United States
Environmental Protection
Agency
Office of Air Quality
Planning and Standards
Research Triangle Park NC 27711
EPA-450/3-83-015b
October 1983
Air
Petroleum Fugitive  EIS
Emissions—       450383015b
Background
Information for
Promulgated
Standards

-------

-------
                                EPA-450/3-81-015b
Petroleum Fugitive Emissions-
     Background Information
   for Promulgated  Standards
        Emission Standards and Engineering Division
        U.S. ENVIRONMENTAL PROTECTION AGENCY
           Office of Air, Noise, and Radiation
        Office of Air Quality Planning and Standards
        Research Triangle Park, North Carolina 27711

                October 1983

-------
This report has been reviewed by the Emission Standards and Engineering Division of the Office 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, N.C.  27711, or 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 in Petroleum Refineries
                               Prepared by:
JacXR. Farmer                                        f  (Date)
Director, Emission Standards and Engineering Division
L)jS. Environmental Protection Agency
Research Triangle Park, North Carolina  27711

1.   The promulgated standards of performance will  limit emissions  of  VOC
     from equipment leaks in new, modified, and reconstructed petroleum
     refinery 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 welfare.

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:

     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

-------
IV

-------
                           TABLE OF CONTENTS
Title                                                           Page

1.0  SUMMARY	1-1
     1.1  Summary of Changes Since Proposal  	  1-1
     1.2  Summary of Impacts of Promulgated Action	1-6
     1.3  Summary of Public Comments	1-8
2.0  STANDARDS	2-1
     2.1  General Discussion	2-1
     2.2  Valves	2-10
     2.3  Pumps	2-35
     2.4  Compressors	2-44
     2.5  Pressure Relief Devices 	  2-46
     2.6  Sampling Systems	2-51
     2.7  Open-Ended Lines	2-53
     2.8  Flanges, Liquid Service Relief Valves, and
          Heavy Liquid Service Valves and Pump Seals	2-54
     2.9  Control Devices	2-56
3.0  APPLICABILITY	3-1
     3.1  Affected Facility 	  3-1
     3.2  Definition of "In VOC Service"	3-8
     3.3  Exclusions	3-12
     3.4  Small Refiners. . .  .	3-15
4.0  MODIFIED SOURCES 	  4-1
     4.1  Emission Increase 	  4-1
     4.2  Capital Expenditures	4-3
     4.3  Small Facilities	4-3
5.0  RECONSTRUCTION	5-1
6.0  LEGAL	6-1
7.0  TEST METHODS	7-1
8.0  RECORDKEEPING AND REPORTING	8-1

-------
                     TABLE OF CONTENTS (concluded)
                                                               Page


APPENDIX A - Incremental  Cost Effectiveness of
             Control Techniques for Equipment
             Leaks of VOC	      A-l

APPENDIX B - Regulatory Decisions Affecting Standards
             for SOCMI	      B-l

APPENDIX C - Evaluation of Available Equipment Leak  Data .  .      C-l

APPENDIX D - Model Unit and Nationwide Impacts	      D-l
                                  vi

-------
                             LIST OF TABLES

Title
1-1   Summary of Individual Component Impacts ........     1-7
1-2   List of Commenters On Proposed Standards of
      Performance for Fugitive Emission Sources in the
      Petroleum Refining Industry ............ • •     1-9
2-1   Comparison of CTG Recommendations and NSPS
      Requirements ................. ....     2-3
2-2   Summary of Individual Component Impacts ........     2-5
2-3   Projected VOC Fugitive Emissions from Facilities for
      1982-1986 Under Uncontrolled, Baseline, and NSPS ...     2-7
2-4   Revised Emission Reductions and Costs for Leak
      Detection and Repair Programs .............     2-13
2-5   Valve Leak Detection and Repair Cost Estimates ....     2-18
2-6   Derivation of Average Component Monitoring Time.  .  .  .     2-30
A-l   Summary of the Individual Component Control
      Impacts ........................     A- 3
A-2   Pressure Relief Device Impacts  ............     A-4
A-3   Compressor Seal Impacts ................     A-7
A-4   Open-ended Lines  Impacts .............  •  •     A- 8
A-5   Sampling Connection  System Impacts  ..........     A- 9
A-6   Valve Emissions and  Emission  Reductions ........     A-10
A-7   Valve Leak Detection and Repair Costs .........     A-ll
A-8   Sealed Bellows Valve Cost  Impacts ...........     A-12
A-9   Cost Effectiveness of Valve Controls  .........     A-13
A-10  Pump Emissions and Emission Reductions  ........     A-14
A-ll  Pump Leak Detection  and  Repair  Costs  .........     A-15
A-12  Dual Mechanical Seal System Costs  for  Pumps ......     A-17
A-13  Cost Effectiveness of Pump Controls ..........     A-18

-------

-------
                               1.0  SUMMARY

      On January 4, 1983, the U.S. Environmental  Protection Agency (EPA)
 proposed standards of performance for fugitive emission sources of
 volatile organic compounds (VOC)  in the petroleum refining industry
 (48 FR 279) under the authority of Section 111 of the Clean Air Act.
 Public comments were requested  on the proposed standards in the Federal
 Register, and 24 commenters responded.  Most of  the commenters  represented
 refining companies or industry  associations.  Other commenters  included
 an environmental  group,  the Department of the Interior, and vendors of
 equipment used  to control  fugitive emissions. This summary of  comments
 and EPA's responses  to these comments serve  as the  basis for the  revisions
 made to the applicability  and the requirements of the proposed  standards.
 1.1  SUMMARY OF CHANGES  SINCE PROPOSAL
      The proposed standards  were  revised  as  a result  of reviewing
 public  comments.   The  major  revisions concern the following:
     •    Leak  Detection and  Repair for Refineries  Located  in the
           North  Slope  of Alaska
     •    Alternative  for  Determining a "Capital  Expenditure"
     •    Clarification of Reconstruction  Provisions
     •    Provision  for Difficult-to-Monitor  Valves in  New  Units
     •    Exemptions for Compressors
     •    Addition of  Reporting Requirements
     •    Open-ended Lines on Double  Block and Bleed  Valves
 1-1.1   Leak  Detection  and Repair  for  Refineries Located  in the North
        Slope of Alaska~"
     Since  proposal, EPA has  reviewed comments concerning refining
 operations  in the North Slope of Alaska and determined 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

                                  1-1

-------
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.  EPA excluded these plants only from the
routine leak detection and repair requirements because the costs of
the other requirements are reasonable.
1.1.2  Alternative for Determining a "Capital Expenditure"
     The General Provisions (Subpart A) of 40 CFR Part 60, require that
increases in emissions of a pollutant covered by applicable standards
trigger the application of standards of performance for existing facilities,
These increases make a source covered by standards a modified source,
as set forth in Section 111 of the Act.  EPA has interpreted Section
111 so that production rate increases accomplished without a capital
expenditure do not trigger these provisions even though they might be
accompanied by an increase in emissions (See 40 CFR 60.14(e)(2)) Capital
expenditure is defined in 40 CFR 60.2.  In the proposed standards, EPA
also excluded increases in emissions resulting from process improve-
ments accomplished without a capital expenditure from being considered
a modification.  The intent was to exclude minor changes in operations
as indicated by changes not accompanied by a capital expenditure.
     The annual asset guideline repair allowance (AAGRA) and the original
cost basis are used to define capital  expenditure (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 changes in 1982.  In
response to the comments concerning the difficulties of using the AAGRA
and the original  cost basis, EPA is providing in the standards for
equipment leaks an alternative procedure for determining capital
expenditure.  The purpose of the alternative is to make the determination
of capital  expenditure more practicable, yet maintain the original
intent of the definition.  This alternative provides that a capital
expenditure would be incurred  if actual  costs exceed the product, P, of
the existing facility's (that  is, the  equipment covered by the standards)
replacement cost A, the AAGRA  basis and an inflation factor,  Y,  as shown
in the following equation:
                                  1-2

-------
           P  =  A  x  Y x 0.07,  where
           A  =  existing  facility replacement cost,
           Y  =  the  percent of the present replacement cost which is
               equivalent to the original cost,
             =  1.0  - 0.575 log (X), and
           X  =  the  year  of construction.
 1.1.3  Clarification of Reconstruction Provisions
     The provisions for reconstruction (40 CFR 60.15) imply that costs
 are accumulated  over an unlimited time period.  Commenters, however,
 objected to  a  continuous accumulation of costs because refineries are
 continually  replacing components.  To clarify the application of Section
 60.15, EPA is  defining  "proposed replacement" under this standard to
 include components which are replaced pursuant to all continuous programs
 of component replacement which commence (but are not necessarily completed)
 within a 2-year  period.  Thus, EPA will count toward the 50 percent
 reconstruction threshold the "fixed capital cost" of all depreciable
 components replaced pursuant to all continuous programs of reconstruction
 which commmence  within  any 2-year period following proposal of these
 standards.
     EPA is  further clarifying the intent of the reconstruction provisions
 based on comments concerning routine equipment replacement.  In response
 to these comments, EPA  is clarifying that certain routine replacements
 are not considered in the basis for reconstruction.  The routine replace-
ments excluded by the final  standards from reconstruction are valve
 packings, pump seals, nuts and bolts, and rupture disks.  Replacement
of equipment pieces, such as valves and pumps, at turnaround or at
other times must be included when considering whether a reconstruction
will  take place.
 1.1.4  Provision for Difficult-to-Monitor Valves in New Units
     At proposal, there was  no exemption for difficult-to-monitor valves
in new units, although difficult-to-monitor valves were exempt from
 routine monitoring in units  covered through the modification or
reconstruction provisions.  Commenters argued that while the number of
difficult-to-monitor valves  can  be substantially reduced in number for
new units,  they cannot be totally eliminated.  Upon reviewing the
                              1-3

-------
comment letters, EPA decided to permit an allowable percentage of
valves in a new unit to be designated as difficult-to-monitor.  Based on
existing units, about 3 percent of the total  number of valves may be
impossible to eliminate without additional  costs.   Therefore, EPA is
allowing up to 3 percent of total  numbers of valves to be treated as
difficult-to-monitor valves for new units.
1.1.5  Exemptions for Compressors
     Commenters were concerned that process streams with a high
hydrogen content would be subject  to the standards.  The commenters
contended that such streams would  have a lower percentage of VOC and,
consequently, the controls required by the  proposed standards would
achieve lower emission reductions  and have  a higher cost effectiveness
($/Mg of VOC emission reduction).
     Upon analyzing the cost effectiveness  of valves and compressors
in hydrogen service (greater than  50 volume percent hydrogen) (Document
Reference No. IV-B-9), EPA determined that  significant emission reduc-
tions are achieved for valves in hydrogen service  at a reasonable cost
($!06/Mg VOC).  However, control of compressors in hydrogen service
results in a cost effectiveness of $4,600/Mg VOC.   EPA, therefore,
decided to exempt these compressors from the standards.
     Commenters also implied that  EPA had provided an exemption from
the standards for existing compressors.  EPA provided no blanket exemp-
tion in the proposed standards even though  EPA discussed that certain
reciprocating compressors might not be covered under the reconstruction
provisions if retrofitting the required equipment  was technologically
or economically infeasible (See 40 CFR 60.15(e)).   To make EPA's intent
clear and to reduce the burden of  reviewing reconstruction determinations,
EPA is explicitly exempting reciprocating compressors that become
affected by the standards through  40 CFR 60.14 or  60.15 from the stan-
dards for compressors provided the owner or operator demonstrates that
recasting the distance piece or replacing the compressor are the only
options available to bring the compressor into compliance.  If an owner
or operator is replacing a compressor or recasting the distance piece
for some other reason than to reduce emissions and comply with the
standards or if these actions occur later,  then a  modified or recon-
structed compressor would not be exempt from the standards.
                                  1-4

-------
 1.1.6   Addition  of  Reporting  Requirements
     The  proposed standards did  not  require  routine  reporting.  The
 preamble  to  the  proposed  standards addressed three alternative  levels
 of  reporting  requirements.  The  alternative  of no routine  reporting
 was selected  because  State or  local  agencies, who usually  are delegated
 the responsibility  for enforcement of the  standards, could  require
 routine reporting.
     In response to comments on  the  enforceability of the  standards and
 comments  on the need  for  routine reporting,  EPA decided to  require
 routine reporting in  these standards of performance  rather  than relying
 on  individual State requirements.  Compliance with the leak detection
 and  repair program and equipment requirements will be assessed through
 semiannual reports, review of  records, and by inspection.   The semi-
 annual  reports provide a  summary of  the data recorded on leak detection
 and  repair of valves, pumps, and other equipment types.  Notifications
 are  still  required as described  in the General  Provisions  for new
 source standards (40  CFR  60.7).  However, the semiannual reports may be
 waived for affected facilities in States where the regulatory program
 has  been delegated, if EPA, in the course of delegating such authority,
 approves  reporting requirements or an alternative means of  source
 surveillance adopted  by the State.   In these cases, such sources would
 be  required to comply with the requirements adopted by the  State.
 1.1.7  Open-Ended Lines on Double Block and Bleed Valves
     The proposed standards required all  open-ended lines or valves
 to be capped except when they are being used.  In some cases, however,
 open-ended valves are installed in a "double block-and-bleed" arrangement
 such that  emissions must occur to the atmosphere through the open end
 of the  bleed valve.    In such cases, the open end of the bleed valve was
 not  required to be capped because they are used to vent the line between
 the  block  valves.  However, when the bleed valve is not open, then it
must be capped.  This was not as clear as it could have been in  the
 proposed standards and, therefore,  a specific provision has been
 added to the standards.
                                 1-5

-------
 1.2   SUMMARY  OF  IMPACTS OF PROMULGATED ACTION
 1.2.1  Alternatives to Promulgated Action
      The  regulatory alternatives discussed in Chapter 6 of the BID for
 the proposed  standards generally reflect the different levels of emission
 control.  They were used to help in selection of the best demonstrated
 technology, considering costs and nonair quality health, environmental,
 and economic  impacts for fugitive emission sources in the petroleum
 refining  industry.  These alternatives remain the same; however, the
 costs, emission  reductions, cost effectiveness, and incremental  cost
 effectiveness for the various levels of control included in the regulatory
 alternatives, which were estimated and then summarized in the preamble
 to the proposed  standards, have been reevaluated and are now summarized
 on a per  component basis as presented in Table 1-1.  These estimates
 served as the basis for determining the impacts of the standards.
 Model Unit and nationwide impacts of the promulgated standards are
 documented in Appendix D.
 1.2.2  Environmental Impacts of Promulgated Action
     Environmental impacts of the proposed standards are described in
 48 FR 279.  The  revisions to the applicability and provisions of the
 proposed  standards (described in Section 1.1) will  have a minimal
 effect on the environmental  impacts of the standards.
 1.2.3  Energy and Economic Impacts of Promulgated Action
     The  energy and economic impacts of the standards  are described in
 Chapters 8 and 9 and Appendix F of the BID for the proposed standards.
 In general, there has been little change in these impacts since  proposal.
      The nationwide cost impacts reported in the preamble to the
 promulgated standards are lower than the impacts reported at proposal.
At proposal, the nationwide  cost impacts were based on  refineries  not
 subject to State or regional  regulations to control  equipment leaks of
VOC.   However, the nationwide impacts  have subsequently been  revised
 (as shown in Appendix D)  based upon the baseline control  costs to
comply with existing regulations for equipment leaks of VOC.
1.2.4  Other Considerations
     1.2.4.1  Irreversible and Irretrievable  Commitment of Resources.
Section 7.6.1 of the BID  for the proposed standards concluded that  the
                                   1-6

-------
                                 Table 1-1.   SUMMARY OF THE  INDIVIDUAL  COMPONENT  CONTROL  IMPACTS6
 I
—I
Fugitive Emission
Source
Pressure relief devices


Compressors

Open-ended valves
Sampling connection
systems
Valves


Pumps




Control Technique
Quarterly LDR
Monthly LDR
Rupture disks0"
Controlled degassing
vents
Caps on open ends
Closed purge sampling

Quarterly LDR
Monthly LDR
Sealed bellows valves
Annual LDR
Quarterly LDR
Monthly LDR
Dual mechanical seal
system
Emission Reduction
(Mg/yr)
4.4
5.3
9.8
16.5

2.8
2.6

66
77
110
3.0
9.8
11.5
13.9

Average Cost
Effectiveness
($/Mg)b
(170)
(110)
410
150

460
810

(110)
(60)
4,700
860
157
158
2,000

Incremental Cost
Effectiveness
($/Mg)c
(170)
250
1,000
150

460
810

(110)
310
16,700
860
(140)
170
10,900

                      (xx)  = Cost  savings
                      LDR = Leak detection and repair

                      aCosts and emission reductions are based on fugitive emission  component counts in Model  B from the BID
                       for the  proposed standards,  EPA-450/3-81-015a, page 6-3, and  from  Tables A-2 through A-13 of Appendix A.

                      ^Average  Cost Effectiveness = net annualized costs per component  +  annual VOC emission reduction  per
                       component.

                      clncremental Cost Effectiveness  = (net annualized cost of the  control technique - net annualized  cost
                        of the  next less restrictive control technique) •» (annual  emission  reduction of control  technique -
                        annual  emission reduction of the next less restrictive control  technique).

                      dllnderlined  control techniques were selected as basis for standards.

-------
standards will  not result in any irreversible or irretrievable  commit-
ment 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.
      1.2.4.2  Environmental and Energy Impacts of  Delayed Standards.
Table F-ll of the BID for the proposed standards summarizes the
environmental and energy impacts associated with delaying promulgation
of the standards.  The emission reductions and associated energy  savings
shown would be irretrievably lost at the rates shown for each of  the
5 years.
1.3  SUMMARY OF PUBLIC COMMENTS
     Letters were received from 24 commenters commenting on  the
proposed standards and the BID for the proposed standards. There was
one request for a public hearing, however, these commenters were  request-
ing a meeting with EPA for clarification of the proposed standards.
Minutes of this meeting are contained in the project docket.   A list
of commenters, their affiliations, and the EPA docket number  assigned
to their correspondence is given in Table 1-2.
     The comments have been categorized under the following  topics:
          Standards  (Section 2)
          Applicability (Section 3)
          Modification  (Section 4)
          Reconstruction  (Section 5)
          Legal  (Section 6)
          Test Methods  (Section 7)
          Recordkeeping and Reporting  (Section 8)
          Appendix A -  Incremental Cost Effectiveness of Control  Techniques
                        for  Equipment Leaks of VOC
          Appendix B -  Regulatory Decisions Affecting Standards
                        for  SOCMI
          Appendix C -  Evaluation of Available Equipment Leak Data

          Appendix D -  Model  Unit and  Nationwide Impacts
                                   1-8

-------
   TABLE  1-2.   LIST OF COMMENTERS ON PROPOSED STANDARDS OF PERFORMANCE
                        FOR FUGITIVE EMISSION SOURCES
                        IN THE PETROLEUM REFINING
	INDUSTRY	
COMMENTER AND  AFFILIATION	      DOCKET ITEM NO.

1.  Mr.  B.T. McMillan                                  IV-D-3
    Allied Chemical
    P.O. Box 1053R
    Morristown, NJ  07960

2.  Mr.  C.H. Barre                                     IV-D-4
    Marathon Petroleum Company
    Findlay, Ohio  45840

3.  Mr.  R.E. Farrell                                   IV-D-5
    Standard Oil Company of Ohio
    Midland Building
    Cleveland, Ohio  44115

4.  Mr.  A.H. Nickolaus                                 IV-D-6
    Texas Chemical Council
    100  Brazos, Suite 200
    Austin, TX  78701-2476

5.  Ms.  Geraldine V. Cox                               IV-D-7
    Chemical Manufacturers'  Association
    2501 M. Street, North West
    Washington, DC  20037

6.  Mr.  Robert N. Harrison                             IV-D-8;
    Western Oil and Gas Association                    IV-D-8a;
    727 West Seventh Street                             IV-D-1;
    Los Angeles, CA  90017                             IV-D-2

7.  Mr. James A. Young                                 IV-D-9
    Independent Refiners'  Association
    900 Wilshire Boulevard,  Suite 1024
    Los Angeles, CA  90017

8.  Mr. Phillip L. Youngblood                           IV-D-10
    Conoco, Inc.
    Suite 2136, Post Office  Box 2197
    Houston, TX  77252

9.  Mr. Roger Noble                                    IV-D-11
    John Zink Company
    4401 S.  Peoria
    Tulsa,  OK  74105
                                   1-9

-------
                TABLE  1-2.   LIST  OF  COMMENTERS  ON  PROPOSED
                         STANDARDS OF  PERFORMANCE
                         FOR FUGITIVE  EMISSION  SOURCES
                         IN  THE PETROLEUM REFINING
                            INDUSTRY (Continued)
 COMMENTER AND AFFILIATION                        DOCKET  ITEM  NO.
10.  Mr.  Herman A.  Fritschen                            IV-D-12
     Cities Service Company
     Post Office Box 300
     Tulsa, OK  74102

11.  Mr.  Alan J. Schuyler                               IV-D-13
     ARCO Alaska, Inc.
     Post Office Box 360
     Anchorage, Alaska   99510

12.  Mr.  Paul M. Kaplow                                  IV-D-14
     Atlantic Richfield Company
     Post Office Box 2679-T.A.
     Los  Angeles, CA 90051

13.  Mr.  William F. O'Keefe                              IV-D-15
     American Petroleum Institute
     201  L Street,  Northwest
     Washington, DC  20037

14.  Mr.  A.G. Smith                                      IV-D-16
     Shell Oil Company
     One Shell Plaza
     Post Office Box 4320
     Houston, TX  77210

15.  Mr.  J.D. Reed                                       IV-D-17
     Standard Oil Company (Indiana)
     200 East Randolf Drive
     Chicago, IL  60601

16.  Mr.  J.J. Moon                                       IV-D-18
     Phillips Petroleum Company
     Bartlesville, OK  74004

17.  Mr. Louis R. Harris                                 IV-D-19
     B S & B Safety Systems, Inc.
     7455 East 46th Street
     Post Office Box 45590
     Tulsa, OK  74145

18.  Mr. R.H. Murray                                     IV-D-21
     Mobil Oil Corporation
     3225 Gallows Road
     Fairfax, VA  22037
                                  1-10

-------
                TABLE 1-2  LIST OF  COMMENTERS  ON  PROPOSED
                         STANDARDS  OF  PERFORMANCE
                         FOR FUGITIVE  EMISSION SOURCES
                         IN THE PETROLEUM  REFINING
                            INDUSTRY  (Concluded)
 COMMENTER AND AFFILIATION	DOCKET  ITEM NO.

19.  Mr. M.W. Anderson                                   IV-D-22
     Kerr-McGee Corporation
     Kerr-McGee Center
     Oklahoma City, OK  73125

20.  Mr. J.H. Leonard                                    IV-D-23
     Beacon Oil Company
     525 West Third Street
     Hanford, CA  93230

21.  Mr. William L. Rogers                               IV-D-24
     Gulf Oil Corporation
     Post Office Drawer 2038
     Pittsburgh, PA  15230

22.  Mr. Joseph M. Macrum                                IV-D-25
     Texaco U.S.A.                                       IV-D-25a
     Post Office Box 52332
     Houston, TX  77052

23.  Mr. Bruce Blanchard                                 IV-D-26
     U.S. Department of the  Interior
     Washington, DC  20240

24.  Mr. David D. Doniger                                IV-D-30
     National Resources Defense Council,  Inc.
     1725 I Street, N.W.
     Washington, D.C.  20006
                                   1-11

-------

-------
                              2.0   STANDARDS

     This  chapter  summarizes  public comments and responses to comments
 pertaining to  the  proposed  standards.  Section 2.1 presents those
 comments and responses that pertain to the standards in general;
 Sections 2.2 through 2.8  present comments and responses that pertain to
 particular requirements for each piece of equipment covered by the
 standards; and Section 2.9  presents comments and responses on control
 devices.   In Chapter 2 and the following chapters, information used in
 responding to the  public  comments  is referenced according to document
 number within the  project docket,  Docket No. A-80-44.
 2.1  GENERAL DISCUSSION
 Comment:
     Commenters (IV-D-5,  IV-D-12,  and IV-D-25) wrote that the standards
 should be  the same as State requirements.  Commenters (IV-D-22 and
 IV-D-24) argued that the  proposed  standards would be redundant and
 conflict with existing State regulations.  For example, States may
 presently  require  leak detection and repair of compressor seals while
the NSPS would require equipment specifications.   One commenter (IV-D-12)
 thought the standards would require the abandonment of existing programs.
 Response:
     The Clean Air Act Amendments of 1977 require each State in which
there are areas where the national  ambient air quality standards (NAAQS)
 are exceeded to adopt and submit revised State implementation plans
 (SIP's) to EPA.  Sections 172(a)(2) and (b)(3)  of the Clean Air Act
 require that nonattainment area SIP's include reasonably available
control technology (RACT) requirements  for stationary sources.   EPA
 issues Control  Techniques Guidelines (CTG) documents to provide State
and local air pollution control agencies with an  initial  information
base for proceeding with their own assessment of  RACT for specific
stationary sources.
                                  2-1

-------
     Most State regulations for fugitive VOC emissions are based on the
"Control of Volatile Organic Compound Leaks for Petroleum Refinery
Equipment," EPA-450/2-78-036, released by EPA in June 1978 (Document
No. II-A-6).  This CT6 was issued by EPA to provide information and
guidance to State and local air pollution control agencies for their
use in regulating VOC emissions in oxidant nonattainment areas.  The
CTG identifies RACT that can be applied to existing refineries to control
VOC from equipment leaks.
     The Clean Air Act requires that standards of performance for
stationary sources reflect the degree of emission limitation achievable
through application of the best adequately demonstrated technological
system of continuous emission reduction (best demonstrated technology,
BDT), taking into consideration the cost of achieving such emission
reduction, any nonair quality health and environmental  impacts, and
energy requirements.  NSPS applies to newly contracted, modified, or
reconstructed facilities in both attainment and non-attainment areas.
Because the purpose of the NSPS and the purpose of the State regulations,
as reflected in the Clean Air Act, differ EPA believes  that it would be
inappropriate for the requirements to be necessarily the same.
     The standards are not redundant, and no substantial  conflict
occurs between the NSPS and State requirements.  The NSPS requirements
and the CTG recommendations are identified and compared in Table 2-1.
The NSPS require monthly/quarterly leak detection and repair of gas and
light liquid valves and monthly leak detection and repair for light
liquid pumps, while the CTG recommends less frequent leak detection and
repair (quarterly leak detection and repair for gas valves and yearly
leak detection and repair for light liquid valves and pumps).   The
increased monitoring frequencies of the NSPS are reasonable because the
incremental  cost and emission reduction are reasonable.
     The standards for pressure relief devices and compressors  are  based
on the use of equipment,  whereas the CTG recommends leak  detection  and
repair for pressure relief devices and compressors.   The  CTG  includes
no recommendation for sampling connections.   The standards require
equipment and work practices  for sampling  connections;  again,  there is
no conflict  between the NSPS  and State requirements.  The standards
                                  2-2

-------
                      Table 2-1.  COMPARISON OF CTG RECOMMENDATIONS AND NSPS REQUIREMENTS
                                     CTG
                                           NSPS
  Source
 Routine      Equipment
Inspection  Specification/
Interval     Work  Practice
 Routine      Equipment
Inspection  Specification/
Interval    Work Practice
Valves
Gas/Vapor
Light Liquid
Open-ended Lines
(purge, drain,
sample lines)
Sampl i ng
Connections
i Pump Seals
00 Light Liquid
Relief Devices
Compressor Seals
Quarterly
Annual
None
None
Annual9
Quarterly
Quarterly
None
None
Caps
None
None
None
None
Monthly3
Monthly3
None
None
Monthly13
None
None
None
None
Caps
Closed purge
None
Rupture disksc
Closed vent to control
device
dThe standards require monthly monitoring of valves, except that quarterly monitoring would be allowed
 for valves which have been found not to leak for two successive months.
     pumps, instrument monitoring would be supplemented with weekly visual  inspections for  .
 liquid leakage.  Pumps with evidence of liquid leakage are to be monitored and if emission concentrations
 are 10,000 ppm or more they must be repaired
cStandard is "no detectable emissions."

-------
also provide alternatives for valves that allow an owner or operator
to continue the CTG control  level for process units with less than
2.0 percent of the valves leaking.  Other alternatives are allowed for
valves, and three alternatives are allowed for pumps.   EPA does not
believe that the standards would conflict with most existing State
regulations.  The standards  apply only to affected facilities (compressors
are one of the facilities covered by the standards) that commence
construction or modification after January 4, 1983.  A conflict may
occur if State agencies require leak detection and repair.   The conflict
is requiring unneeded leak detection and repair for compressors equipped
with the controls required by the standards.   Most State air pollution
control regulations include  variance procedures that owners or operators
can assess.  In the few cases where it occurs, costs would not
be unreasonable.  These procedures could be used to eliminate the
conflict.  After considering the differences  between the CTG and NSPS,
EPA concluded that the NSPS  requirements are  not redundant and do not
substantially conflict with  existing State regulations.
     EPA also judged that the standards do not require or motivate
refiners to abandon existing plant leak detection and  repair programs.
In making these judgments, EPA noted that the CTG was  based on RACT,
whereas the NSPS is based on BDT, considering costs.   In determining
BDT, EPA analyzed the cost effectiveness and  incremental  cost effec-
tiveness of a variety of control  techniques (presented in Table 2-2),
including those presently required by State regulations.   EPA also
included existing programs in assessing the impacts of the NSPS.
Based on this, EPA determined that the costs  of the control  techniques
selected as the basis for the standards are reasonable.   The final
standards and existing programs should work together,  yet to the  extent
that some existing regulations may conflict with the NSPS,  refiners can
request a variance for existing programs as discussed  in the previous
paragraph.
Comment:
    Some commenters (IV-D-5, IV-D-16,  and IV-D-25)  contended that EPA
overstated the emission reductions of  the proposed standards by under-
stating baseline emissions.   The  emission reductions should reflect the
sources covered by State regulations.   Another commenter (IV-D-15)

                                  2-4

-------
         Table 2-2.   SUMMARY  OF  THE  INDIVIDUAL  COMPONENT  CONTROL  IMPACTS*
Fugitive Emission
Source
Pressure relief devices


Compressors

Open-ended valves
Sampling connection
systems
Valves


Pumps




Control Technique
Quarterly LDR
Monthly LDR
Rupture disks'*
Controlled degassing
vents
Caps on open ends
Closed purge sampling

Quarterly LDR
Monthly LDR
Sealed bellows valves
Annual LDR
Quarterly LDR
Monthly LDR
Dual mechanical seal
system
Emission Reduction
(Mg/yr)
4.4
5.3
9.8
16.5

2.8
2.6

66
77
110
3.0
9.8
11.5
13.9

Average Cost
Effectiveness
($/Mg)b
(170)
(no)
410
150

460
810

(110)
(60)
4,700
860
157
158
2,000

Incremental Cost
Effectiveness
($/Mg)c
(170)
250
1,000
150

460
810

(110)
310
16,700
860
(140)
170
10,900

(xx) = Cost savings
LDR « Leak detection and  repair

aCosts and emission reductions are based on fugitive emission component counts in Model B from the BID
 for the  proposed standards, EPA-450/3-81-015a, page 6-3, and from Tables A-2 through  A-13 of Appendix A.

bAverage  Cost Effectiveness = net annualized costs per component + annual VOC emission  reduction per
 component.

Incremental Cost Effectiveness = (net  annualized cost of the control  technique - net  annualized cost
 of the next less restrictive control technique)  4 (annual emission reduction of control technique -
 annual emission reduction of the next  less restrictive control  technique).

dllnderlined control  techniques were selected as basis for standards.
                                                 2-5

-------
wrote that  it was not clear whether the baseline emissions represent
existing levels of control or no control.
Response:
     Promulgation of the Priority List (40 CFR 60.16, 44 FR 4922,
August 21,  1979), reflects EPA's determination that refinery fugitive
emissions are a major source category which contribute significantly to
air pollution.  As discussed in Chapter 7 of the BID for the proposed
standards,  baseline reflects a weighted average of refineries in
attainment  (no regulations) and nonattainment (recommendations of the
refinery CTG) areas and, therefore, does account for sources covered
by State regulations.  Table 2-3, taken from the BID for the proposed
standards,  compares the projected VOC emissions under baseline level of
control for 1982-1986 with projected emissions under both the uncontrolled
level and the level of control obtained by the standards.
Comment:
     One commenter (IV-D-6) stated that the proposed regulations are
difficult to follow because of the exemptions, alternatives, and the
Federal Register format, which requires almost continuous "cross-checking."
The commenter suggested that the requirements for a specific component
appear as a separate section to improve the readability of the regulations.
Response:
     The format of the standards does require some cross-checking as the
commenter mentions.  However, the standards do present  individual  component
requirements in separate sections as requested by the commenter and
discuss common aspects of these individual  component requirements in
other sections (e.g., Section 60.595 Test Method and Procedures).  This
format greatly reduces the redundancy in  presenting the regulations.
Comment:
     One commenter (IV-D-14) maintained that the standards should apply
only to valves because they are the largest single source, and annual
inspection and repair programs are cost effective in reducing emissions
only from valves.   In addition, the commenter wrote that  limiting the
standards to valves would result in a more efficient and  practicable
approach to reducing emissions from new,  modified,  and  reconstructed
sources.  Other fugitive emission components comprise a smaller source
of emissions, and  control  for these components either has not been
demonstrated or has not been shown to be  cost effective.
                                  2-6

-------
    Table 2-3.  PROJECTED VOC FUGITIVE EMISSIONS FROM FACILITIES FOR
            1982-1986 UNDER UNCONTROLLED, BASELINE, AND NSPS*
                        Total Fugitive Emissions Projected (Gg/yr)

Year                 Uncontrolled    Baseline0          NSPSd
1982
1983
1984
1985
1986
12
24
37
51
64
9.2
19
29
39
49
3.4
7.1
11
14
18
a
 The emissions estimates are taken from Table F-10 of the BID for the
 proposed standards.  The estimates are based on projected new, modified,
 and reconstructed model units.
b
 The uncontrolled emissions projection assumes all refineries are
 operating in the absence of regulations (Regulatory Alternative I).
c
 The baseline emissions projection reflects normal existing operations
 in refineries nationwide in the absence of any new regulations.  The
 baseline assumes that refineries in nonattainment areas for ozone are
 subject to regulations similar to those recommended in the refinery
 Control Techniques Guideline document (Document No. II-A-6), Regulatory
 Alternative II in the BID for the proposed standard.  Baseline emissions
 are calculated as the weighted sum of the proportion of refineries in
 attainment and nonattainment areas for ozone:  (0.56)(Regulatory
 Alternative II) + (0.44) (Regulatory Alternative I).
d
 Emissions projected under promulgated standards.
                                  2-7

-------
Response:
     EPA agrees with the commenter that valves represent the largest
source of fugitive VOC by component type in a refinery.  However,
the uncontrolled emission factors presented in the BID for the proposed
standards, Table 3-1, clearly show that pressure relief devices, com-
pressors, and light liquid service pumps also have relatively high
emission rates.  EPA has determined that fugitive emission sources in
petroleum refinery equipment contribute significantly to ozone pollution
and, therefore, were included in a source category on the NSPS Priority
List in 40 CFR 60.16.  Under Section lll(b)(l), EPA is now required to
set standards of performance for all new sources within this listed
category for which EPA can identify BDT (considering costs).  EPA is
selecting BDT (considering costs) based on cost-effective control
techniques for the source.  Several types of refinery equipment may
emit VOC leaks, and, therefore, each is a subset of the entire source.
In selecting BDT, EPA is setting standards for each fugitive emission
component with demonstrated, cost-effective controls (see Table 2-2).
Because EPA maintains that the standards provide cost-effective control
for other sources as well as valves, equipment other than valves will
be covered by the standards.
Comment:
     Commenters (IV-D-12 and IV-D-25) indicated that the proposed
standards are inflexible, manpower intensive, and not cost effective.
Response:
     EPA has expended considerable effort to make these standards as
definitive and flexible as possible.  As discussed in the preamble for
the proposed standards, different formats are required for different
fugitive emission sources because the characteristics of the emission
sources and the availability of the measurement method used for fugitive
emission sources differ among the sources.  Performance standards allow
some flexibility because any control technique may be used if it achieves
the required level of emission reduction.  However, for most refinery
equipment, it is not feasible to prescribe a performance standard.  EPA
has selected performance standards for certain equipment, where practi-
cable, and has provided alternative standards when equipment, work
practice, design or operational standards have been used.  Hence,

                                  2-8

-------
multiple control options are allowed wherever practicable and are
considered equivalent to the control techniques selected as BDT.  In
contrast to the comment, EPA considers the manpower requirements of the
standards worthwhile and a prudent use of resources as evidenced by the
cost effectiveness of the control techniques presented in Table 2-2.
The overall cost effectiveness of the standards is approximately $130/Mg,
which EPA considers reasonable.
Comment:
     A commenter (IV-D-15) expressed some confusion on the cost-
effectiveness information presented in Table 1 of the preamble to
the proposed standards.  The values presented in Table 1, the commenter
said, cannot be generated from the emissions, labor, and cost information
presented in the BID.  The commenter added that since the BID for
the proposed standards does not contain a cost-effectiveness analysis
for each component as given in Table 1 of the preamble to the proposed
standard, a supplemental document should be prepared by EPA showing
calculations and should be made available for public comment prior to
promulgating the standards.
Response:
     Upon reviewing the calculations that were performed to generate the
cost-effectiveness information presented in Table 1 of the preamble to
the proposed standards, it was determined that a mistake was made in the
valve emission reduction calculations.  This is explained in Section 2.2.1,
In addition, the analysis for pumps was changed as discussed in the AID
(Document No. II-A-41) and Section 2.3.1.  The corrected valve and
revised pump calculations are presented in Appendix A and the results
are summarized in Table 2-2.
     The commenter is correct in that individual cost-effectiveness
estimates are not presented in the BID for the proposed standard; instead,
the proposal BID presents cost-effectiveness estimates for regulatory
alternatives.  The proposal BID does, however, present the method for
calculating the cost, emission reduction, and cost effectiveness for
the various levels of control from which individual component impacts
can be derived.  In addition, a previous supplemental information
document (Document No. IV-D-41) has been issued by EPA that explicitly
presents the method for calculating individual component impact estimates.
                                  2-9

-------
 EPA is also providing  the derivation  of  individual  component  impacts  in
 Appendix A.
      EPA does not  believe that  a  supplemental  information  document  is
 warranted.   In September  1980,  EPA  requested public comments  on the
 preliminary model  units and  regulatory alternatives;  in May 1981, the
 preliminary draft  BID  was distributed to the National Air  Pollution
 Control  Techniques Advisory  Committee, industry, environmental groups,
 and other interested persons; and in April  1982, EPA announced the
 availability of and invited  comments on an  additional information
 document  on the emissions, emission reductions, and costs  for control
 of  fugitive emission sources of organic compounds.  Because the commenters
 did not  question EPA's method of analysis and EPA's review of the
 comment  did not change its analysis, an additional information document
 is  not warranted.
 2.2  VALVES
 2.2.1  Basis  for Standards
 Comment:
      One  commenter (IV-D-15) wrote that it was not clear whether the
 cost-effectiveness  values presented in Table 1 of the proposal preamble
 are based on  a  continuing monthly monitoring schedule or the reference
 leak  detection  and  repair program for valves that allows less  frequent
 monitoring  of  non-leakers.
 Responses:
      In selecting the basis of the standards for valves, EPA considered
 different alternative monitoring periods  for valves:  annual,  quarterly,
 and monthly monitoring.   In reviewing the public comments,  EPA re-
 examined the incremental  impacts of the  three monitoring intervals
 (Document No.  IV-B-10).  Each of these intervals was compared  in  terms
 of the emission reduction  achievable and  cost-effectiveness of the leak
 detection and  repair programs as presented  in Appendix F of the BID  for
the proposed standards.  Monthly monitoring  was  selected because  it
 achieves the largest emission reduction,  77  Mg  per year  for a  Model
Unit B.  EPA also judged  that monthly  monitoring has  a reasonable  cost
effectiveness, a credit of $60/Mg,  and that  the  incremental  cost  effec-
tiveness of  $310/Mg VOC for monthly  versus quarterly monitoring is
                                  2-10

-------
reasonable.  Based on these estimates EPA considers  monthly monitoring
BDT for valves.
     Available data (II-A-21 and II-A-26) indicate that leak recurrence
is an important factor in predicting leaks from valves.  That is,  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 is  important in predicting
leaks, EPA considers that the annual cost of monthly monitoring of
valves that leak infrequently would probably be unreasonably high  in
comparison to the annual cost of quarterly monitoring,  considering the
emission reduction achieved by monthly and quarterly monitoring.
Therefore, the standards allow quarterly monitoring  for valves which
have been found not to leak for two successive months resulting in a
hybrid monthly/quarterly monitoring program.
     It is possible that basing the standards on monthly monitoring,
but allowing monthly/quarterly monitoring, has led to confusion.  The
basis of the standards remain monthly so the cost-effectiveness estimates
for valves given in the proposal preamble are based  on  continuing
monthly monitoring.  By basing the cost analysis on  monthly monitoring
rather than monthly/quarterly, a maximum cost impact estimate was
evaluated.  It is important to note, however, that the  actual  cost
effectiveness of the standards for valves is likely  to  be even better
because the standards allow quarterly monitoring for valves that have
been found not to leak for two successive months (monthly/quarterly
monitoring).  EPA expects that most affected facilities would follow
the monthly/quarterly reference leak detection and repair program
and, further, that most valves would be on the quarterly inspection
schedule.  Hence, the actual costs for valves under the standards  is
likely to be more closely represented by the costs estimated for
quarterly monitoring.
     Upon reviewing the calculations that were performed to generate
the information presented in Table 1 of the preamble to the proposed
standards, it was determined (Document No. IV-B-3) that an error was
made in the valve emission reduction calculations.  This is likely what
caused the commenter's confusion on the calculation  of  the impacts of
the valve standards.  Valve impacts are calculated based upon a Model
                                  2-11

-------
 Unit  B  refinery  unit  component  inventory  (260 gas/vapor service valves
 and 500  light  liquid  service valves).  The valve emission reductions
 were  underestimated by mistakenly using a weighted average of the
 emission  reductions for the two types of valves, gas/vapor and light
 liquid, rather than the total emissions from the two types of valves.
 The corrected  emission reductions impacts are presented in Table 2-4.
 In rechecking the cost and emission reduction calculations, rounding
 resulted  in a  slight  differences in the cost effectiveness and incremental
 cost  effectiveness values.  The revised emission reduction and cost
 effectiveness  for monthly leak detection and repair are 77 Mg/year
 and a savings of $60/Mg VOC emission reduction, respectively.  The
 incremental cost-effectiveness from quarterly to monthly monitoring now
 shown is  $310/Mg of VOC emission reduction.  The revised numbers have
 been  incorporated into the analysis of the final standards and did not
 affect any of the decisions on the proposed standards.
 Comment:
      A number of commenters (IV-D-8, IV-D-12, IV-D-14, IV-D-16, IV-D-
 17, IV-D-18, IV-D-21, and IV-D-25) wrote that monthly monitoring of valves
 is not cost effective.  Commenters contended that their experience with
 less  frequent monitoring intervals (quarterly and annual  monitoring)
 shows that these intervals are more reasonable.   Some commenters recom-
 mended annual  monitoring for valves because the  results of programs
 performed at West Coast (California) refineries  indicate that leak
 occurrence rates for valves under annual  monitoring are lower than
 EPA's assumed estimate of 3.8 percent on a quarterly basis.   Two com-
 menters (IV-D-8 and IV-D-14) stated that their refinery-wide  leak
 occurrence rate was only 1 to 2 percent on an annual  basis.   Similarly,
 another commenter (IV-D-21) stated that annual  inspection  programs
 result in leak occurrence frequencies as low as  0.3 percent.
     One commenter (IV-D-15) recommended that EPA obtain  leak detection
 and repair program data generated in California.  Other commenters (IV-
 D-14, IV-D-25a, and IV-D-31) provided data on leak  detection  and  repair
 programs.  In particular,  information was provided  EPA concerning  leak
 frequency, leak occurrence and  recurrence rates, and  monitoring  and
maintenance costs.
                                  2-12

-------
        Table  2-4.   REVISED  EMISSION  REDUCTIONS AND COST  FOR VALVE
                    LEAK DETECTION AND REPAIR PROGRAMS3
                                         Monitoring  Interval
                                    Quarterly              Monthly
Emission Reduction
(Mg/yr)b
Proposed0
Revised9
Average $/Mgd
Proposed0
Revised3
Incremental $/Mge
Proposed0
Revised3

31.7
66

(90)
(110)

—

37.1
77

(40)
(60)

300
310
(xx) = Cost savings.

Memorandum from T.W. Rhoads, PES, Inc., to Docket A-80-44.
 VOC Emission Reduction and Cost-Effectiveness Estimates.
 July 14, 1983.  Document No. IV-B-3.

bBased on Model Unit B component counts, BID for proposed standards.

°From Table 1 preamble for proposed regulation.

^Average dollars per megagram (cost effectiveness) = net annualized
 cost per component * annual  VOC emission reduction per component.

Incremental  dollars per megagram = (net annualized cost of monthly
 monitoring - net annualized  cost of quarterly monitoring) * (annual
 emission reduction of monthly monitoring - annual emission reduction
 of quarterly monitoring).
                               2-13

-------
 Response:
      As  discussed  in  the previous  response, EPA evaluated three
 monitoring  intervals  for valves:   annual, quarterly, and monthly.  Each
 of  these intervals was compared  in terms of the emission reduction
 achievable  and the cost effectiveness of the leak detection and repair
 programs.   Annual  and quarterly monitoring are more cost effective than
 monthly  monitoring, however, the standards require monthly monitoring
 because  it  provided the greatest emission reduction at a reasonable
 cost  effectiveness and incremental cost effectiveness.
      The commenters are questioning the cost effectiveness estimates
 used  by  EPA based mostly on their experiences with monitoring required
 by  State implementation plans in California.  However, the effectiveness
 of  leak  detection and repair programs in California is not strictly
 comparable  with the regulatory alternative used in the BID for the
 proposed standards.  The commenters refer to the South Coast Air Quality
 Management  District (SCAQMD) Rule 466.1 on leakage from valves and
 flanges.  Contrary to the commenter's contention, monitoring under Rule
 466.1 is not strictly on an annual basis, but rather biannual  for the
 first year  and annual  in the following years.  Like the final  standards,
 Rule  466.1  focuses on recurring leakers; Rule 466.1 requires follow-up
 inspections on leaking equipment at 3 months and, if they are  still
 leaking  at  this inspection, follow-up inspections at successively
 shorter  periods, and in addition, Rule 466.1 requires all  repairs to
 be completed within 2 working days unless a variance is obtained.  In
 contrast, the standards require that repairs be made as soon as  practi-
 cable with  an initial  attempt within 5 days and completion  within 15
 days.  The  standards also provide for automatic delay of repairs to  a
 process unit turnaround.  Another important distinction is  distance  at
 which monitoring measurements are taken.  The standards require  measure-
ment at the source, whereas the SCAQMD Rule 466.1  allows measurement at
 1 cm from the surface.  Thus, simple comparison of data from refineries
 subject to the SCAQMD  rules to EPA's data base  is  misleading.
     In response to these comments, EPA reviewed  and  compared  equipment
leak data from current industry leak detection  and  repair programs to
the data used in developing the  standards.   The  results of this  analysis
                                  2-14

-------
 are  presented  in  Appendix C.  Where these data relate to specific
 comments on the standards, they are also incorporated into the response
 to the comment.
      Section C.I.2 of Appendix C presents occurrence rate data obtained
 from  industry  leak detection and repair programs.  An EPA study found
 (Document  No.  IV-B-11 and Section C.4) that the occurrence rates
 observed at two refineries in the South Coast Air Quality Management
 District (SCAQMD) are similar to the occurrence rates EPA used to
 estimate the national average for evaluating the effectiveness of leak
 detection  and  repair programs.  Other existing leak detection and repair
 programs,  however, have resulted in low annual occurrence rates as the
 commenters argued.  However, the occurrence rate data obtained from
 refineries with existing leak detection and repair programs may be
 underestimated as a result of differences, as discussed above, in the
 requirements of existing regulations and those considered in developing
 the  standards.
      To the extent EPA data do not reflect certain process units covered
 by current plant  practices (State regulation or otherwise), the standards
 have  been developed to define BDT (considering costs) appropriately taking
 this  into account.  As discussed in the first response in this section
 and in the presentation before the National Air Pollution Control
 Techniques Advisory Committee (NAPCTAC) in June of 1981 (Document No.
 II-B-34) and in the preamble to the proposed standards, EPA believes
 that monthly monitoring for valves with a history of low leak rates is
 unnecessary.  The final  standards, therefore, allow monthly/quarterly
monitoring, and alternative standards are provided for units with low
 rates.  EPA believes the standards and the alternative standards
 represent BDT for all  units covered by the standards.
Comment:
     One commenter (IV-D-14)  submitted the results of an LDAR Model  run
which calculated an incremental  cost effectiveness from annual to
monthly monitoring of $5,900/Mg.  In addition, this commenter added
that this LDAR Model  run predicted emission reductions from the annual
inspection program at a West  Coast refinery would achieve a greater
emission reduction than  EPA's estimate for monthly monitoring, 72 percent
versus 70 percent, respectively.
                                  2-15

-------
Response:
     EPA reviewed (Document No. IV-B-16) the commenter's LDAR Model run
and found several problems with the data inputs used by the commenter.
The commenter used 1982 costs, rather than May 1980 costs that were
used for the basis of the standards.  Also, the commenter's use of
occurrence rates in the analysis was incorrect.  The commenter substi-
tuted different (2.0 percent annual) occurrence rates without also
correcting the initial leak frequency and emission factors, which are a
function of the occurrence rate.  In lowering the occurrence rate, a
corresponding reduction in the initial leak frequency and average
emission factor should occur (Document Nos. II-B-43 and II-B-7).   The
commenter wrote that the 2.0 percent occurrence rate represents the
West Coast annual inspection program, yet, as discussed in the previous
response, direct comparison of annual monitoring under Rule 466.1 with
the standards for valves is misleading.   Also, the commenter did  not
use the LDAR Model input values that they indicated.  The commenter's
occurrence rates were purportedly 2.0 percent annual  occurrence rates
for annual and monthly monitoring, yet,  in reviewing the commenter's
analysis, it was found that an 8 percent annual  occurrence rate was
used to evaluate monthly monitoring.  In addition, the commenter  failed
to use half the inputs the commenter said would provide a better  estimate
of the impacts of leak detection and repair programs.  Correcting just
the occurrence rate (to 2.0 percent annual  occurrence for both annual
and monthly monitoring) and the cost basis (to 1980 dollars) in the
commenter's data inputs results in an 80 percent emission reduction for
monthly monitoring (significantly more than the 72 percent reduction
achieved by annual  monitoring in the comment)  and a cost effectiveness
of monthly monitoring of $500/Mg and an  incremental  cost effectiveness
between annual  and monthly monitoring of $l,900/Mg.  EPA and the
reported source of the commenter's estimates do  not believe the inputs
the commenter used are representative.
Comment:
     Two commenters submitted that monthly  monitoring would be more
costly than EPA estimated.   One commenter (IV-D-24)  wrote that the
                                  2-16

-------
 costs for monthly monitoring at one refinery would increase  from  $6.40
 per component per year for quarterly monitoring  to approximately  $19
 per component per year for monthly monitoring.   Another  commenter
 (IV-D-25) and IV-D-25a wrote that  the cost  per component  given  in the
 April  1981 preliminary draft BID,  $0.82,  is far  below  the $2.57 per
 component monitoring  cost  experienced at  his  refinery  in  1982.  The
 commenter further stated that they had a  total program cost  of  $4.82
 per component.
 Response:
      EPA  calculated the costs per  valve based on the data and information
 discussed  in  "Fugitive Emission Sources of  Organic Compounds -  Additional
 Information on  Emissions,  Emission Reductions, and Costs"  (AID),  April
 1982,  EPA-450/3-82-010.  EPA  requested public comments on  the monitoring
 labor  requirement and  cost estimates  in the AID.   EPA  has  previously
 received  specific comments at  the  June 1981 meeting of the National  Air
 Pollution  Control Techniques Advisory Committee  (NAPCTAC) on the  information
 in  the BID for  the proposed standards.  After reviewing the NAPCTAC
 comments and comments  on the AID,  EPA added some provisions making them
 more practicable where  possible, however,  the standards remain  essentially
 the same.
     To compare the first commenter1s cost estimates to the EPA's
 estimates, it was necessary to contact the commenter to determine  the
 basis of his costs.  From the  additional information obtained (IV-F-28),
 EPA learned that the leak detection and repair costs submitted were
 mostly for valves (about 90 percent), but  included some components
 other than valves.  Thus, the commenter was comparing the costs  for
 leak detection and repair for  all  components to the EPA's cost estimates
 for valves.  Since the cost to monitor and repair components  other than
 valves are typically higher than that for  valves, the commenter's  costs
 overestimate the actual cost per valve he  incurred.  EPA  also learned
 that the costs provided reflect a contractor's total  labor costs for
monitoring and repairs from a refinery inspection in 1982 dollars. The
costs submitted did not include the refiner's overhead.
     The first commenter's  costs were adjusted for 1980 dollars  (to be
consistent with the EPA's  cost estimates)  by using cost indexes  from
Chemical  Engineering (Document Nos. II-I-58 and  IV-J-2).   An  overhead
                               2-17

-------
rate was also applied to the commenter's costs consistent with EPA's
estimates.  The results, compared in Table 2-5, show that EPA's cost
for quarterly monitoring is somewhat higher, $5.91 per valve compared
to $4.22 per valve.  The commenters adjusted estimate for monthly
monitoring, however, is somewhat higher than EPA's, $17.5 per valve
compared to $16.00 per valve.  The comparison of the commenters1  costs
and EPA's estimates indicates that the EPA's estimates are reasonably
close to the plant's expenditures for leak detection and repair.   This
commenter did not comment on the basis for the EPA cost estimates.
Also, to the extent that an individual plant's costs may be higher than
the EPA estimates, the EPA costs appropriately reflect the nationwide
average costs to comply with the standards, such that any individual
plant costs may be somewhat higher or lower.
       Table 2-5.  COMPARISON OF COMMENTER AND EPA ANNUAL COSTS FOR
                       LEAK DETECTION AND REPAIR
                      Commenter Estimates             EPA Estimates
                        For Components9               For Valvesb


Monitoring
Period
Quarterly
Monthly
1982
Dollars


6.40
19.00
1980
Dollars


4.22
12.50
1980
Dollars
plus
Overhead
5.90
17.50
1980
Dollars


9.20
16.00
 a
  From Document Nos. IV-D-24 and IV-E-28.  Cost are adjusted  to  1980
  dollars using cost indexes (Document Nos. II-I-58 and  IV-J-2).   Costs
  presented in the comment letter are based on 1982 dollars.  .Overhead
  is estimated as 0.40 x monitoring and repair labor cost.
 b
  From Document No. IV-B-10.
     The second commenter's costs are based on first year costs  to
implement a leak detection and repair program similar to Regulatory
Alternative II in the BID for the proposed standards, which specifies,
quarterly inspections of gas service components and annual inspections
of liquid service components.  A review of the commenter's cost data
                                  2-18

-------
(see Section C.I.5 of Appendix C) found the commenter's actual  costs to
be similar to EPA's cost estimates when compared on a common basis
(e.g., adjusting to a 1980 cost basis, overhead rate).  The commenter's
total leak detection and repair costs in 1980 dollars, about $71,100
per year, compare to an EPA estimate of about $60,900 per year for a
similar scenario.  Here again, the comparison is not strictly valid
because the two estimates are based on different component populations.
The commenter's data include an unknown number of component types not
included in the CTG recommendations such as valve flanges and capped
open ended lines.
     The information EPA received from the commenters did not lead EPA
to change the cost estimation methods.  Therefore, EPA has not changed
the basis for the costs and considers the costs of a monthly leak
detection and repair program for valves to be reasonable (i.e., an
average cost effectiveness (credit) of $60/Mg and an incremental  cost
of $310/Mg compared to quarterly monitoring.
2.2.2  Alternative Standards
Comment:
     Some commenters (IV-D-5, IV-D-10, IV-D-24, and IV-D-30) wrote that
they support the alternative standards for valves as they provide
incentive for a facility to maintain a low incidence of leaking sources.
One of the commenters (IV-D-10) wrote that the addition of alternative
standards for valves was an improvement from the proposed standards for
valves in the synthetic organic chemical manufacturing industry (SOCMI).
Another commenter (IV-D-30), however, questioned the basis for allowing
a 2.0 percent leaking valve rate when the objective is to keep the real
leak rate below an average of 1.0 percent because most facilities would
operate with a real  leaking valve rate well above 1.0 percent.
Response:
     EPA believes that monthly monitoring does not have a reasonable
cost effectiveness for process units with a low percentage of valves
leaking.   EPA judged at proposal that for units with less than (on the
average)  1.0 percent valves leaking, monthly monitoring is unreasonable.
EPA has,  therefore, included alternative standards for valves in units
with a low percentage of leakers:  (1) two skip period monitoring
programs  and (2) an allowable percentage of valves leaking (performance
                                  2-19

-------
limit).  This approach addresses  the comment that,  if annual  leak
detection and repair or some other reduction program reduces  leak  rates
to an average of well under EPA's estimates, the cost effectiveness  of
monthly monitoring is unreasonable.
     The alternative standards apply on an affected facility  basis,
(i.e., individual process unit).   As was explained  at proposal,  the
allowable percent (2.0 percent)  of valves leaking was selected  (Document
No. II-B-43) after considering the costs and emission reductions of
monthly monitoring of low leak units and the variability  inherent  in
leak detection of valves.  The variability in the number  of valves
which are found leaking at any one time (e.g., variability  in the
monitoring instrument, instrument operators, the piece of equipment,
leak occurrence, and recurrence)  in leak detection  of valves  can be
characterized as a binominal distribution around the average  percent of
valves leaking.  Inclusion of the variability in leak detection  of
valves is accomplished by straightforward statistical techniques based
on the binominal distribution.  An allowable percent of valves  leaking
of 2.0 percent, to be achieved at any point in time, would  provide an
owner or operator a risk of about 5 percent that greater than 2.0
percent of valves would be determined leaking when  the average  of  1.0
percent was actually being achieved.  Based on these considerations,
EPA considers an allowable percent of valves leaking of 2.0 percent  to
represent about one percent of valves leaking.
     The first alternative specifies two statistically based skip-period
leak detection and repair programs.  Under skip-period monitoring  programs,
an owner or operator can skip from routine monitoring to less frequent
monitoring after completing a number of consecutive monitoring  intervals
with performance levels less than 2.0 percent of valves leaking.  The
first skip-period program provides that after 2 consecutive quarterly
leak detection periods with the percent of valves leaking equal  to or
less than 2.0, an owner may begin to skip one of the quarterly  leak
detection periods (semiannual monitoring).  The second skip-period
program provides that after 5 consecutive quarterly leak detection
periods with the percent of valves leaking equal to or less than than
2.0, an owner or operator may begin to skip 3 of the quarterly  leak
detection periods (annual monitoring).  This skip period alternative
                                  2-20

-------
 standard also requires that, if an affected facility exceeds the 2.0
 percent limit of valves leaking during the semiannual or annual  inspec-
 tion, the owner or operator must revert to the monthly/quarterly leak
 detection and repair program that is specified in the standards.  The
 original criteria for skip monitoring would again have to be met before
 owners and operators can again skip monitoring periods.
     The second alternative standard for valves is a performance standard
 that specifies a 2.0 percent limitation as the maximum percent of
 valves leaking within a process unit.  This alternative standard would
 require a minimum of one performance test per year.  This alternative
 provides an incentive for maintaining a low percentage of leaking valves
 level by implementing any type of leak detection and repair program or
 engineering controls at the discretion of the owner or operator.
 Comment:
     Several (IV-D-8, IV-D-12, IV-D-14, IV-D-16, IV-D-17, IV-D-18,  and
 IV-D-21.) commenters suggested that the regulations should begin with
 annual inspections and require more frequent inspections only if needed.
 Another commenter (IV-D-25) wrote that industry would be reluctant  to
 use the alternative standards due to noncompliance penalties.
 Response:
     Based on evaluation of data that EPA considers representative  of
 petroleum refineries, EPA selected standards for valves that require
monthly/quarterly monitoring.  The standards,  however, also provide
 alternatives for facilities (process units) with relatively low  leak
 frequencies.  The commenters are asking that the standards be structured
to allow increasing the frequency of monitoring in high-leak units
 rather than decreasing the monitoring in low-leak units (as currently
structured).
     The standards for valves (monthly/quarterly monitoring) and the
 alternative standards are structured to assure that best demonstrated
 technology for valves is achieved initially and throughout implementation
 of the standards.   The data (II-A-19) indicate that about 10 percent of
 the valves in a facility would be found leaking on an initial inspection.
 Hence, the standards are structured to identify and control  leaking
 valves through relatively frequent monitoring  initially, and once
 recurring leakers  are identified and controlled, allow less frequent
                                  2-21

-------
monitoring.  If the standards were structured as the commenters propose
(based on increasing monitoring frequency), less emission reduction
would result by allowing longer time intervals before recurring leakers
are controlled.
     EPA agrees with the commenters in that owners and operators might
be reluctant to use the alternative standard for allowable percentage
of valves leaking.  Use of this alternative standard would subject an
affected facility to non-compliance penalties.  The owner or operator of
a process unit selecting the allowable percentage of valves leaking
alternative would have to do performance tests initially, annually, and
at other times requested by the EPA Administrator.  If more than two
percent of the valves are found leaking, the facility would not be in
compliance with the regulation.
     For many facilities it may be impossible to guarantee that the
facility will always have less than 2.0 percent valves leaking.  These
facilities should consider implementing the skip-period monitoring
programs outlined in the previous response.  For facilities following
the skip-period monitoring alternative standard, the "penalty"  for
having greater than 2.0 percent valves leaking is more frequent
monitoring rather than non-compliance with the standard.
2.2.3  Special  Provisions
     The following comments and responses pertain to specific groups of
valves and provisions in the standards for valves relating to them.
2.2.3.1  Difficult-to-mom'tor valves
Comment:
     Two commenters (IV-D-12 and IV-D-15) maintained that difficult-to-
monitor valves can not be eliminated in new units, even though  they can
be reduced in number.  Therefore, the difficult-to-monitor provisions
should be allowed for new facilities as well  as existing facilities.
Conversely, another commenter (IV-D-30) objected "to the exception from
the monitoring requirements for supposedly difficult-to-monitor valves,"
claiming that, "it simply is not a significant burden  for the monitoring
personnel  to use a ladder to reach valves higher than  2 meters  off the
ground." One commenter (IV-D-4) remarked that  requiring monitoring
personnel  to carry equipment and climb  a ladder to inspect  difficult-
to-monitor valves could double the cost of the monitoring program.
                                  2-22

-------
Furthermore, commenters (IV-D-8, IV-D-15, IV-D-21, and IV-D-25)  maintained
building new units without difficult-to-monitor valves will  substantially
increase unit costs because of extra ladders and platforms.   The commenters
added that it is too costly to monitor difficult-to-monitor  valves.
Another commenter (IV-D-24) requested that the EPA include a statement
that would require annual  monitoring "if practicable."
Response:
     The intent of the standards is to monitor those valves  that can  be
reached with the use of portable ladders or with existing supports  such
as platforms and fixed ladders.  EPA defines valves that cannot  be
reached without extraordinary means, difficult-to-monitor valves, as
those valves that cannot be monitored without elevating the  monitoring
personnel more than 2 meters above a support surface.  EPA does  not
consider valves that can be reached from a portable ladder to be difficult-
to-monitor, and hence the standards require operators to use portable ladders
to monitor such valves.
     EPA has estimated the cost for monitoring difficult-to-monitor
valves in existing units (Document No. II-B-46) and determined that the
cost effectiveness of monthly monitoring of difficult-to-monitor valves
may be unreasonable, and that the average cost effectiveness for annual
monitoring is reasonable.   Hence, EPA proposed annual monitoring of
difficult-to-monitor valves in existing facilities.  The provision  was
not allowed for newly constructed affected facilities because commenters
on the proposed standards  for VOC fugitive emission in the synthetic
organic chemicals manufacturing industry (Document No. IV-A-5, Section
4.2.4) wrote that difficult-to-monitor valves can be eliminated  in  new
units.  A refinery design  engineer (IV-E-15) also indicated  that new
units can be designed with no difficult-to-monitor valves.
     Upon reviewing the comments received on the proposed standards,
EPA agrees with the commenters in that eliminating all difficult-to-
monitor valves in new units may substantially increase the costs of
constructing a new unit, for example, due to the necessity for
additional  fixed ladders and platforms.  Yet, 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 difficult-
to-monitor valves (Document No. IV-B-13).
                                  2-23

-------
     A refinery maintenance study (Document No. IV-A-3) 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 on these data and contacts with petroleum refinery design engineers
(Document No. IV-B-13) the final standards allow the owner or operator
of a newly constructed process unit  to designate no more than 3 percent
of its valves as difficult-to-monitor.  The standards require annual
monitoring of those valves.  Limiting the percent 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:
     Other commenters (IV-D-8, IV-D-21, and IV-D-24) wrote that it is
unreasonable to stand on any elevated object and reach overhead to
monitor for leaks because the practice is unsafe.
Response:
     EPA does not believe that it is necessary to include a provision
for valves that require operators to "reach overhead."  The standards
require operators to monitor valves and to repair leakers that can be
reached safely with or without the aid of a ladder.  The practice of
reaching overhead to perform monitoring is not generally unsafe, and,
to the extent this can be unsafe, personnel should be provided proper
equipment (e.g., head and eye protection, ladders) and training as
required by the Occupational Safety and Health Administration and
refinery safety guidelines.
2.2.3.2  Unsafe-to-Monitor Valves
Comment:
     Four commenters (IV-D-8, IV-D-15, IV-D-21, and IV-D-25) wrote that
unsafe valves should never be monitored because they are no less safe to
monitor annually than monthly.  These valves are only safe-to-monitor
when they are out of service, and it makes little sense to monitor
components not in service.  One commenter (IV-D-24) wrote that the
proposed exemption of valves in unsafe locations, §60.5927(g)(2),
should have a qualifying statement added so that it reads "required
                                  2-24

-------
 monitoring of the valve  as  frequent  as  practicable  during safe-to-monitor
 times  but  not more than  quarterly."
 Response:
     EPA agrees  with  the commenters  that valves  should never be
 monitored  during unsafe-to-monitor conditions, that is, during periods
 of  extreme temperature,  pressure, or explosive process conditions that
 make these areas off-limits to all personnel.  Accordingly, the standards
 do  not  require monitoring during unsafe periods.  Valves that are
 considered unsafe-to-monitor are not unsafe-to-monitor all the time.
 Monitoring can conform to the requirements of the standards as much as
 possible,  but monitoring does not need to occur  during unsafe conditions.
 Valves  that are  routinely operated under safe conditions would be
 subject  to the routine monthly monitoring required by the standards.
 Valves  that are  only  safe-to-monitor once per quarter or year would be
 subject  to quarterly  or annual monitoring, respectively.  The standards
 require  an owner or operator to explain why a valve is unsafe-to-monitor
 and to develop a  plan to monitor the valve when  it is safe, as often as
 possible but  not  more than monthly.  For valves  that are safe-to-monitor
 only when  they are out of service (for example,  during a process unit
 shutdown),   pressure testing such as  is specified by API Standard 598 for
 new valves   could  be part of an owner's or operator's monitoring plan.
     The provisions for unsafe-to-monitor valves were included in the
 proposed standards of performance for equipment  leaks of VOC in the
 Synthetic  Organic Chemicals Manufacturing Industry (46 FR 1136,
January 5,   1981)  because a few valves may be unsafe-to-monitor; the same
 provisions  were  proposed in these refinery standards.  EPA believes that
 very few such valves exist in refineries.
Comment:
     Another commenter (IV-D-16)  requested that EPA delete the requirement
to demonstrate that valves are unsafe-to-monitor and the need for a
written plan for monitoring unsafe valves.
Response:
     EPA is providing the exception for unsafe-to-monitor valves and
does not believe  other valves  should  be allowed to use this  exception.
Very few,  if any, valves in a refinery would be considered unsafe-
to-monitor  by EPA.  Thus, a demonstration  that  particular valves  are
                                  2-25

-------
unsafe-to-monitor and a written plan for monitoring these valves  is not
a significant burden upon owners or operators.   These demonstrations
and plans are needed to ensure compliance with  the intent of the  standards,
which is use of best demonstrated technology, considering costs,  on all
new sources.
2.2.3.3  Small Valves
Comment:
     One commenter (IV-D-22) stated that small  valves, in general,  have
lower mass emissions at 10,000 ppm than larger  valves and suggested
that EPA provide a small valve exemption such  as that in the State  of
Texas.  In a refinery as many as 50 percent of  all  valves are 2 inches
and smaller, and eliminating the monitoring requirement for these
valves would lessen the monitoring burden.  Another commenter (IV-D-15)
wrote that a large number of the refinery valves are small  valves
servicing instruments or control system bypasses and that the repair of
these small valves cannot be performed while in service.  Hence,  it was
recommended that the standards apply only to valves 3/4 inch size or
larger since repair costs for small valves is greater than  for large
valves.
Response:
     The first commenter1s request for a small  valve exemption is
predicated on his contention that small valves  have lower emissions at
10,000 ppm than large valves.  EPA contends, however, that the relation-
ship between valve size and mass emissions at 10,000 ppm is not relevant,
although the relationship for the average emissions per valve at  or
greater than 10,000 ppm is relevant in assessing the need for a small
valve exemption.  Nevertheless, the commenter's contention that valve
size relates to valve emissions is not supported in any test data.   On
the contrary, EPA test data indicate that valve emissions are essentially
independent of valve size.  An EPA study (Document No. II-A-19) found
only a slighly positive correlation between mass emissions  from valves
and valve line size (correlation coefficients  (r) equal  to  or less  than
0.150).  Also, data from facilities with existing leak detection  and
repair programs, presented in Section C.I.3 of  Appendix C,  further  demon-
strate that small valves account for a significant portion  of leaking
valves.  This data indicates that small valves  (less than or equal  to
                                  2-26

-------
3.8 cm or 1.5 inches) represent nearly half the valves found leaking.
The commenter noted that the Texas State Implementation Plan exempts
all valves that are 2 inches or smaller; however, he failed to add that
the exemption is contingent upon demonstration that emissions would not
increase by more than 5 percent as a result.
     EPA agrees that some small valves may need to be replaced for repair,
but the cost of repair for these valves is reasonable.  EPA has estimated
valve repair to require 1.13 labor hours based on 75 percent of all
valves being repaired on-line and in service with a repair time of 10
minutes, and 25 percent of the valves requiring off line repair requiring
4 hours per repair (BID for the proposed standards, Chapter 8).  EPA
anticipates that most instrument valves are not in VOC service and
would therefore not be covered by the standards.  However, if they are
in VOC service, small valves servicing instruments and control systems
are normally field repairable.  Most repairs would consist simply of
replacing the stem packing ring.  For valves in corrosive service, the
stem may be deteriorated to the extent that an entire stem assembly
(stem, packing, and stem tip seal) must be replaced.  Repair time in
either case would be less than 30 minutes (Document No. IV-B-8).
Hence, small  valves are no more expensive to repair than large valves.
The data presented and discussed in Section C.I.3 also indicate that
small  valves are as repairable on-line as large valves.  For those
valves in critical  service (i.e., those that cannot be isolated from
the process), the standards provide for delay of repair until  a process
unit shutdown.
2.2.4  Monitoring Time
Comment.
     A number of commenters (IV-D-4,  IV-D-15, IV-D-18) were concerned
with the EPA estimates of monitoring  time.   Two commenters (IV-D-4 and
IV-D-15)  stated that monitoring would require 2 minutes per valve.
Another commenter (IV-D-18) reported  that their experience found
that a two-man  team averages one valve every 3 minutes, so that 160
valves could be monitored in one 8-hour day.  This commenter also noted
problems  with hiring part-time or full-time employees to conduct  moni-
                                  2-27

-------
toring for the alternative standards;  with training  personnel;  and  with
purchasing additional  monitoring instruments.
Response:
     The EPA monitoring time estimate  of 1 minute per  valve was taken
initially from information provided by Exxon Company,  U.S.A.   (Docket
No. II-D-22) based on an "in-depth study to determine  the  monitoring
manpower requirements."  The average monitoring time for a leak detec-
tion survey for valves was found to be 1 minute per  valve  for  a two-man
team (2 man-minutes).   In another study conducted by Union Carbide
Corporation (Document No. II-I-57) an  estimated 400  to 500 sources
(valves and other equipment) were screened per  day.  Although  this
estimate was based on a three-man team, the third person was a unit
operator who provided process data. For a two-man monitoring  team,
this corresponds to 1.9 to 2.4 man-minutes per  source  (0.95 to 1.2  man-
minutes per source per monitoring person).  Information from other
studies, including EPA studies, shows  that monitoring  times are generally
less than 2 man-minutes.  Phillips Petroleum Company conducted a study
(Document No. IV-B-10) of a petroleum  refinery  and petrochemical  complex
in which 70,000 components were screened in about 936  manhours with a
two-man team.  This represents an average of 0.8 man-minutes per component.
EPA also reviewed the results of recent California Air Resources Board
(CARB) inspections of refineries in the South Coast  Air Quality Manage-
ment District (SCAQMD) and the Bay Area Air Quality  Mangement  District
(BAAQMD).  The data (presented in Section C.I.6 of Appendix C), submitted
in part by one commenter (IV-D-31) and in part  obtained from BAAQMD
and SCAQMD (Document No. II-B-18) revealed that monitoring time averaged
about 1 minute per valve for the more  than 6,400 valves monitored in 12
refineries.  In this effort, 2 monitoring instruments  were used and
more information than required under the standard was  recorded, such as
line size, time of monitoring, and valve type  and function.  Considering
these data, the time estimate of 1 man-minute  per valve (2 man-minutes
per valve for a two-person team) used  for costing purposes is  reasonable.
     In further reviewing the commenters's claim that  EPA  underestimated
the time required to monitor a valve,  EPA examined the effect  on the
cost effectiveness of monthly monitoring for valves  assuming that twice
the monitoring time (2 minutes per source) is  needed (Document No.

                                  2-28

-------
IV-B-4).  The results obtained from the LDAR Model  show  that  monthly
monitoring would have a cost effectiveness of $42/Mg and an  incremental
cost effectiveness from quarterly monitoring of  $768/Mg.  Hence, monthly
monitoring would, nevertheless, be reasonable even  if 4  man-minutes  per
valve were required.
     The cost impacts presented in the background information document
for the proposed standards (Chapter 8) included  the cost for  two moni-
toring instruments per model unit plus $3,000 per year (1980  dollars)
for instrument calibration and maintenance.  Therefore,  the cost of
additional monitoring instruments is accounted for  in the cost impacts.
Furthermore, the actual monitoring instrument costs incurred  by a
refinery may be less since monitoring instruments may be used for  more
than one process unit.  The cost impacts are based  on 2  monitoring
instruments per process unit (affected facility).  However,  when there
are several affected facilities in a refinery, it is likely that the
refiner will not purchase two monitoring units for  each  of them.
     Training plant personnel to use the monitoring instruments and
perform equipment monitoring is also considered in  the cost  analysis.
These costs are included as "Administrative and Support" costs
(40 percent of the total monitoring labor and maintenance labor costs).
Owners/operators may, however, choose to employ consultants  to perform
equipment monitoring.  Use of consulting firms would eliminate the need
to hire part-time or full-time employees for a short period  of time  (e.g.,
half a year), for example, if monitoring requirements are reduced
through use of the alternative standards for valves.
     It is noteworthy to reiterate that promulgation of Method 21
(48 FR 37598) provides an alternative monitoring technique,  soap  screening.
Soap screening, although restricted to those valves with moderate
surface temperatures, may significantly lower the average monitoring
time.  By soap screening valves, owners/operators will experience  lower
costs for monitoring instrument maintenance.  Further, the standards
provide for quarterly monitoring of valves which are found not leaking
for two consecutive inspections.  Hence, monthly/quarterly monitoring
will lower costs as a  result of reduced monitoring labor requirements
and instrument wear.
                                  2-29

-------
Comment:
     Some commenters (IV-D-17, IV-D-18 and IV-D-25) contended that the
monthly monitoring requirements will not permit sufficient time for all
of the components in a unit to be monitored and repaired.  One commenter
(IV-D-18) noted that it would be impossible for one two-man crew to
complete all inspection work, process work orders, and complete record-
keeping requirements within the one-month time frame.
     Another commenter (IV-D-25) offered the example of a facility
having about 14,000 components to be monitored.  Based on a crew moni-
toring 200 components per day, it would take 24 days for three crews to
check the components '(not including the time for repairs and rechecks).
Response:
     The basis for the EPA time estimate for performing leak detection
and repair is found in Table 8-3 of the BID for the proposed standards.
An average (for valves, pumps, and pressure relief devices) monitoring
time requirement per component can be estimated based upon the component
distribution for Model  Plant B, as shown in Table 2-6.  The resulting
average component monitoring requirement is about 2.3 man-minutes.
Based on this estimate, a two-man monitoring team can inspect about
420 components in an 8-hour day.  Hence, a single two-man monitoring
team can inspect a Model  Unit A in about 1 day, Model  Unit B in 2 days,
and Model Unit C in 5 days.  Actual  industry and EPA testing (Document
No. II-A-41) has demonstrated that a two-man monitoring team can inspect
between 400 and 500 components per day.
      Table 2-6.  Derivation of Average Component Monitoring Time
Component
type
Valves
Pumps
Rel ief valves
Number of
components
760
14
7
Time to
monitor
(min.)
1
5
8
Persons
2
2
2
Total
time
(man-mi n.)
1520
140
112
Total
      781                                 1772
Average =2.3 man-minutes per component.
                                  2-30

-------
     EPA believes that a month provides more than enough  time  to  complete
leak detection and repair of an affected facility.   As  previously
pointed out, even a large process unit, Model  Unit C,  can be monitored
within a week by a single monitoring team,  allowing three weeks to
complete repairs, rechecks, process work orders,  and recordkeeping
within the one-month period.  There are several  other  factors  that also
indicate that a month permits sufficient time to comply with the  leak
detection and repair requirements:   (1) repairs  are commenced  concurrent
with the onset of monitoring as specified in the repair requirements;
(2) more than one monitoring team may be employed to perform the  inspec-
tions; and (3) the use of soap screening may reduce the monitoring time
required.  In addition, since the standards for  valves  allow quarterly
monitoring of valves not found leaking for 2 consecutive monthly
inspections, most valves are likely to be monitored on  a quarterly
basis.
     EPA recognizes that it is possible for individual  facilities to
expend less time monitoring than the EPA estimate of two man-minutes
per valve, as discussed above, and possible to expend more time monitoring
as the commenters imply.  Nevertheless, EPA maintains that the basis
for estimating labor hour requirements as presented in the BID for the
proposed standards appropriately reflect the monitoring requirements
for the petroleum refining  industry.   Hence, EPA believes that sufficient
time is allowed  in the period of one month to complete the leak detection
and repair  requirements.
2.2.5  Repair
Comment:
     Several  commenters wrote that the 5 and 15 day repair intervals
should be extended.  Some  (IV-D-8 and  IV-D-15) argued that an initial
attempt at  repair within 5  days  should be extended  because of holidays
and weekends.  Others  (IV-D-5,  IV-D-12,  IV-D-16,  IV-D-17, and IV-D-24)
thought that  the 5-day  requirement was unnecessary  provided repair was
accomplished  within  15  days.  Another  commenter  (IV-D-18) requested
that the  repair  intervals  be  extended  to 15 days  for initial  repair and
30 days for final repair.
     Conversely, another commenter  (IV-D-30) contended that the  first
attempt at  repair for  valves  and pumps should be  made within  24  hours

                                 2-31

-------
instead of 5 days.  The commenter wrote that repair personnel  should
either accompany or trail the monitoring personnel  and,  therefore,  the
minor repairs (e.g., tightening a valve bonnet)  could be completed  (and
rechecked) immediately.
Response:
     The standards require that a first attempt  at  repairing a leaking
valve or pump should be accomplished as soon as  practicable but no
later than 5 days after detection of a leak.  Attempting to repair  the
leak within 5 days will help maintenance personnel  identify the leaks
which can be repaired without shutdown of the process unit.  Valves or
pumps 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 leaking valves that require more
than simple field repair.  The 15 days provides  sufficient time to  sche-
dule and effect on-line repairs that a shorter period might not allow.
Provisions have been made for delaying repair of those valves which are
in critical service and cannot be bypassed.   The two repair period
requirements provide efficient reduction of emissions and allow suffi-
cient time for flexibility in scheduling repairs of leaking equipment.
A single period would simply permit delays in repairs that could
otherwise be accomplished quickly.
     Most valve repairs can be done quickly.  This  is evident from
compliance experience of refineries with the South  Coast Air Quality
Management District Rule (Rule 466.1) for valves which requires repair
within 2 working days.  A 5-day period for initial  attempts provides
sufficient time to schedule field repair.  Originally, EPA was considering
a 3-day limit, but decided to increase the limit to 5 days to allow for
holidays and weekends.
     Requiring an initial attempt at repair within  a shorter time
period (e.g., 24 hours as suggested by one commenter) may, however,
pose a significant problem to owners.  With shorter repair periods, a
repair crew would have to accompany or closely follow a monitoring
crew, repairing leakers as they are detected in  order to perform all
initial attempts at repair within the required time interval.  Although
this is a repair technique often employed, some  repairs cannot be
performed as described in the comment.  For example, a pump seal that
                                  2-32

-------
 is leaking may not be  repairable while the pump is in operation due to
 casings which must be  removed or safety hazards due to the shaft motion.
 In other instances, parts (such as a valve bonnet pressure plate) may
 be cracked and require replacement, making "on-the-spot" repair attempts
 impossible.  In addition, shorter time periods may increase the cost
 of leak detection and  repair.  Because very few valves leak (about
 10 percent initially and about 2 percent leaking per month in subsequent
 inspections), repair crews may spend much of the time on an inspection
 with few repairs to perform if they were to accompany the monitoring
 personnel.  The logistics of coordinating monitoring and repair is
 further complicated when considering union regulations that may apply
 and that certain repairs require specially trained personnel  to perform
 (e.g., control  valves).  EPA considers 5 days to be a reasonable time
 constraint for  first repair attempts on leaking valves or pumps.
 Comment:
     One commenter (IV-D-30) indicated that EPA chose the 10,000 ppm
 leak definition because undirected repair attempts for leaks  less than
 10,000 ppm would lead to an increase in emissions.  The commenter
claimed that lowering the 10,000 ppm leak definition and requiring a
directed repair program would not significantly increase the  cost
 impacts of the  standards and would produce a substantial reduction in
emissions.
Response:
     The "leak  definition" is the instrument reading observed during
monitoring that defines which sources require repair.  The best leak
definition would be the one that achieved the most emission reduction
at reasonable costs.  At a leak definition of 10,000 ppm, approximately
 90 percent of the mass emissions from valves would be detected.  EPA
has determined  (Document Nos. II-A-21, II-A-26 and II-A-42) that valves
 found leaking at levels of 10,000 ppm or greater can be brought to
 levels below 10,000 ppm with proper maintenance.  A leak definition
 lower than 10,000 ppm may be practicable in a 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.
                                  2-33

-------
     The commenter suggested that EPA require directed repair,  whereby
the tightening of packing is monitored simultaneously and continued
until no further reduction of leak is observed from the valve.   However,
there is no evidence to support the commenters contention that  directed
maintenance will provide greater emission reductions than the requirements
of the standards.  The standards require owners/operators to continue to
attempt repair if the initial attempt to repair a leaker fails  to reduce
emissions below 10,000 ppm.  The standards require monitoring of a
valve following attempted repair to determine if the repair attempt was
successful.  EPA also believes that requiring directed repair could be
too costly.  Directed repair may unreasonably complicate coordination
of monitoring and repair personnel, especially in refineries where
repair personnel are governed by union regulations.
     Upon reviewing the comments, EPA has maintained the 10,000 ppm
leak definition because it would address approximately 90 percent of
the VOC emissions from valves at reasonable costs and reasonable cost
effectiveness.  Also, the final standards for valve repair remain
unchanged from proposal in requiring the best practices, including
monitoring following repair, because directed repair has not been
demonstrated to be more effective in emission reduction and may have
higher costs.
Comment:
     One commenter (IV-D-30) was concerned that plant owners or operators
may abuse the delay of repair provision that can be used when stocks of
spare valves have'been depleted.  The commenter stated that this provision
invites operators to maintain very small inventories of spare parts.  A
better approach suggested by the commenter is to require the operators
to maintain sufficient stocks, and the inventory required should be
readily determinable after monitoring several times.
Response:
     The commenter appears to misinterpret the intent of the delay of
repair provision.  The standards require that owners or operators must
show that valve assembly supplies had been sufficiently stocked before
the supplies were depleted.  This includes custom-order and unique parts,
as well, to avoid delays of  repair due to unavailability of parts.
Despite what the commenter says, the provision does not invite  operators
                                  2-34

-------
to maintain small inventories of spare parts.  Plant experience with
delays will be considered if delay was reasonable.
2.3  PUMPS
2.3.1  Basis for Standards
Comment:
     One commenter (IV-D-4) recommended that the EPA decrease the
monitoring frequency for pumps based on the lower percentage of pumps
found leaking during a recent refinery inspection.  The commenter also
suggested a skip-period alternative for pumps.
Response:
     The basis of the standards for pumps is monthly leak detection and
repair.  One month provides the most effective leak detection and repair
program for pumps, reducing emissions from a Model Unit B by 11.5 Mg
per year, without imposing difficulties or unreasonable costs in
implementing the program.  EPA has determined (Document No. IV-B-2)
that monthly monitoring has reasonable cost effectiveness, $158/Mg, and
incremental cost effectiveness, $170/Mg between quarterly and
monthly monitoring.
     EPA data collected during screening studies on pump seals represent
plants with and without existing control  programs.  EPA data represent
pumps found in refineries throughout the nation.  No data were submitted
by the commenter to substantiate the contention that some pumps have
distinctly lower leak frequencies.  There may be many reasons that the
lower leak frequency was found in his plant.  One reason may be that a
control program was recently implemented.
     Skip-period monitoring for pumps has not been included in the
standards for two reasons.  The first is that pump seal failures are
sudden events independent of prior leak history.  Valve leaks (where
skip-period monitoring has been used), in contrast, gradually increase
over time, so that leak history is a factor in the leak status of any
one valve.   A skip-period monitoring program for valves achieves
emissions reduction because the number of valves leaking gradually
(and very slowly for process units that can use this alternative)
increases over the monitoring period.  However, skip-period monitoring
for pumps would allow large emitters to leak for a long period of time
                                  2-35

-------
because pump seals begin to leak suddenly.  Secondly, the number of
pump seals that must be monitored is not large enough to develop a
meaningful statistical program.  For example, a large process unit
(Model Unit C) would only have about 40 pump seals in light liquid
service.  Hence, an allowable percentage of pumps leaking, for example
2.0 percent, would not even allow a single pump leaking.  EPA has
provided other alternatives to monthly monitoring, which include
(1) installation of a properly designed dual mechanical  seal  as specified
in Section 60.592(d), (2) installation of an enclosed capture/conveyance/
control system as described in Section 60.592(f), and (3) use of leakless
equipment as provided in Section 60.592(c).
     Since proposal, the cost basis of leak detection and repair programs
for pump seals has been revised to assess pump repair on a consistent
basis with information presented in the AID (Document No. II-A-41).  In
the proposal BID, pump seal repair costs are based on 80 labor hours
per pump seal repair.  This basis has been revised to 16 labor hours
per seal repair plus the cost of a replacement seal ($140/seal, May 1980
dollars).  The cost effectiveness which appears in the preamble for the
proposed standards has been revised to $158/Mg VOC emission reduction
(Document No. IV-B-2).  EPA believes the revised cost effectiveness for
pumps is reasonable.
Comment:
     Two commenters requested that there be exemptions for pumps.  One
of these commenters (IV-D-8) indicated that "some pumps may not be able
to accommodate dual seals.  Accordingly, the standards should provide
exemptions from the requirement to install dual seals if 1) dual seals
cannot  be retrofitted to the existing pump, i.e., the pump would have
to be replaced to install dual seals; and 2) if a compatible barrier
fluid cannot  be found."  The other commenter (IV-D-12) indicated that
certain reciprocating pumps should be exempt due to the prohibitive
cost  of bringing  reconstructed reciprocating pumps into compliance.
Also, if an  owner or  operator installs a dual seal, that pump should be
exempt  from  routine monitoring.
Response:
      EPA  recognizes that some pumps may not be readily retrofitted with
dual  mechanical seals, although circumstances where a dual mechanical
seal  cannot  be  retrofitted without  replacing the entire pump are rare.
                                   2-36

-------
An exemption for these few pumps is necessary.   The standards  do not
require dual mechanical  seals, but only require satisfactory performance
under the leak detection and repair program.   Leak  levels  below 10,000
ppm organic concentration at the surface (which remains  the only require-
ment for pumps) may be obtained by replacing  the original  seal  or the
original seal packing.  Should these measures fail  to reduce the leak
rate to an acceptable level  (an intrument reading of less  that 10,000
ppm), the seal area may be enclosed, and the  enclosure vented  to a
control device.
     Availability of compatible barrier fluids should rarely  pose a
problem.  As discussed in the BID for the proposed standards  (pp 3-4
through 3-6) dual mechanical seals may be arranged in either  of two
configurations, back-to-back or tandem.  The  tandem arrangement utilizes
a barrier fluid pressure lower than the process fluid pressure at the
pump seal, such that any leakage at the primary pump seal  results in  a
leakage of  process fluid into the barrier fluid.  Such a sealing arrange-
ment will prevent contamination of the process fluid by the barrier
fluid.  The barrier fluid must be purged, however, to a controlled
degassing reservoir to prevent the leaked process fluid from eventually
being emitted to the atmosphere.  In the back-to-back arrangement the
two seals provide a closed cavity between them and a barrier  fluid is
circulated through the cavity at an operating pressure greater than the
stuffing box.  Barrier fluid leaking across the primary pump  seal will
enter the stuffing box and mix with the process fluid.  Barrier fluid
going across the secondary seal would release to atmosphere unless
captured by a vent control system.
     Reciprocating pumps may also be maintained in  compliance with the
leak detection and repair program requirements.  EPA recognizes, however,
that maintaining adequate sealing for less than 10,000 ppm organics
concentration around  linear motion shafts may  be difficult.  However,
the  seal area  (or distance  piece) of such pumps may be enclosed, and
the  enclosure  vented  to a control device.  As  such, an exmption  for
these  pumps is  not necessary.   The  commenter provided no  information or
data to support  the statement that costs of compliance for reciprocating
pumps  is prohibitive.
                                   2-37

-------
     If an owner or operator chooses to utilize an alternative control
technique for pumps such as dual  seals or enclosed and vented seal
areas rather than monthly monitoring, these pumps are exempt from the
routine instrument monitoring.  However, several  criteria would need to
be met for these alternative controls such as weekly visual  inspections,
continuous monitoring of the barrier fluid system for seal failure
detection, and daily barrier fluid level checks.   These are required to
ensure the integrity of the alternative control system and to remain
exempt from monthly monitoring.  In order to clarify the intent of this
requirement, the final regulation will include a  definition for
"stuffing box pressure" as "the pumped product pressure at the primary
seal interface."
Comment:
     One commenter (IV-D-14) expressed concern that the alternative
standards for pumps "are essentially a barrier fluid standard because
in the dual seal system, one mechanical seal is still exposed to the
atmosphere, as is the case with a single mechanical seal.  But requiring
a dual seal system rather than defining the standard as the use of a
non-VOC barrier fluid on the seal exposed to atmosphere, precludes the
use of a single mechanical seal with the same barrier fluid even
though the two are equivalent from the standpoint of emissions."
The same commenter suggested that EPA simply establish a no detectable
emissions limit for the barrier fluid system.
Response:
     The commenter appears to be requesting an exemption from the
routine leak detection and repair program for single seal pumps with
non-VOC barrier fluids.  EPA does not have enough information to use in
evaluating such an approach, and the commenter did not suggest a means
of ensuring continued compliance with the standards such as the barrier
fluid requirements of the proposed standards.  EPA does not know how to
do this either.  For these reasons, single seal/barrier fluid systems
are not exempt from the standards.
     The standards, however, allow owners or operators to use equivalent
means of emission limitation as provided for in Section 60.484.  An
owner or operator subject to the standards may apply to the EPA for
determination of equivalence for any means of emission limitation that
                                  2-38

-------
achieves a reduction in emissions of VOC at least equivalent  to  the
reduction in emissions of VOC achieved by the controls  required  by the
standards.  Each owner or operator applying for an equivalence determination
is responsible for collecting and verifying test data to  demonstrate
equivalence.
     EPA did not require a "no detectable emissions"  limit for pump
seals because with the control technique specified (monthly leak detection
and repair) as the basis for the standards, pumps can  still leak.
Several types of pumps with auxiliary equipment (e.g.,  dual mechanical
seals that utilize a barrier fluid system, and enclosure  of the  pump
seal area), however, can achieve emission reductions  of VOC at least
equivalent to that achieved by a monthly leak detection and repair
program for pumps provided that they are operated under certain  conditions.
Seal less pumps do not have a potential leak area and,  therefore, are  at
least equivalent to monthly leak detection and repair  and dual seal
systems.  As with other leakless equipment, seal less  pumps would be
subject to an initial performance test (using procedures  specified in
Reference Method 21) to verify that the piece of leakless equipment
meets the "no detectable emissions" limit, and annual  rechecks to
ensure continued operation with "no detectable emissions."
Comment:
     One commenter (IV-D-21) questioned Section 60.592.2(d)(i) that
he said creates the requirement that barrier fluid be  at  a higher
pressure than the stuffing box.  The commenter said that this requirement
is impossible because the "barrier fluid pressure is  the  stuffing box
pressure."
Response:
     As discussed in Chapters 3 and 4 of the BID for the proposed
standards, dual mechanical seals consist of two seal  elements with a
barrier fluid between them.  By pressurizing the barrier fluid to a
pressure greater than the process pressure, any leakage in the primary
seal would result in the leakage of barrier fluid into the process,
while any leakage in the secondary (outer) seal would result in  leakage
of barrier fluid to the atmosphere.  Provided a non-VOC barrier fluid
is used, no VOC leakage to the atmosphere can occur.
                                  2-39

-------
     The intent of §60.592-2(d)(l)(i)  is that the barrier fluid be
maintained at a higher pressure  than the pumped product  as discussed
above.  Where dual mechanical  seals are used, several  pressures are
associated with the stuffing box,  including the pumped product, the
barrier fluid pressure, and the  atmospheric pressure on  the outside of
secondary seal.  Originally EPA  intended to require  that the barrier
fluid be maintained at a pressure  higher than the pumped product outlet
pressure.  However, the pressure of the pumped product at the primary
seal face (i.e., at the stuffing box)  may be different from the outlet
pressure.  As such, EPA decided  to clarify the requirement by requiring
that the barrier fluid pressure  be greater than the  pressure of the
pumped product at the stuffing box. The terminology used by the EPA has
led to the commenter's confusion.   To  clarify the requirement,  in
response to this comment, EPA will add a definition  of "stuffing box
pressure" to the final standards to indicate that the stuffing  box
pressure, for purposes of the standards, is the pressure of the product
at the primary seal face.
2.3.2  Monitoring
Comment:
     Two commenters (IV-D-8 and  IV-D-21) wrote that  the criteria for
visual pump inspections should be  revised.  One commenter (IV-D-21) stated
that liquid leakage from pumps should  be defined in  a more quantitative
manner.  The commenter suggested that  the criteria be three drops per
minute rather than the subjective  "indications."  Another commenter
(IV-D-8) maintained that monitoring with an analyzer should be  the sole
criteria for determining a leak.  In contrast, another commenter
(IV-D-24) thought that monitoring  of pumps is excessive  and unnecessary
because emission concentrations  measured with a portable analyzer are
erratic for mechanical seals. The commenter held that visual  checks
alone were an adequate means of  detecting leakers.
Response:
     The purpose of visual inspections of pump seals and barrier fluid
systems is to detect leaks.  Liquids dripping from the seal  area indicate
seal wear and may signal the beginning of seal failure or actual failure
of the barrier fluid system.  To prevent excessive wear  that could
possibly result in catastrophic  seal failure, the seal should be repaired
                                  2-40

-------
soon after leakage is detected.   Visual  inspections  detect  the  leaks
associated with such failures  of seal  systems.   Therefore,  any  visible
leakage from the seal area is  considered a  leak.   A  more  quantitative
approach, "such as three drops per minute,"  would  be no more  indicative
of a leak than the approach proposed by  EPA.  However, to define  better
what EPA considers "liquids dripping," a definition  has been  added to
the standards.  "Liquids dripping" means any visible leakage  from the
:eal including spraying, misting, clouding,  and  ice  formation.
     The results of the Western  Oil  and  Gas  Association (WOGA)  testing
of petroleum refining industry pumps (Document No. II-A-42) have  shown
that 44 percent of light liquid  service  pumps found  exceeding the
10,000 ppm action level also had liquid  leaks that were visually  detected.
Hence, visual inspections do find a significant  proportion  of "leakers"
and are an effective supplement  to instrument monitoring.  However,
large quantities of VOC can be emitted from leaking  pump  seals  even
when there is no visual indication of leakage.   Large leaks can occur
without forming liquid drops or obvious  indications  of liquids  dripping.
For example, emissions may be  sprayed as a  fine  mist, vaporize, or may
condense as ice.  Thus, pumps  require a  more precise measurement  method
(i.e., the use of a monitoring instrument)  to determine if  emissions
are equal to or greater than 10,000 ppm.
     There is a relatively high  degree of certainty  whether a pump seal
has an organics concentration  at greater than or less than  10,000 ppm
using Reference Method 21.  Pump seal failures are usually  sudden (not
a deterioration effect), such  that emission concentrations  are  either
well above or below the 10,000 ppm leak  definition for pump seals.
Further, no data were provided by the commenters to support the contention
that instrument measurements are erratic for pump seals.   Hence,  EPA
maintains that instrument monitoring is  effective and necessary to
identify leaking pump seals.
2.3.3  Repairs
Comment:
     Several commenters expressed concern about  the repair of pump
seals.  Two commenters  (IV-D-8 and IV-D-14) were concerned  that a high
portion of pump  seals  continue to leak  in excess of 10,000 ppmv following
attempted repair.  Two  commenters recalled that  in the WOGA study
                                  2-41

-------
 (Document  No.  II-A-42), the previously referenced industry pump seal
 testing, 40 percent of the pumps continued to leak after attempted
 repair, whereas, the cost effectiveness of the control technique is
 based on the presumption of 100 percent successful repair.  Some
 commenters (IV-D-8, IV-D-15, and IV-D-18) requested that the regulations
 allow delay of repair for pump seals.  One commenter noted that more
 than 6 months  (as allowed in the proposed standards) is sometimes
 needed to  retrofit dual seals.  Also, delay beyond unit shutdowns
 should be  allowed for pumps if there is a delay in equipment delivery.
 Response:
     Pump  seal manufacturers have indicated (Document No. IV-E-4) that
 their emissions testing shows 10,000 ppm to be a proper leak definition
 criteria and that properly installed and operated seals should easily
 meet it.  However, EPA recognized before the standards were proposed
 that repairing pump seals to achieve VOC emission concentrations to
 below 10,000 ppm may be difficult in some instances.  The specific
 reference to the WOGA study (Document No. II-A-42),  that "40 percent of
the pumps repaired continued to leak", does not necessarily apply to
 the standards for pumps.  Pumps with new seals, especially seals of
harder material, may have a run-in time of up to 48  hours of operation
 to seat properly (Document No. IV-B-17).  There is no evidence given in
the WOGA study to indicate that some of the pump seals that continued
 to leak following repair were not monitored within the run-in period.
Seal  replacement may in fact have been a much more successful  means of
 repairs, and reported as such, if pump rechecks were measured after the
 run-in period.  Also, in this study, pump seal  repair was not always seal
 replacement (e.g., tightening of seal packing), whereas the EPA cost
analysis was based on seal  replacement.
     EPA analyzed control  techniques for pumps that  might not be
 repairable to below 10,000 ppm.  The cost effectiveness of installing a
dual  mechanical seal  with a barrier fluid system was examined for pump
seals that (1) are known to be leaking and (2)  cannot  be  repaired  by
 relatively simple procedures (such  as replacing a. seal).   Based on
this, EPA found the cost effectiveness  of installing dual  seals to
 reduce pump emissions to be reasonable (Document No.  IV-B-5).   EPA
                                  2-42

-------
expects the use of dual seal  systems to comply with the repair require-
ments of the proposed standards.  However, because retrofitting these
systems cannot be completed in 15 days, EPA provided 6 months to complete
the repair.  Owners and operators also have the option of enclosing the
pump seal and venting the emissions to a control device.
     In response to the second group of commenters, EPA recognizes the
need for delay of repair if the repair necessitates process unit shutdown.
However, delay of repair for pumps beyond unit shutdown is not necessary
because the plant owner or operator can stock (without unreasonable costs)
enough spare seals and seal parts for repair to prevent shortage of
seal parts due to a delay in equipment delivery.  EPA proposed to allow
delay of repair beyond shutdown for valves that require replacement of
the entire valve assembly rf the owner or operator shows that a sufficient
stock of these assemblies had been maintained before the stock was
depleted.  However, there are substantially fewer pumps in process
units than valves, so stocking spare seals is not unreasonable.  In
addition, most refineries have a spare pump in place that can be operated
while the leaking pump is being repaired so it is not clear why many
repairs would ever need to be delayed to a process unit shutdown.  EPA,
therefore, does not consider it necessary to incorporate the delay of
repair provision into the final standards for pumps.
     Commenters were concerned that they would not be able to retrofit
dual mechanical seal systems within the required 6-month period (Section
60.592-2(c)(3)) for leaking pumps that cannot be repaired to achieve
emission concentrations below 10,000 ppm.  However, pump seal manufac-
turers (IV-E-6 and IV-E-8) have indicated that the 6-month requirement
to  retrofit a dual mechanical seal system is  reasonable.  Most dual
mechanical seals can be shipped from the manufacturer the day they are
ordered, and, in the event of an unusual or special order, dual seal
systems can be manufactured in  16 to 18 weeks.  Twenty-four weeks  (6
months) would allow for engineering and installation.   A  refinery  may
have some difficulty with installing more than  10  dual  seal systems in
a  given 6-month period.  However, EPA  does not  expect that more than 10
dual seals will be installed in a single  process  unit during a  given
six-month period.  Thus, EPA considers the decision to  allow owners or
                                   2-43

-------
 operators  to delay  repair of pumps up to 6 months reasonable if a dual
 seal  system must  be used.
 2.4   COMPRESSORS
 Comment:
      One commenter  (IV-D-12) commended EPA for providing exemptions for
 existing reciprocating compressors.
 Response:
      EPA did not  provide a blanket exemption for existing compressors in
 the proposed standards as the commenter implied even though EPA discussed
 that  certain reciprocating compressors might not be covered under the
 reconstruction provisions if retrofitting the required equipment was
 technologically or  economically infeasible (see 40 CFR 60.15(e)).  To
 make  EPA's intent clear and to reduce the burden of reviewing recon-
 struction determinations, EPA is explicitly exempting existing recipro-
 cating compressors  provided the owner or operator demonstrates that
 recasting the distance piece or replacing the compressor are the only
 options available to bring the compressor into compliance.  This exemption
 is necessary because the cost impact of installing the required control
 equipment or replacing the compressor is unreasonable.  These compressors
 will  be exempt from the standards until  they are replaced by new compressors
 or the distance pieces are replaced.
 Comment:
      Another commenter (IV-D-30) was concerned that  a case-by-case
 determination of the feasibility of putting controls on reconstructed
 reciprocating compressors would be burdensome to EPA or the States and
 would probably result in a blanket exemption.  The commenter requested
 that  EPA reexamine whether a specific definition of  the compressors that
 are appropriate to exempt can be written in lieu of  the case-by-case
determination.
 Response;
      It is impossible to fully  define all  applications of reconstruction
considered for existing reciprocating compressors.   EPA maintains,
however, that the determination of technological  or  economic feasibility
 (or infeasibility) to meet  the  standards  would  not be  burdensome  for
EPA (or State agencies delegated enforcement  authority),  considering
                                  2-44

-------
the few compressors that might fit into this exemption.   The exemption
applies only to those specific instances where the seal  area cannot  be
enclosed and vented without recasting the distance piece or replacing
the compressor.
     EPA has evaluated (Document No. IV-B-20) means of controlling
compressor leaks that may comply with the standards for  compressors  at
reasonable cost.  Based on the availability of reasonable control
options, EPA does not believe that the provisions for reciprocating
compressors wil1 result in a blanket exemption of all such compressors.
Comment:
     A commenter (IV-D-14) argued that the standards for compressors
do not provide an incentive to improve existing control  technology.
Another commenter (IV-D-8) wrote that "there is no justification for
establishing a separate and arbitrary definition of leak for compressors,"
thus EPA should use the 10,000 ppm leak definition.  Other commenters
(IV-D-14 and IV-D-16) asked EPA to allow quarterly monitoring as in  the
refinery CT6 (Document No. II-A-6).
Response:
     The standards for compressors do not deter the incentive to
improve existing control technology.  Refiners have the option of
employing mechanical seals with barrier fluid systems and controlled
degassing vents or may alternatively enclose the seal area and vent  the
captured emissions to a control device.  These are generally the only
techniques available to reduce VOC emissions from compressor seals.    In
addition, the  standards provide additional flexibility and  incentive  to
improve upon existing technology through the provisions of  Section
60.592-3(i) that allow an automatic  equivalence  for  no detectable
emissions.  Furthermore, an owner or operator may apply equipment or
procedures that achieve a reduction  in  VOC emissions  at least equivalent
to the  reductions achieved by the compressor control  requirements.
     The standards for compressors  require the  use  of mechanical seals
with barrier fluid systems and controlled degassing  vents.  Leakless
equipment  is allowed  as an alternative  to the mechanical  seal system.
Leakless equipment is considered at  least equivalent  to mechanical
seals  if they  can  be  shown to have  no  emissions.   Method  21 defines  no
emissions, or  "no  detectable  emissions," as  a minimal deflection of  the
                                   2-45

-------
portable instrument meter.  In the case of the proposed standards, this
is a reading of 500 ppm or less.  Hence, the 500 ppm is not a leak
definition as misconstrued by the commenter, but an instrument limited
definition of no detectable emissions as specified in Reference
Method 21.
     Quarterly monitoring is not allowed under the standards because it
would achieve significantly less emission reductions than the mechanical
seal system, other leakless control, or enclosure and venting to a
control device.  Even as noted in the refinery CT6 published in June
1978 (Document No. II-A-6), EPA has concluded that a leak detection and
repair program for compressors in refineries is generally not an effec-
tive control technique and, therefore, EPA did not consider it a viable
option as the basis for the standards.  Quarterly monitoring was recom-
mended in the refinery CTG due to the limitations of retrofitting
equipment controls on existing compressors.  The effectiveness of a
leak detection and repair program for compressors is limited because
repair of leaks for most compressors could not be accomplished without
a process unit shutdown and because some seals must leak to operate
properly.  Because shutdowns generally occur infrequently,  limiting the
emission reduction obtained from maintenance, and because repair of a
compressor seal would often involve the use of mechanical  seal  systems
or enclosure and venting to a control device, equipment controls are
used as the basis for the standards.  These equipment  controls  have a
reasonable cost effectiveness (see preamble for the proposed standards,
Table 1).
2.5  PRESSURE RELIEF DEVICES
Comment:
     One commenter (IV-D-19)  was concerned that the costs  for rupture
disks are overestimated in the BID for the proposed standards.   The
commenter said: (1) EPA costs are based on relief valves with the
rupture disk offset under it;  this practice and cost  is  not  necessary
and violates recommended industry codes;  (2)  the added  cost  for retro-
fitting a disk and valve is only valid for half of  the  field  installations
because the downrating of a valve as an ASME  requirement is  only  mandatory
for new installations and is  not required  for retrofit  installations;
and (3) in a recent API survey,  half of the companies  responded  that
                                  2-46

-------
they already use block valves under relief valves;  therefore,  the  block
valve really should not be included in the cost  of  installing  a  rupture
disk under a relief valve.  Another commenter (IV-D-12)  disagreed  with
EPA's decision to require rupture disks for pressure relief valves.
The commenter thought the incremental  cost of $930  per megagram  emission
reduction from quarterly leak detection and repair  to the use  of rupture
disks was unreasonable.  The commenter recommended  quarterly leak
detection and repair as an alternative control for  pressure relief
valves.
     Other commenters remarked concerning disk sizing.  One commenter
(IV-D-14) urged EPA to consider a case-by-case standard for pressure
relief devices because problems in sizing spool  and piping are likely
to arise as a result of added pressure drop in retrofit installations.
Another commenter (IV-D-19) in contrast wrote that most rupture disks
manufactured have flow coefficients (0.95 and higher) compatible with
relief valves manufactured, so that it is becoming common practice not
to downrate relief valves upon retrofitting rupture disks.
Response:
     The basis of the standards selected  for  pressure relief devices in
gas service is the use of rupture  disks.  Rupture disks eliminate
fugitive emissions of VOC through  the  relief  device unless an overpressure
occurs.  After an overpressure release, replacement of the rupture disk
once again  eliminates fugitive emissions  of VOC through the pressure
relief device.  Therefore,  a  "no detectable emissions" standard was
selected for  pressure  relief  devices.  The proposed standards for
pressure relief devices  require that they be  operated with no detectable
emissions as  indicated by an  instrument  reading of  less than  500 ppm
above  background and that they return  to  this condition within  5 days
following pressure  release.
     At  proposal,  EPA  considered the  incremental cost  effectiveness
between  quarterly  monitoring  (required by State implementation  plans)
and  rupture disks  (see  preamble  for the  proposed standards, Table 1)
and  determined  that  the  resulting  value,  $930/Mg VOC,  was reasonable.
The  commenter disagreeing with the reasonableness  of  this incremental
cost effectiveness has  offered no  specific  information suggesting that
this cost  effectiveness  level is  unreasonable.
                                   2-47

-------
      In  reviewing the  public comments, EPA re-evaluated  (Document No.
 IV-B-2)  the cost effectiveness and incremental cost effectiveness of
 different  levels of control  (quarterly and monthly leak detection and
 repair and the use of  rupture disks) for pressure relief devices as
 shown in Table A-2 of  Appendix A.  The cost effectiveness for quarterly
 and monthly leak detection and repair was estimated to result in savings
 of $170/Mg VOC and $110/Mg VOC, respectively.  The use of rupture disks
 has a cost effectiveness of about $410/Mg VOC.  The incremental cost
 effectiveness between  quarterly and monthly leak detection and repair
 is about $250/Mg VOC.  The incremental  cost effectiveness between
 monthly  leak detection and repair and rupture disks is about $l,000/Mg
 VOC.
     At  proposal, EPA  cost estimates were based on rupture disks with
 offset mounting to prevent damage to the relief valve by disk fragments
 as stated  in the BID for the proposed standards, Table 8-1.  EPA
 recognizes that the offset mounting may not be necessary, and that it
 could present a safety problem if it added significant pressure drop to
 the system.  In these  cases, EPA agrees that  an offset mounting would
 not or should not be used.  However, since owners or operators might
 use the  rupture disks with offset mounting, EPA did not revise the basis
 for the  rupture disk system costs, realizing  that the estimated costs
 to comply with the standards may be overestimated.
     The first commenter was also concerned that block valves should
 not be included in the cost analysis because  they are already installed
 in half  the refineries surveyed by API.  Even though the use of block
 valves may already be widespread, EPA expects that  some refiners would
 use them in the absence of the standards,  and, therefore, EPA decided
to continue to include them in  the cost analysis.  This will  result  in
 an overestimate of the nationwide cost  impact of the standards for
 pressure relief devices.
     Sizing problems in retrofitting rupture  disks  can be avoided
through  the selection  of  compatible disks  and disk  holders;  therefore,
there is no reason to  establish special  requirements for pressure
relief devices in process units  affected through modification or
                                  2-48

-------
reconstruction.   In addition,  refiners  have  the option  of  venting
pressure relief  valve emissions to a VOC control  device, such  as a
flare.
Comment:
     Two commenters (IV-D-8 and IV-D-21) stated that rupture disks
should not be used as the basis for judging  the leak rate  of pressure
relief valves because rupture disks are not  common in the  industry  as
mentioned in the BID for the proposed standards.
Response:
     EPA is required to establish new source performance standards
based upon best  demonstrated technology (BDT), not common  technology.
EPA has determined that rupture disks represent BDT for pressure relief
devices.
     The standards for pressure relief devices are based on the use of
rupture disks.  Because rupture disks eliminate emissions, EPA selected
a performance standard of "no detectable emissions."
Comment:
     Commenters (IV-D-5, IV-D-8, and IV-D-18) were concerned with the
safety of the standards for pressure relief devices.  Monitoring pressure
relief devices is inherently unsafe.  In addition, these components are
frequently difficult-to-monitor.  The practice of employing rupture disks
is unsafe due to the pressure build-up between the disk and relief device,
Response:
     Refineries routinely inspect pressure relief devices approximately
on an annual basis as a part of normal safety and maintenance procedures
to ensure the set pressure  is correct (Document No. II-D-22); therefore,
the standards are not requiring refineries to do a new task.  The
standards implicitly require performance tests  using Method 21 to
verify  that the device  is maintained at no detectable emissions.  This
test  is similar to testing  done by  EPA  and EPA  contractors  in collecting
data  for development of the standards and similar to testing required
by States under implementation  plans.   This  test could be scheduled
during  periodic inspections of  pressure relief  devices, which are
typical  of many industry  safety practices.   Monitoring should be done
by personnel who  understand the precautions  needed when monitoring
pressure relief devices.   Evidence  that operators can  safely monitor

                                   2-49

-------
 pressure relief devices is further indicated by one refiner's  (Document
 No. IV-D-33) practice of removing pressure relief devices  (for repair/
 replacement and testing) following overpressure releases.   Based  on
 this information and EPA's experience in collecting data for pressure
 relief devices, monitoring of these devices  can be done  safely.
      The standards for pressure relief devices  are based on the use  of
 rupture disks;  however, the standards do not require their use.   Alter-
 natively,  pressure relief device emissions can  be routed to a  VOC control
 device, such as a  flare.  If a rupture disk  is  used, a pressure sensor
 should be  installed to warn operators if a pressure increase has
 occurred between the disk and relief valve.   The  cost of a  such a
 sensor (0.6 cm  pressure gauge)  has been included  in the  cost analysis
 that  is presented  in the BID for the proposed standards, Chapter  8,
 Table 8-1.
 Comment:
      Two commenters  (IV-D-8 and  IV-D-21) disagreed with  the  leak  definition
 for  pressure relief  devices  stating  that there  is no justification for
 a  different leak definition  than  10,000  ppm.
 Response:
      The standards  for pressure  relief devices  are based on the use of
 certain  equipment.   This  equipment,  as explained in the preamble to the
 proposed standards,  results  in no  detectable  emissions.   The no detectable
 emissions limit  is 500  ppm according to Method 21 and is  related to
 monitoring  instrument  capabilities.  The 10,000 ppm leak  definition for
 pumps and valves was chosen  based on different considerations and  is
 unrelated to standards  that  require  no detectable emissions, such  as
 the standards for pressure relief devices.  A 10,000 ppm  or greater
 concentration indicates  a pump seal failure or deterioration of a
 valve packing, and concentrations below 10,000 ppm are  allowed.  The  no
 detectable  emissions level (500 ppm) indicates no emissions.
 Comment:
     Another commenter  (IV-D-21) requested the addition of  a qualifying
phrase to the standards, Section 60.592-4(b), such that pressure relief
valve monitoring only be required "after each pressure  relief,  of  which
the operator has knowledge."  The commenter wrote  that this  clause is
                                  2-50

-------
necessary because pressure relief to the atmosphere is not always  known
by the owner or operator.
Response:
     The intent of the standards for pressure relief valves is to
control emissions at all  times except during an overpressure relief.
Therefore, the standards  require that pressure relief valves return to
no detectable emissions as soon as practicable, but no later than  5
calendar days after the pressure relief.  Pressure sensors between the
rupture disk and pressure relief device can alert operators in the event
of a pressure relief.  Owners or operators may also be alerted that
pressure release has occurred from instrumentation in a unit control
room or by visually or audibly detecting a release.
2.6  SAMPLING SYSTEMS
Comment:
     A commenter (IY-D-14) wrote that an exemption in the proposed
standards for sampling systems should be allowed when for example, a
sample might be drawn after a heat exchanger or cooler, and there  is
not enough pressure available to return it to a lower pressure source.
The commenter suggested an alternative to closed loop sampling. He
recommended simply to require accumulation of the purged material  in
another container for proper disposal.  Another commenter (IV-D-4)
maintained that the application of the standards for sampling connection
systems for "low vapor pressure liquid streams is not cost effective
with respect to reduction of VOC emissions."
Response:
     The standards do not require "closed loop" sampling (although
it may be used to comply  with the standards) but do require a "closed
purge system" as one commenter suggested.  Using closed purge sampling,
an owner or operator could simply collect purged materials and properly
dispose of them by any system that collects the VOC and destroys or
recovers the VOC without  emissions to atmosphere.
     EPA recognizes that some sampling connections are located at
points that would have insufficient pressure to return purged fluid in
a closed loop.  For example, low line pressure (resulting from pressure
drop in the final product coolers or a phase change) is characteristic
                                  2-51

-------
 near  plant boundaries.  Hence, EPA expects that owners/operators would
 use closed purge to comply with the standards.  If a plant owner or
 operator chooses to retrofit a closed loop sampling connection in these
 instances at a location of higher pressure (e.g., near a pump), and if a
 sample cooler is not in place at the selected location, a cooler may
 have  to be installed to ensure safe handling of hot materials.
      Retrofitting a closed loop sampling connection at a location
 that  necessitates a cooler, however, is not expected to occur often.
 Most  sampling connections are located near pumps where line pressure is
 not a problem and where cooling systems are already in place (Document
 No. IV-B-6).  Nevertheless, EPA has estimated the additional  cost
 of retrofitting a closed loop sampling system with a cooler.   The
 addition of a sample cooling system increases the cost effectiveness of
 the sampling system from $810/Mg to $l,450/Mg.
     The standards for sampling connections include low vapor pressure
 liquid streams.  Heavy liquid streams have the potential  to emit VOC's
 to atmosphere, particularly from purged sampling materials  that are
 likely at elevated temperatures.   The emission factor developed for
 sampling connections is based on  both light liquid and heavy liquid
 streams.  The cost effectiveness  estimate of $810/Mg is based on closed
 loop sampling.  However, the standards allow closed purge sampling
which would likely be used for low vapor pressure streams at an even
more reasonable cost effectiveness.
Comment:
     Another commenter (IV-D-4)  suggested an exemption for  sampling
connections in units that become  affected facilities through  modifi-
cation or reconstruction if retrofit costs exceed that of a comparable
installation in a new unit.
Response:
     The control  costs presented  in  Chapter 8  of  the BID for  the proposed
standards  for sampling systems  are  likely overstated because  they are
based on closed loop sampling.  These costs included retrofit conside-
rations.  The cost effectiveness  of  closed loop  sampling is estimated
to be $810 per Mg (preamble to  the  proposed standards).  It  is  possible,
however, that in  some  situations  retrofit costs  for  using closed loop
                                 2-52

-------
sampling will exceed that of the cost of a new sampling system.   [The
example given in the previous comment indicated that owners/operators
may retrofit a closed loop sampling system by adding a sampler cooler.
Although it is unlikely that retrofitting a closed loop sampling con-
nection at a location that necessitates a cooler would occur very
often, EPA evaluated the cost effectiveness to retrofit a closed loop
sampling system by adding the cost of a dedicated sample cooler.  EPA
determined that the cost effectiveness of the system is still  reasonable,
$1,450 per Mg VOC (Document No. IV-B-6).]  If a specific plant would
incur extra costs, EPA would not consider this unreasonable.
2.7  OPEN-ENDED LINES
Comment:
     One commenter (IV-D-8) questioned the operational requirement of
closing the inner valve prior to closing an outer valve on open-ended
lines.  The commenter wrote that this requirement is unenforceable and
of no benefit if the inner valve leaks.
Response:
     The standards require open-ended valves to be equipped with a cap,
plug, or a second valve.  If a second valve is used, the upstream valve
is required to be closed first 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.  While it is true that this and many other sound practices
are not 100 percent enforceable, this requirement is enforceable if an
inspector finds that the upstream valve has not been closed at all.
     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.  In this situation the second
valve would control VOC emissions.
Comment:
     Another commenter (IV-D-14) recommended that open-ended valves
and lines be included in the valve standards (i.e., leak detection and
                                  2-53

-------
 repair) with an exemption for pluyged valves.  The commenter was concerned
 that any open-ended valve could result in a violation.  Concern was
 also expressed that requiring plugs on pump case valves could cause
 premature failure of the welded connection at the pump case.  Also, the
 standards would require plugging bleed valves "out-of-service" in a
 block and bleed arrangement.
 Response:
     Open-ended valves are not included in the valve standards because
 leak detection and repair for open-ended valves does not represent BDT.
 Leak detection and repair would achieve less emissions reduction and
may cost more to implement than the equipment and operational standards
 for open-ended valves because of repeated inspections of nonleaking
 sources.  The use of a leak detection and repair program for the control
of VOC emissions from open-ended valves or lines would be inappropriate.
     The standards for open-ended valves provide refineries with the
flexibility to add either a cap, plug, blind flange, or a second valve
depending upon the individual application.  Pump case valves, for
example, could be double valved to avoid the risk of premature failure
of the welded connection at the pump case caused by frequent removal  of
a cap or plug.
     Upon reviewing the comment that the standards would require plugging
bleed valves "out-of-service" in a block and bleed arrangement,  EPA
decided to provide an exemption in the final  standards for open  ended
lines in a double block and bleed arrangement when venting the space
between the two block valves.  However, when the bleed valve is  not
opened, it must be capped.
2.8  FLANGES, LIQUID SERVICE RELIEF VALVES,  AND HEAVY LIQUID SERVICE
     VALVES AND PUMP SEALS
Comment:
     One commenter (IV-D-8)  maintained that  the results of a number of
studies support an exemption for equipment in low vapor pressure  service.
The commenter noted that an  EPA study (Document No.  II-A-19) of  two
refineries in the South Coast Air Basin had  monitored 664  components  in
light and heavy liquid  service (which were exempt from South Coast  Air
Quality Management District  rules)  and found  only four leaking components,
none of which were in  heavy  liquid  service.   Another  study  found  only

                                  2-54

-------
one leaking valve out of 519 in heavy liquid service and  only
3.8 percent of the pumps leaked.  Another commenter (IV-D-15)  also
supported excluding pumps in heavy liquid service,  pressure relief
valves in light liquid service, flanges, and connections  from  routine
monitoring.
     Another commenter (IV-D-21) wrote that there is no demonstrable
cost effectiveness to the inclusion of pumps and valves in  heavy  liquid
service, pressure relief devices in both light and  heavy  liquid service,
and flanges and other connections, and further, these sources  should be
exempt from all requirements as indicated by EPA at the National  Air
Pollution Control Techniques Advisory Committee (NAPCTAC) meeting on
June 3, 1981.
Response:
     The final standards for valves and pumps in heavy liquid  service,
pressure relief valves in liquid service, flanges,  and connections
exempt these sources from routine leak detection and repair.  The low
leak frequency and emission factors for these sources compared to
sources subject to the leak detection and repair programs,  as  discussed
at the June 1981 NAPCTAC meeting, indicate that the cost  of routine leak
detection and repair is not warranted by emission reduction.  However,
Section 60.592-8 provides that if evidence of a potential leak is
found, the piece of equipment must be monitored within 5  days, and
repaired as soon as practicable within 15 days if an instrument reading
of 10,000 ppm or greater is detected.
     For those components that are found leaking, however,  EPA has
demonstrated that the cost effectiveness of repair is reasonable.
(Document No. IV-B-5).  The cost effectiveness of repair  for leaking
flanges, heavy liquid pumps, and heavy liquid valves varies from  a
savings of about $180/Mg to a savings of about $90/Mg.  For pressure
relief devices in liquid service, repair costs are not considered to be
attributable to the standards.  These components should be properly
maintained for safety reasons in the absence of a repair  requirement.
     The SCAQMD regulations for valves, Rule 466.1, do not cover VOC
less than or equal to 1.5 psi RVP.  The NSPS, however, includes valves
that are less than 1.5 psi RVP.   In reviewing the commenter's request
                                  2-55

-------
to exempt these valves, EPA notes that these valves do leak.   An EPA
study [Document No. IV-A-3] found that 3 of the 175 light liquid valves
with vapor pressures less than 1.5 psi RVP leaked.   In addition, raising
the heavy liquid service cutoff to 1.5 psi RVP would affect a small
percentage of valves.  In the study previously cited, the 175 valves
represented only 2.4 percent of the total  valves that would be subject
to the standards.  Hence, considering:  (1) that the basis of the
heavy liquid/light liquid split is easily determined, (heavy  liquids
have vapor pressure equivalent to or heavier than kerosene),  (2) that
light liquid valves servicing less than 1.5 psi RVP streams do leak, and
(3) that a small percentage of refinery valves would be affected, EPA
has retained the light liquid definition.
2.9  CONTROL DEVICES
Comment:
    One commenter (IV-D-21) wrote that since flares are not an affected
facility or a fugitive emission source they should  not be regulated.
Response:
    Flares are one of several  VOC control  devices that might  be used
to comply with the standards.   These control  devices are used to reduce
emissions of VOC that might otherwise be emitted to the atmosphere
uncontrolled.  If flares and the other control  devices were not
specifically regulated they might be operated at conditions which would
result in inefficient combustion and inadequate emission reductions.
The EPA has determined that a flare can be operated at conditions which
assure better than 95 percent emission reduction.   Flares operated in
this way are an acceptable alternative to other control  devices used to
comply with the standards.  Section lll(h)(l) provides that EPA may
promulgate design and operational  requirements (like the requirements in
these standards for other control  devices) to assure BDT - level  control
and that EPA include such requirements as will  assure proper  operation
and maintenance of any such element of design or equipment.   Therefore,
it is appropriate to specify operational  requirements for flares used
to comply with the control device standards developed under Section
Comment:
     Several  commenters (IV-D-4,  IV-D-8,  IV-D-12,  and  IV-D-15)  requested
that the requirements for flares  be  deleted,  including the  provision
                                  2-56

-------
that compliance be determined by Reference Method 22.   They  recommended
that they be replaced with provisions that flares function in  accordance
with good operating practice with an attached flame and no visible
emissions except for periods not to exceed a total  of  5 minutes during
any 2 consecutive hours.
Response:
    Data developed by a Chemical Manufacturers Association (CMA)  -  EPA
flare test program (Document No.  II-A-43) show that some types of
flares meeting certain conditions achieve better than  98 percent emission
reduction.  Consequently, the EPA concluded that design and  operational
standards which require flares to be operated at the conditions determined
by the tests would assure better than the required 95  percent  emission
reduction.  The term "good operating practice" has no  accepted engineering
meaning.  There is no evidence that flares give better than  the required
95 percent emission reduction at all velocities at which the flame
remains "attached." Therefore, requiring that flares function  in accordance
with good engineering practice with an attached flame  does not assure
better than the required 95 percent emission reduction.  The standards
do require that there be no visible emissions except for periods not to
exceed a total of 5 minutes during any 2 consecutive hours.   Reference
Method 22 describes the procedure used to determine whether  the flare
meets this visible emission requirement.
Comment:
    Several commenters (IV-D-6, IV-D-7, IV-D-8, IV-D-15, and IV-D-16)
wrote that the definition of a "flare" was too restrictive and should
be revised to allow many flare designs which are currently in use in
refineries.  The commenters specifically opposed the definition because:
(1) multiple burner arrays are efficient, (2) there is no relationship
between destruction efficiency and flare elevation, and (3)  most flares
operate with turbulent diffusion flames rather than strict diffusion
flames.  One commenter also stated that the definition should allow
automatic  ignition systems.
Response:
    The  definition of a flare is restrictive  in  that the types of flares
permitted  are limited to those on which data  are available.   Other
types of  flares may give better than the  required 95 percent reduction

                                  2-57

-------
 under certain conditions.  EPA has underway a program to determine the
 efficiencies of some other types of flares used in the petroleum and
 SOCMI industries.  As this information becomes available to EPA, the
 requirements for flares could be changed, if it is appropriate.   If an
 owner or operator chooses to use another type of flare, he may use the
 equivalency procedures to demonstrate that the flare should be allowed
 by EPA.   EPA accepts that there is no relationship between flare destruction
 efficiency and flare elevation.   Accordingly,  the  definition of  control
 device has been changed to exclude the term "elevated."  Also, automatic
 ignition systems are not disallowed by the regulation.
 Comment:
     Other commenters (IV-D-6,  IV-D-7,  IV-D-15,  and IV-D-16)  urged
 EPA to revise  the  design and  operational  requirements  for  flares.   The
 commenters noted that the maximum velocity of  22 m/sec  would greatly
 increase flare  costs.   It was  also  suggested that  the assignment of
 minimum  heating  values  for flares be  related to the  relief gas composi-
 tion.
 Response:
     The  maximum  velocity value of 22 m/sec was changed  since  proposal
 to  18  m/sec.  This change  was based on  further evaluation of  the data.
 The  revised exit velocity  (for steam assisted flares),  is the highest
 velocity tested  in the  flare tests sponsored by CMA and EPA  (Document
 No.  II-A-43).  EPA has  underway a program to determine if better than
 the  required 95  percent  emission  reduction can be maintained at higher
 velocities.  If  EPA concludes that high emission reduction can be
 maintained at velocities greater  than 18 m/sec, EPA will change the
 maximum  velocity value accordingly.  An operator would have the option
 of demonstrating to EPA that use of a flare at conditions other than
 those specified, would result in emission reduction equivalent to 95
 percent  control, the level selected as BDT for control  devices.
 Comment:
    One commenter (IV-D-25) stated that flares  should be exempted
 during start-up and shut-down  periods  and for  10 minutes during a
2-hour period.   Even under ideal  conditions, unit upsets may  cause
incidents exceeding the 5 minutes  per  2 hours exemption  for visible
                                 2-58

-------
emissions.  One commenter (IV-D-6)  suggested that EPA clarify  that  a
visible flame does not constitute a visible emission.
Response:
    Start-up and shut-down periods  and any "unit upset"  due  to
malfunctions are covered by the General  Provisions.   Section 60.8(c)
states that "Operations during periods of start-up,  shut-down, and
malfunction 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,
shut-down, and malfuction be considered a violation  of the  applicable
emission limit unless otherwise specified in the applicable  standard."
     The 5-minute limit for visible emissions (within any 2-hour  period)
is consistent with the flare requirement of the State of Texas where many
plants with smokeless flares are located.  EPA has no information,  nor
has any been submitted in this rulemaking, suggesting that  these  plants
cannot achieve this time limit.
     The standards require that Reference Method 22  [Section 114  of the
Clean Air Act as amended (42 U.S.C. 7414)] be used to determine the
compliance of flares.  This method involves the visual  determination of
visible smoke emissions from flares.  Section 3.4 of Method  22 clearly
states that "smoke occurring within the flame, but not downstream of
the flame, is not considered a smoke emission."
Comment:
     Another commenter (IV-D-24) objected to the requirement for
instrumentation to monitor flare operating parameters.
Response:
    With respect to instrumentation to monitor flare operating parameters,
an owner or operator is only required to use a heat  sensing  device  to
indicate the continuous presence of a flame.  Flares and other control
devices are required to be operated at all times when emissions may be
vented to them.  Other measurements for flares are required one-time
only or when requested by enforcement agencies for a compliance
determination.
                                  2-59

-------
 Comment:
      A commenter (IV-D-8) requested that the requirement of Reference
 Method 22 be deleted because fugitive emissions from flares are small,
 and existing flares are designed on a different basis.
 Response:
      As explained in the preamble to the proposed standards,  EPA selected
 design and operational  requirements for VOC  control  devices:   flares,
 enclosed combustion devices, and vapor recovery systems  that reflect
 the application  of the  best technological  system of  emission  reduction
 for these control  devices.   The  design and operation requirements for
 flares require smokeless operation.   Smokeless  operation  of a  flare  means
 that visible emissions  from a flare are to be less than  5 minutes in
 any 2-hour period  as determined  by  Reference Method  22.   Reference
 Method 22, hence,  provides  guidelines  for  assessing  visible emissions
 from flares and  is  included in the  standards to  ensure that flares
 achieve greater  than 95  percent  control.
 Comment:
      One commenter  (IV-D-30)  believed  that EPA has not adequately
 justified rejection  of flares as the sole  basis  for  the standards for
 control  devices.  The commenter  stated  there was  no  cost analysis in
 the BID to support  EPA's  belief  that flares are too costly  if they are
 built solely  for fugitive VOC control.  The commenter contended that,
 where point sources  of YOC  and fugitive sources are controlled by a
 single  flare, 98 percent  control efficiency should be required.
 Response:
      Existing control devices were selected as part of the best
 technological system  of emission reduction for fugitive emission.  EPA
 believes  that most in-place flares and enclosed combustion devices are
 designed and can achieve an average destruction efficiency of about
 98 percent.  Existing vapor recovery systems  can be operated to achieve
 at least 95 percent emission reductions and are an attractive control
 option  in  that some product may be recovered  and realized as an energy
credit  (e.g., process heaters).   Flares were  not selected as the sole
basis for this portion of the standards, as the commenters requested
because the cost  of requiring owners and operators to replace 95 percent
efficient control devices already in place in existing refineries with

                                  2-60

-------
98 percent efficient devices is unreasonably high in light of the
small  additional  emission reduction achievable from these equipment leaks.
The standards, therefore, require 95 percent control (which allows use
of existing vapor recovery systems that can achieve 95 percent,  but not
98 percent control  in all cases)  although EPA expects that most  refiners
will utilize flares and achieve 98 percent or better control.
                                  2-61

-------

-------
                           3.0  APPLICABILITY
3.1  AFFECTED FACILITY
Comment:
     Two  commenters (IV-D-8 and IV-D-21)  questioned whether an  "affected
facility" should be defined as a group of fugitive emission sources  of
VOC, noting that (in their opinion) the definition is inconsistent with
the terms of the Clean Air Act.  A commenter (IV-D-21)  stated that
Regulation 40 CFR 60.2 requires that an "affected facility" be  an
apparatus and "a group of fugitive sources is not an apparatus."  These
commenters implied that all equipment (including equipment not  affected
by the requirements of the standards) should be included in the affected
facility  for process units.  One commenter (IV-D-8) stated that the
control of fugitive emissions is significantly different from controlling
emissions from new stationary sources generally covered by NSPS.
Fugitive  emissions control "involves continuous tightening or repairing
of thousands of individual components, each of which emits relatively
small amounts of emissions," whereas with most other stationary sources
subject to NSPS, "once the control equipment is installed routine
maintenance is generally required."  Based on these positions the
commenters stated that control of fugitive emissions through new source
performance standards is unworkable.
Response:
     In choosing the designation of affected facilities, EPA examined
fugitive emission sources of VOC in light of the terms and purpose of
Section 111 of the Clean Air Act.  The Clean Air Act mandates the EPA
to set standards for any pollutant emitted from a category of new or
modified "stationary sources."  Section lll(a)(3) of the Act defines
the term "stationary source" to mean "any building, structure, facility,
or installation which emits or may emit any air pollutant."  The
                                  3-1

-------
 pieces in VOC service of equipment in  a  process unit, viewed  in the
 aggregate, are a "facility"  that may emit air  pollutants and, therefore,
 are appropriately considered as  a  "stationary  source".* How these
 pieces of equipment are  or are not aggregated  into affected facilities
 is carefully considered  by EPA.
      Since the purpose of Section  111  is to minimize emissions by
 application of the best  demonstrated system of emission reduction at
 new and modified sources (considering  cost, nonair quality health and
 environmental  impacts, and energy  requirements), there is a presumption
 that the narrowest designation (i.e.,  individual pieces of equipment)
 is proper.   However,  EPA rejected  the  equipment component designation
 for fugitive emission sources other than compressors; this decision is
 discussed in response to another comment in this section (see page 3-6).
 Consequently,  the next most  narrow  definition, the group of all  equipment
 components  (except compressors) within a process unit,  was considered.
 Review of the  relevant statutory factors did not lead to the conclusion
 that designating  each group of equipment components in  a process unit
*This agrees with the dictionary definition of "facility," meaning
 something designed, built, installed, ect., to serve a specific function
 or perform a particular service" (The Random House College Dictionary
 Revised Edition, 1975).  The group of equipment in VOC service covered
 by these standards is designed and installed to serve the specific
 function of handling the processing of petroleum products into
 intermediate or more refined materials.
     We note in this regard that the Court of Appeals for the District
 of Columbia Circuit has stated that:
     In designating what will constitute a facility in each particular-
     industrial  context, EPA is guided by a reasoned application of the
     terms of the statute it is charged to enforce, not by an abstract
      dictionary" definition.  This court would not remove this
     appropriate exercise of the agency's discretion.
578 F.2d 319, 324 n. 17 (1978).   EPA's selection of the group of fugitive
VOC emissions-related equipment as the affected facility reflects a
reasoned application of Section  111.   it assures that an identifiable
subset of refinery emissions—equipment leaks of VOC-- is controlled as
soon as the equipment responsible for those emissions is either modified
reconstructed, or newly constructed.   For the reasons explained in the  '
text below, a broader definition (e.g.,  all  the components  of a process
unit)  would simply delay that result.
                                  3-2

-------
 as  an  affected facility would cause adverse impacts.  Defining an
 affected facility as the group of equipment components, other than
 compressors, within a process unit would achieve similar emission
 reductions as designating individual components as the affected facility.
 (See discussion in comment on this point.) Therefore, the affected
 facilities for the standards are (1) compressors in petroleum refineries
 and (2) the group of equipment (pressure relief devices, open ended
 lines, sampling systems, valves, and pumps) in a process unit.
     Some of the commenters appear to be suggesting that the affected
 facilities should include equipment within a process unit even though
 the equipment is not an apparatus to which the standards apply.  Such
 an approach would mean equipment affected by the standards would not be
 required to use the best demonstrated technology (considering costs).
 EPA believes this approach would be inconsistent with Section 111.
 (No evaluation of the best demonstrated technology (considering costs)
 has taken place for these equipment at this time.   EPA may evaluate
 this for these equipment later.) Also, if EPA would follow this approach,
 increases in emissions from emission points not affected by the standards
 and changes in operation not related to the equipment covered by the
 standards could result in modifications.   In contrast, emission reductions
 resulting from the incremental  control of emission points not covered
by the standards could be used to offset increases in emissions resulting
from emission points covered by the standards and* therefore, would
preclude what otherwise might have been a modification.   EPA believes
this approach would be confusing.  Based on this consideration, EPA
rejected this approach.
     The commenter stated that control of fugitive emissions through
standards of performance is unworkable because the fugitive emission
sources covered by the standards do not include all of the equipment
within a process unit.   This is a practical  consideration only when
considering the modification and reconstruction provisions in Part 60.
For newly constructed sources,  the standards are clearly practicable.
The standards are well  defined and will result in the intended purpose
 of requiring the best demonstrated technology for equipment leaks of
VOC (fugitive emission sources of VOC).  For an owner or operator who
might be considering or determining a modification or reconstruction,

                                  3-3

-------
 however, this definition might pose some difficulties.  For example,
 determining the basis  (see definition of capital expenditure--40 CFR 60.2)
 of an existing facility is more difficult for this standard than for
 most other standards of performance (See Section 4.0 for more examples).
 EPA has provided alternative approaches to reduce the burden associated
 with these difficulties.  These alternative approaches were not provided
 to make this definition usable but to make it easier to use.  This
 issue and the alternative approaches are discussed further in Section
 4.0 and, to the extent it concerns the reconstruction provisions, in
 Section 5.0.
 Comment:
     Another commenter (IV-D-30) contended that EPA should define the
 affected facility as each individual fugitive emissions component based
 on "Section Ill's presumption for inclusiveness."  In addition, the
 commenter did not believe that EPA provided convincing reasons to
 support the decision to treat compressors individually and other
 components collectively (process units)  in defining "affected facility."
 The commenter contended that EPA's first reason for rejecting individual
 components (the cost of tracking individually covered sources)  is not
 persuasive because a simple color coding or tagging of new and existing
 components could be used.   Additionally, in response to EPA's second
 reason for rejecting individual  components as the basis for the affected
 facility, the commenter indicated that it does not appear  that there
 would be a significant difference in leak detection and repair costs
 between the "process unit"  definition  and the "individual  component"
 definition of affected facility.   This commenter also stated that he
 found no evidence in the preamble or the BID  for the  proposed standards
 for EPA's assertion that the  "process  unit"  definition  of  affected
 facility would achieve as much  emission  reduction  as  the "individual
component"  definition.   The commenter  believed that as  individual
components  are added to a  unit,  they are  covered earlier and achieve
further emission reductions than  fugitive  emission  sources  within
 process units,  which are not  covered by  the standards until  the entire
unit is replaced.
                                  3-4

-------
Response:
     In selecting the basis for the affected facility,  EPA considered
the effects of keeping track of individually covered sources.   As
discussed in the preamble to the proposed standards, components in
existing plants would be replaced one at a time and, therefore,  would
be covered by the standards one at a time.  Because components  in
existing plants are infrequently replaced, many adjacent components
would not be covered by the standards.  This would mean that a  plant
would be required to inventory all  the components  in a  plant and then
keep track of all activities for each component.  Even  though individual
components could be color coded or tagged, EPA believes that the effort
to keep track of and record activities for a mixture of individual
components within a plant would be tedious and costly.   In addition, as
discussed below, EPA believes that this effort would not likely result
in additional emission reductions, in particular,  during earlier
implementation of this approach.  In contrast to the recordkeeping
effort for individually covered sources, the effort for components
within process units would be less and would still result in more
immediate emission reductions.  Thus, EPA judged that maintaining an
inventory of individual components for an entire plant  would be unreasonably
burdensome, but maintaining an inventory for compressors or evaluating
components occasionally within process units would not  be unreasonably
burdensome.
     The commenter appears to have misunderstood the techniques used to
determine the cost effectiveness for leak detection and repair  programs.
The costs incurred for implementing such a program include fixed costs,
for example, the monitoring instrument and calibration  costs which  are
shared by the components monitored.  The fixed costs can be unreasonably
high if only a few components are monitored.  Additionally, EPA costs
are based on a specific time required for monitoring each component.
These monitoring times are based on the normal physical distribution of
components in petroleum refining process units.  If only a few  components
scattered throughout a plant are monitored, the time required per
component would be greatly increased.  These monitoring costs would be
unreasonable until enough components would be covered within a  certain
                                  3-5

-------
area of a plant.  This area may be smaller than a process unit but not
significantly smaller than a process unit.
     In response to the commenter's concern about emission reductions,
EPA attempted to select a basis for the affected facility definition
that would provide the largest emission reduction that is reasonable.
As discussed above, EPA most seriously considered two  approaches  in
defining the affected facility—the individual  "component" approach and
the "process unit" approach.  The largest emission reductions  are
usually associated with a component basis.  However,  for these standards
the difference between emission reductions of the component and process
unit approaches are unclear.  Based on the "component" approach all new
components—covered by the standards—and all  replaced components would
be affected by the standards.  A modified component would be unlikely.
[The replaced components would be scattered throughout the plant  and
would become affected by the standards one at a time  as existing  components
are replaced.]  In contrast, under the process  unit approach,  all  new
components within "new" process units would be  covered by the  standards,
but individually replaced components would not  be covered.   Most
importantly, many components (not actually increasing  emissions or
being replaced) in modified process units or reconstructed process
units would be covered ("captured") based on the "process unit" approach.
     The difference between the emission reduction potential for  the
two approaches can be based on the difference in the  number of individually
replaced components and the number of components that  are "captured" in
modified or reconstructed process units.  EPA believes that, in this
case, the numbers are similar.  However, there  is no reliable  procedure
to approximate these numbers.  Based on EPA's belief  that the  emission
reductions between the component approach and the process unit approach
are similar and based on the burden associated  with maintaining records
of individual components for an entire plant,  EPA selected the process
unit as the basis of the affected facility for  all  the equipment  covered
by the standards except compressors.
Comment:
     Three comments were received concerning the designation of compressors
as a separate affected facility.  One commenter (IV-D-10)  supported
compressors as a separate affected facility;  however,  others (IV-D-5

                                  3-6

-------
and  IV-D-12) maintained that when compressors are an integral part of a
process unit they should be considered part of the "process unit"
affected facility.  When compressors are not considered an integral
part of a process unit, they should be considered separately.  Also,
commenters said designating compressors as a separate affected facility
could lead to confusion and would likely make them subject to standards
sooner because they are often replaced at shutdown.
Response:
     As discussed in the preamble to the proposed standards,  compressors
(unlike other fugitive emission sources in petroleum refineries)  are
major pieces of equipment and are readily identifiable.  Since compressors
are relatively few in number, tracking of those subject to NSPS requirements
and those not subject to these requirements would not be difficult.   As
mentioned above, there is a presumption that the narrowest definition
of an affected facility is proper unless there is a statutory factor
that leads EPA to a less narrow definition.  Commenters did not present
any of these factors.  The fact that compressors are integral to  the
process unit does not preclude EPA from defining them as separate
affected facilities.  By extension, the commenter's reasoning would
prohibit EPA from defining different emitting sources within  a plant as
separate affected facilities because they are integral  to the plant.
It is clear, however, that EPA has authority under Section 111 to define
each as a separate source.  Moreover,  EPA has often chosen such plant
subsets as separate affected facilities.   (For example, see Subpart  Da
of 40 CFR Part 60 -- each boiler at the utility station is a  separate
affected facility).   Focusing on whether equipment is integral to a
process simply is not helpful  or relevant to the selection of the
affected facility for purposes of standards of performance.
     It should be noted that by making compressors a separate affected
facility, compressors are not likely to be covered by modification
provisions.   However, as one of the commenters stated,  when a compressor
is refurbished or replaced it would likely be a reconstruction and,
therefore, covered by the NSPS.   EPA considers this appropriate.   EPA
considered the comments received regarding designation  of compressors
as separate affected facilities and concluded that the designation
                                  3-7

-------
should remain for the same reasons originally stated in the preamble to
the proposed standards.
Comment:
     One commenter (IV-D-10) recognized that the definition of process
unit includes flexibility and that determinations of an affected facility
may be on a case-by-case basis.  The commenter wrote that the identity
of a "process unit" is not always clear or equally transferable from
one refinery to another.
Response:
     EPA agrees with the commenter that there will be differences in
"process unit" affected facilities, even among processes producing
the same petroleum product.  Thus there is flexibility in the definition
of process unit.  These differences are mainly caused by differences in
design and construction of process units.  Typically, equipment within
a process area is functionally related and associated with a single
process unit.  However, some equipment pieces (generally, very few)
within an area may be functionally associated with a second process
unit that is not located in the area.   Hence, equipment function will
be a determining factor as to which process unit it is considered to be
in.  When a piece of equipment can function in more than one process
unit, its location will  be a determining factor.   It should be noted that
owners and operators may request EPA to review plans for construction
or modification for the purpose of obtaining technical  advice, as
provided in the General  Provisions of  Part 60 (40 CFR 60.6).
3.2  DEFINITION OF "IN VOC SERVICE"
Comment:
     Several  commenters (IV-D-6, IV-D-8, IV-D-16, IV-D-21,  and IV-D-24)
requested that the definition of "volatile organic compound (.VOC)"
specifically state which organic compounds are excluded.   It was also
recommended that the definition include the phrase,  "or as  measured by
the applicable test methods described  in Reference Method 21."
Response:
     Volatile organic compounds (VOC)  are defined as organic  compounds
that participate in photochemical  reactions.   Any organic compound  is
presumed to participate in atmospheric reactions  unless the Administrator
determines that it does  not.   EPA considers several  organic compounds

                                  3-8

-------
to have negligible photochemical  reactivity.   These are methane,  ethane,
1,1,1-trichloroethane, methylene  chloride,  trichlorof1uoromethane,
dichlorodifluoromethane, chlorodifluoromethane,  trifluoromethane,
trichlorotrifluoroethane, dichlorotetraf1uoroethane, and chloropenta-
fluoroethane.
     The standards provide for the exclusion  of  substances considered
non-photochemically reactive  by EPA from the  percent YOC in the  process
fluid when determining whether a  piece of equipment is not in VOC
service.  The purpose of this  is  to avoid covering those sources  that
have only small amounts of photochemically reactive substances in the
line and to establish the standards consistent with the data base.
In determining whether the VOC in a process line is less than 10  percent
of the total  mass in the line  (as a prerequisite to determining  that a
piece of equipment is not in  VOC  service),  quantities of compounds
present in the line that are  considered nonphotochemically reactive by
EPA may be excluded from the  total quantity of organic material.
     Section 60.595(d) of the  standards requires that VOC content is to
be determined by the referenced ASTM methods,  not by Reference Method 21.
The referenced ASTM methods can be used to distinguish among compounds
and, therefore, allow the determination of the amount of photochemically
reactive compounds in a process stream.  In contrast, Reference  Method 21
is a method for determining leaks.  This method  requires that monitors
used in complying with the standards respond  to  the organic compounds
in the process streams.  Thus, there is no reason to include the phrase
requested by the commenter.
Comment:
     Several  commenters (IV-D-8,  IV-D-12, IV-D-18, IV-D-21, IV-D-22, and
IV-D-24) requested that the proposed definition  of "in VOC service" be
revised.  The commenters suggested raising the weight percent cutoff
from 10 to 20 weight percent VOC to exclude coverage of hydrogen service
compressors and to provide more reasonable operating flexibility.
Excluding 75 volume percent or greater hydrogen  streams and changing
the 10 weight percent VOC to 10 volume percent VOC were also recommended.
The commenters contend that such streams would have a lower percentage  of
VOC and, consequently, the controls would achieve lower VOC emission
                                   3-9

-------
reductions and have a higher cost effectiveness ($/Mg VOC emission
reduction).
Response:
     The commenters are suggesting that EPA exempt equipment that,
because they contain so few YOC, are not cost effective to control.
In response to this comment, EPA analyzed the control  of valves
and compressors in hydrogen service (Document No.  IV-B-9).  Most
process streams affected by the standards are clearly above 10 weight
percent VOC, and many are nearly 100 weight percent VOC.  Process  streams
less than 10 weight percent are almost always much less than 10 weight
percent VOC.  Only a few process streams may be near 10 weight percent
VOC, and these are generally those that would be considered in hydrogen
service.  Thus, EPA analyzed the control  of equipment that, based  on
EPA's data, could be found in hydrogen service.  This would allow  EPA
to exempt control  of equipment if it is not cost effective.  In hydrogen
service is defined as greater than 50 volume percent hydrogen based on
EPA's data.  The analysis is explained in docket item IV-B-9.   Emission
reductions are achieved for valves in hydrogen service at reasonable
costs ($106/Mg VOC).  However, application of equipment controls for
compressors in hydrogen service results in a cost  effectiveness of
$4,600/Mg VOC.  EPA, therefore, decided to exclude compressors in
hydrogen service from the standards.
     In EPA's judgment, determination of VOC content in a given stream
is a routine analytical procedure.   The test method, ASTM E-260, gives
quantitative measures of each component proportional  to their concentration.
Hence, the results are expressed as a weight percent.   The commenter
recommending that the VOC content expressed in weight percent be changed
to a volume percent did not provide any basis for  this change.   Hence,
EPA maintains that the VOC content expressed in weight percent is  a
reasonable approach.
Comment:
     Another commenter (IV-D-21) remarked that the definition  of VOC
fails to establish a de minimi's level  for volatile materials which do
not contribute to  atmospheric emissions.   A "heavy liquid"  definition
was considered necessary because it would avoid unnecessary monitoring  of
                                  3-10

-------
 components processing materials that are unlikely  to  register  a  10,000 ppm
 concentration.
 Response:
      The definition of VOC  does not exclude  compounds based on volatility.
 Processes  that  produce relatively  non-volatile  products can involve
 high temperature and pressure conditions,  thus  producing emissions of
 VOC.  These VOC contribute  to ozone formation.  A  de minimi's level
 would not  be appropriate.   EPA has tailored  the standards (in part,
 based on volatility)  to  require the best demonstrated technology.  As a
 consequence,  EPA concluded  that routine  leak detection and repair is
 not warranted for components  in heavy liquid service because they have
 low leak rates  and,  as a group,  control  is not  cost effective (Docket
 No.  IV-B-5).
      Heavy  liquid streams are generally  a mixture  of heavy hydrocarbons
 (e.g., crude  oil)  with very little  light hydrocarbons.  Nevertheless, these
 streams  have  the potential to leak  VOC (determined by concentrations in
 excess of 10,000 ppm), and these VOC would contribute to ozone formation.
 Data reviewed by EPA  (Document  No.  II-A-19) show that a few components
 in heavy liquid  service  do have emission concentrations greater than
 10,000 ppm  and,  therefore, do leak emissions of VOC.  When these leaks
 occur, repair is cost  effective (Document No. IV-B-5).  If an  operator
 sees, hears, smells, or  otherwise suspects a leak,  it is appropriate
 that the component be monitored and, if a leak exists based on  a greater
 concentration, that it be repaired.
 Comment:
     One commenter (IV-D-12) supported the proposed definition  for
 "light liquids"  as it agrees with findings  in their fugitive emissions
 program; however, another (IV-D-17) held that the  definition was too
 restrictive and should include only the  heavy naphthas and lighter
 materials because as defined,  some equipment in light liquid service
 would not significantly contribute to fugitive emissions.   Excluding
 compounds with a Reid Vapor Pressure (RVP)  less than 1.5  psi was
 recommended.
 Response:
     The criterion used by  EPA for the light liquid definition
 (that is, liquids with a vapor pressure  greater than that  of kerosene)
was based on fugitive emission data gathered  in petroleum  refinery
                                  3-11

-------
studies (Document No. II-A-19).  Equipment processing VOC with vapor
pressures greater than kerosene were found to leak at significantly
higher rates and frequencies than equipment processing VOC with vapor
pressures of kerosene or lower.  Therefore, EPA decided to exempt
equipment processing VOC substances with vapor pressures lower than
about the vapor pressure of kerosene from the routine leak detection
and repair requirements of the standards.  This is consistent with the
commenter's request to cover only heavy naphthas and lighter compounds.
     The RVP cutoff of 1.5 psi that was recommended by the commenter is
based on California regulations for the storage of volatile organic
liquids which are at ambient pressure and temperature.  There are no
data to support the 1.5 psi cutoff as it would apply to fugitive emission
sources.  EPA considers control of equipment in light liquid service
(based on the proposed definition) cost effective;  therefore, based
on these considerations, EPA did not revise the definition of light
liquid service.
3.3  EXCLUSIONS
Comment:
     One commenter (IV-D-13) stated that process units with in-place
state-of-the-art hydrocarbon gas detection systems  should be exempted.
This commenter requested that units in an arctic environment be exempted
because of several  unique aspects of refining in the North Slope of
Alaska.  For example, (1) the products are used locally,  (2) process
units are totally enclosed at a high cost because of the  harsh  environment;
therefore, present safety controls (gas detector placed near exhaust
fans with an alarm set at 12,500 ppmv)  are adequate and additional
requirements are unwarranted, (3) requiring rupture disks ahead of
pressure relief devices would compromise safety especially under this
application, (4) repair labor is 2 1/2  to 4 times more costly,  and  (5)
control of VOC has  limited benefit in attainment areas, especially  in
the arctic where cold ambient temperatures, the degree of insolation,
and a low concentration of photochemical  precursors limit ozone
formation.
Response:
     The presence of an in-place state-of-the-art hydrocarbon gas
detection  system does not necessarily ensure  emission  reductions.   Gas

                                  3-12

-------
 detection  systems  set  for  12,500  ppm would permit VOC to be emitted
 without  notice.  Several megagrams of VOC would 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.  Upon request by EPA, the commenter explained the specific
 control  techniques used at their  plant, many of which are identical to
 those  required by  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 final  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.
     EPA has studied the commenter's concerns and acknowledges that
 there  are  several  unique aspects  to refining in the North Slope of
 Alaska.  Accordingly,  EPA  concluded only that the costs to comply with
 the  routine leak detection and repair requirements of the proposed
 standards may be unreasonable.  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.  These
 unique aspects make the cost of routine leak detection and repair
 unreasonable (Document Number IV-B-15).  Therefore, EPA has decided
 that refineries 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
 the cost of those requirements is reasonable.
 Comment:
     One commenter (IV-D-26)  recommended that the definition of "petroleum
 refinery" be clarified to exclude production and intermediate facilities
 such as wells, drill  pads and separation tanks,  that may  be involved  in
onsite processing in oil  fields.   Similarly, another commenter (IV-D-3)
 requested that the definition of "petroleum" be revised to clarify that
coal  tar and refined coal  tar oils that are by-products of coking

                                  3-13

-------
processes are not covered in these standards because they  are not
covered under Subpart J.
Response:
     In Section 60.591 (Definitions),  the proposed standards  defined
"petroleum refinery" as "any facility  engaged in  producing gasoline,
kerosene, distillate fuel oils, residual  fuel  oils,  lubricants,  or
other products through the distillation of petroleum,  through the
redistillation, cracking, or reforming of unfinished petroleum deriva-
tives."  This definition does not include production and intermediate
facilities found in oil  fields, nor does it include  production tar and
tar oils from coal  coking processes.   The standards  apply  only to
process units within petroleum refineries.  New source performance
standards (NSPS), however, are being developed by EPA for  the natural
gas processing industry under another  standards development project
(40 CFR Part 60 Subpart KKK - Standards of Performance for Onshore
Natural Gas Processing Plants:  Equipment Leaks of VOC).   The natural
gas processing industry NSPS may cover fugitive emission sources at
production, and will more likely cover them at intermediate facilities.
EPA has consistently used the term "petroleum"; it does not mean tar
and tar oils from coal  coking processes,  but it does mean  synthetic
petroleum products  from processes that use coal as a raw material.
Thus, EPA has not clarified the term "petroleum."  It  should  be  noted
that the production of some chemicals  (for example,  formaldehyde or
phenols) at coal coking processes, however, is covered by  NSPS for the
synthetic organic chemical manufacturing industry (Subpart VV).
Comment:
     Another commenter (IV-D-8) maintained that the  results of a number
of studies support  a complete exemption for equipment  in low  vapor
pressure service.  The commenter noted that an EPA study (see Document No.
II-A-41, p. 2-38) of two refineries in the South  Coast Air Basin had
monitored 664 components in light and  heavy liquid service (which were
exempt from South Coast Air Quality Management District rules) and
found only four leaking components, none of which was  in heavy liquid
service.  Another study found only one leaking valve out of 519  in
heavy liquid service and only 3.8 percent of the  pumps leaked.   Another
commenter (IV-D-15) supported excluding pumps  in  heavy liquid service,
                                 3-14

-------
pressure relief valves in light liquid service, flanges, and connections
from routine monitoring.
     Another commenter (IV-D-21) wrote, however, that there is no
demonstrable cost effectiveness to the inclusion of pumps and valves  in
heavy liquid service, pressure relief devices in both light and heavy
liquid service, and flanges and other connections, and further, these
sources should be exempt from all requirements as indicated by EPA at
the National Air Pollution Control Techniques Advisory Committee (NAPCTAC)
meeting on June 3, 1981.
Response:
     The commenters suggest that certain types of equipment leak so
infrequently that it is not cost effective to monitor them for leaks.
The final standards for pumps in heavy liquid service, pressure
relief devices in light liquid service, flanges, and connections exempt
these sources from routine leak detection and repair because of the low
leak frequency and emission factor for these sources as discussed at  the
June 1981 NAPCTAC meeting.  However,  the commenters have suggested no
reason that this equipment, if found  to be leaking, cannot be repaired
cost effectively.  EPA has determined that it is cost effective to
repair these components if they are leaking (see Document No. IV-B-5).
Therefore,  Section 60.592-8 provides  that if evidence of a potential
leak is found, the piece of equipment must be monitored within 5 days,
and repaired as soon as practicable within 15 days if an instrument
reading of  10,000 ppm or greater is detected.
3.4  SMALL  REFINERS
Comment:
     Commenters (IV-D-9 and IV-D-23)  accused EPA of incorporating into
the standards a bias against small refiners.  They asserted that small
refiners  will be affected more adversely than will large refiners.
There is  a  bias, they reason, because EPA did not analyze the comparative
impact of the standards on large versus small refineries.  The comparative
analysis  was not done because EPA, citing elimination of the crude oil
entitlements program, decided that relatively little new unit construc-
tion will  occur at small  refineries,  and even considering modified and
reconstructed units, few small refineries would be subject to the
standards.  Thus, the commenters claim, EPA saw no reason to give small

                                  3-15

-------
refineries special attention under the Regulatory Flexibility Act.
Therefore, a bias exists.  One commenter (IV-D-9) said some small
refineries have existed for more than 40 years and are viable without
subsidy programs.  Thus, both commenters conclude, EPA should have
analyzed the differential impact of the standards on small  refineries
compared with large ones.
Response:
     To analyze the economic impacts of the standards, EPA  defined  12
types of refinery units:  crude distillation,  hydrotreating, isomeri-
zation, etc.  (See BID for the proposed standards, Document No.  III-B-1)
Each type was assigned to one of three model  unit categories.  Model
Unit A has a small number of pumps, valves, and other components; Model
Unit B has a larger number; and Model Unit C has the most.   Assignment
of each type of unit to a particular model unit category was based  on
equipment counts averaged over units found at a range of refineries.
EPA then assumed a reasonably small throughput (which might be repre-
sentative of some small refineries) for each type of unit,  because
small-throughput units would show significant adverse economic impacts
much more readily than large-throughput units  would for any given
amount of money to be spent on controlling fugitive leaks.   If the
analysis had revealed potential, adverse economic impacts,  EPA would
have intensified its examination of the units  involved, and possibly
would have changed the standards appropriately.   However, no such
impacts were projected.  EPA concluded that no adverse economic  impacts
would result from the standards and that there was no need  to extend
the economic analysis to cover a wider range  of throughput  levels.
     The Regulatory Flexibility Act (Public Law 96-354, September 19,
1980) requires that special consideration be  given to the impacts of
standards on small firms.  As one criterion for extending loans  and
related assistance, the Small Business Administration defines a  small
petroleum refining firm as one employing fewer than 1,500 workers (13
CFR Part 121, Schedule A).  The 1,500 number applies to the entire
firm, including affiliates, and is tied in with  other criteria relating
to throughput capacity, exchange agreements, and the like.   EPA  accepts
this definition of a small  refiner.   Based on  this definition, EPA
                                  3-16

-------
projected few small firms would be affected by the standards.   However,
even for those affected by the standards, EPA concluded that the impacts
would be reasonable.
     The commenters implied that small process units or refineries may
be owned by small firms.  However, there are no available data that
relate the size of refining firms (number of employees, throughput,
etc.) to the size of their individual  refinery units (number of valves,
pumps, etc.).  In addition, there is no reason to believe that the size
of a refining firm is necessarily related to the size of its individual
refinery units.  EPA, therefore, has no basis to suspect that small
firms will  bear greater compliance costs than large firms on an affected
facility-by-affected facility basis.  In fact, the impacts are similar
for all sizes of units.  The three model units differ only in  regard to
their respective equipment counts.  Compliance costs for Model Unit A,
the smallest, are lower than compliance costs for Model Unit C, the
largest.  Even though the cost effectiveness for Model  Unit A is larger
than it is for Model Unit C.  However, these cost effectivenesses are
not significantly different.  The commenters offered no evidence that
the sizes of their units, measured either by throughput or by equipment
counts, will cause the impact of the standards to fall  disproportionately
on small firms, or that small firms will become non-competitive, or
that small  firms will be forced to raise prices substantially.
     EPA's projection that relatively few units at small refineries will
be affected by the standards by 1987 is still valid.  If, for some
reason not now anticipated, the standards were to place a disproportionate
burden on small refineries, the 1987 projects indicate that comparatively
few such refineries would experience that burden and, more importantly,
the cost estimates indicate that none of those refineries would experience
unreasonable costs.  [EPA regrets that its statement may have been
interpreted by the commenters as a bias against small refiners.]
Comment:
     One commenter (IV-D-22) remarked that economic conditions in the
refinery industry have changed drastically since 1980,  that the projected
number of affected facilities is now too high, and that the benefits
of the proposed standards, therefore, are overstated.
                                  3-17

-------
Response:
     EPA projected that 100 new and 182 modified or reconstructed units
would become affected facilities during the first 5 years of implemen-
tation of the standards.  The commenter offered no alternative
projections to supplement the claim that 282 units are too many.   There
are three reasons why EPA believes the projections should not be  changed.
     First, the projections in no way affect the need for standards or
the selection of the final standards.  The projections are offered only
as a guide for understanding the future aggregate costs and emission
impacts of the standards.  Ideally, the projections should be conservative.
However, being conservative requires EPA to project minimal growth when
estimating emission impacts, and to project maximal growth when estimating
aggregate costs to the nation.  Low projections can understate possible
economic costs, but high projections can overstate emission savings.
The middle ground reflects EPA's best judgment, considering these two
conflicting uses for the projections.  Furthermore, if all projected
growth does not occur by the end of the fifth year, it will occur
sometime shortly thereafter.
     Second, the projections were made for the calendar years 1982
through  1986.  For reasons unforeseen when the projections were prepared,
the proposal of the standards was delayed a year.  This means that the
projections should now be interpreted as applying to the years 1983
through  1987.  This shift of 1 year moves the projection interval
completely out of the 1981-1982 economic downturn that the commenter
believes caused the projections to be overly optimistic.  As general
economic recovery proceeds, there is every reason to believe the recovery
will be  felt throughout  the refinery industry.
     Third, the projection methodology used by EPA excludes modification
and reconstruction at refineries with crude distillation capacity under
2,226 m3 (14,000 bbl) per calendar day.  This exclusion was made as a
way of accounting for the possible effect of elimination of the crude oil
entitlement program.  Nevertheless,  two commenters (IV-D-9 and IV-D-23)
representing small refiners complained that there would in fact be more
modification and reconstruction than EPA projected.  Thus, there is an
indication that the projections are, if anything, too low in this area.
     For the above reasons, EPA is not revising the projection of new,
modified, and reconstructed refinery units.
                                  3-18

-------
                          4.0  MODIFIED SOURCES

 4.1  EMISSION INCREASE
 Comment:
      A number of commenters (IV-D-14,  IV-D-15, IV-D-16,  and  IV-D-21)
 requested that the EPA allow an  increase  of  a  de  minimis  level  of
 emissions before an existing facility  would  be considered to  have
 modified.  De minimis values of  5  tons per year and  40 tons per year
 (as in  40 CFR 51.18(j))  were suggested by the  commenters.
 Response:
      Under the definition  in Section lll(a)(4), any  physical  or
 operational  change resulting in  an  increase  in  emissions  constitutes a
 "modification."   EPA has exempted  certain small emissions increases
 from  consideration in  deciding whether there has  been an  increase in
 emissions constituting a "modification" for  purposes of PSD applicability
 (40 CFR 52.21(b)(2)  and  (b)(231)).  This  action followed  the  decision
 in  Alabama  Power  Co.  v. Costle,  636 F.2d  323 (D.C. Cir. 1979),  in which
 the D.C.  Circuit  held  that  EPA has authority to interpret the definition
 of  "modification"  so as to  exempt sources with  small emissions  increases
 from  PSD  review on grounds  of administrative necessity (Jd_. at 400).
      The  Alabama  Power decision does not  require EPA to provide a de
minimis exemption  from application of the "modification"  definition for
 NSPS applicability purposes.  Nor has EPA's experience in implementing
the NSPS program suggested an administrative need for relieving existing
sources from NSPS applicability when they undergo changes resulting in
only a small increase in emissions.  This differs somewhat from EPA's
implementation of the definition  of "modification" for PSD applicability
purposes.   In that area, the Agency has determined that the administrative
burden of  applying the full preconstruction  review process to a source
with only  a small  emissions increase may be  unreasonable  (45 FR 52705;
August 7,  1980).   The administrative burden  associated with the NSPS
program, however, is relatively minimal.  In  contrast to  PSD requirements,
                                  4-1

-------
 NSPS's  are categorically applicable technology-based requirements only;
 they do not  involve an assessment of ambient effects and do not require
 case-by-case  review.
     Furthermore,  EPA believes that the current straightforward
 application  of the "modification" definition for NSPS purposes best
 serves  Section Ill's intent.  One key purpose of the NSPS program is to
 prevent new  pollution problems from arising.  One way that the statute
 seeks to achieve this is by requiring application of the best demonstrated
 technology at, and thereby minimizing emissions from, existing facilities
 with increased emissions.  The current NSPS approach of not providing
 an exemption  from the "modification" provision based on the size of
 the emissions sources is not intended to cover existing plants making
 routine and minor additions.  The "modification" provisions in the
 General Provisions of 40 CFR Part 60 exempt changes such as additions
 made to increase production rate (if they can be accomplished without
 capital expenditure, as defined in the General  Provisions) and routine
 replacements  (40 CFR 60.14(e)).  In addition, these standards would
 exempt  additions made for process improvements if they are made without
 incurring a capital expenditure.
 Comment:
     Other commenters (IV-D-8 and IV-D-14), concerned about the complexity
 of the modification provisions, endorsed revising the modification
 provisions such that a modification occurs  when the number of components
 exceeds 10 percent of the total  number of the same type and there is a
 net increase  in emissions from the process  unit.
 Response:
     As discussed in Section 4.2, EPA is promulgating an alternative
 procedure that will reduce the complexity of the modification provisions
 (in particular, how to determine a capital  expenditure).  In EPA's view
 40 CFR 60.14 of the General  Provisions adequately specifies the categories
 of changes to an existing facility that  should  bring the facility under
 NSPS as a  "modified"  source.  It should  be  noted that,  under Section
 60.14(e),  certain changes made in an  affected facility  without a  capital
 expenditure are not considered "changes  in  operation"  by EPA and,
therefore, are not modifications.  See,  e.g., 40 CFR 60.14(e)(2)--
 production rate increases.
                                  4-2

-------
     As proposed in the standards, certain changes (process improvements)
made in an affected facility without a capital  expenditure are not
considered "changes in operation" by EPA and, therefore, are not modi-
fications.  This generally excludes coverage from industry practices
that involve adding a few valves and maybe a sampling system and making
other minor changes in equipment configurations.  The 10 percent increase
in the number of fugitive emission components as suggested by the
commenter, would most likely be associated with a VOC emissions increase
of about  10 percent.  Making such a change would likely be associated
with capital expenditure and, therefore, EPA considers this a modification
Therefore, EPA did not revise the modification provision as requested.
Comment:
     Another commenter (IV-D-8) maintained that once a  "modification"
has occurred, the  NSPS requirements should be applied only to those
types  of  components which trigger the definition of modification.
Response:
     Under  Section 111 of the Clean Air  Act, the  application  of
modification is  inextricably tied to  the designation of "new  source,"
or  in  NSPS  terminology,  an  affected facility.   Section  lll(a)(2) defines
the  "new  source"  subject to NSPS as a source on which modification  has
commenced after  proposal,  not the  portion of the  source actually changed.
 Stated differently, modification provisions  are triggered  with  respect
to  the affected  facility;  therefore,  applicability is to all  components
 affected  by the  standards within the  affected  facility.  The  commenter
 is  implicitly  requesting EPA to define the affected facility  as a  group
 of  one type of  equipment within a process unit.   EPA,  as discussed in
 Section 3.1, reasonably  concluded that affected facilities to which the
 standards apply should remain:   (1) the group of all  fugitive emission
 sources (pumps, valves,  sampling connections, pressure relief devices,
 and open ended  lines) within a process unit and (2) compressors.
 4.2  CAPITAL EXPENDITURES
 Comment:
      Commenters requested that the capital expenditure determination
 (as it relates  to the modification provisions) be revised so that  it is
 more practicable.  Commenters  (IV-D-8 and IV-D-15) remarked that the
                                   4-3

-------
capital expenditure guidelines are outdated and would be difficult to
use because units have been substantially rebuilt over the years and
records of costs for determining the cost basis may be kept on a process
unit basis rather than for individual  pieces of equipment, or simply
may not exist.  It is, therefore, very difficult to reconstruct original
costs.
     In support of their concern about using original  costs, the commenters
stated that, based on EPA's current interpretation of capital  expenditure,
1 to 3 percent of the current replacement costs would subject units to
modification.  Another commenter (IV-D-14) claimed that component costs
represent 5 percent of the total original costs and that the addition
of a new pump with several valves could easily exceed the "capital
expenditure" definition.  This commenter provided the hypothetical
example of a unit with a total original cost of $16 million and component
cost of $815,000.  The addition of a pump with several valves would
exceed 4 percent of the total component costs, around $56,000.
     One commenter (IY-D-10) suggested that replacement costs rather
than original costs be used to determine the basis for capital expenditure.
A few commenters (IV-D-4, IY-D-15, and IV-D-16) suggested that capital
expenditure be defined as 7 percent of the replacement cost (based on
the Chemical Engineering Construction  Index or other suitable index) of
an affected facility at the time of process improvement.
Response:
     After reviewing the comment letters concerning the difficulties
with using the capital expenditure definition, EPA agrees that the
definition for capital expenditure may be difficult to use for some
refineries.  Accordingly, EPA decided  to provide an alternative to
the procedures in 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  costs to determine capital
expenditure.  The adjusted AAGRA is determined by a formula and is
based on a ratio that reflects inflation of costs over the last several
years.  The adjusted AAGRA is multiplied by the replacement costs of

                                  4-4

-------
 the  equipment  within  the  facility to determine the value of a capital
 expenditure.
      The  burden  associated with using the capital expenditure definition
 in the  proposed  standards was not quantified by the commenters; however,
 if some of  these problems can be resolved without changing the application
 of the  modification provisions, EPA finds no reason not to do so.
 Accordingly, EPA is providing an alternative method to the General
 Provisions.
      As mentioned above, the alternative method for determining capital
 expenditure enables refiners to use replacement costs rather than
 original  costs.  An inflation index can be applied to the replacement
 value of  an affected facility to approximate the original  cost basis of
 the  affected facility.  The relationship between replacement and original
 costs has been determined (Document Nos. IV-B-4 and IV-B-14) as:
           Y = 1.0 - 0.575 log (X), where:
           Y = the percent of the present replacement cost which is
               equivalent to the original  cost, and
           X = the year of construction.
 Using the above  equations and the annual asset guideline repair allowance
 (AAGRA) of 7 percent (see IRS Publication 534,  page 20), capital  expenditure
can be expressed in replacement dollars as:
           Capital Expenditure =  R x Y x 0.07, where:
           R = existing facility replacement cost.
     Another alternative method that was considered is  similar to that
of the first in that an inflation index, Y (as  defined  above), and the
AAGRA basis of 7 percent are used to allow refiners to  use replacement
costs.  However, this second alternative would  also allow refiners to
use the cost of the entire process  unit rather  than the affected facility
 (the fugitive emission components).   The second alternative would
reduce the number of units that would use  a  detailed costing of equipment.
However, the equipment covered by the standards represent a variable
portion of the total  costs of all  the equipment in  a process unit.
Therefore, it is not practicable to assign a single percentage that
would reflect the modification costs contributed  by fugitive emission
pieces.  Thus,  EPA is not adopting  this alternative.
                                  4-5

-------
     Even though EPA was unable to establish an alternative that would
allow refiners to use the cost of the entire process unit, EPA would
consider estimations from refiners that clearly show that an expenditure
would be less than the quantity associated with a capital expenditure.
There may be a variety of ways that these estimations can be ma.de.  For
example, a refiner may have proof that in certain units 5 percent of
the total replacement value at the process unit is the value of the
equipment covered by the standards.  If an estimation clearly demonstrates
its results, EPA could quickly decide whether a process improvement
involves a capital expenditure.  Based on the example, if the value of
the process improvement may be 0.04 percent of the replacement value of
the process unit, this would be clearly less than 12.5 percent of
5 percent of the value of the total unit.  If an estimation does not
clearly show its results, then the time and effort required by EPA in
evaluating the estimation would not provide the owner or operator a
quick response and, therefore, a more-detailed costing of equipment
(either by estimating replacement or accounting existing equipment)
would be the owner or operators best approach.  If EPA can judge easily--
through review of a clear demonstration that a process improvement does
not involve a capital  expenditure, it will  do so.   In contrast, if
EPA's review raises concerns or questions,  EPA will  reject the estimation
unless further convincing support is presented.
Comment:
     One commenter (IV-D-8)  wrote that the  General  Provisions exempt
"process improvements"  from  being considered modifications if made
without incurring a capital  expenditure;  however,  using the proposed
definition of "capital  expenditure" limits  the exemption.   Another
commenter (IV-D-21) recommended deleting  the modification provisions
which require that process  improvements be  accomplished without a
capital  expenditure.
Response:
     The General  Provisions  do not include  a "process improvement"
exemption.  However,  in the  proposed standards EPA stated its  intent
that minor modifications would not be covered by the  standards:  "addition
or replacement of fugitive emission sources  for the purpose  of process
improvement which is  accomplished without a  capital expenditure  shall

                                  4-6

-------
not by itself be considered a modification under this subpart."   The
capital expenditure criterion was included so that minor process
improvements in a process unit that cause an increase in emissions
would not subject an existing facility to this NSPS.   After reviewing
these comments, EPA has maintained the same exemption.  EPA considers
any increase in emissions that results from a process improvement with
a capital expenditure a "modification" unless one of  the other exceptions
in the General Provisions applies.
     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.
Comment:
     One commenter (IV-D-30) argued that the "no capital expenditure"
exemption for modifications could be construed by a plant as including,
for example, the addition of fugitive components from existing inventory
of spare parts.  The commenter requested that EPA make it clear that
the addition of equipment already in stock is still considered in
determining a capital expenditure.
Response:
     As discussed in the response to the previous comment, the capital
expenditure criterion applies to process improvement  or production rate
increase exemptions that are considered when determining whether  an
increase in emissions at a facility results in a modification. This
criterion is used to judge if the activity results in a change in
operation.  As such, a capital expenditure is determined by what  is
added to a process unit, not by what is purchased.  Accordingly,  it
makes no difference whether the item was already in stock when the
process improvement occurred.
                                  4-7

-------
4.3  SMALL FACILITIES
Comment:
     Commenters (IV-D-9 and IV-D-23) stated that small refiners more
easily trigger the modification and reconstruction provisions of the
standards than do large refiners.  A small capital expenditure on a
small unit would cause the unit to be classified as "modified" more
easily than the same expenditure would cause a large unit to be so
defined.  Some commenters (IV-D-8 and IV-D-15) also remarked that the
definition of "capital expenditure" would impact small facilities more
severely.
Response:
     The provisions of the standards can be triggered by several  different
actions.  Some of these actions are relative changes that are considered
important because they involve a certain percent of the cost of the
unit.  Other triggering actions are absolute changes that are considered
important because they involve an absolute increase in air pollutant
emissions from the existing unit.  Even if the reconstruction and
modification provisions are more burdensome on small refineries,  the
overall impact of the standards is still  reasonable, however.  If the
commenter's claim is true, then there must be a difference between
units at small  refineries on the one hand, and those at large refineries
on the other.  The difference could be related to size, age, ability to
respond to today's changing markets, or myriad other factors.  The
question of unit size, measured by throughput or equipment count, is
discussed in a previous response; the relationship between unit size
and firm size is not clear.  In general,  small changes can trigger the
provisions for units with comparatively few pumps and valves.  However,
it is not clear that the capital  expenditure criterion would be exceeded
quicker by a small process unit than by a large process unit.  It is
not necessarily true that the cost of a given set of equipment would be
the same for a small  and large process unit.   Large process  units can
use large equipment or small  equipment (the costs of which would  be
related, in a very general sense, to the  size of equipment)  and small
process units can use large or small equipment.   The value of any one
pump in a process unit may be relatively  small  or large depending on
the specifications in a particular application,  not solely on the size

                                  4-8

-------
of the pump.  The commenters did not mention specific small refinery
characteristics that would explain why small refiners might suffer a
greater burden than large refiners.  Age and obsolescence of equipment,
the most obvious characteristics, do not appear to be significant
factors.  There are no data to indicate that units in need of modernization
are situated predominantly at small refineries.  Even if there are
differences and small  refineries are disproportionately affected, EPA
does not consider this unreasonable because EPA believes that the
standards are appropriate for all existing facilities that become
affected by the standards.
     For these reasons EPA concludes that the modification and reconstruction
provisions of the standards will  not subject small refiners to unreasonable
adverse impacts.
                                  4-9

-------

-------
                          5.0  RECONSTRUCTION

Comment:
     Several  commenters (IV-D-4,  IV-D-14,  IV-D-15,  and  IV-D-21) wrote
that reconstruction costs should  not be accumulated.  Some  commenters
suggested that costs should be considered  over a 1-year period.
Response:
     Since in enacting Section 111 Congress did not define  the  term
"construction," the question arose whether NSPS would apply to  facilities
being rebuilt.  Noncoverage of such facilities would have produced the
incongruity that NSPS would apply to completely new facilities, but not
to facilities that were essentially new because they had undergone
reconstruction of much of their component  equipment.  This  would  have
undermined Congress's intent under Section 111 to require strict  control
of emissions as the Nation's industrial  base is replaced.
     EPA promulgated the reconstruction provisions  in 1975, after notice
and opportunity for public comment (40 FR  58420, December 16, 1975), to
fulfill this intent of Congress.   Since this turnover in the industrial
base may occur independently of whether emissions from  the  rebuilt
sources have increased, the reconstruction provisions do not focus on
whether the changes that render a source essentially new also result in
increased emissions.
     Congress did not attempt to  overrule  EPA's previous promulgation
of Section 60.15 in passing the Clean Air  Act Amendments in 1977.  This
indicates that Congress viewed the reconstruction provisions' focus on
component replacement, rather than emissions level, as  consistent with
Section 111.  See, e.g., Red Lion Broadcasting Co.  v. FCC.  395  U.S. 367
(1969); NLRB v. Bell Aerospace Division, 416 U.S. 267  (1974).   Nor has
any Court questioned the Agency's authority to subject  reconstructed
sources to new source performance standards.  In fact,  in ASARCO  v.
EPA, 578 F. 2d 319, 328 n.31 (D.C. Cir. 1978), the D.C. Circuit suggested
that the reconstruction provisions may not go far enough toward preventing
possible abuses by owners seeking to avoid NSPS by perpetuating the
useful lives of their existing facilities  indefinitely.
                                  5-1

-------
     Finally, coverage under §60.15 of establishing  petroleum refinery
facilities comports well  with the intent underlying  Section  111.   In
such cases, the refurbishment may transform the existing  facility  into
an essentially new facility.  A key goal  of Section  111 is to enhance
air quality over the long term and minimize the potential for long-term
growth by minimizing emissions through application of  the best demonstrated
technology to new emission sources, concurrent with  the turnover  of
the Nation's industrial  base.  If owners are permitted to replace  most
of the equipment in their existing facilities without  applying the best
demonstrated technology,  they will  be installing new equipment without
minimizing emissions and maximizing the potential  for  long-term industrial
growth, as Congress sought in enacting Section 111.  For  this reason,
NSPS coverage of facilities that undergo substantial component replacement
through conversion accords with Section 111, even where some decrease
in emissions results from the conversion.
     EPA promulgated the reconstruction provisions because failure to
require best control at sources that have become essentially new  through
extensive component replacement would have undermined  Congress's  intent
that best technology be applied as the Nation's industrial base is
replaced.  Failure to cover facilities that have undergone extensive
component replacement over a long period of time similarly postpones
the enhancement of air quality Congress sought under Section 111.   The
D.C. Circuit recognized this when it expressed concern in the ASARCo
case that, absent a provision for aggregating replacement expenditures
"over the years," owners could evade the reconstruction provisions by
continually replacing obsolete or worn-out equipment.  578 F.2d 319,
328 n.31 (D.C. cir. 1978).
     Section 60.15 currently defines "reconstruction"  as  the replacement
of components of an existing facility to such an extent that "the  fixed
capital cost of the new components" exceeds 50 percent of the "fixed
capital cost" that would be required to construct a  comparable entirely
new facility and EPA determines that it is technologically and economically
feasible to meet the applicable NSPS.  Subsection (d)  indicates that
the "new components" whose cost would be counted toward the  50 percent
threshold include those components the owner "proposes to replace."  It
                                  5-2

-------
 is  unclear  under  this wording whether a reconstruction has occurred in
 the  case  of an owner who first seeks to replace components of an existing
 facility  at a cost equal to 30 percent of the cost of an entirely new
 facility  and then, shortly after commencing or completing those
 replacements, seeks to replace an additional 30 percent.  Specifically,
 it is uncertain whether the owner should be deemed to have made two
 distinct  "proposals," or instead a single proposal.
     For  example, assume that a refinery owner refurbishes part of a
 facility  and six months later refurbishes other parts of the facility.
 If the two  actions were interpreted as separate "proposals" under
 Section 60.15, neither might exceed the 50 percent replacement
 cost threshold.  Under this general  interpretation, owners could avoid
 NSPS coverage under Section 60.15 simply by characterizing their
 replacement projects as distinct "proposals," even where the component
 replacement is completed within a relatively short period of time.
     EPA  did not intend, in promulgating the reconstruction provisions,
 that the  term "propose" exclude from NSPS coverage facilities undergoing
 extensive component replacement.   Failure to cover these sources serves
 to undermine Congress's intent that air quality be enhanced over the
 long term by applying best demonstrated technology with the turnover in
 the Nation's industrial  base.
     To eliminate the ambiguity in the current wording of Section
60.15 and further the intent underlying Section 111,  the Agency is
clarifying the meaning of "proposed" component replacements in Section
60.15.   Specifically, the Agency  is  interpreting "proposed" replacement
components under Section 60.15 to include components  which are replaced
pursuant to  all  continuous  programs  of component replacement which
commence (but are not necessarily completed)  within the period of time
determined by the Agency to be appropriate for the individual  NSPS
 involved.  The Agency is selecting a 2-year period as the appropriate
period for purposes  of the  petroleum refinery equipment leak NSPS
 (Subpart 66G).   Thus,  the Agency  will  count toward the 50 percent
reconstruction  threshold the "fixed  capital  cost"  of  all  depreciable
components (except those described above)  replaced pursuant to all
continuous programs  of reconstruction  which  commence  within any 2-year
period following proposal  of these standards.   In  the Administrator's
                                  5-3

-------
judgment, the 2-year period provides a reasonable, objective method of
determining whether an owner of one of these facilities is actually
"proposing" extensive component replacement, within the Agency's original
intent in promulgating Section 60.15.
     EPA realizes that the petroluem refinery industry is constantly
changing; however, the Agency believes that this 2-year limit will
assure that the owner would have to make a substantial change to the
facility to reach the 50 percent threshold.
     The administrative effort to keep the required records should not
be a burden on industry.  The recordkeeping required under a cumulative basis
interpretation of reconstruction is the same as the recordkeeping that
would be required under a strictly project-by-project basis interpretation.
In either case, the dollar amount of the component replacement taking
place at the affected facility must be determined and recorded.  Once
this dollar amount has been determined for each change and conversion,
the additional requirement of keeping this information on file at the
refinery does not appear to be an excessive burden.
Comment:
     Two commenters (IV-D-8 and IV-D-14)  requested EPA to exclude from
the reconstruction provisions the costs of equipment replacement done
for routine maintenance purposes.   Similarly, commenters (IV-D-4 and
IV-D-15) expressed concern with the reconstruction provisions as they
apply to process unit turnarounds.   Commenters stated that process unit
turnarounds are maintenance procedures performed to assure efficient
and safe operation and,  therefore,  turnaround replacements should be
excluded from reconstruction provisions.   Another commenter (IV-D-4)
requested that replacements of equipment  due to fire,  explosions, or
other accidental  causes  should be  exempt  from reconstruction.
Response:
     As discussed above, reconstruction costs are the  fixed capital  cost
or the capital  needed to provide all  the  "depreciable"  components,
while most routine maintenance practices  involve the use of non depreciable
components.
     Because routine maintenance items (valve packings,  pump  seals,
replacement rupture disks,  nuts and bolts)  cost very little compared  to
                                  5-4

-------
the cost of equipment (covered by the standards)  in a process unit,  it
is very unlikely that routine maintenance would trigger a reconstruction
even if accumulated over several  years.  The cost of these items is
relatively small.  In EPA's judgment, maintaining records of the repair
or replacement of these items may constitute an unnecessary burden.
Moreover, EPA does not consider the replacement of these items an
element of the turnover in the life of the facility.  Therefore, in
accordance with 40 CFR 60.15(g),  the final standards (Subpart GGG) will
exempt certain frequently replaced components from consideration in
applying the reconstruction provisions to petroleum refinery process
unit facilities.
     The costs of these frequently replaced valve parts will not
be considered in calculating either the "fixed capital  cost of the new
components" or the "fixed capital cost that would be required to construct
a comparable, entirely new facility" under Section 60.15.  In EPA's
judgment, these items are pump seals, valve packings, nuts and bolts,
and rupture disks.  Replacements  of pumps, valves, and other fugitive
equipment at turnarounds or at other times are included in reconstruction
costs.  For turnarounds that involve significant refurbishment of a
process unit, EPA would likely consider this a reconstruction.  EPA
also considers it appropriate to  include in reconstruction costs the
replacement of equipment due to the accidental  loss of an original
component, since the reason for an owner's refurbishing a facility has no
bearing on whether the facility itself is comparable to a new source
for which application of the best control systems is reasonable.
Comment:
     One of the commenters (IV-D-14) requested that the reconstruction
provisions not apply at all, or only when the number of replacement
valves exceeds 50 percent of the  number of existing valves.  The commenter
reasoned that there is an economic justification  for requiring compliance
with NSPS if, for example, reactors, towers, or heaters are replaced,
but not fugitive emission sources.
Response:
     The standards apply to fugitive emission sources only.  EPA considers
it appropriate to cover process units that are essentially new.   The
costs considered are only those associated with the equipment covered
                                  5-5

-------
by the standards.  The commenter offered no support for his statement
that the standards are not economically justified when applied to
fugitive emission sources only.  EPA considers the standards to reflect
BDT (considering costs) for sources that become affected through recon-
struction or modification.  In response to the commenter1s preference
that the affected facility include valves only, EPA does not disagree
with the concept of using the number of components as the basis for
reconstruction.  However, since there are several  types of components
covered by the standards, this approach would ignore replacements of
other key portions of the facility.  Thus, EPA will  use the cost of
replacements for all the equipment covered by the standards to determine
when a reconstruction has occurred.
                                  5-6

-------
                               6.0  LEGAL

Comment:
     One commenter (IV-D-21) requested that facilities commencing
modification or reconstruction should be subject to compliance on the
date of final promulgation rather than January 4, 1983, the date
of proposal.
Response:
     Under Section lll(a)(2) of the Clean Air Act, a "new source"
subject to applicable standards is defined to be any source on which
construction or modification commenced after the date on which the
standards were proposed.  The standards for equipment leaks of VOC
within petroleum refineries were proposed on January 4, 1983.   A
group of process unit equipment (specified in the standards) or a
compressor on which construction or modification commenced after that
date is, therefore, a new source under the Act and subject to  the
standards.  The commenters suggest that EPA change the applicability
date to the date on which EPA promulgates the standards.  Changing the
applicability date of the standards would be inconsistent with the
plain language of the Clean Air Act.
Comment:
     Two commenters (IV-D-21 and IV-D-22) argued that the proposed
standards are unnecessary because hydrocarbons generally do not affect
human health as reflected in the EPA's rescinding of the national ambient
air quality standards (NAAQS) for hydrocarbons (HC), and because there is
no consistent quantitative relationship between the concentration of
ambient air ozone and hydrocarbons.  Commenters added that there
is no need to regulate VOC in attainment areas.
Response:
     The revocation notice for the NAAQS for HC does not directly
affect the development of this NSPS.   As explained in the revocation
                                  6-1

-------
notice, the NAAQS for HC were intended only as guides in the development
of State implementation plans (SIP) to attain the original NAAQS for
photochemical oxidants (recast as NAAQS for ozone in 1979).  EPA revoked
the NAAQS for HC because EPA determined that there is no single, univer-
sally applicable relationship between HC and ozone and that HC as a
class apparently do not produce any adverse health or welfare effects
at concentrations at or near ambient levels.  However, the revocation
was in no way intended to restrict EPA or State authority to limit VOC
emissions (including HC as a class) where necessary to limit the for-
mation of ozone.  Since VOC are precursors to ozone, and ozone has been
determined to be harmful  to public health and welfare, significant
sources of VOC are subject to regulation under Section 111 of the Clean
Air Act (46 FR 25656; May 9, 1981).
     EPA clearly documented the need to regulate VOC in order to protect
public health and welfare in the EPA publication "Air Quality Criteria
for Ozone and Other Photochemical  Oxidants" (Docket No. IV-A-1).  VOC
emissions are precursors  to the formation of ozone and other oxidants
(ozone).  Ozone results in a variety of adverse impacts on health and
welfare, including impaired respiratory function, eye irritation,
necrosis of plant tissues, and the deterioration of synthetic rubber.
     In setting new source performance standards, 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 NAAQS under Sections 109 and 110 of the Clean Air Act.
The intent of Congress in establishing NSPS was to establish a single
level  of stringency for all  State  limits, thereby preventing States
from soliciting industry  with lenient air pollution requirements and
causing increased air pollution from new sources.  The standards will
limit VOC emissions from  newly constructed, modified, and  reconstructed
refinery process units and will  result in emission reductions well  into
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 preventing future air quality problems.
Clearly, residents in both attainment and nonattainment  areas would
benefit from these standards.  The NSPS complements  the  ambient  air
quality-based rules as a  means of  achieving and maintaining  the  NAAQS,

                                  6-2

-------
while on a broader basis it prevents new sources  from making  air
pollution problems worse regardless of the existing  quality of  ambient
air.  Therefore, while new source standards may help in  the attainment
of NAAQS, the consideration of attainment or nonattainment  of the  NAAQS
does not influence directly the decision to set standards of  performance.
Comment:
     The same commenters (IV-D-21 and IV-D-22)  stated that  the  standards
would place an unreasonable burden upon the industry.  The  standards
should be considered a major rule and not be exempt  from provisions of
Executive Order 12291.
Response:
     Executive Order 12291 requires that a regulatory impact  analysis,
thoroughly examining costs and benefits of a rule, be prepared  in
connection with every major rule.  A major rule is any regulation
which 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
          United States-based  enterprises to compete with foreign-based
          enterprises in domestic or export markets.
     An economic analysis of these standards was  prepared.  Economic
impact estimates presented in  the background information document  for
the proposed standards, and summarized in the preamble to the proposed
regulation (48 FR 279; January 4, 1983), showed that no unreasonable
economic impacts are expected.  Because no unreasonable economic  impacts
are expected and none of the criteria for a major rule has  been met,  no
additional regulatory impact analysis has been prepared.
Comment:
     One commenter (IV-D-21) further stated that  EPA acknowledged  that
no new major refineries are likely to be constructed in the U.S.  in the
coming decade and, thus, all the emission reductions quantified in the
background information document would occur at existing refineries.
                                  6-3

-------
Response:
     Projections of units affected by new source performance standards
are discussed in Appendix E of the BID for the proposed  standards.   EPA
projected that up to 100 new units and 182 modifications/reconstructions
of existing process units will  be subject to the standards.   EPA recognizes
that few, if any, grass root refineries will  be built.   However, EPA
also recognizes that it is appropriate to cover the  industry as  it
rebuilds through modification and reconstruction and through the addition
of new processing units at existing refineries.
                                  6-4

-------
                           7.0  TEST METHODS

 Comment:
      One  commenter  (IV-D-4) requested that EPA propose the entire
 rulemaking, including appropriate reference methods, at one time.
 The  commenter  said  that the "failure to provide either Reference Method
 21 or Reference Method 22 as appendices to these proposed rules prevents
 an accurate analysis of the impact of the proposed rulemaking."  The
 application of Reference Method 22 to refinery flares was questioned.
 Response:
      EPA  proposed Reference Method 21 on January 5, 1981, as an appendix
 to the proposed standards of performance for fugitive VOC emission
 sources in the synthetic organic chemicals manufacturing industry (SOCMI).
 EPA  generally proposes reference methods in conjunction with the first
 standards that use  the method.  Method 21 would normally have been
 promulgated with the SOCMI standards.  However, after reviewing the
 comments  and incorporating changes, it was decided to promulgate Method 21
 before promulgation of the SOCMI NSPS because several  additional regulations
 were scheduled for  promulgation in the near future that specified the
 use  of Method 21.   EPA considered the comments received during the comment
 period for the proposed refinery fugitive standards and decided that no
 additional changes  to Method 21 were needed.   Method 21 was promulgated
 on August 18, 1983  (48 FR 37598).
      Reference Method 22 was initially promulgated on August 6, 1982.
 In the January 4, 1983,  preamble to the proposed petroleum refinery
 standards, EPA stated that revisions to Method 22 would be published
 soon in the Federal  Register to broaden its applicability to flares.
This method was revised on October 18, 1983,  in the rulemaking on
SOCMI.
                                  7-1

-------
Comment:
     Two commenters (IV-D-4 and IV-D-15) requested that EPA clarify
the use of hexane or methane in calibrating the portable analyzer.  It
was suggested that a correction factor be provided to put all measurements
on a consistent basis using hexane as the primary standard.  The use of
an unconnected methane calibnation would nesult in a highen numben of
leaking components being detected.
Response:
     The basis fon selection of the calibnation gases fon the analyzen
was evaluated befone pnoposal.  It was necognized that thene ane a
numben of potential pnocess stneam components and compositions that can
be expected.  Since all analyzen types do not nespond equally to diffenent
compounds, it was necessany to establish a nefenence calibnation matenial.
Based on the expected compounds and the infonmation available on instnument
nesponse factons, hexane was chosen initially (see Contnol  of Volatile
Onganic Compound Leaks fnom Petnoleum Refineny Equipment, EPA-450/2-78-036,
Document No. II-A-6) as the nefenence calibnation gas fon EPA test
pnognams.  At that time, the measunement distance was 5 centimetens
(cm), and calibnations using hexane wene conducted at appnoximately 100
on 1,000 ppm levels.  Aften initial equipment leak data wene collected
at 5 cm, pnoblems wene identified with the nepnoducibility of nesults
at this distance, as discussed in Appendix D of the BID fon the pnoposed
standands.  The monitoning pnocedune was nevised so that measunements
wene made at the sunface of the intenface, on essentially 0 cm.  This
change was accompanied by a change in the leak definition to 10,000 ppm.
At this concentnation hexane calibnation standands wene not neadily
available commencially.  Also based on a neview of the data, it appeaned
that methane was a mone nepnesentative nefenence calibnation matenial
at 10,000 ppm levels.  Based on this conclusion, and the fact that
methane calibnation standands ane neadily available at the  necessany
calibnation concentnations, methane was added as an acceptable calibnation
gas.
     Since then, studies have been completed  that measuned  the nesponse
factons fon sevenal  instnument  types.  The nesults of these studies
show that the nesponse factons  fon methane and hexane ane similan
enough fon the punposes of this method fon these two gases  to be used

                                  7-2

-------
as calibrants interchangeably.  Therefore, the accepted calibration
materials remain as hexane and methane.  In response to the commenters,
EPA will likely use methane as the calibration gas.   Because EPA does
not consider the difference in the number of leaks found using  either
calibration gas to be substantially different, a correlation factor  to
put all measurements on a consistent basis was not provided.
Comment:
     One commenter (IV-D-22) warned that when used at a concentration
of 10,000 ppm, hexane could condense on the walls of the container,
resulting in distorted calibration results.  Also, since the lower
explosive limit for hexane is 12,000 ppm, calibrating with hexane could
be a safety hazard.
Response:
     There are a number of difficulties with using hexane as a  calibrant.
Based on EPA's experience, methane is the preferred  calibrant.   The  use
of hexane may lead to operators finding more leaks during monitoring
because, if hexane condenses on the walls of the container storing the
instrument calibration gas, the concentration of the gas may fall below
10,000 ppm.  In this instance, an instrument would signal that  a leak
has occurred although the actual concentration is below 10,000  ppm.
The fact that 10,000 ppm as hexane is close to 12,000 ppm (lower
explosive limit of hexane) can be added to the factors that led EPA  to
require the instrument to be intrinsically safe for  operation in explosive
atmospheres.
Comment:
     Two commenters (IV-D-6 and IV-D-12) remarked on the reasonableness
of the instrument calibration requirements.  It was  argued that the
"zero" calibration could be performed adequately with ambient air.
Also, daily instrument calibration was deemed too burdensome and it
was felt that weekly calibration should suffice.
Response:
     The specification of air (less than 10 ppm hydrocarbon) as the
zero air calibrant was intended to allow the use of  relatively  clean
ambient air.  Method 21 now specifies 10 ppm, whereas it specified
3 ppm when the standards were proposed.  The use of  air with less than
10 ppm hydrocarbon does allow calibration of the instrument at  essentially
                                  7-3

-------
zero reading.  This is particularly important for "no detectable emission"
requirements and to ensure that the monitor is functioning properly.
Zero air calibrants can be purchased or generated easily (e.g.,  carbon
filtered drawn air).  Thus, there is no need to change the standards  to
require the zero air calibrant as ambient air only.   There may be
occasions when the ambient air within a refinery could have a significant
VOC (organics) concentration, and calibrating with that ambient  air
would be inappropriate.
     Instrument calibration is required on each day  of its use.   This
is not burdensome.  The procedure is relatively simple, does not require
a laboratory, and takes only 15 to 30 minutes.  The  cost of calibrating
the instrument on a daily basis was included in the  cost of leak detection
and repair programs.  Moreover, the proper use and calibration of the
monitor is vital to effective leak detection and repair.  A semiannual
performance evaluation of the instrument is also required by Method 21.
EPA has no reason to believe that weekly calibration would provide
sufficiently stable readings from the monitor.  EPA's experience indicates
that daily calibrations are sufficient, and that less frequent calibrations
may not be adequate.  Thus, no change in instrument  calibration  requirements
was made.
Comment:
     One commenter (IV-D-22) noted that proposed Reference Method 21
refers to monitoring techniques which do not distinguish between VOC
and non-VOC hydrocarbons.  This may result in a component having a
monitor reading greater than 10,000 ppm while actual  VOC emissions
would be less.
Response:
     The commenter is correct that Method 21 responds to non-reactive
organic compounds (e.g., methane).  However, the monitor reading is not
intended to be a quantitative measure of the reactive organic compounds
(VOC) in the leak.  Its purpose is rather to indicate whether a  leak
exists of sufficient magnitude to warrant remedial  action.  EPA  is
using the "in VOC service" definition to exclude equipment that  would
not contain enough reactive organic compounds to warrant coverage by  the
standards.  Thus, if a piece of equipment is in VOC  service and  a leak
of 10,000 ppm is detected, EPA judges that repair is warranted.   For
                                  7-4

-------
 this reason, correcting the 10,000  ppm leak  definition  to  "VOC only"
 is unnecessary.
 Comment:
      A commenter (IV-D-22)  suggested  that  there was  inconsistency in
 EPA's decision not to  allow leak detection using a soap solution because
 the magnitude of leak  rates is  difficult to  assess,  yet the VOC
 monitoring  instrument  "would yield  qualitative indications of leaks."
 Response:
      Since  proposal, an alternative screening procedure has been added
 to Method 21 for those  sources  that can be tested with a soap solution.
 These sources are restricted to those  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  appli-
 cable sources  and  the potential leak sites are observed to determine if
 bubbles are formed.  If no  bubbles  are formed, then  no detectable
 emissions or leak  exists.   If any bubbles are formed, then the instru-
 ment  measurement  techniques must be used to determine whether a leak
 exists, 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 valve 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  regu-
 lation 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  cor-
 rectly identify and score a standard series of test bubbles.   It  has
been reported that trained observers can  correctly  and repeatedly
                                  7-5

-------
 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 the applicable regulation.  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;
     A commenter (IV-D-8) argued  that  it is unreasonable to require
 annual monitoring  of "entire vent or  control systems."  These piping
 systems  are  typically  operated at low pressure (which minimizes the
 amount of potential  leakage) and  are  routed overhead in pipe racks and
 are, therefore, inaccessible.
 Response:
     Method  21 is  used to monitor closed vent systems used in complying
 with the standards.  Method 21 requires the use of an organic compound
 monitor only where  leaks might occur.   Where no leaks can  occur (like
 header-pipes), Method  21 requires only a visual  inspection to ensure
 the closed vent system has not deteriorated and is  not leaking  where
 leaks are not expected.
     Closed vent systems used to comply may be  operated at low  pressure
 or high pressure.   Either type of system may leak at  connections and,
 therefore, the annual test is appropriate.   If  an owner or operator
 uses an approach of ensuring a leak-free system,  such  as monitoring
 oxygen in a vacuum  system, EPA will  consider whether  this  approach can
 be used rather than Method 21, as specified in  §60.13(i) of the  General
 Provisions.  Like other sources that are difficult-to-monitor,  annual
monitoring, if needed,  in  a  pipe  rack  is not unreasonable  in  light  of
the emissions that  would occur from  such a  leak.
                                 7-6

-------
                    8.0  RECORDKEEPING AND REPORTING
Comment:
     A number of commenters (IV-D-8, IV-D-12, IV-D-21, and IV-D-22)
wrote that the proposed recordkeeping requirements are needlessly
complex and burdensome.  One commenter (IV-D-22)  estimated that the
additional paperwork burden would amount to 2 person-months per year
per affected refinery.  Two commenters (IV-D-8 and IV-D-12) listed
specific recordkeeping requirements that should be deleted.  These
requirements include records of: (1) identification numbers for instru-
ments, operators, fugitive emission components, leaking components,
components in vacuum service, and components designated as difficult-to-
monitor or unsafe-to-monitor; (2) repair methods; (3)  logging shutdown
and startup for closed vent systems; (4) signature of  owner or operator
(or designee) whose decision it was that repair could  not be effected
without a process shutdown; (5) expected date of repair; (6) explanation
for unsafe or difficult-to-monitor designation; and (7) schematics,
design specifications, and operations records on  flares used as control
devices.
Response:
     Before the standards were proposed, EPA considered three alternative
levels of recordkeeping.  The proposed recordkeeping requirements are
considered the minimum consistent with adequate enforcement; thus, the
paperwork burden on owners and operators is the minimum amount necessary
to enforce the standards adequately.  At proposal, EPA weighed the
paperwork burden against the enforcement authority (Federal, State and
local) to determine compliance with the standards and  selected the
proposed requirements.
     Compliance with the final standards will be generally determined
through inspection.  However, because the intent  of the standards is a
                                  8-1

-------
continuous reduction in equipment leaks of VOC and continuous inspection
by enforcement authorities is not possible, records must be maintained
if an inspector is to determine retrospectively whether a facility is in
compliance with the standards.  EPA considers the required records for
an owner or operator's leak detection and repair program necessary to
document the operator's compliance efforts.  These records would  likely
be maintained by a prudent owner or operator, and should therefore add
little additional  recordkeeping burden.
     Commenters did not explain why specific records would not be needed
by enforcement authorities.  EPA considers the required records for an
owner or operator's leak detection and repair program necessary to docu-
ment the operator's compliance efforts.  For example, when an unsuccessful
repair attempt is made, a record of the attempted or anticipated  methods
of repair shows what effort was made by the operator and the reason for
delay.  Without such records, EPA and other enforcement authorities
would not be able to determine compliance with the standards.  Addi-
tionally, an expected repair date is obviously required in such cases
to prevent a known leak from being allowed to persist.  Records,  such
as identification numbers for components in vacuum service, can be used
to check compliance with the standards.  The same reasons are applicable
to records for operation of control devices.  Obviously, a control
device (including flares) serves no function when not operating.   As
such, demonstration of shutdown or flame-out periods is necessary to
show compliance.  These records would likely already be maintained by a
prudent owner or operator, and should therefore add little additional
recordkeeping burden.
     The records required for identifying fugitive emission components,
and control device schematics and design data are not unreasonably
burdensome.  This information would be developed only once, and would
require changing or updating only if the facility were changed.  The
control device schematics and design data should be available to  plant
engineers already, and as such do not represent an added burden.   For
new facilities, the reasons why a component must be installed in  a
location which makes it difficult or unsafe to monitor must be documented
prior to installing the component in such a position.  The number of
difficult-to-monitor or unsafe-to-monitor components will  be small  and,
                                  8-2

-------
therefore, should not create an excessive recordkeepiny burden.  After
considering the comments that the recordkeeping requirements are needlessly
complex and burdensome, EPA decided to promulgate the recordkeeping
requirements as proposed.
Comment:
     One commenter (IV-D-22) complained that the demonstration required
for a variance from the 15-day repair requirement is a recordkeeping
nightmare due to the many different types and sizes of valves.
Response:
     The proposed standards permit delay of repair beyond the 15-day
period as provided in Section 60.592-9.  The provisions for delay for
repair are automatic, and not a variance (as the commenter implied)
that must be applied for as suggested by the commenter.  EPA recognizes
that some repairs cannot be performed on line, and that not all compo-
nents can be isolated without a process unit shutdown.  These repairs
should be readily understood by the operator.  Therefore, a relatively
straightforward response (e.g., the seal must be replaced at a shutdown--
pump cannot be bypassed; there is no spare) sufficiently informs EPA
why repair is delayed for a particular component and, accordingly, can
be used to determine whether compliance with the standards has been
maintained.  The intent of the recordkeeping provision is to ensure
that all technically feasible repairs are performed within 15 days.
Comment:
     A commenter (IV-D-14) wrote that the actual cost for a leak
detection and repair program as required by the standards would be
higher than EPA estimated because daily recordkeeping of components
replaced frequently is not included, nor are the associated costs
necessary to determine when a process unit becomes an affected facility.
Response:
     The recordkeeping associated with frequent replacements and
evaluating changes in operation would be something a plant would typi-
cally do on its own for purposes other than complying with the standards
(e.g., tracking the cost of production or assuring that adequate spare
parts or components are stocked).  A small  additional increase might be
                                  8-3

-------
experienced by the owner as a result of these standards.  However, the
entire cost of recordkeeping should not be attributed to the standards
as indicated by the commenter.  The exclusion of certain routine replace-
ment items from reconstruction calculations (see Section 5) and the
addition of alternative methods for determining capital expenditure
(see Section 4) eases the recordkeeping needs.  The cost of pre-
construction efforts, pre-modification efforts, or pre-reconstruction
efforts are not accounted for explicitly in the impacts of the standards
because they are considered part of the overall cost of the owner's
decision to construct, modify, or reconstruct.  As such, it would be
very difficult to make reasonable estimates of the cost to determine when
a process unit becomes an affected facility.  But, EPA believes that
those costs are insignificant in comparison to the costs associated
with other activities that occur during these efforts.  In any cases,
these costs would not raise the overall cost effectiveness ($130/Mg) to
an unreasonable level.
Comment:
     Two commenters (IV-D-30 and IV-D-22)  remarked that the standards are not
enforceable.  As a result one of the commenters (IV-D-22)  concluded that
those facilities not following the NSPS would have a competitive advantage
of reduced cost over those facilities complying with the regulations.
Additionally, the other commenter (IV-D-30) strongly opposed the proposal
of no reporting requirements which, the commenter noted, undermines the
effectiveness of the standards and the ability of EPA and  the States to
enforce them.  Reporting requirements, according to the commenter,
would enhance self-enforcement.  The commenter speculated  that the
reason for EPA's excluding reporting requirements was OMB's role in
administering the Paperwork Reduction Act,  which, the commenter asserted,
does not give OMB or EPA the authority to  compromise the effectiveness
of the standards.  The commenter was also  concerned that no records are
required for equipment not found to be leaking, adding that a much
better incentive to comply with the regulation would exist if records
were required to be kept on all  monitored  equipment.
Response:
     Reports, records, and inspections will  be used to ensure compliance
by all  facilities subject to these standards.   State and EPA Regional

                                   8-4

-------
air quality control authorities have successfully implemented  regulations
similar to the standards.  At proposal  EPA stated that routine reporting
was not required.  Reporting requirements were limited to notifications
of construction, anticipated startup and actual  startup, and an intention
to comply with one of the alternative standards.  As stated in the
preamble for the proposed standards, these reporting requirements would
not provide a mechanism for checking the thoroughness of the industry's
efforts to reduce fugitive emissions of VOC.  As stated in the preamble
to the proposed standards, compliance would be assessed through in-plant
inspections.
     EPA has decided that reporting is necessary to assess implementation
of the work practice and equipment requirements of the standards.  EPA
agrees with the commenter that facilities not complying with the stan-
dards might have an unfair advantage (albeit, somewhat small).  More
importantly, facilities not complying with the standards would not be
using BDT as required by the Clean Air Act, the purpose of which is to
prevent new air pollution problems.  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 enforceability of the standards.  EPA's conclusion that reports are
useful  is also based on the experience of the State  and  local  air quality
control boards.
     As explained  at proposal, three alternatives were  considered for
reporting  requirements.  The three alternatives  represented trade-offs
among  varying amounts of in-plant  inspections and report  preparation
for enforcement.   The first  alternative  required minimal  reporting and
relied  on  inspections for enforcement.   The third alternative relied
almost  totally  on  reports and  would  require minimum inspections  to
judge  compliance.   The  second  alternative  represented a compromise with
some  reporting  and some inspections  required  and is included  in  the
final  regulations.  These  reporting  requirements,  however, have  been
 streamlined to  include  reporting of data on  leak detection and repair
of pumps,  valves,  and  other equipment  types  only.   In addition,  periodic
 reports are on  a semiannual  rather than quarterly basis, and  the require-
ment  for certification  of  reports has  been eliminated.  The semiannual
                                   8-5

-------
reporting requirements may be waived for affected sources in any State
that is delegated authority to enforce these standards, provided EPA
approves reporting requirements or an alternative means of source
surveillance adopted by the State.  Such sources would be required to
comply with the requirements adopted by the State.
     The following reporting requirements were added to the standards
since proposal:
     Each owner or operator must submit semiannual reports
beginning 6 months after the initial startup date.  The initial  semi-
annual  report includes:
     (1)  Process unit identification,
     (2)  Number of valves subject to the requirements excluding those
valves designated for no detectable emissions,
     (3)  Number of pumps subject to the requirements excluding those
pumps designated for no detectable emissions and those pumps enclosed
and vented to a control device, and
     (4)  Number of compressors subject to the requirements excluding
those compressors designated for no detectable emissions and those
compressors enclosed and vented to a control device.
     All subsequent semiannual reports must include:
     (1)  Process unit identification, and
     (2)  For each month during the semiannual reporting period,
the number of valves, pumps, and compressors for which leaks were
detected and the number of valves, pumps, and compressors for which
leaks were not reported.
     The semiannual reports will present the facts that explain each
delay of repair and, where appropriate, why a process unit shutdown was
technically infeasible.  In addition, the semiannual reports will  give
dates of process unit shutdowns which occurred within the semiannual
reporting period, and revisions to items reported according to paragraph
(b) if changes have occurred since the initial report or subsequent
revisions to the initial report.
     The Paperwork Reduction Act of 1980 (PL-511) requires clearance
from the Office of Management and Budget (OMB) of reporting and
recordkeeping requirements that qualify as an "information collection
request" (ICR).  For the purposes of OMB's review, an analysis of the
                                  8-6

-------
burden associated with the reporting and recordkeeping requirements of
this regulation has been made.  During the years 1984 and 1985, the
average annual  burden of the reporting and recordkeeping requirements
of this regulation to industry would be about 20 person-years.
                                  8-7

-------

-------
                       APPENDIX A

INCREMENTAL COST EFFECTIVENESS OF CONTROL TECHNIQUES FOR
                 EQUIPMENT LEAKS OF VOC
                          A-l

-------
                               APPENDIX A
          INCREMENTAL COST EFFECTIVENESS OF CONTROL TECHNIQUES
                       FOR EQUIPMENT LEAKS OF VOC

     Table A-l summarizes the individual component control  impacts and
the incremental cost effectiveness for each individual  component and
control technique.  The individual component control  impacts are derived
in Tables A-2 through A-13.  The net annualized cost, emission reduction,
cost effectiveness, and incremental  cost effectiveness  of the control
techniques are discussed in Chapter 2.0.
                                 A-2

-------
       Table A-l.    SUMMARY  OF  THE  INDIVIDUAL COMPONENT  CONTROL  IMPACTS3
Fugitive Emission
Source
Pressure relief devices
Compressors
Open-ended valves
Sampling connection
systems
Valves
Pumps

Control Technique
Quarterly LOR
Monthly LDR
Rupture disks'3
Controlled degassing
vents
Caps on open ends
Closed purge sampling
Quarterly LDR
Monthly LDR
Sealed bellows valves
Annual LDR
Quarterly LDR
Monthly LDR
Dual mechanical seal
system
Emission Reduction
(Mg/yr)
4.4
5.3
9.8
16.5
2.8
2.6
66
77
110
3.0
9.8
11.5
13.9
Average Cost
Effectiveness6
($/Mg)
(170)
(110)
410
150
460
810
(110)
(60)
4,700
. 860
157
158
2,000
Incremental Cost
Effectiveness0
($/Mg)
(170)
250
1,000
150
460
810
(110)
310
16,700
860
(140)
170
10,900
(xx)  •  Cost savings
LDR " Leak detection and repair.


 Costs  and emission reductions are based on fugitive emission component counts in Model  B from the BID for the proposed
 standards, EPA-450/3-81-015a, page 6-3, and from Tables  A-2 through A-13  of this appendix.
b
 Average Cost Effectiveness * net annualized costs per component + annual  VOC emission reduction per component.

 Incremental Cost Effectiveness - (net  annualIzed cost of the control technique - net annual 1 zed cost of the  next less
 restrictive control  technique) « (annual emission reduction of control technique - annual emission reduction of the
 next less restrictive control technique).
d
 Underlined control techniques were selected as basis for standards.
                                                  A-3

-------
Table A-2. PRESSURE RELIEF DEVICE IMPACTS
(May 1980
Per Pressure
Relief Device
Item/Control Technique

Installed Capital Cost
Annual i zed Capital Costs0

A. Control Equipment
B. Initial Leak Repair^
Annual i zed Operating Costs
A. Maintenance6
•' B. Miscellaneous^
1. Monitoring9
2. Leak Repaird
3. Administrative
and Supporth
Total Annual Costs
Before Credit
Recovery Credit1
Net Annual i zed CostsJ
Total VOC Emission
Reduction (Mg/yr)^
Cost Effectiveness
($/Mg VOC)1
Incremental Cost
Effectiveness"1
($/Mg VOC)
dollars)
Quarterly
LDR
a
0

a
0

__a
__a
19
0

7.6

27
135
(110)

0.63

(170)


(170)

Monthly
LDR
a
0

a
0

__a
..a
58
0

23

81
161
(80)

0.75

(110)


250

Rupture
Disks
b
3,100


600
0

160
120
0
0

0

880
300
580

1.4

410


1,000
(XX) = Cost Savings
(LDR) = Leak Detection and Repair

Model Unit B:  7 pressure relief valves
  Emission Reductions
        Quarterly LDR = 7 x 0.63 Mg/yr = 4.4 Mg/yr
        Monthly LDR = 7 x 0.75 Mg/yr = 5.3 Mg/yr
        Rupture Disks  = 7 x 1.4 Mg/yr = 9.8 Mg/yr
                                A-4

-------
          Table  A^2,  PRESSURE RELIEF DEVICE IMPACTS
                       (May 1980 dollars)
                         (Continued)

aCost of monitoring instrument is not included in this analysis.

bCapital cost for rupture disk is from BID for proposed standards
 Table 8-1.

C0btained by multiplying capital recovery factor (2 years, 10 percent
 interest = 0.58) by capital cost for rupture disk and capital recovery
 factor (10 years, 10 percent interest = 0.163) by capital cost for all
 other equipment (rupture disk holder, piping, valves, pressure relief
 valve).  Based on new installation cost 0.163 (3100 - 230) + 0.58
 (230) = 600.

dLeaks are corrected by routine maintenance in the absence of the
 standards; therefore,  no cost is incurred for repair.

e0.05 x capital  cost.  From BID for proposed standards, Table 8-5.

fQ.04 x capital  cost.  From BID for proposed standards, Table 8-5.

9Mom'toring labor hours (i.e., number of workers x number of components
 x time to monitor x times monitored per year) x $18 per hour.  Assumes
 2-man monitoring team per relief valve, 8 minutes monitoring team per
 valve.

n0.40 x (monitoring cost + leak repair cost).  From BID for proposed
 standards, Table 8-5.

1Recovery  credit based  on uncontrolled VOC emission factor of 3.9 kg/day
 (BID Table 3-1) and 44 percent control  efficiency for quarterly
 inspections (BID Table F-7), 53 percent control efficiency for monthly
 inspections, and 100 percent for rupture disks.  Control  efficiency
 for monthly inspections is estimated based on the method used to
 calculate control  efficiency for quarterly inspections in Table F-7
 (footnote f) of the BID for the proposed standard.  [Ratio of estimated
 control  efficiency for gas/vapor valve  ABCD model  (monthly inspections)
 to gas/vapor valve LDAR model estimate  (BID Table F-3) multiplied by
 safety/relief valve ABCD model  untrol  effectiveness for monthly
 inspections (0.68) based on Table 7-1,  BID for proposed standard,
 ABCD factors:   A = 0.74, B = 0.95,  C =  0.98, D = 0.98].  Therefore,
 control  efficiency for monthly inspections = (Q 53) (0.703) - o 53
                                                     (0.91)

 Recovered product  valued at $215/Mg VOC (from BID Table 8-5).
 Recovered emissions:
 Quarterly LDR = 3.9 kg x 0.44 x 365 days x 1 Mg    =  0.63 Mg	
                  day               yr      1000 kg    yr-relief device

 Monthly  LDR = 3.9  kg x 0.53 x 365 days  x 1 Mg    = 0.75 Mg
                 day              yr      1000 kg   yr-relief device
                                 A-5

-------
            Table A~2.  PRESSURE RELIEF DEVICE IMPACTS
                         (May 1980 dollars)
                            (Concluded)
 Rupture Disk  = 3.9 kg     x    365 days x 1 Mg    =  1.4 Mg
                  day               yr      1000 kg    yr-relief device

JTotal annual cost (before credit) minus recovery credit.

^Based on uncontrolled VOC emission factor and control efficiencies  for
 each control technique in footnote i.

1 Obtained by dividing net annualized cost by total VOC emission
reduction.

•"Incremental cost effectiveness =

   Net annualized cost of          _    Net annualized cost  of
   control technique	    ^	next less restrictive control
   Annual VOC emission reductionIAnnual VOC emission  reduction
   of control technique                 of next less  restrictive control
                                 A-6

-------
            Table A«3.    COMPRESSOR  SEAL  IMPACTS
                          (.May  1980 dollars)
   Per Compressor  '
 Item/Control Technique                        Closed Vent  and Seal System


 Installed Capital Cost3                               3,000

 Annualized Capital Costs
   Control Equipment11                                  1,300

 Annualized Operating Costs
A. Maintenance6
8. Miscellaneous*1
Total Annual Costs
Before Credit
Recovery Credit8
Net Annualized Costs'*
Total VOC Emission
Reduction (Mg/yr)9
Cost Effectiveness
($/Mg VOC)h
400
320
2,020
1,180
340
5.5
150
Model Unit 8:   3  compressors

  Emission Reductions:
      3 x 5.5  Mg  VOC/yr  • 16.5 Mg/yr.


•Capital  cost  from BID for proposed standards,  Table 8-1.

b0.163 capital  recovery factor x capital  cost;  from BID
 for proposed  ,-sndards. Table 8-5.

C0.05 x capital cost.  From BID for proposed  standards, Table 8.5.

do.04 x capital cost.  From BID for proposed  standards. Table 3.5.

Recovery credit  is based on uncontrolled VOC emission factor of
 15 leg/day (8IC for proposed standards, Table 3-1} and 100 percent
 control  efficiency.  Recovered product valued  at S215/Mg (from BID
 for proposed  s:andards, Table 8-5).  Recovery  credit assumes captured
 emissions are recycled to a process line or  used for process Heater
 fuel at  a similar value.

^Total  annual  cost (before credit) minus  recovery credit.

SBased on uncontrolled emission factor of 15  leg/day (BIO Table 3-1) and
 100 percent control efficiency for a closed  vent and seal system:

 15 kg/day x 365  days/yr x 1 Mg/1,000 kg  • 5.5  Mg/yr.

"Obtained by dividing net annualized cost 6y  total VOC emission reduction.
                                          A-7

-------
            Table A-4,   OPEN-ENDED  LINES  IMPACTS
                          (May  1980  dollars)
  Per Open-«nded Line
 Item/Control Technique                                  Caps


 Instal led Capital Cost*                                  53

 Annual lied Capital Costs
  Control Equipment0                                      3.5

 Annual Ized Operating Costs

  A.  Maintenance0                                        3 7
  B.  Miscellaneous*1                                      2^1

 Total Annual Costs
Before Credit
Recovery Credit*
Net Annual 1 zed Costs'
Total VOC Emission
Reduction (Mg/yr )9
Cost Effectiveness
(J/Mg VOC)n
13.4
4.3
9.1
0.020
460
Node! Unit 8:   140 open-ended lines.

  Emission Reductions:
    140 .x 0.020 Ng/VOC/yr • 2.8 Mg/yr.
  •From BIO for proposed  standards. Table 3-1.

  bO.163 (capital recovery  factor) x capital cost;  from BIO  for proposed
   standards. Table 8-5.

  CO.05 x capital cost.   From 310 for proposed standards, Table 3.5.

  dO.irt x capital cost.   From 810 for proposed standards. Table 8.5.

  •Recovery credit based  on uncontrolled VOC emission factor of
   O.GS5 leg/day (from 810 for proposed standards, Table 3-1).  Based on 100
   percent control
   efficiency for caps and  $215/Mg VOC emission reduction (from 310 for
   proposed standards. Table 8-5).

   (.mission Reduction:
   0.055 kg/day/open-ended  line  x 365 day/yr x 1 Mg/1,000 kg • 0.020 Mg/yr

   Recovery Credit:
   0.020 Mg/yr x S215/Mg  VOC - $4.3/yr/open-ended  line.

  'Total annual cost (before credit) minus recovery credit.

  9Based on uncontrolled  emission factor and control  efficiency in
   footnote e.

  ^Obtained by dividing net annualized cost by total  VOC emission reduction.
                                        A-8

-------
    Table A-5.    SAMPLING  CONNECTION  SYSTEM  IMPACTS
                          (May 1980  dollars)
Per Sampling Connection System
   Item/Control  Technique                    Closed Purge Sampling System
Installed Capital Cost*                                 530
Annuallzed Capital Costs
  Control Equipment0     •                                36
Annuallzed Operating Costs
  A.  Maintenance0                                       26
  8.  Miscellaneous4                                     21
Total Annual  Costs
  Before Credit                                         133
Recovery Credit*                                        28
Net Annuallzed Costsf                                   105
Total VOC Emission
  Reduction (Mg/yr)9                                      0.13
Cost Effectiveness
  (S/Mg VOC)n                                          810

Model Unit 3:  20 sampling connections.
  Emission Reductions:
    20 x 0.13 Mg/VOC/yr « 2.6 Mg/yr.
    *From BID for proposed standards.  Table  3-1.
    ^Capital  recovery  factor (0.163)  x capital cost; from 810 for proposed
     standards.  Table  3-5.
    CO.05 x capital  cost.  From BID for proposed standards. Table 8.5.
    d0.04 x capital  cost.  From BID for proposed standards, Table 8.5.
    Recovery credit eased on uncontrolled VOC emission factor of
     0.36 kg/day (from BID for proposed standards, Table 3-1).  Based
     on 100 percent  control efficiency.
     Recovered Emissions:
     0.36 kg/day x 365 day/yr x 1 Ng/1,000 kg » 0.13 Mg/yr
     Recovery Credit:
     $215/Mg x 0.13  Mg/yr  « $28/yr.
    ^Total annual cost (before credit) minus recovery credit.
    9Based on uncontrolled VOC emission factor and 100 percent control  as
     shown in footnote e.                                                  .
    "Obtained by dividing  net annual!zed cost by total VOC emission reduction.
                                  A-9

-------
                    Table  A-6,    VALVE  EMISSIONS AND EMISSION REDUCTIONS
EMlsslon Per Valve Emission Reduction Per Valve
(kg/day)
(Mg/yr)*
control -service Gas Light liquid lias Light liquid Weighted Averaqec
Uncontrolled 0.64 0.26
Quarterly LDR 0.262 0.098 0.14
Monthly LOR 0.192 0.072 0.16
Sealed Bellows
Valves* 0 0 0.23
LDR « Leak Detection and Repair
•Emission Reductions Per Valve Calculated as:
/Uncontrolled emissions Controlled emissions
1 per valve (kg/hr) - per valve (kg/hr)

0.059 0.087
0.069 0.10

0.095 0.14


\ x 365 days x 1 Mg
I year 1000 kg
Emission Reduction Perk
Model Unit 8 (Mg/yr
Gas Light liquid

35.9 29.6
42.5 34.3


otald

65.5
76.8

60.7 47.4 108








      \ (fro* BID Table 3-1)
                    (from BIO Table F-7)
 Example Calculation:  gas service, quarterly LOR • (0.64 - 0.262) x 365 + 1000  •  0.14 Mg/yr
       on  Model Unit B:  260 gas service  valves and 500 light liquid service  valves.
 Example Calculation: gas service, quarterly LDR - (0.64 - 0.262) x 365 * 1000 x 260 • 35.9 Mg/yr

cuelghted  average 1s based on Model  Unit  B  valve population.  Weighted average 1s calculated
 by using  the  formula:
                  260
               ___   __
               260 + W»
                gas service j
                Mission   /
                reduction
         500     x  light  liquid service)
        0 + 500     Mission reduction  J
 Example calculation:  quarterly LDR  •
   260     x  0.14 Mg/yr\ +  /   500
260 > 500             J    \250 * 500
                                                      x  0.059 Mg/yr)  «   0.087 Mg/yr
 The emission reductions reported In  the proposal preamble are on a per component basis, therefore, 1t Is necessary to
 derive per valve Impacts by weighting  the gas and light liquid service Mission reductions by their relative component
 counts In Model Unit B.  The resulting Mission reductions per valve represent a weighted average.
       Unit 8 Mission reductions presented In the proposal preamble Table  1 represent a weighted average of the gas  and
 light liquid Mission reductions, calculated as:
        Quarterly LOR
260   x  35.9 Mg/yrj
                                          500  x  29.6 Mg/yr
                                         I 760
                                            • J  •  31
                          .7 Mg/yr
        Monthly  LOR
                / 260   x  42.5 Mg/yr
                \ 750
                     H
 500  x  34.3 Mg/yr  •   37.1 Mg/yr
-7HT              /
 However,  total emissions from gas and light  liquid service valves In Model  Unit B should have been reported rather than  a
 weighted  average.   Individual valve weighted average Impacts are used to  determine the per component cost effectiveness.

 From BID  for  proposed standards, Section  4.3.3.
                                               A-10

-------
         Table  A-7.    VALVE LEAK DETECTION  AND  REPAIR  COSTS'
                                  (May  1980 dollars)
Per Valve
Cost Item/LDR-Service
Initial Leak Repa1r<=
Monitoring Labor*1
Recurring Leak Repair Labor8
Administration and Support f
Total Annuallzed Cost
Product Recovery Cred1t9
Net Annuallzed Cost"
Quarterly
Gas
0.46
2.4
3.8
2.5
9.2
(30)
(21)
Light
liquid
0.51
2.4
3.8
2.5
9.2
(13)
(«)
Wei ghted
Average''
0.49
2.4
3.3
2.5
9.2
(19)
(10)
Gas
0.46
7.1
3.9
4.4
16
(34)
(18)
Monthly
Light
liquid
0.51
7.1
3.9
4.4
16
(15)
1

Wei ghted
Average1*
0.49
7.1
3.9
4.4
16
(22)
(6)
 (XX) • Cost Savings
 LOR » Leak Detection and Repair

 'Cost of monitoring Instrument Is  not Included In this  analysis.

 bBased on Model Unit 8:  260 gas service and 500 light  liquid service valves.

       Weighted average calculated by using the formula:

                           x
          /    260
          \260 + 500
gas service]  +
cost Item  J
                                                        500      x  light  liquid service)
                                                      260 + 500          cost Item      I
CFrom BID for proposed standards.  Tables 8-3. 8-5.  and  8-6.  Calculated as:

       Initial leak frequency  x   1.13 hrs/valve  x   18/hr  x  1.4  x  0.163

dFrom 810 for proposed standards.  Tables F-4 and F-12.  Calculated as:
              Fraction of Sources Screened
              (from BID Table  F-4, 2nd
               turnaround Annual Average)
                                         x  1 mln   x   1 nr  x  U8.00  x  2 workers
                                            valve     60 mln      Fir
Example  Calculation:  Gas Valves, Monthly LDR  •  11.8  x 1/60  x 18.00  x   2  •  $7.1


 *Fron BID  for proposed standards,  Tables F-4 and F-12.  Calculated as:

                                      x
Fraction of  Sources Operated on,
from BIO Table F-4, 2nd
turnaround annual average
        1.13  hrs  x  S18.QO
         valve          hr
  Example  calculation:  gas service, quarterly LDR  *

     [(0.1762  +  0.1970)/2]  x  1.13   x  $18  *  $3.80

 ^From BID for proposed standards,  Table 8-5.
  Calculated as:  0.4  x (monitoring labor + recurring leak  repair  labor)

 ^Calculated as:  $215  x  emission reductions (given in Table  6).

 "Total  annual cost (before credit) minus recovery credit.
                                     A-n

-------
      Table A-8.  SEALED BELLOWS VALVE  COST  IMPACTS
                     (May 1980 dollars)
Per Valve Cost Item

  Capital Cost3                                 2,730
  Annualized Costb
    Capital recovery                              440
    Maintenance                                   140
    Miscellaneous                                 110
    Total annualized cost                         690
    Product recovery credit0                      (30)
    Net annualized cost                           660
(xx) » Cost Savings
a
 From BID for proposed standards,  Table 8-1.
b
 Basis for annualized costs from BID for proposed standards, Table 8-5.

 Calculated as:   $215 x emission reductions (given in Table 6)
                         A-12

-------
          Table A-9.   COST EFFECTIVENESS OF VALVE CONTROLS
                         (May 1980 dollars)
Per Valve
Item/Control
Net Annual i zed Cost3
VOC Emission Reduction
(Mg/yr)b
Cost Effectiveness
($/Mg VOC)C
Incremental Cost
Effectiveness ($/Mg VOC)d
Quarterly
LDR
(10)
0.087
(110)
(110)
Monthly
LDR
(6)
0.10
(60)
310
Sealed Bel
Valves
660
0.
4,700
16,700
lows

14


(xx) = Cost Savings
LDR = Leak Detection and Repair
 From Tables 7 and 8.

 From Table 6.
 Calculated as:
 Calculated as:
                                   Net annualized Costs ($/yr)
                             annual VOC emission reduction (Mg/yr)
                       Net annualized cost of
                          control technique
  Net annualized cost of next
   less restrictive control
                       Annual VOC emission
                       reduction of control
                       technique
  Annual VOC emission
- reduction of next less
  restrictive control
                                 A-13

-------
         Table A-10.  PUMP EMISSIONS AND  EMISSION REDUCTIONS
Control
Uncontrolled
Annual LDR
Quarterly LDR
Monthly LDR
Dual Seal System
Emissions Per
Per Pump (kg/day)
2.7b
2.12C
0.79C
0.45C
Od
Emission Reductions Per
Pump (Mg/yr)a

0.21
0.70
0.82
0.99
LDR » Leak Detection and Repair
Model Unit B = 14 pump seals in light liquid service.
  Emission reductions:
    Annual LDR           3.0 Mg/yr
    Quarterly LDR        9.8 Mg/yr
    Monthly LDR         11.5 Mg/yr
    Dual Seal System    13.9 Mg/yr
a
 Calculated as:
   Uncontrolled          Controlled
   emissions per    -   emissions per   x 365 days  x   1  Mg
     pump seal            pump seal         year      1000 kg
      (kg/hr)              (kg/hr)

 Example calculation:  annual LDR =
         (2.7 kg/day - 2.12 kg/day) x 365  *  1000  = 0.21 Mg/yr
b
 From BID for proposed standards, Table 3-1.

 From BID for proposed standards, Table F-5.
d
 From BID for proposed standards, Section 4.3.1.1.
                           A-14

-------
         Table A-ll.   PUMP  LEAK DETECTION  AND REPAIR
                   COSTS  (May 1980 dollars)a
Per Pump Seal
Cost Item/Control
Initial Repair
labor15
Replacement Seals0
Maintenance-ongoing
Replacement Seals'1
Monitoring Labor
Instrument6
Visual f
Recurring Repair Labor9
Administration and Support11
Total Annualized Cost
Product Recovery Credit1
Net Annualized CostJ

Annual
16
5.5
48
3
7.8
98
44
220
45
180
LDR
Quarterly
16
5.5
55
12
7.8
110
52
260
150
110

Monthly
16
5.5
57
36
7.8
120
66
310
180
130
 (XX)  »  Cost Savings
 LDR  »   Leak Detection and Repair

 aPump seal  repair  costs are based on 16 labor  hours per pump repair and  includes
 $140 per repair for a replacement seal.   This  analysis does not include the cost
 of monitoring instrument.

 bFrom BID for proposed standards, Tables  8-3 and 8-5.  Calculated as:

 Estimated  percent of   x   labor hours  x labor  x  Administration   x  Capital
 initial  pumps leaking          per         rate           and         recovery
   (BID Table 8-3)          seal  repair              support costs      factor

Example calculation:  annual LDR  *

       (0.24)  x  (16) x ($18) x (1.4)  x  (0.163)  -  $16

clnitial  replacement seal  cost is calculated as:

       Estimated percent of   x   $140/     x    0.163
       initial pumps leaking     replacement     Capital
         (BID  Table 8-3)              seal     recovery factor

Ingoing  replacement seal  cost calculated  as:

       Fraction of Sources    x   $140/
           operated on            replacement
       (from BID Table  F-6)          seal
                                A-15

-------
           Table A-ll.  PUMP LEAK DETECTION  AND  REPAIR COSTS
                        (May 1980 dollars)3
                           (Concluded)
  Example calculation:   annual  LDR

        (0.3397)  x ($140)   =  $48

 Calculated as:
        fraction  of sources    x     5 min
             screened             pump  seal
        (from BID Table  F-6)

  Example  calculation:   annual LDR  =

        (1)  x (5/60) x  ($18) x (2)  =   $3

 Calculated as:
                 1 hr
                60 min
        0.5 min   x  £2  x  1 hr
      pump seal     yr    60 min
   x  $18  x  1 worker
       hr
dCalculated as:
       fraction of sources
          operated on
       (from BID Table F-6)
x    labor hours
   per seal repair
 Example calculation:  annual LDR =

       (0.3397) x (16 hrs/seal repair) x ($18)
                    $98
 $18
 ~W
x  2 workers
$7.8
$18
~hT
"Administration and support  =  0.4  x (monitoring labor + rec-rrinq
 leak repair labor)

Calculated as:

       Emission Reduction  x  $215/Mg
         (from Table 10)

JTotal annualized costs (before credit) minus recovery credit.
                             A-16

-------
       Table A-12.  DUAL MECHANICAL SEAL SYSTEM COSTS  FOR PUMPS
                           (May 1980 dollars)
   Per Pump Seal
     Cost Item
 Capital  Cost3

   Seal                                                 970
   Seal  installation                                    288
   Barrier fluid  system                               1850
   Barrier fluid  degassing  vent                        4000
   Total  capital  cost                                  7110

 Annual i zed Cost'5

   Capital  recovery
                                                       560
    other capital"                                    1000
  Maintenance  charges6                                 360
  Miscellaneous  charges*7                               280
  Total annual i zed  cost'                              2200

Product Recovery Credit9                              (210)
Net Annuali zed Cost  n                                 1990
 (xx) = Cost Savings
 a
 From BID for proposed standards, Table  8-1.
 b
 From BID for proposed standards, Table  8-5.
 c
 Calculated as 0.58 x capital cost for seal.
 Capital cost for seal =  cost for new seal  ($1250)
 minus credit for old seal  ($255 x 328.9/266.6) =  $970.
 Capital recovery credit per seal = 0.58 x $970 =  $560.
 d
 Calculated as:  0.163 x capital cost.
 e
 Calculated as:  0.05 x capital cost.
 f
 Calculated as:  0.04 x capital cost.
 g
 Calculated as:
   Emission Reduction x $215/Mg.
   (from Table 10)
h
 Total  annualized cost (before credit) minus product recovery credit.
                             A-17

-------
                 Table A*13,  COST EFFECTIVENESS Of PUMP CONTROLS
                                 (Way 1980 dollars)
Per pump
Item/Control
Net Annual i zed Cost*
VOC Emission Reduction
(Mg/yr)b
Cost Effectiveness
($/Mg VOC)c
Incremental Cost
Effectiveness ($/Mg VOC)<*
Annual
LDR
180
0.21
860
860
Quarterly
LDR
110
0.70
157
(140)
Monthly
LDR
130
0.82
158
170
Dual
Seals
1,990
0.99
2,000
10,900
(xx) = Cost Savings
LDR - Leak Detection and Repair
a
 From Tables 11 and 12.
b
 From Table 10.

 Calculated as :
 Calculated as:
                                 Net annualized costs ($/yr)
                         annual VOC emission reduction (Mg/yr)
                       Net annualized cost of
                        control technique
net annualized cost of next
  less restrictive control
                       Annual VOC emission
                       reduction of control
                       technique
 Annual VOC emission
 reduction of next less
 restrictive control
                                    A-18

-------
                               APPENDIX B

           Regulatory Decisions Affecting Standards for SOCMI

    Several of the decisions made on these standards (since they were
proposed) affect EPA's position on standards of performance (Subpart
VV) for equipment leaks of VOC within the Synthetic Organic Chemical
Manufacturing Industry (SOCMI).  These decisions are the result of new
or additional analysis of the control techniques considered in the
standards for petroleum refineries and SOCMI and, therefore, should be
made consistent for these two standards.  The decisions concern:
              (1)  alternative for determining a "capital  expenditure,"
              (2)  clarification of reconstruction provisions,
              (3)  difficult-to-monitor valves in new units, and
              (4)  double block and bleed valve exemption.
The discussions of these decisions are found in Sections 2.2.3.1, 2.7,
4.2, and 5.0 of the BID for promulgated standards as they  apply for
petroleum refineries.  The basis for the revisions to Subpart VV is
consistent with these discussions.  EPA knows of no reason not to make
these revisions to Subpart VV and, therefore, based on a prudent use of
its resources, is promulgating these revisions.
                                  B-l

-------

-------
                      APPENDIX C



EVALUATION OF AVAILABLE LEAK DETECTION AND REPAIR DATA
                         C-l

-------
  Appendix C - EVALUATION OF AVAILABLE LEAK DETECTION AND REPAIR DATA

C.O  INTRODUCTION
     As a part of their comments on the proposed NSPS for VOC fugitive
emissions from petroleum refineries, two commenters (IV-D-25, IV-D-14)
presented summary data from their plants and asked EPA to review these
data.  EPA requested the raw data for these summaries and requested other
data to evaluate the NSPS in light of data from refinery leak detection
and repair programs.  EPA analyzed these data by comparing and con-
trasting, where possible, with the estimates used in preparing the BID
for the proposed standards.
     Sections C.2, C.3, and C.4 of this appendix are memoranda that
provide summaries of the data, and describe the techniques used by EPA
in analyzing the data.  All of the data received were generated either
as a result of or as a measure of the effectiveness of state or local
regulations and, accordingly, an understanding of these regulations is
needed.  Therefore, the regulations on which the data are based are
described in the memoranda.  In most cases, data were submitted that
allow direct quantitative comparison to the NSPS estimates.   Where
qualitative comparison of these data and the estimates used in the
proposal BID are made, the reasons why the data are not directly related
are presented and the uncertainties with the qualitative comparison are
discussed.
     Section C.2 presents data obtained from Texaco, U.S.A.  (IV-D-25a,
IV-D-33, IV-D-36).  These data were generated under the requirements of
the Louisiana State Implementation Plan, which requires quarterly moni-
toring of gas service components and annual monitoring of liquid service
components.  Texaco monitors additional  components to those required by
the proposed NSPS, and the data are therefore not directly comparable to
the data used to support the NSPS in many cases.  The Texaco data
memorandum includes assumptions made by EPA in analyzing the data.
     Section C.3 presents data obtained from facility inspections made
by the South Coast Air Quality Management District (SCAQMD)  and the Bay
Area Air Quality Management District (BAAQMD)  during a California Air
Resources Board (CARB) petroleum refinery valve inspection program
                                  C-2

-------
 (IV-D-31, IV-B-18).  These data,  as received by EPA,  were the  field
 data sheets from valve inspections at 12  refineries.   The data  are
 not based on inspection of entire process units,  but  only selected
 valves within selected process  units.   Again,  EPA made several  assumptions
 in analyzing the data, and these  assumptions are  stated in  the  memorandum.
      Additional  data were  available from  a  study  of the effectiveness of
 the South Coast Air Quality Management District (SCAQMD)  Rule 466.1,
 performed by the EPA Office of  Research and Development (ORD).  The
 results of this study are  summarized  and  evaluated in  Section C.4.
 C.I  DATA SUMMARY
      Data are available from the  memoranda  in  Sections  C.2  through C.4
 on initial  leak  frequency,  leak occurrence  rates,  small  valves, repair
 effectiveness,  program costs, and monitoring time.  This  section
 summarizes  these data  and  provides  comparisons  with the  values  used
 by EPA  in estimating  the impacts  of the NSPS.   The data  discussed
 in this  section have  been arranged  by  topic rather than  data source to
 provide  for  ease in  locating  specific  information.
 C.I.I Initial Leak Frequency
      Information on  the initial  leak frequencies at a  few facilities
 can be obtained, by making  assumptions, from the data supplied by Texaco
 (C.2) and the EPA analysis  of the SCAQMD data  (C.4).   These data are
 shown in Table C-l.
     For the Texaco data, initial  leak frequency for each process unit
 was derived by assuming that the first period  for which leak monitoring
 data was reported was  indeed the first time the unit was monitored.
 Therefore, the percent of components found leaking at the first monitoring
 period is the initial  leak   frequency.  The initial leak frequency for
 gas components varied  from  0.0 percent to  14.8 percent, with a weighted
 average value of 6.5 percent.  For liquid  service components, the initial
 leak frequency varied  from  0.0 percent to  17.0 percent, with a weighted
 average value of 2.8 percent.  It should be noted that Texaco screens
components not included in  EPA's NSPS leak detection  and repair program,
 such as valve flanges and valve  bonnets.  Since these sources are normally
considered by EPA to have low leak frequencies, and they may represent
                                  C-3

-------
a significant portion of the total  number of components, EPA expects the
initial leak frequency determined from the Texaco data to be understated
compared to data based solely on testing of NSPS sources.
     The EPA analysis of the SCAQMD data determined initial  leak frequencies
for five process units.  These data show an initial leak frequency varying
from 1.5 percent to 14.5 percent with an average value of 6.2 percent.
     Since the CARB inspection data presented in Section C.3 is from
inspection screening of facilities which have been performing leak
detection and repair routinely, no initial leak frequencies can be
generated.
     In estimating the emission reductions for the refinery NSPS, EPA
used an initial leak frequency of 10.5 percent.1  While the Texaco data
shows initial leak frequencies of 6.5 percent for gas service and
2.8 percent for liquid service components, the data can not be expected
to be comparable to the EPA estimate due to the inclusion of infrequently
leaking components as discussed above.  The initial leak frequency of
6.2 percent found in the SCAQMD study is the result of a valve popu-
lation identical to the EPA estimate basis, and is the result of 7,263
valve screenings.  As the 6.2 percent initial  leak frequency calculated
indicates that EPA may have overstated the initial leak frequency for
these plants, the LDAR model was run again using the 6.2 percent value.
This run, which included other deviations from the original  EPA estimates,
is discussed in detail in Section C.4, and showed that leak detection
and repair was a cost-effective control  technique even with the lowered
initial leak frequency.
3.1.2    Leak Occurrence Rate
     Leak occurrence rates may be calculated for all  three data sets.
Table C-2 provides a summary of the average monthly occurrence rates
for all data provided in the three memoranda.
     Several factors must be considered when comparing these data with
other information on leak occurrence rates.  For the data provided by
Texaco, the occurrence rates stated are for all  components screened,
which, as mentioned earlier, include a significant quantity  of low leak
frequency components.  Therefore, the Texaco occurrence rates  are
probably understated significantly.   The data presented for  the CARB
inspections represent valves only.   However,  CARB inspections  are
performed by monitoring component leaks at a distance of 1 centimeter

                                  C-4

-------
 from  the  component  surface, rather than at the surface as required by
 Method  21.   Since some  leaks causing an instrument reading less than
 10,000  ppm  organics  at  1 cm would likely read greater than 10,000 ppm
 at  the  surface, these occurrence rates are likely to be understated.
 In  addition, the data from Texaco was presented by quarters and for
 the purpose  of estimating occurrence rates, it was assumed that
 monitoring  occurred  on  the first day of each quarter although monitoring
 may have  actually occurred at any time during a 3-month period.  Again,
 this  discrepancy could  cause understatement or overstatement of the
 occurrence rates.
      In estimating the  impacts of a leak detection and repair program,
 EPA used  a leak occurrence rate of 1.27 percent/month, as discussed
 in  Section C.4.  As  shown in Table C-2, the occurrence rates determined
 in  the  data memoranda varied from 0.05 to 0.6 percent/month for valves.
 The occurrence rates developed from the Texaco data are not compared
 to  EPA estimates as  they include measurement of sources that generally
 have  a very low leak frequency.   Although these occurrence rates
 indicate  that EPA may have overestimated occurrence rates in the
 impacts analysis, it should be noted that EPA analyzed and consequently
 provided  alternative standards for valves where low leak  occurrence is
 found.  Additionally, as discussed in Section C.4, EPA re-estimated the
 impacts of a leak detection and repair program based on a lower (0.6
 percent/month)  leak occurrence rate,  and found the leak detection and
 repair program to still  be cost effective.
 C.I.3  Small Valves
     The  data provided by Texaco can be used to derive information
on the leak characteristics of small  valves.   Table C-3 provides a
listing of the leak incidences for small  valves for line  sizes less
than or equal to 1  1/2 inches.  For the three monitoring  periods for which
Texaco provided data, small  valves accounted for 48 percent,  49 percent,
and 32 percent of all valve leaks found,  or an average for the three
monitoring quarters  of 45 percent of  all  valve leaks.   Since  it has
been shown that valve size is  relatively  unrelated to  the mass emission
rate from a leaking  valve,2 the  small  (<1  1/2")  valves in the Texaco
                                  C-5

-------
refinery accounted for nearly one half of the valve leak  emissions.
Texaco did not provide small  valve and large valve  equipment  counts,
and therefore, the fraction of sources leaking could not  be determined.
     Data can also be obtained on the difficulty of performing  on-line
repairs on small valves.  As described above, nearly one  half of the
leaking valves were 1 1/2" sizes or smaller.  Table C-3 presents a
summary of components in the Texaco program which required off line
repair.  As shown in the table, 68 small valves (£2" line size) and 107
large valves (>2" line size) required off-line repair.  Therefore, it
appears that small valves are as repairable on-line as their larger
counterparts.  It should be noted that Texaco does  not attempt to bypass
components for off-line repair prior to unit shutdown, as will  be
required by the NSPS, but allows the leakers that are not repaired
in two attempts at simple maintenance in service to continue to leak
until a process unit shutdown occurs.  Therefore, some of the components
which were delayed until shutdown for repair in the Texaco program
would be repaired under NSPS by bypassing the leaking valve for off-
line  repair.   It  should also be noted that some of the small valves
listed in the table as requiring off-line repair were listed as large
valves by Texaco  due to the cutoff difference of 1 1/2" by Texaco and
2" by EPA.  Therefore, the table actually shows a disproportionate
amount of small valves requiring off-line repair.
C.I.4  Repair  Effectiveness
      Information  can also be obtained from the Texaco data on the
various aspects of repair effectiveness.  Table 2 of memorandum C-2
provides data  summarized from Texaco's  listing of components for which
repair was delayed (as explained in C.I.3).  As can be seen, 7.4 to
26.5  percent  of all  leaks were delayed  until turnaround for repair during
the first five monitoring periods.  For the sixth and seventh monitoring
periods, 42.5  and 91.0 percent of all repairs were listed for turnaround,
apparently because most of the leaks occurred in process units  for
which  turnarounds were scheduled within the monitoring quarter.  For
the five quarters within which shutdowns of major units were not
scheduled, the weighted average  percentage of repairs which required
shut  down was  16.0 percent, while the weighted average percentage of
repairs  requiring shutdown  for all quarters was 24.7  percent.   As such,
                                  C-6

-------
  it would appear that  simple,  on-line  repair was approximately 75.3 percent
  effective in  the Texaco  program.   In  estimating the annual repair labor
  cost1,  EPA assumed  75 percent of all  valves can be repaired with simple
  on-line maintenance,  and 25 percent of all valves require off-line repair.
  This  basis  seems  to agree very well with Texaco's experience.
  C.I.5   Program  Costs
      Texaco provided  data on the cost of operating a Control  Techniques
  Guideline  (CTG)3  based leak detection and repair program for  a 1-year
  period.  The program  underway at Texaco is required by the Louisiana
  State Implementation  Plan,  and includes quarterly leak detection and
  repair  of gas service components and annual  leak detection and repair
  of liquid service components.   As  such, the  program is similar to
  Regulatory Alternative II in the BID for proposed  standards.   Section
 C.2 provides a comparison of the costs provided  by Texaco with the cost
 estimates provided in  the BID  for  proposed standards.   Due to differences
 in costing techniques, the costs cannot be directly compared.   However,
 by adjusting both the  Texaco and EPA costs slightly, as shown  in the
 memo,  some comparisons between the  two costs can be made.
     For example,  Texaco  reported a monitoring labor cost  (in  1982
 dollars) of $72,215/year  for the entire  refinery.  Since EPA  costs are
 estimated on a model unit basis, the Texaco refinery was broken  down
 into model  units as  shown in the memo,  and the EPA monitoring  labor
 costs were  totaled for the resulting model units.  The  resulting EPA
 labor  cost  for monitoring was approximately 50 percent  lower than the
 Texaco monitoring  costs.  This is expected, however, as EPA labor
 cost estimates normally have other  costs added to them to determine the
 total monitoring costs, such as recovery of the monitoring instrument
 capital  costs and  instrument calibration/maintenance costs.  Since
 Texaco's monitoring program was performed by  a contractor, the monitoring
 cost reported by Texaco would include these additional  costs.
     Texaco reported an annual  repair cost (in 1982 dollars) of
 $5,301/year.  While EPA estimates are normally much higher than this
 figure, Texaco's  costs  do  not include any off-line  repairs, as do the
 EPA repair cost estimates.  Therefore,  the EPA estimate for these costs
were made by assuming a 10-minute repair time  for the simple on-line
                                  C-7

-------
repairs attempted by Texaco,  and  calculating  the  EPA  cost  estimate  on
the number of leaks encountered  by Texaco.  Texaco's  reported  repair
cost of $5,301/year results in a  unit cost  (for 630 repairs) of  $8.41
per component repair.  Using  the  EPA estimate of  10 minutes at $18  per
hour for simple repairs, with 40  percent overhead,  results in  a  unit
cost of $4.20 per repair.  Again, the EPA estimate  is about one-half
the Texaco cost.  Texaco made two repair attempts where  necessary,
which is not accounted for the EPA estimate.   EPA normally assumes  that
25 percent of all valves require off-line repair  at 4 hours  per  valve,
and 16 hours of repair labor  is  required for  every  pump  seal.  Hence,
the EPA estimate used for comparison with Texaco's  repair costs  is
significantly lower than the  costs presented  in the BID  for the  proposed
standards.
     The "overhead" cost reported by Texaco ($57,922) included the
costs of tagging the components  and setting up the  monitoring  program.
Obviously, this cost is not recurring and, as such, is amortized over a
10-year period by EPA.  As mentioned above, EPA estimates normally
amortize non-recurring costs, including the capital costs of  the monitoring
instruments which were included in the "monitoring  costs" by Texaco.
As  such, the EPA estimate in this case is significantly  higher than the
Texaco cost and corrects for the lower monitoring cost in the  EPA
estimate.
     As shown  in Table 4 of Section C.2, the Texaco total program costs
were close to the EPA cost estimates for a similar  program ($61,000 EPA
vs. $71,000 Texaco  after adjustment to reach the same cost basis),
especially considering the differences in costing techniques  employed.
C.I.6  Monitoring Time Data
     The  data  provided on the California Air Resources Board  (CARB)
refinery  valve inspection program included the "time monitored"  for
each component.  The CARB data as presented  in Section C.3 includes
data from inspections in 12 refineries with  a total of 93 process
units, or 6,497  components.   For these process units, monitoring time
varied from  0.61 to 1.1 minutes/valve, with  a weighted average value of
0.9 minutes/valve.   In  the BID for  proposed  standards, EPA estimated
1  minute/valve,  which is nearly  identical to that found in the CARB
inspections.
                                  C-8

-------
C.I.7  References for Section C.I

     1.  VOC Fugitive Emissions in Petroleum Refining Industry -  Background
         Information for Proposed Standards, EPA 450/3-81-015a, November
         1982.  Docket Item Number II-B-1.*

     2.  Memo, T.L. Norwood, PES, Inc., to Docket A-80-44,  Small  Valve
         Repair Cost Effectiveness.  September 26, 1983.  Docket  Item
         Number II-B-8.*

     3.  "Control of Volatile Organic Compound Leaks from Petroleum
         Refinery Equipment"  EPA450/2-78-036.  June 1978.   Docket
         Item Number IV-A-6.*

     4.  Assessment of Atmospheric Emissions from Petroleum Refining:
         Volume 3 Appendix B.  EPA-600/2-80-075c, April  1980.   Docket
         Item Number II-A-19.*

     5.  Fugitive Emission Sources of Organic Compounds  - Additional
         Information on Emissions, Emission Reductions,  and Costs.
         EPA-450/3-82-010, April  1982.  Docket Item Number  II-A-41.

*Document numbers refer to entries in Docket A-80-44,  which can be
 found at the U.S. Environmental  Protection Agency Library, Waterside
 Mall, Washington, D.C.
                                  C-9

-------
                        Table C-l.  COMPARISON OF NEW INITIAL LEAK FREQUENCY DATA TO INITIAL
                                        LEAK FREQUENCY DATA USED IN NSPS DEVELOPMENT
o
I
Source
Texaco^
SCAQMD6
Refinery^
Assessment
Serivce
Gas
Liquid
Gas and
Liquid
Gas
Liquid
Component
.Type
All
All
Valves
Valves
Valves
Number of
Sources Screened
4,736
10,082
7,263
570
995


Range(%)a
0.0 -
0.0 -
1.5 -
0.0 -
0.0 -
14.8
17.0
14.5
27.3
19.2
Initial Leak
Average 1
6.5
2.8
6.2
10
11
Frequency
;%)k 95% C.I
0.0 -
0.0 -
0.0 -
3.2 -
5.0 -

. (%)<-
15.5
12.8
15.7
19.8
17.0
a.  Range of leak frequencies found for individual  process  units  in  the  studies.
    #screened) X 100%.

b.  Weighted average leak frequency, calculated  as:
Leak frequency = (#leaks/
                   AVG. =
                           i=0
                                  N
             Where n-j  =  number of components in the ith process unit
                   Pi  =  percent leaking in the ith process unit
                   N   =  number of process units

    c.  95 percent confidence interval for average, calculated as:

                   95% C.I. = AVG. +_ 2 S.D.
                   where S.D. = Standard Deviation
                           i=0
                                  (AVG. -
    e,

    f.
    From Section C.2

    From Section C.4

    From Reference 5.  Basis for NSPS analyses.

-------
                    Table C-2.   AVERAGE LEAK  OCCURRENCE RATES
Refinery
Texaco; Convent, La.
Texaco; Convent, La.
Tosco; Bakersfield, Ca.
Exxon; Benicia, Ca.
Chevron; Richmond, Ca.
ARCO; Carson, Ca.
Mobil ; Torrance, Ca.
Fletcher Oil; Carson, Ca.
Champ! in Oil; Wilmington, Ca.
Shell Oil; Carson, Ca.
Chevron; El Segundo, Ca.
Newhall; Newhall, Ca.
Power! ne; Sante Fe Springs, Ca.
SCAQMD Summary
Proposal BID Basis
Component Type
All
All
Valves
Valves
Valves
Valves
Valves
Valves
Valves
Valves
Valves
Valves
Valves
Valves
Valves
Service
Liquid
Gas
Gas and Liquid
Gas and Liquid
Gas and Liquid
Gas and Liquid
Gas and Liquid
Gas and Liquid
Gas and Liquid
Gas and Liquid
Gas and Liquid
Gas and Liquid
Gas and Liquid
Gas and Liquid
Gas and Liquid
No. of
Sources
10,082
4,736
326
858
803
591
338
317
157
816
602
152
576
7,263
—
Occurrence
Rate3
(%/month)
0.08
0.6
0.05C
0.6C
O.ic
0.2C
0.3=
0.4C
0.2C
0.2C
0.2=
O.ic
0.4C
0.6
-------
                 Table C-3.   SMALL  VALVE  DATA  SUMMARY
Monitoring
Quarter
4Q 1981
1Q 1982
2Q 1982
3Q 1982
1Q 1983
2Q 1983
3Q 1983
Small Valvesb
Requiring
Small Valve3 Large Valve3 Off Line
Leaks Detected Leaks Detected Repair
5
15
93 102 11
51 53 4
2
24 50 11
20
Large Valvesb
Requiring
Off Line
Repai r
10
14
21
2
7
19
34
a
 From Section C.2 Table 3;  small  valves  are  those valves less than or
 equal  to 1 1/2".  "-" indicates  that no data was provided by Texaco.

b
 Small  valves defined as less than or equal  to 2".  From Section C.2,
 Table 2.  Therefore, the "small  valve"  listings in these columns would
 include some of the "large valves"  from the "leaks detected" listings
 (those valves greater than 1 1/2" and less  than or equal to 2".
                                  C-12

-------
         Section C.2



TEXACO DATA SUBMITTAL SUMMARY
             C-13

-------
                                                          A-80-44

                          MEMORANDUM             IV-B-22

                                            DATE:   November 11, 1983

TO:       Docket A-80-44

FROM:     T.L. Norwood, P.E., Pacific Environmental  Services,  Inc.  /tj

SUBJECT:  Review and Summary of Leak Detection and Repair Program
          Data Supplied by Texaco, U.S.A.



Introduction

      This memorandum summarizes the data supplied by Texaco,  U.S.A as
comments on the proposed new source performance standards for  fugitive
VOC emissions from petroleum refineries1 and in response to EPA requests
for additional information to clarify their original submittal.2»3
This memorandum summarizes only those data supplied by Texaco  that were
sufficiently detailed to allow comparison with estimates provided by
EPA in the background information document (BID) for the proposed
standards.4  The data supplied by Texaco were not  supported by listings of
component types for the process units (i.e., Texaco specified  how many
components were in each unit but did not break the unit totals  into types
of components).  Some of the Texaco data, however, were sufficiently
detailed to allow analysis.

Data Assumptions and Deviations with NSPS Programs

      The Texaco facility for which the data were  generated is  subject
to the State Implementation Plan for the State of  Louisiana, which is
based on the refinery CTG5 requirements.  The leak detection and repair
program underway at the Texaco facility is, therefore, based on quarterly
leak detection and repair of gas service components, annual  leak
detection and repair of liquid service components, and weekly  visual
inspection of pump seals.  This program corresponds roughly to Regulatory
Alternative II in the BID for proposed standards.   There are,  however,
differences in the Texaco program and the BID Regulatory Alternative II
program, as follows:

      o   The Texaco program includes screening of certain flanges,
          capped lines and other components not required in the
          NSPS alternatives.  Since EPA believes these components have
          very low leak frequencies, their inclusion in a leak  detection
          program would lower the overall leak frequency from  that
          expected for a normal NSPS component mix.   These "non-leakers"
          could represent a very significant portion of the total
          components monitored.
                            C-14

-------
        •   The Louisiana SIP defines leaking  components  as  those  components
            with surface organic  concentrations  of greater than  10,000 ppm
            while the refinery NSPS  defines  leaks  as  greater than  or       '
            equal  to 10,000 ppm surface  organic  concentrations.  Those
            sources reading 10,000 ppm are thus  considered leaks by the
            NSPS and not considered  leaks by Texaco.  Although this
            appears to be a minor difference,  operators monitoring
            components may record the monitoring instrument  readings in
            rounded numbers,  with sources reading  across a range being
            recorded as  "10,000"  ppm.  These sources would be considered
            non-leakers  by Texaco, reducing  the measured leak incidence rate.

       •    Texaco  uses  OVA®  leak detectors,  which are calibrated
            with hexane  at approximately 5,000 ppm.  Method 21, as used
            in the  refinery NSPS,  currently specifies calibration with
            methane or hexane at approximately 10,000 ppm.  The difference
            in calibration technique may result in small  differences in
            the leak readings (and therefore  the number of leaks)
            However, EPA feels that the differences in leak readings
           caused  by the calibration differences should be small  if
           any.                                                 '

 Data Summary

       Using the assumptions described in  the  footnotes  to the  tables
 EPA was  able to summarize the data  collected  by Texaco  for  comparison
 with EPA estimates for several leak detection and repair program
 parameters.  These data must be  used judicially,  with careful attention
 given to the assumptions and program specifics  stated.   These data are
 presented in Tables 1 through  4, as follows:

       Table 1  -  Leak  frequency and  occurrence rate data

       Table 2  -  Delay of repair  data

       Table 3  - Small  valve  vs.  large valve leak  count data

       Table 4 - Program annual costs.

 References

 1.  Letter, J.M. McCrum, Texaco USA; to Central Docket Section  EPA-
    Comments on Proposed NSPS for Fugitive VOC Emissions from Petroleum
    Refineries.  April 22, 1983.   Docket Item Number IV-D-25a.

 2.  Letter, J.J. Lennox, Texaco,  USA to R.E. Rosensteel,  U.S. EPA-
    September 2, 1983.  Docket Item Number IV-D-33.*

3.  Letter, J.J. Lennox, Texaco,  USA to R.E. Rosensteel,  U.S. EPA-
    October 14, 1983.  Docket Item  Number  IV-D-36.*

4.  BID for proposed standards.   Docket Item Number III-B-1.*
                                  C-15

-------
5.  "Control  of Volatile Organic Compound Leaks  from  Petroleum Refinery
    Equipment."  EPA-450/2-78-036.   June 1978.   Docket  Item Number
    II-A-6.
*Docket items refer to Docket Number A-80-44  in  the EPA Central Docket
 Section, Waterside Mall,  Washington, D.C.
                            C-16

-------
Table 1.   ESTIMATED RATE OF LEAK OCCURRENCE FOR TEXACO DATA - GAS  SERVICE  COMPONENTS3
Process Unit Date h Percent
Number of Screened0 Leaking
Components
viyvw/tou 'V l-°
ion l/n /.i
V?l /<<3
7/n -*
y>3 o.'*
Yn 0.7
V« 1.3
**« /*>/ /.*
7/f I/to *•«
*/»! ft«
7/n 0.*?
//M /•»
¥/?J <5Jf
7/73 °-?^
HTH-1 /ff/j| 49
6ol '/H At0
Vsi. /,5"
7/8X /.i
I/O /.«
^3 °-3
y»j i>.r
Months
Between
Tests
-
3
3
-
6
3
3
-
3
3
3
£
3
•3
—
3
3
3
6
3
3
Test-to-test Cumulative Percent
30- day Leaking from A
Occurrence1* Beginning of Test0
-
o.to 1.1
0.13 2 2
-
O.I51 o.l
0-2.3 |.6
o.60 3.4-
_
*-'1 0.56
o.n /./i
o.ZJ a<;

-------
Table 1.  ESTIMATED RATE OF LEAK OCCURRENCE FOR TEXACO DATA
            GAS SERVICE COMPONENTS (continued)
Process Unit
Number of
Components
CRU






ALXr
2*3




HTU-i
1*4




AUI
Iff?





Date b
Screened
10/d
i/n
1/11
7/fi
'/H
»/»
7/U
/$/?»
i/fc
(r/*I
7/Ji
J/W
•*/«
i/9 2
i Q /& I
u/M
7/fc
1/93
f/w
7/tt
/'/•I
I/U
v»
7/fl
I/M
f/o
7/X3
Percent
Leaking
JO.l
8.9
i.f
A.1
/.3.
0,
1.0
/.r
i
o-o
9.0
0.0
0,0
0,(,
o.l,
0.0
Months Test-to-test Cumulative Percent Months from 30-day
Between 30- day Leaking from d Beginning of Occurrence from a
Tests Occurrence1" Beginning of Test0 Test Beginning of Test6
. - -
3 3.0 *.S 3 3'°
3 O.I M.3 i /.9
3 /•" 14.^ 9 /^
i *.-* /4.« /^ /./
3 o* /4.9 /j (J.9¥
3 o.l /"*.3 ^/ o.^f
- - -
3 a.o 6.» 3 &.o*
3 /.f H.i 3 /.4-
6 o.fe ?.o 9 o.3?
s i.f 15.1 12. i.n
3 o.1? n.ff if /.I7*

3 /./ 3.3 3 /.;/"
_ - -
- '
3 0-6 /.? 3 0.^
- _ -
• .
^ 0.0 o.o 3 0,00
•3 0,0 0,0 £ ,jg00
3 «•" 
-------

6
Table 1. ESTIMATED RATE OF LEAK OCCURRENCE FOR

Process Unit Date b
Number of Screened
Components
T&-TU /0/?l
&f ~ i/n
7/«
'/W
7/U '
7/83
ETU/CoQ /0/fl
¥/*Z
7/fc
l/?3
tfa
7fa
• LOG. 10/91
<+t l/fe.
Lf/?z
7/fc
//S3
7/M
TfCG- /*/?/
a1*© . //«.
¥/K
7/fi
//W
7AJ
GAS SERVICE COMPONENTS (continued)
Percent Months Test-to-test Cumulative Percent
Leaking Between 30- day Leaking from
Tests Occurrence Beginning of Test
0.0 3 °-° °-°
o.o 3 o.o o.o
o.o 3 o.o o.o
0.0 3 0.0 °'°
o.o 2 o.o °-°
0,0
o.fo 3 0 ,w o,$o
l.o 3 0.33 /.r
o.o 3 o.o /.r
o.o 6 o.o i.s
os 3 o.n l.o
0.0 3 0.0 2.0
^
o.o 3 0,0 o.o
o.o 3 0.0 0,0
-fr - -
5,3 K 3 c.2.1 O.H ,
63 6 /./ 7,/3
S.3 3 ^.2 /r.if
7.7 3 O.r? I7.J

TEXACO DATA

Months from 30-day
Beginning of Occurrence from
Test Beginning of Test
3 O.O*
3 0.0
7 o.o
JZ o.o
IS o.o
3 O.il
< 0.iS
1 0.17
is- 0,10
/? o.i i
11 0.01 f
-
3 O.O
(, o.o5'
-
6 0.38

-------
               Table 1.   ESTIMATED  RATE  OF LEAK OCCURRENCE FOR  TEXACO  DATA
                                GAS SERVICE COMPONENTS  (continued)
Process Unit    Date  b   Percent     Months     Test-to-test  Cumulative Percent    Months from         30-day
 Number of    Screened    Leaking     Between     30- day       Leaking from  d    Beginning of   Occurrence from
                                           Occurrence1-  Beginning of Testa      Test      Beginning of Test6
Components
             |o/f|
              I/?Z
i.l
 -»
5". 3
t.«f
0.0
3
3-
0.70
0.70

O.w
2.1
0.0
                                                             J.3
                                                            11.7
                                                            11.7
                                                                                           0.70
                                                                                          1.30
                                                                           li
                                           C-20

-------
         Table  1.   ESTIMATED RATE OF  LEAK  OCCURRENCE  FOR TEXACO  DATA  -  LIQUID SERVICE
Process Unit    Date  h   Percent
 Number of    Screened    Leaking
 Components
                                   Months     Test-to-test  Cumulative Percent
                                   Between      30- day       Leaking from  H
                                              Occurrence.  Beginning of Test
                                  Tests
              Months from        30-day
              Beginning of   Occurrence from
                 Test      Beginning of Test
VPS/VBU/C.OU
7/13       0.10
                                    /2.
                                              0.06
0,70
                                                                                                      f
   PCCU
              f/13
                       <7.J/
   HTU-1
           1%

           7/J3-
   CRU

   fOl
           \o/n
   ALKY
                                                                f.f
  WTU-2.
    4 IS"
                        0.0
                                            0.0
                                                               0.0
  TCTU    M./W.       -I-
                                              0.0
                                                                                                     -r
                                          C-21

-------
                        Table  1.   ESTIMATED RATE OF  LEAK  OCCURRENCE  FOR TEXACO  DATA

                                           LIQUID SERVICE  (concluded)
Process Unit    Date  h   Percent
 Number of   Screened    Leaking
 Components
                              Months    Test-to-test  Cumulative Percent    Months from         30-day
                              Between     30- day       Leaking from H    Beginning of   Occurrence from
                               Tests     Occurrence   Beginning of Test       Test      Beginning of Test
em/cos    1/91        o.o
   91         1/93        0.0
                                                0.0
                                                            0.0
                                                                                 o.
LOG-

2.0.1
7/12.

I/?3
                         -f
                                                0.3
                                                             /.?
   TK&      H/tt
   .pw
                                                0.0
                                                            0.0
                                          C-22

-------
                                    10
                          Footnotes  to  Table  1


 aThe number of gas and liquid  components  during the monitoring periods
  October 1981  through  September 1983 are  assumed to be constant and
  are based on  the  component  summary in Table 1 of Docket Number A-80-
  44-IV-D-25a (page 8).

 bThe date screened is  based  on the assumption that components were
  monitoring the first  month  of the quarterly monitoring period. In
  reality, the  components  could have been monitored at any time during
  the quarter.

 cBased  on all  leaks  from  the previous  inspection being repaired and
  an  assumption of  linear  leak  occurrence.  The leak frequency (percent
  leaking) divided  by the  number of months between tests estimates the
  30-day  leak occurrence rate.

 dBased  on all  leaks  at initial  inspection being repaired and assuming
  that if the other inspection  had not  occurred, the leaks could have
  accumulated from  inspection to inspection (leaks found are new leaks
  at  each  inspection).

 eSame methodology  as discussed  in footnote b, except based on the
  initial  inspection.

 'These are  the  overall  average monthly occurrence rates for a continuous
  series  of  screenings.

 9These data could  not be determined because the breakdown of gas  and
  liquid components that were found leaking was not available.
  Therefore, this screening date begins  a  new monitoring period for
  the purpose of calculating the cumulative leak  occurrence rate.

 "Unit not in operation  this quarter.

 ""Unit was shut down or  partially shut  down.

JNo liquid components monitored in 1983.
                                 C-23

-------
                                              Table 2.   TEXACO DELAY OF REPAIR  DATA  SUMMARY*
o
 I
QUARTER
10/81
1/82 -
4/82 -
7/82 -
1/83 -
4/83 -
7/83 -
- 12/81
3/82
6/82
9/82c
3/83e
6/83
9/839
TOTALS
NUMBER OF
LEAKS DETECTED
256
132
168
74
139
87
78
934

NUMBER
19
35
40
13
16
37
71
231
REPAIRS SCHEDULED FOR TURNAROUND
PERCENT j
7.4
26.5
23.8
17.6
11.5
42.5
91.0
24.7
1 SMALL VALVE5D-
5
15
11
4
2
11
20
68
I LAftGE VALVES^
10
14
21
2
7
19
34
107
TURNAROUND REPAIRS
NUMBER
17
22
35
11
4
8
-_f
85"
PERCENTd
90
63
88
85
25"
22"
-_f
79n
a ~ — 	
                 Includes all process units and gas and liquid  service components.
                D
                 Small valve Is defined as  2 Inches or less.
                c
                 Large valve Is defined as  greater than 2 Inches.
                d
                ^Percent repair effectiveness; number of turnaround repairs divided  by number of repairs scheduled for turnaround.

                 New leaks for this quarter were determined by  examining leak data for previous quarters.
                 Recurring leaks are considered to be new leaks for the purposes  of  this analysis.

                 Cannot be determined from  data
                9

                 Monitoring occurred during  shutdown of ARO. TGTU. and HTU-2 (which  contain 2.142 components).

-------
                              Table 3.   NUMBER OF  LEAKS  FOUND BY VALVE SIZE GROUP
Period
UNIT
VPS
FCCU
HTUI
CRU
ALKY
HTU-2
ARU
TGTU
ETU/COB
TK CAR/
TRUCK/
DOCK
TANKAGE
FLARE/
ADD.
P/W
TOTAL
PLANT
2nd Quarter
1982
Valves
1-1/2"
and Under
No. I*
5 71
2 50
15 41
53 57
6 18
8 67
0
1 50
0
1 50
1 33
1 100
93 48
Valves
Over
1-1/2"
No. X»
2 29
2 50
22 59
40 43
28 . 82
4 33
0
1 50
0
1 50
2 67
0
102 52
3rd Quarter
1982
Valves Valves
1-1/2" Over
and Under 1-1/2"
No. *« No. %»
3 27 8 73
2 67 1 33
5 63 3 37
21 64 12 36
0 ~ 12 100
17 63 10 37
0 —o
0 0
0 0
0 —o
3 43 4 57
0 2 100
51 49 53 51
1st Quarter
1983
Valves Valves
1-1/2" Over
and Under 1-1/2"
No. S» No. %»
2 15 11 85
4 57 3 43
5 71 2 29
4 31 9 69
3 27 8 73
0 —o
0 -o
0 0
1 100 0
1 10 9 * 90
1 50 1 50
4 36 7 64
24 32 50 68
a
X values Indicate percentage of all valves leaking In that quarter.

-------
                                        13
             Table 4.   COMPARISON  OF  REPORTED TEXACO  LEAK DETECTION
                      AND  REPAIR PROGRAM COSTS WITH EPA COST ESTIMATES
      ITEM
    TEXACO
(1982  dollars)1
    TEXACO
(1980 dollars)2
 EPA ESTIMATE
(1980 dollars)
 Monitoring Labor
   Repair Labor
  Overhead/Setup

Total Program Costs
  $ 72,215/year
     5,301
    57.922

  $135,488/year
   $59,060/year
     4,335c
     7.7205

   $71,115/year
$30,740/year3
  2,646/year4
 27,500/year6

$60,886/year
 From Reference.

"1980 dollars calculated using Chemical  Engineering Cost Indices:
       December 1982 * 316.1; May 1980 = 258.5; Ratio = 0.8178

 EPA Monitoring labor based on 4 each of Model Units A, B, and C
 per Table 4a, with Proposal BID Table F-12 labor estimates of
 33.7 hours/year (A). 68.8 hours/year (B); and 202 hours/year (C).
 Labor at $18/hour, with 40% overhead added.

             $/year = (33.7 + 68.8 + 202) hours x 4 units x $18/hour x 1.4

                    = $30,740

*EPA estimate based on 630 leaks detected in first 4 quarters (Table 2)
 by Texaco, at 10 minutes for each repair attempt.  Labor at $18/hour
 with 40% added overhead.

             $/year = 630 x  18 x 1.4 x 10/60 = $2,646/year

 EPA estimate based on amortizing initial setup costs for 10 years at
 10% interest (capital cost  x 0.163) and deflating to 1980 dollars
 per footnote 2 method.

 EPA estimate based costs of operating and maintaining 5 pairs of
 monitoring  instruments  (at  $5,500/year  each  per BID for proposed
 standards Table 8-9) =  $27,500/year.
                                     C-26

-------
                                Table 4a
                 FUGITIVE EMISSIONS COMPONENTS SUMMARY9
UNIT
Vacuum Pipe Still, Visbreaking
Unit & Gas Oil Unit (VPS)
Fluid Catalytic Cracking
Unit (FCCU)
Hydrotreating Unit #1
(HTU-1)
Catalytic Reforming Unit
(CRU)
Alkylation Unit
(ALKY)
Hydrotreating Unit #2
(HTU-2)
Amine Regeneration Unit
(ARU)
Tail Gas Treating Unit
(TGTU)
Effluent Treating Unit & Co
Boilers (ETU/COB)
Tank Car & Truck Loading
& Dock (LOG)
Tankage, Flares & Additives
(TKG)
Pipeways
(PW)
TOTAL
No,
Gas
1016
715
607
450
263
926
157
24
200
44
240
94
4736
. of Components
Liquid
2440
1934
951
401
1203
542
425
68
81
221
1688
128
10082
Total
3456
2649
1558
851
1466
1468
582
92
281
265
1928
222
14818
EPA
Model
Plant
C
C
C
B
B
B
B
A
A
A
C
A

a.
This Table is presented by Texaco as Table 1 in Reference 1.
Model Plant designations added by EPA are based on the similarity
of these model plants to the number of pieces of equipment
shown for the Texaco units.


                               C-27

-------
                 SECTION C.3

SUMMARY OF AVAILABLE CALIFORNIA AIR RESOURCES
            BOARD INSPECTION DATA
                     C-28

-------
                                                          A-80-44

                            MEMORANDUM           IV-B-18

                                               DATE:  November 11, 1983
  TO:       Docket A-80-44

  FROM:
          T.L. Norwood, P.E., Pacific Environmental  Services, Inc.   /\ tf

 SUBJECT:  Review and Summary of California Air Resources Board (CARB)
          Refinery Valves Inspection Program Data


 Introduction

    This memorandum summarizes the California Air Resources  Board (CARB)
petroleum refinery fugitive emissions inspection  data  received by EPA.
St«2rSJ%Sr ii? 5°"?"* peMod.for the Pr°P°sed "ew  source performance
standards for VOC fugitive emissions from petroleum  refineries (48 FR  279)
commenters on the standards (IV-D-8, IV-D-14, IV-D-15,  IV-D-21) either
requested that EPA obtain and analyze data on California  leak  detection
and repair programs or the commenters used the California program results
as a basis for their comments.   EPA received  the  results of  several CARB
inspections,  in three separate submittals,  as follows:

    o     Letter  from ARCO1  with data from four facilities.

                    frora the Ba* Area A1r Quality Management District
                    with data  from  four  bay area  facilities.
                      -,      Coast A1 r Qual1t* Management District
                    with data from eight facilities.
 Data  Description
seefromthpfn--      •   CARB insPecti°ns "*re the field data
JniSSin  H I   facilUy inspections.  For each valve inspected, the
following data were tabulated:

    1.  Source ID #
    2.  Time monitored
    3   Date Monitored (for entire process unit)
    4.  Date last monitored (for entire process unit)
    5.  OVA« leak detector reading
    6.  TLV« leak detector reading
    7.  Component line size, rounded to nearest inch
    8.  Component type and service
    9.  Pre-and post-calibration results.

These inspection  data were found to provide  information  that could be used
r£ic 1211   A     occurrence ™tes and estimate average  monitoring times.
This memorandum presents  these data analyses.


                             C-29

-------
CARB Data Analysis

     CARB is currently performing a separate analysis of these data,  as
they relate to local  rules.  CARB's analyses will  include the following
differences from this analysis:

    o   CARB is to "correct" the leak readings based on the measured
        calibration drift in the monitoring instrument between pre-test
        and post-test calibrations.

    o   Those sources measuring 10,000 ppm are not considered leaks by
        CARB, while those measuring greater than 10,000 ppm are leakers.
        EPA defines greater than or equal  to 10,000 ppm a leak.

    o   CARB is attempting to change the calibration basis from methane
        to hexane for some data.

CARB should publish the results of their analysis in the near future.

EPA Data Analysis

    From the field data received, EPA calculated the leak frequency
based on those valves reading greater than or equal to 10,000 ppm organics
and leak occurrence rate on a process unit and overall refinery basis.
Leak frequency was calculated as the number of leaks divided by the
number of components monitored, and expressed as a percentage.  The 30-day
leak occurrence rate was calculated by dividing the leak frequency by the
number of months elapsed since the last inspection, and again expressing
the result as a percentage.  It should be noted that the sources were
monitored with the instrument probe at 1 cm from the source.  Under NSPS
requirements  (Method 21),  the instrument probe must be placed at the
surface of the component.  While the relationship between the organics
concentration and the distance form the source is not known precisely,
readings at  1 cm  from the  source should normally be lower, reducing the
measured leak frequency compared to measurements taken at the source.

    Table  1  provides a summary of  the data received by EPA on a refinery
basis.  These data are developed in Tables 2 through 13 for each refinery.
These  data should be used  judiciously, as suggested by the following
general comments  on the CARB data.

    Four Century  Model 108 OVA were used along with four Bacharach TLV
for this survey.  The AQMD expressed concerns on the accuracy of TLV
readings as  calibration knobs are  easily moved, precalibrations of
TLY's  were not always done with gas (however, post-calibration were), two
teams  did  not use gas for  either pre or post-calibrations, probe tips were
sometimes contaminated, TLY's were not allowed to stabilize, and some TLV
readings were potentially  questionable due to be saturation of instrument
on  low scale.  One CARB diTutor had leaks in the dilutor probe resulting
in  questionable  readings and another required odd computations to obtain
a reading.   OVA  data from  the  first day at Shell and the second day at
Fletcher are questionable  because  one OVA did not hold calibration well.
These  data are footnoted on Tables 8 and 10.  Some leaks are noted in
excess of  10,000  ppm when  the  instrument was calibrated on methane.


                              C-30

-------
Equivalent readings would be 15,800 ppm for AQMD rule based on 10,000 ppm
hexane.  Also, the determination of the background reading varied from
CARB team to team.

    As mentioned above, EPA also calculated the monitoring time required
for each process unit.  These times were calculated by subtracting the
start time for a given process unit from the end time, and dividing the
result by the number of components in the unit.  Where gaps in the
monitoring time from one component to the next of more than 15 or 20
minutes were noted, the time of these gaps was subtracted from the
monitoring time.  Shorter gaps (< 15 minutes) were not subtracted, however,
to allow for normal operator breaks, instrument flameouts, and other
normal break periods.


References

1.  Letter, P.M. Kaplow, ARCO to Central Docket Section, U.S.E.P.A.;
    Results 6f Refinery Inspections.  June 15, 1982.   Docket Item
    Number IV-D-31.*

2.  Memo, T.L.  Norwood, PES, Inc., to Docket A-80-44, Local  Air Quality
    Management District Refinery Inspection Data.   November 21, 1983.
    Docket Item Number IV-B-18.*

3.  Letter, D.M. Newton, SCAQMD to S.R.  Wyatt, U.S.  E.P.A.;  Refinery
    Inspection  Data.   October 24,  1983.   Docket Item Number IV-D-37.*

*Document numbers refer to entries in Docket A-80-44, which  can be
 found at the U.S.  Environmental  Protection Agency Library,  Waterside
 Mall, Washington,  D.C.
                                   C-31

-------
       Table  1.   SUMMARY  OF CALIFORNIA AIR  RESOURCES  BOARD  INSPECTION DATA
REFINERY
LOCATION
Tosco "Corp.
Bakersfield
Shell Oil Co.
Martinez
Exxon Co.
Benicia
Chevron USA
Richmond
Arco
Carson
Mobil Oil Corp.
Torrance
Fletcher Oil
Carson
Champlin Oil
Wilmington
Shell Oil
Carson
Chevron USA
El Segundo
Newhall
Newhall *
Power ine
Santa Fe Springs
TOTAL 6
NUMBER OF
SOURCES
INSPECTED
326
816
858
803
591
338
317
435
683
602
152
576
,497
MONITORING
TIME
(min/source)
0.61
1.1
0.72
1.02
0.8
1.1
1.1
0.9
0.9
1.1
1.1
0.8
0.9

PERCENT
LEAKING
0.31
3.7
3.6
2.1
1.5
3.6
2.2
1.4
1.3
3.2
0.7
2.1
2.4
OVAa
b 95
C.I
0 - 0
2.4-
2.3 -
1.1 -
0.5 -
1.6 -
0.6 -
0.3 -
0.4 -
2.4 -
0.0 -
0.9 -
2.0 -

o/
"c
.93
5.0
4.9
3.1
2.5
5.6
3.8
2.5
2.2
4.0
2.1
3.3
2.8

30 DAY .
OCCURRENCE0
0.02
1.27
0.50
0.17
0.21
0.31
0.42
0.27
0.24
0.25
0.14
a. 51
0.41
PERCENTb
LEAKING
0.61
4.3
4.0
1.7
1.5
3.3
1.9
1.1
1.2
2.3
0.7
1.6
2.3
TLVa
95%
C.I.C
0-1.47
2.9 -5.7
2.7-5.3
0.8-2.6
0.5-2.5
1.4-5.2
0.4-3.4
0.1 -2.1
0.4-2.0
1.1 -3.5
0.0-2.1
0.6-2.6
1.9-2.7

30 DAY
OCCURRENCE0
0.05
1.61
0.56
0.14
0.18
0.29
0.36
0.21
0.21
0.13
0.14
0.39
0.41
a.  OVA  refers to the  Foxboro, Inc. organic vapor  analyzer;  TLV  refers to the  Bacarach, Inc.  "Sniffer"
   organic vapor analyzer.
b.  Leaks are defined  as those sources measuring greater than  or equal to 10,000 ppm organics  concentration
   at a distance of one centimeter.

c.  95%  C.I. = 95 percent confidence  interval  of percent leaking. This is estimated as P t '2SD, where P =
   percent leaking;                            	
                  SO = standard  deviation =  ./ P(IOO-P)   , and N = number  of sources inspected.
                                           V    N	
d.  The  30-day occurrence rates are calculated using the number  of sources inspected, the number of days from
   the  last plant inspection to the  CARB inspection, and the  measured percent leaking:
30 day occurrence rate =
                                             and  plant-weighted rate =  £(Neach Process un1t X 30  day occurrence)
                               periods
                                                                                 N
                                                                                  total sources  inspected
                                                    C-32

-------
            PROCESS
             UNIT
                        Table 2.  CARB FACILITY  INSPECTION SUMMARY FOR TOSCO,  BAKERSFIELD
           DATE OF
         LAST PLANT
         INSPECTION
                     6/f/FZ
DATE OF
CARB
JNSPECTION
*/AVfe*
*/fe.>,
—

NUMBER OF
SOURCES
INSPECTED
/*y
/
-------
Table 3.  CARB FACILITY INSPECTION SUMMARY FOR SHELL OIL,  MARTINEZ
DATE OF DATE OF
PROCESS LAST PLANT CARB
UNIT INSPECTION INSPECTION
LPG Storage 9/3/82 6/20/83
LPG Loading 2/1/83 6/21/83
LPG Storage 9/3/82 6/21/83
FCCU 5/23/83 6/21-22/83
CFH a 6/22/83
CGH b 6/22/83
CFH c 6/23/83
Alkylatlon d 6/23/83
Utilities e 6/24/83
Catalytic f 6/24/83
Reformer
TOTAL
SOURCE: Reference 2.
NUMBER OF
SOURCES
INSPECTED
50
72
31
162
63
15
85
160
120
58
816

NOTE: Terms and calculation procedures are defined In the
a5-l-83 62 sources
No date available 1 source
D5-23-83 6 sources
6-15-83 9 sources
C2-1-83 37 sources
5-1-83 48 sources
d2-l-83 18 sources
5-1-83 142 sources
e2-l-83 14 sources
6-13-83 6 sources
6-15-83 89 sources
For 11 sources, last plant date unknown.
f 2- 1-83 4 sources
3-31-83 21 sources
5-23-83 33 sources





MONITORING OVA
TIME NUMBER PERCENT 30-DAY NUMBER
(m1n/source) LEAKING LEAKING OCCURRENCE LEAKING
1.9 5 10.0 1.03 7
2.1 5 6.9 1.50 6
1.8 6 19.4 2.00 6
0.99 1 0.62 0.62 1
0.76 0 0 0.0 0
1.13 0 0 0.0 1
0.73 0 0 0.0 0
1.00 7 4.4 1.47 7
0.88 3 2.5 2.87 3
1.31 3 5.2 2.60 4
1.1 30 3.7 1.27 35
•
footnotes to Table 1.





TLV
PERCENT 30-DAY
LEAKING OCCURRENCE
14.0 1.45
8.3 1.77
19.4 2.00
0.62 0.62
0 0.0
6.7 12.4
0 0.0
4.4 1.47
2.5 2.87
6.9 3.45
4.3 1.61







Average of remaining 109 sources used.




-------
CO
en
                       Table 4.  CARB FACILITY INSPECTION SUMMARY  FOR  EXXON  USA,  BEiJICIA
            PROCESS
            UNIT
         SOURCE:   Reference 2.

         NOTE:  Terms  and calculation procedures are  defined in the footnotes  to  Table 1.

-------
                        Table 5. CARB FACILITY  INSPECTION SUMMARY FOR CHEVRON,  RICHMOND
o
CO
01
PROCESS
UNIT
<"6f (LP(,
fflK£iy/*j&)
feeo
tfcvfty
%%%£*
r«»
/SJkr««,
Mirth * /t
tyeMtfrfftnr*
'(iJofO/l^t.
n$ yf^Af,€
Tur/t-t.
DATE OF
LAST PLANT
INSPECTION
W/&1
5/4/91
y/77/£2.
V/fc/22^
V*//fe»
7/76/52,
4//S/M-
Wrfa

—
DATE OF
CARB
INSPECTION
L/tf/v*
6//V/J33
6//r/bs
*/*/**
Lfir/&3
^/^r/ss
*//
%>
&03
MONITORING
TIME
(•In/source)
1-2-
f-o
o.e>o
o.lr
O.W
/.2-
0.&?
I.I
/ ^/
/' ^
£>
r
o
z
^
v
/*-
OVA
PERCENT
LEAKING
O
O
0
o
3-1
o
I.Z
3-^r
V-V
^. /

30-DAY
OCCURRENCE
0
O
O
O
ait
o
O.OJ
o.tj
0,36
O.I*

NUMBER
LEAKING
O
O
O
O
$-
O
Z-
r
^
it-
TLV
PERCENT
LEAKING
0
O
O
O
3.1
O
/•3
z.f
z.*
1*

30-DAY
OCCURRENCE
O
O
o
0
0,26,
0
o.o?
0.2?
ai&
o,«t
             SOURCE:  Reference 2.

             NOTE:  Terms and calculation procedures  are  defined in the footnotes to Table  1.

-------
                          Table 6.  CARB FACILITY INSPECTION SUMMARY  FOR  ARCO,  CARSON
o
I
CO
PROCESS
IIMIT

"ZZ*
ffcu
«***.
£rf^&
3/4&/t4 ۥ
t/\C A* *^^
«|MrW
'EF*&fieR_
2
FAiwtxfEvfc
*Wfc
DATE OF
1 ACT Dl AUT
LAo 1 rLftW 1
INSPECTION
• »/»/«.
9I«H*
?/*>/«-
«/«/,*.
• Vtn.
afrh*
i*t,fi*
1/2V/71.
-
DATE OF
CARB
INSPECTION
s/6/as
6/Jfgj)
&to
S/3/73
S/5/8*
S»/9*
V/3S/9*
4/*r/?3
—
NUMBER OF
SOURCES
INSPECTED
?0
V8
5a
/dto
^
/jo
?/
20
&t
MONITORING
TIME
(•In/ source)
. <5>6
/^
af
. 0.2
0.<
A^
tf.f
1.0
£>.&

NUMBER
LEAKING
/
/
d>
3
^?
V
O
O
7
OVA
PERCENT
LEAKING
I.I
2 /
O
$.0
O
5-3
o
0
t.s-

36-DAY
OCCURRENCE
0,16
0.l!T
O
at,?
o
0,1*
£>
O
O.ll

NUMBER
LEAKING
1
O
0
3
o
3
o
0
*
TLV
PERCENT
LEAKING
l.f
0
0
3.0
O
z.r
o
0
AS"

30-DAY
OCCURRENCE
o.u,
O
0
t.fr
0
O.J.I
o
o
o><*
            SOURCE:   Reference |.

            NOTE:  Terms and calculation  procedures are defined in the footnotes to Table 1.

-------
                     Table 7.   CARB FACILITY  INSPECTION SUMMARY FOR MOBIL, TORRANCE
I
oo
CO
PROCESS
UNIT
Kt* ,A/
HypHo -
fSFotflSt.
'BbJAAt
i-Pc,
•VTA**
DATE OF
LAST PLANT
INSPECTION
6/?/^
/fa
i/i//fri-
S/^/g^
7/fc/g^
W*"
—
DATE OF
CARB
INSPECTION
• *j/ZL/&3
i(/}t>/&~l
4fo-»fo
4fi ?/V3
4/ZB/&*
«fa»*
—
NUMBER OF
SOURCES
INSPECTED
*r
JS-
%
^r
sd
sr
33H
t«WITORING
TIME
(•In/ source}
/.*
hi.
I.B
/-*
ay-
0,1,
/./

NUMBER
LEAKING
1
0
4
«r
^
o
ft.
OVA
PERCENT
LEAKING
Z.Z
0
9-.?
^?
O
0
5-*

30-DAY
OCCURRENCE
o.ao
c?
0,70
C.59
(9
0
o,3/

NUMBER
LEAKING
/
0
5-
5"
0
O
II
TLV
PERCENT
LEAKING
z.z-
^
^?-
4-f
<9
,^D
0,^)
O
O
«.*>
           SOURCE:  Reference \ .

           NOTE:   Terms and calculation procedures are defined in the  footnotes  to Table 1.

-------
i
OJ
                            Table 8.   CARB FACILITY INSPECTION SUMMARY FOR FLETCHER OIL, CARSON
               PROCESS
               'UNIT
             SOURCE:   Reference 1  .


             NOTE:  Terms and calculation  procedures are defined in the footnotes  to Table 1.

             * The OVA did not hold calibration  well during testing of these  units

-------
                   Table 9.  CARB FACILITY  INSPECTION SUMMARY FOR CHAMPLIN OIL,  WILMINGTON
o
I
o
PROCESS
UNIT
CcKft.cn>
<«*,*»*
rc«
CJ32*
»****.

Futi &A*
**«ess//i^
^ Vo»«
701*1 S.
DATE OF
LAST PLANT
INSPECTION
«/z//fe.
K/V*^
'/63
tfofa
r/c/ti
/'A/te
l///5/fe-
"//r/fc.
—
DATE OF
CARB
INSPECTION
454-/^
V/feve.
V/2^/i£j
V/^6/V4
4^?A?
^V^
4fofa-
ffrfi*
—
NUMBER OF
SOURCES
INSPECTED
Ik
zf

-------
               Table  10.   CARB FACILITY  INSPECTION SUMMARY  FOR  SHELL, CARSON
PROCESS
UNIT
FZCU

}£££
II A Tit*)
6/.SW,

**«#£*'
HyoeaHt&ttt
a>tet-
-7*0*.
DATE OF
LAST PLANT
INSPECTION
/. • „
n/^/s^
nlisfgi
l! •'<;• *-.
VAV«*
/'/"A*
3/*?/g3
"/T/3Z
"/J3/Z2.
—
DATE OF
CARB
INSPECTION
V/£?/&*
4fa/Bi
Sfo/**
5/2./B3
S/Z/83
sfifys
S/3/tlS
5/r/23
r/r/s
—
NUMBER OF
SOURCES
INSPECTED
43
nc,
23
<}
NUMBER
LEAKING
/ *
0*
/
i.
0
3
0
O
O
1
OVA
PERCENT
LEAKING
O
l.y-
J.f
O
?.t>
0
O
o
1.5
30-DAY
OCCURRENCE
0
*./1
• .' ,"
0
110
o
o
o
•J ' '

NUMBER
LEAKING
O
O
/
2-
0
s-
£>
O
o
?
TLV
PERCENT
LEAKING
O

,7
oU
O
?(0
0
0
f>
t.1
30-DAY
OCCURRENCE
O

0,11

0
1,20
o
o
0

SOURCE:   Reference 1.
NOTE:  Terms  and calculation procedures are defined  in  the footnotes to Table 1.
*The OVA  did not  hold calibration well during testing of these units.

-------
                           Table 11.  CARB FACILITY INSPECTION  SUMMARY  FOR CHCVRON,  EL SEGUUDO
o
I
ro
              PROCESS
               UNIT
                 CAT.
              fffU
              ffjcu
              2, AM
               ft*ar*HL
  DATE OF
LAST PLANT
INSPECTION
                       3/S/25
 DATE OF
  CARB
INSPECTION
                                  */»./«
NUMBER OF
  SOURCES
INSPECTED
                                             3/
                       vr
                                             111
                                             58
 MONITORING
   TIME
(•In/ source)
                                   //J •
                                  l.t
                                  0$
                                                        I.I
 NUMBER
LEAKING
                                                                      0
         OVA
PERCENT
LEAKING
                                                                              Z.f
                                                       $.0
                                              1.?
                                                        0
                                                                               O
                                                                               0
                                                       3.3L
  30-DAY
OCCURRENCE
                                                     o.lg
                                                                                      0,77
                                                     0,31
                                                                 O
                                                     0.1S
                           TLV
 NUMBER
LEAKING
                                                                                                0
                                                   r
                                                                o
                                                                         0
                                                    It
PERCENT
LEAKING
                                                                                                        o
                                                                                                       o
                                                o?.3
  30-DAY
OCCURRENCE
                                                                                        0, 33
                                                                                                               O
                                                                   O
                                                                              o
                                          O.I*
             SOURCE:   Reference  2.
             NOTE:  Terms  and calculation procedures are defined  in the  footnotes  to Table 1.

-------
o
I
GO
           PROCESS

            UNIT
              Table 12.  CARB FACILITY INSPECTION SUMMARY FOR MEWHALL. MEUHALL. CALIFORNIA



                                                                                TLV
         c*
SOURCE:  Reference  2.


NOTE:  Terms and calculation  procedures are defined  in  the  footnotes to Table 1.
                                                                                                                    cn

-------
                 Table 13.  CARB FACILITY INSPECTION SUMMARY FOR POWERINE,  SANTA  FE  SPRINGS
o
I
PROCESS
UNIT
tATFfffnet.
F^
C«S£*
fyt*t>-
Hyc*o-
LP6
£fi**l
"£££*
FV6L4K
-ami
DATE OF
LAST PLANT
INSPECTION
IZ/I7/&Z'
t/a/13
I'fa/fr-
tfx/93
I/ZI/OZ
H/tr/fr
ll/,.f/**-
'/*l/u
I/ if /QLU
// ty^
—
DATE OF
CARB
INSPECTION
5/2/63
Sf*f83
S/3/&
S/3/S&
C/z/rs
S/i/g>
s/3/n
slv/ii
s/v/u
f 	
NUMBER OF
SOURCES
INSPECTED
&>
41
*¥•
So
31
$0
Co
loi
113,
STL
MONITORING
TIME
(mln/ source)
0.3
O.f
0,1
o. r
0. (,
/.Oi
I'S
0,1
o.sr
M
OVA
NUMBER
LEAKING
/
£
1
0
0
i
3
3
1
/t-
PERCENT
LEAKING
AZ-
4.1
/?
O
o.
Z.o
6,0
3-o
0.1
Z.I
30-DAY
OCCURRENCE
0.3
1.1-
o,y
O
o
°
O
/
/
3
1
1
PERCENT
LEAKING
/• 2-
t.y
/•?
0
a
1,0
Z.0
3.0
0.7
•/.(,
30-DAY
OCCURRENCE

O.b
o.y
0
o
o.*/
0.*
0.7
0-3
o.v
SOURCE


NOTE:
                 :   Reference 2.


                  Terms and calculation procedures  are  defined in the footnotes to Table  1.

-------
           SECTION C.4

SOUTH COAST AIR QUALITY MANAGEMENT
   DISTRICT STUDY DATA SUMMARY
               C-45

-------
                  UNITED STATES ENVIRONMENTAL PROTECTION AGENCY
                          Office of Air Quality Planning and Standards
                         Research Triangle Park, North Carolina 27711
    """'                                                               A-80-44
                                   SEP 1  5 1983
                                                                   ,  iv-jB-n
MEMORANDUM

SUBJECT:  Review of ORD Study of South Coast Air Quality Management
          District Fugitive Emissions
FROM:     K. C. Hustyedt
          Petroleum Section, CPB/ESED (MD-13)

TO:       James F. Durham, Chief
          Petroleum Section, CPB/ESED (MD-13)


BACKGROUND AND CONCLUSIONS

     Several years ago, we requested that the Office of Research and
Development (ORD) analyze the effectiveness of local rules at reducing
fugitive emissions.  The ORD contracted with Radian and GCA to perform
fugitive emission testing at two refineries operating under the South
Coast Air Quality Management District (SCAQMD) Rule 466.1  on leakage from
valves and flanges.  As possible, the contractors also gathered historical
data on the implementation of Rule 466.1 and other information.  The
results of this study are reported in "Evaluation of the Maintenance
Effect on Fugitive Emissions from Refineries in the South  Coast Air
Quality Management District," EPA-600/7-82-049, January 1982.1

     An analysis of the absolute effectiveness of local rules based on
the historical Rule 466.1 implementation data is not possible because
of changes in the sample populations from test to test. However,
several general observations can be made from this study.   The rate of
leak occurrence found in this study is similar to what the results of the
maintenance study2 would predict for the given initial leak frequencies.
At one of the refineries where maintenance was performed whenever the
contractor found a leak, greater than 99 percent of the leaks were
repaired within 2 days with a resulting calculated emission reduction of
95.7 percent.  Further, as an indicator of the reliability of portable
detectors at identifying leaking sources, 97.3 percent of  the sources
identified as leaking by the original inspection team were also identified
as leaking by a second inspection team.  Overall only 2 of the 521 sources
screened  (99.6 percent) by two independent screening teams had different
leak/no leak determinations.  Regarding the ability to screen all valves
in a process unit, 12.3 percent of the overall valves were not screened
for various reasons, including high background organics concentration,
location, and instrument problems.  Only 3.3 percent of the overall valves
could not be monitored because they were difficult to monitor without
                                     C-46

-------
 extraordinary aids such as scaffolding or a cherry  picker.   Finally, by
 assessing pump seal  maintenance records,  it was determined  that  pump seal
 concentration is essentially independent  from seal  age  and  that  both
 operating and spare mechanical  pump seals are replaced  on average  every 1
 to 1.5 years with 90 percent of the seals replaced  within 3 years.  All
 of these data support the estimates we made in the  development of  the
 proposed refinery equipment leaks  new source performance standard  (NSPS).
 Using the results of the SCAQMD study in  Radian's leak  detection and
 repair (LDAR) modelJ,  I estimated  the cost effectiveness of routine
 monitoring programs.   This analysis shows that replacing their present
 programs with monthly  monitoring would have a  cost  effectiveness of about
 50 dollars per megagram ($/Mg)  and that the incremental cost effectiveness
 of monthly monitoring  over quarterly monitoring for these plants would be
 less than 1000 $/Mg.

 DISCUSSION

      The analysis of the effectiveness of Rule  466.1 in the SCAQMD study
 relies upon historical  data  developed by  the refineries in  their imple-
 mentation of Rule 466.1.   Reviewing  the trend  in percent of sources
 leaking from inspection to inspection should give an indication of the
 effectiveness of  Rule 466.1  at  reducing fugitive emissions.   In this
 testing,  however,  the number of  sources tested  often changed more than
 the number of sources found  leaking,  indicating that some unknown portion
 of the leaks  detected after  the  first inspection may have been leaking at
 the first inspection but were not  screened.  For this reason, no attempt
 has  been  made to  estimate  the overall  effectiveness of Rule 466.1 at
 reducing  fugitive  emissions.  There  are data in the SCAQMD  study that can
 be compared to numbers  we  used in  the  proposed  refinery equipment leaks
 NSPS  and  these are discussed in  the  following sections.

 Occurrence Rate - Table 4-16 in  the SCAQMD report presents historical
 data  on the implementation of Rule 466.1.   These data are used to estimate
 leak  occurrence in Table 1 based on the following assumptions:

      1.   All  leaks are  screened  and repaired on the last day of the month
 (to develop the time intervals).

      2.  Leak recurrence is insignificant because of the follow-up screening
 and maintenance performed under Rule 466.1 for repaired  leaks.

      3.  Changes in instrument,  instrument operator, calibration gas,  and
 number of sources screened have a small effect on the percent of sources
 found leaking.

     4.  If no leak detection and repair were performed, leaks would
 accumulate from the previous inspection to the next  inspection  (to  estimate
 overall occurrence rates from the beginning to the end of the test  period).

All of these assumptions appear  reasonable except possibly that  there is
 only a small effect from the change in number of sources screened.


                                   C-47

-------
     Table 1 shows that the 30-day occurrence rate from test to test ranged
from 0.24 to 1.80 percent and that the 30-day occurence rate for the
whole test ranged from 0.35 to 0..94 percent for the five units tested.
Table 2 shows a comparison of these overall  occurrence rates to the ones
used in the proposed refinery equipment leak NSPS.  As you can see, these
rates compare quite favorably, indicating refineries operating under the
SCAQMD Rule 466.1 have similar occurrence rates to those we estimated on
the national average in developing the refinery equipment leaks NSPS.

Maintenance Effectiveness - Chevron used Radian's testing as their annual
check under SCAQMD Rule 466.1.  For this Rule, all leaks must be repaired
within 2 days.  Of the 347 valve leaks detected by Radian, 344 were
repaired by means ranging from simple packing adjustment to sealant
injection and valve replacement.  The remaining three valves were taken
out of service.  This equates to a maintenance effectiveness of 99.1
percent within 2 days, as opposed to the estimate of 90 percent repaired
within 15 days in the refinery leaks NSPS.   The reduction in mass emissions
due to maintenance was estimated based on screening valves to be 95.7
percent.  In the refinery leaks NSPS, successful  leak repair was estimated
to result in an emission reduction of 97.7  percent.  As with occurrence
rates, the estimates of maintenance effectiveness in the SCAQMD study
compare favorably with those used in the refinery leaks NSPS.

Test Method Reliability - As in all research efforts, a quality assurance
(QA) check was performed during the SCAQMD  testing by the EPA contractors.
As a part of this QA effort, approximately  5 percent of the sources were
independantly screened by a second screener.  In  this QA testing, 37
sources were found to be leaking during the initial screening and 36 of
these sources, or 97.3 percent, were also found to leaking during the
second screening.  A total of 521  sources were screened in the QA effort,
and all but 2 of the sources were either found to be leaking by both
screeners or found not to be leaking by both screeners.

Difficult-to-Monitor Sources - During fugitive emission screening programs
by the EPA contractors, several sources are not monitored for various
reasons.  To assess the magnitude of this problem, Radian identified the
reasons sources were not screened during their testing.   Table 3 summarizes
these results from Table 4-1 of the SCAQMD  report for gas and light-liquid
service valves only.   As shown in Table 3,  most of the sources not screened
(8.6 percent of the overall sources and almost 70 percent of those not
screened) could have been screened if ladders had been provided.   An
additional 3.3 percent of the 12.6 percent  not screened (about a fourth)
were also not screened because of location,  but these sources  would have
needed extraordinary aids such as scaffolding or  a movable crane (cherry
picker) to have been screened.  The remaining few sources,  about 0.6 percent
of the total, that were not screened were for temporary  reasons,  such as
the source out of service, sampling problems, or  the plant not allowing
the contractor access to certain areas.
                                     C-48

-------
 Pump Seal  Replacement -  The  maintenance history of pump seals was acquired
 by examining refinery records.  Data were  available on 98 pumps with a
 total  of 544 seal  replacements.  A comparison of the screening value and
 the months since most recent seal replacement indicated a slight positive
 relationship (concentration  increasing with time), but age was not found to
 account for a very significant  portion of  the total variation in screening
 values.  These data are  consistent with our approach to controlling pump
 seal  emissions, which is based  on random accidents, mistakes, or catastrophic
 seal  failures causing pump seal leaks rather than gradual deterioration of
 the seals.   Only a routine inspection program can quickly identify these
 unpredictable leaks for  replacement.  Routine seal replacements or infrequent
 seal  inspections would not be as cost-effective because routine replacement
 would cause properly operating  seals to be replaced with no emission reduc-
 tion and because infrequent  inspections would allow seals to remain leaking
 for long periods of time.

      An  analysis of the  historical average length of time between seal
 replacements  was also  performed.  It was found that the average length  of
 time  between  seal  replacement was 1 to 1 and 1/2 years and that 90
 f!fnoent  °f  the pump seals are rePlaced within 3.years.  In the refinery leaks
 NSPS,  we estimated that  pump  seals are routinely replacaed on the average
 every  2  years,  which compares favorably with these findings.

 LDAR Analysis -  Several  inputs  for the LDAR model can be derived from the
 SCAQMD study  of valves.  These inputs were used, along with other necessary
 inputs as documented  in  the AID-3, to assess the cost effectiveness of
 additional  controls  for  valves  in SCAQMD refineries.   The new inputs based
 on  the SCAQMD  study  results are compared to the refinery NSPS inputs in
 Table 4.  The initial  percent leaking of 6.2 percent is the average for the
 five SCAQMD units  tested.  The emission factor was derived from this initial
 leak  frequency  using the leak/no leak emission factor calculation techniques
 developed in  the AID.  The leak occurrence rate is the average of the
 overall  occurrence  rates shown in Table 2.   The emission reduction and
 repair rate are  also from the SCAQMD study averages as discussed elsewhere
 in  this memo.

     Although the SCAQMD and NSPS inputs appear similar,  the LDAR model  was
used to determine the extent to which the inputs could effect the cost
effectiveness of routine monitoring programs.   The complete inputs and
outputs of this analysis are attached and the  results  are summarized in
Table 5.   As you can see, all of the cost effectiveness numbers are rather
low, with the highest one, the incremental  cost effectiveness of monthly
over quarterly monitoring, less than 1000 $/Mg.
                                     C-49

-------
REFERENCES;

1.  Evaluation of the Maintenance Effect on Fugitive Emissions From Refineries
in the South Coast Air Quality Management District, EPA 600/7-82-049,
December 1981.

2.  Evaluation of Maintenance For Fugitive VOC Emission Control, EPA-600/2-
81-080, May 1981.

3.  Fugitive Emission Sources of Organic Compounds--Additional Information
on Emissions, Emission Reductions, and Costs, EPA-450/3-82-010, April 1982.

4.  VOC Fugitive Emissions  in Petroleum Refining Industry--Background
Information for Proposed Standards,  EPA 450/3-81-015a, November 1982.

Attachments

cc:  Fred Dimmick, SDB
     Tom Rhoads, PES
     Refinery Leaks Docket
                                    C-50

-------
                             TABLE  1.    ESTIMATED RATE OF LEAK OCCURRENCE FOR SCAQMD VALVE DATA
   PROCESS
    UNIT
                   DATE
                 SCREENED
PERCENT
LEAKING*
MONTHS
BETWEEN
TESTS
TEST-TO-TEST
   30-DAY
 OCCURRENCEb
CUMULATIVE PERCENT
LEAKING FROM BEGIN-
   NING OF TESTC
MONTH? FROM
 BEGINNING
  OF TEST
  30-DAY
OCCURRENCE
FROM BEGIN-
ING OF TEST<1
Alkylation
 Isomax
FCCU
Crude Unit
Platformer
                   2/79
                   7/79
                   9/80
                   2/81

                   4/79
                   9/79
                  10/80
                   2/81

                   3/79
                   8/79
                  10/80
                   2/81

                  11/79
                   9/80
                   3/81

                  11/79
                   9/80
                   3/81
 14.5
  5.4
  8.2
  9.0

  2.6
  2.2
  3.6
  2.4

  4.1
  1.2
 11.7
  6.6

  1.5
  3.3
  2.3

  8.2
  3.5
  5.9
                                             5
                                            14
                                             5
                                             5
                                            13
                                             4
                                             5
                                            14
                                             4
                                            10
                                             6
                                            10
                                             6
             0.93
             0.59
             1.80
             0.44
             0.28
             0.60
             0.24
             0.84
             1.65
             0.33
             0.38
             0.35
             0.98
                      5.4
                     13.6
                     22.6
                      2.2
                      5.8
                      8.2
                      1.2
                     12.9
                     19.5
                     3.3
                     5.6
                     3.5
                     9.4
                          5
                         19
                         24
                          5
                         18
                         22
                          5
                         19
                         23
                         10
                         16
                         10
                         16
                 0,93
                 0.72
                 0.946
                 0.44
                 0.32
                 0.376
                 0.24
                 0.68
                 0.856
                 0.33
                 0.356
                0.35
                0.596
SOURCE:  SCAQMD STUDY  (Reference 1)
a
b
All but the last sreening test were performed by the refinery.
Based on all leaks from the previous inspection being repaired  and an assumption  of  linear  leak occurrence,
the leak frequency divided by the number of months between tests  estimates  the 30-day leak  occurrence rate!
Based on all leaks at initial inspection being repaired  and assuming  that if  the  other inspection had not
occurred the leaks would have accumulated from inspection to inspection.
Same methodology as discussed in footnote b except based on the initial  inspection.
These are the overall average monthly occurrence rates for the  units  studied.

-------
       TABLE 2.   COMPARISON  OF SCAQMD AND REFINERY NSPS OCCURRENCE RATES




o
1
Ul
ro

PROCESS
UNIT*
ALKYLATION
ISOMAX
FCCU
CRUDE UNIT
PLATFORM ER
INITIAL
LEAK
FREQUENCY a
14.5
2.6
4.1
1.5
8.2
SCAQMD 30 -DAY
OCCURRENCE
RATEb
0.94
0.37
0.85
0.35
0.59
REFINERY NSPS 30-DAY
OCCURRENCE
RATE
1.68
0.52
0.66
0.41
1.06
a  From SCAQMD study (Reference 1)
b  From Table 1 - average 30 day occurrence  rate  from  the  beginning  to  the  end
   of the test.
c  Occurrence rate calculated as occ  • 0.0976  x (initial leak  frequency) +  0.264
   based on a least squares analysis  of  the  occurrence rates from  the Maintenance
   study (Reference 2)  as discussed in AID (Reference  3).

-------
                  TABLE 3.   REASONS VALVES WERE NOT SCREENED DURING RADIAN SCAQMD TESTING
                                                              PERCENT NOT SCREENED SPLIT BY
                                                                  REASONS NOT SCREENED*
SERVICE
Gas
Light Liquid
OVERALL
NUMBER OF
SOURCES
2630
. 5677
8307
NUMBER
SCREENED
2337
4926
7263
rCKULNI
NOT
SCREENED
11.1
13.2
12.6
1 2
3.3
0.4 3.3
0.3 3.3
345
7.3 0.04 0.4
9.2 ~ 0.3
8.6 0.01 0.3
6 7
0.04 --
0.04 --
0.04 --
o
I
en
co
        Source:  SCAQMD  Study  (Reference 1)

     *Reasons  valves  not  screened:
00
    1.  Temporary  factors  such as sources taken out of service for repair.
    2.  Permanent  factors  such as sources that could not be reached without extraordinary aids.
    3.  Location,  such  as  sources that could not be screened using the probe extension, but could
        be  reached by using a stepladder.
    4.  High  background concentration.
    5.  Accessible,  but climbing not permitted.
    6.  Possible fouling of probe by visible leak to atmosphere.
    7.  Possible fouling of probe by visible leak to drain or sump.

-------
        TABLE 4,   COMPARISON OF SCAQMD STUDY AND REFINERY NSPS  LDAR INPUTS


PARAMETER                                     SCAQMD STUDY       REFINERY NSPSa
Emission Factor (kg/hr)
Monthly Leak Occurrence Rate (%)
Initial Percent Leaking (%)
Emissions Reduction for Successful Repair (%)
Unsuccessful Repair Rate (%)
O.OlQb
0.6C
6.2
95.7
0.9
0.0163
1.27
10.7
97!7
10
a - From Reference 4.

b - Calculated based on the average percent leaking and the leak/no leak technique
    developed in Reference 3.

c - Average occurrence rate from Table 2.  The quarterly occurrence rate would be
    3 times the monthly rate and the annual occurrence rate 4 times the quarterly
    rate.
                                    C-54

-------
                                   10
     TABLE 5.  COST EFFECTIVENESS OF ROUTINE MONITORING BASED ON
                       SCAQMD STUDY LDAR INPUTS
Monitorine
Interval
(Mo)
12
3
1
12-3*
12-1*
3-1*
) Emission
Reduction
(percent)
20.7
47.9 •
54.5
27.2
33.8
6.6
(Mg/yr)
18.1
42
47.8
23.9
29.7
5.8
Net
Cost
($/yr)
- 407
-2870
2580
-2463
2987
5450
Cost
Effectiveness
($/Mg)
- 23
- 68
54
-103
100
940
a - These denote the incremental  emission reduction,  cost,  and cost
    effectiveness between the two monitoring intervals shown.
                              C-55

-------
                                                 INPUT   DATA

                                                       PLANT SCV
                                                (FOR GCU. ANNUAL VALVES >
           FOR EXAMINING EMISSION REDUCTIONS DUE TO LDARI
                MONITORING INTERVAL (MONTHS)                              12
                TURNAROUND FREQUENCY (MONTHS I                             24
                EMISSION FACTOR (KG/HR/SOURCE)                         0.01
                LEAK OCCURRENCE RATE U PER PERIOD)                      7.4
                INITIAL * LEAKING                                        6.2
                EMISSIONS REDUCTION FOR UNSUCCESSFUL REPAIR (SI         62.6
                EMISSIONS REDUCTION FOR SUCCESSFUL REPAIR 1%)            93.7
                EARLY LEAK RECURRENCE (* OF REPAIRS)                    14.0
                UNSUCCESSFUL REPAIR RATE ill                             0.9
                UNSUCCESSFUL REPAIR RATE HI AT TURNAROUND               0.0
           FOR EXAMINING THE COSTS OF LOARl
                TOTAL NUMBER OF SOURCES                                1,000
                MONITORING TIME PER SOURCE INSPECTION  (MINUTES)           2.0
        i        VISUAL MONITORING TIME PER SOURCE (MINUTES)              0.00
        ^       NUMBER OF VISUAL INSPECTIONS PER YEAR                       0
                REPAIR TIME PER SOURCE (MINUTES)                          68
                LABOR RATE U/HOUR)                                       jfl
                PARTS COST PER SOURCE (*)                                   0
                ADMINISTRATIVE t SUPPORT OVERHEAD COST FACTOR  I*)        40.0
                CAPITAL RECOVERY FACTOR m                             16.3
                RECOVERY CREDIT FOR EMISSIONS REDUCTION  U/MG)            213
 o
 o
'o
 o
 o

-------
                                                 INPUT   DATA

                                                      PLANT SCV
                                                (FOR CtLL MONTHLY VALVES  )
 •

 o
FOR EXAMINING EMISSION REDUCTIONS DUE TO LOARI
               MONITORING  INTERVAL  (MONTHS)
               TURNAROUND  FREQUENCY (MONTHS)
               EMISSION FACTOR  (KG/HRXSOURCE)
               LEAK OCCURRENCE  RATE (X PER PERIOD)
               INITIAL X LEAKING
               EMISSIONS REDUCTION FOR UNSUCCESSFUL REPAIR (X)
               EMISSIONS REDUCTION FOR SUCCESSFUL REPAIR (XI
               EARLY LEAK RECURRENCE (X OF REPAIRS)
               UNSUCCESSFUL REPAIR RATE (?)
               UNSUCCESSFUL REPAIR RATE (X) AT TURNAROUND
                                                                1
                                                               24
                                                            0.01
                                                              0.6
                                                              6.2
                                                             62.6
                                                             95.7
                                                             14.0
                                                              0.9
                                                              0.0
      o   FOR EXAMINING THE COSTS OF LDARt
               TOTAL NUMBER OF SOURCE*                                1*000
               MONITORING TIME PER SOURCE INSPECTION (MINUTES!          2.0
               VISUAL MONITORING TIME PER SOURCE (MINUTES)              0.00
               NUMBER OF VISUAL INSPECTIONS PER YEAR                      0
               REPAIR TIME PER SOURCE (MINUTES)                          60
               LABOR RATE U/HOUK)                                       IB
               PARTS COST PER SOURCE ($)                                   0
               ADMINISTRATIVE t SUPPORT OVERHEAD COST  FACTOR  IX)        40.0
               CAPITAL RECOVERY FACTOR (X)                             16.3
               RECOVERY CREDIT FOR EMISSIONS REDUCTION U/MG) .          215
9

9

-------
                                                INPUT   DATA

                                                      PLANT SCV
                                               IFOR GCLL QUATERLY VALVES )
          FOR EXAMINING EMISSION REDUCTIONS DUE TO LDARl
               MONITORING INTERVAL IHONTHS)
               TURNAROUND FREQUENCY (MONTHS I
               EMISSION  FACTOR (KG/HR/SOURCE)
               LEAK OCCURRENCE RATE (X  PER PERIOD)
               INITIAL X LEAKING
               EMISSIONS REDUCTION FOR  UNSUCCESSFUL REPAIR  IS)
               EMISSIONS REDUCTION FOR  SUCCESSFUL REPAIR (X)
               EARLY  LEAK RECURRENCE  (X OF REPAIRS)
               UNSUCCESSFUL RbPAIR RATE IX)
               UNSUCCESSFUL REPAIR RATE IX) AT TURNAROUND
                                                                     3
                                                                    24
                                                                 0.01
                                                                   1.9
                                                                   6.2
                                                                 62. 6
                                                                 95.7
                                                                 14.0
                                                                   0.9
                                                                   0.0
I

I

ft
o
I
CJ1
oo
         FOR EXAMINING THE COSTS OF LDARl
         TOTAL NUMBER OF SOURCES                                1,000
         MONITORING TIME PKR SOURCE INSPECTION (MINUTES)          2.0
         VISUAL MONITORING TIME PER SOURCE (MINUTES)             0.00
         NUMBER OF VISUAL INSPECTIONS PER .YEAR                      0
         REPAIR TIME PER SOURCE (MINUTES)                          68
         LABOR RATE (S/HOUR)                                       18
         PARTS COST PER SOURCE (*)                                  0
         ADMINISTRATIVE C SUPPORT OVERHEAD COST FACTOR IX)        40.0
         CAPITAL RECOVERY FACTOR (X)                             16.3
         RECOVERY CREDIT FOR EMISSIONS REDUCTION U/MG)            215

-------
                                                                 •N
                                            PLMT SCV SUMMARY*


                                    AVERAGE  ANNUAL COST EFFECTIVENESS

                                               1 YEARLY LDAR)
SOURCE TYPE
VALVES
GCLL
GCLL
GCLL
ANNUAL
MONTHLY
QUATERLY
EMISSION
REDUCTION
(MG/YRI
18.1
47.8
42
RECOVERY
CREDIT
» 3,890
10,300
9,030
NET
COSTS
$ -407
2,580
-2,870
GROSS COST
EFFECTIVENESS
(PER HG)
$
193
269
147
NET COST
EFFECTIVENESS
(PER MG)
* -23
54
-68

    PLANT TOTAL
o
I
CJi
                           108
                                          23,200
-«92
                                                                           209
                                                                                            -6

-------

-------
                               APPENDIX D

                   MODEL UNIT AND NATIONWIDE IMPACTS
      Attached as Appendix D is a memorandum dated December 13,  1983
that documents the calculation of model  units and nationwide impacts  of
the promulgated standards.
                                  D-l

-------
                                                       A-80-44
                                                       IV-B-24

                          MEMORANDUM

                                     DATE:  December 13, 1983

TO:       Docket A-80-44, Petroleum Refinery VOC Fugitive Emissions
          NSPS

FROM:     Thomas Rhoads, Pacific Environmental  Services, Inc.

SUBJECT:  Calculation of Model Unit and Nationwide Impacts
     This memorandum documents the calculation of the capital  cost,
net annualized cost, and emission reduction resulting from implementation
of the standards.  The impacts are presented for each model  unit on a
yearly basis and nationwide in the fifth year of implementation of the
standards.  The basis and method for calculating the model unit and
nationwide impacts are from the background information documents for
the proposed standards and promulgated standards.

     Tables 1 and 2 show model unit and nationwide emission  reductions
achieved between baseline and uncontrolled scenarios.  Uncontrolled
means the level of control implemented by refineries in the  absence of
any regulations to control equipment leaks of YOC.  Baseline,  however,
reflects a nationwide average level of control implemented as  a result
of existing regulations (i.e., State and regional) to control  equipment
leaks of VOC.  Tables 3 and 4 show model unit and nationwide emission
reductions resulting from implementation of the final standards as the
increment between baseline emissions and the level of emissions following
promulgation of the standards.  As shown in Table 4, 31,100  Mg VOC
emission reduction would be achieved in the fifth year of implementation
of the final standards.  The cost impacts of the final  standards, likewise,
do not include the baseline product recovery credits and costs for
monitoring instruments (those costs incurred by the industry due to
existing regulations).  Net annualized costs to implement the  final
standards are derived in Tables 6, 7, and 8.  Implementation of the
final standards would cost approximately $4.14 million (1980 dollars)
with a cost effectiveness, therefore, of about $130/Mg VOC emission
reduction.  The cumulative nationwide capital  costs, calculated in
Tables 8 and 9, are projected at about $17.9 million in the  fifth year
of implementation of the final standards.
                                  0-2

-------
                      Table 1.  MODEL UNIT EMISSION REDUCTION BETWEEN
                                 BASELINE AND UNCONTROLLED
Equipment
Pressure
Relief
Devices
Compressors
Open Ended
Lines
Sampling
Connections
Val ves
Gas
Light
Liquid
Pumps
Total Emission
Alternative II
Control
Quarterly
LDR
Quarterly^
LDR
Cap
No Control
Quarterly
LDR
Annual d
LDR
Annual
LDR
Regul atory
Alternative II
Emission
Reduction per
Component
(Mg/yr) a
0.63
4.3
0.020
0
0.14
0.019
0.21
Reduction Regulatory
(Mg/yr)
Baseline Emission Reduction
(Mg/yr)e
Regulatory Alternative II
Model Unit Emission Reduction
(Mg/yr)b
A
Compo- Sub-
nents total
3 1.9
1 4.3
70 1.4
10 0
130 18.2
250 4.8
7 1.5
32.1
18.0
B
Compo-
nents
7
3
140
20
260
500
14



Sub- Compo-
total nents
4.4 20
12.9 8
2.8 420
0 60
36.4 780
9.5 1,500
2.9 40
68.9
38.6
C
Sub-
total
12.6
34.4
8.4
0
109
28.5
8.4
201
113
LDR = leak detection and repair

Footnotes on next page
                                            D-3

-------
         Table 1.  MODEL UNIT EMISSION REDUCTION BETWEEN
                   BASELINE AND UNCONTROLLED (concluded)

aFrom BID for the promulgated standards,  Appendix A,  pages A-4,  7,  8,
 9, 10, and 14.


bModel unit equipment counts are found in Table 6-1  of the BID for  the
 proposed standards.  The model unit emission reductions are obtained  by
 summing the products obtained from the per component emission reductions
 and the number of components per model unit.

cFrom Table 7-1, BID for the proposed standards.  Uncontrolled emission
 factor = 15.0 kg/d.  Controlled emission factor for quarterly LDR  =
 3.2 kg/d.
      Emission
      Reduction
(15.0 kg/d -  3.2  kg/d)  (365  d/yr)f  1 Mg  \
 =4.3 Mg/yr                      V1000 kg/
dFrom Table F-3, BID for the proposed standards.   Uncontrolled  emission
 factor = 0.26 kg/d
 Controlled emission factor for annual  LDR  =  0.209 kg/d

      Emission    = (0.26 kg/d - 0.209 kg/d)  0.365
      Reduction      = 0.019 Mg/yr
eBaseline emission reduction is  achieved  by  industry  using  existing
 levels of control.  About 44 percent of  the petroleum  refining  industry
 is located in attainment areas  for  ozone and not  subject to  equipment
 leak regulations (uncontrolled),  and about  56 percent  is located  in
 non-attainment areas and subject  to State or local regulations
 (Regulatory Alternative II). Baseline emission reduction  is, therefore,
 calculated as 56 percent of the emission reduction achieved  under
 Regulatory Alternative II.
                                    D-4

-------
                Table 2.  NATIONWIDE EMISSION REDUCTION BETWEEN
                           BASELINE AND UNCONTROLLED

Model
Unit
A
B
C

Emission Reduction
Per Model Unit
(Mg/yr)*
18.0
38.6
113


Projected
Model Unitsb
96
106
80
TOTAL

Subtotal
(Mg/yr)
1,730
4,090
9,040
14,900
Emission reductions from Table 1.

bFrom BID for the proposed standards,  Table  7-4.
                                    D-5

-------
                     Table 3.  MODEL UNIT EMISSION REDUCTION BETWEEN
                                    NSPS AND BASELINE
Equipment
Pressure
Relief
Devices
Compressors
Open Ended
Li nes
Sampling
Connections
Val ves
Pumps
Control
Disk
Barri er
Fluid
System
Cap
Closed
Purge
Monthly
LDR
Monthly
LDR
Emission Reduction From
to NSPS
Emission Reduction From
to Basel i neb
Emission Reduction From
Emission
Reduction Per
Component
Between
NSPS and
Uncontrolled
(Mg/yr)a
1.4
5.5
0.020
0.13
0.10
0.82
Uncontrolled
Uncontrolled
Baseline to NSPSC
Emission Reduction by Model Uni
A B
Compo- Sub- Compo- Sub-
nents total nents total
3 4.2 7 9.8
1 5.5 3 16.5
70 1.4 140 2.8
10 1.3 20 2.6
380 38.0 760 76.0
7 5.7 14 11.5
56.1 119
18.0 38.6
38.1 80.4
t (Mg/yr)
C
Compo- Sub-
nents total
20 28.0
8 44.0
420 8.4
60 7.8
2,280 228
40 32.8
349
113
236
aFrom BID for the promulgated standards Appendix A,  pages  A-4,  7,  8,  9,  10,  and 14.

bFrom Table 1.

cRepresents emission reduction per model  unit resulting from  promulgation  of the
 standards.
                                            D-6

-------
              Table 4.  NATIONWIDE EMISSION REDUCTION BETWEEN
                       NSPS AND BASELINE

Model
Unit
A
B
C
Emission Reduction
Per Model Unit
(Mg/yr)a
38.1
80.4
236

Projected
Model Unitsb
96
106
80

Subtotal
(Mg/yr)
3,660
8,520
18,900
                                              Total               31,100
Emission reductions from Table 3.

bFrom BID for the proposed standards,  Table  7-4.
                                   D-7

-------
                     Table 5.  MODEL UNIT NET ANNUALIZED COST BETWEEN
                               BASELINE AND UNCONTROLLED





f* • J_
Equipment

Pressure
Relief
Devices
Compressors^

Open Ended
Li nes
Sampl i ng
Connections
Valves
Gasc

Lightd
Liquid
Pumps







Net Annual i zed
Cost Per
Component
Between
Regul atory
Alternative II
Control and uncontrolled*1

Quarterly
LDR

Quarterly0
LDR
Cap

No Control


Quarterly
LDR
Annual d
LDR
Annual
LDR
Regulatory Alternative II
Instruments

Regulatory Alternative II
Instruments6
Net Annual i zed
Uncontrolled^

($/yr)
(170)


(690)

9.1

0


(21)

2.04

180

Costs Without

Costs With

Cost Between Baseline and

Regulatory

Alternative I

I
Net Annual i zed Cost
Per Model Unit ($/yr)
A

Compo- Sub-
nents total
3 (510)


1 (690)

70 637

10 0


130 (2,730)

250 510

7 1,260

(1,520)

3,980

2,230
B

compo- Sub-
nents total
7 (1,190)


3 (2,070)

140 1,270

20 0


260 (5,460)

500 1,020

14 2,520

(3,910)

1,590

890

C

compo- Sub-
nents total
20 (3,400)


8 (5,520)

420 3,820

60 0


780 (16,400)

1,500 3,060

40 7,200

(11,200)

(5,700)

(3,190)
LDR = leak detection and repair
(  ) = cost savings

aRegulatory Alternative II costs per component are  from the  BID  for  promulgated  standards
 Appendix A, pages A-4, 7, 8,  11, and 15.

                                             D-8

-------
         Table 5.  MODEL UNIT NET ANNUALIZED COST BETWEEN
                   BASELINE AND UNCONTROLLED (continued)


bQuarterly leak detection and repair for compressors,  from  BID  for
 the proposed standards, Table F-12.

                            (Initial   \  /RepairA / LaboA /OverA /Capital \
                            Leak      ]  I  Time   1 I  Rate 1  Head  [ Recovery I
                            Frequency/  \       /\    •/ \     / \Factor   /

                        = (0.35) (40 hrs)  ($18/hr)  (1.4) (0.163)

                        = $57.50

    Monitoring Labor = (Monitoring  labor hours)  (Labor  rate) =  (1 hr) ($18/hr)

                     = $18

    Repair Labor = (Repair Labor hours)  (Labor rate) =  6 hr ($18/hr) = $108

    Administrative  = 0.4/Monitoring    RepaiA =  0.4 (18 + 108) = $50.40
    and Support          I Labor      + Labor  1

    TOTAL ANNUALIZED COST =  $234                '

    RECOVERY CREDIT = ($215/Mg)  (4.3 Mg/yr)  = $924

    NET ANNUALIZED COST per  compressor = is  a cost savings of $690

GGas Service Valves,  from BID  for promulgated standards Table A-6.

dLight  Liquid Service Valves,  from BID for  proposed  standards Table  F-27.

      Initial  Leak Repair =[Initial  leak]  /Repair\/Labor\/Over-\ /Capital \
                           iFrequency    /  I  Time  )\ Rate jlhead If  Recovery)
                           \          /  x      '  \    /\     / I  Factor  /

                         =  (0.11) (1.13 hr) ($18/hr) (1.4)  (0.163)  = $0.51

    Monitoring Labor  =/Fraction  of\ /MonitoringW LaborN
                      I Sources     ) I Time       )\ Rate J
                      VScreened    / x          ' x      '

                     = (0.99)  (1/60  hr)  ($18/hr)   (2) = $0.59

    Repair Labor =/Fraction of\  /RepairX /LaborN
                  I Sources      |  I Time   ) I Rate  )
                  \0perated on /  x      '         '

                = 0.168  (1.13)  ($18) =  $3.42
                                 D-9

-------
         Table 5.  MODEL UNIT NET ANNUALIZED COST BETWEEN
                   BASELINE AND UNCONTROLLED (concluded)
                           /*
    Administrative   =0.4 [Monitoring   +   Repair
    and Support            I Labor            Labor
                     = 0.4 ($0.59 + $3.42) = $1.60
    TOTAL ANNUALIZED COST = $6.12
    RECOVERY CREDIT = (0.019/Mg/yr) ($215) = $4.08
    NET ANNUALIZED COST = $2.04 per valve
6Annualized instrument cost from BID for the proposed standards, Tables 8-1
 and 8-5.  Annualized cost = Capital recovery Cost + Maintenance Cost +
 Miscellaneous Cost.
                            /                 A    /^Capital        ^
     Capital Recovery Cost =l$9,200/Model Unity  X I Recovery Factor J
                           = $9,200/unit x 0.23
                           = $2,100/unit
     Maintenance Cost = $3,000
     Miscellaneous Cost * 0.04 x $9,200 = $368
                  Total  = $5,500/model  unit
 fSee footnote e from Table 1.  Baseline costs =  0.56 x Regulatory
  Alternative II.
                                    D-10

-------
                      Table 6.  MODEL UNIT NET ANNUALIZED COSTS BETWEEN
                                NSPS AND BASELINE
Net Annual i zed Cost



Equipment Control
Pressure Rupture
Relief Disk
Devices
Compressors Barrier
Fluid
System
Open Ended Cap
Li nes
Sampling Closed
Connections Purge
Valves Monthly
LDR
Pumps Monthly
LDR
Costs from Uncontrolled to
w/o Instrument
Costs from Uncontrolled to
Instrument15
Costs from Uncontrolled to
Net Annuali zed
Costs From
Uncontrolled
To NSPS
($/yr)a
580


840


9.1

105
(6)

130

NSPS
NSPS with
Basel inec
Costs from Baseline to NSPSd

A

Compo- Sub-
nents total
°3 1,740


1 840


70 637

10 1,050
380 (2,280)

7 910

2,900
8,400
2,230
6,170
Model Unit
B

Compo- Sub-
nents total
7 4,060


3 2,520


140 1,270

20 2,100
760 (4,560)

14 1,820

7,210
12,700
890
11,800
Per

c

Compo- Sub-
nents total
20 11,600


8 6,720


420 3,820

60 6,300
2,280 (13,700)

40 5,200

20,000
25,500
(3,190)
28,700
LDR = leak detection and repair
(  ) = cost savings
     basis for the control  costs for the individual  components represent the costs from
 uncontrolled to the control  required by the standards.   From Appendix A of the BID for
 the promulgated standards,  pages A-4, 7,  8, 11,  and 15.

 bSee footnote e of Table  5.

 CFrom Table 5.


 dRepresents model  unit  net  annual i zed costs between NSPS  and baseline.

                                           D-ll

-------
                 Table 7.  NATIONWIDE NET ANNUALIZED COSTS FROM
                           BASELINE TO NSPS

Model
Unit
A
B
C

Net Annual ized
Cost Per
Model Unit ($/yr)a
6,170
11,800
28,700


Projected
Model Units5
96
106
80
Total

Subtotal
(Mg/yr)
592,000
1,250,000
2,300,000
4,140,000
aNet Annualized Costs per model unit from Table 6.

bFrom BID for the proposed standards, Table 7-4.
                                            D-12

-------
                      Table 8.   MODEL  UNIT CAPITAL  COSTS
                                      Model  Unit  Capital  Cost
                                               ($)
Control
Regulatory Alternative IIa
Basel ineb
NSPS (from uncontrolled)
New Unit0
Modi f ied/Reconstructedd
Unit
NSPS (from Baseline)6
New Unit
Modi f i ed/Reconstructed
Unit
A
13,000
7,280
35,000
39,000
27,700
31,700
B
17,000
9,520
73,000
81 ,000
63,500
71,500
C
31,000
17,400
190,000
210,000
173,000
193,000
aFrom BID for the proposed standards, Table 8-2.

bSee footnote e of Table 1.  Calculated as 56 percent of Regulatory
 Alternative II.  Baseline capital costs are incurred by industry using
 existing levels of control and, therefore, represents a weighted average
 between uncontrolled (no cost) and Regulatory Alternative II, not the
 actual  capital  cost incurred by an individual  model unit.

°From BID for the proposed standards, Table 8-12.  Regulatory Alternattvef 7T
 capital costs are the same as that for the standards.

dFrom BID for the proposed standards, Table 8-13, corrected to include
 closed  loop sampling under Regulatory Alternative III.

Calculated as the capital  cost from uncontrolled to NSPS minus baseline
 capital costs.   Represents capital  cost incurred per model  unit regulting
 from promulgation of the standards.
                                     D-13

-------
                                           Table 9.  NATIONWIDE CAPITAL COSTS
o
I
New Units
Capital Cost
Model Projected Per Model Un1tb Subtotal Model
Unit . Model Unitsd ($) ($) Unit
A 49 27,700 1,360,000 A
B 27 63,500 1,710,000 B
C 24 173,000 4,150,000 C
Total 7,220,000
Modified/Reconstructed Units
Projected Capital Cost Subtotal
Model Units3 Per Model Unitsb ($)
47 31,700 1,490,000
79 . 71,500 5,650,000
56 193,000 10,800,000
Total 17,900,000
*" aFrom BID for the proposed standards, Table 7-4.


   bFrom Table 8.

-------
                                   TECHNICAL REPORT DATA
                            (Please read Instructions on the reverse before completing)
1. REPORT NO.
   EPA-45073-81-015b
                                                            3. RECIPIENT'S ACCESSION NO.
4. TITLE AND SUBTITLE
   Equipment Leaks of VOC  in the Petroleum  Refining
   Industry—Background  Information for Promulgated
   S tandards
                                                            5. REPORT DATE
                                                              December 1983
             6. PERFORMING ORGANIZATION CODE
7. AUTHOR(S)
             8. PERFORMING ORGANIZATION REPORT NO.
9. PERFORMING ORGANIZATION NAME AND ADDRESS
   Office of Air Quality Planning and Standards
   U.S.  Environmental  Protection Agency
   Research Triangle Park,  North Carolina   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,  North Carolina  27711
              13. TYPE OF REPORT AND PERIOD COVERED
              14. SPONSORING AGENCY CODE

                EPA/200/04
15. SUPPLEMENTARY NOTES  This  document presents  the background information used by tne
   Environmental Protection Agency in developing the promulgated  new source performance
   standards for equipment of VOC in the petroleum refining  industry.
16. ABSTRACT
        Standards of  performance for the  control of volatile  organic compound (VOC)
   equipment leaks from the petroleum refining 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 petroleum refinery  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.
17.
                                KEY WORDS AND DOCUMENT ANALYSIS
                  DESCRIPTORS
                                               b. IDENTIFIERS/OPEN ENDED TERMS  C. COSATI Field/Group
   Air Pollution
   Petroleum Refining
   Pollution Control
   Standards of Performance
   Volatile Organic Compounds (VOC)
  Air Pollution Control
13b
18. DISTRIBUTION STATEMENT
   Unlimited
                                               19. SECURITY CLASS (ThisReport!
                                                     Unlimited
                            21. NO. OF PAGES
                              214
20. SECURITY GLASS (This page I
      Unlimited
                                                                          22. PRICE
EPA Form 2220-1 (R«y. 4-77)   PREVIOUS EDITION is OBSOLETE

-------

-------

-------
 United States
 Environmental Protection
 Agency
Office of Air, Noise, and Radiation
Office of Air Quality Planning and Standards
Research Triangle Park NC 27711
Official Business
Penalty for Private Use
$300
                                            Publication No. EPA-450/3-81-015b
                                                                                                                               Postage and
                                                                                                                               Fees Paid
                                                                                                                               Environmental
                                                                                                                               Protection
                                                                                                                               Agency
                                                                                                                               EPA 335
                                                                          If your address is incorrect, please change on the above label;
                                                                          tear off; and return to the above address.
                                                                          If you do not desire to continue receiving this technical report
                                                                          series, CHECK HERE D ; tear off label; and return it to the
                                                                          above address.

-------