy reserves.

     For standards requiring add-on control devices for vented process
streams (such as condensers  or adsorbers), energy expenditures for
operation of the control  equipment would be included in the analysis of

                                  7-1

-------
the regulatory impacts.  For example, a  condenser system for control  of
a VOC vent stream, while recovering energy potential,  would  expend
energy for operating recovery pumps, the refrigeration system,  etc.
For the equipment leaks NSPS, however, no additional  equipment  is
required in order to prevent losses of hydrocarbons  from valves,  relief
devices, pumps, or open-ended lines.  The energy values "recovered"  by
preventing leak losses from these sources, therefore,  are not offset by
operational energy requirements.

     EPA concurs with the cormnenter that the energy  values saved  do  not
completely offset the monetary cost of performing the  programs  required
by the standards.  However, the net cost effectiveness, as presented in
Appendix H of the BID for the proposed standards, is  reasonable.

     Comment:  One commenter (II-D-10) stated that in  using  recovery
credits for demonstrating cost effectiveness, it is  misleading  to  include
quarterly or monthly monitoring results  with total  recovery  on  an  annual
basis.  The commenter further stated that credits gained after  the  initial
baseline screening and repair are relatively small  and are only 10  percent
per quarter or 5 percent per month (according to EPA estimates  of  percent
leaks initially detected for quarterly and monthly monitoring).

     Response:  API submitted this comment prior to  proposal and  requested
that it be included in the rulemaking (IV-F-le).  API  contends  that,
since fewer leaks are found during follow-up inspections than during
the initial inspections, the recovery credits should be based on  the
quantity of emissions reduced during a follow-up test, rather than
the reductions from the initial inspection.

     As explained in Section 6.1, the leak detection and repair programs
required by the standards are intended to prevent leakage from  occurring.
The initial inspection of a facility demonstrates the  emission  rate
from the facility in the absence of a routine leak and repair program.

     A leak detection and repair program providing nearly 100 percent
control would have almost no leaks detected during follow-up inspections
and, therefore, would provide an emission reduction  equal to the  uncon-
trolled emissions on a continuous basi.s.  The EPA considers  this  reduction
from the uncontrolled leak rates to be the proper basis for  the calculation
of recovery credits.

     Comment:  One commenter (II-D-30) remarked that "front-end costs"
should not be equated with "capital costs."  For example, double  valving
open-ended lines and initial leak repair are front-end expenses that
should be considered as operating costs.  The commenter added that only
the VOC analyzer and the compressor piping should be defined as capital
costs and that remaining costs incurred are classified as expenses.
According to  the commenter, expense and capital costs cannot be combined
without using an amortization schedule.

     Response:  API submitted this comment prior to proposal and  requested
that it be included in the rulemaking (IV-F-le).  Although the control
cost of open-ended lines and initial  leak repair could be treated as
operating expenses, they are treated  as though they were capital  costs

                                  7-2

-------
and are amortized for the purposes of these standards, because they are
one-time, start-up costs.  This assumes capital would be borrowed
to pay these initial costs.

     Comment:  One cornrnenter (IV-D-30), stating that most new, small
plants are designed to operate unattended, claimed that the costs
associated with repair have been grossly underestimated in the BID for
the proposed standards.  Similarly, another cornmenter (IV-D-29) stated
that the effect of the proposed standards on the small, partially-
attended plants is not cost effective.  The commenter quoted an annual
monitoring cost of $39,191.52 compared to EPA's estimate of $970 for
the direct monitoring cost for Model Plant A and indicated that EPA
failed to include costs for training labor, scheduling, preparation,
travel, direct management, supply personnel, reporting, recordkeeping,
setting up a system of compliance, tagging, and problem-solving.   The
commenter said that, given that gas processing plants do not contribute
significantly to air pollution, these anticipated related costs are
disproportionate to the marginal  benefit derived and, hence, are  not
cost effective.

     Response:  The commenters contend that, since small plants are
minimally attended, the costs of the standards presented by EPA are
unrealistic.  The cost analyses presented in Chapter 8 of the BID for
the proposed standards were based on leak detection and repair programs
being performed by plant personnel.  The EPA recognized, prior to
proposing the standards, that small gas plants may not have adequate
staffing to perform the leak detection and repair program.  Therefore,
based on industry supplied data, EPA performed a separate cost analysis
for small, unattended plants based on leak detection and repair programs
being performed by corporate office personnel.  These analyses were
presented in Appendix F of the  BID for the proposed standards and
showed that the cost effectiveness of monitoring programs deteriorated
rapidly for plants smaller than 10 MMscfd.  Consequently, plants  with
capacities of 10 MMscfd and less were exempted from routine leak  detection
and repair programs at proposal.

     The commenters mentioned numerous administrative items not
specifically included in the EPA cost analyses.  In preparing cost
analyses for regulatory alternatives, EPA includes a percentage add-on
to all  control  labor costs to account for administrative costs.  Although
EPA added 40 percent of the direct labor cost  for general  and adminis-
trative costs in analyzing the cost of the standards, the Agency  believes
that once the program is established, administrative costs will be much
1ower.

7.2  COMPRESSORS

     Comment:  One commenter (IV-D-21)  thought that the costs to  control
emissions from compressor seals  are unreasonable, considering the small
return  (emissions  reduction)  from emission control  and, therefore,
recommended  that compressor seal  control  system requirements  be deleted.
The commenter also argued that the costs  of controlling compressors are
particularly high, considering the nature  of the  emission  sources which
include several  small  compressors.   The  commenter also  remarked that

                                  7-3

-------
the burden of compliance will  fall  on existing  as well  as  new -jar;
plants since replacement of a  single compressor will  result in dn
affected facility.   Existing plants with some portion of a control
system, such as a flare, will  require new systems since the existing
equipment is not designed for  VOC emission control.   Thus, the marginal
operating economics of these facilities makes the costs unreasonably high.

     Response:  The commenter  contends that all compressors should  be
excluded from the promulgated  standards because the  emission reductions
achieved do not justify the control cost.  The  selection of the require-
ments of the standards was based on control  cost effectiveness, that
is, the cost per unit mass of  emission reduction.  At proposal, the
control cost effectiveness for compressors was  based  on a  weighted
average emission factor assuming 66 percent of  all  compressors are  in
wet gas service and 34 percent of all  compressors are in natural  gas
liquids (NGL) service.  These  weighting factors were  based on component
inventories observed during testing at operating gas  plants.  Given
compressor emission factors of 0.7 Mg/yr and 5.5 Mg/yr  for wet gas  and
NGL service, respectively, a weighted average emission  factor of
2.3 Mg/yr was used  in analyzing the compressor  control  cost effectiveness
(Docket Item II-B-35).

     Since control  costs for reciprocating compressors  were higher  than
those for centrifugal  compressors, EPA averaged the  costs  for the  two
compressor types to obtain an  average compressor control cost of $6,400/yr
based on one-half of all compressors requiring  an add-on control  device
as explained in Appendix H of  the BID for the proposed  standards.   Using
these cost and emission reduction values, the cost effectiveness of compressor
seal  vent control systems was  $460/Mg, which EPA considered reasonable.

     Since proposal, as discussed in the response to  the next comment,
EPA decided to exempt all wet  gas reciprocating compressors from the
promulgated standards.  As shown in Table 7-1,  the control cost-
effectiveness values of all other compressors are reasonable.  Therefore,
all compressors except wet gas reciprocating compressors are covered by
the promulgated standards.

     Comment:  Several cornmenters (II-D-30; IV-D-13;  IV-D-23; IV-D-26;
IV-D-29; IV-D-33; IV-F-la; IV-F-lb) indicated that EPA had underestimated
the average cost of compressor seal vent control systems by assuming
that 50 percent of the compressors used will be of the  centrifugal  type.
One of these commenters (IV-D-26) provided a recent industry survey
showing that out of over 3,000 compressors, centrifugal compressors are
utilized in only 9.8 percent of the applications of the companies
surveyed.  Another commenter (IV-F-lb) stated that small plants will
almost exclusively utilize reciprocating compressors  for technical  and
economic reasons.  One commenter (IV-D-33) estimated  that the cost  for
a  reciprocating compressor seal vent system would be about $28,000
(including installation, foundation, and flare support).  When factored
into EPA's revised cost analysis (Docket Item  II-B-37 and BID Appendix
G), the commenter felt his estimate will justify the  exemptions for
reciprocating compressors  in NGL or 100 percent VOC service when no
control device is present.  Another commenter  (IV-D-34) stated that
control costs were extremely high and unjustified and recommended  that
all compressors be exempt  from the standards.
                                  7-4

-------
                                TABLE 7-1

                  COST EFFECTIVENESS OF COMPRESSOR VENT
                    CONTROL SYSTEMS FOR MODEL PLANT B*

                            Control Device
     Compressor Type	Present	Cost Effectiveness  ($/Mg)

 Centrifugal  - Wet Gas           No                      710
                                 Yes                     280

 Centrifugal  - NGL               No                       91
                                 Yes                      36

 Reciprocating - NGL             No                      280
                                 Yes                     200
 *From the BID for the proposed  standards,  Appendix  G,  Table  1,  page  G-14.


      These comrnenters recommended  that compressor requirements  be
 eliminated from  the  standards,  since  a cost  analysis based on  recipro-
 cating compressors only  would show the cost  effectiveness of compressor
 controls  to be marginal.

      Response:   In determining  the costs and cost effectiveness
 associated with  controlling  VOC  emissions  from compressors at  proposal,
 EPA  considered the costs  for eight individual compressor control
 configurations.   Specifically,  EPA calculated the costs for  two types
 of compressors (reciprocating and  centrifugal) in two  different VOC
 services  (wet gas or  NGL  at  plants  either  with or without an already
 existing  control  device).  The  resulting compressor control  configurations
 and  the cost  effectiveness for each one are  presented  in Appendix G,
 Table  1 of the BID for the proposed standards.  The cost effectiveness
 for  each  configuration ranged from  $36/Mg, for a centrifugal  compressor
 in NGL service at a plant with an  existing control device, to $2,200/Mg
 for  a  reciprocating compressor in  wet gas  service at a plant without an
 existing  control device.  The $2,200/Mg cost includes  the cost of
 installing  and operating a flare as the control  device.

     The  cost effectiveness of controlling reciprocating compressors in
 wet  gas service at a  plant without a control  device ($2,200/Mg) was
 judged to  be unreasonably high and, therefore, these compressors were
 exempted  from the proposed standards.   The cost effectiveness of
 controlling this same type of compressor at a plant with an existing
 control device was $l,700/Mg.  This cost effectiveness was considered
 to be  borderline in terms of whether it was reasonable or unreasonably
 high.  In making this judgment,  EPA assumed that typical  gas  plants
 would  employ both centrifugal and reciprocating  compressors  (about
 50 percent centrifugal and 50 percent  reciprocating).   About  one-
 third of these compressors would be used  in NGL  service and  the other
 two-thirds would be used  in wet  gas service.   Because  these  other
compressors can  be controlled at much  lower costs, the  average  cost

                                    7-5

-------
effectiveness for the plant to control  compressor emissions  would  be
closer to $460/Mg than it would be to $l,700/Hg.   This  cost  effectiveness
was judged to be reasonable.   Thus, EPA decided  not  to  exempt  from the
proposed standards reciprocating compressors  in  wet  yas service  at a
plant with a control  device present.

     Since proposal,  the comrnenters have indicated that many plants,
especially small plants, may have significantly  more reciprocating
compressors than centrifugal  compressors,  and, in some  cases,  an
individual plant may  have only reciprocating  compressors.   In  addition,
other commenters stated that all compressors  at  some plants  would  be
in wet gas service (i.e., there would be no compressors in NGL service
at some plants).  For plants like the examples cited by the  commenters,
the overall compressor control cost effectiveness would be $l,700/Mg
rather than the $460/Mg assumed at proposal.   The $l,700/Mg  cost
effectiveness was judged to be unreasonably high as  an  average control
cost effectiveness for wet yas reciprocating  compressors at  plants with
an existing control  device.  For this reason, all wet gas  reciprocating
compressors, including those located at a plant  without a  control  device,
are exempt from the promulgated standards.

     However, reciprocating compressors used  in  NGL service  and  all
centrifugal compressors in wet gas or NGL service are still  required  to
be equipped with closed vent systems because they can be controlled at
a reasonable cost effectiveness, as shown in Table 7-1.

     The average cost-effectiveness values for the compressor config-
urations  (wet gas reciprocating, wet gas centrifugal, NGL reciprocating,
and NGL centrifugal)  are shown  in Table 3-1.   The values in Table 3-1
are based on averaging the costs for each compressor configuration
(average cost of compressor with a control device and cost of compressor
without a control device) as  presented in Appendix G, Table 1, of the
BID for the  proposed standards.  These values show that controls are
cost effective  for all compressors except wet gas reciprocating
compressors.  The promulgated  standards, therefore, exempt all reciprocating
compressors  in  wet gas service.

7.3  PRESSURE RELIEF DEVICES

     Comment:   One commenter  (IV-D-2)  indicated  that the offset-mounted
pressure  reli ef valve/rupture  disk configuration  shown  in Figure 4-1 of
the BID  for  the proposed standards is  rarely  used in industry.  The
commenter  said  that  current  industry practice is  to mount the rupture
disk directly under  the  relief valve except where fragmenting graphite
disks  are  used.   The commenter noted that  the offset was used before
the availability of  nonfragmenting rupture disks  and that nearly all
disk manufacturers now  recommend  a nonfraginenting rupture disk.   This
mounting  method is also  recommended  by American  Petroleum Institute,
American  Society of  Mechanical  Engineers,  and Insurance Service Office
standards  for rupture  disk  installation.

     Response:   The  commenter, a  rupture  disk manufacturer, is  concerned
that  rupture disks may  have  been  considered  too  costly by EPA due  to
the  assumption  that  offset mounting  of the rupture  disk and relief

                                   7-6

-------
 valve  would  be  necessary.   The cominenter  is correct that  EPA did assume
 offset mounting  of a  rupture  disk and that offset mounting would not
 be  required  in  many cases.  However, the  capital costs shown in Table 8-1
 of  the BID  for  the proposed standards show that the additional capital
 cost  for  offset mounting  is $21, which  is less than 1 percent of the
 total  average capital cost  ($3,100) for new rupture disk  installations.
 Since  many  rupture disks  are  of the fragmenting type, and offset mounting
 would  be  required, EPA  included the minor cost of offset  mounting in
 the cost  estimates for  pressure relief devices.

      The  EPA maintains  the  use of rupture disks as an alternative to leak
 detection and repair, however, as provided in Section 60.632(d) of the
 proposed  standards.   The  use  of rupture disks does allow  for designation
 of  pressure  relief devices  equipped with  rupture disks as leakless
 equipment.   In  certain  situations, plant  operators may find rupture
 disks  preferable to routine monitoring of relief devices.  Pressure
 relief devices  equipped with  rupture disks require annual compliance
 testing.  The compliance  test for relief devices can coincide with the
 annual  relief device  safety tests normally performed by most gas plants.

      Comment:   One comrnenter  (IV-D-2) requested that EPA  change the
 language  for rupture  disk system costs to delete the phrase "at
 relatively high  cost" in  the  preamble to  the proposed standards:
 "installation of the  rupture  disk controls an additional   500 kg/yr
 but at  the relatively high cost of $6,700/Mg" (49 FR page 2639).
 The commenter agreed  with EPA's cost calculations for rupture disk
 systems,  but preferred  that plant owners or operators be  allowed to
 decide  what  is  relatively high.

     Response:   The phrase "at relatively high cost" relates the cost
 of  rupture disk  installations to that of routine leak detection and
 repair  for pressure relief devices.   The EPA does not intend to
 imply  that rupture disks are overly expensive or of high cost, but
 merely  that  rupture disks have relatively high cost compared to a leak
 detection and repair  program.

 7.4  CONTROL DEVICES

     Comment:  Several commenters  (IV-D-21,  IV-D-23, IV-D-24, IV-D-29,
 IV-D-31,  IV-D-33, IV-F-la and  b)  indicated that EPA had substantially
 understated the cost of a suitable dedicated  flare system.  One commenter
 (IV-F-lb) indicated that a suitable  flare system would  cost over $60,000
 (1980 dollars)  compared to EPA's  estimate of  $11,000 shown on page G-9
 of  the  BID for the proposed standards.   Another (IV-D-23) provided a
minimum cost of $100,000 to install  a  new flare including a smokeless
 flare tip, flare stack, foundation for  the flare stack, and related
valving and piping based on actual  new  smokeless flares that have  been
 installed at his company's facilities.   Other commenters  (IV-D-24,
 IV-F-la and IV-F-lb)  stated that  the  cost analysis in  the BID for  the
proposed standards does not account  for  the  flare velocity requirements
and fails to consider costs for the  additional  utility  consumption
(i.e., steam, pilot and purge  gases) and ancillary control systems.
One commenter (IV-D-33)  could  not  determine  from  the EPA  cost analysis
whether the  cost of the  larger flare had been  included.

                                  7-7

-------
     Response:  As discussed in Chapter 4,  the velocity requirements
for flares do not apply during malfunctions.   The EPA believes  that most
emergency flares will  have velocities below 60 ft/sec during  normal
operating periods.  The EPA has costed other  available control  devices  for
compressors.  The results of these analyses (Docket Item IV-B-7)  show
that flares and other  control  devices are cost effective when used to
control emissions from wet gas and NGL centrifugal  compressors, as well
as NGL reciprocating compressors.

     Comment:  One commenter (IV-D-21) pointed out  limitations  in using
process heaters as a VOC control  device at  his company.  The  commenter
noted that, as a control device,  heaters would have hiyh operating
costs (since heaters normally operate only  about 50 percent of the time
during the heating cycle of the regeneration  process) and would not
meet the design and operating standards specified in Section  60.632-9(c)
of the proposed standards for efficiency, minimum residence time, or
minimum temperature.  In addition, since heaters usually operate  at 25
to 35 psig, all vents  to the heater would have to be compressed,  creating
greater costs and safety hazards.

     Response:  The EPA recognizes that, in many cases, process heaters
operate on a cyclical  basis and are unsuitable for  use as VOC control
devices.  In those cases, the owner or operator would probably utilize
either a flare or other control device as opposed to continuously firing
the process heater (Docket Item IV-B-7).  The 25 to 35 psig pressures
mentioned by the commenter are typical of process heater fuel lines.
However, the heater combustion chambers are normally close to atmospheric
pressures, and the vent stream may be routed  directly to the  combustion
chamber for control.

     As discussed in Chapter 4, the temperature and residence time
parameters for incineration are provided as acceptable design parameters
to show compliance with the combustion efficiency requirement.   These
parameters are not required; other temperatures and/or residence times
may be used.

     Comment:  One commenter (II-D-30) said that costs for adding double
valves on open-ended lines are underestimated because these costs should
include recordkeeping, vehicle use, and source identification and
tagging.  In addition, the commenter wrote  that the cost estimate for
capping open-ended lines is based on the price of a 1-inch, screw-on
type globe valve and the incorrect assumption that  any lines larger
than 1 inch can be reduced to 1 inch.  The commenter suggested that EPA
review the 721 open-ended lines tested as reported  in Appendix A of the
CTG for equipment leaks of VOC in natural gas plants and base the costs
on a distribution of line sizes.

     Response:  API submitted this comment prior to proposal  and requested
that it be included in the rulemaking  (IV-F-le).  Double valving an
open-ended line does not require additional recordkeeping or tagging
because the second valve would not be subject to the valve leak detection
and repair requirements.
                                  7-8

-------
     Complying with the standards does not necessitate installation of
a second valve.  Open-ended lines may be capped or plugged at much
lower costs.  The basis for the cost estimate is the price of a one-inch,
screw-on type globe valve, which reflects the maximum cost likely to be
incurred for control of open-ended line emissions.  Larger lines would
likely have a blind flange installed at a similar cost, and smaller
lines would be capped or plugged at a much lower cost.

7.5   LEAK DETECTION AND REPAIR

     Comment:  One commenter (II-D-30) stated that the costs for the
labor associated with leak detection are severely underestimated, they
do not -apply to gas plants, and they are out of date.  Monitoring time,
according to the commenter, is a function of such factors as plant
configuration, the monitoring method, the personnel, the weather, and
the location of the component.  Monitoring labor charges of $4 per
source for contractor labor and $3.50 per source for plant personnel
(not including leak repair, resampling after repair, or initial  design,
acquisition, or implementation of the monitoring network) were offered.
It was also argued that front-end set-up costs, equipment depreciation,
and instrument maintenance should be included, as well as cost of
platform for inaccessible sources [Note:  This comment was made in
March 1982; therefore, the costs reflect that date].

     Response:  API submitted this comment prior to proposal  and requested
that it be included in the rulemaking (IV-F-le).  The monitoring time
estimates for plant equipment are based on the results of refinery
inspections and have been corroborated in chemical plant testing.
Since gas plants are similar in construction, and even more compact
than refineries, the monitoring time estimates are valid for gas plants.
Set-up costs, equipment depreciation, and instrument maintenance costs
are included in the cost analysis.  Leak detection costs account for
field labor time only.  Administrative, support, and instrument costs
to implement the standards are itemized separately.  The leak detection
and repair costs are based on field monitoring under all  weather conditions.
For Model Plant B, EPA's estimated costs fall  within the range of costs
the commenter quotes.   With 750 valves maintained at 2 man-minutes per
inspection and one-fourth the annual  instrument cost of $5,500, the
cost per valve inspection is $2.67.  The date the estimates were made
is unimportant, as a fixed-year basis is always used for making cost
comparisons.  A comparison of costs for leak detection and repair
between industry and EPA estimates is presented in Docket Item IV-B-5.

     Comment:  Four commenters (IV-D-20; IV-D-21; IV-D-26; IV-D-36)
stated that the cost effectiveness of monthly monitoring for valves and
pumps is unreasonable  and recommended that the monitoring frequency be
reduced to quarterly or annual.   One commenter (IV-D-20)  noted that
Table 1 of the preamble to the proposed standards shows that the
incremental  emission reductions from quarterly to monthly monitoring
would be 7 percent for valves and 15 percent for pumps.  The commenter
believed that this increase in emission reduction does not justify a
300 percent increase in monitoring time and  costs.   A second  commenter
(IV-D-36) argued that  a monthly program cannot be justified as cost
effective based on the costs incurred at Conoco's Cdshion plant, the

                                  7-9

-------
significantly lower repair rate achieved,  and the low leak  rate  which
occurred at the second test.   A third commenter derived  cost  effective-
ness values of $1984/Mg (monthly),  $768/Mg (quarterly),  and  $160/Mg
(annually) from monitoring program  net annual  costs  and  emission
reductions.  The commenter based his cost  estimates  on the  average  bid
price of several  leak detection contractors, and the emission reductions
(in part) for Regulatory Alternative III  in Table 8-8 of the  BID for
the proposed standards.

     A fourth commenter (IV-D-21)  said that an annual  leak  detection
and repair program for valves, relief devices, and pumps with a
corresponding annual  report,  would  reduce  industry's compliance  cost
and still meet the goals of the Clean Air  Act.  The  commenter thought
that a monthly leak detection and  repair  program is  excessively  burden-
some without much VOC emission reduction.   In particular, scheduling
monthly monitoring is difficult because it would depend  upon  the
availability of outside personnel,  weather problems, and unscheduled
plant problems.

     Response:  The goal of the Clean Air  Act is to  apply best
demonstrated technology (BDT) to newly constructed,  modified, or
reconstructed sources.  The Act requires  that standards  of  performance
reflect the degree of emission limitation  achievable through  application
of the best adequately demonstrated technological  system of  continuous
emission reduction, taking into consideration the cost of achieving
such emission reduction, any  nonair quality health and environmental
impacts and energy requirements.  In selecting the basis of  the  standards
for valves, EPA considered quarterly and monthly monitoring  as discussed
in Section 3.2.  Each of these intervals was compared in tenns of the
emission reduction achievable and  cost effectiveness of  the  leak detection
and repair programs as presented in Appendix H of the BID for the
proposed standards.  Monthly  monitoring was selected at  proposal  as the
basis for the standards for valves  because it achieves the  largest
emission reduction at reasonable costs and cost effectiveness.
Based on these estimates EPA  considers monthly monitoring BDT for
valves.

     Available data (Docket Items  II-A-25  and II-A-37) indicate  that
leak recurrence is an important factor in  predicting leaks  from  valves.
If a valve leaks, then it is  more  likely  to leak in  the  future than a
valve that has not leaked.  These  data also show that some  valves leak
less frequently than others.   Because leak recurrence can be  important
in predicting leaks, EPA considers  the annual  cost of monthly monitoring
of valves that leak infrequently to be unreasonably  high in  comparison
to the annual cost of quarterly monitoring, considering  the  emission
reduction achieved by monthly and  quarterly monitoring.   Therefore, the
proposed standards allowed quarterly monitoring for  valves  that  have
been found not to leak for 2  successive months, resulting in  a monthly/
quarterly implementation program.   The EPA expects that  most  affected
facilities would follow the monthly/quarterly schedule and  that  most
valves would be on the quarterly inspection schedule.   The  annual
emission reduction achieved by monthly/quarterly monitoring  is 37.7 Mg
for a Model Plant B, and the  cost  effectiveness is a credit  of $100/Mg.
The incremental cost effectiveness  of monthly/quarterly  from  quarterly

                                  7-10

-------
leak detection and repair is $240 per Mg (see Table 3-1).   As stated
above, the selection of the standards was based on emission reduction,
cost effectiveness, and incremental  cost effectiveness and not on the
percent increase between monthly and quarterly monitoring.  Additionally,
the actual costs for valves under the standards is likely  to be more
closely represented by the costs estimated for quarterly monitoring.
Consequently, the actual percent increase in cost will be  far less than
the 300 percent increase claimed by one commenter.

     In reviewing the data submitted by the commenters (Docket Item
IV-B-5), it was found that the commenters estimated total  annual  cost
to perform the leak detection and repair requirements of the standards
were generally in agreement with EPA's cost estimates.  Costs projected
by commenters IV-D-21 and IV-D-22 were lower than the costs estimates
using the calculation method presented in the BID for the  proposed
standards.  The projected costs presented by commenter IV-D-26 from
contractor bids and the actual  costs for one-time leak survey are
higher than costs estimated by EPA.   However, these costs  are similar
after correcting the GPA projections for several  misconceptions in
the requirements of the standards and correcting  accounting errors.
Table 7-2 presents a summary of the  total annual  cost estimates to
perform a leak detection and repair  program at a  500-component gas plant.

     The commenter included the costs to tag each component, record and
report more information than required, and other  activities that would
not occur at a typical plant inspection.  The contractor bids were also
based on 1984 dollars and did not consider that most components (i.e.,
valves) would be monitored on a quarterly rather  than monthly basis.
Some of the contractors providing bids may not have had any experience
in leak detection and repair work (Docket Item IV-E-1), and could have
overestimated the cost of performing the work.

     Comment:  One commenter (IV-F-lb) stated that EPA had failed to
include several cost elements in their analyses of leak detection and
repair program costs.  These items included:

     t  The initial setup costs for  special  flowsheets and component
        identification,

     •  The additional costs of the  "more elaborate" recordkeeping and
        reporting contained in the proposal, and

     *  The cost of lost product associated with  any special  shutdowns
        required to repair minor valve leaks.

     Response:  The EPA disagrees with the commenter's contention that  EPA
failed to include several  cost elements in their  analysis.  The
comrnenter's first example cites initial setup costs for special  flow-
sheets and component identification.  As discussed in Chapter 11,
detailed schematics, design specifications, and piping and instrumenta-
tion diagrams are included in the recordkeeping requirements for closed
vent systems and control devices as  specified in  Section 60.636(d)(l)
of the proposed standards.  However, this information should already be
available to operators and, consequently, little  burden would be incurred
by plants as a result of the standards.

                                  7-11

-------
  TABLE 7-2.   SUMMARY OF TOTAL ANNUAL  COSTS TO PERFORM  LEAK
       DETECTION AND REPAIR PROGRAM AT A 500-COMPONENT  GAS PLANT
Source of Estimate3       Total  Annual  Cost,  1980 dollars
EPA Proposal  BIDb                36,000
GPA Testing at Conoco
    Cashion gas plant0            34,800
Union Oil  Companyd               28,000
EPA-corrected value for
    GPA estimate                 35,250
 Docket Item IV-B-5.
b
 Costs are based on monthly leak detection and repair for pumps,  quarterly
 for pressure relief devices,  and monthly/quarterly for valves.   Inspections
 are performed by a two-person team (one  plant person and one  contractor).
 Based on 500 total components, the equipment distribution is  proportionate
 to Model  Plant A equipment count:  490 valves, 5 pressure relief devices,
 and 5 pump seals.  On the average, EPA estimates about 1.1 minutes  to
 monitor a component.  Hence,  about 3 hours would be required  to  monitor
 the gas plant during a typical inspection.  This assumes contractor or
 central office will  purchase  two monitoring instruments and use  them to
 inspect 5 plants.
c
 Itemized cost submitted by GPA represents second-year costs (corrected to
 1980 dollars) which are based on the ratio of Chemical Engineering  cost
 indices (263.2/319.3.  First-year costs  (corrected to 1980 dollars) are
 $37,100, which are higher than subsequent years because the cost of
 the first test is not amortized by GPA.
d
 Docket Item IV-D-22.  Costs adjusted to  1980 dollars using cost  index  factor
 of 263.2/319.3 from Chemical  Engineering.  Costs submitted by Union
 are based on a monthly monitoring program for an average plant (35  MMcfd),
 508 components tested at a rate of 2.6 man-minutes per component.
e
 Docket Item IV-B-5, Table 3.   Based on Chemical Engineering cost indices
 (263.2/319.3) calculated from 1984 dollar values"
                                  7-12

-------
      Component  identification  is also  required under the  recordkeeping
 requirements.   Section  60.636(b) of the  proposed standards clearly
 states,  however,  that an  identification  number is  to be attached to
 leaking  equipment and is  not required  for all equipment.  This
 requirement  has  not  changed since  proposal.   Identification can be
 accomplished  by  monitoring teams affixing inexpensive tags to leaking
 components.   The  cost for tags  for leaking components is  very small.

      The commenter's reference  to  "more  elaborate" recordkeepiny and
 reporting  contained  in  the proposal may  be interpreted as EPA's failure
 to  attach  a  cost  to  gathering  each individual piece of information that
 must  be  recorded  or  reported.   In  this case,  EPA has responded to
 similar  comments  in  Chapter 11, Recordkeeping and  Reporting.  However,
 the commenter may also  be referring to very elaborate recordkeeping and
 reporting  practices  presently  implemented in  the petroleum refining
 industry.  The commenter  would  be  correct by  inferring that facilities
 may implement more elaborate information systems than required under
 the standards.   Several refineries collect far more information than
 required by  existing State and  local  regulations because  their
 experience has demonstrated the usefulness of the  information.  Added
 benefits include minimizing product loss, information to make better
 future purchases of  equipment  by manufacturer and  type, and developing
 more  efficient maintenance practices.

      Furthermore, EPA has not  accounted  for lost product  associated
 with  special shutdowns  because  the standards do not require special
 shutdowns  as the commenter remarked.    The EPA has  included provisions
 (Section 60.632-8 of the  proposed standards) allowing the delay of
 repair for equipment leaks whenever repair is technically infeasible
 without a  process unit  shutdown.

      Comment:  One commenter (II-D-10) expressed concern that monitoring
 time  estimates do not include time for personnel  travel  to the work
 site  or  for  reaching the  component once  on site.    As an example, the
 commenter stated that relief valves located on top of gas plant vessels
 might be 50 to 120 feet high, accessible by a ladder.   According to the
 commenter, climbing up and down a 50-foot column  requires more than
 8 mi nutes.

      Response:  The EPA recognizes that some components will  require more
 time  to monitor than the estimates used in the cost analysis.   The  cost
 of monitoring is based on average monitoring times, measured during
 actual test programs.  Most sources will  require  less  than average
 times to monitor, so that a few sources requiring greater than average
 time will result in the total  monitoring time being that  presented  in
 the BID for the proposed standards.

     Comment:  Two commenters  (IV-D-11; IV-D-34)  maintained  that the
 proposed standards would be  burdensome in view of the  instrumentation
 involved in monitoring.   As  an  example, the commenter  wrote  that the
cost of a typical VOC analyzer  is  approximately  $5,000,  with  an  annual
operating cost for supplies  and labor of $4,000.   This  $9,000  first-year
cost is exorbitant compared  to  the  cost of current  (audible,  visual,
and olfactory)  leak  detection  practices in the commenter's opinion.'

                                      7-13

-------
     Response:   The EPA based  the  cost  for  instrument monitoring  on  one of
the available VOC analyzers  at $4,600 per instrument  (in  1980 dollars).
This cost is similar to the  $5,000 (assumed  to  be  in  1984  dollars)
presented by the commenter.   Although less  expensive  VOC analyzers are
available, EPA used the higher cost to  estimate the cost  impacts  of  the
standards.  In  addition, the cost  analysis,  presented in  Chapter  8 of
the BID for the proposed standards, includes the cost of  two  instruments.
The standards only require one instrument;  however, a second  instrument
is included in the cost analysis because  plant  operators may  need to
purchase a second instrument as a  spare in  the  event  the  first  becomes
inoperable.  Alternatively,  operators could  simply stock  spare  parts
for the instrument.  The total estimated  capital cost  for  instrument
monitors per gas processing  unit is $9,200.

     Annual costs for monitoring instruments include  annualized capital
costs and annual operating costs.   The  EPA  estimated  annualized capital
costs at $2,100 per gas processing unit.  This  cost  is  based  on a
6-year life for the instruments and an  annual  interest  rate of  10 percent.
Estimated annual operating costs include  $3,000 for materials and labor
to maintain and calibrate the instruments and  $370 for  rni seel laneous
costs.  The EPA's estimated  total  operating  costs  are  $3,370  which compares
closely with the commenter's estimate  of  $4,000 (assumed  to be  1984
dollars).  The EPA's total annual  cost, therefore, is  approximately  $5,500.

     Given the commenter's operating costs  and  annualizing the  commenter's
purchase price for two instruments adjusted for the  cost  year basis
results in a lower annual cost than estimated  by EPA.   The Chemical
Engineering cost index for June 1980 represents about 75  percent  of  the
cost index in December 1983 dollars.  Therefore, the  commenter's  purchase
price for  two instruments amounts to $7,500 in  1980 dollars,  and
annualized over 6 years at 10 percent interest at  $1,725.   The  commenter's
operating costs would  equate to about $3,000 (assuming  that the operating
costs would be the same for one instrument  as two).   The  commenter's
total annualized cost  would be nearly $4,700 or only 85 percent of  the
cost EPA used as the basis for the cost and economic  impact analysis of
the standards.

     Actual annual  instrument costs may be  significantly  less than  EPA
estimated.  As mentioned, less expensive VOC analyzers  are available
and operators may  use  a soap  solution to screen components as provided
in  Method  21.   Soaping may reduce  labor time during inspections and
reduce wear on  instruments.   Operators may also find it more economical
to  stock  spare  parts than to  purchase a spare  instrument.  Nevertheless,
the purchase of  a  VOC  analyzer  is  necessary to detect for  leaking
equipment  in order  to  determine compliance with the 10,000 ppm leak
definition  or  the  500  ppm no  detectable emissions limit.    The commenter's
current  practices  of audible,  visual, and olfactory leak detection  are
ineffective  in  discerning all  equipment leaks.  Also, soaping  is not
applicable  to  all  equipment types  (i.e., rotating or reciprocating
shafts).   As mentioned,  the cost  analyses in the  BID for  the proposed
standards  assume  two instruments  are purchased  for each plant.   Most
plants  would  be able to  share monitoring instruments with  other  plants
within  the same  corporation or, where contractors are utilized,  with
other  companies.   A second  possibility is for  a company to purchase one

                                   7-14

-------
instrument for each plant, and one or two spare instruments to be
shared by all  of its plants.  Where monitoring is  to be performed by
central office personnel, the company would probably purchase two or
three instruments to be used for all  plants.

     Although the instrument cost may appear high  to the commenter, the
majority of the costs are offset by the value of the products saved
through loss prevention.  For example, Model  Plant B has annual  recovery
credits due to leak detection and repair programs  of $13,230/yr  with
annual program gross costs of $13,944/yr (based on values for labor and
recovery credits presented in Appendix H of the BID for the proposed
standards).  Therefore, the leak detection repair  program net cost is
less than $800/yr, while reducing emissions by 45.7 Mg/yr.
                                  7-15

-------
                          8.0  ECONOMIC IMPACTS
8.1  PRICES

     Comment:   One commenter on two occasions (IV-F-la; IV-U-26)
contended that natural  yas processors will  have to bear the full costs
of compliance with the standards and will  not be able to recover these
costs.  The comrnenter cited that natural  gas processing economics is
based on the margin between the value of natural gas liquids components
as a fuel in natural  gas compared to the value as a petrochemical
feedstock or liquid fuels.  The economics  is not based on general
energy costs as assumed in the EPA analysis.  The commenter added that
these margins are quite small, particularly for ethane and propane
components, such that additional operating  costs could make new plant
construction uneconomical.  The comrnenter  went on to indicate that
natural  gas is sold on long-term, fixed-price contracts.

     Response:  The commenter raises two issues that are discussed below:

     (1) the role of long-term contracts in predetermining natural
         gas prices,  and

     (2) the relevant process economics for an economic impact analysis.

     1.   Natural Gas  Pricing

     The NSPS will affect primarily new sources of natural  gas, but
some existing sources of natural gas are also expected to be affected.
Long-term, fixed contract prices have been  negotiated for some of the
output of existing wells.  The term of some of these contracts, from
date of origin to date of termination, is  as much as 25 years.  However,
in recent years (since the rnid-1970's) a great deal of new natural gas
output has been sold  on the basis of prices negotiated in contracts
with a much shorter term of completion (2  to 3 years).  Furthermore,
over the same  time period, increasing amounts of natural gas have been
sold on  the spot market at spot, or current, market prices.  In addition,
many of the long-term gas contracts under  which natural gas used to be
marketed (which were  negotiated during the  1950's and early 1960's)
have been or will shortly be terminated.   Thus it is reasonable to
assume either  that the output from natural  gas wells will  be marketed
in the future at current prices or at prices negotiated under medium
term contracts that are closely tied to expected or actual  future spot
prices for natural yas (or perhaps linked  to a general energy price
index; for example, a 6-month moving average index for oil  prices).

     2.   Process Economics

     The commenter contends that the relevant process economics for
natural  gas plants should involve a comparison of the profitability of
natural  gas liquids in its alternative end  uses.  The commenter
identifies three general  categories of end  use:
                                  8-1

-------
     •  Natural gas liquids components as a fuel  in natural  gas

     •  Natural gas liquids as a petrochemical  feedstock

     •  Natural gas liquids as a liquid fuels

     If the standards were likoly to have a measurable effect on the
cost of natural gas as a feedstock to any of the  above uses, then the
relative economic impacts of the regulation on  each of these indirectly
affected markets would have to be estimated.  However, the economic
impact analysis indicates that the market price for natural  gas is  not
likely to increase measurably as a result of the  fiSPS.  Therefore,  the
process economics issues raised by the commenter  concerning  processes
that utilize natural  gas as a feedstock do not  have to be taken into
account in the economic analysis as the NSPS is estimated to have no
significant impacts on the economic conditions  under which firms involved
in those markets operate.

     It is possible that the commenter is concerned about the effects
of the NSPS on vertically integrated firms involved in gas exploration
and well head production in addition to downstream  activities, such  as
the manufacture of petrochemicals or liquid fuels.   These firms, however,
have the option of selling or buying natural gas  on the open market.
Thus the opportunity cost to them of natural gas  as a feedstock in
their downstream operations is in fact the market price of that gas.
Thus the process economics of and profitability of  downstream activities
involving natural  gas will be unchanged by the  NSPS for the  reasons
discussed above.

     Comment:  One commenter (IV-D-3) claimed  that  the assumption by
EPA (page 9-12 in the BID for the proposed standards) is untrue that
gas deregulation would be in effect and gas prices, along with liquid
prices, will  increase.  The commenter maintained  that if prices decrease
or usage of natural  gas decreases, the impact  of  the standards would be
greater and perhaps could slow or stop growth  in  the industry.

     Response:  The EPA acknowledges that the  price of natural gas  in the
future will  likely fluctuate.  In all likelihood  the price will be
relatively low in the 1985-1990 time frame but  will then begin to slowly
increase into the mid-19901s.  In any case, the price of methane-ethane
(Table 8-5 in the BID for the proposed standards) used to calculate
recovery credits is based on a natural gas price  of $1.46/Mcf.  This
price is very low and is, in fact, considerably lower than even the  low
projections for the 1985-1990 time frame.

     Comment:  One commenter (IV-D-12) expressed  concern that the
additional  cost burden imposed on the gas industry  by the proposed
standards must be paid eventually by the consumer.   The commenter added
that the negative effect on the environment of  switching from gas to
other fuels (because of increased gas costs),  which would occur mainly
in populated areas, would more than offset the  environmental benefit to
any remote rural areas.
                                  8-2

-------
      Response:   The  EPA economic  analyses,  prepared  prior  to  proposing  the
 VOC  equipment  leak standards, are  presented  in  Chapter  9 of the  BID  for
 the  proposed standards.   These  analyses, as  discussed  in Section  9.2  of
 the  BID  for the  proposed  standards are  used  to  determine the  maximum
 price increases  required  to maintain  the same profitability as would  be
 present  in the absence of the standards.  Page  9-30  of  the BID for the
 proposed  standards indicates that  changes in natural gas prices  would
 be less  than 0.5  percent  for the  most stringent of the  regulatory
 alternatives considered.   Based on the  ratio of the  cost of the  actual
 standards to the  cost of  the most  stringent  regulatory  alternative,  the
 standards will cause less than  a  0.07 percent increase  in  gas prices.
 The  actual natural gas cost increase due to  the effects of the standards
 is expected to be less than 0.3 <|:/Mcf.

      Since the price increases  are very small, and the  demand for
 natural  gas is relatively inelastic, it is  highly unlikely that  such
 small  increases  in prices would have any effect on natural gas usage.
 Alternative fuels (as shown in  Table 9-12 of the BID for the  proposed
 standards are more expensive than  natural gas, even  if  the price
 increases are understated.  Therefore, no adverse environmental  impacts
 from fuel switching are likely.

      Comment:  One cornmenter (IV-D-30) stated that EPA's assessment of
 the  economic impacts of the standards based on natural   gas prices was
 not  accurate. The effect  of the proposed standards on certain petro-
 chemical  feedstocks, propane for  rural heating, and  butane and natural
 gasoline  for motor fuels  should be considered.  The  commenter noted
 that gas  plants are built  because  natural gas liquids are worth more in
 a separate liquid phase than in a  vapor phase mingled with lighter
 natural gas components, and that this difference (the margin) is often
 slight and changes rapidly.

      Response:  The commenter contends that, since the  purpose of
 natural gas processing plants is to produce natural   gas liquids by
 extraction from the natural gas, the economic analysis  should be based
 on natural gas liquids prices rather than natural  gas prices.  The EPA
 realizes that in some situations, natural  gas liquids prices could be
 affected by the proposed  standards.  However, most of the impacts would
 be very slight due to the  small  compliance costs with the standards.
 It should also be pointed  out that one of the primary goals of examining
 natural gas prices was the determination of product   recovery credit.
 The  recovery credits  calculated are lower than actual,  since they are
 based on a combination of  natural  gas and natural  gas liquids prices, and
 natural gas prices are expected to be higher than the $1.46/Mcf used  by
 EPA.

8.2   INDUSTRY IMPACTS

      Comment:   One commenter (IV-D-3) stated that the annual  costs of
 implementing the  standards would be in the  neighborhood of 2 percent
 to 13 percent of  annual  revenues for two plants  owned by his company,
compared to  less  than 0.01 percent claimed  by EPA in  the preamble to
the  proposed  standards.   The commenter did  not provide  any explanation
of the calculation of the  relative cost  values  claimed.

                                  8-3

-------
     Response:  The EPA indicated in the preamble to the proposed btdndurd
a value of 0.1 percent as the maximum increase in consumer gas prices,
rather than 0.01 percent as indicated by the commenter.   The estimated
cost to industry of less than 0.1 percent of revenues is based on
industry average values and, therefore, could be higher  or lower for
individual plants.  The EPA expects that costs could be  as high as
2 percent of revenues at only a few plants; however, EPA does not think
that the costs of the standards could be as high as 13 percent of
revenues at any plant.

8.3  SMALL BUSINESSES

     Comment:  Two commenters (II-B-23; IV-D-3) stated that, contrary
to EPA assertions, many gas plants would qualify as a small  business
based on a criterion of less than 500 employees.

     Response:  The EPA recognizes that, based on a criterion of less than
500 employees, some companies producing natural gas qualify as small
businesses.  However, the standards are based on cost-effective control
techniques for significant emission sources, with the cost effectiveness
based on the product values of gas plant products.   In general, the
standards are as cost effective for small  businesses as  for large
businesses since the costs and emissions reductions are  independent of
business size.

     Small nonfractionating plants (<10 MMscfd) are excluded from leak
detection and repair programs due to poor cost effectiveness.  This
exclusion applies to small plants owned by large businesses as well as
those owned by small businesses.

     The Regulatory Flexibility Act states that if there are a significant
number of "small  gas companies" that will  incur adverse  economic impacts
as a result of the regulation, a  regulatory flexibility  analysis is
required.  Since small businesses are not expected  to incur adverse
impacts, no analysis is required  for these standards.

     Comment:  One commenter (IV-D-23)  remarked that a large majority
of compressors are fitted with open distance pieces and  cannot be
fitted with closed and sealed distance  pieces.  Some of  the smaller
compressor manufacturers do not offer the accessory equipment, such as
double distance pieces.  Consequently,  the standards for compressors
will cause the smaller compressor manufacturers to  go out of business.
Since only large manufacturers will  be  able to supply compressors to
the gas industry, the standards will force a limited competition
si tuation.

     Response:  The EPA recognizes that some existing compressor designs
cannot be technologically or economically retrofitted with double
distance pieces.   Since proposal, EPA has decided to exclude all wet
gas reciprocating compressors, so that  the only reciprocating compressors
subject to the standards are those in natural jas liquids service.

     Compressors of any size may be fitted with double distance pieces
since the physical size of the enclosed compartment may  bo small.

                                  8-4

-------
Alternative sealing techniques, such as doubly-vented barrier fluid
packings or single compartment distance pieces with improved seals may
also be used.   Threfore, small compressor manufacturers are provided
with sufficient leeway to adapt compressors to meet the requirements
of the standards.
                                8-5

-------
                  9.0  MODIFICATION AND RECONSTRUCTION

9.1  CAPITAL EXPENDITURE

     Comment:  Two commenters (IV-D-14 and IV-D-19) recommended that a
definition for "capital expenditure" should be added to the standards
before promulgation.  The commenters suggested that "capital  expenditure"
be defined as 10 percent of the replacement cost of the affected facility
at the time the process improvement is made.   Replacement cost should
be based upon the Chemical Engineering Construction Index or another
suitable cost index.  The comenters thought that this approach would make
the modification provision in the NSPS more comprehensible and workable.

     Another commenter (IV-D-35) recommended  raising the fixed percentage
AAGRA for natural  gas plants from 4.5 percent to 10 percent.   According
to the commenter, this would bring the repair allowance threshold in
line with onshore drilling and avoid discouraging process improvement
modifications at existing gas plants.  The commenter stated that the
application of a tax rule which sets an arbitrary break point between
an expense and capitalization requirements for these facilities is
impractical and ignores real world problems.   The commenter reasoned
that gas plants are typically designed for declining production rates
and, therefore, are not frequently expanded or added to as are other
facilities, and thus the cost basis is normally low.

     Response:  The EPA agrees with the comrnenter's suggestion that
capital  expenditure can be based, in part, on a certain percentage of
the replacement cost of the affected facility.  After reviewing similar
comments on other standards of performance (e.g., Equipment Leaks of
VOC in SOCMI, Subpart VV, and Petroleum Refineries, Subpart GGG), EPA
decided to provide an alternative for the definition of "capital  expenditure"
in 40 CFR 60.2 of the General Provisions.  Although the implementation
of the capital expenditure definition has been made more practicable,
the original  intent of the definition has been maintained.

     The alternative uses an adjusted annual  asset guideline repair
allowance (AAGRA)  and the replacement cost to determine capital
expenditure.   Details of the alternative, which is applicable to plants
built prior to 1982, are discussed in the BID for the promulgated
standards for equipment leaks of VOC in petroleum refineries
(EPA-450/3-81-015b).  A definition for capital expenditure was added to
40 CFR Part 60, Subpart VV (Standards of Performance for Equipment
Leaks of VOC in the Synthetic Organic Chemical Manufacturing  Industry),
Section 60.481 (49 FR 22607, Hay 30, 1984) and, therefore, is incorporated
by reference into the promulgated standards for natural  jas processing
plants.   "Capital  expenditure" means, in addition to the definition in
40 CFR 60.2,  an expenditure for a physical or operational  change to an
existing facility that exceeds P, the product of the facility's replacement
cost, R, and  an adjusted AAGRA, A, as reflected by the following equation:

                                P = R x A

The replacement cost, R, means the capital needed to purchase all  the
depreciable components in a facility.   The adjusted AAGRA, A, is a

                                  9-1

-------
 product of the age-adjusted percentage, Y, of the replacement cost,
 and the applicable basic AAGRA, B, as reflected by the following equation:

                            A = Y x (B -i 100)

 For Subpart KKK, B is equal to 4.5.  The percent Y is determined from
 the following equation:

                          Y = 1.0 - 0.575 log X, where
                     X = 198Z - the year of construction
                               and X > 0.

 The relationship between replacement and original  costs was derived
 using inflation indexes (Docket Items 1I-B-50 and 1I-B-51).

     In response to the request of the second commenter, the 4.5 percent
 AAGRA for natural  gas processing plants is established by the Internal
 Revenue Service, and its application to NSPS is established by the
 General  Provisions for NSPS.   EPA recognizes that other industries use
 different AAGRA percentages (for example, the AAGRA for petroleum
 refineries is 7.0 percent).  The AAGRA values are industry specific
 based on the type of operations performed by the industry and the
 expected repair expenses an industry is likely to incur.  Consequently,
 the AAGRA would require adjustment by the IRS and not EPA.

 9.2  MODIFICATION OF EXISTING SOURCES

     Comment:  One commenter  (II-B-23) suggested that "modification" be
defined  and included in the preamble due to several  uses of the term.

     Response:  The definition for modification is included in the
 General  Provisions, 40 CFR  60.14; therefore, it does not need to be
defined  in each separate subpart.

     Comment:  One commenter  (IV-D-37) suggested that the provisions
 for process improvements without a capital  expenditure (Section 60.630(c))
be revised to "... process  improvement that is accomplished without an
increase of emissions shall  not by itself be considered a modification."
The commenter explained that  this would give older facilities the
incentive to modernize and  reduce emissions if modernization did not
trigger  NSPS.

     Response:  The EPA stated its intent in Section 60.630(c) of the
proposed standards that minor modifications would  not be covered by the
standards:

             Addition or replacement of equipment for
             the purpose of process  improvement that is
             accomplished without capital  expenditure
             shall  not by  itself be  considered a modifi-
             cation under  this subpart.

The capital  expenditure criterion was  included so  that minor process
improvements in a  process unit that  cause an increase in emissions

                                  9-2

-------
would not subject an existing facility to the requirements of the
standards.

     It should be noted that any potential  emission increase that results
from changes in operation that require the  addition of a few fugitive
emission sources could be offset or nullified by controlling existing
equipment or installing components with no  fugitive emissions.
Accordingly, there would be no modification in such a case even if the
emissions occurred with a capital  expenditure.  The standards do not
require that process improvements be made without a capital  expenditure.
They merely provide an exemption when the process improvements  are made
with such an expenditure.

     As described in Chapter 5 of the BID for the proposed standards,
Section 60.14 of the General Provisions limits the modification
provisions to changes which result in an increase in emissions.

9.3  MISCELLANEOUS

     Comment:  One commenter (IV-D-19) wrote that it would be helpful if
EPA would clarify that facilities under different ownership are separate
sources.  The commenter gave as an example  the following situation.  If
company X modifies its C02 removal plant (sweetening unit) which is
adjacent to and serving company Y's gas plant, the NSPS should  not be
triggered.

     Response:  The NSPS defines an affected facility as either an
individual compressor or all equipment in a process unit.    If  an
individual process unit (such as a dehydrator) is modified, the process
unit is then subject to the requirements of the standards.  However,
the remainder of the plant would not become subject to the standards.
In the commenter's example, the sweetening  unit would become subject to
the standards, but the remainder of the plant would not.
                                  9-3

-------
                            10.0  TEST METHODS
 10.1  LEAK DETECTION METHODS

      Comment:   Numerous  commenters  (IV-D-35;  II-B-23;  II-D-20;  II-D-30;
 II-E-8)  stated  that EPA  should  recognize  the  validity  of  less costly,
 faster,  easier,  and more accurate  leak detection  using soap scoring.
 One commenter  (IV-D-35)  noted that  several  factors  lead to the conclusion
 that soap scoring  is better than  instrument monitoring:

      •  Cost -  a  small trigger-activated  pump bottle ( $1.75) can be
         used for  soap scoring as opposed  to a $6,000 instrument for
         instrument  monitoring.

      •  Ease of  testing  -  the soap  solution bottle  is much lighter
         (  1/2 to  1  kg) than the monitoring  instrument  ( 7 to 8 kg),
         which is  important when the  operator must climb over pipes,
         climb ladders, or  traverse walkways.

      •  Accuracy -  the commenter noted that industry demonstrated, by
         statistically comparing the  soap  score with volumetric leak rate
         and concentration measurements, the absolute accuracy of the
         soap test (Eaton, et. al . 1980).  The EPA, according to the
         commenter,  attempted to show that the detection instrument is
         as accurate as the soap test (DuBose, et. al. , 1982).   However,
         EPA performed their analysis by comparing instrument readings
         with soap scores rather than actual  leak rates.  Thus, the
         validity of EPA's claims of accurate leak detection using the
         instrument  is not supported by gas plant data.

      In  one test (Eaton, 1980), commenter IV-D-35 reported that 81 percent
 of  all OVA-detected  leaks were found to have a soap score of three or
 greater.  Union Oil   Company data (Anderson,  1984) showed identical
 results  for the two  methods, but soap scoring was almost twice as fast.
 Based  on  these reasons,  the commenter suggested the proposed standards
 be modified to allow  use of soap testing, with a soap score of three or
 more  indicating a leak.

      Another commenter (IV-D-10) also recommended that  EPA allow the
 use  of soap scoring  at natural  gas plants to quantify VOC leaks from
 components for which soap scoring is effective.   The commenter wrote
 that,  while soap scoring  is subjective, the  method has  been shown  to
 correlate accurately with leak  rate and concentration measurements.
 Commenters IV-D-30 and IV-D-35  provided results  of a 1982 Western  Oil
 and  Gas Association  (UOGA) training session  in which 500 measurements
 were made using  10 hexane-calibrated OVA's.   The data  indicated that
81 percent of all  OVA-detected  leaks received  soap scores  of 3  or
 greater,  while  92 percent of all  OVA-detected  leaks  received soap
scores of 2 or  greater.   Commenter IV-D-10 felt  that due  to the vari-
ability of OVA  readings,  the  correlation  between OVA readings  and  soap
scores is probably better than  the data indicate.

[Note:  OVA is  a  registered  trademark,  use of  which  does  not  imply
        endorsement  by LPA.J

                                  10-1

-------
     Response:   Prior to promulgation  of Method  21,  an  alternative  soap
screening procedure was added  for sources that can  be  tested  with a
soap solution.   These sources  are restricted  to  ones with  non-moving
seals,  moderate surface temperatures,  without large  openings  to  atmosphere,
and without evidence of liquid leakage.   The  soap  solution is sprayed
on all  applicable sources,  and the potential  leak  sites are observed  to
determine if bubbles are formed.   If no  bubbles  are  formed, then no
detectable emissions or leaks  exist.  If any  bubbles are formed, then
the instrument measurement  techniques  must be used  to  determine  if  a
leak exists, or no detectable  emissions  exist, as  defined  in  the regulation,

     The alternative soap solution procedure  does  not  apply to pump
seals,  components with surface temperatures greater  than the  boiling
point or less than the freezing point of the  soap  solution, components
such as open-ended lines or valves, pressure  relief  device horns, vents
with large openings to atmosphere, or any component where  liquid leakage
is present.  The instrument technique specified  in  the  method must  be
used for these components.

     The alternative of establishing a soap scoring leak definition
equivalent to a concentration-based leak definition is  not included
in the method and is not recommended for inclusion in  an applicable
regulation because of the difficulty of  calibrating and normalizing a
scoring technique based on  bubble formation rates.   A  scoring technique
would be based on estimated ranges of volumetric leak  rates.   These
estimates depend on the bubble size and  formation  rate, which are
subjective judgments of an  observer.  These subjective  judgments could
be calibrated or normalized only by requiring that the observers
correctly identify and score a standard  series of  test bubbles.   It has
been reported that trained  observers can correctly and repeatedly
classify ranges of volumetric leak  rates.  However,  because soap scoring
requires subjective observations and since an objective concentration
measurement procedure is available, a soap scoring equivalent leak
definition  is not recommended for this standard.  The  alternative
procedure that has been included will allow more rapid identification
of potential leaks for more rigorous concentration measurement using a
monitoring  instrument.


     Comment:  One commenter  (II-D-30) listed limitations of  the portable
hydrocarbon detector  in pointing  out that  the instrument  is  not the best
survey method, as EPA claims:

     • Extreme delicacy of the instrument;

     • Sensitivity  to correct calibration;

     • Weight and  inconvenience  of  the  instrument;

     • Poor repeatability;

     • Lack of  demonstrated accuracy;

     • Time delay  in  achieving a  reading;  and

     • Difficulty  in  receiving timely repairs of  the instrument.

                                   10-2

-------
       Response:   The EPA recognizes  that  portable  hydrocarbon  detectors
  have limitations;  however,  none  of  the  limitations  prevent  the use  of
  instrument monitoring  from  being  the  recommended  method  of  leak
  detection.   Although  typical  monitoring  instruments can  be  damaged  by
  accidental  dropping or misuse, they are  not considered "extremely
  delicate."   These  instruments have  been  in routine use in refineries
  subject  to  State or local leak detection and  repair requirements with
  few  problems.

       The  absolute  accuracy  of the instruments  is  not a real  concern.
  Since  the  purpose  of the instrument monitoring is to locate failures of
  seal  mechanisms, the readings observed by the  instrument operator are
  normally much lower or much higher than the 10,000 ppm leak definition.
  Consequently, a minor  change  in the calibration of the instrument makes
  little difference  in the number of leaks detected.

       The response  time of the hydrocarbon detector is required by
  Method 21 to be less than 30 seconds.   However, 3 to 10 seconds is more
  representative of currently available instruments.  The EPA  recognizes
  that  the instruments may be cumbersome in certain situations.   However,
  as mentioned before, use of the instruments has been demonstrated
  through routine use in regulated  refineries.   Additionally,  since this
  comment was written, EPA has added soap screening procedures to Method 21
  The use of soap screening will allow the operator to minimize  the
  use of the instrument,  since instrument monitoring is  only required  for
  sources producing visible bubbles during soap  screening.   Limited use
  of the instrument can result in  extending the  instrument  life  as  well
 as minimizing operator  fatigue.

      The instrument can be  repaired  by either  the  manufacturer or
 qualified repair  services.   Most  instruments are  relatively  simple  in
 design; therefore,  most problems  can be  repaired  in  the  field  by  the
 operator.   The  EPA  expects  the plants  to  maintain  spare  parts  as
 recommended by  the  instrument  manufacturer.  Costs for  a  second
 (backup)  instrument are included  in  the  cost analyses of  the BID  for
 the proposed  standards.

      Comment:   One  commenter (IV-D-20) mentioned that hydrocarbons are
 measured  using  the  Flame  lonization  Detector (FID) principle.   Such
 analyzers  are not capable of distinguishing methane and ethane  from
 heavier  hydrocarbons.   Thus, there is  no  way to determine the actual
 level  of  regulated  hydrocarbons in either the  background  air or from a
 leak.   Consequently, the commenter stated that  industry and  EPA will
 have  a  difficult time complying with and enforcing the proposed standards.

      One commenter  (IV-D-25) requested that EPA revise the definition
 of  VOC  in Section 60.2 of the General  Provisions to exclude  specifically
 methane and ethane.   The commenter reasoned that Reference Method 21
 detects any organic  compound including  methane and ethane.   In fact,
 methane is used as a calibration gas  in Reference Method 21.

     Response:  In selecting the affected facilities as  only  equipment
 in VOC service or wet jds service, the  program  limitb testing of
equipment to those sources with the potential  to emit VOC, that is,

                                  10-)

-------
those components in wet gas  service  or  ones  containing  at  least
10 weight percent VOC as determined  by  ASTM  Methods  E-260,  E-168, or
E-169.  The purpose of Method  21  is,  therefore,  to detect  whether a
regulated component is leaking, and  not to quantify  either  the leak
rates or the composition of  the leaking material.  Therefore, there
is no reason to require VOC  leaks  to  be distinguished  from  methane/ethane
leaks, as any component found  leaking will be  leaking  VOC.

     The 10,000 ppm leak definition  represents methane/ethane and VOC
combined.  The leak rates presented  throughout the BID  for  the proposed
standards are based on measurements  of  total  hydrocarbon concentrations
to determine leaking/nonleaking status.   As  such, the  testing requirements
for the standards are also  based on  total  hydrocarbon concentrations.

10.2  LEAK DEFINITION

     Comment:  One commenter (II-D-10)  stated  that there is no justifi-
cation for selecting 10,000  ppm as  the  action  level.

     Response:  API submitted  this  comment prior to  proposal and
requested that it be included  in the  rulemaking  (IV-F-le).   The selection
of the action level, or "leak  definition," for the definition of a  leak
was based on several factors,  which  are discussed in Section 4.2.3.1  of
the BID and in the preamble  to the  proposed  standards.   The rationale
for the selection of the action level for  valves is  different from  the
rationale for pumps because  of differences in  the cause of leakage.

     The best leak definition  would  be  the one that  achieved the most
emission reduction at reasonable costs.   Within  practical  limits, the
emission reduction achieved  would  increase as  the leak  definition
decreased, due to the increasing  number of components  that would be
found leaking and, therefore,  repaired.   At  a  leak definition of
10,000 ppm, approximately 90 percent of all  VOC  leaks  from valves would
be detected.  It is well documented  that valves  that have  been  found
leaking at levels of 10,000  ppm or  greater can be brought  to levels
below 10,000 ppm with proper maintenance.  Also, as  a  practical matter,
most commonly available hydrocarbon  detectors  that are  considered
intrinsically safe have a maximum  reading  of 10,000  ppm.   Leak definitions
higher than 10,000 ppm could,  nevertheless,  be selected (and dilution
probes could be used with portable  detectors); however, there would be
less emission reduction than with  the 10,000 ppm definition and  no  substantial
associated cost savings.  Consequently, there  is no  basis  for selecting
a leak definition greater than 10,000 ppm.   A leak definition lower
than 10,000 ppm may be practicable  in the  sense  that leaks can  be
repaired to levels less than 10,000 ppm.  However, EPA  is  unable  to
conclude that a leak definition  lower than  10,000 ppm  would provide
additional emission reductions and,  therefore, would be reasonable.
Because the 10,000 ppm leak  definition  would address approximately
90 percent of the VOC leaks  from  valves at  reasonable  costs and  at
reasonable cost effectiveness, and  because  safe, available hydrocarbon
detectors can read 10,000 ppm, the  10,000 ppm  level  was selected  as the
leak definition for valves.   In  contrast to  valves,  which  generally
have zero leakage, most pump seals  leak to a certain extent while
operating normally.  These seals  would  tend  to have  low instrument

                                  10-4

-------
 meter readings.   With  time,  however, as the seal begins to wear, the
 concentration  and  emission rate are likely to increase.  At any time,
 catastrophic seal  failure can occur with a large increase in the
 instrument  meter  reading and emission rate.  As shown in Table 4-2 of
 the  BID for the proposed standards, over 80 percent of the emissions
 from pump seals are  from sources with instrument meter readings greater
 than or equal  to  10,000 ppm.  Properly designed, installed, and operated
 seals should have  low  instrument meter readings.  Furthermore, the bulk
 of the pump seal  emissions are from seals that have worn out or failed
 so that they have  a  concentration equal to or greater than 10,000 ppm.
 Therefore,  10,000  ppm  was chosen as a reasonable action level.

      It should be  noted that the purpose of instrument monitoring is to
 detect seal failures (leaks), and not to quantify the leak rate.
 Although any leak  definition could be selected, EPA for the reasons
 given believes 10,000  ppm is the appropriate level  of hydrocarbon
 concentration  to  distinguish between leaking and nonleaking sources.

      Comment:  Two commenters remarked about the safety of testing flares.
 One  commenter  (IV-D-35) recommended deleting the requirement for compliance
 testing of  flares  using Methods 2, 2A, or 2C.  The commenter stated
 that Methods 2, 2A or  2C expose personnel  to risk of death if, during
 the  test, a component  attached to the flare relieves.   The commenter
 claimed that requiring this test is likely to bring a  challenge from
 the  Occupational  Safety and Health Administration.   The second commenter
 (IV-D-31) added that since flares are designed to handle intermittent
 releases from relief valves, it is uncertain how representative velocity
 and  gas  stream Btu content determination could be made and how such
 measurements could be  performed without endangering the safety of the
 testing personnel.

      Response:   Testing of flares should not require any measurements
 to be  made at the flare, other than monitoring for  the presence of a
 flame  with a permanently mounted  thermocouple.   The velocity measurements
 and  no  substantial associated cost savings.   Consequently, there is no
 made  in Method  2  or its alternates can be  made in the  flare header at a
 distance from the flare.  Therefore, no climbing or exposure to the flare
 is necessary.   Since Method 22 requires only that the  flare be observed
 from a  distance to determine if visible emissions are  occurring, no
 exposure to  the flare  is required.   The standards require that flares
 be operated  within specific  velocity guidelines.   The  stream Btu content
can be  based on engineering  calculations  of vessel  or  equipment contents.
                                  10-5

-------
                    11.0  RECORDKEEPING AND REPORTING

 11.1  GENERAL

      Comment:  Several commenters (IV-D-3, IV-D-5, IV-D-19, IV-D-20
 IV-D-21, IV-D-22, IY-D-25, IV-D-26.  IV-D-29, IV-D-30, IV-D-34,  IV-D-36,
 IV-F-la) urged EPA to reconsider the proposed recordkeeping and reporting
 requirements.  Some of these commenters (IV-D-5; IV-D-21; IV-F-la)
 expressed concern that the proposed  standards included more extensive
 recordkeeping and reporting requirements than the previous (NAPCTAC)
 draft.  One commenter (IV-D-25) thought the requirements should be
 eliminated because they are far too  complicated for the small  emission
 reductions that will  result from a leak detection and repair program.
 Another commenter (IV-D-30) said that the  recordkeeping and reporting
 requirements are totally counterproductive to the Federal government
 goals of reducing paperwork.   Other  commenters  thought the requirements
 were excessive.  Some commenters (IV-D-3,  IV-D-34, and IV-F-la)
 indicated that it will be unusually  burdensome  for the small plant
 operator to comply with these requirements because of minimal  staffing.
 One (IV-D-3) argued that small  plants would need to hire additional
 employees just to comply with the recordkeeping, and these additional
 employees could make  the difference  between operating and shutting
 down.   Another commenter (IV-F-la) did not think that EPA considered
 increased costs for recordkeeping, reporting, and allowances for setting
 up a system to comply adequately with these requirements.

      Several  of the commenters  took  exception to the semiannual  reports
 and thought that annual  reports would be  sufficient.   The commenters
 questioned  the ability of EPA to review and analyze the  data submitted
 over the next several  years and noted that annual  reports would be
 consistent  with State  and other Federal  regulations.

      Response:   The commenters  are incorrect in  stating  that record-
 keeping  requirements  of  the proposed  standards are more  extensive than
 the recordkeeping  requirements  of  the NAPCTAC draft.

      In  both  NAPCTAC draft  and  proposal  standards,  EPA  selected  the
 same  recordkeeping  requirements that  would  provide  the  necessary records
 for managing  implementation of  the required  programs  while  ensuring
 effective implementation  and  maintenance of  the  proposed  standards.
 The commenters  are  correct  in stating  that  reporting  requirements at
 proposal are  more extensive than  the  reporting requirements  at  NAPCTAC.
 After evaluating  three alternatives,  EPA selected  the  first  alternative
 as  the reporting requirement  at NAPCTAC, which provided  no  routine
 reporting of  compliance other than the  requirements  of  the  General
 Provisions  of  Subpart A of  40 CFR Part  60  (notifications  of  construction,
 anticipated  startup and actual  startup), the requirements  for notification
 of  intention  to comply with one  of the alternative  standards for valves,
 and the  reports necessary for determination of equivalence of alternative
 means of emission limitations.  Compliance under  this alternative would
 be assessed through in-plant  inspections.

     Since NAPCTAC, however, EPA re-evaluated the alternatives  for
 reporting requirements of the proposed standards.  The reporting require-
ments under the first alternative (no routine reporting) would not

                                  11-1

-------
 provide  a mechanism for checking the thoroughness of the industry's
 efforts  to reduce equipment leaks of VOC.  Facilities not complying
 with  the standards would not be using BDT as required by the Clean Air
 Act.   The EPA believes that reporting is necessary for the effective
 enforcement of the standards.  Reporting will reduce the necessity for
 many  in-plant inspections, while improving the enforceabi 1 ity of the
 standards.  The EPA's conclusion that reports are useful  is also based
 on  the experience of State and local air quality control  boards.  The EPA
 concluded, therefore, that reporting is necessary to assess implementation
 of  the work practice and equipment requirements of the standards without
 requiring excessive resources from industry and enforcement personnel
 and proposed that certain information be reported (i.e., submittal of
 semiannual reports to summarize information on leaking equipment and
 the number of leaks detected and repaired).

      One of the commenters asserted that EPA did not consider increased
 costs  for recordkeeping and reporting and for setting up a system for
 compliance with these requirements.  On the contrary, these costs are
 included in the cost analysis of the standards as administrative costs.
 In  addition, contrary to the commenter's remarks, EPA did make allowances
 for the costs of setting up a system to ensure compliance with these
 requirements.  The EPA is required under the Paperwork Reduction Act of
 1980  (PL-511) to submit a request to the Office of Management and
 Budget (OMB) for approval of recordkeeping and reporting requirements
 that  qualify as an "information collection request" (ICR).  For the
 purposes of OMB's review, an analysis was presented of the need for
 information, methods of ensuring the quality of data obtained, alternatives
 to  the information collection,  and estimated burden hours and cost to
 the industry and to EPA of complying with the requirements.   The costs
 of  recordkeeping and reporting  for both industry and EPA at proposal
 are presented in Docket Item II-F-3.  Since proposal, EPA costs of
 reviewing reports and compliance inspections have been revised because
 of errors made in the estimates at proposal and because of revisions in
 ICR collection procedures as discussed in the next response to comment.
 The revised costs are presented in Docket Item IV-H-2.

      The commenters also expressed concern that the recordkeeping and
 reporting requirements would be burdensome for small plant operators
 and that additional  employees would need to be hired just to comply
 with  the recordkeeping requirements.  The EPA considered the staffing
 needs  of plants in implementing the standards and analyzed the impacts
 of the standards on small plants.   The EPA recognizes that some small
 plants operate without technically trained personnel because of the
 type of process that is performed there.  In particular, small nonfrac-
 tionating plants often operate  either unmanned or without a staff
 having the technical  ability to conduct a leak detection and repair
 program and to maintain the associated records and reports.   In these
cases, EPA analyzed the additional  cost of hiring an outside consultant
or using central  office personnel  that would be needed and judged the
costs  to be unreasonable at nonfractionating plants having capacities
below  10 million standard cubic feet per day (MMscfd).   The EPA decided,
therefore, to exempt any nonfractionating plant whose capacity is
 10 million scfd or less of field gas from the routine monitoring requirements
                                  11-2

-------
  (and the associated  recordkeeping  and  reporting  requirements)  for
  valves,  pumps,  and  pressure  relief devices.   All  fractionating  plants,
  however, regardless  of  capacity  are required  to  implement  the  routine*
  monitoring  requirements because  these  plants  require  the presence of
  technically trained  personnel.   The recordkeeping and  reporting requirements
  have little impact on the cost of  the  standards;  therefore, EPA does
  not  anticipate  any disruption  in plant operations or  closings attributable
  to these requirements.   These  requirements are considered  the minimum
  consistent  with  adequate enforcement;  therefore, the  burden on owners
  and  operators is  the minimum necessary to enforce the  standards adequately.

       Comment:   One commenter stated  that the  reporting and recordkeeping
  requirements of  0.15 person-years  per  plant given in the preamble to
  the  proposed standards  (p. 2648-2)  are low.   The commenter offered what
  he considered more realistic impacts than the EPA estimates.   The
  commenter estimated that the initial monitoring, recordkeeping, and
  reporting for a typical  cryogenic  plant will   require 0.5 person-years
  The  commenter also offered that  a  plant representative of the 90 MMscfd/
  day  gas  plant referenced in the  regulations includes over 1,000 valves
  5 pumps, and 4 compressors.   The commenter also disagreed with the
  nationwide recordkeeping and reporting burden estimate given  on page
  2648 of  the preamble to  the proposed standards.   The commenter noted
  that the recordkeeping and reporting impact of 6.6 person-years reported
  in the preamble is based on the number of plants (44)  per year affected
  in the first 2-year period.   The  commenter  argued that the  impacts  are
 cumulative and  would, therefore,  affect 220 plants in  1987.  The
 commenter also  used his  0.5  person-years  per  plant impact to  estimate
 the nationwide  annual  burden at 22  person-years;  the first  5-year
 average at 55 person-years,  and that at the end  of 5 years, the
 annual burden would be  110 person-years.

      Response:  The 0.15  person-years per  plant estimate shown  in the
 preamble  to  the  proposed standards  represents  312 labor hours/year  for
 recordkeeping and reporting,  or 26  hours  per  monthly monitoring  period
 Since most required  records  are generated during  the actual monitoring"
 period by the monitoring team,  little additional  recordkeeping and
 reporting effort is  required.   The  amount of  labor  (26 hours) costed in
 the  preamble for  the  proposed standards should be more  than adequate
 for the average  plant.

      As discussed  above,  EPA  is required to submit recordkeeping and
 reporting burden  estimates to OMB for approval.  At  proposal, these
 estimates were required  to be performed for the first  2 years of the
 regulations.  The  commenter is correct  in recognizing  that  the fifth-year
 impacts would be greater, but the total burden for 220  plants would be
 33 person-years, based on 0.15 person-years per plants.

      It should be  noted  that, since  proposal,   OMB has revised the
 information collection reporting  period from 2 to 3 years, so that EPA
 is required to submit recordkeeping  and reporting burden estimates
calculated over  the first 3 years  of the regulation.   This change, how-
ever,  does not affect the industry burden estimates presented  at proposal
because they were calculated  on  an annual  basis
                                  11-3

-------
11.2  REPORTING

     Comment:  Two commenters (IV-D-22 and IV-D-26)  suggested  eliminating
the following information from the reporting requirements:

     1.  Instrument Identification Number - EPA would need  to  verify
         each daily calibration of the instrument for absolute control
         over instrument use,

     2.  Operator Identification Number - it should  be sufficient  for
         the operator to sign a log or report often-taken  readings,

     3.  Date each repair was attempted if unsuccessful  -  only if  the
         leak was not repairable within the 15 days  allowed should any
         notation be made,

     4.  Repair method - the information EPA derives from  this will  not
         justify the burden on industry, and

     5.  Expected date of repair - in small plants and many large  ones,
         there are no scheduled shutdowns or the date is unknown.

     Similarly, another commenter (IV-D-23) held that the  information
reported should be kept as simple as possible to be  useful.  The
commenter provided an example of information a report should contain:

     1.  The number and size of valves, including relief valves,  found
         leaking, the number repaired and the number of valves that
         were unable to be repaired without shutting down  the  plant.

     2.  The number of leaks found in affected equipment,  the  number of
         valves that were unable to be repaired without shutting  down
         the plant.

     The commenters also thought only leaking components should be
reported.  One commenter (IV-D-19) thought that this information  could
be maintained at the plant and made available for inspection on request
as is done in many other federal and State regulations.   Another  commenter
(IV-D-20) stated that it was unnecessary to report immediately repaired
components because such information does not affect  the emission  reductions
obtained by the inspections.  In addition, exemption of the reporting
requirement would reduce the compliance burden on industry.  One  commenter
(IV-D-22) added that the reporting requirements should be  kept to  a
minimum because they reduce morale by taking time away from more  productive
pursuits, and that the greater the volume of detail  required,  the  lower
the quality of the information.

     Response:  All of these commenters contend that EPA requires  excess
information to be included in the required reports.   In selecting  the
reporting requirements, EPA carefully considered what information  would
be required to assess adequately a plant's compliance with the standards.
The resulting requirements, as proposed, represent the minimum acceptable
reports.

                                  11-4

-------
      For example,  the instrument  number  is  necessary  to  show  that  the
 calibrations  reported are  done  on the  same  instrument  used  for monitoring.
 The number may be  assigned by  the plant,  as instrument 1, 2,  3, etc. or
 may be an actual serial  number.   Similarly, the  operator  identification

      It should be  noted  that,  since  proposal,  OMB has  revised the
 will  serve to indicate if  one  operator detects consistently fewer  (or
 consistently  more)  leaks than  other  operators.   The dates of  unsuccessful
 repair show that repair  was  indeed attempted and, combined with the
 repair method, assist both the  owner or  operator and  EPA  in determining
 which repair  techniques  are  the most successful.  The  expected date of
 repair assists the  appropriate  enforcement  agency in  assessing the
 environmental  impact  of  the  standards  by  quantifying  the  long-term
 emissions that cannot be prevented.

      The EPA  concurs  with  one  commenter  that the required reports  alone
 do  not increase the emission reduction of the  standards.  However,
 they  indicate to EPA  the diligence of  the plant  owner  or  operator  in
 complying with the  standards and  thus  help  EPA judge the  necessary
 frequency of  inspections.   The  EPA also  concurs  with  keeping  the reports
 as  simple as  possible and  has made every  effort  to do  so.  Therefore,
 elaborate reports with extraneous information  are neither required nor
 desired.

 11.3   RECORDKEEPING

      Comment:   Several commenters offered specific examples to be
 deleted from  the recordkeeping requirements.   One commenter (IV-D-15)
 suggested deleting  details such as schematics  and design  specifications
 for  flare and  alarm sensors.  The commenter suggested  that an annual
 certification  of compliance  by the operator would be sufficient.

      The  commenter  wrote that since  plant design criteria are generally
 available through State agencies, EPA, and  at  each gas plant, there
 is no  need  to  include  such information in a log.

      The  commenter  also thought that devices found leaking that are
 repaired  by immediate  action (such as  tightening of packing) should be
 exempt from the recordkeeping requirements  of  the standards.  The
 commenter claimed that such an exemption  would not affect the emission
 reductions  obtained by the inspections and would reduce the compliance
 burden  on  industry.

      Response:  The details of the recordkeeping and reporting require-
 ments  were  discussed  in the response to the previous comment.
     The commenter questions the necessity of maintaining schematics
and design specifications for control devices and closed vent systems
required by Section 60.636(d) of the proposed standards.  The VOC
control effectiveness for compressors, relief devices, pumps, or other
equipment that may be controlled with a closed vent system in lieu of
routine leak detection and repair is dependent on the integrity of the
vent system and control  device.  The vent system and control  device
must be designed and operated properly for the emission reduction of

                                  11-b

-------
 the  standards  to  be  realized.   In order to demonstrate compliance with
 the  standards, the operator must be able to show that the vent system
 and  control device were designed for operation in accordance with the
 requirements of the  standards.  The required design specification and
 schematics will provide the necessary demonstration.  Likewise, the
 design information for the necessary monitoring systems will serve as a
 demonstration of  the proper operating conditions.

     It should be noted that few plants, if any, would purchase either
 control or monitoring equipment without first examining the design
 specifications and flow schematics prepared by the prospective equipment
 vendor.  Consequently, the necessary records are available to the
 operator when the required equipment is installed, and no expenditure
 is necessary to generate these records.

     Records should be kept of all  leaks,  including those repaired
 immediately. The standards allow for quarterly monitoring of valves
 that were found to be nonleakers for 2 consecutive months.  Therefore,
all leakers must be recorded so that monthly monitoring is maintained.
The leaker records may also help the operator identify troublesome
components for replacement, as well  as assessing  the emission reduction
achieved  and the potential for skip  period  or alternative standards.
                                 11-6

-------
                            12.0  MISCELLANEOUS

       Several comments were received that were not related to the other
  topic areas presented in the previous sections.   These comments are
  discussed in this section.  They include comments on determination of
  equivalent means of emission limitation, determination of "major rule "
  technology transfer, plant construction, energy  requirements, effects'
  of other regulations, extension of the public comment period, and
  request for reproposal  of the standards.

       Comment:   One commenter (IV-D-29) claimed that the procedures for
  determining equivalent  means of emission limitations are  impractical  for
  gas processing  plants.   Approval  of equivalence  would take  a year  when
  many gas plants  need to  be designed and  constructed within  2 to  6
  months.   Blanket generic  equivalency approvals,  according to the
  commenter,  should  be allowed since  gas plants  are  relatively small
  primarily  rural  sources.

       Responses:   The EPA  has provided  all of  the  known  blanket  generic
  approvals  of adequately demonstrated control  techniques.  Plant  owners
  or  operators may apply to  EPA any  time  for other alternatives to either
  the  equipment or work practice  standards.  Where "general use" alterna
  tives are approved,  a blanket approval has been  provided.   Other requests
  must be  reviewed,  evaluated, and processed according  to the  procedures
  required by the  Clean Air  Act.

      Comment:  One commenter (II-B-23) remarked that  the guides  in the
  regulation Tor demonstrating equivalence of work practices are unclear as
  to how they would be implemented.

      Response:    Section 60.634 of the proposed standards has been
 revised to simplify the wording of the equivalency determination.  The
 tenn  equivalent" has been changed to "alternative" even though the
                          that *  Pl*nt owner  or operdtor must stn]  show
       ,           .          - -  _ r . — »"""t-iuiufciai,u[iiiu:>iai,iii
      the alternative can achieve a  reduction in VOC emissions that  is
      ast equivalent to the VOC emission reduction achieved  by the
 standard specified  in the regulation.   For example, if an  owner or
           .                  j --_-.„...   . v .  *_^\ VAIM p i \_ ,  ii  U M  UVYM C (  U I
 operator wishes to implement an  alternative program  for reducing valve
  eaks  he must collect,  verify,  and  submit  sufficient  data (e.g., at
 least 12 months of monitoring  data)  to  show that the alternative
 technique(s)  used  to  control  valve  leaks  achieves  the  same or  better
 emission reduction than  the  required  leak detection and repair program.
 The  data and  a written commitment by  the  owner  or operator to  implement
 the  alternative to achieve equivalence  (or  better) would be  submitted
 as an application  to  the Administrator  for  his  consideration.   If the
 Administrator  determines that  the alternative is at least  equivalent to
 the  required  standards, then a notice to  announce the opportunity for a
 public  hearing  would  be published in the  Federal Register.   After an
 opportunity for  a  public hearing, the AdmTTmtrTtoTllSuTd  publish in
 the federal Regj^ter a notice  permitting  the use of the alternative for
 the  purpose of  compliance with the standards.

      Coiranent:    One commenter (IV-D-35) recommended changing  the  language
of Section 60.632-1(c)(2).   The proposed  section reads:  "If the
                                  12-1

-------
 administrator ...owner or operator shall  comply  with  the  requirements
 of....", which, according to  the  commenter,  forced compliance by the
 owner or operator with the equivalent  means  of emission reduction.
 Since the owner or operator may wish to comply with the standards as
 written, the commenter suggested  changing  the wording of  Section
 60.632-l(c)(2)  to read "owner or  operator  may substitute  the requirements
 of that determination."

      Response:   The owner or  operator  may  always elect to comply with the
 standards as written.   Proposed Section 60.632-l(c)(2) has been replaced
 by Section 60.634.

      Comment:   One commenter  (IV-D-20) stated that he believed that it
 was inappropriate for  EPA to  determine that  the  proposed  standards
 are not a major rule.

      Response:   The proposed  standards are not a major rule because
 they do not meet the criteria  outlined in  Executive Order 12291 for a
 major rule.   In the Order a major  rule is  any regulation  that is likely
 to result in:

      (1)   An annual  effect  on  the  economy of $100 million or more;

      (2)   A major increase  in  costs or prices for consumers, individual
           industries,  Federal, State, or local  government agencies, or
           geographic regions;  or

      (3)   Significant  adverse  effects on competition, employment,
           investment,  productivity, innovation, or on the ability of
           the United States-based  enterprises to compete with foreign-
           based  enterprises in domestic or export markets.

      As  stated  in Section  1.3 of this document,  the industrywide net
annual  cost  for  all new, modified, and  reconstructed  facilities  will be
approximately $1.6 million  in  1987.  This will  not result in a major
increase  in  costs or prices and will not create  a significant adverse
effect  on  competition, employment, investment,  productivity, innovation
or  foreign competition.  These costs represent  a  small impact on the
industry  and are  not expected  to deter  construction of gas processing
plants.

      Comment:   One commenter (IV-F-3)  stated  that a monthly leak
detection and repair program does  not  constitute  a "demonstrated tech-
nology" since he was unable to find any industry  or local  regulation
that  has  been effectively implemented  at  such a  frequent  interval.
Another commenter (II-B-23) requested  that "best  demonstrated"  be
clearly defined  in order to state  the  intent  of  the proposed regulations.

      Response:   A "demonstrated technology" is  one which  is effective
in reducing emissions to the atmosphere.   Essex  Chemical  Corp.  v.
Ruckelshaus 486  F. 2d 427 (D.C. Cir. 1973); Portland  Cement Ass'n  v.
Ruckelshaus 486  F. 2d 375 (D.C. Cir. 1973).  Control  technology  can be
considered "best demonstrated" technology  (BDT)  if it can  be shown  to
be the best system demonstrated for the category  of sources, not

                                  12-2

-------
 necessarily on the category of sources.  The EPA's interpretation of
 demonstrated technology means that a technique used in an entirely
 different industry using a different process from the one being regulated
 can be BDT if its performance would not be affected by the differences
 in the sources.

      Leak detection and repair programs are currently implemented in
 over 100 petroleum refineries throughout the country.  These programs
 have been demonstrated by industry as a workable, effective technology
 to reduce VOC equipment leaks from refineries.   The EPA and refinery
 personnel have shown that substantial emission reductions are possible
 through on-line repair for all valves.   These programs are basically a
 reflection of the refinery control technique guideline document issued
 by EPA in 1978 (EPA-450/2-78-036, OAQPS No. 1.2-111).  State and Regional
 air pollution control  agencies have followed those guidelines to issue
 regulations  to achieve and maintain the National  Ambient Air Quality
 Standards for ozone.

      The EPA recognizes that there are  differences between refineries
 and gas plants;  however,  these differences  do not preclude the  transfer
 of control  technology  to  the gas  processing industry.   In API testing
 of the natural  gas processing industry, the process  type, operating
 temperature  and  pressure,  and line size were determined  to be unrelated
 to the frequency  and magnitude of equipment leaks.

      The frequency of  the  leak detection and  repair  programs  followed
 by existing  programs varies.   Most State Implementation  Plans require
 quarterly inspections  of gas  service  valves and annual  inspections of
 light  liquid  service valves,  whereas  the Air  Quality  Management  Districts
 in California  require  less  frequent (annual)  valve  inspections,  while
 requiring more stringent follow-up inspections on  leaking valves.  The
 frequency of  leak  detection  and repair  required in the  standards  is
 considered reasonable  as discussed in other responses  to  comments and
 is based  on valve  leak  occurrence, leak  recurrence, costs,  and  emission
 reductions.

     Comment:  One commenter  (IV-D-20)  stated that increased  emphasis
 will be  placed on  extended gas gathering  systems  in the  future,  resulting
 in the  construction of  fewer  plants than  estimated by EPA.   Extending
 gas gathering systems would maintain  the  throughput at existing  plants
 and would be done  in preference to  building new plants due  to economic
 pressures, according to the commenter.

     Response:  While it is true  that some  companies may  elect to extend
 gas gathering systems to continue  operation of existing plants, many
 new plants will be constructed.   A recent article in the  Oil and Gas
 Journal states that 21  new U.S. plants were built  in  1983, with 20
 being built in 1984.   Since new plants are often more energy efficient
 than older plants, and  older plants may be located too far from new
 fields, EPA expects the industry  to continue to replace older plants
with new plants.   Alternately, older plants may be revamped with new
processes or new equipment  and become subject to the NSPS requirements
through modification  and reconstruction  provisions.


                                  12-3

-------
      Comment:   One  commenter  (II-B-23) disagreed with EPA's claims that
 the  proposed  standards do not  require energy.  The commenter stated
 that energy  is  required  for maintaining pressure on compressor barrier
 seals,  maintaining  vacuum on  the closed-vent systems, maintaining a
 pilot flare,  transportation of monitoring equipment by employees, and
 additional use  of office equipment  for increased administrative duties.

      Response:   API  submitted  this  comment prior to proposal and requested
 that it  be included  in the rulemaking (IV-F-le).  Although the proposed
 standards state  that, in general, the required controls do not require
 energy,  EPA acknowledges that energy is also required for standards
 requiring add-on control devices for vented process streams in addition
 to some  of the  examples cited by the commenter.  However, the increase
 in energy for these  requirements will be minimal.  Furthermore, the
 effect  of the final  standards will  be to maximize natural gas production
 for  a given amount  of raw material  (wet gas), resulting in a net energy
 savings.  Therefore, the standards  potentially save energy rather than
 expend  energy.

      Comment:   One  commenter  (IV-D-12) stated that certain provisions
 of the  proposed standards are in direct conflict with the requirements
 of other Federal agencies with jurisdiction over his facilities.  The
 commenter indicated  that enclosure  of compressor distance pieces creates
 an explosion hazard  and is a violation of Department of Transportation
 rules.  The commenter claimed that, since his plant was located on U.S.
 Bureau of Land Management property, use of a flare is specifically
 prohibited.

      Response:  The  EPA knows of no requirements in the standards that
 are  in conflict with other Federal  standards.  The Department of
 Transportation regulations apply to pipeline transmission of natural
 gas  and other petroleum products and do not specify requirements for
 in-plant compressors.  The Bureau of Land Management does regulate
 flares burning in some wilderness areas, but flares represent only one
 control device option available to  the plant owner or operator.
     Comment:  Nine commenters (IV-D-1, IV-D-4, IV-D-5, IV-D-6, IV-D-7,
IV-D-8, IV-D-9, IV-F-2, IV-F-3) requested that the comment period be
extended to allow more time to review the proposed standards, to collect
additional  data to enable commenters to judge the impact of the proposed
standards, or to submit more detailed comments.

     Response:  In response to the commenters1 requests, EPA extended
the comment period by 60 days.  The extension notice was published in
the Federal Register (49 FR 13392, April 4, 1984).  The new deadline
for comments was changed to June 6, 1984, to allow time for industry
representatives to obtain additional data and cost estimates for
implementing VOC leak detection and repair programs as well as for
using flares to control compressor seal leakage.

     Comment:  One commenter (IV-D-32) requested  that the regulation be
reproposed  if LPA finds the standards to be justified.  According to
the commenter, reproposal  would afford an accommodation of changes in
the proposed regulation anticipated as a result of comments.

                                  12-4

-------
     Response:  Reproposal  is not necessary, since the purpose of the
public comment period is to allow for changes in the proposed standards.
The comments received have been considered carefully, and as a result
many changes have been made to the standards as discussed in Section
1.2, "Summary of Changes Since Proposal".  The need for national  standards
has been documented in Section 2.  Because State requirements vary, the
national standards are needed to prevent States with lenient air  pollu-
tion requirements from attracting new industries that might deteriorate
the State's air quality, as well as to prevent migration of pollutants
from State to State.

-------
                                    TECHNICAL REPORT DATA
                             (Please read Instructions on the reverse before completing)
  I. REPORT NO.
  EPA-450/3-82-Q24h
                               2.
                                                             3. RECIPIENT'S ACCESSION NO.
  . TITLE AND SUBTITLE
  Equipment Leaks of  VOC  in  Natural  Gas Production
  Industry - Background  Information  for Promulgated
  Standards
                                5 REPORT DATE
                                      May 1985
                                6. PERFORMING ORGANIZATION CODE
 7. AUTHOR(S)
                                                             8. PERFORMING ORGANIZATION REPORT NO
 _. PERFORMING ORGANIZATION NAME AND ADDRESS
 Office of Air Quality Planning  and Standards
 U.S.  Environmental Protection Agency
 Research Triangle Park, NC   27711
                                10. PROGRAM ELEMENT NO.
                                11. CONTRACT/GRANT NO.

                                   68-02-3060
 12. SPONSORING AGENCY NAME AND ADDRESS
 Director for Air Quality  Planning and Standards
 Office  of Air, Noise, and Radiation
 U.S.  Environmental Protection Agency
 Research Triangle Park, NC   27711
                                13 TYPE OF REPORT AND PERIOD COVERED
                                14. SPONSORING AGENCY CODE
                                  EPA/200/04
    UPPLEMENTARY NOTES
     .                 ---is document  presents the background  information used by the
 Environmental  Protection Agency  in developing the promulgated  new source performance
 standards  for equipment leaks of VOC  in the natural gas production industry
      Standards  of performance  for  the control of volatile organic compound (VOC) equip-
 ment leaks  from the natural gas production industry are being  promulgated under Section
 111 of the  Clean Air Act.  These standards will apply to equipment leaks of VOC within
 new, modified,  and reconstructed gas  plant compressors and process units.  This
 document  summarizes the responses  to  public comments received  on  the proposed standards
 and the basis for changes made in  the standards since proposal.
                                KEY WORDS AND DOCUMENT ANALYSIS
                  DESCRIPTORS
 Ai r Po 11 uTTorT
 Fugitive  Emissions
 Natural  Gas Production
 Pollution Control
 Standards of Performance
 Volatile  Organic Compounds
(VOC)
                                               b IDENTIFIERS/OPEN ENDED TERMS
                    Air Pollution Control
                                                                            COSATI I-tcld/Group
      13b
Unlimited -  Available to  the  public free of
charge  from  U.S.  EPA Library  (MD-35)
Research  Triangle Park, NC  27711
                  19 SECURITY CLASS (This Report)
                     Unlimited
                  20 SECURITY CLASS /This page>
                     Unlimited
21. NO. OF PAGES
    122
22. PRICE
EPA Form 2220-1 (Rev. 4-77)   PREVIOUS EDITION i s OBSOLE Tt

-------