United States
Environmental Protection
Agency
Office of Air Quality
Planning and Standards
Research Triangle Park NC 27711
EPA-450/3-83-019b
October 1990
 Air
Polymer
Manufacturing
Industry -

Background
Information for
Promulgated
Standards
               Final
               EIS

-------

-------
                                           EPA-450/3-83-019b
       Polymer Manufacturing Industry
          Background Information For
             Promulgated Standards
          Emission  Standards  Division
     U.S.  ENVIRONMENTAL  PROTECTION AGENCY
         Office of Air and Radiation
Office of Air  Quality Planning  and  Standards
Research Triangle Park,  North Carolina  27711
                October  1990

-------
                               DISCLAIMER

This report has been reviewed by the Emission Standards Division, Office
of Air Quality Planning and Standards, Office of Air and Radiation,
Environmental Protection Agency, and approved for publication.  Mention
of company or product names does not constitute endorsement by EPA.
Copies of this report are available free of charge to Federal employees,
current contractors and grantees, and non-profit organizations -- as
supplies permit -- from the Library Services Office, MD-35, U.S.
Environmental Protection Agency, Research Triangle Park, N.C. 27711; or
may be obtained,  for a fee, from the National Technical Information
Services, 5285 Port Royal  Road,  Springfield, Virginia 22161.
                                  ii

-------
                      ENVIRONMENTAL PROTECTION AGENCY

                     Background Information and Final
                      Environmental Impact Statement
                    for Polymer Manufacturing Industry
 rack R.  rarmer-                                           /   (DateT
 Director,  Emission Standards Division
 U.S. Environmental Protection Agency (MD-13)
 Research Triangle Park,  North Carolina  27711

 1.    The emission standards will  limit emissions of volatile organic
      compounds from new,  modified,  and reconstructed polymer
      manufacturing facilities.   The standards implement Section 111 of
      the Clean Air Act (42 U.S.C.  7411),  as amended, and are based on
      the Administrator's  determination of August 21, 1979 (44 FR 49222)
      that  polymer manufacturing  plants contribute 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  Air 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:

      Fred Dinunick
      Standards  Development  Branch (MD-13)
      U.S. Environmental Protection Agency
      Research Triangle Park, North Carolina  27711
     Telephone:   (919) 541-5625

4.   Copies of this document may be obtained from:

     U.S. Environmental Protection Agency Library  (MD-35)
     Research Triangle Park, North Carolina  27711

     National Technical Information Service
     5285 Port Royal Road
     Springfield,  Virginia  22161

-------

-------
                             TABLE OF CONTENTS

 Chapter                                                            paqe

 1.0   SUMMARY                                                      i-1

       1.1   Summary of Changes Since Proposal                      l-l
             1.1.1   Definition of Intermittent Releases
                     to be Controlled                               1-2
             1.1.2   Monitoring of Vent Streams                     1-3
             1.1.3   VOC Weight Percent Exemption                   1-3
             1.1.4   Definition.of Concurrently Constructed,
                     Modified, and Reconstructed Affected
                     Facilities                                     1-4
             1.1.5   Reduced Testing Program for PET Facilities     1-4
             1.1.6   Applicable Standards and Uncontrolled
                       Threshold Emission Rates                     1-5
             1.1.7   Miscellaneous Changes                          1-6

       1.2   Summary of Impacts of Promulgated Action               1-6
             1.2.1   Environmental Impacts of Promulgated Action    1-6
             1.2.2   Energy and Economic Impacts of Promulgated
                     Action                                         1_6
             1.2.3   Other Considerations                           1-7

       1.3   Summary of Public Comments                             1-7


 2.0    BASIS  FOR  THE STANDARDS                                     2-1

       2.1   Coverage of Processes                                 2-1
       2.2   Definition of Affected  Facility                        2-5
       2.3   Model  Plant  Descriptions                               2-21
       2.4   Selection  of Basis  of the Proposed  Standards           2-27
             2.4.1   Model  Plant  Approach                            2-27
             2.4.2   Polystyrene  Processes                           2-30
             2.4.3   Polyethylene  terephthalate) Processes          2-37
       2.5   Presentation  of the Standards                          2-39
       2.6   Appropriateness and Applicability of Proposed
             Standards  to  Current  Polymer  Production  Processes      2-40
             2.6.1    Dilute VOC Streams                             2-40
             2.6.2    Low  Flow  Streams                               2-50
             2.6.3    Polyolefin Plants                              2-55
             2.6.4    Polystyrene Plants                             2-60
             2.6.5   Threshold Levels                               2-62
             2.6.6    Emergency Vent Stream Exemption                2-66


3.0   CONTROL TECHNOLOGY - PROCESS EMISSIONS                       3-1

      3.1   General                                                 -,
      3.2   Flares                                                 3.7
      3.3    Intermittent Emissions                                 3-12
      3.4   Condensers                                             3_jg
      3.5   Polyethylene terephthalate) Facilities               3-22

-------
                      TABLE  OF  CONTENTS  (continued)


Chapter                                                           Page

4.0   CONTROL TECHNOLOGY - EQUIPMENT LEAKS                        4-1

      4.1   General                                               4-1
      4.2   Assumptions and Effect on Emission Estimates          4-7
      4.3   Cost Effectiveness                                    4-18
      4.4   Applicability of Leak Definition                      4-27


5.0   MODIFICATION/RECONSTRUCTION                                 5-1


6.0   MONITORING REQUIREMENTS                                     6-1

      6.1   General                                               6-1
      6.2   Need for Flow Monitors                                6-3
      6.3   Thermocouples and Flame Monitoring                    6-8
      6.4   Incinerators                                          6-11
      6.5   Carbon Adsorbers                                      6-12
      6.6   Costs                                                 6-13
      6.7   Modification/Reconstruction                           6-14


7. a   TEST METHODS AND PROCEDURES                                 7-1

      7.1   General                                               7-1
      7.2   Flares                                                7-2
      7.3   Boilers                                               7-3
      7.4   Condensers                                            7-3
8.0   ECONOMIC IMPACTS, COSTS, AND COST EFFECTIVENESS
      CALCULATIONS                .                               8-1

      8.1   Economic Impacts                                      8-1
      8.2   Cost Impacts                                          8-3


9.0   REPORTING AND RECORDKEEPING REQUIREMENTS                    9-1

      9.1   General                                               9-1
      9.2   Flow Rates                                            9-2
      9.3   Pilot Light Flame                                     9-4


10.0  MISCELLANEOUS                                            .   10-1

      10.1  Definitions                                           10-1
      10.2  Other                                                 10-6

-------
                      TABLE OF CONTENTS (continued)
APPENDIX A  Flow Diagrams Illustrating Procedures for Determining
            Which Polypropylene and Polyethylene Emission Streams
            Are to be Controlled

-------

-------
                             LIST OF TABLES

Table                                                              Page

1-1   List of Commenters on Proposed Standards of Performance
      for VOC Emissions from Polymer Manufacturing Plants          1-9

4-1   Comparison of LDAR Programs                                  4-14
4-2   Comparison of Emission Reduction for Various LDAR
      Programs                                                     4.15
4-3   Fugitive Emission Reduction for the Sources in Polymers
      and Resins Model Plant                                       4-16
4-4   Fugitive VOC LDAR Program Costs                              4-21
4-5   Comparison of LDAR Programs by Equipment Component           4-22
                                 viii

-------

-------
                               1.0  SUMMARY

       On September 30, 1987, the U.S. Environmental Protection Agency
 (EPA) proposed standards of performance for volatile organic compound
 (VOC) emissions from certain facilities located in polypropylene,
 polyethylene, polystyrene, and polyethylene terephthalate) (PET)
 manufacturing plants (52 FR 36678) under the authority of Section 111 of
 the Clean Air Act.  A second Federal  Register notice was published on
 January 10, 1989,  (54 FR 890).   The purpose of the second Federal
 Register notice was to reopen the public comment period for purpose of
 allowing public comment on a new approach for determining which process
 emissions from polypropylene and polyethylene production would be
 subject to the proposed standards.
       Public comments were requested  on  the proposed standards as
 presented in both  Federal  Register notices.   There were 14 commenters on
 the September 30,  1987,  Federal  Register notice  and 11  commenters on  the
 January 10,  1989,  Federal  Register notice.   All  of the  commenters were
 either polymer manufacturers  or  trade  associations representing  polymer
 manufacturers.
       This  summary of comments and  EPA's  responses to these  comments
 serve  as  the basis for  the  revisions made to  the standards between
 proposal  and promulgation.  The  proposed  standards  from which  revisions
 are  identified  are those that reflect the September 30, 1987,  Federal
 footer  nt1ce and the January  10, 1989, Federal  Register notice.

 1.1  SUMMARY OF CHANGES SINCE PROPOSAL

      The proposed standards, which reflect the combined proposed
standards presented in the two Federal  Register notices, were revised as
a result of reviewing public comments.   The primary changes were made in
the following areas:
        Definition of Intermittent Releases to be  Controlled
        Monitoring Requirements  of Vent  Streams
        VOC Weight Percent Exemption
        Definition of Concurrently Constructed,  Modified,  and
         Reconstructed Affected Facilities
                                  1-1

-------
        Reduced Testing Program for PET Facilities
        Applicable Standard and Uncontrolled Threshold Emission Rates
         for Polystyrene and Poly(ethylene terephthalate) Plants

1.1.1  Definition of Intermittent Releases to be Controlled
      In the September 30, 1987, Federal Register notice, emergency vent
streams from certain polypropylene and polyethylene affected facilities
were proposed to be excluded from the control requirements for
intermittent emissions.  Public comments were received that suggested
that this exemption be broadened to include other emergency releases
from other facilities.  The EPA revised this exemption in the January
10, 1989, Federal Register notice.  In the January 10, 1989, Federal
Register notice, the EPA proposed that decomposition emissions be
exempted and all other intermittent emissions, including those that
occur as a result of attempts to prevent decompositions, be controlled.
Commenters on this latter notice expressed strong concern that the
proposed control of certain intermittent streams under the definitions
presented would be cost ineffective, would present severe safety
problems, and could force certain plants or technologies out of
business.  The commenters recommended that EPA reconsider which
Intermittent streams are to be controlled and indicated that the
September 30, 1987, proposed language was much closer to identifying
which intermittent streams are appropriate to control or not control.
      The EPA has reexamined both Federal Register notices, the
information available on intermittent releases,  and the nature and
causes of such releases.  The EPA has carefully considered the
commenters' comments and concerns.  The EPA agrees with the commenters
that the language in the September 30, 1987, Federal Register notice is
better at identifying the appropriate intermittent streams that are to
be controlled or exempted than the language in the January 30, 1989,
Federal Register notice.  Therefore, the rule for determining which
intermittent emissions are to be controlled as presented in the January
10, 1989, Federal Register notice has been revised for the promulgated
standard.  The revisions extend the exemption for decompositions to
other emergency type intermittent emissions.  For additional information
and details, please refer to Section 2.6.6, Emergency Vent Stream
Exemption, in Chapter 2.
                                   1-2

-------
  1.1.2  Monitoring  of  Vent  Streams
        In  the  proposed standards, flow  indicators  were  proposed to  be
  inserted  in each vent stream going  to  a  control device to  monitor  the
  presence  of flow.  This requirement was  intended  to  ensure that the vent
  stream was ducted  to  the control device.  This requirement has changed
  in the standard.   To  ensure that a  vent  stream is ducted to  a  control
 device, the standard  requires the submittal of a detailed  engineering
 report depicting the  piping used to vent the stream  to  a control device.
 Any valves in the  piping are to be  car-sealed opened or closed,  as
 appropriate, for ensuring that the  vent stream does  not bypass  the
 control device.  An owner or operator then elects to either:   (1)
 monitor the car-seals and valves on a regular basis and report  any
 changes in the position of the valves whenever they occur or (2) install
 a flow indicator immediately downstream of each valve that when opened
 would allow any portion of the vent stream to bypass the control and
 vent to the atmosphere.  Where flow indicators  are installed, the
 reporting  of all  times when flow is indicated is  required.   For
 additional information and  details,  please refer  to  Section 6.2, Need
 for Flow Monitors,  in  Chapter  6.

 1.1.3   VOC Weight Percent Exemptinn
       In the promulgated  standard,  all  affected facilities  may  use  the
 Individual  stream exemption  that allows individual vent streams with a
 VOC concentration of less than 0.1 weight  percent  VOC to be exempted
 from control.   Previously, this exemption was proposed  for  individual
 vent streams from modified or reconstructed affected  facilities  only,
 not for vent streams from new affected facilities.  This change  was
 based on information supplied by the commenters and reconsideration by
 EPA that there would be sufficient information available in any
 situation to determine when an owner or operator is intentionally
diluting a vent stream so as to take advantage of this exemption.  For
additional  information and details,  please refer to Section 2.6.1,
Dilute VOC Streams,  in Chapter 2.
                                  1-3

-------
 1.1.4  Definition of Concurrently Constructed.  Modified,  and
        Reconstructed Affected Facilities
       In the January 10,  1989,  Federal  Register notice,  it was proposed
 that affected facilities  that are constructed,  modified,  or
 reconstructed within a specified  period of time of each  other be
 considered  "concurrent."   By  considering  such  facilities  as concurrent
 the  continuous process emissions  within the same, weight  percent range
 from all  such concurrently constructed, modified,  or  reconstructed
 affected facilities  would be  added  together and the control/no control
 determination under  the new approach would then be made.   In the
 promulgated  standard,  the EPA has deleted  the  term "concurrent",  but has
 retained and expanded  its concept.  Under  the  promulgated  standard,  a
 vent stream  from  an  affected  facility in a polypropylene or polyethylene
 plant would  always be  subject to  potential- control  regardless  of when
 the  next affected facility was  constructed,  modified, or reconstructed.
 Once controlled to meet the standard, a vent stream would  not  be
 considered again  in  the control/no  control  decision for other  vent
 streams.  The EPA feels this  procedure  accurately  reflects  the new
 approach's generic decision making  process  that  is  dependent on  vent
 stream VOC concentration  and  annual emissions,  but  not on when a  vent
 stream becomes subject  to  these standards.   For  additional  information
 and  details,  please  refer  to  Section 10.1,  Definitions, in  Chapter 10.

 1.1.5  Reduced Testing  Program  for  PET Facilities
      Where  steam-jet ejectors  are  used to  provide vacuum to the
 polymerization reaction section at  PET plants, the proposed standards
 required daily testing  of  ethylene glycol  concentrations in either the
 cooling water of  the cooling tower that provides cooling water to  the
 vacuum systems servicing the polymerization reactors or in the liquid
 effluent exiting  these  vacuum systems.   It  is possible that some PET
 plants may be operated  in  such a manner that the ethylene glycol
 concentration would remain sufficiently below the standard  (0.35 or  6.0
weight percent of ethylene glycol) so that daily testing is not
 necessary to ensure that the standard is met.  Therefore,  the
 promulgated standard identifies the specific circumstances that are
 required for an owner or operator to conduct a reduced testing program.

-------
  In the promulgated  rule,  the  reduced  testing  program  requires  one  14
  consecutive day sample  program  (one sample  per  day) over  a  two-month
  period.  The final  rule also  allows an owner  or operator  to measure  an
  alternative parameter,  such as  chemical or  biological oxygen demand,
  provided that the measurement of such parameter allows the  owner or
  operator to demonstrate compliance with the ethylene  glycol
  concentration standard.   These  provisions are found in Section 60.564(j)
  of the final rule.

  1'1'6  Applicable Standards and Uncontrolled Threshold Emission Rates
       Several  commenters pointed out that the analysis for  polystyrene
 and PET plants failed to consider the potential  presence of moisture
 that could freeze up a condenser operating at the temperatures proposed.
 The EPA reevaluated the condenser system and recosted the control  to
 determine whether the proposed standards and uncontrolled threshold
 emission  rates needed to be revised for material recovery sections  in
 polystyrene plants and PET plants.   (For a discussion  on  the cost
 reanalysis, please refer to pages 2-32 to  2-37 in Chapter 2.)
       The results  of this  analysis  resulted  in a final standard of
 0.0036 kilograms  of total  organic compounds  (minus  methane and  ethane)
 per megagram of product  (kg TOC/Mg  product)  for  the material recovery
 section from polystyrene plants  or  a maximum outlet gas temperature from
 the final  condenser  of -25 degrees  Celsius(C) [-13 degrees  Fahrenheit
 (F)].  These values are the same as those that  were proposed.  The
 final  standard for the material  recovery section from  a PET  facility
 using  a dimethyl terephthalate (DMT) process  is  0.018  kg TOC/Mg product,
 which  is  an  increase from  the  proposed standard  of  0.0027  kg TOC/Mg
 product.  The corresponding maximum outlet gas temperature also
 increased,  from -24C  (-1TF) to +3C  (+37F).
      As  a result of the new analysis, the uncontrolled threshold
emission  rates for material recovery section from both types  of plants
also increased.  The uncontrolled threshold emission rate is 0.05 kg
TOC/Mg product for the material recovery section in a polystyrene plant
and 0.12 kg TOC/Mg product for the material recovery section  in a PET
plant using a DMT process.
                                   1-5

-------
1.1.7  Miscellaneous Changes
      Several change were made to the calculated threshold emission
(CTE) equations presented in the January 10, 1989, Federal Register
notice.  These changes reflect changes in the costing for catalytic
incinerators and flares.  In addition, several  definitions have been
revised.  The process section definitions were revised to provide
clearer guidance on the placement of equipment in each process section.
The definition to polypropylene was also clarified by including the term
"thermoplastic."
      One change was made to the application of Subpart VV to polymer
manufacturing plants since proposal.  Since proposal, it was learned
that certain polymer pumps are designed to purge polymer fluid from
bleed ports, thereby allowing small quantities of VOC emissions to
escape to the atmosphere.  These pumps must use the polymer fluid to
provide lubrication and/or cooling of the pump shaft.  The EPA believes
that the bleed ports from such pumps should be covered by Subpart VV if
they meet the definition of a leak based on an instrument reading of
10,000 parts per million or greater, but should not be covered based on
Indications of liquids dripping from the pump seal (as that definition
of a leak would otherwise be applied to the bleed ports).  Therefore,
the final rule contains a limited exemption for the bleed ports from
these types of pumps.

1.2  SUMMARY OF IMPACTS OF PROMULGATED ACTION

1.2.1  Environmental Impacts of Promulgated Action
      Environmental impacts of the proposed standards were described in
the September 30, 1987, Federal Register notice and in the January 10,
1989, Federal Register notice.  The revisions to the proposed standards
will have a minimal effect on the environmental impacts of the
standards.

1.2.2  Energy and Economic Impacts of Promulgated Action
      The energy and economic impacts of the standards are described in
Chapters 7, 8, and 9 of the proposal background information document
(BID Vol. I).  The changes made in the standards since proposal have
                                   1-6

-------
 essentially no effect on the energy impacts.  In addition, the costs and
 economic impacts for these standards have changed slightly since
 proposal.  Capital costs have increased for polystyrene and PET plants.
 This reflects revised condenser costs.  Annual costs were also revised
 to reflect revised unit price costs for natural  gas and electricity.
 The net effect of these changes is marginal.  Total nationwide capital
 costs in the fifth year following the promulgation of these standards
 from all four polymers is estimated to be approximately $4.3 to $4.5
 million (compared to $4.5 million at proposal) and annual  costs to be
 approximately $1.3 million (down from $1.4 million at proposal).
       The economic impacts of the promulgated standards are estimated to
 be essentially the same as for the proposed standards.   Adverse economic
 impacts would be minor.

 1-2.3  Other Considerations
       1-2.3.1  Irreversible and  Irretrievable Commitment of Resources.
 Section 7.5.2 of the BID for the proposed  standards  (Docket Item  III-B-1)
 concludes  that the standards will  not  result in  any  irreversible  or
 irretrievable commitment of resources.   It was also  concluded  that the
 standards  should  help to save resources  due to the  energy  savings
 associated  with  the  reduction in emissions.   These  conclusions  remain
 unchanged  since  proposal.
       1-2.3.2  Environmental  and Energy  Impacts  of Delayed  Standards.
 Delay in implementation  of  these standards would adversely  affect  air
 quality.   Lost emission  reductions  for each year the standards  are
 delayed  are estimated to  be  approximately  600  megagrams  per year
 (Mg/yr).  No  adverse  solid waste, water  pollution, or energy impacts are
 expected from delaying regulatory actions.

 1.3   SUMMARY OF PUBLIC COMMENTS

      Letters were received from 14 correspondents commenting on the
September 30, 1987, Federal Register notice and from 11 correspondents
commenting  on the January 10, 1989, Federal Register notice.  A public
hearing was requested on the September 30,  1987,  notice, but not on the
January 10, 1989, notice.  A list of commenters,  their affiliations, and
                                   1-7

-------
the EPA docket numbers assigned to their correspondence are given  in
Table 1-1 of this document.
      This document presents comments pertaining to the preamble and
regulation resulting from the proposed standards.  The comments from
interested parties and EPA's responses to those comments have been
categorized, and they are presented under the following topics:
        Basis for the Standards (Chapter 2)
        Control Technology - Process Emissions (Chapter 3)
        Control Technology - Equipment Leaks (Chapter 4)
        Modification/Reconstruction (Chapter 5)
        Monitoring Requirements (Chapter 6)
        Test Methods and Procedures (Chapter 7)
        Economic Impacts, Costs, and Cost Effectiveness Calculations
         (Chapter 8)
        Reporting and Recordkeeping Requirements (Chapter 9)
        Miscellaneous (Chapter 10)
      The procedure for determining which process emissions from
polypropylene and polyethylene affected facilities are to be controlled
is relatively complex.  Appendix A presents several  flow diagrams to
assist the reader in understanding this procedure.  The flow diagrams do
not contain all of the details associated with the procedure.  Thus, the
reader should refer to the final rule for additional  details.
                                   1-8

-------
                 TABLE  1-1.   LIST  OF  COMMENTERS ON PROPOSED
                 STANDARDS OF PERFORMANCE  FOR VOC EMISSIONS
                     FROM POLYMER MANUFACTURING  PLANTS
 Commenter and Affiliation     	Docket  Item No.

 L.G. Lund                                        IV-D-1
 HIMONT U.S.A.,  Inc.                              IV-D-40
 PO Box 1687
 Lake Charles, LA  70602

 L.J. Tabary, II                                  IV-D-2
 Copolymer Rubber and Chemical Corp.
 PO Box 2591
 Baton Rouge, LA  70821

 W.C. Dedeke                                      IV-D-3
 Union Carbide Corporation                        IV-D-42
 222 Pennbright Drive, Suite 109
 Houston,  TX  77090

 R.R. Kienle                                      IV-D-4
 Shell Oil  Company                                IV-D-48
 One Shell  Plaza
 PO Box 2463
 Houston,  TX  77001

 B.F. Ballard                                    IV-D-5
 Phillips  Petroleum Company                      IV-D-39
 Bartlesville,  OK  74004

 P.P.  Cash                                        IV-D-6
 Mobil  Chemical  Corp.                             IV-D-44
 211  College  Road East
 Princeton  Forrestal Center
 Princeton, NJ   08540

 6.E.  Matula                                      IV_0  7
 Texas Chemical Council
 1000  Brazos, Suite 200
 Austin, TX   78701

 G.V. Cox                                         IV-D-8
 Chemical Manufacturers Assoc.                    TV-n-
-------
                          TABLE  1-1.   Concluded
 Commenter and Affiliation	   Docket Item No.

 C.M.  Moffitt                                    IV-D-10
 Cain  Chemical
 Eleven Greenway Plaza
 Suite 2700
 Houston,  TX  77046

 A.T.  Roy                                         IV-D-11
 Allied Fibers                                   IV-D-41
 PO  Box 831
 Hopewell,  VA  23860

 P.  Meitner                                      IV-D-12
 E.I.  duPont de Nemours  &  Co.
 DuPont Building,  Room 70ISA
 Wilmington,  DE  19898

 J.C.  Edwards                                     IV-D-13
 Eastman Kodak Co.                                IV-D-47
 Kingsport,  TN  37662

 W.E.  Tessmer                                     IV-D-14
 International  Institute of Synthetic
  Rubber Producers, Inc.
 2077  South  Gessner Road
 Suite 133
 Houston, TX   77063

 C.T.  Gonzales                                    JV-D-46
 Exxon Chemical Americas
 PO Box 1607
 Baton Rouge,  LA  70821

S.J. Roadcap                                     IV-D-49
Chevron Chemical Co.
PO Box 5047
San Ramon, CA  94583
                                  1-10

-------
                       2.0  BASIS FOR THE STANDARDS

 2.1  COVERAGE OF PROCESSES

 Comment:
       Two commenters (IV-D-2, IV-D-14) were concerned over EPA's
 definition of polypropylene (PP), in which copolymers containing at
 least 50 percent propylene would be covered under the standard as
 proposed.  Commenter IV-D-2 requested EPA to reevaluate the statement
 that as long as the proportion of propylene used in the production of
 polypropylene copolymers is at least 50 percent by weight in the
 copolymer product,  the production processes used to manufacture both the
 polymer and the copolymer are essentially the same.  This commenter felt
 this assumption and the resulting definition of polypropylene would
 include inappropriately their copolymer production facilities that
 produced ethylene-propylene terpolymers rubber products,  which are
 considered "elastomers" under industry definitions.
       Commenter IV-D-14 stated that  several  of their member companies
 produce synthetic rubbers,  which contain  ethylene/propylene/diene
 monomers (EDPM)  in  the  range 50-75/25-50/0-12,  using processes entirely
 unlike  those  used for production of  plastics  composed of  polyethylene,
 polypropylene,  polystyrene  or  polybis(2-hydroxyethyl)terephthalate.   The
 commenter noted  that the  presence of the  diene  monomer at  low  levels  and
 the  mixing  of the ethylene  and propylene  units  in  the polymer  molecule
 result  in  physical  characteristics which  are clearly  those  of  a rubber
 rather  than a "thermoplastic."   Technically, the EPDM products are well
 above their glass transitions at normal service temperatures,  and they
 have a  very low  fraction  of  crystallinity at all temperature ranges
 relative to polyethylene  and polypropylene.  The commenter pointed out
 that none of  their member companies' EPDM production .facilities were
 included in the  surveys conducted by EPA prior to proposal of these
 rules.   In view  of the major differences between the polymerization and
 finishing operations of the EPDM rubber plants and those of the
thermoplastic resin  plants,  this commenter requested that  the proposed
standards for polyethylene and polypropylene apply to certain sources in
polymer manufacturing plants that produce copolymers consisting of at
                                  2-1

-------
 least 80 percent (rather than 50 percent) by weight of ethylene or
 propylene, respectively.
 Response:
       The intent of these standards is to apply to certain thermoplastic
 or thermoset resins.  The Agency agrees that the definition of
 polypropylene in the September 30,  1987, Federal Register notice would
 have subjected certain elastomer and synthetic rubber producers to these
 standards, because the definition did not contain the clarifying term
 "thermoplastic".  Therefore,  the Agency has revised the definition of
 polypropylene to include the  term "thermoplastic."
       One commenter suggested that  the definition of polypropylene and
 polyethylene copolymers be revised  so as to require at least 80% by
 weight of propylene and ethylene,  respectively,  before being subject to
 the standards.  This suggestion  was made to ensure exclusion of member
 synthetic rubber companies.   The Agency does not believe that such a
 revision is necessary because the definitions refer to thermoplastic
 resins,  which are not rubbers, and  thus synthetic rubber producers are
 already  excluded.
 Comment:
       One commenter (IV-D-2)  requested that their company be exempted
 from the effects of the proposed regulation until  further studies  are
 performed on synthetic rubber producers.   This commenter felt  that the
 current  regulation  would cover their facilities,  which produce products
 and polymers more appropriately  covered under a  synthetic rubber or
 rubber manufacturing industry standard,  not under the  currently proposed
 polymer  manufacturing standards.  This commenter stated  that the intent
 of these regulations is to cover thermoplastics  and  their manufacturers,
 and that there appears to be  a lack of documentation evidencing studies
 of the processes of rubber manufacturers  in  the  Agency's  published
'documentation.   The commenter also  noted  several  differences  between
 their processes  and those described in BID  Vol.  I  for  polypropylene
 production,  and  claimed that  their  processes  produced  a  relatively small
 amount of discharge and emitted  certain emissions  not  conducive to
 flaring.   Thus,  this commenter concluded  that given, in  their  view,  the
 relatively small  amount of discharge combined with the deficiency  of
 information  relative to the synthetic  rubber manufacturing  industry,  any
                                   2-2

-------
 regulation of their facilities without a more comprehensive  study  that
 includes the specific nature of design and manufacture of synthetic
 rubber producers is premature and could harm any further expansion
 projects now under consideration.
 Response:
       The intent of these standards is to cover certain producers of
 thermoplastic or thermoset resins and copolymers of these resins that
 are in turn themselves thermoplastic or thermoset resins.  The Agency
 does not intend for these standards to cover synthetic rubber producers,
 including manufacturers of thermoplastic or thermoset elastomers, such
 as ethylene-propylene copolymers and terpolymers that are elastomers or
 rubbers.   Resins are thermoplastic or thermoset  polymers  that are
 essentially synonymous with the  term plastic.   Elastomers are also
 thermoplastic or thermoset polymers,  but  are  capable of returning to
 their initial  form following deformation.   The Agency believes that the
 revision  to the  definition of polypropylene limits  the scope  of these
 standards appropriately.   Synthetic  rubber  and elastomer  producers,
 however,  may  still  be  regulated  in the  future under this  standard or a
 new standard  should the Agency decide such  regulation  is  warranted.
 Comment!
       One commenter (IV-D-1)  stated that there are  more alternative
 polypropylene technologies  than the two listed in BID  Vol. I  and  in  the
 proposed  regulations.  This  commenter assumed that  the new source
 performance standards  (NSPS)  should be applicable to all  polypropylene
 processes.  The  commenter then concluded that the NSPS should  be  amended
 to  include, at least, the "bulk (liquid-phase) polymerization
 technology."  The commenter  stated that bulk plant technology, although
 in  principle a liquid-phase polymerization process,  is unlike  the
 "traditional" slurry process and hence,  conclusions drawn on the  basis
 of  the "slurry process" do not apply directly to the bulk process.
      Commenter  IV-D-45 also noted that in order to determine
 applicability dates for affected facilities  (using Tables  1 and 2 of the
 proposed regulation), one must still  have a  clear understanding of which
 "production process" applies (the applicable model  plant)  and  the
definition of "process section" (which equipment  falls into which
process section), especially where plants  have commenced construction,
                                  2-3

-------
 modifications,  or reconstruction between  the two proposed regulation
 dates.
 Response:
       The  commenter is  correct  in assuming  that  the  standards  proposed
 in the  September 30,  1987,  Federal  Register notice were  to be  applicable
 to all  polypropylene  processes.   At that  time, the Agency understood
 that  all such processes could be described  as either liquid phase  or gas
 phase processes  and that the model  plants described  in BID Vol.  I  were
 reasonable representations  of those processes upon which  to base
 standards.   Comments  received on the September 30, 1987,  Federal
 Register notice,  however, indicated that  the polypropylene model plants
 may not be adequate representations of all  processes.  As  a result  of
 this  and other comments,  the Agency undertook an analysis  to examine
 alternative  ways  to determine which process emissions from polypropylene
 (and  polyethylene)  plants should be controlled.  The  results of this
 analysis were presented  in  a January 10,  1989, Federal Register notice
 for public comment.   The  approach selected by the Agency and
 Incorporated into the final rule is  independent of the particular
 process technology  used  to  produce  polypropylene or polyethylene.   Thus,
 the non-representativeness  of the polypropylene,  "liquid phase model
 plant presented in  BID Vol. I as  it  applies to the bulk process is  no
 longer a concern.
      For  polypropylene  facilities  that are constructed,  modified,   or
 reconstructed after September 30, 1987, and on or before January 10,
 1989, the owner or operator of such facilities must still determine
which process--!iquid phase or gas phasehis or her facility falls
 under for purposes of determining the affected facilities.  The final
rule requires an owner or operator to select one  of the production
processes listed in Table 1 of the final  rule to  apply to his or her
facility (see Section 0.560(b)(l)(ii)  of the final  rule).  The
determination of which emissions from these  affected  facilities would be
controlled is made using the new approach, which  is  independent of
process type, rather than the control/no  control  decisions that were
based on the model plant.  However,  an  owner or operator  can still  use
the uncontrolled threshold emission rates  proposed  in the September 30,
                                  2-4

-------
  1987,  Federal Register notice to  exempt  these  affected  facilities  from
  control,  if he or  she so elects.

  2.2    DEFINITION OF AFFECTED FACILITY

  Comment:
        One commenter (IV-D-13) stated that the  selection of process
  sections as affected facilities results  in unreasonable cost and gives
  unfair trade advantage to patent holders of these processes which the
 Agency selected as model  plants.  The commenter stated that the proposed
  standards will require companies with existing polymer plants to divide
 up their plants into process sections and process lines, creating a need
 to develop expensive new accounting and recordkeeping methods and
 procedures to determine if modifications occur.  The commenter claimed
 that the Agency recognized this  problem in the preamble (52 FR 36682)
 when evaluating whether the affected facility should be individual
 process emission  points,  in "attempting to maintain  records of which
 process or equipment were  subject  to the standard ...  .  "   The commenter
 stated that  plants  are  not  designed,  estimated, justified  and built,  and
 cost centers  are  not usually set up by  process  sections.   A large
 polypropylene  or  polyethylene polymer plant may have as  many  as  10  to  15
 affected  facilities,  according to  the commenter,  and over  the  course of
 a year, many  small  changes  may be  made  to these plants.  To evaluate
 whether a modification that  has  been  made falls under  the  proposed
 rules,  the commenter stated  that completely new accounting  systems must
 be developed and  implemented.  For  existing polypropylene  and poly-
 ethylene  plants,  the commenter believes a more  reasonable  choice would
 be to designate each process line as  the  affected facility.  The com-
menter  stated that  a problem recognized by EPA  concerning the choice of
process units, i.e., "a process  line cannot be  determined clearly ...  ."
 (52 FR  36683), is even more of a problem with the selection of process
section as the affected facility.  Specifically, the commenter pointed
out, material recovery is being utilized more and more by the industry
as a means of reducing overall  emissions.  According to the commenter,
in the model  plants studied by the  Agency, material  recovery occurred
primarily after the reaction section, while modern plants continue the
                                  2-5

-------
 recovery  of  raw materials  throughout  the  process  line, wherever  it  can
 be  done economically.  The commenter  concluded  that  the  standards need
 to  be  revised  to a  process line concept.
 Response:
       The main  considerations  in selecting the  definition of  affected
 facility  is  the consideration  of the  application  of  best demonstrated
 technology and  the  degree  to which replacement  equipment is brought
 under  the standards.  As a result, narrower definitions  are preferred.
 This preference can be overcome if analysis concludes that a  broader
 designation  would result in greater emission reductions or avoid
 unreasonable impacts (i.e., costs, energy, or other  environmental
 impacts).  The  commenter's main point for changing the definition of
 affected facility is the need  by industry to develop and implement
 completely new  accounting  systems to  track costs  for as many  as 10 to 15
 affected facilities in a large polypropylene or polyethylene  plant.  The
 commenter also  believes that defining process sections is even more
 difficult than  defining a  process line and refers to changing practices
 of material  recovery in the industry.
      The Agency disagrees with the commenter on both points.   The
 imposition of new regulations on the  industry would likely require some
 plants to develop new accounting systems to track modifications and
 reconstructions whether the definition of affected facility is "process
 section" or  "process line."  Furthermore,  under either definition, the
 owner or operator would still  need to track all  changes.   The only
 difference is the number of affected facilities that would be tracked.
 On this issue,   the commenter claims that 10 to 15 affected facilities is
 too many.   The Agency simply disagrees with this statement.   The Agency
 rejected designating each individual  emission point as an affected
 facility because a typical  plant may have  as  many as 40 individual
 process emission points;  a large plant may have substantially  more.   In
 terms of process sections,  the commenter notes that a large  plant may
 have 10 to 15 process sections; a typical  plant would likely have fewer.
The number of process lines at any plant will  always be less than the
 number of process sections; but,  the Agency is not convinced that at
 large plants  10 to 15 affected facilities  is  unreasonable.
                                  2-6

-------
        With regard to the second point, the commenter claims  that  the
  definition of "process section" is even more of a problem than that of  a
  "process line."  The Agency again disagrees.  The example provided by
  the commenter may not match the material recovery section of the  model
  plant, but that is irrelevant to the definition.  The concept of
  material recovery is the key aspect of identifying the material recovery
  process section or material recovery process sections at a plant.
  Further, the Agency identified a major conceptual  problem with trying to
  define a "process line," where equipment was shared between two
  otherwise distinct process lines.   The commenter offered no suggestions
  as to how to deal  with that problem.
        In summary,  the Agency finds no reason to change the definition  of
  affected facility for process  emissions.
  Comment:
       'Several  commenters  (IV-D-43/IV-D-50,  IV-D-45,  IV-D-47)  noted that
  the current  proposed  regulation  is  based  on  identifying  "process
  sections"  and  continuous  or intermittent  emissions  in  "affected
  facilities," and  that  identifying  these process  sections  in  affected
  facilities requires making  base  assumptions  that are critical for
  consistent interpretation throughout  the  industry.  Commenters  IV-D-47
  and IV-D-50 stated that plant technical and management personnel
  continue to experience problems  in discerning the specific requirements
  of  the proposed rules applicable to each process section  and  affected
  facility.  These commenters encouraged EPA to further clarify its
.  definitions of "process section" and "affected facility"  to ensure
  proper compliance with the regulation.
       Two other commenters (IV-D-8, IV-0-9) were concerned over
  potential confusion that the definitions of the various process sections
 may create in the placement of particular pieces of equipment.
 Commenter IV-D-9 stated that the use of "process section" as the
 definition, of affected facility may likely result in over or under.
 regulation if a piece  of equipment  is  misclassified  as  to the section it
 belongs.   Commenter IV-D-8 noted  that,  as  currently  defined,  "Product
 Finishing Section" means  "the equipment that  treats,  shapes,  or  modifies
 the polymer or  resin to  produce the finished  end product  of the
 particular facility.   Product finishing equipment may  accomplish
                                   2-7

-------
 extruding, cooling and drying, blending,  additives introduction, curing,
 or annealing."  The commenter pointed out that for many of the processes
 covered by the regulation,  the product finishing section should consist
 of the extruder and the pelletizer.   Equipment downstream of the
 pelletizer includes check hoppers,  blenders,  and silos,  which do not
 "treat, shape, or modify the polymers."   They are used to transfer
 product for quality control, general  operations,  and  shipping,  and
 according to the commenter  should  be  a part  of the product storage
 section,  since there is no  intentional modification of the
 characteristics of the polymer as  specified  in the "product storage
 section"  definition.
       Both commenters  noted  that the  use  of  the terms  "blending"  and
 "blenders" can confuse the delineation of product  finishing and  product
 storage section.   In one of  the commenter's  process, check hoppers
 collect polymer pellets as they leave  the dryer and then  blenders  are
 used  to physically mix the pellets in  order to obtain  a homogeneous
 mixture over time.   This commenter points out' that  these  check hoppers
 and blenders meet  the  definition of equipment  in the product  storage
 section since there  is  "no intentional modification of the
 characteristics  of the  polymer."  However, the  commenter  argues, a
 regulator  could  easily  place the blenders  under the product finishing
 section because  this section is defined as "equipment  [that]  may
 accomplish extruding,  cooling  and drying,   blending, additives
 introduction,  curing, or  annealing."  Both commenters note  that
 "blending" in  this case  is clearly the mixing of 2 or more compounds
 (I.e. polymer  and  additives) in order to produce a product that has
 different properties.  Commenter IV-0-9 pointed out that  if the
 emissions from the product finishing section were subject to controls,
 controlling  these very dilute VOC streams   (vents from blenders) would
 not be  practical or economical  and believes there should be a way to
prevent this situati-on from occurring.
      Both commenters suggested that this  problem may be addressed by
adding  a statement under the definition of product finishing which
states that  "blending does not mean the physical mixing of polymer
pellets in order to obtain a homogeneous mixture over  time."  However,
Commenter IV-D-9 stated that this  would eliminate their concern for
                                  2-8

-------
 their blenders, but still would not solve the overall concern of
 potential over or under regulation of vents from equipment due to
 misclassification of the equipment within a process section.  Therefore,
 the commenter urged the Agency to add a general provision to the
 regulation that would allow a company to petition the Administrator to
 show that controlling the emissions from a vent within an affected
 process section is not technically feasible or not economical.  Since
 some form of incineration or carbon adsorption are typically the only
 alternative controls available, the commenter suggested that a concept
 like the Total  Resources Effectiveness (TRE)  index used in the proposed
 SOCMI distillation NSPS would be appropriate  as a test for cost
 effectiveness,  and would eliminate much of industry's  concerns
 associated with trying to delineate process  sections.
 Response:
       The  Agency agrees that some  of the  process  section  definitions
 proposed could  create  some  confusion  as to where  particular  pieces  of
 equipment  are placed.   Therefore,  in  the  definitions  in the  final  rule,
 the Agency has  delineated more  clearly where  a  process  section  ends and
 another  process  section  begins.   In  the specific  example  cited  by  the
 commenters,  the  Agency  agrees that  the product  finishing  section ends
 after  the  last  piece of  equipment  that modifies the characteristic of
 the polymer  and  that the physical mixing  of the pellets to obtain a
 homogenous mixture is not part  of product finishing, provided such
 equipment  follows the last piece of equipment in the product finishing
 section.
      A  commenter still felt that some "misclassification" may still
 occur.   The commenter appears to define "misclassification" in terms of
 some streams being required to be controlled that are either not
 technically feasible or that are uneconomical  to do so.  The new
 approach for determining which process streams are to be controlled,
which was presented in the January 10,  1989,  Federal  Register notice,  is
designed to take these items into consideration.  All  process sections
that become affected facilities  either on, before,  or after January 10,
1989, can use this approach  and  thereby exclude from  control  emissions
that are excluded from control when the new approach  is applied.
                                  2-9

-------
 Comment:
       One commenter (IV-D-5)  stated that it is possible for equipment to
 be designed to produce more than one of the polymers covered by the
 proposed  standards.  The commenter pointed out that the preamble
 indicates that merely switching production from one type of polymer to
 another would not be classified as a "modification" as  long as the
 original  equipment was designed to accommodate both products.   The
 commenter requested clarification as to which  category  of polymer
 process (and thus standards)  would be applicable to new facilities that
 are designed to produce more  than one polymer  using the same process
 equipment.
 Response:
       The situation described by the commenter had  not  been envisioned
 by the Agency when the standards were proposed on September 30,  1987.
 Where  a new facility is constructed,  modified,  or recorvstructed  after
 January 10,  1989,  the  situation  described  by the commenter  does  not
 exist  because the same procedure is  applied for  determining control
 regardless  of whether  a low density  polyethylene  (LDPE),  high  density
 polyethylene (HOPE), or PP product  is  being produced.   However,  if a
 facility  is  built after September 30,  1987, and cm  or before January 10,
 1989,  the standards  to be met  could  be  different where  that facility is
 designed  to  produce  more than  one polymer  (e.g., HOPE and LDPE)  as
 different process  sections (and  their emissions) were designated  as
 affected  facilities  depending  upon whether HOPE or  LDPE was  being
 produced.  As  presented  in the January  10, 1989, Federal Register
 notice, the  new approach did  not  guide  an operator-  or owner in
 determining  which  process sections are  to be considered affected
 facilities for those process sections constructed, modified, or
 reconstructed before January 10,  1989, where two types of polymers
 (e.g.,  LDPE  and HOPE)  are produced in the same equipment.
       In addition, the uncontrolled threshold emission rate  to be
 applied in exempting from control emissions from an affected facility
differ depending on whether HOPE or LDPE is being produced.  The
January 10,  1989,  Federal Register notice allowed owners or operators
the option to exempt from control emissions that under the new approach
would require control but could be shown under the model plant approach
                                  2-10

-------
 to be exempt from control through the use of the uncontrolled  threshold
 emission rate exemption.

       The Agency considered the following options to clarify which

 process sections and uncontrolled threshold emission rates  are  to be

 used where a facility is designed to produce more than one  polymer  using
 the same process equipment:

       Option 1.   Apply each model plant's affected facilities  and
                   uncontrolled threshold emission rates as  each polymer
                   is produced.

       Option 2a.  Designate as affected facilities all  process  sections
                   (and the emissions) that are in either model  plant and
                   use the most restrictive uncontrolled threshold
                   emission rate where there are more than one for a
                   process section.

       Option 2b.  Designate as affected facilities all  process sections
                   (and the emissions) that are in either model  plant and
                   use the least restrictive uncontrolled threshold
                   emission rate where there are more than one for a
                   process section.

       Option 3a.  Designate as affected  facilities only those process
                   sections (and the  emissions)  that  are common  to both
                   model  plants and use  the most restrictive uncontrolled
                   threshold  emission rate.

       Option 3b.  Designate  as affected  facilities only those process
                   section (and the emissions)  that are  common to both
                   model  plant  and  use the  least restrictive uncontrolled
                   threshold  emission rate.

       Option  4.    Provide an applicability date for  all  "hybrid"
                   facilities of January  10,  1989.

       Option  5.    Have the owner/operator  select  one model  plant (pre-
                   sumably  the  one  that most closely matches  the  hybrid
                   facility)  for the  purpose of  determining  the  affected
                   facilities with  a  September 30,  1987  applicability
                  date and use  the uncontrolled threshold emission rates
                   for those process  sections as identified  in the
                  September 30, 1987  Federal Register notice.

Of these options, the Agency has selected Option  5.  Option  1 could  lead

to a process section being an affected facility on a "part time" basis,

which could lead to control devices being turned on and off  depending on

which polymer is being produced.  Further,  it would be an inefficient
                                  2-11

-------
 use of control  devices.   For these reasons,  the Agency rejected
 Option 1.
       Option 2  would be  to identify as  affected facilities all  of the
 process sections identified in  the September 30,,  1987,  Federal  Register
 notice as  affected  facilities for  either  polymer.   Option 2 may be
 overly broad in that more process  sections are  identified as affected
 facilities than would have been identified if the  facility of concern
 had been specifically examined.  While  the Agency  feels  that the new
 approach would  not  result the control of  inappropriate  streams,  and  each
 affected facility could  still be exempted from  control  by using  the
 uncontrolled threshold emission rate cutoff  for that process section,
 the Agency feels this option may treat  owners or operators  of such
 facilities unequally compared to others.
       Option 3  would be  to identify those process  sections  as affected
 facilities on the basis  of being identified  as  affected  facilities for
 both types of polymers.   The Agency feels that  this approach is
 unnecessarily restrictive  in that  it potentially excludes  from control
 emissions  from  process sections that would have been identified  as
 affected facilities  had  the facility of concern been specifically
 examined.
       Further,  under both  Options 2 and 3, the threshold  emission  rate
 to  be  applied would  need to be determined where two rates have been
 identified  for  the same  process section.  Should the higher  of the two
 rates  be used or the  lower?  Without specific data on each such
 facility,  the Agency  has no technical basis for selecting one or the
 other.
      Option 4 essentially sidesteps the issue by allowing all such
 facilities to have a January 10, 1989,  applicability date.   The Agency
does not believe this is necessary.  Option 5 allows the owner to most
closely match one of the model  plants to his  or her facility.  This
 incorporates a'Concept from a similar comment where a modified HOPE,
solution process now produces LDPE, but  the emission characteristics
remain similar to when HOPE was  being produced.   Thus,  the Agency has
selected Option  5 (see Section  60.560(i) of the  final  rule).
                                  2-12

-------
 Comment:
       One  commenter  (IV-D-6)  requested  clarification  on  what  equipment
 is to be included  in  the material  recovery  section  for a continuous
 process polystyrene unit.  The commenter pointed out  that  "material
 recovery section"  is  defined  as equipment which recovers unreacted or
 by-product material from any  process section;  "process sections"  are
 defined such that  product finishing is  separate from  material recovery;
 and "recovery system" is defined as units such as condensers...which are
 recovering VOCs.   From this,  the commenter concluded  that the affected
 facility of a polystyrene unit under this regulation  would only include
 the condensers and associated equipment for cooling and  recovery, and
 would specifically not include the devolatilizer or preheater, which is
 part of the process for product purification.
 Response:
       As noted in an earlier response,  the Agency has clarified  more
 specifically the definitions of the process sections to determine which
 process  equipment falls  into which process  section.   With regard to the
 particular  equipment referred to  by this commenter,  the deyolatilizer
 was specifically shown in  BID Vol.  I as  part of the  material  recovery
 section.  Although  the preheater was not specifically  shown  in BID Vol.
 I,  1t  Is associated with the  equipment used  to  "purify" the  product.
 None of  the process sections  specifically  includes the concept of
 "product purification."  Therefore,  such equipment needs  to  be allocated
 either to a material recovery  section or to  product  finishing.   In the
 polystyrene industry,  the recovery  of unreacted styrene is such  a
 critical part of  the economic  aspects of the process that equipment used
 to  separate  the unreacted styrene from the product through the final
 recovery of the styrene were considered  part of the material recovery
 process section.  Thus, in this case, the devolatilizer and preheater
would be part of the material  recovery section.   For other processes,
where no material recovery is  practiced  after the polymerization
reactor,  equipment that,  in a  sense, prepares the product for product
finishing (e.g., removes  unreacted or byproduct  material  from the
polymer produce) would be included in the product finishing process
section.
                                  2-13

-------
 Comment:
       One  commenter  (IV-D-6)  requested  that  vapor  recovery  systems
 servicing  emergency  vents  in  continuous  polystyrene  plants  be  excluded
 from  the definition  of material  recovery  section.  The  commenter  pointed
 out that some  of  their facilities have  emergency vent vapor recovery
 systems  (add-on pollution  control facilities) which  attempt to reduce
 these intermittent emissions  by  spray water  contact.  For other design
 reasons, the continuous vent  from the material recovery system serving
 the process devolatilizer  is  also passed  through the same system.   There
 can also be material recovered from the surge vessels in this  system,
 although most  of  the material recovered will likely be  from emergency
 vents.  The commenter then stated that a  strict interpretation  of the
 definition of  material recovery  section might subject emergency vapor
 recovery units of this type to the performance standards.   The  commenter
 claimed that this would not be technically or economically  feasible  due
 to the very large gas volumes and peak rates intermittently  incurred.
 Response:
      The standards for polystyrene facilities apply only to continuous
 emissions and  not to intermittent vents.  The emergency vapor  recovery
 systems described by the commenter are not part of the material recovery
 process section considered in the analysis as they are designed
 primarily to control  intermittent vents that occur due to
 overpressurization of process vessels relieving through rupture disks.
 Such  systems and the intermittent emissions that are emitted to them are
 not subject to these standards.   However, continuous vents from the
material recovery system,  which  includes the process devolatilizer,  are
 subject to these standards, regardless of whether or not they pass
through an emergency vapor recovery system.   For new facilities, it is
 up to the owner or operator to decide on whether or not to route such
 streams through emergency vapor  recovery systems.   For existing
facilities, the Agency has determined that where such continuous
emissions are routed  through an  existing emergency  vapor recovery
system, control to the standard  level  is required  when the existing
emergency vapor recovery system  undergoes modification,  reconstruction,
or replacement.  Of course, if such  continuous  emissions are below the

-------
  uncontrolled  threshold  emission  rate,  then  no  additional  control  is
  required.
  Comment:
       One commenter  (IV-D-11)  stated that the  proposed  regulation  was
  unclear as to which  process  section a  second stage  reactor would  fall.
  The commenter requested that the proposal be modified to  clearly
  indicate that second stage reactors in polyethylene wax manufacturing
  facilities are to be considered a part of the  "polymerization reaction
  section" and not the "product  finishing section."  The commenter pointed
  out that the model plant used  as the basis of  the proposed standard uses
  a single fluidized bed reactor to trigger the  desired chemical reaction.
  The commenter noted, however,  that at their polyethylene wax facility
  one of the two methods used employs a second stage reactor in series
 with the first reactor to produce a product that is emulsifiable in
 water.  The commenter states that in this process a portion of the
 molten homopolymer from the initial  reactor is  diverted to the second
 reactor where the polymer is oxidized  under elevated temperature and
 pressure.   According to the commenter,  the reaction causes a  breakdown
 of the molecule  and  the addition  of functional  polar groups  including
 carboxyls,  alcohols,  aldehydes  and  ketones,  and the molten oxidized
 material  leaving  the reactor is then sent  to various finishing steps.
 It is  on  this  basis  that the commenter  recommends that  second stage
 reactors  used  to  chemically modify  the  material it  receives be
 classified  in  the polymerization  reaction  section.
 Response:
       The Agency  disagrees  with the commenter.   Information provided  by
 the commenter  clearly indicates that there is no  polymerization
 occurring in the  second  stage reactor.  Rather, the commenter  clearly
 indicates that the polymer  undergoes a chemical reaction (oxidization)
 and then proceeds to  (other) product finishing  steps.  The oxidation  in
 the second stage  reactor falls  within the definition of "product
 finishing."  (This comment arose as a result of the September 30, 1987,
 F^dera1 Register notice.  The commenter was concerned that if this
reactor were considered part of the product finishing section, the cost
of control,  based on which streams would require control  under the
September 30,  1987, Federal  Register,  would be unreasonable.   This
                                  2-15

-------
 concern  arose,  in  part,  because  of"the  model  plant  approach originally
 used  to  develop these  standards.   Under the  new  approach,  which was
 presented  in  the January 10,  1989,  Federal Register notice,  the
 determination of control/no control  is  based  on  generic  emission stream
 characteristics.   This new approach  covers the commenter's  situation  in
 a  more adequate manner and enables  a more reasonable  control/no control
 decision to be  made.  As stated  in  their comments on  the January 10,
 1989, Federal  Register notice  (see  Docket Item IV-D-41), the  commenter
 agreed that their  concern has  been  alleviated.)
 Comment:
      One  commenter  (IV-0-6) stated  that there are  no  intermittent  (non-
 emergency) vents from the product finishing and  product storage process
 sections in high pressure, low density  polyethylene plants.   The
 commenter  pointed  out that continuous emissions  from these  two  process
 sections were not  proposed for control.  The commenter felt,  therefore,
 that these sections  should not be considered as  affected facilities.
 Response:
      The commenter made this comment in response to the September  30,
 1987, Federal Register notice.  In the January 10,  1989, Federal
 Register notice, a new approach for determining which process emissions
 from all  polypropylene and polyethylene plants would be subject  to
 control.   The new approach encompasses all  emission streams and  process
 sections in these types of polymer plants as  an  integral part of the new
 approach.  It may be possible that certain  processes do not have one
 type of emissions from certain process sections or lack one of  the  basic
 process sections.  It is not the intent of  the new approach to  identify
 such specific situations, thereby potentially limiting its
 applicability,  if not now, then in the future as  processes change.
 Therefore,  the Agency has retained as designated  affected facilities all
process sections and emissions from polypropylene and polyethylene
plants,  as was presented in  the January 10,  1989, Federal  Register
notice.
Comment:
      One commenter (IV-D-44)  stated that the model  plant approach,
contained in the September 30,  1987, proposal,  excluded certain process
sections  from being affected  facilities  for both  continuous and
                                  2-16

-------
 intermittent emissions, and that the January 10, 1989 proposal  appears
 to provide an exemption for only facilities exempted by Table  1.
 According to the commenter, it would be unfair to penalize projects
 which commenced construction, modification, or reconstruction  under this
 September 30, 1987, guidance and on or before the January 10,  1989,
 proposal, which were not designated as affected facilities in  the
 September 30, 1987 proposal.  The commenter then stated that the Agency
 recognized (54 FR 905) that certain emissions and process sections not
 required to be controlled under the standards proposed on September 30,
 1987, may be required to be controlled under the new approach  and
 therefore, the Agency proposed to resolve this potential  compliance
 problem by proposing a new applicability date (i.e., January 10, 1989)
 for those facilities that would have been excluded  under the original
 proposal, but subject under the new approach.   The  commenter recommended
 that the model  plant approach (September 30,  1987,  proposal)  should be
 the governing standard for polypropylene and  polyethylene construction,
 modification, reconstruction projects  which can  be  shown  to  have
 commenced after September 30,  1987,  but  on  or  before January  10, 1989.
 Response:
       The January  10,  1989,  Federal  Register  notice  does  what the
 commenter recommends  in  Table  2  of  the regulation portion  of  that
 notice.   Table  2 lists all  the  emission  and process  sections  that were
 excluded  as  affected  facilities  in  the September 30,  1987, notice and
 applies  a January  10,  1989,  applicability date to the'se emissions and
 process  sections.   Thus,  as  indicated  in the January  10,  1989,  Federal
 Register  notice and as provided  for  in the  final rule, only those
 process  sections identified  in the September 30, 1987, Federal  Register
 notice are affected facilities when  constructed, modified, or
 reconstructed after September 30, 1987, and on or before January 10,
 1989, for the appropriate continuous or intermittent emissions.   It
 should be noted that the procedure for determining control or no  control
 of the emissions from these process  sections is the "generic" or  new
 approach.  The uncontrolled threshold emission rates proposed in  the
September 30, 1987, Federal Register notice, however, can still  be used
for these affected facilities, if an owner or operator so elects, to
exempt individual  process sections from control.
                                  2-17

-------
 Comment:
       One commenter (IV-D-40)  stated  that  with the development of their
 new processes since the model  plant concept  was first proposed several
 years ago,  the process  sections  may not  always be a perfect match with
 the old definitions,  especially  as regards the relative weight of VOC
 emissions from corresponding sections.   The  commenter then  noted that
 from their  point  of view,  they do not  foresee  any major difficulties in
 applying  the  new  "generic  rules" to any  of their  existing processes, or
 future modifications  thereof.
 Response:
       The Agency  agrees with the commenter that the  emission
 characteristics identified  in  BID Vol. I for the  process sections in the
 model  plants  may  no longer  be  representative of all  current  stream
 characteristics in  the  polypropylene and polyethylene  segments.   The new
 approach  for  determining which process emissions  are  to be  controlled
 was  designed  to take  this problem into account.   The Agency still
 believes  that  the five  generic process sections identified  under the
 model  plants  remain a useful and representative way to describe  the
 industry  and  upon which  to  apply the standards.
 Comment:
       Two commenters  (IV-D-44,  IV-D-50) stated that the two columns
 marked  "continuous" and  "intermittent" should be dropped from Table  2 of
 the  regulation portion  in the January 10,  1989, Federal Register  notice
 because the definition of affected facility no longer differentiated
 between continuous and  intermittent sources.   Commenter IV-D-50  stated
 that they do not see the significance of the Table 2 designation  of
 continuous and intermittent sources in process sections.  This commenter
 noted  as an example that Table  2 does  not indicate a continuous emission
 source in a material recovery section  of a high density polyethylene
 liquid phase slurry process.  This  commenter then asked:  Does this mean
 that if such sources should occur and  meet all  of the criteria which
would require control, that they would not actually require  control
because Table 2 does not mark them  as  an  affected facility?
Response:
      The commenter is correct  to state that  for polypropylene and
polyethylene processes the  affected  facilities  are the same  for
                                  2-18

-------
 continuous and intermittent emissions.  However, the distinction in
 Table 2 of the regulation portion of the January 10, 1989, Federal
 Register notice must be retained because it identifies those process
 sections and emission combinations that have a January 10, 1989,
 applicability date.  This is necessary for affected facilities
 constructed, modified, or reconstructed after September 30, 1987, and on
 or before January 10, 1989.   For example,  if a new polypropylene, liquid
 phase plant was constructed  after September 30,  1987,  and on or before
 January 10, 1989, the plant  would have two applicability dates for its
 process sections:  September 30, 1987,  applicability date for the raw
 materials preparation, polymerization reaction,  material  recovery,  and
 product finishing section for continuous emissions  and the
 polymerization reaction section for intermittent emissions; and a
 January 10, 1989, applicability date for the- raw materials preparation,
 material  recovery,  product finishing,  and  product storage process
 sections  for intermittent emissions,  and the  product storage  process
 section for continuous emissions.   As  explained  in  January 10,  1989,
 Federal Register  notice (54  FR  905),  this  distinction  was necessary
 because certain emissions and process  sections not  required to  be
 controlled  under  the  standards  proposed  on  September 30,  1987,  may be
 required  to be  controlled under the  new  approach.   Thus,  the Agency has
 retained  the distinction  in  the final rule.  The  Agency,  however, has
 replaced  Table  2  from the January  10, 1989, Federal  Register notice with
 Section 60.560(b)(l)(i) and  has  revised  the reference to  Table  1.  This
 revision  provides a clearer  presentation of the appropriate
 applicability dates.
 Comment:
      One commenter (IV-D-44) stated that the affected facilities
 defined in Section 60.560  are not consistent with those shown in Table 2
 of the regulation portion  of the January 10, 1989, Federal Register
 notice.  The commenter stated that wording in the regulation lists
process sections as affected  facilities, but Table 2 does not mark as
affected facilities some sections under continuous or intermittent.
Response:
      Table 2 in the regulation portion of the January  10, 1989, Federal
         identifies only those process sections and emissions that have
                                  2-19

-------
 a January  10,  1989,  applicability  date.   All  other  process  sections and
 emissions  are  identified  in  Table  1  of  that  notice's  regulation  portion.
 Together,  these  two  tables  identify  all  process  sections  and  emissions
 identified in  Section  60.560 of  the  January  10,  1989,  Federal  Register
 notice.
 Comment:
       One  commenter  (IV-D-40)  referred  to Table  2,  Section  60.560(d)  in
 the  January 10,  1989,  Federal  Register  notice, which  lists  continuous
 and  intermittent  emissions for the various process  sections that will
 become subject to control, and pointed  out that  in  the category of  poly-
 propylene  processes  (liquid  phase),  "raw materials  preparation" and
 "material  recovery"  are not  listed in the "continuous" column, although
 In some of the newer processes these would be considered  continuous
 operations.
 Response:
       Polypropylene, liquid  phase processes that have continuous emis-
 sions  from "raw materials preparation" and "material recovery" have an
 applicability date of September 30,  1987.  This was indicated  in the
 January 10,  1989, Federal Register notice by Section 60.560(b), which
 said "any  facility under paragraph (a) of this section which commences
 construction, modification, or reconstruction after September  30, 1987,
 is subject  to the requirements of this subpart except as  provided in
 paragraphs  (c) through (f) of this section," and by Table 1 of paragraph
 (c).   As noted in the above response, Table 2 lists those process sec-
 tions  and  their emissions that have a January 10, 1989, applicability
 date.  Those process sections and emissions not shown in Table 2 are not
 excluded from the standards,  but have September 30,  1987,  applicability
dates.
Comment:
       Two commenters (IV-D-44, IV-D-50)  stated that "raw materials
preparation" should be added to the list of process sections under  "low
density polyethylene, low pressure process and high density poly-
ethylene,  gas phase process"  in Table 2  of the regulation  portion  of the
January 10, 1989, Federal  Register notice.   Commenter IV-D-50  also
stated that Table 2 does not indicate all  continuous and intermittent
                                  2-20

-------
 emissions, and illustrated this statement by noting that no  intermittent
 emissions are shown from the LDPE, high pressure process.
 Response:
       Table 2 shows only those process sections that have a January 10,
 1989, applicability date.  The raw materials preparation section refer-
 red to by the commenter was proposed for control in the September 30,
 1987, Federal Register notice and, thus, has a September 30, 1987,
 applicability date.  Therefore, it would have been inappropriate to
 include this process section in Table 2.

 2.3  MODEL PLANT DESCRIPTIONS

       Comments in this section express concern over the representa-
 tiveness of the model  plants presented in  BID Vol.  I  to processes in
 general  and to individual  plants  specifically.   Commenters  pointed to
 emission streams  that  may have been overlooked and  to  process changes
 that affect emission characteristics.
 Comment:
       One commenter (IV-D-8)  stated that the generic model  plants,
 proposed on September  30,  1987, do not  represent  plants using non-
 olefinic comonomers at a  wide  range of  concentrations.   The  commenter
 stated that there has  been  a  proliferation of plants using  non-olefinic
 comonomers and  VOC emission  controls  for these  copoVymer plants  must  be
 tailored  individually.  In  a follow-up  exchange,  this commenter
 indicated that  copolymers of organic  acids and  ethylene produce  emission
 streams  of varying concentrations  of the non-olefinic comonomers  and
 that  these comonomers  may damage certain conventional control equipment
 or may not be abated by it at all.
 Response:
       In  response  to this comment,  the Agency requested  specific
 emission  stream information from the commenter that would allow specific
 identification of  "problem" streams and allow analysis  of such streams
 (see Docket Item  IV-C-20).  The response by the commenter (see Docket
 Item  IV-D-35) was general  in nature and did not provide the specific
 information needed to determine if special  consideration beyond the
generic approach was warranted.  Based on the information that was
                                  2-21

-------
 provided,  the  Agency believes  the  generic  approach  provides reasonable
 control/no control  decisions  in  all  instances,  and  therefore has  not
 modified  the rule  in response  to this  particular  comment.
 Comment;
       One  commenter (IV-D-1)  stated  that both bulk  processes and  gas
 phase  processes  used in  the manufacture of polypropylene generate a
 purge  stream to  rid the  process  of non-reactive impurities,  and that
 other  "controlled"  releases from the process may  be combined with this
 purge  stream.  The  commenter pointed out that common practice is  to
 subject the purge stream to "resource  recovery"  either through  direct
 recovery  for reuse  of its components or through combustion  for heat
 recovery.  The commenter then  stated that  this stream may qualify for
 inclusion  in the "regulatory baseline" on  the basis of the  definitions
 used.
 Response:
       The  Agency has  decided not to revise  the model plant  descriptions
 for polypropylene plants to incorporate this purge stream.   The approach
 for determining which  process  emissions from polypropylene  plants  are  to
 be controlled does  not require this revision.  For affected  facilities
 that are constructed, modified,  or reconstructed after September  30,
 1987,  and  before January 10, 1989,  the control/no control decision is
 based  on the "new"  approach that was presented in the January  10,   1989,
 Federal Register notice.  Such affected facilities may be still exempted
 from control based  upon  the uncontrolled threshold emission  rates
 presented  in the September 30, 1987,  Federal Register notice.  This was
 allowed to prevent  an owner or operator from being placed in  a position
 of having  not controlled emissions  based upon the September  30, 1987,
 proposed standards  that would now require control  if the new  approach
was applied.  An explicit consideration of this  purge stream  by the
Agency at  this time would not have  any effect on the control/no control
decision for emissions from the affected facility.
Comment:
      One commenter (IV-D-7) stated that,  in many cases,  the model
plants used to develop this  standard  do not adequately reflect current
operation of manufacturing plants.   The commenter then referred to the
model  plant for LOPE based on the UNIPOL process.   The commenter noted
                                  2-22

-------
 that  the model  plant did not consider the modified UNIPOL process that
 has a pellitizer section added  and the linear LDPE solution process.
 The commenter stated that emission characteristics differ significantly
 from  the model  plant,  but are controlled  the  same.  The commenter then
 stated that  the gas  phase process  for polypropylene failed to consider
 emergency atmospheric  vents  that  are  used on  newer plants.   Therefore,
 the commenter suggests that  these  model plants  be  reviewed to ensure
 that  they are representative of operating plants throughout the
 industry.
 Response:
       The Agency  understands and agrees with  the commenter's  concern.
 As a  result  of  concerns  such as this,  the Agency developed  a  new
 approach  for determining  which process emissions from polypropylene and
 polyethylene plants  would be controlled and presented this  approach in  a
 January  10,  1989, Federal  Register notice.  The new approach  does  away
 with  the  need to  define model plants  and  the Agency determined it  was
 unnecessary  to  revise  the model  plants.
 Comment!
      Two  commenters (IV-D-6, IV-D-8) stated that  there are several
 aspects of the  fluid bed  gas phase polyethylene process model plant
 (used to describe the  LDPE low pressure and HOPE gas phase processes)
that do not  reflect actual current operations.  These discrepancies (and
the changes  suggested by  Commenter IV-D-8) are identified below.

      1.  The process model does not include a polymer finishing  section
          which was consistent with the early  expectations of this
          process.  However,  the process  evolution  over the past  seven
          years has grown to  include a Finishing,  Extrusion, and
          Blending section in essentially  all  plants of this type.   This
          important aspect of the  process  should be included to ensure
          clarity.  It  is important to note that this  finishing section
          will  continue to have  the Vent  "J" shown  on  the  diagram.   This
          is  typically  a  bin  vent  and  will contain  residual  VOCs  in a
          very dilute air stream.   The revisions shown  in Table  3
          reflect  this  change and  shows a  typical range  of emission
          rates  with  operations on  HOPE and LDPE.
                                 2-23

-------
       Commenter IV-D-6  noted  that  this  section takes the
 granular  product  from  storage  bins  (vent  "J"  in  BID Vol.  I)
 and  treats,  shapes and  modifies the  polymer to the finished
 palletized and  product,  and that there  should  be  no
 incremental  emissions generated by this  additional  material
 handling facility.
 There  are several  changes  that  are required to correct  the
 stream "I"  (Product Discharge Vent)  in Table 3-9  of BID Vol.  I
 to reflect  actual  operations and to  ensure  clarity:   (a) This
 stream should be correctly indicated as  intermittent  in
 nature;  (b)  This'stream  has been incorrectly included  in the
 Process  Finishing  Section of the plant for  process  evaluation.
 It is  actually  part of a Material  Recovery  section  according
 to the definitions  included in  the proposed  rule,  and that
 process  section definition should  be corrected; and  (c) The
 magnitude of the emission rate  assumed for  vent stream  "I"  in
 Table  3-9 appears  to also include  the recycle  from  the  Product
 Discharge system back to the Reactor system, which  is not
 consistent with actual operations.   For clarity this emission
 rate should  be adjusted to reflect typical   rates.
       This commenter also noted  that footnote  "d"  is the vent
 stream summary in Table 3-9 in BID Vol.  I incorrectly
 indicates that stream "I" is not routed  to a flare as part of
 baseline control.  This should be changed to reflect typical
 plant  operations.
       Commenter IV-D-6 noted that Vent "I"  from the Polymer
 Discharge Tank,  which is described  as a  continuous vent in the
 Product Finishing section,  does  not exist on its own but is
 routed through the Polymer Purge Tank, which is in the
 Polymerization Reaction Section, and  is  combined with vent
 "H".   Also, this commenter stated that the  Polymer Discharge
Tank, which was  classified in  BID Vol.  I as product finishing,
should be part of the Polymerization  Reaction Section.
Stream "H" in the process shown  in  Figure 3-5 of BID Vol. I
also requires several  changes  to reflect actual operations
according to Commenter IV-D-8:   (a) A more  correct name is
                        2-24

-------
           "Polymer Purge Vent", rather than "Reactor Purge Vent", to be
           consistent with the equipment name shown on the diagram in
           Figure 3-5 of BID Vol. I; (b) This purge stream is continuous
           in plant operations rather than the intermittent nature
           indicated in Table 3-9 of BID Vol. I; (c) The process section
           for this stream should be "Material  Recovery", not
           "Polymerization Reaction" as indicated in Table 3-9 of BID
           Vol. I.  This classification would more closely fit the
           definitions included in the proposed regulation (Subpart ODD)
           for this type of operation;  (d)  The  part of the process
           emitting stream "H" has typically undergone optimization since
           the initial  process development  and  may now include raw
           material  recovery technologies  for certain  LDPE applications.
           This important phase of emission reductions should  be noted on
           the process  diagram;  and  (e)  The magnitude  of this  material
           recovery stream emission  rate is highly dependent on  the type
           of polymer  being  produced,  HOPE  or LDPE,  and  the  processing or
           recovery equipment employed.  These  aspects result  in a  wide
           range  of emissions depending  on  the  operations.  A  "typical"
           range  is  between  5.0  and  20.0 kg VOC/Mg  product.
                 Commenter IV-D-6  stated that vent  "H"  from the  Polymer
           Purge  Tank,  which  is  described as  an  intermittent vent  in  the
           Polymerization  Reaction Section,  is  a  continuous vent  in
           current operations, and in general is  routed to a flare.   In
           some cases,  raw materials are recovered  from vent "H" prior to
           it  being  vented to  a  flare system.
      Finally, Commenter  IV-D-8 noted that the basic  processing steps in
a fluid bed gas phase process are identical for HOPE  and LDPE, with  only
the finer  operating parameters and emission levels being different.  A
range of emissions was provided by the commenter for three of the eleven
streams, resulting  in a total emission rate from the entire facility
ranging from 7.36 to 22.5 kg VOC/Mg product.
Response:
      The  aspects of the polyethylene process described by the
commenters as being "discrepancies" reflect changes that have  occurred
to the process upon which the model  plant  was based.  It was this type
                                  2-25

-------
 of comment,  in  part,  that  led  the  Agency  to  adopt  the  "new"  approach for
 determining  which  process  emissions  from  polypropylene and  polyethylene
 plants  are to be controlled  rather than continuing  to  rely  on  model
 plants  that  may change.  The Agency  has determined  that  it  is
 unnecessary  to  develop new model plants or to  revise those  that  served
 as the  basis of the September  30,  1987, Federal Register  notice;  the new
 approach  takes  such changes  into account.
 Comment:
      One commenter (IV-D-8) referred to  the characterization  of
 intermittent versus continuous streams.   In their  initial letter,  the
 commenter indicated that in  polystyrene plants changing processes  often
 change  flow  streams from continuous  in nature  to intermittent  as
 recovery  processes are applied to  those streams..  A follow-up  exchange
 (see Docket  Item IV-D-35)  indicated  that  the commenter was referring not
 to the  emission stream changing from continuous to  intermittent, but to
 either  higher or lower levels of emissions being released as a result of
 changes in the conditions  of the closed system or in the raw materials
 to make a product change.
      The commenter also stated that the Agency accounted for  such items
 as start  up, shutdown and maintenance purges by grouping them  as
 "intermittent bin vent" emissions,  and that these purges can come  from
 any area  of  the process.   The commenter then stated that it is unlikely,
 however,  that purges would come from the product storage area  itself.
 With regard  to bin vents, the commenter stated that it is generally
 agreed that  bin vents are continuous when the bin is in use and that
 there are no emissions when the bin is not in use.   Lastly,  the
 commenter stated that in  actuality in LDPE and HOPE units, the
 intermittent process emissions (startup -  shutdown  and maintenance and
 process emissions)  for safety reasons must be separated from product
 storage bin  units which normally contain oxygen (air)  in control  device
 collection systems.
 Response:
      These comments are  directed at  making explicit certain aspects of
various polymer manufacturing practices that  affect emission levels and
control.  None of the specific comments require the Agency to revise the
                                  2-26

-------
 polystyrene model plants or the approach for determining which process
 emissions from polyethylene processes are to be controlled.

 2.4 SELECTION OF BASIS OF THE PROPOSED STANDARDS

 2.4.1 Model Plant Approach
 Comment:
       One commenter (IV-D-8) stated that by using reaction mechanisms as
 the models for developing the NSPS control  requirements EPA has tailored
 the rule to these particular patented processes and that this provides
 an unfair advantage for both the licensors  and licensees of those
 technologies because the sale of licenses is due,  in part,  to the
 ability of a process to comply with applicable environmental  standards.
 The commenter pointed to the low pressure linear low density
 polyethylene (LLDPE) process as the best example.   The commenter  stated
 that when the model  plant for this process  was developed,  the only
 process in use was  a fluid  bed gas phase facility  licensed  by a major
 U.S. chemical  company.   Since that time,  the commenter continues,  both a
 slurry and a solution  process have been  revised to produce  the linear
 low density product,  and a  new gas phase facility  licensed  by a major
 British petrochemical  company is currently  under construction.
 According to the  commenter,  these  newer  processes  differ greatly  from
 the low pressure  process  that  the  Agency  considers  state-of-the-art.
 The commenter  notes  that  if  the NSPS  is  promulgated  as  written, it will
 require that similar new  installations,  as well  as  existing units that
 are modified or reconstructed,  be  equipped with  emission controls
 designed  specifically for the  fluid bed gas  phase  facility.   The
 commenter suggests that  the  economic penalty resulting  from this action
 be  considered  before the regulation is finalized.
      This commenter then suggested that  it  might be more appropriate
 for  EPA to establish control requirements based on the  system pressure
 and  process  type rather than the specific technology used because system
 pressure  (not reaction mechanism)  is the primary factor influencing
 emissions from polyolefin manufacturing facilities and will  determine
the relative ease at which unreacted raw materials are removed from
reacted mixtures.   The commenter also stated that system pressure will
                                  2-27

-------
 also be a factor in defining the process  step from which the unreacted
 raw material  will  be emitted,  noting that the higher pressure processes
 tend to hold  on to the unreacted materials longer,  thereby yielding
 greater emissions  in later process  steps  (i.e.,  product storage).
       The commenter illustrated  this suggestion  by breaking the
 polyethylene  processes into two  broad classes -  high and low pressure.
 Under the low pressure classification,  the commenter indicated that
 there are presently three  commercial  processes --  (1)  gas  phase,  (2)
 liquid slurry and  (3)  liquid solution.  According  to the commenter,  each
 of these  processes in  the  newer  and  more  recently  modified plants  has
 the capability of  manufacturing  both  HOPE  and LDPE.   Under the high
 pressure  classification, the commenter  indicated that  there are
 presently two commercial processes  --  (1)  tubular  and  (2)  autoclave
 -both of  which manufacture  conventional LDPE.
       The commenter pointed  out  that  under  the proposed  regulation,  if
 these low pressure solution  or slurry process plants manufacture LLDPE,
 then  they are grouped  with  the gas phase model plant, while if they
 manufacture HOPE,  they are  grouped with either the HOPE  solution model
 or the HOPE slurry model.   The problem or  inconsistency  with  this
 result, according  to the commenter,  is that from a product,  process, and
 emissions  point of view, when a  solution plant manufactures  LLDPE  it
 actually  more  closely  resembles  the HOPE solution model  and when a
 slurry plant  manufactures LLDPE,  it more closely resembles  the  HOPE
 slurry model  than  the  gas phase model plant.  Finally, the  commenter
 pointed out that the product finishing areas of modern solution and
 slurry LDPE plants  resemble the HOPE solution and HOPE slurry model
 plants much more than  they do the gas phase model plant.
 Response:
       The Agency took  this comment under consideration and  as a result
 of extensive analysis  presented a new "generic"  approach for determining
which  process emission streams from polypropylene and polyethylene
 plants would be subject to control.   The new approach effectively deals
with the commenter's concern for  process sections that become affected
 facilities either before, on, or  after January 10,  1989.
                                  2-28

-------
 Comment:
       One commenter (IV-D-8) stated that the Agency failed to consider
 not only a "worst case" scenario in determining the economic impacts but
 also failed to consider the "real case."  The commenter referred to page
 36691 in the September 30, 1987, Federal Register notice in which the
 Agency states that its survey was based on "economic impacts associated
 with a worst case growth distribution."  The commenter referred to
 Tables 4, 5, and 6, in Docket A-82-19, Item II-B-92, Calculations of
 Average, or Typical.  Cost Effectiveness Selected as the Basis for the
 Proposed Standards and The Next More Stringent Regulatory Alternatives.
 which show expected growth distribution by process section, process
 line, and new plants  and the costs  associated with the proposed
 standards and the next more stringent  regulatory alternatives weighted
 by the expected growth distribution.   The commenter states  that since
 this analysis was projected from 1984  to 1988,  it  is now possible to '
 check if the worst case was actually considered.   The  commenter claims
 that during  this  time  period,  no new polyethylene  and  polypropylene
 process  lines or  plants have  been brought on  line  and,  therefore,  all
 capacity increases have been  the result  of process section
 modifications.  Process sections were  shown to  have  the  highest  cost of
 the three mechanisms considered.  Since  EPA in  their analysis did  not
 weight growth by  process  section, the  growth  with  the  highest costs
 associated with it, as  heavily  as they did the  growth  by process  line or
 new plants,  the commenter  believes the true growth distribution would
 have led  the  Agency to  the  conclusion  that different regulatory control
 alternatives  should have been chosen.
 Response:
      The commenter has misunderstood  the use of the projected growth
 distribution  in relationship to  the selection of regulatory alternatives
 as  the basis  of the proposed standards in the September- 30, 1987,
 Federal Register notice.  The regulatory alternatives selected as the
 basis of the  standards were done so,  in part,  based on the costs and
 cost effectiveness values reported in Chapter 8 of BID Vol.  I, and not
 on the basis of the analysis in Docket  Item II-B-92.  The costs in
Chapter 8 reflect costs incurred on  a per process section per process
line basis.  For several of the model  plants,  the cost  effectiveness of
                                  2-29

-------
 the  selected  alternative  was  felt  to  be  somewhat  high  if control  would
 be  implemented  by  a  control device dedicated  to that particular  process
 section.  The Agency considered whether  emissions  from these  process
 sections might  not be controlled in a less expensive manner.   For these
 process sections,  the Agency  concluded that the emissions  to  a common
 control device  or  that modifications to  multiple process  sections of the
 same  type (e.g., product  finishing sections)  were  more  likely  to  occur
 than  to a single process  section (e.g.,  one of three product  finishing
 sections).  These  conclusions were noted in footnote b  to  the
 appropriate tables in  Chapter 8 of BID Vol. I.  Furthermore,  the
 economic impacts reported in Chapter 9 of BID Vol. I were  based on the
 costs found in  Chapter 8  and not in Docket Item II-B-92.   Thus, the
 regulatory alternatives selected were based on consideration of the
 worst-case scenario  (control of individual process sections) as were the
 economic impacts.  The purpose of Docket Item II-B-92 was  to obtain
 (hopefully) a more realistic picture of the impact of the  already-
 selected regulatory  alternatives on the  industry as a whole over  the
 next five years.  While it is now possible to check how accurate  the
 growth distribution  projection was for that document, its  accuracy is
 not relevant as it pertains to the selection of regulatory alternatives.

 2.4.2 Polystyrene Processes
 Comment:
      One commenter  (IV-D-8) stated that the use of refrigeration
 condensers are not technically feasible for polystyrene processes and do
 not meet the criteria  for standards of performance as stated in the BID:
 "standards of performance must (1)  realistically reflect best
demonstrated control  practices;  (2) adequately consider the cost, the
nonair quality,  health and environmental  impacts,  and the energy
requirement of such control; (3)  be applicable to  existing sources that
are modified or reconstructed  as  well  as new installations; and (4) meet
the conditions of all variations  of operating  conditions being
considered anywhere in the country."
      The commenter wrote that the  Agency is  correct  in stating in BID
Vol.  I that condensers are cost  effective for  recovery  of compounds with
relatively high  boiling points like styrene and  that  a  refrigerated
                                  2-30

-------
 condenser is not feasible when moisture is present in the stream which
 might cause freezing in the condenser.  The commenter then pointed out
 that the latter holds true for continuous polystyrene processes.
       The commenter then gave a number of reasons why freezing will
 occur.  First and most important, according to the commenter, is the
 water in the vent.  The commenter noted that the Agency assumed that
 when plants switched from steam ejectors to vacuum pumps, the freezing
 problem associated with water disappeared with the steam.  However, the
 commenter stated,  there is water entering with the raw materials and
 with the air leakage into the system (especially true in hot, humid
 climates).   The commenter pointed out that most of the water comes from
 the water content  of the styrene and the rubber used  in  the  manufacture
 of high impact  polystyrene (HIPS).   To illustrate the level  of water
 content, the commenter noted  one company reported an  average of 0.08
 weight percent  or  800  parts per million (ppm)  water in the feed to the
 process, while  another company has  measured  167,  453,  and 284 ppm
 moisture in  styrene  in  three  different months  with a  maximum of 700 ppm
 (saturation).
       According  to the  commenter, the rubber used in  making  HIPS
 typically has 0.4  percent  water in  it,  and since  rubbers  can  make  up  5
 to 10  percent of the finished  product,  the overall  water  content in  the
 raw materials could be  as  high  as 0.1  weight percent.  The commenter
 pointed  out  that while  some water will  be driven  off  through  the heating
 reactor  and  any  water draws, saturated  water vapor  in  the 30-50 ppm
 range will remain in the devolatilizer  vent stream  and cause  freezing
 problems  at  the  control temperature  (-25C) proposed  in the regulation.
      Another potential problem with  using a refrigerated condenser,
 according to the commenter, is the presence of additives  in the process,
 some of which have high freezing points.  One company reports freezing
 of the primary condensers coming off the devolatilizer when they ran the
 condenser's glycol  system at -2 to -4C.  Thus, if certain additives are
 present, the freezing is more likely to take place.
Response:
      After further investigation, the Agency agrees that freezing  may
be a problem at  sub-zero temperatures.  Therefore, the Agency has
reanalyzed the regulatory alternatives for polystyrene plants using
                                  2-31

-------
 spared heat  exchangers  with  defrost  capabilities  and  a  refrigeration
 system for sub-zero  applications.  This  reanalysis  is presented  in
 Docket Item  IV-B-18.
 Comment:
       One commenter  (IV-D-6)  stated  that  it  is  inappropriate  to  impose
 compliance by  an  alternate indication technique (i.e.,  temperature
 measurement).
 Response:
       The Agency  disagrees with the  commenter.  The use of condensers to
 control emissions allows for  an essentially  equivalent  standard  to  be
 set  in terms of outlet  gas temperature.  The Agency is directed  by  the
 Clean  Air Act  to  set emission limits unless  otherwise impracticable  to
 do so.  The Agency set  an emission standard  of 0.0036 kg TOC/Mg  product.
 At the same time, the Agency  saw an  opportunity to provide for
 equivalent emission control by limiting the  temperature of the outlet
 gas  stream from the condenser.  Temperature  is a good indicator  of
 performance in this case because at  a given outlet temperature the
 concentration of styrene that can be vented  is essentially the same
 regardless of concentration of styrene entering the condenser.   As a
 temperature standard allows continuous compliance determinations to be
 made and is much less costly for compliance testing, the Agency  feels it
 is entirely appropriate to set a temperature standard as an alternative
 standard for condensers.
 Comment:
       Two commenters (IV-D-6, IV-D-8) expressed concern  over the costs
 basis  used to select refrigerated condensers as the basis of the
 proposed standards for polystyrene plants.
       Commenter IV-D-8 stated that the Agency did  not adequately
determine the costs and cost-effectiveness associated with using
refrigerated  condensers in the polystyrene continuous process.
Specifically, this commenter  made the following points:
       1.  The Agency did not  account  for  the cost  of removing  any water
          or  additives that could freeze  the condenser.   This  cost
          clearly was not considered  in the justification  of the
          refrigerated condenser since  the  Agency  assumed  there were no
          high freezing point compounds present.   Figure 4-6  on page
                                  2-32

-------
           4-28 of BID Vol. I shows the condensation system with a
           dehumidification unit in it.  However, no economics were done
           on this control technology.
       2.  The Agency grossly understated the installed capital cost of
           the condenser and the refrigeration unit for the regulatory
           alternative chosen.  A recent cost estimate was done by
           contracting a vendor to determine the overall cost for the
           condenser system specified in BID Vol. I.  The installed cost
           for the proposed system was $15,600.   The commenter believes
           this number represents a more accurate installed cost than the
           $1,717 cost reported in BID Vol.  I.
       3.  Because of present  State regulations  for VOC emissions from
           polystyrene units and the desire  of companies to lower the
           content of unreacted styrene in  the finished product,  both new
           and older plants have lowered their primary  condenser
           temperature to  pull  lower vacuum  and  remove  (condense)  more
           styrene.   The Agency used 27'C to determine  the amount of
           styrene saturated in  the air going to  the vacuum pump.   The
           temperature now runs  around 5 to  10'C.   Therefore,  less
           styrene is  being emitted.   This establishes  a new baseline  to
           determine  cost  effectiveness  that should  be  considered  before
           a  final  rule can be determined.
       4.   The Agency  calculated  a  $3,000 credit  for the styrene
           recovered  by the use of  a refrigerated condenser.   This
           assumes the condensate is usable.  One company  has  chilled
           condensers  (2-5'C) on  the discharge of the vacuum pumps.
           (They also run their primary  condensers at the  level stated  in
           point 3.)  This  condensate  consists of dimers and trimers of
           styrene and other heavy materials, and is not suitable to
           return to the recycle tank.  The condensate  is disposed of by
           incineration.  The commenter believes that the Agency should
          not allow a credit for material that cannot be reused.  The
          only credit should be for the British thermal unit  (BTU)
          content of the condensate when it  is burned.
      Commenter IV-D-6 stated  the cost of the refrigeration condenser
system for the model plant in  BID Vol. I appears to be  totally
                                  2-33

-------
 unrealistic and grossly underestimated and the cost of the specified
 technology to achieve the indicated reductions does not appear to
 justify the additional  control  above 0.12 kg TOC/Mg product.   The
 commenter pointed out the following cost  factors  they felt might have
 been overlooked:
       1.   moisture content of the  stream  would require drying systems;
       2.   poor heat transfer  coefficients due to  the high  nitrogen and
           noncondensible  content of the stream;
       3.   explosion-proof requirements must  be Class I division  2;
       4.   refrigeration system  would  be non-standard (probably
           propylene)  due  to the temperature  requirements;
       5.   higher  metallurgy (stainless steel)  required due to the low
           temperature requirement;  and
       6.   cost  for new process  condenser  and  associated  piping appears
           not  to  have been considered  in  development  of  the proposal.
           Existing chilled water condensers  are riot  rated  for this low
           temperature application  and refrigeration  service.   Therefore,
           new  process condensers would be required.
       The  commenter provided  cost  estimates  for controlling two
different  existing polystyrene  plants and one  new plant.   The
commenter's cost  estimates yielded cost effectiveness  values  of
approximately $24,000/Mg of VOC removed for the two existing  facilities.
The plants and the cost basis used by the commenter were as follows.
       The  first existing plant was assumed to  (1)  have a production rate
of 75  million pounds  per year,  (2)  emit 0.12 kg VOC/Mg product from
the material recovery section,  (3)  use chilled water condenser system,
and (4) achieve compliance with the 0.0036 kg TOC/Mg product standard  by
installing a new  (add-on)  process  condenser,  refrigeration/condensing
system and associated piping.   The  capital cost for this equipment was
estimated at $125,000.  This cost  was based on equipment designed  for
explosion proof (Class I division  2) installation, two refrigeration/
condenser package units to ensure  and maintain unit operations due to
anticipated reliability and maintenance problems with the refrigeration
units, a vent stream design basis  of 5 cubic  feet  per minute (cfm) at
140*F, 5 to 10 ton compressor  units, and  a propylene refrigeration
system.  The commenter stated  that  a non-standard  refrigeration (i.e.,
                                  2-34

-------
 propylene system) appears to be necessary due to the low temperature
 requirements.  The commenter also stated that a new process condenser
 would be required since the existing chiller is not rated for low
 temperature applications and refrigeration service.  The cost breakdown
 provided by the commenter was: $60,000 for two refrigeration (propylene)
 chillers; $8,000 for new process condenser; $5,000 for associated
 piping; $10,000 for industrial rated controls; $18,000 for labor for
 installation; $5,000 for engineering and drafting; $4,000 for freight;
 $5,000 for associated expense; and $10,000 for material and labor
 contingencies.
       The second existing facility was assumed to produce 120 million
 pounds (Ibs) of HIPS and 360 million Ibs of general purpose polystyrene
 and emit 0.025 Ibs of VOC per Mg of product.   The commenter estimated
 that it would cost $300,000  to install  the required incremental
 refrigeration/condensing system in order to comply with the proposed
 standard.
       For the new facility,  the commenter stated  that  an  engineering
 contractor has estimated the incremental, costs  associated with just
 complying  with these  new provision to  be $754,000.
       On  the basis of these  comments regarding  the  cost and
 cost-effectiveness of refrigerated condensers,  as well  as the
 questionable viability  of the  technology  due to the presence of water,
 Commenter  IV-D-8  stated that EPA must redetermine the best  technology
 system for polystyrene  continuous  processes.
 Response:
       The commenters  have brought  up a number of points concerning the
 use  of and cost estimates for refrigerated condensers on emission
 streams from polystyrene plants using continuous processes.  For the
 response to these comments,  the reader is referred  to Docket Item
 IV-B-18.  The  following summarizes the Agency's response to the
commenters concerns.
      The Agency agrees that consideration of subzero condenser
applications requires dealing with moisture level.  Based on
conversations with a number of vendors,  the Agency has  Devaluated
control below 32*F using a system composed of two  heat  exchangers  each
equipped with a defrost unit  and a refrigeration unit  to service  both
                                  2-35

-------
 heat exchangers.   This  system now incorporates  stainless steel
 construction.   The system previously considered noncondensibles and thus
 the Agency feels  proper heat  transfer coefficients  have been used.
 Vendor contacts indicated that ethylene  glycol-water solutions  and  Freon
 502 coolants would suffice; non-standard refrigerants would not be
 needed.   In addition, the explosion  proof requirements,  which refer to
 electrical  wiring requirements,  have been directly  considered.   Based on
 these assumptions,  the  Agency has  recalculated  the  costs of the control
 alternatives.
       Based on  the new  costs,  the  Agency has  calculated  a total
 Installed  capital  cost  of between  $8,600 and  $9,900  (1980$)
 (approximately  $11,400  to $13,100  in  1989$) for a single and a  spared
 condenser  system,  respectively.  While these  costs  are  higher than  the
 cost the Agency estimated at  proposal, they are still  substantially
 below those estimates provided by  the commenter.  The  first  example
 provided by the commenter had  a  total installed capital  cost of
 $125,000.   This estimate  was  based on replacing existing process
 condensers  with ones that would  meet the  standard.  After discussions
 with the commenter, it  became clear  that  while  this route can be used,
 ft  1s  unnecessary.
       The commenter reestimated  the  installed capital cost of this
 system to be $48,000.   This is still much larger than the  new estimate
 by  the Agency.  The difference is due, in part,  to the costing  of a
 condenser that  is much  larger than needed (60 square feet  versus a
 needed 10 square  feet or  less) and an "exotic"  refrigeration system.
 Based  on information from  a vendor for a  system  nearly identical to that
 being  considered  by the commenter, purchase costs were between  $7,000
 and  $12,000 (1989$), for  an installed capital  cost of between $9,700  and
 $16,700 (using  an EPA installation factor of 1.39) for a  single and
 spared condenser  system,  respectively.
      The commenter's cost estimate for the second existing  facility
 included costs  for an expensive water removal  system.  The Agency
 believes this  is unnecessary,  that a spared condenser system, which is
much less expensive, will handle potential freezing  problems that a
 single condenser might encounter.  Further, the costs for the
 condensers/refrigeration units appear to be again based on units much
                                  2-36

-------
 larger than necessary to handle the emission stream during normal
 operating conditions.  Finally, the commenter's cost estimate for the
 third facility is not an incremental cost estimate and, when asked for
 additional details, the commenter requested that the Agency "ignore"
 that cost estimate in their comments.
       In summary, the Agency has carefully reevaluated the condenser
 system necessary to meet the standards.   These costs are higher than
 originally estimated because of the technical, aspects identified by the
 commenters.   However, the cost estimates provided by the commenters
 appear,  based on the information provided,  to be based on systems that
 are much larger than needed, that in some instances use techniques that
 are more expensive than other equally  effective techniques,  and  may
 still  reflect unnecessary changes to existing process or recovery
 equipment.  Thus, the Agency believes  its costs now reflect  more
 accurately the costs of condenser system that would be incurred  if
 designed specifically to meet these standards.
       Commenter IV-0-8 suggested that  the Agency establish a  new
 baseline to  determine cost  effectiveness before determining  the  final
 rule.  The Agency provides  for an uncontrolled  emission  rate  threshold
 level  that protects  against non-cost effective  control  of facilities
 referred to  by the  commenter.   Therefore, the Agency  has  retained  the
 baseline as  presented in  BID Vol.  I.

 2.4.3  Polyethylene  terephthalate)  Processes
 Comment:
       One  commenter  (IV-D-13)  stated that currently highly efficient
 materials  recovery process  technology has greatly reduced the rate and
 diluted  the concentration of  the  affected vent  stream so  as to reduce
 the value  of the  condensate  and cause the installation and the operating
 cost of  refrigerated  condensers to be unreasonable.
 Response:
      This commenter was referring to the lower quantity of methanol
 being recovered in PET plants and thus the recovery credit would be
 lower.  The Agency has taken into account facilities that already
substantially reduce their methanol rate by the provision of an
uncontrolled threshold emission rate.  Based upon this analysis,  the
                                  2-37

-------
 Agency has determined that additional  control  using refrigerated
 condensers is unreasonable if uncontrolled emission rates at existing
 facilities are less than 0.12 kg TOC/Mg product,  but that such control
 is reasonable if the emission rate is  higher.
 Comment:
       One commenter (IV-D-11) stated that  the  limit of 0.04 kg TOC/Mg
 product from esterification  vessels for high viscosity PET using
 multiple  end finishers  (Section  60..562-l(c)(4)(iii)  of the
 September 30,  1987,  Federal  Register notice) appears to be in  error and
 is not supported by BID Vol.  I.   The commenter stated that the
 appropriate limit should be  0.15 kg TOC/Mg  product.
       The commenter referred  to  Section 3,  page 3-61 of BID Vol.  I  where
 emissions from the various PET processes were  discussed,  and Table  6-lib
 on page 14 of Section 6 where the model  plant  process  emissions  for the
 PET terephthalic acid (TPA)  process using multiple  end  finishers  are
 summarized.   The commenter pointed out  that the information contained in
 these  sections of BID Vol. I  show the baseline emission  rate from the
 raw materials  preparation section  of the PET-TPA multiple  end  finisher
 process to be  0.15 kg TOC/Mg  product.   The commenter then  referred  to
 the various  regulatory  alternatives for control of this  process which
 are discussed  throughout Section  6 of BID Vol. I and  are summarized  on
 page 36690 of  the  preamble to the  September 30, 1987, Federal  Register
 notice.   This  information, the commenter pointed out, indicates that
 control of emissions beyond the  baseline level  in the raw  material
 preparation  section  is  not required.  Therefore, the commenter concluded
 the 0.15  kg  TOC/Mg product baseline level shown in BID Vol.  I  should  be
 the limit  mandated by Section 60.562-l(c)(4)(iii).
 Response:
      The  Agency reviewed the information in BID Vol. I and  the docket
 concerning this comment.  The sources show inconsistent treatment of
 esterifiers  from high viscosity PET plants using multiple  end  finishers.
 For example, Appendix E states that baseline control costs  for these
 facilities were estimated assuming reflux condensers on the  esterifiers,
which are  associated with an emission rate of 0.04 kg TOC/Mg of product
 (see page  3-63 of BID Vol. I).  The commenter,  who uses a different type
 of  condenser on their esterifiers, has  stated that they would expect
                                  2-38

-------
 their condensers to be as efficient as reflux condensers (see Docket
 Item II-B-73).  While a previous estimate based on sampling conducted in
 1978 at the commenter's facility showed an estimated emission rate of
 0.15 kg TOC/Mg product, a more recent test conducted by the commenter
 shows that the controlled emissions from the esterifiers are below 0.04
 kg TOC/Mg product (see Docket Items IV-D-70, IV-E-78, and IV-E-80).  In
 developing the baseline control  costs, the Agency incorporated reflux
 condensers as baseline control.   Unfortunately,  this was neither stated
 nor shown in Chapter 6, page 6-41,  where the contradictory,  and
 erroneous, statement was made that  baseline control  consisted,  in part,
 of distillation columns on the esterifiers.   For new plants, it was the
 Agency judgement that reflux condensers represented  best available
 technology and should serve as baseline for new, grass  roots plants.   As
 noted above,  the more recent test by the commenter shows that their
 condensers are achieving equivalent levels  of control.   The  Agency also
 conducted a new analysis specifically estimating the cost  of controlling
 the commenter's 0.15 kg TOC/Mg product stream to 0.04 kg TOC/Mg  product
 (see Docket Item IV-B-20).   This  analysis  showed the cost  of control  to
 be reasonable.   Thus,  while  the commenter  is  correct in  pointing  out
 discrepancies  in the BID for the  proposed  standards,  the final  rule
 retains  the proposed standard of  0.04  kg TOC/Mg  product.

 2.5   PRESENTATION OF  THE  STANDARDS

 Comment:
      Two  commenters  (IV-D-6,  IV-D-8)  stated  that  plant  technical  and
 management  personnel  in  a major polyethylene  plant using both high  and
 low pressure technology to manufacture low density polyethylene have
 encountered problems in determining and understanding the specific
 requirements applicable to each process unit  and section.  The
 commenters  felt that their misinterpretations appear to  be from the fact
 that the facility is faced with a maze of requirements which vary
 between product lines, affected process sections, and type of vent.  One
 commenter  (IV-D-8) further felt that this was complicated by the use in
 BID Vol. I of a low pressure LDPE model plant that did not adequately
describe that type of facility.  The commenters thus  recommended that
                                  2-39

-------
 the  regulation  be  simplified  and made more clear.  The  commenters
 provided  an example of  a table that described the  specific  requirements
 and  controls by process, affected sections, and types of  emissions
 (continuous vs. intermittent).  One commenter (IV-D-8)  believed  that
 this will provide  the clarity needed to ensure proper interpretation,
 and  can be used either  to clarify or replace Table 1 in Subpart  ODD of
 the  regulation.
 Response:
      The Agency has revised the final rule and incorporated  several
 figures to help clarify the requirements applicable to each facility and
 the  procedures  for determining which process emissions are  to be
 controlled.  These and other similar figures are presented  in Appendix A
 of this document.
 Comment:
      One commenter (IV-D-8) suggested that the flare regulations be
 referenced to the flare standard as it appears in 40 CFR  Part 60 rather
 than being repeated in the proposed regulation.   The commenter pointed
 out that by incorporating the flare standard by reference any changes
 made 1n the future to the flare standard would be automatically  incor-
 porated into the polymer NSPS, rather, than requiring a special effort
 to incorporate  changes to the flare standard in  this proposed rule, and
 such an approach would then be consistent to the way the Agency handled
 equipment leaks in Section 60.562-2 of the proposed standard.
 Response:
      The Agency agrees with the commenter.   Therefore,  the final rule
 now references 40 CFR 60.18 for the specific requirements for using
 flares to meet the standards.

 2.6   APPROPRIATENESS AND APPLICABILITY OF PROPOSED STANDARDS TO CURRENT
      POLYMER PRODUCTION PROCESSES

 2.6.1  Dilute VOC Streams
Comment:
      Three commenters (IV-D-6,  IV-D-8,  IV-D-13)  expressed concern over
requiring control  of dilute streams.   Commenters  IV-D-6  and IV-D-8
 stated that the Agency has not demonstrated that  using  a device that is
                                  2-40

-------
 98 percent efficient or reducing the TOC concentration to 20 ppm by'
 volume (ppmv) on a dry basis {corrected to three percent oxygen content)
 is either possible or economically reasonable for dilute gas streams.
 The commenters noted that control of dilute streams has become much more
 significant since the industry now relies more heavily on material
 recovery for control -- far less material is left to be emitted and the
 remaining VOC is more dilute.  In this connection,  Commenter IV-D-8
 stated that there has been a recent trend in the polypropylene and
 polyethylene industries,  which developed since EPA  reviewed the industry
 to reduce emissions by maximizing raw material  recovery.   The comment
 pointed out that nowhere  in BID Vol.  I or in the proposed rule did the
 Agency consider material  recovery as  a possible control  mechanism for
 these industries.   As a result, the extent to which low concentration
 (TOC) streams now exist in these processes was  not  anticipated and the
 difficulty and expense involved in  controlling  these streams was  not
 considered by EPA.
       The commenters felt  that  the  Agency failed to address  adequately
 the problem of availability and cost-effectiveness  of control  for dilute
 gas streams (TOC concentrations <2,000 ppmv).   One  commenter (IV-D-8)
 stated that EPA gave only  cursory attention  to  the  availability and cost
 of controls for these dilute  streams.   This  commenter referred to  the
 preamble  to the NSPS (52 FR 36696), where  the Agency  states  that,  "Data
 summarized  in  BID Vol.  I showed that  20  ppmv  is  the  lowest VOC
 concentration  achievable by combustion  of  gas streams  containing  less
 than  2,000  ppmv  VOC."  However,  as  the  commenter  continued,  BID Vol.  I
 only  dealt  with  one  source  containing  an uncontrolled  VOC concentration
 of less than 2,000 ppmv (Table  C-6);  these are being  one containing 950
 ppmv  of maleic  anhydride.    According  to  this commenter, there are
 insufficient data to  establish  the  cost  and effectiveness of controlling
 vent  streams from polyethylene  and  polypropylene  facilities at all
 concentrations less than 2,000  ppmv.  Commenters  IV-D-8 and IV-D-13
 concluded, that, unless EPA can establish the availability and cost
 effectiveness of controls  for dilute gas streams, controls should not be
 required.
      One commenter (IV-D-8) believed that the insufficiency of these
data  is not coincidental.   According to this commenter, control (of
                                  2-41

-------
 these  dilute  streams)  is not currently required  under  any 'State  or
 Federal  program.  The  commenter claims EPA alluded to  this  on  page  4-21
 of  BID Vol. I where  it is stated that "... most  of the streams involved
 in  the polymers and  resins  industry are high enough  in heating value  to
 self-combust without using  auxiliary fuels, virtually  no advantage  is
 achieved by using a  catalytic unit and their applicability  in  this
 industry is very limited."
       This commenter (IV-D-8) also stated that if BID  Vol.  I did  in fact
 establish that 20 ppmv  is the lowest VOC concentration  achievable by
 combustion of gas streams containing less than 2,000 ppmv VOC, then the
 lowest achievable VOC  concentration is controlled far  greater  than the
 best technology system  called for by Section 111 of the Clean  Air Act.
 One commenter (IV-D-8)  pointed out that many dilute streams will require
 controls due to the  misclassification of various effected processes (see
 previous comments on LLDPE  units).  However, if these  processes in
 question are reclassified as suggested by this commenter, then the
 problem of controlling  these dilute concentration streams should go
 away.  Finally, the  commenters suggested that EPA should also  establish
 a procedure in the regulation that would exempt any dilute concentration
 stream in any affected  facility from control  if it can be shown that
 control of that stream  is not cost effective.   Commenter IV-D-8
 suggested as one possible way to do this is to consider the application
 of a Total  Resource  Effective (TRE) Index to the subject vent  streams.
 This was also suggested by the other two commenters.    If a suitable
 alternative test cannot be provided,  the commenter (IV-D-8) continued,
 EPA should exempt all low concentration  streams from control in this
 standard using the same rationale it used to exempt the continuous vent
 streams from the product storage bins.
 Response:
      The BID for the proposed standards summarized information
contained in Docket  Items II-B-4 and II-B-5.   These docket items contain
the conclusions reached by EPA concerning thermal incinerator
performance over a number of tests.  The Agency believes that the
conclusions reached in those docket items are  still  valid and support
the proposed standard.   Furthermore,  the new approach presented in the
January 10, 1989,  Federal  Register notice and  adopted in the final  rule
                                  2-42

-------
 considers a wide range of low VOC concentration streams and the cost of
 their control for determining the level of emissions necessary for
 control to be cost effective.  The final rule incorporates a low VOC
 concentration exemption for new and existing affected facilities.
       With regard to specific points raised by the commenters, the
 Agency encourages industry to seek ways to maximize raw material
 recovery.  The analysis performed by the Agency is still valid,
 nevertheless, for cost-effective control of emissions that remain after
 the material  recovery.
       One commenter (IV-0-8)  suggested that 20 ppmv is more stringent
 than the best technology system called for by Section 111  of the Clean
 Air Act [presumably calculating 99 percent reduction (2000-20/2000 =
 0.99)], which is more stringent than the 98 percent destruction.   The
 commenter is  incorrect,  but  the Agency understands the confusion.   In
 evaluating the performance capabilities of incinerators, the Agency
 examined a large number  of streams that had combustion air added  to them
 prior to being combusted.  For  these streams,  the  tests  showed  a
 leveling off  at  20  ppmv  at the  outlet  when  the  concentration  of streams
 with combustion  air fell  below  1,000 ppmv.   In  other words,  98  percent
 destruction was  still being achieved by emission streams with  combustion
 air that had  VOC concentrations  down to 1,000 ppmv.   Many  of  the  streams
 examined,  however,  required combustion  air  to be added to  them.  The
 amount  of combustion  air  required  typically reduced  the  VOC
 concentration  by one-half.  Thus,  2,000 ppmv of VOC  before combustion
 air is  required  could be  incinerated by 98  percent  (on a weight basis).
 In  summary, the  20  ppmv standard does not require  a greater degree  of
 control  than the  best technology system, which is  98  percent destruction
 by  weight.
 Comment:
      One commenter (IV-D-39) stated that they attempt to control all
 continuous VOC streams that are available at elevated pressures and it
 has been only the very low pressure, dilute VOC streams that are
continuous (i.e., dryers, pellet tanks) which have been vented to the
atmosphere.  The commenter stated that the proposed regulations would
possibly bring these streams  under control, but the cost/benefit would
be extremely high.
                                  2-43

-------
Response:
      The Agency has carefully considered the VOC reduction obtained  and
control costs  incurred when controlling dilute VOC streams.  The
procedure presented in the January 10, 1989, Federal Register notice  for
determining which process emission streams should be controlled examined
existing data  on both polypropylene and polyethylene emission streams,
Including very dilute streams.  The control/no control determination
procedure developed by the Agency is designed to avoid extremely high
costs when compared to the resulting VOC emission reduction.  Where
other comments have identified specific instances of potential problems,
the Agency has reexamined the procedures and the cost analysis.  The
final rule incorporates some changes.  Nevertheless, the Agency believes
the procedures in the final rule for determining which low VOC
concentration  streams should be controlled is a more than adequate
procedure for  avoiding the general situation described by the commenter.
Comment:
      Four commenters (IV-D-39, IV-D-43/IV-D-50,  IV-D-.44, IV-D-45)
expressed concern over the lack of an exemption for streams with less
than 0.1 weight percent VOC from new affected facilities.  Commenter
IV-D-39 stated that the Agency's reason for excluding new affected
facilities from the low VOC concentration exclusion was to preclude
operators from purposefully diluting streams to benefit from the
exemption.  This commenter pointed out that "intentional  dilution is
circumvention of control,  which is already forbidden."  This commenter
then stated that EPA should not promulgate rules  which could possibly
Impose non-cost effective controls on some law abiding operators to
hinder those who may otherwise circumvent the regulations."
      Commenter IV-D-44 stated that the Section 111 of the Clean Air Act
requires NSPS's to be economically justified.  The  commenter stated that
the Agency's "only justification for proposing control  of sources less
than 0.10 weight percent is the difficulty in determining if dilution to
the low concentration was  appropriate design or circumvention."   The
commenter believes that this "difficulty" in evaluating sources  does not
justify violating the Clean Air Act by imposing controls  that are not
economically justified.   The commenter stated further that there has
been no justification for control  of low concentration sources from new
                                  2-44

-------
 facilities as proposed and 'that,  if new sources have a different cost of
 control,  the minimum concentration at which controls are justified may
 be less than the 0.10 weight percent VOC proposed for sources from
 modified  facilities, but not zero percent.   This commenter concluded by
 stating that low concentration streams from new facilities should meet
 the same  economic justification criteria (as for those streams from
 modified  facilities) and that, unless a difference in economic
 justification is shown by new data,  no sources  less  than 0.10 weight
 percent VOC should be controlled.
       Commenter  IV-D-43/IV-D-50,  states that EPA,  in the support
 material,  has not provided justification or established  cost
 effectiveness for control  of low  concentration  hydrocarbon  emissions
 from new  sources.   The commenter  stated that if for  very dilute  streams
 in  modified plants it is  true,  as  EPA stated, that "below 0.10 weight
 percent VOC,  the cost of  control  becomes  so  large  that control is  not
 cost-effective regardless  of the  amount of  emissions, then  it  should
 also  be true  for such streams  in  new  facilities  because  there  is  still
 no  real cost-effective  control  technology available.  The commenter
 states  that the  lack of a  low  concentration  exemption for new sources  is
 a serious omission.   This  commenter identified  several reasons as  to why
 very  low VOC  concentrations  [0  to 5 percent  of  the lower explosive level
 (LEL)]  are  found  in  the industry and  stated  that EPA should not require
 companies to  compromise safety  by requiring  higher VOC concentrations
when  history  has led  industry to use more dilute levels to ensure
 employee safety.   This commenter also provided the following suggestions
 in response to EPA's concern that new.facilities could use excessive
dilution to escape regulation:

(1)  Due to the large amounts of purge air needed to operate existing
     plants safely, the horsepower requirements  to provide the purge
     air is very  large.  Since it costs a great  deal  of money  (300
     to 400 horsepower units at approximately 25 to 30 dollars per
     horsepower equals as much as  $12,000 annually) to operate the
     blowers,  it  is unlikely that  an  operator would want  to
     unnecessarily increase that cost.
                                  2-45

-------
 (2)   Most  new  facilities  are  being  built  by  companies  that  already
      produce polyethylene.  If  a  company  wanted  to  intentionally
      circumvent  the  law by  using  more  air than needed  for a new
      process line  or plant, it  would be hard to  justify  if  the
      company were  already operating an existing  line at  much lower
      purge  air flows.  The  EPA  or the  state would only have to  compare
      the new versus  the old air flow rates.  If  the flows are much
      different,  the  owner should  have  to  justify the difference.

      Commenter  IV-D-45 recommended that  EPA use the same procedures  for
determining control  of dilute VOC streams  (<5.5 weight percent  VOC) for
new affected facilities as  it did for modified and reconstructed
facilities, and  that  EPA promulgate the same provisions  for determining
control required for  new facilities as it  has proposed for  modified and
reconstructed  facilities.  The  commenter  stated that not to do  so  would
subject the proposal  to allegations that  it is not representative  of  all
affected facilities  in the categories being regulated,  hence defective.
This  commenter addressed EPA's  request concerning excessive dilution  of
streams to take  advantage of a  low VOC concentration cutoff for new
facilities.
Response:
      The Agency has carefully considered this issue.   The Agency  is
still concerned that new facilities could be designed so as to  "take
advantage" of a low VOC concentration cutoff, and at less expense  than
the cost of controlling the streams.  One commenter (IV-D-43/50)
indicated that the Agency could compare the purge air rates of a new
facility with existing rates.   The Agency agrees  that this could be a
useful mechanism to evaluate whether a company is trying to
intentionally dilute the stream to circumvent the rule.  The Agency is
concerned that this is not necessarily applicable in all  cases,  either
because it is a first time facility for an owner  or operator or because
it represents a sufficiently different technology that  a comparison is
inappropriate.   Another commenter (IV-D-45) pointed out that the nature"
of storage bin purges are such that the VOC concentration is initially
high and drops over time,  and  that such equipment and  air purges are
designed so that the maximum concentration would  be between  20 and 25 of
                                  2-46

-------
 the LEL.  The Agency believes that recognition of this design feature is
 a useful tool for evaluating whether "too much" dilution is taking
 place.  This again is somewhat limited in that not all dilute streams
 will have VOC concentrations that vary as dramatically as might occur in
 storage bins.  On the other hand, designing storage bin air purge rates
 so as to have maximum VOC concentrations at 20 and 25 percent of the LEL
 makes it more difficult to justify streams that are diluted so that the
 maximum VOC concentration is significantly below this level.  By
 examining these and other items (e.g.,  design criteria for pneumatic
 conveyors), the Agency now believes that there are a sufficient number
 of indicators that can be used to judge whether intentional
 circumvention is being practiced at new facilities.   Therefore,  the
 Agency has extended both the VOC concentration exemption  and the
 calculated threshold emission equations used for existing affected
 facilities to new affected facilities  (see Section 60.560(g) and
 Table 3 of the final  rule).
 Comment:
       Two commenters  (IV-D-44,  IV-D-50)  pointed out  that  low
 concentration continuous emission  sources  from modified and
 reconstructed facilities in  the  range of 0.10  < 5.5  weight percent  VOC
 are  required  to  be controlled  only  if the  total  emission  from  these
 sources  exceeds  that  amount  calculated  by  the  calculated  threshold
 emissions (CTE)  formulae,  but  the  same  sources  from  new construction are
 required  to be controlled  if the total  emission  exceeds a fixed  amount
 of 47  Mg/yr.
       Commenter  IV-D-44  stated that assuming that the CTE formulae were
 derived to reflect that  lesser concentrated  sources are not  justified to
 control unless the total emissions are greater, the same should hold
 true for  cost analyses of economic justification regardless  of whether
 the source is in new or modified facilities.  According to this
 commenter, if control costs are different for new vs. modified
 facilities, the CTE formulae may be different, but not a fixed threshold
value.  Commenter IV-D-44 indicated that no economic justification was
shown in the docket items for control  above the fixed limit  (47 Mg/yr)
for sources in the range of 0.10 < 5.5  weight percent VOC.
                                  2-47

-------
       Commenter  IV-D-50  supported  the  use  of  the  CTE  formulae saying
 "such  a  procedure,  as  required  by  law,  consider:;  the  cost  of abatement
 and  the  problems  associated with controlling  dilute VOC  streams  when
 determining whether controls  are justified."  This commenter then  stated
 that using these  formulae will  not discriminate against  the  size of the
 facility since the  amount of  emissions  is  considered  in  the  equations.
 Commenter IV-D-50 then stated that despite  admitting  that  the emission
 amount must be considered to  determine  whether controls  are  cost
 effective for dilute streams  (<0.60 weight  percent),  EPA has  abandoned
 the  rationale previously stated for modified  sources  by  setting  a
 47 Mg/yr VOC emission cutoff  for all streams  with a VOC  concentration
 < 5.5 weight percent from new plants, and that this approach  does  not
 recognize the problems and costs associated with controlling  dilute VOC
 streams.  According to this commenter,  a 47 Mg/yr cutoff discriminates
 against  larger sources.
      Both commenters recommended that  the control of both new and
 modified  facility low level concentration streams (0.10  <  5.5 weight
 percent)  be determined by the same CTE  formulae.  Commenter  IV-D-44
 noted that this would also greatly simplify the regulation by
 eliminating the difference between new  and modified sources and  would
 satisfy  the economic justification requirement of the Clean Air  Act.
 Response:
      As  noted in the previous response, the CTE formulae  for existing
 facilities are now applied to new affected facilities.
 Comment:
      One commenter (IV-D-44)  stated that a new equation is needed for
 calculating threshold control  levels for VOC streams in the 0.10 < 0.12
weight percent range, because the upper end point calculated for this
weight percent range and the lower end point calculated for the  0.12 <
 0.2 weight percent range are not the same (and they should be).
 Response:
      The Agency agrees that the end points of the equations referred to
by the commenter should agree.  The equations  presented in the final
rule have end points that agree.
                                  2-48

-------
  Comment:
       One  commenter  stated  that  for  streams  containing  less  than
  0.1 weight percent VOC  concentration,  control  to  98  percent  destruction
  is not  always technically feasible and  is generally  not  economic.   The
  commenter  then stated that  catalytic  incinerator  vendors,  for  example,
  will often not guarantee above 95 percent efficiency, and  thermal
  incinerators which can  achieve 98 percent are  not economic for low  VOC
  concentrations.  The commenter then suggested  that to solve  the problem
  of a lower weight percent cutoff for new affected facilities at
  virtually  the same level of environmental protection, the Agency modify
  the rule to require 95  percent control for individual streams  from  new
  process sections with more than 47 Mg/yr total combined emissions.
 According  to the commenter,  this modification addresses the two main EPA
 concerns:
 --  EPA did not want to extend the 0.1 wt weight percent cutoff allowed
     for modified facilities  to new facilities because new designs, given
     no existing baseline operating data,  might simply adjust  dilution to
     achieve the cutoff.   However, a 95 percent control  would  prevent
     this particularly since  the capital  and  operating costs of such
     controls  would increase  with  flowrate (dilution).
 --  For  environmental  impact,  the difference  between  95  and 98 percent
     control for  a stream with  <0.1  weight percent  VOC is  quite
     negligible.   Thus,  setting the  control requirement to 95  percent for
     such dilute  streams  would  allow the  owner to  select  the most cost-
     effective control  (catalytic  versus  thermal  incineration,  for
     example) without  materially contributing  to air emissions of VOC.
 Response:
      As noted in  an  earlier response,  the Agency  has decided  to extend
 the 0.1  weight percent VOC cutoff to new  affected  facilities,  although
 not for  the reasons suggested  by  this commenter.   Further,  in comments
 received on the January  10,  1989, Federal Register notice and  in
 conversations with individuals in the industry, the Agency continues to
 find that catalytic incinerators  can be and are designed for 98 percent
destruction.  Therefore, the Agency has retained 98 percent destruction
as the standard.
Comment:
      One commenter (IV-0-44) stated that the Agency's review of cost
effective controls appears to assume that all  low concentration VOC
                                  2-49

-------
 streams  are  equally  cost  effective  to  control  by  catalytic  incineration
 if the sum of  their  emissions  exceeds  the  CTE  level.   The commenter
 pointed  out  that  some  streams  may require  additional  pretreatment  to
 assure that  particulates  (polyethylene) will not  foul  the catalyst.
 Alternate control  devices  (i.e., spare multi-bed  systems),  or  allowances
 for equipment  defouling should be considered in the control  cost
 effectiveness.  The  commenter  recommended  that each potential  candidate
 stream for control be  reviewed to see  if additional cost for particulate
 control  and  equipment  spares is required,  thus changing the  economic
 justification  threshold in order for a proper cost effective evaluation.
 Response:
      As a result  of further investigations, the Agency agrees that
 additional pretreatment to ensure particulates will not foul the
 catalyst is  required in order  to install a catalytic  incinerator.  The
 additional investigation revealed that "primary" particulate control
 equipment (e.g., baghouses) is already in  place on the emission streams
 of  concern,  so that  the cost of the catalytic incinerator system would
 now include  filtering  units to ensure that the particulates  that remain
 in  the stream  after  the primary particulate equipment are reduced to
 avoid fouling  of the catalyst.   The final  rule incorporates  these
 additional costs for those streams that are likely to be controlled by
 catalytic incinerators (i.e.,   streams with VOC concentrations less than
 5.5  percent VOC by weight).

 2.6.2 Low Flow Streams
 Comment:
      One commenter  (IV-D-43/IV-D-50) pointed out that, under the
 approach presented in the January 10, 1989, Federal Register notice,
 individual continuous emission  streams, with flows of 8 standard cfm
 (scfm) or less, require control.   The commenter stated that  using a
 flare to control  certain 8 scfm continuous  vent streams (those that
 contain dangerous levels of oxygen)  is  not  appropriate, because
 introducing these vent streams  into  a flare can cause an explosion.  The
 conmenter referred to vent streams  from compressor lubrication pots,
which may contain oxygen in which case  they could not be sent to a flare
 because of the risk of explosion.   The  commenter then stated that EPA
                                  2-50

-------
  should  reconsider  whether  control  of  those  continuous  process  streams  of
  8  scfm  or  less  that  could  contain  oxygen  is  appropriate.
       Another commenter  (IV-D-49)  stated  that  the  requirement  for
  control of streams with  flowrates  below 8 scfm (regardless  of
  concentration)  seems to  penalize owners with small emission  streams
  compared to those  owners with larger  emission  streams.  This commenter
  then stated that control of 8 scfm streams may  not be  economic  in all
  cases:  for example, if  the stream is at a pressure below the flare
  backpressure, or if  it contains a high oxygen  content.  This commenter
  recommended deleting the 8 scfm rule  entirely.
       A third commenter  (IV-D-44) also stated that some streams may
  contain oxygen,  making them unsafe to flare, and recommended that any
  stream that cannot be controlled due to safety  considerations be
 exempted.
 Response:
       The  Agency has not changed the rule  for low flow streams.   An
 analysis undertaken for the January 10,  1989, Federal  Register notice
 (see Docket Item IV-B-7)  showed  that low flow streams  with air would
 constitute such  a small  portion  of  the total  volume of emissions being
 combusted  that the  composition of the  combined  stream  is  changed only
 marginally.  Further, there are  various  methods available  to minimize or
 alleviate  safety concerns,  such  as  design  features  for  flares,
 substitute  inerts as  purge  materials,  or purging with  inerts  downstream.
 As  a generic rule,  there  will always be  some  streams that  are more
 costly to  control as  well as some streams  that  are  less costly to
 control  than  the typical  stream.  The  Agency  finds  no reason  to  further
 complicate  the final  rule with specific provision aimed at
 distinguishing between low  flow  streams.
 Comment:
      One commenter (IV-D-44) recommended that  the Agency  include cost
 effectiveness to control as site specific criteria for  all continuous
 8 scfm or less streams.  The commenter pointed out that all streams of
 8 scfm or less are required to be controlled, either by (1) 98 percent
 destruction or to 20 ppmv, whichever is less  stringent,  or (2) in an
existing flare,  incinerator, boiler or other  control device located at
the plant site.   The commenter then  noted that the Agency determined  the
                                  2-51

-------
 control  of 8  scfm or  less streams  in flares to be economically
 justifiable,  but that no such economic justification was made for  other
 control  devices.  The commenter stated that NSPS are required to be  the
 best documented technology  (BDT) economically achievable (less  than
 $2,000/Mg of  pollutant removed) and asked:  "What basis is the  Agency
 using to require control in devices other than flares?"  According to
 the commenter, these alternate control devices may yield less pollutant
 destruction and be more expensive  to gain access.  The commenter then
 stated that flare lines are usually readily available to tie-in
 additional streams, but this would not be the case for incinerators,
 boilers, and  furnaces located at the plant site.  According to  the
 commenter, costs for piping of the 8 scfm or less streams to these
 devices  may be significant and, as the requirement in Section .60.562-
 l(3)(a)(l)(11) of the January 10,   1989, Federal  Register notice
 specifies that if a boiler or process heater is used to comply,  the vent
 stream shall  be introduced into the flame zone of the bo'iler or process
 heater,  this  provision could also  be additionally expensive.
 Response:
      An owner or operator may elect to send such streams to any control
 device that is available.   As flares are the most commonly available
 control devices, the Agency examined the costs to send such emissions to
 them.  The Agency then allowed the use of control devices other than
 flares to give owners and operators specific options.   The cost  of using
 other control  devices may or may not be more expensive.
      The commenter states that incinerators,  boilers,  and furnaces may
yield less pollution destruction.   These devices can only be  used to
 control 8 scfm or less streams in  certain situations,  as  explained in an
 earlier response and in the January 10, 1989,  Federal  Register notice.
 In such situations,  BDT is not 98  percent destruction.   Furthermore,  in
 such situations,  the stream need not be introduced  into  the flame zone
 as that requirement is associated  with  98-percent destruction.   This
latter point  is clarified in the final  rule.
      In summary,  the Agency does  not need to  perform cost  calculations
 for each and every conceivable situation.   The justification  of  a
commonly available control  technology that will  work in  all  cases is
 sufficient.   The Agency believes that the cost basis used  to  justify
                                  2-52

-------
  control  of  all  8  scfm  or  less  streams  is  reasonable.   As  a  generic  rule,
  there will  always  be some situations where  the  actual  cost  may  be higher
  as well  as  lower.   The Agency  sees  no  compelling  reason to  further
  complicate  the  rule by trying  to  incorporate  the  commenter's  suggestion.
  Comment:
       One commenter (IV-D-50)  stated that the Agency's suggestion that
  there is adequate  excess  capacity to accept and control all 8 scfm  (or
  less) streams may  not  be  true  on a  pi ant-by-plant basis.  The commenter
  recommended that EPA include cost effectiveness of control  as a
  criterion for controlling 8 scfm or less continuous process streams.
  Response:
       The Agency has not  revised the final rule to incorporate a
 criterion for determining when it may or may not be cost effective to
 control  all  8 scfm or less streams on a plant-by-plant basis.   In a
 sense,  the problem presented by the commenter is that if there is not
 sufficient excess capacity in a control device,  then the cost  of a new
 control  device to control  all  8 scfm or less emission streams  is not
 cost  effective.   The Agency reexamined  expected  excess capacities and
 the amount of excess capacity required  before  the  threshold  emissions
 levels  are reached for  various  weight percents (see  Docket Item
 IV-B-24).   It  is important to  remember  that  (I)  all  streams  in an
 affected facility  that  has annual  emissions  equal  to  or greater  than the
 calculated threshold emissions  (CTE) level for the particular  weight
 percent  of VOC being examined are  to be controlled regardless  of  the
 flow  rate of the individual  stream;  (2)  the  concern  is  where the  total
 emissions are less  than the  CTE level;  and (3) total emissions come  from
 streams  with flows  greater than 8 scfm  as well as  those with flows less
 than  8 scfm.
      The Agency's  reexamination showed  that as the VOC concentration of
 the combined stream of  all 8 scfm or less streams  decreased, the  amount
 of excess capacity  needed  increased.  To illustrate this, suppose the
 VOC concentration of the combined streams was 10 percent VOC by weight.
 Then,  an existing control  device would need 12 to  19 scfm of available
 excess capacity to  accept this combined  stream.  Suppose instead, the
concentration of this combined stream was 4 percent VOC by weight.  The
existing control  device  would now need  62 to  93 scfm of available excess
                                  2-53

-------
 capacity to  accept the stream.   If the total flow of the combined  stream
 is higher than that  indicated for ei-ther of the two VOC concentrations,
 then the total annual emission of the combined stream  is greater than
 the CTE level for that weight percent, and control in  a new control
 device is cost effective.  Thus, if the existing control device did  not
 have the available excess capacity needed, the combined stream could be
 sent to a newly constructed control device, and to do  so would be  cost
 effective.
      If the total flow of all 8 scfm or less streams  is close to  the
 scfm that is equivalent to the CTE level, then the cost of control of
 the combined stream  in a newly constructed control device is only
 slightly higher than that cost effectiveness used to guide the selection
 of the CTE levels.   If the total flow of all 8 scfm or less streams  are
 further away from the flow that is equivalent to the CTE level, then the
 total flow is more likely to fit into available excess capacity.   In
 addition, the required excess capacities calculated by the Agency,  in
 the example above, assumed that the each stream being considered for
 control in an existing control device has a flow of 8 scfm or less.  The
 regulation does not require control  of individual  streams with flows
 greater than 8 scfm in affected facilities whose total  emissions are
 less than the CTE level.   Thus,  the available excess capacity that is
 actually needed could be substantially less once streams with flows
 greater than 8 scfm are factored out.   Finally,  the rule proposed and
 adopted is a general  rule.  The commenters seek to exclude streams  only
where there may be insufficient excess capacity,  but do not propose to
 include larger streams (e.g.,  all 16 scfm streams)  at plants where  there
may be sufficient excess  capacity.   If the rule were made specific, it
would also catch other streams that  would be greater than 8 scfm.   As a
general  rule, there will  be some situations in  which it will  be more
costly to comply and others where it will  be less  costly.   The Agency
continues to believe that the  8 scfm or less rule  is an appropriate one
that will  not cause undue hardship  on  any one facility  and has not
changed this provision in the  final  regulation.
                                  2-54

-------
 2.6.3 Polyolefin Plants
 Comment:
       Two commenters (IV-D-8,  IV-D-9)  expressed concern that the gas
 phase model  plant used as the  basis of the proposed standards for LDPE
 including linear LDPE (LLDPE)  does not accurately represent plants that
 produce LLDPE using the liquid solution process.
       Commenter IV-D-8 stated  that burning emissions from the product
 finishing area of plants using a liquid phase solution process for
 producing HOPE and LLDPE is not practical  due to  the low concentrations
 of hydrocarbons in the stream.   The commenter pointed out that some
 plants,  depending on process design or plant  layout,  may be able to send
 these dilute VOC streams to a  boiler or process heater as a part of the
 air supply.   If a plant must install a boiler to  meet the proposed
 limits for dilute streams,  the  commenter continued,  then the energy cost
 would be excessive for a small  VOC reduction.   Commenter IV-D-9  stated
 that  a solution plant  manufacturing LLDPE  more  closely resembles the
 HOPE  solution  model  plant used  to  determine the regulatory  alternatives
 than  it  does the gas phase  model plant.  This commenter  recommended  that
 processes  that make  LLDPE by the liquid  phase solution  process be
 classified,  for control  purposes,  with  the HOPE liquid  solution  process
 units.   Commenter IV-D-9 proposed  the  following wording  for  Sections
 60.560(a)(l)(v)  and 60.560(a)(2)(vi) of  the September  30, 1987,  Federal
 Register notice:

 Section 60.560(a)(l)(v)
       For the manufacture of high density polyethylene and low
      density  polyethylene  (including linear)  using a liquid phase
      solution process:  each raw material  preparation section and
      each material recovery section;

Section 60.560(a)(2)(vi)
      For the manufacture of high density polyethylene and low
      density polyethylene (including linear)  using a liquid phase
      solution process:  Each raw materials preparation section,  each
      polymerization reaction section,  and  each  material  recovery
      section.
                                  2-55

-------
       Commenter  IV-D-9  stated that  if this change  is not made,  the  pro-
 posed  regulation would  not be cost  effective  for LLDPE  liquid  solution
 units.  The commenter pointed out that under  the proposed  regulation  a
 plant  manufacturing LLDPE is required to  "achieve  a 98  percent  reduction
 of emissions of TOC contained in the gas  streams continuously discharged
 to the atmosphere from  each new, modified, or reconstructed raw material
 preparation and product finishing section."  The commenter stated that
 the product finishing section in their LLDPE units, by  definition,
 potentially could consist of mixers, extruders, dryers, check hoppers,
 blenders, and silos and their emissions would have to be controlled.
 The commenter pointed out that emissions  from these pieces of equipment
 come from residual solvent left in the polymer which slowly exudes  from
 the polymer bead, and to prevent buildup of flammable hydrocarbon
 mixtures, they dilute the emissions from the polymer beads with an  air
 purge before venting the emissions to the atmosphere.  According to the
 commenter, these vents have the same type emission characteristics  as
 vents from product storage bins, which in the proposed regulation,  the
 Agency determined were not practical or cost effective to control
 because of their emission characteristic (VOC content is very dilute).
 The commenter indicated that the Agency had determined,  for the same
 reasons,  that the volatile organic compounds in the dilute air  streams
 found in the product finishing area of the HOPE liquid solution
 processes need not be regulated.  The commenter then concluded that
 dilute vents from the LLDPE liquid solution process should not be
 controlled either,  and for determining proper emission controls, LLDPE
 liquid solution units should be grouped with the HOPE liquid solution
model  plants,  not the LOPE gas phase process.
 Response:
      The Agency agrees with the commenters that the situation described
could result in control  of emission streams that would be unreasonable
given the costs and emission reductions involved.   As noted earlier, the
new approach for determining which process emissions to  control  resolves
these concerns.
Comment;
      One commenter (IV-D-6)  stated that  the regulation  should recognize
the existence of a polymer finishing section in many current low
                                  2-56

-------
 pressure low density polyethylene plants, but should not require
 additional vent control because there should be no incremental emissions
 generated in a plant with this additional material handling facility.
 The commenter stated that the VOC emissions should be the same as shown
 for vent "J" in the model plant.
       Another commenter (IV-D-8) stated that in order to be consistent
 with other sections of the standard, the proposal  must exclude Stream J,
 bin vents, in the low density gas phase polyethylene process from
 regulation in Subpart ODD.  The commenter noted that this exclusion
 would be consistent with other process models that have similar
 finishing sections in other process facilities shown in the regulation,
 such as the polypropylene gas phase process.
 Response:
       The Agency is aware, as a result of this and other comments,  that
 the model plants that formed  the basis of the September 30,  1987,
 .Federal  Register notice are outdated for certain existing plants.   The
 new approach  presented  in  the January 10,  1989,  Federal  Register  notice
 addresses the concern raised  by the  commenters.  The  Agency  did not
 revise the  model  plant  description  because  it  would  have no  effective
 impact on the application  of  this standard  to  such facilities.  In  other
 words,  suppose  a product  finishing  section was  added  to  the  low pressure
 LDPE model  plant  at  this time and an analysis  showed  that control should
 be  required of  the emissions.   An applicability date  of  the  Federal
 Register  notice  for  the final  rule would have  to be used.  However, this
 process section  and  its emissions already have a January  10, 1989,
 applicability date.   Thus,  it  is unnecessary to revise the model plant.
 The  new approach would require  control of or exclude  from control  these
 emissions as  appropriate.
 Comment:
      One commenter  (IV-D-6) stated  that the uncontrolled emission limit
 of 10 ppm (by weight) of VOC in Table 1 of the proposed regulation for
 the Product Finishing Section  in low pressure, low density polyethylene
 plants should not be applicable to the product storage hoppers and bins
 in the Product Finishing Section of modernized low  pressure,  low density
polyethylene plants because the Polymer Storage Tank (vent "I"), for
                                  2-57

-------
 which  this  uncontrolled  emission  limit  exemption  was  provided,  does  not
 exist  in  a  modernized  plant.
 Response:
       In  this  comment, the  appropriateness of  a specific  uncontrolled
 threshold emission  rate  from  the  September 30,  1987,  Federal  Register
 notice  is questioned.  The  Agency agrees that  if  the  model  plant
 approach was retained, a reanalysis would be required based upon  a
 modernized  LOPE, low pressure plant.  However, as discussed in  other
 responses in this section,  the Agency has determined  that no  useful
 purpose is  served by undertaking  such a reanalysis.  A reanalysis might
 show a  lower uncontrolled threshold emission rate is  appropriate  for the
 product finishing section in a modernized LDPE, low pressure  plant.
 Since this  would be a more  stringent rule, a new  applicability  date
 would be required.  The new applicability date, however, would  be after
 the January 10, 1989, applicability date, which already applies to all
 facilities.  Thus, the new  applicability date would be inappropriate.
 The reanalysis might show that a higher uncontrolled threshold  emission
 rate is applicable.  The Agency believes that this result is  also
 irrelevant.  Under the rule for determining which process emissions are
 to be controlled, the uncontrolled threshold emission rates come into
 play only where the new approach shows control, but the uncontrolled
 emissions from the affected facility are less than the uncontrolled
 threshold emission rate exemption for that affected facility.   This
 applies only to affected facilities that are constructed,  modified,  or
 reconstructed after September 30,  1987,  and on or before January 10,
 1989.  A stream with an uncontrolled emission rate higher than the
 uncontrolled threshold emission rate in  the September 30,  1987, Federal
 Register notice might have been exempted if its emissions  are less than
 the reanalysis' result.  The Agency does not believe it appropriate to
 extend such further potential  exemptions.   The new approach is the
 procedure to be used for controlling appropriate streams without the
 need to develop new model plants or modify existing ones.   The retention
of the uncontrolled threshold  emission rates was only to prevent owners
 and operators who proceeded to make control/no control decisions based
on the September 30, 1987,  from being penalized from an enforcement
perspective.
                                  2-58

-------
 Comment:
       One commenter (IV-D-11) requested the Agency to more closely
 review the impact of the proposed standards on the commenter's
 polyethylene wax facility, if second stage reactors are to be considered
 part of the product finishing section.  The commenter stated that the
 plant's process and products are unique to the industry, but the Agency
 failed to take into account this type of facility during the development
 of the proposed rule.   Consequently the proposed rule, according to the
 commenter, does not adequately address the unique characteristic
 associated with second stage reactors and may impose unreasonable and
 unrealistic standards  on the plant.   Therefore,  the commenter believes
 the Agency may need to consider developing separate standards for
 polyethylene wax facilities.
 Response:
       As stated in  an  earlier response,  the Agency considers  the second
 stage reactors described by this commenter to be  a part  of the
 polymerization reaction section.   In  response to  the  January  10, 1989,
 Federal  Register notice,  this commenter  agreed with EPA's  preliminary
 decision that  the model  plants  used  in the original September,  1987,
 proposal do  not accurately  reflect many  existing  polyethylene processes.
 The commenter  noted  that  the  original September,  1987, proposal
 classified their  low molecular weight polyethylene  wax manufacturing
 facility as  a  low density low pressure polyethylene process,  and that
 their comments  submitted  characteristics associated wi'th a polyethylene
 wax manufacturing facility  and described the  numerous differences
 between  their plant  and EPA's model plant, which was based on Union
 Carbide's  UNIPOL  process.   The commenter then  concluded by stating that
 EPA's  proposal  to drop the  model plant approach for polyethylene
manufacturing facilities  is supported by the information supplied in
their  original  comments.  Therefore, the concerns of this commenter with
regard to their specific process have been addressed in that Federal
Reqister notice, and .the Agency has not developed separate standards for
polyethylene wax facilities.
Comment:
    One commenter (IV-0-3) believes that  small vents from affected
sections of the UNIPOL  polyethylene and polypropylene plants should not
                                  2-59

-------
 require  control  if  below  a  deminimus  level.   This  commenter believes
 that  the control  of all ven-ts  from  affected  sections  of  these  UNIPOL
 processes is  unjustified  because  they  are  very  small  vents  (containing
 more  than 20  ppm  VOC,  but only a  few  tons  per year of VOC)  that  would
 need  to  be compressed  to  overcome flare header  pressure  and therefore
 represent a significant cost.   The  commenter suggested an allowance for
 a deminimus amount  of  VOC,  such as  5  tons  per year per point source or
 20 tons  per year  per affected  section, or  a  deminimus allowance  total of
 0.15  lbs/1,000 Ibs  of  production.
 Response:
      The Agency  has incorporated a deminimus level into the standards.
 The deminimus level was presented for public  comment  on January  10,
 1989, and has been  retained in  the  final rule.

 2.6.4 Polystyrene Plants
 Comment:
      One commenter (IV-D-6) stated that smaller polystyrene facilities
 of the size assumed in the regulatory impact  analyses would typically
 have emission rates too low to measure if emitting less than the current
 0.12 kg TOC/Mg product standard.  Therefore,  the commenter  stated
 exemption  rates proposed would not be available to these facilities.
 Response:
      As  noted elsewhere,  the Agency disagrees that emission rates would
 be too low to measure streams from "smaller" polystyrene facilities.
The uncontrolled threshold emission rate,  however,  has changed (it has
been increased to 0.05 kg  TOC/Mg product)  due to a cost reanalysis (see
Docket Item IV-B-18).
Comment:
      One commenter (IV-D-6) claimed that  the proposed emission limit of
0.0036 kg TOC/Mg product or outlet gas temperature of -25C  from each
final  condenser in the material recovery section of a polystyrene plant
is not appropriate for vent gas streams that contain  significant
quantities of water.  The  commenter stated  that  in their experience
moisture will  be present in a polystyrene  plant  material  recovery vent
stream and that this moisture comes from the water content of the
styrene and rubber used in HIPS manufacture.   The commenter  stated that
                                  2-60

-------
 the moisture  content  in  the  rubber  used  in  HIPS  manufacture  can  be  as
 high  as 4,000 ppm, with  the  rubber  content  of  HIPS  ranging from  5 to  10
 percent of the  finished  product.  The commenter  pointed  out  that, while
 some  of this  moisture will be driven off through  heat of reaction or
 collected in  water draws  in  the process, saturated  water vapor  (30  to  50
 ppm)  will remain  in the  devolatilizer vent  stream (approximately 140F)
 and cause freezing problems  at the  specified refrigeration/condenser
 control temperature of -25C.  Thus, the commenter  concluded, the
 technology and  economics  for which  these standards  were  based are not
 achievable without inclusion of water removal  systems.
 Response:
       The Agency has reevaluated the regulatory options  for polystyrene
 emissions from the material recovery section, taking into account the
 presence of water or other condensibles.   The results of the new
 analysis show that the proposed standard  of 0.0036 kg TOC/Mg product is
 still  achievable at a reasonable cost considering the amount of emission
 reduction (see Docket item IV-B-18).  The new analysis,  however,  showed
 that if emissions from existing plants  where less than 0.05 kg TOC/Mg
 product,  control to 0.0036 kg TOC/Mg product would not be cost
 effective.   Previously,  this  level was  found to occur at  0.016 kg TOC/Mg
 product.  Thus,  on the basis  of the  new analysis,  the final  rule  sets
 the emission  limit at  0.0036  Kg  TOC/Mg product  with  an uncontrolled
 threshold emission rate  for exempting modified  and reconstructed
 affected  facilities of 0.05 kg  TOC/Mg product.
 Comment:
       One commenter  (IV-E-82) requested the  Agency consider allowing
 emission streams from  polystyrene plants the option  to meet a  98  percent
 destruction standard in addition to  the emission  limit  (of 0.0036 kg
 TOC/Mg product)  and the outlet temperature alternatives.   According  to
 the commenter, the emission stream from a subject facility contains
 components that  are not as  easy to condense  as  styrene and the most  cost
 effective way  to control  such streams is through combustion.
 Response:
      The commenter has indicated the presence of certain compounds  of
which the Agency had not been  previously aware existed in the material
recovery emission stream for a polystyrene plant.  The Agency agrees
                                  2-61

-------
 that  for  streams  containing  components that  are difficult  to  condense,
 the cost  of meeting  the  current  standards  through  the  use  of  recovery
 techniques, such  as  condensers,  could be very expensive  considering  the
 emission  reduction achieved.   In  such a situation  as presented  by  the
 commenter, the owner or  operator  of the affected facility  could
 demonstrate compliance with  the  emission limit by  ducting  the emission
 stream to a combustion device  (e.g., a thermal incinerator),  as the
 standard was proposed.   The  owner or operator could measure the
 destruction efficiency of the  combustion device and apply  that
 destruction efficiency to the  inlet emission level of  the  material
 recovery stream to calculate the expected outlet emission  level of the
 stream.  The Agency  believes that there is little, if  any, adverse
 environmental impact by  allowing these emission streams  to meet a 98
 percent reduction or to  20 ppmv standard.   Therefore,  the  final rule
 specifically adds this alternative standard for polystyrene plants.
 Comment:
      One commenter  (IV-D-6) stated that the proposed  standards for
 continuous polystyrene plants  should include fugitive  emission
 requirements only.
 Response:
      The Agency has determined that process emission  from continuous
 polystyrene plants can be controlled beyond what many plants are already
 achieving.  Therefore, the standard continues to require control of both
 process and fugitive emissions.

 2.6.5 Threshold Levels
 Comment;
      Three commenters (IV-D-7, IV-D-8,  IV-D-13)  recommended that any
 existing facility that becomes modified  should not  have to meet the re-
quirements of Section 60.562-1 if the  facility's existing emissions
 (controlled or uncontrolled)  are already  equal  to  or less than the rates
 in Table 1 of the proposed standards.  The  commenters stated that the
proposed exemption rates failed to consider that there  are some existing
 facilities,  which will  become modified,  that are already achieving
 significant emission reductions through  existing state  and PSD permits,
and that additional  controls are likely  to  be installed on existing
                                  2-62

-------
  facilities  as  a  result  of  State  Implementation  Plan  revisions  for  ozone
  nonattainment  areas.
       Another  commenter (IV-D-47)  stated  that the  criteria  for
  determining applicability  should not  be termed  "uncontrolled emissions,"
  but  "federally enforceable emissions  limits" in order to allow credit
  for  the emissions controls that have  already been  required  by  State Air
  Pollution Control Agencies.
  Response:
       The Agency agrees that the regulation needs  to take into  account
  emission streams that are already controlled as a  result of State
  regulations, especially those that are Federally enforceable.   The new
  approach presented in the January 10, 1989, Federal Register proposed
  that "uncontrolled" emission streams be combined.   The September 30,
  1987, Federal  Register notice proposed "uncontrolled threshold emission
 rates," where  uncontrolled referred to the emissions that would be
 "emitted to the atmosphere in the absence  of any add-on  control
 devices...  ."   The promulgated regulation  requires  an owner  or operator
 to examine  the  uncontrolled emissions (i.e.,  those  that  would be emitted
 to the atmosphere in  the absence  of an add-on  control  device).   Where  an
 emission stream in an affected facility is controlled by an  existing
 control  device  (i.e.,  one that was  operating  before September 30,  1987;
 one that was operating between September 30,  1987,  and January  10,  1989,
 on emissions from a process  section  that was not identified  as  an
 affected facility for those emissions  in the September 30, 1987, Federal
 Reqister notice),  the  inlet conditions  to  the control device  would  be
 examined to  determine  whether  that emission stream  is required  to be
 controlled by BDT.  Individual  streams  that are vented to the same
 control device constitute a single stream.   The following describes how
 control determinations are to  be made for  controlled streams.
      For polypropylene  and polyethylene affected facilities with an
 applicability date of January  10,  1989, or earlier,  the inlet emission
 rate  is compared to the uncontrolled threshold emission rate for the
 appropriate process section and type of emission (i.e.,  continuous or
 intermittent).   If the inlet emission rate  is equal  to or less than the
corresponding uncontrolled threshold emission rate,  no further control
 is required.   However, if the inlet emission rate were to exceed the
                                  2-63

-------
 uncontrolled threshold emission  rate  at  some  time  in  the future,  then
 the new approach for determining which process  emissions would be
 required to  meet the standards,  as  discussed  in  the following
 paragraphs,  would be used  to  redetermine whether these  emissions  need to
 be  controlled to meet the  standards in the  final rule.   The  new approach
 would  also be used in those instances where the  inlet emission rate is
 greater than the corresponding uncontrolled threshold emission rate.
       For polypropylene and polyethylene affected  facilities with a
 January 10,  1989,  applicability  date and for  those facilities identified
 above,  the new approach would be used.   Under the  new approach, the
 annual  emissions of the inlet emission stream would be  compared to  the
 CTE level, which would be  calculated based  on the TOC weight percent  of
 the inlet stream.
       If the emissions (Mg/yr) of the stream  entering the control device
 are greater  than or equal  to the CTE level, then that stream is subject
 to  BDT  (98 percent  reduction, 20  ppmv, control in flare  meeting the
 specified operating conditions).  If the existing control device  is
 meeting  BDT,  then, no  further control of  the stream (or combined streams)
 is  required.   If the  existing control  device  is not meeting  BDT (e.g.,
 only achieves  90 percent reduction), then the emission stream is
 required  to  be  controlled  to BDT  at the  next  available opportunity.   The
 next available  opportunity constitutes the next time the existing
 control device  is reconstructed or replaced or its operation  is changed
 as  the result  of changes in State or local  requirements.  At  such time,
 any  uncontrolled emissions in the same weight percent range  from  any
 affected facility are also required to be controlled.
      If the emissions (Mg/yr) of the  stream entering the control device
 are  less than the CTE level,  BDT  is not required at that time.  Whenever
 the  existing control device is reconstructed,  replaced,  or changed  (as
 discussed above), the controlled  stream is  reevaluated to see  if  BDT  is
 required by  combining its annual  emissions  (measured at  the  inlet)  with
 the  annual emissions of any uncontrolled  vent stream within  the same
weight percent range and comparing it  to  the CTE level.   If  these
 combined emissions  are now equal  to or greater than the  CTE  level,  BDT
 is required  for the controlled and uncontrolled vent streams.
                                  2-64

-------
        It  is  important to note that the  "delay"  in  applying  BDT  to  a
 controlled stream does not affect the timing for applying BDT to
 uncontrolled emission streams.  Application of  BDT is required  for all
 uncontrolled emissions as soon as the total annual  emissions for a
 combined  stream  (or single stream) are  equal to or greater  than the CTE
 level  for the weight percent.
        For all polystyrene and PET affected facilities, if the inlet
 emission rate is less than the uncontrolled threshold emission  rate,
 then the existing control does not need to be BDT.  If the  inlet
 emission rate is greater than or equal to the uncontrolled  threshold
 emission rate,  then the stream is required to meet BDT at the next
 available opportunity (as discussed above).
 Comment:
       One commenter (IV-D-1)  stated that the 98 percent emission
 reduction required under Section  60.562-l(a)(l)(i)  would translate into
 approximately 0.2 kg/Mg (see  Table 1  in  the proposed regulations,
 "Maximum Uncontrolled  Emission  Rates")  for a bulk plant with a  purge
 stream equal  to  1 percent of  the  total  propylene feed,  using the
 rationale for "destruction  efficiencies" developed  in  the preamble.
 This commenter recommended  that the same should  also apply to a  gas
 phase process using  a  feed  gas  stream  of equal quality,  unless of
 course, these continuous  gas  streams would come  under  the rules  for
 "waste gas disposal" and  thus "qualify for inclusion in  the  regulatory
 baseline."
 Response:
       In  a follow-up phone call to the commenter (see Docket Item  IV-E-3),
 it was  discovered that the commenter did not fully  understand Table 1
 and  how the numbers were developed.  The commenter  was identifying  the
 remaining  emissions that would occur as  a  result of 98 percent destruc-
 tion  of purge stream for a bulk plant, which was equal to 1  percent of
 the  total propylene feed, and thought this was what was being reported
 for  the other process section emissions  in Table 1.  The  "maximum
uncontrolled emission rates" refer to uncontrolled  emissions that are
vented  to the atmosphere without passing through an air pollution
control device;  thus, no percent reduction in applied to these streams.
                                  2-65

-------
2.6.6  Emergency Vent Stream Exemption
       Many commenters commented on the exemption of  intermittent
releases as proposed in the September 30, 1987, Federal Register  notice
and as revised in the January 10, 1989, Federal Register notice.   The
comments made were extensive.  Both sets of comments are reported,  in
detail, and then the Agency response is given.
Comment:
       In responding to the September 30, 1987, Federal Register notice,
a number of commenters (IV-D-3, IV-D-4, IV-D-7, IV-D-8, IV-D-13)
requested that an exemption be made for emergency vent streams from
polypropylene plants.  In general, the commenters; pointed out Section
60.562-l(a)(2)(i) of the September 30, 1987, Federal Register notice
requires that each vent stream that emits intermittent emissions  in
affected sections of polyethylene and polypropylene plants to combust
the emissions in a flare, and that most emergency vent streams from
polyethylene plants are exempt from this requirement, but there is no
similar exemption for polypropylene plants.   Three of the commenters
(IV-D-7, IV-D-8, IV-D-13) pointed out that to direct emergency vents
(i.e., relief valves and rupture discs) from the polymerization reaction
and materials recovery sections will  compromise safety due to the
potential for polymer buildup in the discharge to the flare.  One
commenter (IV-D-8) pointed out further that  while Table 1,  (page 36688
of the September 30, 1987, Federal Register  notice), which shows the
vent streams that were intended by the Agency to be controlled by the
new standard, indicates control  only of non-emergency intermittent vent
streams for polypropylene, gas phase and high density polyethylene,
solution and slurry processes, the regulatory language itself does not
exempt the emergency vents from these processes.   To resolve this
inconsistency, this commenter suggested the  exemption for emergency
vents be extended to all  intermittent vents, except those for
which EPA can establish that the cost of control  would meet the same
definition of economic reasonableness as is  applicable to all other
controlled emission sources, and meet the necessary standards of safety
to run the facility.  Therefore,  polypropylene plants should be exempt
from this requirement.
                                  2-66

-------
       Two of the commenters (IV-D-3 and IV-D-4) believe that the control
 of UNIPOL emergency reactor blowdowns from either polyethylene
 production or polypropylene production is not justified.  These
 commenters pointed out that the proposed NSPS does not require control
 of emergency reactor blowdowns from the low pressure gas phase (UNIPOL)
 polyethylene process,  but appears to require control of emergency
 reactor blowdowns from the low pressure gas phase polypropylene process.
 The commenters felt that the process evaluated by the Agency has
 characteristics that are different from the UNIPOL polypropylene
 process, and that the  emergency blowdowns from the UNIPOL polypropylene
 process are of the same nature as those from the UNIPOL polyethylene
 process.  The commenters estimated that the annualized cost of
 controlling these emissions of about 25 tons per year of VOC would  be
 about $325,000,  which  is at a  cost per ton considered unreasonable
 according to the preamble discussion.   Thus,  these commenter believe
 that the UNIPOL  gas phase polypropylene plants should not  be required to
 control  emergency reactor blowdowns.
       Three commenters  (IV-D-7,  IV-D-8,  IV-D-13)  suggested  that  Section
 60.562-l(a)(2)  of the September  30,  1987,  Federal  Register  notice 'be
 revised  to read  as  follows:  "For each  vent  stream  that  emits
 Intermittent  emissions  in an affected  facility,  as  defined  in  Section
 60.560(a)(2)  except for  emergency vent  streams  in  low density
 polyethylene  plants using either  a high pressure or  low  pressure process
 and  high density  polyethylene  plants using a gas phase or a  slurry or a
 solution process, or polypropylene plants:   (i) combust  the  emission  in
 a flare."
      Another commenter  (IV-D-5)  recommended that the exemption provided
 emergency  gas streams from low density polyethylene plants using either
 a high pressure or  low pressure process and high density polyethylene
 plants using a gas  phase  process be extended to all other polymer
 processes  that employ multistage relief systems.  This commenter felt
 that requiring flare systems-to be designed to handle the very
 infrequent, large volume releases would result in unreasonably large and
expensive  flares.  This commenter also felt that extending such an
exemption would provide control (flaring)  of most releases while
enabling the safe relief of infrequent releases.
                                  2-67

-------
Comment:
       In responding to the January  10, 1989, Federal Register  notice,
most commenters  (IV-0-39, IV-D-42,  IV-0-43,  IV-D-44, 45, 46, 47, 48)
expressed concern over the costs of controlling certain  intermittent
streams that would be incurred because of the way the types of
intermittent streams were defined in the January 10, 1989, Federal
Register notice.  Commenter IV-D-39 pointed out that their HOPE and PP
plants do not contain decomposition streams, but they do have  runaway
reactions.  The commenter pointed out that they have in the past
installed rupture discs under pressure safety valves to vent to the
atmosphere emissions from these runaway reactions.   According  to the
commenter, these emissions would not be allowed to be vented to the
atmosphere under the proposed rule and significantly higher equipment
costs  (e.g., flare header size) would be incurred.   (IV-D-39.6)
      Two commenters (IV-D-42 and IV-D-48) expressed concern that the
definition of "decomposition" does not appear to include the UNIPOL
emergency reactor blowdown.   These commenters indicated that these
emergency blowdowns are very infrequent and occur as a result of a
runaway reaction.  The commenter stated that the consequences of not
having an emergency blowdown are quite costly (because of a resin
meltdown in the reaction).  Commenter IV-D-42 stated that the
alternative of using a very  large flare to handle these very short
duration emergency blowdowns is extremely expensive,and has been
documented in their comments to the  September 30,  1987,  proposed rule.
This commenter then provided a description of the UNIPOL emergency
reactor blowdown:
          An emergency blowdown is necessary when cooling gas
          circulation  in  the reactor is lost -  to stop the reaction,  to
          stop temperature rise,  and to prevent meltdown of the resin.
         An emergency reactor blowdown happens because of a  cycle
          compressor outage  predominately caused by  electrical  outage
          and infrequently by  instrument  or mechanical  malfunction.
          False instrument readings  can cause this but are rare because
          of reliability  of  today's  instrumentation.
          The consequences of  not taking  an emergency  reactor blowdown
          is meltdown  of  the resin in  the reactor into a large  chunk.
          This results in a  2  to  3 week shutdown costing approximately
          $500,000 ($100,000 for  remedial  maintenance  and $400,000 in
          lost income).
                                 "2-68

-------
       Commenter  IV-D-42  concluded by recommending  that  the UNIPOL
 emergency reactor  blowdown  be  excluded  from  control  in  the same manner
 as  "Decompositions",  and offered  the following  definitions for
 consideration:
       1.   "Decomposition" means for  the purposes of  this  standard an
           event  in a  polymerization  reactor  that advances  to  the point
           where  the polymerization reaction  becomes  uncontrollable,
           resulting in runaway temperature or pressure  increase.
       2.   "Decomposition" means for  the purposes of  this  standard an
           event  in a  polymerization  reactor  that advances  to  the point
           where  the polymerization reaction  becomes  uncontrollable, or
           the polymer begins to break down (decomposes),  or  it becomes
           necessary to relieve the reactor instantaneously in order to
           avoid  catastrophic equipment  damage or serious  adverse
           personnel safety  consequences.

       This commenter  stated that  based  on their knowledge  of
 polyethylene and polypropylene processes  (including  the high  pressure
 process which is the  only one  that has  what  Union Carbide  Corporation  in
 the past  has defined  as  a decomposition), the first  definition is
 adequate,  simple and  to  the point, while the second  definition is  an
 alternative that contains more words, and that the commenter  believes
 would  also include  the UNIPOL  emergency reactor blowdown.
       Commenter  IV-D-43  stated that  the proposed regulation,  which now
 requires  that intermittent vents  in  an  affected facility be combusted  by
 a flare,  allows an  exemption for  decompositions, but no longer exempts
 emergency  intermittent streams (Section 60.562-l(a)(2) of  the
 January 10, 1989,  Federal Register notice).   The commenter stated
 that the exemption  for emergency  vents  should be retained  for  reasons of
 personal safety,  lack of available technology,  and cost.
      Commenter IV-D-44 stated that the requirement that each
 intermittent vent stream other than decompositions in an affected
 facility be controlled does not take   into account that some intermittent
 streams for safety reasons cannot be  controlled  in a flare system.  This
commenter pointed out  that relief valves are  extensively used  in the
chemical industry for  protecting  pressure vessels from the over-
pressurization  which may  occur for a  number  of  reasons (i.e.,   fire,
exothermic reaction, inadequate cooling).  The  commenter noted that,
                                  2-69

-------
based on the definition of "Intermittent Emissions" (52 FR 36707), the
emissions from relief valves would be included as an intermittent
emission source, thus requiring control.  This commenter stated that
relief valves in the polyethylene industry primarily vented to flare
systems and secondary relief devices to atmosphere are only used when
normal venting to flare is too slow to protect plant personnel from
injury or prevent mechanical damage to the plant.  As an example, the
commenter noted that should for any reason the flare header become
plugged (which has happened) and a plant experiences an emergency
situation, the secondary relief valve would be the only avenue of
process relief to protect personnel and equipment.  In the commenter's
opinion, to require these to be controlled would compromise the safety
and integrity of the process unit.
      Commenter IV-D-44 also noted that some intermittent vent streams
may contain oxygen which is not safe to include in a flare system.  As
an example, the commenter referred to the manufacture of some catalyst
grades for the UNIPOL process in which air is used to activate the
catalyst.  According to the commenter, nitrogen cannot be used in these
cases and thus introducing air (oxygen) into a flare header would create
an unsafe condition.  Thus, the commenter recommended that the
regulation be revised to exempt secondary emergency releases which
protect equipment and personnel from injury; and intermittent sources
which contain oxygen and are therefore unsafe to be included in a flare
system.
      Commenter IV-D-44 also requested that EPA review the definitions
of "Decomposition" and "Decomposition Emissions" to clarify that the
emissions from attempting to prevent a decomposition should also be
excluded.  According to the commenter, there is no system that can
distinguish between an out-of-control condition and a false indication,
therefore the same preventative measures must be taken quickly in either
case to protect equipment and personnel.  The commenter also stated that
routing of other emergency vent streams, which currently utilizes an
automatic control mechanism for atmospheric venting, to a control device
may be significantly different in cost for retrofit of an existing unit
versus cost for a new unit, that the cost for retrofitting the control
and associated hardware would be excessive, and therefore the regulation
                                  2-70

-------
 should allow for a cost effective-justification review for retrofit
 requirements.

       Commenter IV-D-44 concluded  by recommending that:

       1.   The  definition of "Decomposition  Emissions"  include those
           emissions which cannot  be  controlled  in attempting to prevent
           a decomposition,  because actual  and  indicated  decompositions
           cannot be distinguished.

       2.   The  regulation allow  decomposition venting,  and  other
           emergency venting which  is necessary  to prevent  exposure  to
           decomposition hazards,  or  which  is not  economically justified,
           to go directly to atmosphere;  and

       3.   Emergency vent streams on  existing units  be  excluded from
           being required to go  to  a  control system  where the cost can be
           shown to exceed that  which would provide  cost  effective
           control  benefits  (cost greater than $2000 per  Mg/yr).

       Commenter IV-D-45,  which  focused their comments  on the high

 pressure  polyethylene  process,  requested that the Agency change  the

 definitions  of "decomposition"  and "decomposition emissions"  to  read as
 follows:

       "Decompositions"  means for the purposes of  this  standard  an event
       in  a  polymerization reactor  that advances to  the point  where  the
       polymerization reaction becomes uncontrollable and it  becomes
       necessary to  relieve  the  reactor instantaneously in  order  to  avoid
       catastrophic  equipment damage  or serious  adverse personnel safety
       consequences.

       "Decomposition emissions" refers to only  those emissions  released
       from  a polymer production process as the  result of a
       decomposition.   Except for high pressure  polyethylene  processes,
       this term does not  include emissions that may occur during
       attempts  to  prevent a decomposition.

       In  support of this  request,  Commenter IV-D-45 provided  the
 following information:  The high pressure polyethylene process differs
 from the  low pressure processes in that the reaction is a free radical,
 random polymerization at  very high pressures.   The reaction releases
 about  1500 BTU/lb of polymer produced with  the heat sink being the

 unreacted ethylene  in the reactor.   The kinetics  of the reaction is  such

 that the  hotter  it gets,  the faster  it reacts  and (because of the heat

generated) the  faster it  reacts, the hotter it  gets.  Basically, you are
controlling an explosion.   If the  reaction  runs away completely, you
                                  2-71

-------
have a decomposition reaction all  the way to carbon.   This must be
released through emergency valves  to the atmosphere.
      According to the commenter,  during the past decade, computer
control of the reactor has led to  a minimum of decomposition reactions.
The commenter stated that one factor that has made this possible (and
that must be considered in the NSPS) is the situation where the computer
senses the onset of a potential  decomposition reaction and prevents it
by releasing a minimal amount of ethylene to the atmosphere.  The
commenter pointed out that this  release is still considered to be under
emergency conditions since the decomposition reaction moves very fast
(less than a few seconds) and that the reason this release must be to
the atmosphere is that the only  way to know to prevent the decomposition
is an immediate and complete drop  of the pressure to  a very low level.
The commenter indicated that the increase in volume of ethylene from
30,000 pounds per square inch (psi) to near atmospheric pressure is in
the order of several thousand times and therefore, it is not practical
to have a vessel  (a line is out  of the question) to depressurize into.
An additional problem pointed out  by the commenter would be if, in fact,
the depressurization did not stop  the decomposition,  then the catch
vessel would be receiving a runaway reaction, thereby creating a severe
and totally unacceptable safety  problem.
      The commenter concluded by stating that industry will continue to
minimize both decompositions and the type of emergency vents described
above due to the loss of expensive monomers, damage to relief valves,
and the loss of production and by  noting that normal  start-up, shutdown,
and miscellaneous purges from the  reactor section are contained and sent
to a destruction device.
      Commenter IV-D-46 stated the control high pressure emergency
releases as would be required by the proposed procedure outlined in the
January 10, 1989, Federal Register notice would create intolerable
safety risks and the cost of such  control, if achievable, would be
prohibitive.  This commenter stated that the control  facilities required
to capture a high pressure, non-decomposition emergency release or a
high pressure decomposition emergency release would be similar in design
and cost, the control of either would be not cost effective, and
                                  2-72

-------
 therefore all  emergency releases should be excluded from the need to be
 controlled.
       This commenter also presented several  safety arguments related to
 the control  of non-decomposition emergency vents.   According to the
 commenter, most decomposition  (including the emergency release attempt
 to  prevent them)  from a high pressure,  low density polyethylene process,
 occur in  a fraction  of a second  and that when  a  decomposition has
 occurred,  the  non-decomposition  emergency release  has  already been
 attempted, but was unsuccessful  in  preventing  the  decomposition.   The
 commenter pointed out that  decompositions are  secondary emergency
 emissions  (explosions)  that are  directed upward  from the top of the
 polymerization reactor pump or rupture  disc  stacks  into the  atmosphere
 away from personnel  for their protection.
       The  commenter  stated  that  an  emergency release to prevent a
 decomposition  occurs  because emergency  action  is required  to ensure the
 safety of  personnel.   Requiring  burning  of high  pressure emergency
 safety releases  in a  flare  would  defeat  the  purpose of  the safety device
 because to send the  high  pressure emergency  safety  release to a flare
 would  slow the  depressurization  process  down,  because  the  high  pressure
 release needs  to be  fed  to  the flare at  a  significantly (several  orders
 of magnitude)  slower  rate than would otherwise be released to the
 atmosphere.  Depressurization needs to be  instantaneous  to avoid
 compromising safety,  and  the technology  to accomplish this is  not
 available.
       Another  safety  hazard, according to  the commenter, would  be
 created by polymer pluggage of the  flare  system piping.  When  an
 emergency  release occurs, the contents of  the equipment  is dumped,  and
 if the release  is from the  reactor,  gas  and polymer will be  released.
 Flare  piping system pluggage would  cause overpressurization  of  the  flare
 system and also would not allow the emergency safety vent on  the
 equipment  (Polymerization Reaction  and Material Recovery process
 sections) to function properly  because of backpressure.  If  the
emergency vent collection system  feeds a common flare header that serves
other parts of the plant, pluggage  in this header would cause safety
hazards in flare tie-in piping  throughout the plant.
                                  2-73

-------
      To minimize these problem, the commenter stated that an enormous
vent collection system would be required to collect the emergency
release and feed it to a flare, and each emergency vent collection
system would require a sophisticated polymer filtering system that would
not create backpressure in the system.   The commenter stated that they
were not aware of any such system in this service.
      The commenter also stated that the facilities required to capture
non-decomposition emergency releases in a high pressure plant would
likely have to be designed to the criteria standards that facilities to
capture decomposition emissions would have to be designed and built to,
because decomposition emissions are going to exit through the already
opened non-decomposition emergency release system (i.e.,  open dump
valves or blown rupture discs).  According to the commenter, a
containment system to capture a high pressure emergency release would
create an undesirable safety risk because of the enormous size of the
system and the possibility of overpressurization or internal ignition
that could cause a catastrophic explosion.
      In summary, Commenter IV-D-46 stated that the exemption for
emergency vent streams found in the September 30, 1987,  Federal Register
notice should be retained for all intermittent vents in polyethylene and
polypropylene plants, except those for  which EPA can establish that the
cost of control would meet the same definition of economic
reasonableness as is applicable to all  other controlled emission
sources, and meet the necessary standards of safety to run the facility.
Therefore, Section 60.562-l(a)(2) of the January .10, 1989, Federal
Register notice should be changed to reflect this and should therefore
read as follows:
      60.562-l(a)(2)          For each  vent stream that emits
                              intermittent emissions in an affected
                              facility  as defined in Section
                              60.560(a)(2), except for emergency vent
                              streams in polyethylene and polypropylene
                              plants:  control the emissions according
                              to Section 60.562-l(a)(1) (ii}  or (iii).
      Commenter IV-D-46 also stated that the definition of
"Decomposition" should be changed to include the break down  of raw
materials in addition to the break down of polymer since in  some polymer
processes, the raw material break down  is the only significant part of
                                  2-74

-------
 the decomposition.   Therefore,  Section 60.561-Definitions
 "Decomposition" should be changed to read:
       60.561 - Definitions
                 * * *
           "Decomposition"  means  for the  purposes  of this  standard an
           event in a polymerization reactor that  advances to  the point
           where the polymerization  reaction becomes uncontrollable,  the
           polymer or raw materials  begin to break down  (decomposes),  and
           it becomes necessary to relieve the  reactor instantaneously in
           order to avoid catastrophic  equipment damage  or serious
           adverse personnel  safety  consequences.
       Commenter IV-D-47  stated that the  distinction between
 decomposition  releases and planned  releases to prevent  decomposition  is
 a  matter  of  semantics more than  a technical  difference  as  the
 technological  difficulties for controlling  each release are similar.
 This  commenter stated that trying to throttle  a highly  pressurized
 gas/polymer  mixture to "control" a  release  is  essentially  a go/no go
 decision  to  dump  the reactor, the purpose of which  is to  moderate
 pressure  spikes  affecting  reactor life and  distortion of  internal parts.
 According  to the  commenter,  industry, generally,  does not  presently
 attempt to control  releases  during  attempts  to prevent  decomposition.
 The commenter  pointed out  that the  time  interval  between detection of a
 need  to release  and  the  time that decomposition actually occurs  is on
 the order  of one  second  and  stated  that  until technology to control such
 releases  is  developed safety concerns should mandate venting such
 releases to  the air.
      Commenter  IV-D-47  also stated that  the Agency's proposed rules
 should provide an  exemption for upset operations  emissions in the
 polypropylene process.   Commenter IV-D-47 noted that current
 polypropylene technology is such  that runaway reactions do not occur.
 The commenter described  the technology as self-limiting in that as
 temperature  increases, the reaction  rate  increases up to the point that
 the higher temperature causes the fluidized polymer to melt.   The heat
of crystallization/melting  absorbs  sufficient energy and provides for
coating the catalyst particles  so as to quench  the reaction.   The
commenter noted that there  is another concern.   In the event  of a power
failure affecting the recirculation  compressors,  the commenter stated
                                  2-75

-------
that the reactor beds would no longer be fluidized and unless the
reactor is vented the reaction will continue to the point of making one
large polymer "chunk" in the reactor.  According to the commenter, the
clean up procedure following such an occurrence would require personnel
to enter the reactor vessel where potential  pockets of decomposition gas
remain.  The polymer "chunk" would have to be cut into small blocks
using air-driven saws.  The downtime required for this situation may
take two weeks and be very labor intensive and costly.  The commenter
pointed out that while such an occurrence is very infrequent, upset
operations emissions do occur and requested  that they be addressed in
the proposed regulations.
      Lastly, Commenter IV-D-47 suggested that Table 3 may be an
appropriate spot to incorporate alternative  control options for
intermittent emissions.  Eastman Chemicals Division utilizes a catalytic
oxidation unit for emergency blowdown emissions as required by the State
Air Pollution Control Board.  The commenter  indicated that they would  
like to use the existing catalytic oxidation unit for the intermittent
emissions requiring control (Table 3) rather than install a new flare
system (the estimated cost of the new flare  system exceeds one-million
dollars).
Response:
      Intermittent releases from polymer manufacturing plants can be
among the largest sources  of VOC emissions at a plant.  Some of these
releases have high volumes released in very  short periods of times.
Because of the high volumes, the cost of controlling such releases is
very expensive.  Releases  of this nature include emissions that occur
during a decomposition event, during attempts to prevent decompositions,
and during reactor dumps to minimize the adverse consequences of a
runaway reaction (other than a decomposition).  Based on the information
provided by the commenters and previously by the industry as whole,
these releases are related to operating conditions that are abnormal and
abnormal to the point that the design of the process cannot return
conditions to normal operations.  In at least one instance, in the case
of emissions during attempts to prevent decompositions, these releases
are triggered to prevent a decomposition from occurring although there
is no guarantee that a decomposition would actually have occurred if the
                                  2-76

-------
release was not made.  However, the explosive nature of a decomposition
and the rapidity with which it occurs makes it virtually impossible to
distinguish between situations.
      Some polymer producers use pressure relief valves or other
mechanisms to vent emissions from process vessels as part of the process
design for operating the vessels under normal operating conditions and
for maintaining normal process conditions.  Examination of the
information available indicates that these emissions have been
controlled and can be controlled cost effectively.  These intermittent
releases are designed to keep the process vessel in normal operating
conditions; they are not releases that occur because of abnormal
operating conditions or because releases are necessary to prevent
equipment damage or personnel  safety hazardous because the operating
conditions can no longer be returned to normal operating conditions.
      The Agency would like to relate the control/no control decision
for all  intermittent releases in the same manner as has been done
for continuous emissions.  However,  for the reasons stated in the
January 10, 1989, Federal Register notice, this approach is not
feasible.   Therefore,  the promulgated standard continues to define the
categories of intermittent releases  that are exempt from the standards.
In the promulgated standard,  the definition of "emergency vent stream"
has been revised and is used,  rather than decomposition, for exempting
individual  intermittent releases from control  requirements.   In
addition,  the definition of "decomposition emissions"  has been revised
to include those emissions that occur as a result of attempts to prevent
decompositions.
                                  2-77

-------

-------
               3.0  CONTROL TECHNOLOGY - PROCESS EMISSIONS

 3.1   GENERAL

 Comment:
       Four commenters (IV-D-7,  IV-D-8,  IV-D-10,  IV-D-13/47)  stated their
 support for the  philosophy of allowing  freedom of  choice of
 technological  methods to the  manufacturer  as  long  as emission
 requirements are met, and for a  performance oriented approach to all
 aspects of the standards,  from control  devices utilized  to monitoring
 techniques.   Commenter IV-D-8 stated  that  specifying the specific
 control  techniques  that  can be used  is  unnecessarily restrictive,
 especially as  the Agency recognizes  that various devices are  capable of
 achieving  essentially equivalent  VOC  destruction efficiency.
       Two  commenters  (IV-D-7  and  IV-D-13)  expressed  concern over "EPA's
 growing attempts to mandate technology  to  industry."   One commenter  (IV-
 D-13/47) stated  that  the Agency  should  consider technical feasibility
 and  reasonableness  of cost during  its efforts  to set  realistic  standards
 and  that the equipment specifications and  the  process  technology
 employed to  meet reasonable standards should be at the discretion  of the
 manufacturer.  This commenter expressed concern over  the Agency's
 tendency to  mandate specific  technology to the polypropylene  and the
 polyethylene industries,  stating that specifying equipment (i.e.,
 flares)  instead  of  the emissions standards discourages the development
 of new  processing technology  and control equipment and that the  Agency's
 determination  of where available capital must  be spent tends  to  stymie
 long-term  growth  and  stability of the polymers industry.  The commenter
 then recommended  leaving the  freedom of choice of control or  process
 technology to  the industry.
      The  commenters  believe  that EPA should set the emission
 requirement  and  allow industry to choose the appropriate technology to
 achieve that requirement.  Examples of restrictive wording cited by the
 commenters are presented below.
      Three  commenters (IV-D-7, IV-D-8,  IV-D-13) referred to Section
 60.562-l(a)(2)(i) of the September 30, 1987,  Federal  Register notice,
which requires vent streams from affected facilities  that emit
 intermittent emissions to combust the emissions in a  flare.   The
                                  3-1

-------
commenters stated that other control options that achieve a 98
percent/20 ppm reduction of VOC should also be allowed.  Commenter
IV-D-8 indicated that some existing facilities may currently use other
control devices (such as catalytic incineration or boiler fire box
combustion) to control intermittent emissions from nonemergency vent
streams.
      Additionally, one commenter (IV-D-8) noted the option of control
alternatives is highly dependent on the location of the process unit to
other units.  For example, if an ethylene plant is on the same site as a
LDPE unit, then the LDPE intermittent emissions could be routed back to
the ethylene plant for recovery, a solution far preferable to flare
combustion.
      Commenter IV-D-8 noted that examples of this type of restrictive
working are Sections 60.562-l(a)(2); 60.562-l(a)(l)(ii)(C); 60.562-
l(c)(l)(iii), 60.562-l(c)(3)(iv); and 60.562-l(c)(4)(iv).
      Specific revisions (indicated by the words underlined) suggested
by one commenter (IV-D-13) were as follows:

      1.    Section 60.562-l(a)(l)(ii)(C):  "Flares used to comply with
            provisions of this subpart shall be steam-assisted, air-
            assisted, non-assisted, or pressure (kinetic)  flares."

      2.    Section 60.562-l(a)(l)(ii)(D):  ". . . .; or with the net
            heating value of the gas being combusted being 7.45 MJ/scm
            (200 Btu/scf) or greater if the flare is non-assisted; or
            with the net heating value of the gas combusted being 5.6
            MJ/scf (.15.0 Btu/scf or greater i_f the f 1 are i_s a piloted,
            non- assisted flare."

      3.    Section 60.562-l(a)(2), add (ii):  ". . . or use another
            demonstrated satisfactory method of achieving emissions
            reductions of 8 percent p_r 20 ppmv VOC. ifl the final vent
            stream."

      4.    Sections 60.563(e)(l)(ii) and 60.563(e)(2)(ii) should be
            revised to state:  "An organic monitoring device used to
            indicate the concentration level of organic compounds based
                                  3-2

-------
             on  a  detection  principle  such  as  infrared,  photoionization,
             thermal  conductivity,  or  other demonstrated equally
             effective  device,  or monitoring scheme."

       5.     Section  60.563(e)(3)(i) should be  revised  to state:  "An
             organic  monitoring device  used to  indicate  the  concentration
             level of organic compounds  based on  a detection  principle
             such  as  infrared,  photoionization, thermal  conductivity, or
             other equally effective device or  monitoring scheme  for  the
             outlet of  the carbon bed."
Response:
       The Agency  agrees with the commenters that any control device  that
meets  the emission requirements of a standard  should be  allowed  to be
used.  This  is  in fact a part  of the Clean  Air Act as amended  [Section
lll(h)(3)],  and it was never the intent of this  rule to  countermand  the
Clean  Air Act.  The  Agency  also agrees  that any monitoring technique
that provides equivalent indication of  control device performance and
provides the same reporting and recordkeeping  information required by
the rule can be used subject to the approval of the Administrator.   The
particular control devices  and monitoring  techniques specified in the
proposed rule were selected based  upon the Agency's experience and
analysis as to which control devices are the most effective, and then as
to which monitoring  techniques were the most appropriate for those
control devices.
      One commenter  is incorrect in his implication that the Agency did
not consider technical  feasibility and reasonableness of cost during the
standard setting process.   The Agency throughout the development of this
standard, as well  as for all standards, constantly concerns itself with
the technical feasibility of control  devices as they would be applied to.
emission streams associated with,   in this instance,  the polymer
manufacturing industry.  In addition,  the Agency checks and double-
checks its cost information with equipment vendors and tries to identify
industry-specific  characteristics  that may affect the cost of the
control devices.  In most  instances,  the primary source of emission data
available to the Agency to  assess  the  technical feasibility of control
is the industry itself.  In the course of this information exchange,
sometimes pertinent  data that affect  control device  feasibility and cost
                                  3-3

-------
are not transmitted by the industry or are sometimes inadequately
considered by the Agency.  For example,  initial  industry data did not
indicate the presence of water in certain polystyrene and poly(ethylene
terephthalate) emission streams that would then  affect the use of
subzero refrigerated condensers.   Since  proposal,  industry has alerted
the Agency to this potential  problem, the Agency has contacted
additional information sources, and, as  a result,  has revised its
analysis to take this factor into account.  Thus,  in summary, the Agency
always considers technical feasibility and costs in its standard setting
procedures.
      The Agency also disagrees,  in part, with the commenters'
suggestion that the "equipment specifications and  the process technology
employed to meet reasonable standards should be  at the discretion of the
manufacturer."  The Agency has developed much expertise over the years
on the performance capabilities of various control devices.  This
expertise has been gained in part from working closely with many
different equipment manufacturers.  This association allows the Agency
to consider and rely on the collective wisdom of many individual
experts.  To the extent that certain types of control devices (in this
instance, flares) require equipment specifications to ensure that a
standard is met, the Agency has the experience and expertise to specify
what those equipment specifications should be and  will continue to
specify such in the future as necessary.  As noted earlier, the Agency
does not wish to restrict the manner in  which the  proposed standards are
met, whether the industry or manufacturer selects  the control technique
to use, so long as the standards are met.
      Five specific examples of restrictive wording are noted by the
commenters.  Each of the five examples is dealt with as specific
comments latter in this document.  In general, the first two suggestions
cannot be accommodated at this time because of insufficient data to show
that the standard of 98 percent reduction would be met.  The last three
comments deal with specifying in the rule the allowance of equivalent
control technologies or monitoring devices and,  in general, the Agency
has incorporated these suggestions into the final  rule.
Comment:
      One commenter  (IV-D-6) wrote that the preamble states that a flare
or other combustion device is appropriate, but that the regulation
                                  3-4

-------
 [Sections 60.562-l(a)(l)(i) and (ii) of the September 30, 1987, Federal
 Register notice] appears not to allow a flare or other combustion device
 as a control mechanism by requiring 98 percent reduction or a
 concentration level of 20 ppm or less of VOC be maintained.   This
 commenter stated that the discussions in the preamble and preamble Table
 1  indicate control  options for continuous emission sources from
 polyethylene processes to include (1) 98 percent control efficient
 device, (2)  reduce  to 20 ppmv,  or (3) flares.   The commenter then
 pointed out  that, in the actual  regulation [Section 60.562-l(a)(l)],  the
 inclusion of a statement about  boilers or process heaters in paragraph
 (i)  before the final "or" confuses the intent  to include the third
 option of using flares.   The commenter suggested that it would be less
 confusing if the boiler/process  heater statement was  separated as
 paragraph (iii)  with reference  to  paragraph  (i).
 Response:
       The intent of Section  60.562-l(a)(l)  is  to require continuous
 process emissions to be  reduced  by 98 percent  reduction  or to  a
 concentration of 20 ppmv,  whichever  is  less  stringent.   This  requirement
 was  stated in Section  60.562-l(a)(l)(i)  of the  standards  proposed  on
 September 30,  1987.  Any  control device,  including  flares, that can be
 demonstrated  to  meet this  standard may  be  used  to  comply  with  the
 standard.  Performance tests  for flares, however,  are very expensive  and
 difficult.  Through  a  history of flare  efficiency  tests  conducted  by
 EPA, the John  Zink  Company,  and  industry,  flares have been demonstrated
 to achieve 98  percent  reduction when  operated under certain conditions'.
 These conditions  were  spelled out  in  Section 60.562-l(a)(1)(ii)  of the
 standards as  proposed  and  are in Section 60.18 of the General
 Provisions.  The  intent of separating flares from Section 60.562-
 l(a)(l)(i) was to delineate the operating conditions under which they
 could be used  to  comply with the standards and not to infer that they
 could not be used in complying with Section 60.562-l(a)(l)(i) as
 originally proposed.  This portion of the regulation has been reworded
 to avoid misinterpretation (see Section 60.562-l(a)(l) of the final
 rule).
Comment:
      One commenter  (IV-D-43) stated that the January 10, 1989, Federal
Register notice contains  inconsistencies regarding flare, incinerator,
                                  3-5

-------
and boiler control and that the intended extent of control is unclear.
The commenter referred to a statement in the preamble to the Federal
Register notice (54 FR 896) regarding the control  of the low flow
streams (less than 8 scfm), in which EPA states "... such control can
take place in a flare, incinerator, boiler,  or other control device
located at the plant site and the destruction efficiency of the device
would not be specified."  The commenter recommended that the language in
the proposed rule be modified to incorporate the concepts in the
preamble.
Response:
      The preamble portion of the January 10, 1989, Federal Register
notice required control of all streams with  flows  of less than or equal
to 8 scfm and annual emissions equal to or greater than 1.6 Mg/yr.  The
level of control differed according to whether or  riot the stream was
located in an affected facility whose total  emissions are required to be
reduced by 98 percent.  As stated in the preamble, if the former were
the situation, then the low flow stream would also be required to be
reduced by 98 percent.  On the other hand, if the  latter were the
situation, then the low flow stream could be vented to a control device
without regard to that control device's VOC  control efficiency.
      This distinction was carried over into the standard portion of the
January 10, 1989, Federal Register notice in Table 3.  In that table,
under the "Control/No Control Criteria," all streams are required to be
controlled by 98 percent (or to 20 ppmv), which was indicated by
reference to Section 60.562-l(a)(l)(ii) under the  column titled
"Applicable Standard," or in a flare that meets specified operating
requirement, which was indicated by reference to Section 60.562-
l(a)(l)(iii) under the column titled "Applicable Standards," if
emissions are equal to or greater than the specified calculated
threshold emission (CTE) level.  If emissions are  less than the CTE
level, then no control is required except for streams with flows of 8
scfm or less.  The control requirements for these  streams were also
indicated under the "Applicable Standard" column.   Reference to Section
60.562-l(a)(l)(i) allowed these streams to be vented to a control device
located on the plant site.  Section 60.562-l(a)(l)(i) does not mention
how much VOC reduction must be obtained; thus, the standard presented  in
the January 10, 1989,  Federal Register notice was consistent with the
                                  3-6

-------
 preamble and the Agency has  kept this provision in the final  rule (see
 Section 60.562-l(a)(l)(i)(D)  of the final  rule).

 3.2   FLARES

 Comment:
       One commenter (IV-D-8)  suggested  that  the proposed  standards
 (Section  60.562-1)  allow the  use of pressure (kinetic)  flares and a
 section on pressure (kinetic)  flares be included.   The  commenter
 proposed  the following  language for Section  60.562-l(a)(l)(ii)(C) of the
 September 30,  1987,  Federal Register notice:   Flares  used  to  comply  with
 provisions of this  subpart shall  be steam-assisted, air-assisted,
 nonassisted,  or  pressure (kinetic)  flares.
 Response:
       In  order to accommodate  the commenter's  suggestion,  EPA must be
 able  to establish minimum heat  content  and maximum exit velocity
 parameters  for pressure  (kinetic) flares that  would ensure  98 percent
 destruction  of VOC's.   Unfortunately, EPA  is not able to define  these
 conditions  for pressure  assisted  flares.  The  EPA  is  aware  of one set  of
 tests  of  two  pressure assisted  flareheads  (see  Docket  Item  IV-A-2).
 Neither flare  was piloted.  One  of  the  flares  (flare  F) was able to
 maintain  flame stability  at about 200 feet per  second  (ft/sec) exit
 velocity  (maximum velocity tested)  with only 150 Btu/scf (propane -  N2
 mix).   At  60  ft/sec, the  head required  about 130 Btu/scf.
       The  second flare tested  (flare  E) required much higher  gas  heating
 value  to maintain flame  stability.  At  60 ft/sec,  this head required
 about  400  Btu/scf (propane - nitrogen mix).  This  head was tested  at
 velocities up to about 900 ft/sec.
       For both of these flares, combustion efficiency was correlated to
 flame  stability using the ratio of gas  heating value to that minimum
 heating value required to maintain flare stability.  Combustion
 efficiencies greater than 98 percent were achieved when this ratio was
greater than 1.3.
      The difference in performance of these  two flares is not
understood.  Both were equipped with flame retention devices,  yet one
required a heat content almost three times as high as  the other.   Other
factors (in addition to gas  heating value and exit velocity) must be
                                  3-7

-------
important in determining flame stability.  Additional testing would be
required to define these factors.  Therefore, due to the inconclusive
nature of the tests that have been done on pressure assisted flares to
date, EPA has not incorporated the commenter's suggestion.
      Pressure (kinetic) flares can still be used if they can
demonstrate equivalency.  The flares tested achieved at least 98 percent
reduction under certain operating conditions (see Docket Item IV-A-2).
In determining equivalency, the Agency will consider the design of the
pressure (kinetic) flare and its operating conditions.  If the flare can
be demonstrated to the Agency's satisfaction to be similar in design to
one of the two flares tested and that when operating will achieve those
operating conditions that were found to achieve at least 98 percent
reduction for the similar flare, then the Agency would consider such a
flare as demonstrating equivalency.  If these things cannot be
demonstrated to the Agency's satisfaction, an owner or operator can
still demonstrate equivalency by conducting a performance test.
Comment:
      One commenter (IV-D-44) pointed out that the Agency is currently
evaluating comments suggesting that control devices other than flares be
allowed (54 FR 890) and that Section 60.562-l(a)(l)(ii)(C) of the
September 30, 1987, Federal Register notice states; that only air-
assisted, steam-assisted, or non-assisted flares shall be used to comply
with this standard.  The commenter then stated that ground flares, which
are multi-stage, pressure-velocity assisted flare systems, are also
currently being developed and used by the polyethylene manufacturing
industry, but that this type of flare does not appear to meet any of the
above flare type definitions or their associated requirements [proposed
Sections 60.562-l(a)(ii)(0, E, F)].  The commenter stated further that
the Agency has not proposed equivalency or use of this type of flare
technology, although in the new (January 10, 1989) approach, the Agency
is specifying control of "low flow" streams in some cases in a flare,
incinerator, boiler or other control device located at the plant site
and the destruction efficiency of the control device would not be
"specified (54 FR 896 and 908).  The commenter recommended that these
inconsistencies be corrected, formal guidance on acceptable control
devices be issued, and procedures for approval of new innovative and
alternate technologies be provided by the Agency.
                                  3-8

-------
 Response:
       In the January 10, 1989, Federal Register notice, the Agency
 stated that we were evaluating control devices other than flares with
 regard to the control of intermittent emissions as a result of comments
 received on the September 30, 1987, Federal Register in which the
 proposed standards required the use of flares to control intermittent
 emissions.  For continuous emissions,  the September 30, 1987, Federal
 Register notice indicated that 98 percent destruction by weight or 20
 ppm by volume was the standard.  If an owner or operator elected to
 comply with this standard for continuous emissions by using a flare,
 then the flare would be required to be an air-assisted, steam-assisted,
 or non-assisted flare meeting certain  specified operating conditions.
       If an owner or operator wishes to use a multi-stage,  pressure-  
 velocity assisted flare to control  continuous emissions that require  98
 percent destruction,  then the owner or operator would be required to
 demonstrate equivalency or perform  a compliance test,  as noted in the
 response to the previous comment.   The Agency has  inconclusive and
 insufficient data to  prescribe those operating conditions  that would
 ensure 98 percent destruction using such  flares.   Since control  of
 intermittent streams  and certain  low flow streams  can  be controlled
 without  regard  to the destruction efficiency  of the control  device, then
 such flares  can  be used  without  the need  to demonstrate equivalency to
 98  percent  destruction.
       In  conclusion, the Agency believes  the  standards  are  internally
 consistent with  the application of  flare  technology in  accordance to
 whether  a specific VOC destruction  percentage  is or is  not required.
 Furthermore,  it  is impossible  (and  inappropriate)  to  identify  in  this
 rule  all  procedures an owner  or operator  must  follow to  demonstrate
 equivalency  because each  technology will  have  its  own unique aspects  and
 the General  Provisions allow  a case-by-case determination to be made.
 Comment:
      One commenter (IV-D-1) stated that  the requirements proposed on
 September 30, 1987, for combustion of continuous gas streams in a flare
 [Section 60.562-l(a)(l)(11)(A)] and for combustion of emergency releases
 in a flare [Section 60.562-l(a)(2)(i)(A)] are identical to existing
rules for "visible emissions" and are straightforward  enough.  However,
this commenter felt that requirements (B) through (F)  (for continuous
                                  3-9

-------
 emissions)  and  (B)  and  (C)  (for  intermittent  emissions)  do  not  seem to
 belong  in the proposed  regulations  as  these requirements deal with  the
 technical operation of  flares  or  other matters  that  are  usually resolved
 in  cooperation  with the flare  manufacturer.
 Response:
      Because of the technical difficulty of  testing for the VOC
 reduction efficiency of flares, EPA determined  a set operational
 specifications  that would ensure  98 weight percent reduction efficiency
 for continuous  emissions.  These  specifications are  identified  in 40  CFR
 60.18 of the General Provisions (see Federal  Register, January  21,  1986,
 pp. 2701-2702).
      The EPA has not identified  a complete similar  set  of  operational
 specifications  for intermittent emissions controlled in  a flare that
 would ensure 98 percent  destruction.   However,  the Agency believes  that
 the destruction efficiency of  a flare  controlling intermittent  emissions
 is  greater when a flame  is present at  all times and when the flare  is
 designed to maintain a  stable  flame than when either of  these two
 conditions are  absent.   Flares controlling continuous or intermittent
 emissions could be designed or operated in a manner other than
 prescribed by the regulations  in the absence of these regulations.  Such
 flares would potentially achieve less  emission reduction.  Thus, the
Agency disagrees with the commenter that such provisions do not belong
 in  the regulations, and  these provisions are retained in the promulgated
 rule.  (Note:  The provisions for controlling continuous emissions  in
 flares is now incorporated in the promulgated rule by referring to  40
CFR 60.18 of the General Provisions.)
Comment:
      One commenter (IV-D-8) suggested that the standard should allow
5.6 megajoules per standard cubic meter (MJ/scm) (150 Btu/scf)  instead
of 7.45 MJ/scfm (200 Btu/scf) as the minimum net heating value  for
nonassisted flares with pilots.  The commenter supported this suggestion
by referring to flare tests performed by EER for EPA that have  shown
nonassisted flares with pilots reduce TOC by at least 98 percent if the
net heating value of the flared gas is not less than 5.6 MJ/scm.  The
commenter provided estimates that such a reduction could reduce energy
demand for five flares combusting five vent gas streams with a  total
flow of 514.5 x 103 cubic feet  per hour by  17.5  x  106 Btus per hour.
                                  3-10

-------
 The commenter proposed the following language for Section 60.562-
 l(a)(l)(ii)(D) of the September 30, 1987, Federal Register notice:
 "...; or with the net heating value of the gas being combusted being
 7.45 MJ/scm (200 Btu/scf) or greater if the flare is nonassisted; or
 with the net heating value of the gas combusted being 5.6 MJ/scm (150
 Btu/scf) or greater if the flare is a piloted, nonassisted flare."
 Response:
       The EPA believes that 200 Btu/scf is the correct cut off for all
 nonassisted pipe flares.   The commenter refers to a series of tests
 which determined the flame stability of unassisted and piloted 3 inch
 pipe flares (see Docket Item IV-A-3).   A plot of "Flare Exit  Velocity
 (ft/sec) vs Gas  Heating Value (Btu/scf)," Figure 4.2 page 4.4 of Docket
 Item IV-A-3,  shows  the "stability limit" for  these conditions passes
 through  the 60 ft/sec line at 150 Btu/scf.  The stability limit plotted
 here was determined by the appearance  of the  flame,  not by combustion
 efficiency  test.  This plotted  line corresponds  to the flare  operating
 conditions  at  which point the "faint orange flicker"  was  observed to
 disappear as gas  heating  value  was  decreased  or  velocity  increased.   The
 line drawn  in  this  figure represents the average  of  the points  observed.
 Some points  are  above  the line  and  some  below.  The  conclusion  of all
 the  flare studies has  been  that  flares firing  gases with  heating  values
 near the stability  limits are susceptible to  perturbations  and  poor
 combustion efficiency  (see  Docket  Item IV-A-1).   Combustion efficiency
 is almost surely greater  than 98  percent when  the  ratio of  the  actual
 heating  value  to the heating  value  required for stability  is  1.2  or
 above  (see Docket Item  IV-A-3).   If  the  stability  limit (at 60  ft/sec)
 was  determined to be 150  Btu/scf, the heating  value of  the gas  combusted
 must be  180 Btu/scf (150  Btu/scf x  1.2)  for 98 percent efficiency.  This
 is in close agreement with the value of 200 Btu/scf determined  in
 previous  studies and required for nonassisted pipe flares and provides a
 small margin of protection that perturbations will not cause poor
combustion efficiency.  Therefore, the Agency has retained 200 Btu/scf
as the minimum net heating value for all  nonassisted pipe flares.
                                 3-11

-------
 3.3   INTERMITTENT  EMISSIONS

 Comment:
      One  commenter  (IV-D-8)  suggested that  the  proposed  standard  for
 controlling  intermittent emissions  [Section  60.562-l(a)(2)(i)  of the
 September  30,  1987,  Federal Register notice] have  a  second  part added
 that  allows  the  use  of other  control technology  for  intermittent flows
 provided such  control technology can achieve 98  percent reduction  in
 emissions  or 20  ppmv.  The commenter suggested the following language be
 added:
             60.562-l(a)(2)(ii):  Or use some other or provable method
             achieving emissions reduction of 98  percent or  20 ppmv VOC
             in the final oases.
 Response:
      In addition to flares,  several other combustion devices have the
 potential  to effectively control intermittent emissions.  Both
 Incinerators and boilers can  reduce emissions by 98 percent reduction or
 to 20 ppmv if properly designed, installed, and operated.    There are
 capacity and safety considerations that limit the applicability of these
 devices for  controlling intermittent emissions.  It is generally
 accepted that flares can control effectively a much wider range of flow
 rates than incinerators or boilers.  For these reasons,  EPA considered
 only flares  as the control  technology for reducing intermittent
 emissions.  However, the Agency sees no reasons not to allow other,
 equally effective control  devices,  such as incinerators, boilers,  or
 process heaters from being used to control intermittent  emissions.
Therefore, the promulgated rule now specifies that incinerators,
 boilers, or process heaters may be used to control  intermittent
emissions  (see Section 60.562-l(a)(2)(ii)  of the final  rule).   Other
control  technologies not specifically identified in this rule may  be
used if they are demonstrated to be equivalent  in emission reduction.
      Although the Agency  is not requiring that intermittent emissions
be shown to be reduced 98  percent  (or to  20 ppmv) when controlled  in a
flare, incinerator, boiler, or process  heater,  such devices are to  be
always operated in such a  manner as to  achieve  optimal  destruction.  For
example, where a flare or  incinerator is  used to control continuous
emissions that are subject  to these standards as well as intermittent
                                 3-12

-------
 emissions, then the owner or operator is required to operate  the  flare
 according to the operating parameters specified for the control of
 continuous emissions or, where an incinerator is used, demonstrate that
 continuous emissions are reduced by 98 percent (or to 20 ppmv, whichever
 is less stringent) both with and without the intermittent emissions.
 Comment;
       One commenter (IV-D-39) stated that the control  requirements for
 intermittent emissions are clearly stated in the preamble on page 898
 and in Table 3 on page 894 of the January 10, 1989,  Federal Register
 notice, but that there is no such clear statement in the actual proposed
 rule section.
 Response:
       While the Agency believes the proposed rule section in the
 January 10,  1989.  Federal  Register notice clearly presented the
 requirements for intermittent emissions,  the language  has been revised
 in the final  rule  and  provides  clearer statements  of the required
 controls.   As discussed above,  revisions  have been made  to  the control
 requirements  for intermittent emissions  specifying the use  of certain
 control  devices other  than  flares,  and these revisions are  also
 incorporated  in the  final  rule.
 Comment:
       Several  commenters  (IV-D-39,  IV-D-42,  IV-D-48) requested that  the
 1.6  Mg/yr  exemption  be  extended to  intermittent vents.   Commenters
 IV-D-42  and  IV-D-48  indicated that control of small  intermittent vents
 is more  costly  and difficult  than the  control of small continuous  vents.
 Commenter  IV-D-42 identified  the high  temperature and the use  of air
 (fluidizing)  as  aggravating the problem of control for two small
 intermittent  vents in the UNIPOL process.
      Commenter  IV-D-39 referred to the preamble statement that control
 is required for  all "start-up, shutdown,  maintenance purges, and other
 normal process releases."  This commenter then stated that venting these
 streams to control  devices is extremely difficult.  According to the
 commenter,  to lower the VOC concentration to a level  where maintenance
can be performed on a vessel,  for example, a purge stream must be  routed
to the lowest available backpressure,  which is the atmosphere, and there
is a point  where vessels must  be purged to the atmosphere prior to
                                 3-13

-------
 venting.   For this reason,  the commenter concluded,  there should be an
 exemption for low level  intermittent  streams.
 Response:
       The Agency recognizes that  some individual  intermittent streams
 may be more costly to  control  than  individual  continuous  streams,  and
 that the  same logic for  providing a small  stream  exemption  should  be the
 same whether the stream  is  continuous or intermittent.  The Agency
 believes, however,  that  there  is  an overriding consideration that  makes
 such an exemption for  intermittent  streams  impracticable.   The  purpose
 of the small  stream exemption  is  to identify individual streams whose
 annual  emissions are sufficiently small  that it is not cost effective to
 send the  emissions  to  an existing control device.  In order to  properly
 apply this  exemption,  one must  be able to determine  accurately  total
 annual  emissions.   Because  they are continuous, this can  be done for
 continuous  emissions.  For  intermittent  emissions, this same certainty
 is lacking  because  the annual emissions  of  an  intermittent  vent is
 determined  in  part  on  the number  of vents that occur each year, which
 can vary.   Thus,  it  is possible that  a small intermittent vent  would
 have annual emissions  lower than  an exemption  level one year and higher
 than the  exemption  the following year.  This could result in  an
 inconsistent decision  from one year to the next.  Furthermore,  the
 testing of  intermittent emissions to determine accurately total annual
 emissions is much more complicated  and costly than for continuous
 emissions.  For  these reasons, the Agency does not believe  that an
 individual  stream exemption for intermittent streams is practicable, and
 the  final  rules  does not contain a low annual  emission exemption for
 intermittent streams.
 Comment:
      One commenter  (IV-D-8) requested that the allowable  period of
 visible emissions for flares controlling intermittent emissions be
 increased and the period extended because intermittent gas  streams are
 by definition much less frequent and are more  difficult or  impossible to
 control (reduce) safely.   The commenter pointed out that  intermittent
gas  streams differ from continuous gas streams  because they are highly
variable,  including emergency vents  and usually have  much  higher flow
rates.  The commenter noted that flares controlling intermittent
emissions  can be designed to operate smokelessly  for  most  of the gas
                                  3-14

-------
 flow range they are expected to handle; however, there are  some  flows
 that cannot be handled smokelessly  (e.g., vents resulting from equipment
 breakdown, a decomposition, or power failure).  In addition, the
 commenter stated that tests done by CMA, EPA, and John Zink have shown
 that the carbon (smoke) emissions from the smoking flare amount to  1
 percent or less of the TOCs.  Therefore, according to the commenter, a
 smoking flare, although more visible, will  still achieve a high
 reduction of emissions.  The commenter suggested visible emissions  could
 be limited,  for example,  to less than 10 to 15 minutes in any 3 to  4
 hour period for intermittent process emissions.
 Response:
       As proposed,  the standards for flares controlling intermittent
 emissions required, in part, that such flares be designed and operated
 with no visible emissions except for periods  not to exceed a total  of 5
 minutes during any  2 consecutive hours.   The  commenters are  requesting
 that this period  following visible  emissions  be  extended  to  10  to 15
 minutes in any 3  to 4  hour period  because some flows  cannot  be  handled
 smokelessly.   Example  flows cited  by the  commenters were  those  that
 result  from  equipment  breakdown,  decomposition,  or  power  failure.
      The Agency  notes that each  example  cited by the commenter  is  a
 type of malfunction.   The  General Provisions  allow  emissions  in excess
 of the  level of applicable emission  limits that  are caused by
 malfunctions without such  emissions  being considered  a  violation  of  the
 applicable emission  limit  unless otherwise specified  in the  applicable
 standard [Section 60.8(c)].  It  is not the intent of  these standards to
 require control of  emissions that occur due to power  failures,
 decompositions, or  other process upset that require emission releases  in
 order to  avoid  catastrophic equipment damage or personnel safety
 hazards.  For all other types of intermittent emissions subject to these
 standards, flares can be and have been designed so as to operate within
 the standards as proposed.  The Agency finds no reason to change the
 standards.
Comment:
      One commenter (IV-0-46) stated that, while  it  wouTd be  beneficial
to allow control of intermittent emissions by  devices  other than flares,
particularly for energy recovery,  they know  of no technology  to  control
                                 3-15

-------
high pressure emergency releases as would be required by this proposal,
whether the control device be a flare or other control device.
      The commenter stated that emergency releases in high pressure
plants typically occur at 35,000 to 45,000 psi on tubular reactors and
15,000 to 30,000 psi on autoclave reactors with initial gas flow rate
(peak flow) of a non-decomposition emergency release exceeding 300,000
scfm at a peak velocity of 1 mile/sec, and at temperature exceeding
600F.  The commenter then stated that the proposed revisions to the
regulations do not take into consideration that there is no technology
available to send this magnitude of gas flow to a flare.  According to
the commenter, this flow would immediately destroy the flare system and
thus an intermediate step to add a gas collection system would be
required to collect high pressure releases and then send them to a flare
and at the same time, depressure the process equipment instantaneously
to avoid compromising safety.  The commenter also noted that on high
pressure process equipment where decompositions can occur in
polyethylene manufacturing facilities, the decomposition emissions are
going to exit through the already opened non-decomposition emergency
release system (i.e., open dump valves or blown rupture discs) and thus
the control system to handle non-decomposition emergency releases would
be exposed to decomposition emissions and must have the capability to
handle decomposition emissions.
      The commenter also stated that since technology does not exist for
control of high pressure releases from dump valves,  the proposed
regulations may result in the generation of decompositions that would
otherwise be avoidable and this would increase the net emissions to the
atmosphere because, unlike emergency releases to prevent decompositions
in which a dump valve typically may stay open for 5 to 20 seconds (small
releases), decompositions result in equipment damage and venting of the
polymerization reaction and material  recovery systems until  they are
empty.
Response:
      One company has in the past controlled high pressure emergency
releases from a LDPE,  high pressure process.  'Thus,  there is technology
that not only exists,  but has been  used,  to control  such releases.
However, the Agency does not believe that the cost of controlling such
releases is reasonable given the infrequency of such releases,  the high
                                 3-16

-------
 flows, and the resulting lower level of emissions.  As noted  in the
 response to the previous comment, the final rule does not require the
 control of emergency intermittent releases as described by the
 commenter.
 Comment;
       One commenter (IV-D-46) stated that the requirement to capture
 high pressure non-decomposition emergency releases would create safety
 risks that are unacceptable and would,  therefore, force a phase out of
 existing high pressure processes.  According to the commenter, existing
 facilities would not be able to compete with other processes because
 when it becomes necessary to modify or  reconstruct,  the existing
 facilities would be required to be retrofitted to send this relatively
 small quantity (the commenter estimated emergency emissions including
 decompositions to be less than 10 percent  of total  intermittent
 emissions)  of non-decomposition emergency  releases  to  a flare.   New
 facilities  would also  be required to send  the  non-decomposition releases
 to a flare.   If technology  were available  to do  this,  it  would be  cost
 prohibitive  for the typical  high  pressure  line or plant as  seen in
 Docket Item  IV-B-12.   Typical  high pressure  lines have a  peak  flow of
 50,000 scfm  of 100 VOC  weight  percent to 300,000 scfm  of  100 VOC weight
 percent.  Based on their experience, the 50,000  scfm lines  typically
 have emergency releases of  less than 10 Mg/yr  and the  300,000  scfm lines
 typically have emergency releases  of less  than 40 Mg/yr.  The  cost
 analysis in  Docket Item IV-B-12 apparently has made the assumption that
 technology is  available to  send high pressure  peak flows to a  flare.
 Hypothetically,  if one  could accomplish this,  then the cost developed  in
 Docket  IV-B-12  would be approximately $15,000/Mg  for the 50,000 scfm,
 100  VOC weight  percent  lines (from Table 2 of Docket Item IV-B-12).  No
 data  are available  in that docket  for the 300,000 scfm, 100 VOC weight
 percent lines.  According to the commenter, such  costs, technological
 uncertainty regarding required controls, and safety risks would likely
 lead  to the phase out of high pressure processes, although high pressure
 process plants currently:  (1) operate safely,  (2) have an inherent low
 percentage of fugitive  emission leakers  relative to other processes
 because in high pressure service,  even  a "pin hole"  leak cannot be
tolerated,  (3) can effectively control  continuous emissions  by  using
currently available technology, (4) can  effectively control  non-
                                  3-17

-------
 emergency intermittent releases (approximately 90 percent of total
 intermittent releases) by using currently available technology, and (5)
 offer products that cannot be produced by other processes.
 Response:
       As noted in the response on the previous page,  the "high pressure
 non-decomposition emergency releases" have been controlled by at least
 one company in the past and,  as far as the Agency is  aware,  without
 incident or compromise of safety.   However,  as also noted in a previous
 response, the Agency does not believe it  is  reasonable to require
 control  of such releases.  The final  rule has  been clarified as to this
 point.   Thus, the commenter's concern has been alleviated.
 Comment:
       One commenter (IV-D-46) stated  that the  costs of control  for
 emergency releases ($15,000/Mg)  are significantly understated because
 they do  not include the cost  of an  emergency vent collection system to
 protect  the flare system.  The commenter  estimated the preliminary cost
 of the required system to be  $50,000/Mg to $110,000/Mg depending  on
reactor  size and stated that  safety considerations lead them to believe
 that even these estimates understate  true costs.
 Response:
      The costing of the separa-te flare system to control  emergency
 releases from a LDPE,  high  pressure process  included  the  flare, ducting,
 and particulate removal  equipment per information provided by the
 company  that was controlling  these  emissions at  that  time  (see  Docket
 Item II-D-105).   Thus,  the  Agency believes it  has adequately included
 the necessary equipment for controlling such releases.  Regardless,  the
 Agency continues to believe that the  cost of controlling these  releases
 is  unreasonable in  light of the emission  reduction  achieved,  and has
 clarified  the wording  in the  final  rule to exclude  these emissions  from
 control.

 3.4  CONDENSERS

 Comment:
      Two  commenters  (IV-D-8,  IV-D-13) stated  that while the control
 technology exists  to achieve  final  condenser outlet temperatures of
 -24*C during  steady state operation of a  polyethylene  terephthalate)
                                  3-18

-------
 plant,  there  are routine stages  of the operation of a plant which make
 this  temperature unachievable.   The commenters  pointed out that the
 composition of the vent  stream during  startup,  shutdown,  and process
 upsets  varies  to the  point  that  maintaining  an  outlet temperature of
 -24*C would lead to freeze  ups and further process  upsets.   The water
 content,  polymer carryover,  and  other  contaminants  in the stream during
 startup or shutdown,  the commenter continued, will  affect the
 temperature at  which  the condenser outlet can be operated without
 freezing  during this  portion of  the  processing.   Therefore,  the
 commenters stated,  if a  gas temperature  is specified  in the  standard  and
 is included as  a permit  parameter,  there will be  times when  the process
 must  perforce violate the permit and adequate recognition  that  the
 standards do not  apply during startup, shutdown,  or malfunction
 conditions must  be given.
      The commenters  then suggested that the regulatory language  be
changed to reflect that  if refrigerated condensers are used  for control,
then when the process  runs at steady state the outlet gas temperature
should be -24C.  The  commenters  provided the following suggested
language for the following paragraphs from the September 30, 1987,
Federal  Register notice:

      60.562-l(c)(l)(ii):  If refrigerated condensers are used for
      emissions control,  not allow the  outlet gas temperature from
      each final condenser in  the materials recovery section (i.,
      methanol  recovery)  to  exceed  -24'C  (-ll'C), at steady state.

      60.562-l(c)(3)(iv):  If  the dimethyl terephthalate  process is
      being used, not  allow  continuous  TOC emissions  from  the material
      recovery  section (i.e., methanol  recovery)  to  be greater  than
      0.0027 kg  TOC/Mg product; or  if refrigerated condensers are
      used for emissions  control, not allow the outlet gas  temperature
      from each  final  condenser in  the  material recovery section (i.e.,
      methanol recovery)  to exceed  -24'C  (-ll'F),  at steady  state.

      60.562-l(c)(4)(iv):  If the dimethyl terephtalate process  is being
      used, not  allow  continuous TOC emissions from the material
                                 3-19

-------
      recovery section (i.e., methanol recovery to be greater than
      0.0027 kg TOC/Mg product; or if refrigerated condensers are
      used for emissions control, not allow the outlet gas temperature
      from each final condenser in the material recovery section (i.e.,
      methanol recovery) to exceed -24C (-HF), at steady state.
Response:
      Periods of startup, shutdown, and malfunction are not considered
to be in violation if they exceed the expressed emission limits, as
provided for in the General Provisions, Section 60.8(c):

            "...nor shall emissions in excess of the level  of the
            applicable emission limit during periods of startup,
            shutdown, and malfunction be considered a violation of the
            applicable emission limit unless otherwise specified in the
            applicable standard."
Since the proposed standards do not specify otherwise,  the General
Provision section was assumed to be prevailing.  However,  the
alternative temperature standard does not necessarily fall  within the
definition of "emission limit" as used in Section 60.8(c).   The Agency
Intends the same treatment to be accorded the alternative  temperature
standard as for a true emission limit.  Owners and operators are,
however, required to maintain and operate any affected  facility
Including associated air pollution control  equipment in a  manner
consistent with good air pollution control  practice for minimizing
emissions, to the extent practicable,  at all times including periods of
startup, shutdown,  and malfunction [General  Provisions,  Section
60.11(d)].  Therefore, the promulgated standard includes the commenter's
suggestion.
Comment;
      One commenter (IV-0-6) stated that the uncontrolled  emission  rate
of 0.016 kg TOC/Mg of product for continuous polystyrene plants needs to
be reconsidered to take into account  the presence of water  in the
material recovery condenser vent stream.   The commenter stated further
that it appears inappropriate that the newest facilities built with the'
latest devolatilizing vacuum and refrigeration technologies cannot  meet
this uncontrolled emission level.   The commenter suggested  that this
                                 3-20

-------
 value, if appropriate at all, needs to be in the 0.050 kg TOC/Mg product
 range.
 Response:
       The uncontrolled emission rate for the material recovery section
 from polystyrene plants has been recalculated (see Docket Item IV-B-18)
 based on the new data concerning water in the material recovery
 condenser vent stream.  The new uncontrolled threshold emission rate has
 increased to 0.05 kg TOC/Mg product.  This increase reflects the use of
 a spared condenser system to bypass the potential  freezing problem of
 using subfreezing temperatures in the condenser.
       The commenter refers to the inappropriateness of the proposed
 uncontrolled emission rate by referring to the newest facilities  with
 the latest devolatilizing vacuum and refrigeration technologies not
 being able to meet that level.   The commenter appears to  presume  a
 relationship between the level  of uncontrolled emissions  from a facility
 that installs the latest process equipment and the level  of emissions
 that can  be achieved when air pollution control  is sought.   The Agency
 disagrees with this apparent assumption.   The uncontrolled  emissions
 from an  industrial  facility in  the  absence of environmental  regulation
 is  typically determined  by  a different  set of economic and  cost criteria
 than the  criteria used in setting environmental standards.   The lower
 level  of  emissions  required by  the  standards  does  not  say  anything  about
 the technical  capabilities  of th,e latest  equipment  installed  by
 industry,  but reflects the  use  of emission control  equipment  that  allows
 further reduction in  emissions.  As  noted  above, the Agency  reevaluated
 the appropriateness  of the  control  technique  used  and  as a result of
 this  revaluation has  increased  the  uncontrolled emission rate.
 Comment;
      One commenter  (IV-D-12) stated that the vapor streams  from the
 material  recovery (methanol  recovery) section of PET processes, both
 high and  low viscosity dimethyl terephthalate (DMT), are laden with
water vapor.  The commenter  pointed out that the concentration of TOC
emissions and condenser temperature are regulated in Sections 60.562-
 l(c)(l)(i) and (ii) and 60.562-l(c)(4)(iv) of the September 30, 1987,
Federal Register  notice.  According to the commenter,   if a refrigerated
condenser were used as the final condenser in the material  recovery
section,  the vapor stream would have to be dried before entering the
                                 3-21

-------
 condenser or the  condenser would  freeze and plug.  The  commenter  stated
 that  such a drier for  that large  a flow and concentration would be
 prohibitively expensive  in terms  of capital and operating cost and
 should be excluded.
 Response:
      The Agency  has reevaluated  the regulatory alternatives  for  the
 material recovery section from PET/DMT processes to taken into account
 potential freezing problems  (see  Docket Item IV-B-18).  However,  rather
 than  using a drier on  the stream, the Agency used a lower cost approach
 of analyzing the  potential emission reduction and cost  using  a spared
 condenser system.  This  has  resulted in a revision to the standard for
 this  process section.  Based on the revised analysis, the final rule
 sets  an emission  limit of 0.018 kg TOC/Mg product for material recovery
 sections.  Alternatively, an owner or operator of an affected facility
 may limit the outlet temperature  of the final  condenser to +3C (+37F).
 At proposal, these limits were 0.0027 kg TOC/Mg product and -24C
 (-1TF), respectively.   In addition,  the uncontrolled threshold emission
 rate  increased to 0.12 kg TOC/Mg  product.

 3.5   POLYETHYLENE TEREPHTHALATE) FACILITIES

 Comment:
      One commenter UV-D-12) stated  that  several  features of their PET
 process technology render the proposed standards either inappropriate or
 technically impossible.  Specifically,  the commenter referred to the
 ethylene glycol  vapor stream that exits the steam jets (which are used
 to draw the vacuum necessary to drive the  polymerization reaction to
 completion)  is first cooled,  and then sent to  process waste treatment.
The commenter indicated that this stream would  be (presumably)
 regulated,  according to the September 30,  1987,  Federal  Register notice,
by Sections 60.562-l(c)(l)(iv)  and (4)(ii) for  DMT processes and by
Section 60.562-l(c)(2J(ii) for  TPA processes.
      According  to the commenter,  (1)  no appropriate limits can be set
for the ethylene glycol condensate from the vacuum system servicing the
polymerization reaction because that  condensate  stream cannot be
analyzed separately from the vastly larger steam jet condensate stream,
                                 3-22

-------
 and (2) the very dilute ethylene glycol in the combined condensate
 stream is not recovered, but is biodegraded harmlessly.  As such, the
 commenter continued, its release into the environment is adequately
 regulated under provisions of the Clean Water Act, and its vanishingly
 low vapor pressure in wastewater solutions at ambient conditions results
 in insignificant air emissions from trade waste basins.
       The commenter stated that distillation of such a dilute stream to
 recover the ethylene glycol is economically prohibitive,  and use of a
 refrigerated condenser is impossible,  as there is no practical  way to
 dry the ethylene glycol  vapor stream.   Therefore, the commenter
 requested that this stream be excluded from the regulation because it is
 already regulated by National  Pollutant Discharge Elimination System
 (NPDES) permits, and it  is neither cost-effective nor technically
 feasible to regulate it  as proposed.   The  commenter pointed out that
 their stream is  analogous to the "extruder quench vent stream"  excluded
 in the preamble  (see 52  Fed.  Reg.  36691, Col.  2)  for cost  reasons and
 should be similarly treated.
 Response:
       Further  conversation with  the commenter  has clarified this  comment
 (see Docket Item IV-E-39).   The  commenter  had  misunderstood as  to where
 the ethylene glycol  was  to have  been tested.   The proper testing  point
 for the commenter's  systems  is  in  the  vacuum system's  liquid  effluent
 (e.g.,  in  the  hotwell) prior to  the cooling tower.   Thus,  the commenter
 agrees  that  the  stream can  be analyzed  (see Docket  Item IV-E-54).  The
 Agency  points  out that for  plants producing high  viscosity  PET  using
 multiple end finishers this testing is only required when the effluent
 is  recirculated  through  a  cooling tower or similar  facility.  If  a once-
 through  (one-pass) system  is used for the cooling water at  these  plants,
 then the liquid  effluent from the vacuum system is  not required to be
 tested.  (Note,  however,  that the rule requires testing of  the liquid
 effluent at  all  other PET facilities,  including those that  produce high
 viscosity PET using a single end finisher,  whether or not a cooling
 tower is used.)
      The second point by the commenter really deals with the need for
daily testing of this stream.  The Agency has reconsidered this point,
and has included in the final rule a reduced testing program for those
facilities that meet certain criteria  (see  Sections 60.564(j)(l)(ii)  and
                               .  3-23

-------
(iii) of the final rule).  The final rule also allows an owner or
operator to use an alternative approach that relates ethylene glycol
concentration to another parameter, such as chemical oxygen demand  (COD)
or biological oxygen demand (BOD) (see Section 60.564(j)(2) of the  final
rule).
                                 3-24

-------
                 4.0   CONTROL  TECHNOLOGY  -  EQUIPMENT  LEAKS

 4.1  GENERAL

 Comment:
       One commenter (IV-D-8)  stated that in order to achieve consistency
 with Subpart VV [Section 60.480(c) and Section 60.480(d)(2)], Section
 60.560 should contain the fugitive equipment modification  provision and
 a design  capacity exemption for applicability of affected  facilities.
 Two other commenters (IV-D-7  and IV-D-13)  also recommended adding the
 design capacity exemption.   Commenter IV-D-8 suggested the following
 paragraphs be added:

       60.560(a)(3):   Addition or replacement of equipment  for the
       purpose of process improvement  which is accomplished without
       a capital  expenditure shall  not by itself be considered a
       modification under this Subpart.

       60.560(a)(4):   Affected Facilities with a design capacity to
       produce less than  1000  Mg/yr shall be  exempt from Sections
       60.562-1  and 60.562-2.
 Response:
       The  Agency is  aware of  no  differences  between  the polymer
 manufacturing industry and  the  synthetic organic  chemical  manufacturing
 industry  (SOCMI)  that would form a  basis for not  including in these
 standards  the two  sections  recommended by  the  commenters as they apply
 to  equipment  leaks.  Therefore,  the Agency has  revised  the standards  to
 include these two  paragraphs  for equipment leaks  (see  Sections
 60.560(a)(4)(i)  and  (ii) of the  final rule).   In  addition,  a definition
 of  "capital expenditure" has  been  added.
 Comment:
      One commenter  (IV-D-1)  pointed out that Subpart VV does not make
 any distinction  between valves of different  sizes and suggested  that
valves 1"  or  less  in size be  exempted because there  is  very little
return in  terms of reduced emissions for the increases  in  the volume of
work tagging  and monitoring these size valves.
                                   4-1

-------
 Response:
       In developing  similar  standards for petroleum  refineries,  EPA
 analyzed both  the emission factor dependency on valve  size  and  the costs
 of  small valve repair  as compared to the average repair cost  estimates
 (Docket Item II-B-82).  These analyses indicated that  the emission rates
 from small valves have not proven to be lower than those for  larger
 valves, and that repair of small valves is no more difficult  or  costly
 than repair of large valves.  The EPA concluded that,  since both the
 environmental  benefits and cost of control for small valves are
 comparable to  those for larger valves, small valves  are cost  effective
 to  control.
      The  above conclusion was also reached in a similar standard for
 the natural gas production industry.  In that project, emission  source
 test data were collected at  six natural  gas/gasoline processing  plants
 by  EPA and industry.  Valves of all  sizes were monitored in these tests,
 and a relationship between valve size and emissions  could not be
 determined. The emission factors developed for the natural  gas
 production industry were developed based on these data for all valves in
 gas processing plants.  The  basis of the gas production standards for
 valves included all  valve sizes and was  determined to be cost effective.
 Therefore, the Agency does not believe an exclusion  for small  valves is
 warranted, and such an exclusion is  not  in the promulgated standards.
 Comment:
      One commenter (IV-D-40) asked  where in the proposed standards one
 was to quantify fugitive emissions.   The commenter noted that the EPA
 standards for  "equipment leaks"  are  now  being proposed for inclusion in
 the regulations (Section 60.562-2 Standards:   Equipment leaks of VOC),
 but these standards  as such do not address the question of quantifying
 the leaks or fugitive emissions  in terms  of kg/Mg of product.   The
 commenter pointed out that when  submitting an application  for a "permit
 to construct" under individual  state rules (especially Louisiana and
Texas), all emissions have to be quantified,  including the  fugitive
emissions.  The corresponding rules  for  equipment leaks (Part 60--
Subpart VV) have already been applied to  VOC emissions from polymer
plants in the states referred to above (in LAC Paragraph 2121  and TAC
 115.271, respectively).  This means,  of  course,  that emissions defined
 in Table 1, as  well  as leaks  based on quarterly  measurements (usually)
                                  4-2

-------
 have  been  declared in  recent  permit  applications.   When actual
 measurements  are  available, they should  be  used for estimating  overall
 leak  rates.   Permits  issued,  at  least  by the  Texas  Air Control  Board,
 also  specify  compliance  with  the proposed "Subpart  ODD" as  a permit
 requirement.
 Response:
      The  application  of this  standard is different from procedures or
 requirements  associated  with  obtaining permits  from State or local
 agencies.  Whereas those permits  may require  an  owner  or operator  to
 quantity all  emissions,  quantification of fugitive  emissions are not
 required under Subpart VV or  under these  standards,  which adopt Subpart
 VV for polymer manufacturing  facilities,  because the control
 requirements  are work practices  and equipment requirements  and,
 therefore, do not  require quantifying the emissions.
 Comment:
      Two commenters (IV-D-7,   IV-D-8) stated that the  fugitive emission
 requirements should be modified to improve cost effectiveness of the
 control and that the required control program should be  flexible enough
 to account for the effect of new technology, state  regulatory programs,
 and the chemical  characteristics of the compounds that could be emitted.
The commenters proposed several changes to the valve monitoring
 provisions in order to achieve this flexibility.  Commenter  IV-D-8
 stated that these changes would still  obtain maximum emission control
 and desirable cost effectiveness.  The changes proposed by the
commenters were:

      1.   When a  facility (new, modified, or reconstructed)  becomes
          subject  to the  fugitive emissions  portion  of the regulation:
          (1)  require  testing  of the  valves  within  180 days  to determine
          the  percentage  of  leakers,  (2)  do  not  allow any special  or
          additional maintenance  (other than that normally practiced) to
          be done  45 days prior to the  test,  and (3) use the results of
          the  test  (percent  leakers)  to determine the level  of future
         monitoring.

      2.   For  valves, consider a  plant with  less  than 2 percent  leakers
         to be a well  controlled  plant and  require  annual monitoring.
                                  4-3

-------
           If during a subsequent annual test, the percent leakers  exceed
           2 percent, require the plant to monitor quarterly or monthly
           depending on the percent leakers.  If the plant has between 2
           percent and 10 percent leakers, consider it to be an average
           performer and require quarterly monitoring.  If the plant has
           more than 10 percent leakers, require it to do monthly
           monitoring and follow the existing regulation. Commenter IV-
           D-8 stated that the economics of the proposed regulation are
           based on emissions determined through the use of average SOCMI
           factors for valves.  Therefore, the commenter reasoned, a
           plant that has greater than 10 percent leakers should be
           required to monitor valves once per month as presently
           proposed (the percent of leaking valves used to determine the
           average SOCMI factor was 11 percent).   Finally, allow plants
           exceeding 2 percent leakers to work their way to annual
           monitoring by following the requirements of Subpart VV Section
           60.483-2 (Alternate standard for valves--skip period leak
           detection and repair).
Response:
      As adopted in these standards and as promulgated under Subpart VV,
valves in  gas/vapor service or light liquid service are to be monitored
monthly to detect leaks (Section  60.482-7).   A valve that has not been
found to leak for 2 consecutive months may then  be monitored quarterly
until a leak is detected,  in which case monitoring reverts to a monthly
basis [Section 60.482-7(c)].   Alternatively,  the monitoring  interval  can
be increased to quarterly and then to annually if the percent of leaking
valves is  equal  to or less than 2.0 for a specified number of detection
periods [Sections 60.483-2(b)(2)  and (b)(3)].   If more than  2.0 percent
leakers are detected,  the monitoring interval  reverts to  monthly leak
detection  and repair (Section 60.482-7).   Alternatively,  an  owner or
operator may elect to comply with an alternative standard that limits
the percent leakers to a 2.0 or less (Section  60.483-1).   Monitoring
under Section 60.483-1 would be conducted annually.
      The Agency believes  that these standards  are already flexible with
regard to  specifying and adjusting monitoring  intervals  on the basis  of
how well  a plant is controlled.   In contrast,  the  commenters are
requesting that  the standards be  revised  so  that the  initial  monitoring
                                  4-4

-------
 interval  (monthly, quarterly or annually) is based upon the results of
 the initial performance test.  In order to avoid companies from
 purposely minimizing the number of leakers and thus to start with a
 longer monitoring frequency, the commenters proposed that "special or
 additional" maintenance be prohibited 45 days prior to the initial
 performance test.
       The Agency does not believe that the initial monitoring period can
 be reasonably set based on a single initial  performance test.  (Note
 that the annual monitoring interval under Section 60.483-1 is based upon
 the decision to comply with this alternative standard and not upon the
 results of an initial performance test.)  Some facilities may test out
 at less than 2.0 percent leakers (when in fact their historical  leak
 rate may be higher),  and the next test would not be for 12 months under
 the commenters suggestion.   The purpose of the skip period (Section
 60.483-2) is to allow for the development of a historical  basis  upon
 which longer monitoring intervals can  be justified.   The  commenters
 recommendations could result in an  increase  in emissions  that  would be
 otherwise controlled  in a cost  effective manner under  the  standards as
 currently written.   Furthermore,  the Agency  has determined that monthly
 monitoring is  incrementally cost  effective versus  quarterly monitoring
 regardless of  whether the leak  rate is between  2  and  10 percent or over
 10 percent.  The commen.ter  did  not  provide data to  show otherwise.
 Thus,  the Agency sees no  reason to  create an  artificial range  of  2  to  10
 percent  and  allow initial quarterly monitoring.   Finally,  the  current
 skip  period  procedures  require  relatively few  additional monitoring
 periods  before  the  intervals  recommended by the commenter  are  reached.
 For example, since quarterly  monitoring may be  allowed after 2
 consecutive  quarters  showing  2  percent or less  leakers, a  plant showing
 over 2 percent  leakers  in its initial   test could perform just 4
 additional monitoring months  before-instituting a quarterly monitoring
 program.
      The  standards do not need to try to prohibit special maintenance
 prior to the performance test.  If an  owner or operator truly believes
 his facility is  so well controlled as  to have less than 2 percent
 leakers, then the owner or operator can comply with Section 60.483-1,
which provides for annual monitoring.   It is then incumbent on the owner
                                   4-5

-------
 or operator to  continue  sufficient maintenance so as to ensure
 compliance with  the  alternative standard.
       For these  reasons, the Agency has decided not to revise the
 relevant portions of Subpart VV as requested by the commenters.
 Comment;
       Commenter  IV-D-8 felt that a plant whose valve population has less
 than 2 percent leakers is well controlled and should not be penalized
 along  with the lesser controlled plants.  But if its performance slips
 (has greater than 2  percent leakers among its valves),  then it should be
 required to correct  its mistakes, not be punished by the threat of a
 non-compliance penalty.  In Section 60.483-2 (Alternative standard for
 valvesskip period  leak detection and repair)of Subpart VV, EPA allows
 a plant that has "paid its dues" (advanced to annual monitoring by first
 doing  monthly and quarterly monitoring) to do more frequent monitoring
 if 1t  fails the annual 2 percent leaker test for valves,  but without
 threat of non-compliance.  The regulatory penalties should be the same
 [for a facility] that has complied prior to regulation  and one that
 compiles (through the skip monitoring alternative)  after promulgation of
 a regulation.  This commenter thus recommended that the non-compliance
 part of Section 60.483-1 be eliminated.
 Response:
      The Agency disagrees that a plant electing to comply with Section
 60.483-1 is penalized relative to a plant electing  to comply with
 Section 60.482-7.  Both plants are subject to non-compliance penalties.
The difference lies in the basis for determining compliance/non-
compliance,  which reflect the two different standards.
      The Agency agrees that plants which constantly achieve a low level
of emissions should not have to bear the cost of and adhere to a
mandated inspection repair program (i.e.,  Section  60.482-7).  The
average level  below which-leak detection and repair (LDAR)  is not
required is  1 percent leakers.   In order to take into account test-to-
test variability, the performance level  is 2 percent.   The  Agency is
aware of no  data showing that a plant which averages 1  percent will  have
a significant probability of producing test results  over  2  percent.
However,  if an owner or operator believes that  the  leak rate may  exceed
2 percent,  then he may select LDAR to avoid the chance  of a non-
compliance penalty under Section 60.483-1.
                                  4-6

-------
 4.2  ASSUMPTIONS AND EFFECT ON EMISSION ESTIMATES

 Comment:
       Several  commenters  (IV-D-1,  IV-D-7,  IV-D-8) expressed concern over
 a  number  of assumptions made in BID Vol.  I  concerning  fugitive
 emissions,  and in turn the  estimates of fugitive  emissions  from polymer
 manufacturing  plants.
       Commenter IV-D-7 specifically referred  to the  Agency's  assumption
 that  there  are no leak detection and repair programs in  place in the
 polymer industry and that the  SOCMI  factors underestimate fugitive
 emissions.   According to the commenter, the Texas Air  Control  Board  has
 for several  years required  fugitive  monitoring and repair programs  on
 permits for  new plants and  major modifications to existing  plants.   The
 commenter noted that experience with  monitoring in polymer  plants and
 other  SOCMI  plants indicates that the factors overestimate  fugitive
 emissions.
       Commenter IV-D-1 stated that the application of Subpart  VV
 (fugitive emission standards) to polypropylene plants does  not appear to
 add any major  new requirements to these plants because equipment leaks
 are already regulated under  Louisiana and Texas rules,  which basically
follow the directives of Subpart VV, and most polypropylene plants are
located in Louisiana and Texas.
      Commenter IV-D-8 indicated the following assumptions and
statements as being  inaccurate:

      1.   The assumption  that SOCMI  and  polymer plants  are
          technologically  similar and that  average SOCMI  factors
          understate  emission reductions.   The EPA assumes the major
          monomers used  in polymer  production  are  among the  chemicals
          found to have the'highest  leak and emission rates  using the  '
          SOCMI fugitive data base.   While  it  is true that the ethylene
          valves and  pumps leaked twice  as much as the  average SOCMI
          plant, there are major differences between  the  ethylene SOCMI
          plant and the entire  polymer manufacturing  industry.

          a.  Ethylene is the only compound  common to both the  SOCMI
             studies and polymer manufacturing.   There are  many  gases,
                                  4-7

-------
    monomers and solvents used in polymer processes  that  have
    very different physical and chemical properties  from
    ethylene and are used more during the process.   These
    compounds have similar properties as other chemicals  in
    the SOCMI study that leaked approximately an order of
    magnitude lower than ethylene.

b.  Since the operating pressures for the majority of polymer
    manufacturing units are much lower than for an ethylene
    plant (0-400 pounds per square inch gauge (psig) vs.  1000-
    2000 psig respectively), it is reasonable to expect that
    they leak less.  This was determined from the EPA six
    plant maintenance study on SOCMI plants (including two
    ethylene plants), which discovered a correlation between
    line pressure and-fugitive emission rate.  The EPA states
    "and in almost every case examined,  higher leak
    frequencies (i.e.,  emissions)  were associated with higher
    line pressures."  (EPA Document 450/3-82-010.  "Fugitive
    Emission Reductions, and Costs",  p.  2-30.)  On the other
    hand, the high pressure LDPE processes operate at about
    20,000 psi.  Whenever these plants leak, the high noise
    level and visibility of the escaping chemicals will  result
    in an almost immediate process shut down for repair.
    Therefore,  leaks will  be detected quickly.

c.  Some of the compounds  used in  the manufacture of polymers
    have distinctive odors (i.e.,  styrene)  which make leak
    detection and equipment repair much  easier.   The compounds
    in the 24 SOCMI plant  study that  had low odor thresholds
    had very low emissions (<2.0%  leakers vs.  23% leakers for
    ethylene service).   This is not  taken into consideration
    in the proposed rule.

On page 6-2 of BID Vol.  I,  EPA states that  "...  fugitive
emissions are proportional  to the  number of potential sources,
but are not related to  capacity, throughput, or age."  It has
been 7-9 years since testing was done to determine average
                        4-8

-------
           SOCMI factors.  In those 7-9 years many advances have been
           made in various components to reduce fugitive emissions.  This
           includes better valve designs and packings, better mechanical
           seals for pumps, and improved seating surfaces for safety
           valves.  All these advances contribute to create new
           components that have a lower potential to leak.   Comparisons
           of screening data from old and new components in the same
           service clearly confirm this.

       3.   The EPA states that no fugitive emissions sources are
           currently subject to any sort of routine LDAR program.    This
           may have been true when BID Vol. I was written,  but is  not the
           case now.   Most of the states where polymer plants  are  located
           have or will have in place a LDAR program for new and existing
           polymer manufacturing plants.   Most of these states require
           quarterly  monitoring of valves and monthly monitoring of
           pumps.
           For units  under a State LDAR program,  the  proposed  NSPS  would
           only achieve a 9.0  percent  reduction  in  gas  valve emissions
           instead  of the 73  percent  stated in  BID  Vol.  I (73  percent
           reduction  in emissions  for  monthly monitoring minus  64 percent
           reduction  in emissions  for  quarterly monitoring).   Similarly
           for light  liquid  valves, the  reduction would only be  15
           percent  instead of  the  59 percent  EPA  assumes (59 percent
           minus 44 percent).  There would  be no  emission reductions
           achieved for pumps  since the State programs  already require
           monthly monitoring  of pumps.  Using these  percentages,
           reductions  in  Table 8-5 of  the BID results in only a
           15.1 Mg/yr reduction in VOC emissions, not a 58.7 Mg/yr
           reduction.
Response
      For  the reasons discussed below, the Agency continues to believe
that the assumptions and methodology used in the BID to estimate
emissions  from the polymer manufacturing industry is a reasonable
approach.   Each of the specific points made by the commenters are
addressed below.
                                  4-9

-------
Applicability  of  SQCMI  Factors  to  Polvmer Manufacturing  Plants.
Commenters  IV-D-7 and  IV-D-8  basically questioned the  applicability  of
using SOCMI  fugitive emission factors for estimating fugitive  emissions
from polymer manufacturing plants.  Commenter  IV-D-8 identified  three
specific reasons  as to  why it would be inappropriate to  apply  SOCMI
factors to the polymer  manufacturing industry  and how  in doing so
fugitive emissions from polymer manufacturing  plants are overestimated.
The points are related  to:  (1) the types of compounds used  in polymer
plants relative to those in SOCMI  plants; (2)  the operating  pressure of
polymer units; and (3)  the distinctive odors of certain polymer
compounds.   The specific points are addressed  following our  general
response.
General.  The development of  VOC emission factors for equipment  leaks  is
founded on the concept  that for a given screening valve equipment leaks
VOC at the same rate regardless of the industry or process unit.  Total
emissions or average emission rates may vary,  however,  based upon the
relative distribution of screening valves (i.e., components  leaking at
different rates)  found  in different industries or process units.  For
example, the original SOCMI emission factors were derived from petroleum
refineries.  After further study,  the SOCMI fugitive emissions data
showed a difference in  the number of leaking and non-leaking sources
(leak frequency)  compared to the petroleum refinery data.  The original
SOCMI fugitive emission factors (which were derived from the emission
factors from the  Petroleum Refinery Assessment Study)  were therefore
adjusted to more  accurately'reflect the number of leaking and non-
leaking sources found in the SOCMI process units tested (the SOCMI 24-
Unit study).
      In gathering data on the SOCMI,  EPA sampled a number of vastly
different chemical process units.   Characteristics  of individual
chemicals were considered in -selecting the process  units sampled during
the 24-Unit Study.  Hi-volume, low-priced chemicals  were included, as
were lower volume, higher-priced chemicals.   Chemicals  with widely
divergent volatilities were included along with chemicals that are
particularly odoriferous.   Not surprisingly, the frequencies of leaks
found ranged from nearly zero  to thirty percent.  Leak  frequencies
provide an indication of the  relative  quantity of mass  emissions and it
is apparent from this range of leak frequencies that fugitive emissions
                                  4-10

-------
  from  some  units will  be  higher  and  some  lower  than  estimates  based  on
  the SOCMI  factors.  The  mass  emissions estimates generated  by EPA using
  the average SOCMI emissions factors  represent  an average  that is
  applicable to  industry-wide emissions estimates.  Without conducting a
  rigorous Method 21 survey to  determine the leak frequency (thereby
  generating average emission factors  based on the leak/no-leak factor
  components), the average SOCMI  emission  factors stand as the  best
  estimators of  fugitive VOC emissions currently available.
       The major factor affecting the percent of leaks detected  (or  leak
  frequency) from any equipment type is the vapor pressure of the
  substance in the line.  This  finding forms the basis for separation of
 different types of equipment  by service  (gas, light liquid,  heavy
 liquid).  The substance in the line does not necessarily mean the
 substance produced as the final  product.   Thus, the primary concern  is
 whether or not the leak frequency distribution  found in polymer
 manufacturing plants  is sufficiently similar to those found  in the SOCMI
 plants.   The commenters have identified  three specific areas that they
 believe  invalidate the use of the SOCMI  emission  factors.   These are now
 discussed.   For the  reasons  described below,  the  average leak
 frequencies found  in  the  SOCMI study are  considered  to be  representative
 of frequencies  in  polymer manufacturing plants.
       Compounds Used  in the  Industry.  The commenters  stated that
 ethylene is  the only compound  common  to both  the SOCMI  studies and
 polymer  manufacturing.  This  is  in error.  As pointed  out  in Table 1 of
 Docket Entry No. II-B-82, Docket No.  A-82-19, there  are  at least  seven
 major  organic compounds used in  polymer manufacturing  that are listed in
 SOCMI.   These chemicals are ethylene, propylene, butene, isopropyl
 alcohol, ethylbenzene,  styrene,  and vinyl acetate.   As shown  in Table 2
 of Docket Entry No. II-B-82, leak frequencies for five of the  compounds
 are available.  These are propylene, ethylene, ethane, propane, and
 vinyl  acetate (a common comonomer used in polyethylene copolymer
 production).  SOCMI leak frequency data are also available for three
 chemical  units -- ethylene, vinyl acetate, and high density polyethylene
 (HPDE) -- the latter of which  is a polymer manufacturing plant.
      The leak frequencies for propylene and  ethylene chemical types  and
for ethylene and HOPE units are much higher overall  than those for the
average SOCMI model  unit.   The leak frequencies  for vinyl acetate
                                  4-11

-------
 chemicals  and  units  are  much  lower  than  for  the  average  SOCMI  model
 unit.   Similarly,  the  emission  factors for ethy'lene  ire  consistently
 higher  than  the  respective  average  SOCMI  emission  factors,  while  vinyl
 acetate's  emission factors  are  consistently  lower.   This  range of leak
 frequencies  and  emission  factors  for the  individual  chemicals  found  in
 the  polymer  manufacturing industry  is similar to the range  of  leak
 frequencies  and  emission  factors  for the  individual  chemical components
 used to derive the average  SOCMI  parameters.
      Considering the  available data on specific polymer  chemicals,
 there is no  evidence on  this  basis  to suggest that it is  unreasonable to
 use  the average  SOCMI  leak  frequencies and emission  factors for the
 purpose of analysis.
      Operating  Pressure.   In the past, industry has postulated that
 certain process  parameters  affect leak frequency and emission  rates.
 The  three  most commonly mentioned process parameters have been line
 size, process temperature,  and process pressure.  To the  extent,  then,
 that these parameters differ  between SOCMI and polymer plants, the
 appropriateness  of using the  SOCMI data for polymer  plants may be
 questionable.
      As pointed out in Docket Item II-B-82,  Docket No.  A-82-19,  EPA
 data currently available on line size support the argument that line
 size is unrelated to emission rates and leak frequencies and that line
 temperature has  been found  in general  to have no consistent effect on
 leak frequency.
      With regards to operating pressures (the factor identified by the
 commenter  in this instance), data available on line pressure does show
 that in almost every case higher leak frequencies for fugitive emission
 sources were associated with higher pressures.  Operating pressures of
 polymerization reactors may range from near atmospheric  (in polystyrene
 plants)  to very high pressures (up to about 50,000  psig  in high pressure
processes for low density polyethylene)  with  a number of processes
 between 100 and 500 psig.  These operating pressures  in  polymer plants
 fall  within the range of operating pressures  found  in SOCMI plants.
Comparison with ethylene plants only overlooks the  fact  that SOCMI
 factors and leak frequencies are average  numbers derived from data on
this wide range of operating pressures.
                                  4-12

-------
       Odor.  The commenters imply that the low rate of leaks in
 polystyrene plants is due solely to the distinctive odor of styrene that
 makes leak detection and repair much easier, and refer to the 24 SOCMI
 plant study.  The percent leakers in a plant is due to a number of
 factors, one of which may very well be odor.  If odor (or other factors)
 result in lower average percent leakers for a particular plant, fugitive
 emissions may well be lower and the use of average emission factors may
 overestimate emissions for certain facilities.   Nevertheless,  the cost
 of the standard,  on average,  will be reasonable across the polymer
 industry.
 Advances in Equipment.   The commenters stated that advances in  valve
 designs and packings,  better mechanica.l  seals for pumps,  and improved
 seating surfaces  for safety valves have led to  lower level  of  fugitive
 emissions.   The Agency agrees  that such advances  would be expected to
 reduce fugitive emissions.   Unfortunately,  there  is  no data to  quantify
 the magnitude of  the effects  of such advances.  Thus,  in  the absence of
 site-specific Method 21  testing,  the Agency believes it  is  using  the
 best data available and  that  these data  are reasonable to estimate
 emissions.
 State LDAR  Programs.  As pointed  out by  the commenters, State LDAR
 programs  now exist that  affect  polymer manufacturing plants.  The  Agency
 recognizes  that where  such  programs  are  in  place,  actual  baseline
 emissions for affected plants will  be  reduced from that in  the  BID.
 Thus,  as  a  result  of programs being  put  in  place  since the  BID, the
 emission  reduction  attributable to the standard is reduced.
       The amount by which the projected reduction would be  reduced
 depends on  whether  the State in which  the constructed, modified, or
 reconstructed  facility is located  has  a LDAR program and, if so, on  the
 particular  State LDAR program.  Table 4-1 compares the LDAR program
 under  Subpart  VV with the LDAR program identified by Commenter  IV-D-8
 and the LDAR programs descri-bed in Texas' and Louisiana's -State
 regulations.  The expected emission reduction, in terms of percent
 reduction from an uncontrolled plant, from these various LDAR programs
 are shown in Table 4-2.
      The Agency agrees with the commenters that where LDAR programs
exist the emission reduction attributed to these standards do decrease.
Table 4-3 shows the estimated emission reduction (from uncontrolled
                                  4-13

-------
                                      TABLE  4-1

                          COMPARISON  OF LDAR PROGRAMS
Component
Valves
- gas service
- light liquid
service
Pump seals
- light liquid
service
Safety valves
Compressor seals
Sampling
connect ions
Open-ended lines
Reauirement
Subpart W

monthly monitoring
monthly monitoring

monthly monitoring;
weekly visual
inspections
no detectable
emissions
no VOC emissions
to atmosphere; no
detectable emis-
sions; vent to
control device
closed purge system
or closed vent
system
cap, blind flange,
plug, or second
valve
Commenter"3

quarterly
monitoring
quarterly
monitoring

same as
Subpart W
none
none
none
none
Texas"

quarterly
monitoring
quarterly
monitoring

annual
monitoring;
weekly visual
inspections
quarterly
monitoring
quarterly
.monitoring
none
same as
Subpart W
Louisiana1"

quarterly
monitoring
annua I
monitoring

annua I
monitoring;
weekly visu
inspections
quarterly
monitoring
quarterly
monitoring
none
same as
Subpart VV
As identified in Table 4,  p. 23 of commenter's letter (Docket  Item IV-D-8).

Texas Regulation V.  Fugitive Emission Control in Synthetic Organic Chemical,  Polymer, and
Resin Manufacturing Plants in Harris County.  115.271 through  115.272.

Louisiana Air Pollution Regulations.  22.21, "Fugitive Emission Control."
                                        4-14

-------
                                    TABLE  4-2


           COMPARISON OF EMISSION REDUCTION FOR VARIOUS LDAR PROGRAMS
Component
Valves8
- gas service
- light liquid
service
Pump seals6
- light liquid
service
Safety valvesd
Compressor sealsd
Sampling
connections6
Open-ended linesd
Percent Emission Reduction
Subpart VV

73
59

61
100
100
100
100
Commenter

64
44

61
0
0
0
0
Texas

64
(0)b

(0)b
59
72
0
100
Louisiana

64
(0)b

(0)b
59
72
0
100
'  See Table 4-12, Docket Item II-A-32.


   A "negative" control efficiency was estimated for these monitoring
   intervals (see Tables 4-12 and 4-19 in Docket Item II-A-32).  Assumed zero
   for this analysis.


c  See Table 4-19, Docket Item II-A-32.

d  See Table 4-21, Docket Item II-A-32.
                                        4-15

-------
                              TABLE 4-3

         FUGITIVE EMISSION  REDUCTION  FOR THE SOURCES  IN
                  POLYMERS AND  RESINS  MODEL  PLANT


Emission
Source
VALVES
Gas
Light Liquid
Heavy Liquid
PUMP SEALS
Light Liquid
Heavy Liquid
SAFETY VALVES
OPEM-ENOEO LINES
COMPRESSOR SEALS
SAMPLING CONNECTIONS
FLANGES
TOTAL
Annual
Uncontrolled
Emissions,
Mg/ypa

19.7
32.6
1.1

12.6
5.6
10.0
0
4.0
3.4
17.5
106.4


Subpart Wb

14.4
19.2
0

7.6
0
10.0
0
4.0
3.4
0
58.7


Conwenter's
LDAR Program"

12.6
14.3
0

7.6
0
0
0
0
0
0
34.6


Texas/ LA
LDAR Program

12.6
0
0

0
0
5.9.
0
2.9
0
0
21.4
From Table 8-5, BIO Vol. I,  p. 8-21.


Calculated by multiplying the annual uncontrolled emissions by the percent  reductions shown in
Table 4-2.
                                    4-16

-------
 levels) for Subpart VV, the LDAR program described by the commenter, and
 the LDAR programs of Texas and Louisiana.  The incremental emission
 reduction attributable to these standards decreases from 58.7 Mg/yr to
 approximately 24 Mg/yr where the LDAR program described by the commenter
 is in place1 and to approximately 37 Mg/yr where  the Texas or Louisiana
 LDAR program is in place.  However, as discussed in the response to the
 next comment, the reduced emission reduction attributed to these
 standards does not make these standards cost ineffective.
       Lastly, the commenters stated that no new major requirements are
 applied to polypropylene plants in Louisiana or Texas.   As seen in Table
 4-1,  there are differences between Subpart VV and the LDAR programs
 (e.g.,  quarterly monitoring versus monthly)  described by the commenters
 and in  the Texas and Louisiana regulations.   In addition,  the main
 purpose of standards of performance is to require new sources,  wherever
 located,  to reduce emissions  to the level  achievable by  the  best
 technological  system of continuous emission  reduction  considering  the
 cost  of achieving such  emission reduction,  any  nonair  quality health and
 environmental  impact, and  energy  requirements...[(Section  lll(a)(l)].
 Where State LDAR programs  are  already  in  place, the  impact of the  NSPS
 will be smaller  than calculated.   The  State  LDAR  programs  and the
 systems chosen as  the best demonstrated technology for this  industry's
 standards  of performance for new  stationary  sources  are  not conflicting
 types of control;  therefore, where  State  LDAR programs apply, Subpart VV
 will supplement  them.   The EPA  has  determined that existing State  LDAR
 program-level facilities that become subject to Subpart  VV (e.g.,
 through modification) can  achieve the  additional  reduction required at a
 reasonable cost.
      Finally, it  is important  to note that Subpart VV is designed to
 take into account the effect of other control programs and equipment
 designs. It does this by providing  alternate standards for low leak rate
        This estimate is higher than the 15 Mg/yr stated by the
commenter.  This difference is due to an inconsistency in Table 4,
p: 58i ?f the commenter's letter.  A "differential control efficiency"
of 1.0 is indicated for safety valves.   This means that the standard is
100 percent more effective than LDAR program referred to by the
commenter, and the emission reduction due to additional control should
from 159tor24VMr/U? ^ ^^'  * increases the "mmenter's estimate

                                  4-17

-------
 equipment  and  plants.   Thus,  even  if the  overall  reductions  are
 overstated,  the standards  are reasonable  and  appropriate.

 4.3   COST  EFFECTIVENESS

 Comment:
      One  commenter  (IV-D-8)  stated  that  as a result  of  LDAR programs
 currently  in place or  soon to be in  place the emission reduction  to  be
 obtained under  Subpart VV will be  less and concluded  that  the
 incremental cost of  further reducing  already controlled  leaks through a
 more  stringent  LDAR  program (i.e., Subpart VV) is substantially higher
 than  the cost of reducing emissions  by initially applying  Subpart VV to
 an unregulated  unit.   The commenter  stated that the cost effectiveness
 of the regulation will  increase due  to this decrease  in  estimated
 reductions obtained  by the proposed  regulation.  The  commenter estimated
 annualized cost to now be $35,200 due to less credit  received for the
 VOC recovered (credit  - 15.1  Mg/yr x  $528/Mg), with incremental cost
 effectiveness being  $2,330/Mg  reduced ($35,200/yr divided  by  15.1
 Mg/yr).  This cost,  the commenter pointed out, is 11  times greater than
 the $208/Mg the Agency calculated in  the proposed standard and is not
 cost effective.
 Response:
      The incremental  cost effectiveness of applying  Subpart VV to
 polymer manufacturing  plants will vary depending on whether or not there
 is a current LDAR program in place.   In addition, where a  LDAR program
 is in place, the incremental  cost effectiveness of applying Subpart VV
will  vary depending upon the particular program in place.  The Agency
agrees that the incremental cost effectiveness of applying Subpart VV to
polymer manufacturing plants  complying with  a  LDAR program as described
by the commenter will be higher than if no LDAR program was in place.
However,  as shown in the analysis below,  the Agency also found that the
incremental cost effectiveness was  lower when  Subpart  VV was applied to
polymer manufacturing plants  complying with  a  LDAR program as described
in Texas'  and Louisiana's  regulations than if  no  LDAR  program was  in
place.  (This unusual result  occurs because  EPA's analysis shows that
annual monitoring of valves and pumps in  light liquid  service is
ineffective, providing no  emission  reduction.)   In  either case, the
                                  4-18

-------
 incremental  cost effectiveness of applying Subpart VV to polymer
 manufacturing facilities is still reasonable even given the LDAR
 programs referred to by the commenter.   Therefore, no changes to the
 fugitive emission standards have been made.
       Analysis.   As a first step in analyzing the incremental cost
 effectiveness of applying Subpart VV to polymer manufacturing plants,
 EPA compared Subpart VV with the LDAR program indicated by the commenter
 and the LDAR programs described in the  Texas and Louisiana regulations.
 (Texas and Louisiana were selected because of the large number of
 polymer manufacturing facilities located in  the two states.)   This
 comparison was shown earlier in Table 4-1.   The effectiveness of these
 different LDAR programs on emission reductions were also shown earlier
 in  Table 4-2.   As noted in the previous response,  EPA calculated the
 incremental  emission reduction for each component that Subpart VV would
 achieve over the other LDAR programs.   In  the absence of any  LDAR
 program,  Subpart VV was estimated to reduce  fugitive  emissions by
 approximately  59 Mg/yr from the model plant  described in BID,  Vol.  I.
 The Incremental  emission reduction attributable to these standards
 decreases  to about  24 Mg/yr when compared  to the LDAR program described
 by  the commenter and to about  37 Mg/yr  when  compared  to  Texas'  or
 Louisiana's  LDAR program.
       Having identified the incremental  emission  reduction  attributable
 to  these  standards,  EPA next evaluated  the incremental costs  of  applying
 Subpart VV.  With the one  apparent exception  for  safety  valves,  the
 commenter  appropriately calculated the  incremental  emission reduction
 from quarterly to monthly monitoring.   However,  the commenter  failed  to
 do  the same  appropriate calculation  for  costs.   In  calculating their
 cost effectiveness,  the commenter  divided the  cost  of monthly monitoring
 vs. uncontrolled baseline situation  by  the incremental emission
 reductiorr  between monthly and  quarterly monitoring.  This is  incorrect.
The incremental  costs between  quarterly and monthly monitoring must be
calculated.  This result is then divided by the incremental emission
reduction  to calculate  the  incremental cost effectiveness of going from
a quarterly  LDAR program to a monthly LDAR program.
      The Agency calculated the costs of the LDAR programs described by
the commenter and as  identified in the Texas and Louisiana regulations.
                                  4-19

-------
(Note:  All costs are in 1980$).  These costs are compared with LDAR
program costs associated with the proposed standards in Table 4-4.  As
seen in Table 4-4, the cost of Subpart VV's LDAR program is more
expensive ($12,200 per year) than either the commenter's LDAR program
($4,000 per year) or the Texas/Louisiana LDAR program ($6,600) (and has
a higher average cost effectiveness than the commenter's LDAR program,
but a lower average cost effectiveness than the Texas/Louisiana LDAR
program).  The incremental  cost attributable to Subpart VV when the
commenter's LDAR Program is in place is about $8,200/yr and about
$5,600/yr where the Texas/Louisiana LDAR program is in place.
      Using the incremental costs and incremental  emission reductions,
the Agency calculated an incremental cost effectiveness of about $240/Mg
for applying Subpart VV where the commenter's LDAR program is in place
and about $150/mg where the Texas/Louisiana LDAR program is in place.
In both cases, the incremental cost effectiveness  is still  reasonable in
the Agency's view.
      The following paragraphs describe in detail  the before "VOC
Recovery Credit" cost differences between Subpart  VV, the commenter's
LDAR program,  and the Texas/Louisiana LDAR program.   These differences
are summarized in Table 4-5.

      1.  Gas  and light-liquid valves.
               The installed capital cost ($2,300)  represents the
          initial  leak repair.  Since all  three LDAR programs require
          initial  leak repair, this cost ($400/yr  on an annualized
          basis) is incurred by each program and the cost is the same.
          Thus,  there is no incremental  cost associated with going to
          monthly monitoring.
               The operating labor for monitoring,  leak repair,
          administration,  and  support for the monthly program was
          estimated at $14,200/yr.  The quarterly  program would  require
          less frequent monitoring and  a lower repair effort.   For
          quarterly monitoring of both  valves in gas service and in
          light-liquid service,  this cost was calculated to be
          $8,000/yr.   For quarterly monitoring of  valves in gas  service
          and  annual  monitoring of valves in light-liquid service,  this
          cost was estimated to be $6,440/yr.   The  incremental  cost of
                                  4-20

-------
                                   TABLE  4-4

                        FUGITIVE VOC  LDAR PROGRAM COSTS
Cost Item
1. Installed Capital Cost
2. Annual Costs
a. Operating labor
including adminis-
tration and support
b. Maintenance
c. Miscellaneous
d. Annual i zed Capital
Cost
Total
3. VOC Recovery Credit3
4. Net Annual ized Cost
5. VOC Emission Reduction,
Mg/yr
6. Average Cost
Effectiveness, $/Mg
7. Incremental Cost
C ****&> ,..__ 	 ** / K
BID, Vol. I
$66,800

$20,800
$5,700
$2,600
$14,100
$43,200
($31,000)
$12,200
58.7
208
-
EPA Analysis
Commenter's
LDAR Proaram
$12,900

$14,500
$3,000
$400
$4,400
$22,300
($18,300)
$ 4,000
24.1
166
237
Texas/Louisiana
LDAR Proqram
$13,600

$10,500
$3,000
$400
$4,100
$18,000
($11,300)
$6,700
21.4
313
147
(NOTE:   All  costs are in 1980 $.)

   Emission  reductions from Table 4-3 times $528/Mg of VOC reduced.

                                                          Net
     Vol. 1 Net Annual i zed Cost - Subject LDAR Program Net Annual i zed Cost.
(BID Vol. 1 Emission Reduction - Subject LDAR Program Emission Reduction)

For Commenter's LDAR Program:  ($12,200 - $4,000)/(58.69 - 24.1)

For Texas/Louisiana LDAR Program:  ($12,200 - $6,600)/58.69 - 21.4)
                                        4-21

-------
                    TABLE 4-5



COMPARISON OF LDAR PROGRAMS BY EQUIPMENT COMPONENT
                               LDAR PROGRAM
Cost Item
Valves
Initial Leak Repair
- Initial cost, $
- Annual i zed cost
Operating Labor
- Labor- hours/yr
- $ (at $18/hr)
Admi ni strati on/Support
- $ at 40% of
operating labor
Subtotal
Liaht-liauid oumo seals
Installed Capital Cost
Annual ized Capital Cost
Seal replacement cost
Operating Labor
- Labor-hours/yr
- $ (at $18/hr)
Administration Support
- $ at 40% of
operating labor
Subtotal
(Subpart VV


$2,300
$400/yr

562.7
$10,100/yr .

$4,040/yr
$14,500/yr

$1,400
$230/yr
$l,650/yr

260
$4,680/yr

$l,870/yr
$8,400/yr
Commenter's
LDAR Program


$2,300
$400/yr

317
$5,700/yr

$2,280/yr
$8,400/yr

$1,400
$230/yr
$l,650/yr

260
$4,680/yr

$l,870/yr
$8,400/yr
Texas/Louisiana
LDAR Program


$2,300
$400/yr

255.2
$4,600/yr

$l,840/yr
$6,840/yr

$1,400
$190/yr
$l,330/yr

71
$l,300/yr

$510/yr
$3,300
                       4-22

-------
                                TABLE 4-5 --  continued

                  COMPARISON OF LDAR PROGRAMS BY  EQUIPMENT COMPONENT
 Cost  Item
ISubpart  VV
                                                 LDAR PROGRAM
Commenter's
LDAR Program
                                                                   Texas/Louisiana
                                                                   LDAR  Program
 Gas  Safety/Relief  Valves

   Installed  Capital  Cost

   Annualized Capital Cost

   Maintenance  (at 5% of
     capital  cost)

   Miscellaneous (at  4% of
     capital  cost)

   Operating  Labor

      - Labor-hours/yr
      - $ (at $I8/hr)

   Administration/Support
      (at 40% of operating
      1abor)

      Subtotal
  $33,730

  $6,550/yr

  $l,690/yr



  $l,350/yr
 $9,590/yr
                                        44.8
                                      $806/yr

                                      $323/yr
                                                                       $l,130/yr
Compressor seals

   Installed capital  cost

   Annualized capital  cost

   Maintenance (at 5% of
      capital cost)

   Miscellaneous (at  4% of
      capital cost)

   Operating Labor

      -  Labor-hours/yr
      -  $  (at il8/hr)

   Admi n i strat i on/Support
      (at  40% of operating
      labor)

   Subtotal
  $6,400

 $l,040/yr


   $320/yr


   $260/yr
$l,620/yr
                          $720

                         $120/yr
                                            42.7
                                         $770/yr

                                         $310/yr
                                                                      $l,200/yr
                                       4-23

-------
              TABLE 4-5 -- continued
COMPARISON OF LDAR PROGRAMS BY EQUIPMENT COMPONENT

Cost Item
Sampling Connections
Initial Capital Cost
Annual i zed Capital Cost
Maintenance (at 5% of
capital cost)
Miscellaneous (at 4% of
capital cost)
Subtotal
Leak Monitoring Equipment
Capital Cost
Annual ized Capital Cost
Maintenance
Miscellaneous
Subtotal
TOTALS
Capital Costs/Initial
Costs
Annual ized Capital Cost
(Control equipment,
initial leak repair)
Annual operating
(operating labor, admin-
istration and support,
maintenance, monitoring)
GRAND TOTAL

(Subpart VV

$13,800
$2,200/yr
$690/yr
$550/yr
$3,490/yr

$9,200
$2,100/yr
$3,000/yr
$400/yr
$5,500/yr

$66,800
$14,200/yr
$29,000/yr
$43,200/yr
LDAR PROGRAM
Commenter's
LDAR Proqram

-
-
-
-


$9,200
$2,100/yr
$3,000/yr
$400/yr
$5,500/yr

$12,900
$4,380/yr
$17,930/yr
$22,300/yr

Texas/Louisiana
LDAR Proqram

-
-
_
_
-

$9,200
$2,100/yr
$3,000/yr
$400/yr
$5,500/yr

$13,600
$4,140/yr
$13,860/yr
$18,000/yr
                       4-24

-------
     the monthly program is thus estimated to be $6,200/yr for the
     commenter's LDAR program and $7,800/yr for the Texas/Louisiana
     LDAR program.  Thus, for gas and light-liquid valves, the
     annualized costs from the BID Vol. I estimates decrease by
     $6,600/yr ($400 for annualized capital costs and $6,200 for
     annual operating labor) for the commenter's LDAR program and
     by $8,200/yr for the Texas/Louisiana program.

 2.  Light-Liquid pump seals.
          According to the commenter's LDAR program,  pump seals are
     already monitored monthly.  Since there is no  difference,  the
     companies are already incurring the full  cost  of this portion
     of the standards ($8,400/yr),  and no incremental  cost is
     attributable to these standards.   (The breakdown is $1,400 for
     installed capital  cost;  $1,800  for annualized  capital cost;
     and $6,600 for operating  labor,  administration,  and support).
          The LDAR program described in Texas'  and  Louisiana's
     regulations  require  annual monitoring  (rather  than  monthly
     monitoring).   A total  annualized  cost  of  about $3,300 was
     estimated for annual  monitoring  ($1,400 for  installed capital
     cost;  $1,500  for annualized  capital  costs; and $1,800 for
     operating labor,  administration,  and support).   The
     incremental cost  of  the monthly program is about  $5,100/yr.

     3.   Gas  safety/relief valves.
         Under the  commenter's LDAR program, these valves  are not
     controlled.   In this case, the $9,600 annual costs  of  control
     are  attributable to these standards.  The Texas/Louisiana LDAR
     programs  do control these valves, requiring quarterly
     monitoring.  The cost of quarterly monitoring is about $1,130
     per year  (all operating labor, administration,  and  support).
     Thus, the incremental cost of the standards is  about
     $8,500/yr.

4.  Compressor seals.
         The commenter's  LDAR program does  not control these
    sources of fugitive emissions.   Thus, the  $1,600  annual costs
                            4-25

-------
           are  attributable to the standards.  The Texas/Louisiana  LDAR
           programs do control compressor seals, requiring quarterly
           monitoring.  The cost of quarterly monitoring  is  about
           $l,200/yr  ($120 annualized capital costs, $1,080  for
           operating, administration and support). Thus,  the  incremental
           cost of the standards is about $400/yr.

      5.   Sampling connections.
               Since the other LDAR programs do not address  these
           fugitive emission sources, no costs are attributed to their
           programs, and the entire $3,500 annual costs are  attributed to
           these standards.

      6.   Leak monitoring equipment.
               Since companies are currently subject to  a LDAR program,
           they will have already purchased leak monitoring equipment.
           Thus, these programs incur the associated costs ($2,100/yr for
           annualized capital  cost; $3,000/yr for maintenance; $400/yr
           for miscellaneous;  and a capital  cost of $9,200).   The cost of
           this equipment can not be attributed to these  standards on an
           incremental basis.
Comment:
      One  commenter (IV-D-8)  requested that the monitoring frequency be
reduced for already controlled polymer plants in order to make the
regulation more cost effective.   This change, the commenter said,  should
specifically address those plants already covered by a state program or
that use newer technologies.
Response:
      The Agency does not disagree that reduced monitoring frequencies
have, in general,  lower costs.  However,  as seen in the response to the
previous comments, lower monitoring frequencies do not necessarily have
more favorable cost effectiveness levels.   In addition, as seen in the
response above, monthly monitoring programs for each type of emission
source are incrementally cost-effective even where existing  LDAR
programs are in-place.   Furthermore with  regards to valves in
particular, companies that use newer valve  designs (and thus presumably
have very  low leak frequencies)  can perform annual  monitoring for valves
                                  4-26

-------
 if they chose to comply with Section 60.483-1.  For the reasons stated
 earlier, the Agency believes the skip period provisions for valves are a
 reasonable way to allow companies the opportunity to extend their
 monitoring frequency if they do not wish to comply with Section 60.483-
 1.  Finally, Section 60.484-8 provides owners and operators with the
 opportunity to demonstrate equivalent means of emission limitation for
 demonstrating compliance.  For these reasons, the Agency has retained
the monitoring frequencies as currently stated in Subpart VV.

 4.4  APPLICABILITY OF LEAK DEFINITION

       In Section 60.482-2(b) of Subpart VV, there are two definitions of
 when a leak is detected:  (1) "If an instrument reading of 10,000 ppm or
 greater is measured" and  (2) "If there-are indications of liquids
 dripping from the pump seal."  Since proposal,  the Agency has  learned of
 certain polymer pumps are designed  to purge polymer fluid from bleed
 ports, thereby allowing small quantities  of VOC  emissions to escape to
 the atmosphere (see Docket'Item IV-B-25).   These pumps must use the
 polymer fluid to provide  lubrication and/or cooling of the pump shaft.
 While the  Agency believes that "indications of liquids dripping from the
 pump seal" should not be  applied to such  pumps,  the 10,000 ppm or
 greater definition should be retained to  ensure  these  pumps are not
 emitting significant quantities  of  VOC.   Further,  the  Agency does  not
 believe that  this exemption  should  be applied  to  "new"  or  replaced
 pumps,  because there are  pumps  available  that do  not have  this  designed-
 in purge.   Therefore,  the final  rule exempts  purging from  bleed ports  in
 existing pumps  that  must  have such  ports  from the  "indications  of  liquid
 dripping"  definition until the pump  is replaced  or  reconstructed [see
 Section  60.562-2(a)(l)].
                                  4-27

-------

-------
                      5.0  MODIFICATION/RECONSTRUCTION

 Comment:
       Two commenters  (IV-D-8,  IV-D-10) stated that, in determining
 whether a process change is considered a "modification," the emission
 rate should be based on the mass of emission per mass of product.  One
 commenter (IV-D-10) noted that this definition should be consistently
 maintained throughout the standards and the preamble.   The other
 commenter (IV-D-8) noted that, in the Federal Register notice on page
 36700, the definition of modification is "...  change  to an existing
 facility that increases the emission rate of any pollutant to which a
 standard applies."  This commenter pointed out that this wording is not
 consistent with the definition of modification contained in 40 CFR Part
 60, Subpart A - General  Provisions,  whereby the rate is not included.
 The commenter noted that the use of emission rate throughout the
 Preamble and Subpart ODD provides a definition of mass of emission/mass
 of product for emission  rate rather than  mass of emission/unit  time.
 According  to the  commenter,  this distinction is  important in determining
 an "affected facility"  for  process  changes  that  may increase a  unit's
 production capacity without  increasing  the  emission rates on a  mass/mass
 basis.   To ensure clarity,  the commenter  suggested  that  the language  in
 the preamble and  in  Subpart  ODD (Section  60.561  - Definitions")  be
 worded to  include this view  of emissions  with  the focus  on  rate  from a
 mass of  VOC/mass  of  product  produced basis.
 Response:
      Commenter IV-D-8 is only partially  correct concerning  the
 inclusion  or exclusion of the  term "rate" in the General  Provisions for
 determining when  a modification has occurred.  "Rate"  is  not included
 under Section 60.2, Definitions, where modification  is defined,  in part,
 as  "any physical  change in, or change in method of operation of, an
 existing facility which increases the amount of any air pollutant
 emitted into the  atmosphere ... ."  However, under Section 60.14,
Modification, increases in the amount of any air pollutant is further
clarified by stating, in part  "... any physical or operational  change to
an existing facility which results in an increase in the emission rate
to the atmosphere  of any pollutant ...." (Emphasis added.)  Emission
                                  5-1

-------
 rate  is  then  defined  in  Section  60.14(b)  as  being expressed  as  kilograms
 per hour (kg/hr).   Thus,  it  is the  intent of the General  Provisions  to
 have  those  situations where  the  absolute  levels of a pollutant  (as
 expressed in  terms  of kg/hr)  increase become subject to the  applicable
 NSPS.
      The commenters  seem to  presume that the emission levels used for
 determining when a  modification  has occurred and the emission limits
 associated with a standard must  be expressed in the same  format  (i.e.,
 mass  per unit of time or mass per mass of production).  This is
 incorrect.  The formats used  are based on different objectives.  Under
 the Clean Air Act,  changes in the absolute level of emissions to the
 atmosphere are an important factor for determining whether a NSPS should
 be applied to an existing source.  The intent is to prevent  increases in
 the emissions from  existing sources that are increasing emissions.
 Thus, a  format that identifies the absolute  amount of emissions being
 emitted, such as kg/hr, achieves this objective.  On the other hand,  a
 format that uses kg of VOC per kg of product, as suggested by the
 commenters for this standard, would not achieve this objective,  because
 it would allow existing sources to increase  emissions to the atmosphere
 while keeping its mass emissions per mass production rate constant.
 (Note that under either format, an increase  in  production rate of an
 existing facility is not a modification if that increase is accomplished
 without  a capital  expenditure on that facility  [Section 60.14(e)(2)]).
      The formats associated with performance standards,  on the  other
 hand,  are selected to reflect the performance capabilities of the
 applicable control  technologies.   As discussed  in  the preamble to these
 proposed standards,  several  formats were considered.   The Agency
 determined that,  in  some instances,  a percent emission reduction format
 and, in other instances, a mass VOC per mass production rate format  best
 reflected the performance capabilities  of the applicable  control
 technologies.  In addition,  the uncontrolled threshold emission  levels
were also expressed  in terms  of mass VOC per mass  production.  Since
 then,  the Agency has revised  the threshold levels  associated with the
 polyolefins  to reflect mass  emissions per unit  of  time emission  rate.
      In summary,  the emission level  or limit formats associated with
determination of modifications and standards of performance do  not need
                                  5-2

-------
 to be the same.  The format suggested by the commenter for emissions for
 determining whether a modification has occurred could result in less
 emission reduction than provided by the General Provision's definition
 and it is not the intent of these standards to do so.  For these
 reasons, the Agency has rejected the commenter's suggestion to redefine
 modification in these standards.
 Comment:
       One commenter (IV-D-47)  referred to 54 FR 893  in which is stated:
 "Under the new approach, any existing process section that is modified
 or reconstructed becomes an affected facility subject to  the proposed
 standards.  Similarly,  any newly constructed process  section at an
 existing plant or a new plant  would be an affected  facility...".   The
 commenter then stated that the impact of these statements  is unclear
 where an existing process section that is modified  or reconstructed
 becomes an affected facility subject to the proposed  standards,  yet  that
 modification or reconstruction results in increased emissions only in
 another section.   The commenter asked how the standards are to  be
 applied in this situation.
 Response:
       Modification or reconstruction to a process section  only  affects
 that  process section  regardless of  the effect on emissions  in other
 process sections.   A  process section is  "modified" if a physical change
 occurs to  the  facility  or there is  an  operational change to  the  facility
 either of  which results  in  an  increase in  the emission rate.  In the
 example provided  by the  commenter,  the modified process section would
 not become an  affected  facility because  there  is no increase  in
 emissions  from  that process  section.   Assuming  the other process section
 is not  modified,  as defined, then it  is  not  considered to be  an affected
 facility even though  there  is  an  increase  in  emissions.
      A process section  is  "reconstructed"  if  the replacement of
 components  in the  existing  facility  occurs  so  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 process
 section and  it  is technology and economically  feasible to meet the
 applicable standards.  The definition of reconstruction does not depend
on an increase  in emissions.  Thus,   in the example provided by the
                                   5-3

-------
commenter, if the first process section is reconstructed, then it
becomes an affected facility subject to the standards regardless of the
fact there has been no increase in emissions.   (Note:  This is also true
even if a decrease in emissions occurs.)  As before, if the process
section in which emissions do increase does not undergo replacement of
components so as to constitute a reconstruction,  then that process
section is not an affected facility and is not s;ubject to the standards.
                                  5-4

-------
                        6.0  MONITORING  REQUIREMENTS

 6.1  GENERAL

 Comment:
       One commenter  (IV-0-8) stated that the Agency is requiring very
 specific types of monitoring devices in the proposed standards and
 certain types of devices (for example, concentration monitors for carbon
 absorbers) that do not have proven performance standards and their
 output is not used for compliance monitoring.  The commenter believes
 that the choice of not only what and how to monitor but where to monitor
 in the process should be left up to the owner/operator and can result in
 a control strategy that anticipates and prevents mistakes,  rather than
 just providing for alarms or after-the-fact information gathering.   The
 commenter further believes  that the selection of monitoring devices is
 better left to the discretion of the manufacturers.   In cases where
 performance proven devices  are required for compliance, the commenter
 recommended reworking the standard to  encompass a specific  device of
 other proven or provable technology.
 Response:
       The General  Provisions  [40 CFR 60.13(i)]  state that the
 Administrator may  approve the  use of "alternatives to  any monitoring
 procedures  or requirements  in  this  part...  ."   This is  applicable to all
 NSPS  and  the citing of specific  monitoring  requirements in  this subpart
 does  not  preclude  an  owner  or operator  from seeking approval  of
 alternative  monitoring procedures or requirements.
      The commenter makes several statements that need  to be  addressed.
 First.  The  commenter refers to  concentration monitors  for  carbon
 adsorbers as not having proven performance  standards and that their
 output is not  used for compliance monitoring.  The commenter  is correct
 on both points, but misses the intent of monitoring requirements in  this
 Instance.  If  the purpose of the concentration monitors was to indicate
 the compliance status  of the affected facility, the Agency agrees that
 it would be an inappropriate compliance monitoring technique because it
 lacks proven performance standards.  However, the intent of the
monitoring requirement is to provide an indication that an affected
                                  6-1

-------
facility mav be out of compliance and not that it is out of compliance.
Historical monitoring data can be used to identify control devices that
may not be operating properly and this information may be used by EPA,
State, or local enforcement officials to decide when a performance test
is warranted.  Given the intent of the monitoring requirement, the fact
that the output is not used for compliance monitoring is irrelevant.
Concentration monitors and their output are sufficient in quality and
type to provide an indication that an affected facility may be out of
compliance.
      Second.  The commenter stated that they believe the choice of
what, how, and where to monitor should be left up to the owner or
operator and that the selection of monitoring devices is better left to
the discretion of the manufacturers.  The Agency has considerable
experience with the control devices that can be used for complying with
these and other standards, and with various monitoring devices that can
be used to monitor the performance of the control device.  Thus, the
Agency believes it is fully capable of identifying appropriate
monitoring devices and locations.  Nevertheless,  new technologies come
along or unique situations occur that may not be covered by a rule.  In
such situations, the owner or operator may request (as noted earlier)
for approval of alternative monitoring procedures or requirements.
      Third.  The commenter suggests that monitoring procedures or
requirements can be implemented by the owner or operator that anticipate
and prevent mistakes, rather than provide for "alarms or after the fact
information gathering."  The Agency encourages all control  schemes,
including monitoring procedures, that help prevent process equipment and
control devices from malfunctioning.  However, while such control
schemes may reduce the times when a facility is out of compliance, they
may not necessarily guarantee compliance at all times.  As stated
earlier, the purpose of the monitoring requirements is to provide an
indication that a facility may be out of compliance.  This purpose may
not be accomplished by schemes tha-t "anticipate and prevent mistakes."
                                  6-2

-------
 6.2  NEED FOR FLOW MONITORS

 Comment:
       Four commenters  (IV-D-5,  IV-D-6,  IV-D-7,  IV-D-8) questioned the
 need and desirability  of requiring flow measuring devices on vents to
 control equipment.  According to the commenters, this will add
 significant cost, provide no useful data, and add to the burden of
 recordkeeping.  One commenter (IV-D-6) specifically wondered how the
 flow monitoring data would affect the control of emissions.  This
 commenter also stated  that the flare requirement portion of the fugitive
 emission standard (40  CFR Part 60 - Subpart VV) does not require flow
 monitoring devices.
       Two commenters (IV-D-7, IV-D-8) maintained that flow measurement
 on intermittent vents'will  be especially valueless since it will be
 difficult to differentiate between true no-flow situations and
 instrument problems.   Two of the commenters (IV-D-7,  IV-D-8)  stated that
 installation of flow devices in a process flare system that also serves
 as a safety  flare is  not a good practice from a maintenance or safety
 standpoint.   One commenter (IV-D-8)  stated further that  the temperature,
 density,  pressure,  and fouling  or corrosive characteristics of flared
 gases tend to cause maintenance and  reliability problems  on flow
 measuring elements  that are  placed  into the flare  gas  line.   Commenter
 IV-D-8  noted  that while external  measurements  devices  are available,
 they are  expensive  and certainly should not  be required on  each
 individual vent  stream into  the  flare gas  header.  This commenter  also
 claimed that  measuring every  flare gas  vent  stream flow is  not possible,
 with  even  single  stream flare gas flow  measurements being notoriously
 difficult  to  measure.
      One  commenter (IV-D-6)  stated that technical problems exist  for
 retrofit of modified facilities where emergency  vent systems are
 integrated with normal  process vent streams, because the  range of  flow
 rates makes accurate measurement of lower  flows  impossible without
causing excessive restrictions to emergency ventings.
      Commenter (IV-D-8) suggested that Sections 60.563(a)(2), (b)(2),
and (c)(2) of the September 30, 1987, Federal Register notice, which
require the installation of a flow indicator to provide a record of the
                                   6-3

-------
vent stream flow to the incinerator or flare, be deleted.  All  four
commenters felt that engineering estimates and design calculations of
the vent flows should be adequate to ascertain compliance with  flare
flow allowable ranges.
      One commenter (IV-D-5) recommended that a requirement for an
engineering piping report be substituted for the flow instrument
requirements for flares.  This commenter believes that an engineering
report describing the piping arrangement for the vent streams would
provide assurance that these streams will be continuously flared.  Such
a report, the commenter said, would achieve the same objective  as the
flow instrument requirement by showing that the vent streams are "hard
wired" (no physical possibility of an atmospheric release prior to the
flare) without the burden of installing, operating, and maintaining a
large number of flow recorders.
Response:
      The EPA considers it very important to ensure that vent streams
are continuously vented to the flare (or other control  device).  The
primary intent of the flow monitoring requirement was to provide a means
for Indicating when vent streams were bypassing the flare or other
control device.  In the September 30,  1987,  Federal Register notice,
flow indicators were proposed.   Flow indicators envisioned by the Agency
would simply provide an indication of flow/no flow, and need not provide
quantitative estimates of flow rates.   The Agency has reevaluated the
use of flow indicators as proposed and in light of the  comments
received.  This revaluation has led the Agency to the  following
conclusions.
      1.   Flow meters,  which provide quantitative estimates  of flow
          rates,  could be one way to ensure  emissions are vented to a
          control  device.   However,  as pointed  out by the commenters,
          there may be technical  problems and less expensive ways to
          achieve  the same goal.
      2.   Flow i-ndicators  located on the vent pipe between the emission
          source  and the control  device by themselves may be
          insufficient  to-meet  the intent (even though  this  was what  was
          proposed).
      3.   Engineering  reports that show an emission stream is  "hard
          piped"  to a  control  device is a less  expensive  method than
          flow meters  to ensure the  entire flow will  be  vented to the
          control  device.   Other piping arrangements  can  be  used,  but

                                  6-4

-------
           car seals on valves or flow indicators located immediately
           downstream of each valve that could divert a portion of the
           flow to the atmosphere,  either directly or indirectly, become
           necessary.

       Considering the above conclusions, the Agency is now requiring an
 engineering report that describes  the piping arrangement for venting the
 affected emission streams to the control device.  If any valves are
 present in the line between the source and  the control  device,  the rule
 requires them to be car-sealed opened.   In  addition,  all  valves that
 allow emissions to bypass the control  device are required to be car-
 sealed closed.   The monitoring requirements have been revised now that
 this  engineering report is  required.   An owner or operator may  elect to
 follow one of two methods for monitoring the vent system.   One  method
 would require monthly inspection of  the valves to inspect the car seals,
 the reporting and recording of any time the car seals are broken,  and
 the recording and reporting of any time the valve position  has  changed.
 The other method would  require the installation of a  flow indicator,
 which gives an  indication of flow/no  flow,  at  the closest downstream
 point of each valve that  is required  to  be  car-sealed closed.   The  owner
 or operator is  to record  all  periods  of flow (which  indicates a  portion
 of the emission  stream  is bypassing the  control  device)  and  report  such
 periods  of flow.
 Comment:
       One  commenter (IV-D-6)  pointed out  inconsistencies  in  the  wording
 of the proposed  standard  as  stated in Section  60.563.  The commenter
 pointed  out  that  (1)  the  preamble  states that  if  a flare  is  used to
 comply with  the  proposed  standards, a flow  indicator  must be  installed
 and operated  to  provide a record of vent stream flow  to the  flare for
 each  vent  stream  (52  FR 36701);  (2) the regulation (Section 60.563(b)(2)
 of the September  30,  1987,  Federal  Register notice) states that a flow
 indicator  be provided to  record vent  stream flow to the flare at least
 once every hour  for each  affected facility;  and (3) the regulation
 (Section 60.563(c)(2) of  the September 30, 1987, Federal Register
notice) states that a flow  indicator  be provided to record vent stream .
flow to the flare for each affected facility.  This commenter felt that
                                  6-5

-------
 these  inconsistencies  in wordings would cause considerable  ambiguity  and
 inconsistency  in  application of the regulations.
       Another  commenter  (IV-D-5) stated that the actual  requirement for
 the  installation  of  an indicator of vent stream flow to  continuously
 monitor  flow rates to  a flare contradicts the preamble which  says  the
 installation of an indicator of vent stream flow is not  to  continuously
 monitor  exit gas  velocity, but rather to ensure that the vent  streams
 are  continuously  flared.
 Response:
       The commenters are referring to Sections 60.563(b)(2) and  (c)(2)
 of the proposed rule concerning monitoring requirements  when  a flare  is
 used for the installation of a "flow indicator that provides  a record of
 vent stream flow  to the flare for each affected facility."  As noted  in
 the  previous response, the record of vent stream flow to the  flare (or
 other  control  device)  is no longer required.  Although the  placement  of
 the  flow indicator has been changed, the record of vent  stream flow need
 only be  a yes/no  type of indication and not a measurement of velocity or
 other  quantitative measurement.
 Comment:
       One commenter (IV-D-6) stated that the proposed vent  stream
 monitoring requirement under Section 60.563 specifying flow monitoring
 on each  vent stream at a point closest to the inlet of the  control
 device and before being joined with any other vent stream is ambiguous
 in that  it is  unclear if multiple vent streams from a process section
 listed as an affected facility must each be monitored or collected and
 monitored, or  if  all  vent streams from all  affected facilities be
 collected separately for a single point monitoring.
 Response:
      This comment is no longer relevant as the  location of the flow
 indicators has been changed.   To ensure that vent  streams do not bypass
 the flare, the EPA now requires an  engineering report depicting the
 piping used to vent each affected emission  stream  to the flare or other
 control device.  This report will  show,  in  part,  all individual vents
 and where they are joined.   Where flow indicators  are used for
monitoring,  they are  now placed in  a different location to identify
 periods when there has been flow away  from  the control  device.
                                  6-6

-------
 Comment:
       One commenter (IV-D-49)  stated that flow monitoring requirements
 need greater flexibility due to the lack of instrumentation that is
 accurate  over a wide range of flowrates.   This commenter suggested that,
 rather than  hourly recordings  at individual  points,  the  Agency should
 allow some choice  in the placement  and  accuracy of  the flowmeters.
 According to the commenter,  this would  still  provide EPA with  the
 information  needed for enforcement  and  environmental  protection,  yet
 give the  affected  plants a more practical,  less-costly means of ensuring
 compliance.
 Response;
       As  noted  above,  engineering reports  now replace flow  indicators as
 the  primary  means  for  ensuring  emissions  are  vented  to the  control
 devices.   Where flow indicators are  used  for  monitoring  the vent  system,
 they are  now located downstream of  valves  that  could allow  emissions  to
 bypass  the control  device.   These flow  indicators need only show  a
 yes/no  type  of  indication,  and  not  a quantitative measurement  of  flow.
 Whether in this  new location a  wide  range  of  flow rates  would  be
 observed,  "the  lack of instrumentation  that is  accurate  over a  wide
 range of  flowrates"  is  not  a concern.
      At  proposal,  the  Agency proposed  that hourly recordings  of  flow  to
 the  control  device  be  required.   In the final rule,  the  location  of the
 flow indicators  has  been moved  so that  the flow indicators provide an
 indication of times  when flow is  being  diverted to the atmosphere.  The
 Agency  believes  that more frequent collection of flow/no  flow data is
 appropriate when the purpose of the monitoring  is to detect flow  to the
 atmosphere than  to a control device.  Thus, the final rule requires flow
 indicators to be equipped to indicate and record whether  or not flow
exists  at  least once every fifteen minutes (rather than  once every
hour).  If an owner or operator feels that an alternative frequency or
placement of a flow  indicator is as  appropriate, then the owner or
operator can  petition the Administrator  as provided  in the General
Provisions to use an alternative monitoring scheme.
                                  6-7

-------
6.3  THERMOCOUPLES AND  FLAME MONITORING

Comment:
      Two commenters  (IV-D-6,  IV-D-8) pointed out that the preamble
clearly states that thermocouples is the only acceptable monitor, while
the standard allows for a thermocouple or similar device.  The
commenters requested  that this confusion be eliminated from the rule.
      Four commenters (IV-D-6, IV-D-7, IV-D-8, IV-D-49) requested that a
provision for any other equivalent devices capable of detecting a flame
be allowed with the regulations (Sections 60.563(b)(l) and (c)(l) of the
September 30, 1987, Federal Register notice).  One commenter  (IV-D-8)
referred to the general provisions of the National Emission Standards
for Hazardous Air Pollutants (NESHAP) rules where thermocouples or
equivalent devices are required.  This commenter suggested that visual
inspection combined with an assessment of the reliability of  the fuel
supply to the pilot be allowed as an equivalent pilot flame detection
system.  The commenter also pointed out that remote sensors of various
types are available and can be used, but should not be required as these
types of sensors have reliability and citing problems as well.
      Commenter (IV-D-6) stated that there are other techniques (e.g.,
flame ionization detectors and remote infrared scanners) that have been
proven to be just as reliable as thermocouples.   The commenters stated
that the final  rule should allow individual  plants to select alternate
flame sensors as the point of the regulation should be to require a
pilot detection system (i.e.,  thermocouples, flame ionization detectors
and remote infrared scanners)  capable of detecting a flame.  Commenter
IV-D-6 pointed out that other NSPS have already allowed these options.
      Commenter (IV-D-8) suggested the following  language to replace the
current language in Sections 60.563(b)(l) and (c)(l)  of the
September 30, 1987, Federal  Register notice:  The presence of a flare
pilot flame shall  be monitored using a thermocouple or any other
equivalent device to detect the presence of a flame.
Response;
      The preamble for the proposed  rule should  not have stated so
distinctly that thermocouples  were the only acceptable monitor.   Other
                                  6-8

-------
 similar devices are acceptable provided they, provide the necessary
 recordkeep i ng requ i rements.
       The presence of a flame is obviously critical to the operation of
 a flare as a control device.  The intent of the flare monitoring
 regulation is to require a reliable monitoring device on the flare that
 will indicate when there is no flame present and thus when the flare is
 not operating; or in the case of intermittent emissions, not in a ready
 state to control emissions.   For flares controlling continuous
 emissions, monitoring of the flare flame or pilot light flames is
 appropriate to ensure the vent stream is being destroyed.   For flares
 controlling intermittent emissions,  a flare flame will  not always be
 present.  Thus,  for these flares,  it is important to monitor the pilot
 light flames.
       Thermocouples are  generally  accepted  as  the most  reliable means to
 monitor the presence of  a pilot  flame.   For flares controlling
 intermittent emissions alone,  it  is  important  to  ensure that  the pilot
 lights  are lit (i.e., have  a flame present).   Thus,  the standards
 require such flares  to monitor the pilot light  flames using  a
 thermocouple or  equivalent monitoring device.   For flares  controlling
 both intermittent  and continuous emissions  or  continuous emissions
 alone,  the EPA has  decided that the  use  of  certain  optical devices  is
 also acceptable  to  indicate  the presence of a  flame  (either the  flare
 flame or pilot light flames).  Ultra-violet or  infrared beam  sensors  may
 be  used  in lieu  of thermocouples for these  flares.  These devices offer
 an  advantage over thermocouples because they may be installed  remote
 from the flare tip thereby allowing maintenance to be done without
 shutting down the flare.   It is important that these optical devices  be
 installed  properly to minimize the effects of solar radiance.  Although
 these devices may have difficulty in distinguishing the pilot flame from
 the main flame, the detection of a flame fulfills the intent of the
 regulation for flares used to control both continuous and intermittent
 emissions or continuous  emissions alone.
      The detection of flame presence by visual means or by remote video
camera is not a suitable  method of monitoring.   If a flame  is operating
smokelessly,  it can be difficult  to determine if a flame is present.
                                  6-9

-------
       Flame  ionization detectors are not considered as reliable  as  other
monitoring technologies.  The experience of one flare manufacturer  (see
Docket  Item  IV-0-54) showed major problems with the accumulation of
moisture on  the flame rod, which tended to ground the flame rod  and then
lock up the  system.  Further, this manufacturer found that the formation
of small amounts of carbon in the pilot flame and its accumulation
around the base of the flame rod also tended to "ground out" the flame
rod and lock up the system.
      The EPA is willing to take into consideration any operating
records or test data for alternative monitoring devices.
Comment;
      Four commenters (IV-D-6, IV-D-7,  IV-D-8, IV-D-49)) stated that
thermocouples are known to be unreliable when placed in the severe
operating environment at the top of a flare and that the flare tip
maintenance period can typically be much longer than the service period
for a thermocouple.  The commenters then asked what needs to be done
when a pilot flame thermocouple burns out.   Is the flare to be shut down
prior to the regular maintenance to replace the pilot thermocouple?  The
commenters pointed out that since flares are emergency relief devices,
taking a flare out of service can not usually be done without taking the
entire process which the flare services out of service and that more
emissions would undoubtedly result from premature flare maintenance
related to thermocouple burnout.   Commenter IV-D-49 also stated that
thermocouples can be difficult to replace.
Response:
      Recent improvements in  thermocouple  installation technology have
extended the operating life of thermocouples in flare monitoring
service.  If a thermocouple is sheathed within a thermowell,  the
thermocouple is protected from the severe  flame environment,  and the
thermocouple operating life can  be extended to approximately the same
length of time as the flare tip  maintenance period.   Installing a
thermocouple with thermowell  will  reduce significantly the number of
times an operator must decide whether to shut  down upon  thermocouple
failure.
      Any breakdown or malfunction of the  thermocouple should  be
repaired as soon as practicable  as stated  in Section  61.14(b)  of the
                                  6-10

-------
 General Provisions.  The operator is expected to determine the best time
 to shut the flare down after considering how to minimize emissions both
 for safety and environmental reasons.

 6.4  INCINERATORS

 Comment:
       One commenter (IV-D-8) stated that the regulation should not
 attempt to limit the option of using any proven or provable technique
 for monitoring emissions from incinerators.
 Response:
       As  noted earlier,  the General  Provisions  [40 CFR 60.13(1)]  state
 that the  Administrator may approve the  use  of "alternatives to any
 monitoring procedures  or requirements in this part...".   This is
 applicable to  all  NSPS and the citing of specific  monitoring
 requirements  in this subpart does  not preclude  an  owner  or  operator from
 seeking approval  of alternative monitoring  procedures  or requirements.
 Comment:
       One  commenter (IV-D-7)  believes that  the  requirement  for measuring
 Incinerator temperature  is  sufficient to ensure  that the incinerator  is
 operating  as required, and  that material  balance and engineering
 calculations can  be  used  to  determine if residence  time  is  sufficient.
 Response:
      The  commenter  appears  to  be  referring to the  monitoring
 requirements from  incinerators  rather than suggesting  an alternative to
 the performance test.  The monitoring requirements  proposed for
 incinerators were  a  temperature monitoring device and  a  flow  indicator
 In each vent stream  to the incinerator.   The Agency has determined that
 firebox temperature  alone provides an adequate indication of  incinerator
 performance.  The  final rule now requires that only the temperature of
 the incinerator be monitored.  An engineering report that describes the
piping arrangement for venting the vent  stream to the incinerator has
replaced the placement  of flow indicators within the piping that vents
the vent stream to the  incinerator.
                                  6-11

-------
6.5  CARBON ADSORBERS

Comment:
      One commenter  (IV-D-8) concurs with the concept of leaving
flexibility in the standards for case-by-case selection of alternative
for adsorbers.  However, the commenter believes that since the required
organic concentration monitors with recording devices on carbon bed
adsorbers are not intended for compliance determinations, but rather as
status indicators, it makes more sense to leave the type of
concentration monitor to the owner/operator of the process.  The
commenter listed such alternatives as in-line gas chromatography, bagged
grab samples, or time or flow proportional sampling as all providing
equivalent information of status of a carbon bed.  The commenter stated
that these techniques can also be used for developing the required
maintenance history.  The commenter also suggested along the same lines
that the record of testing and monitoring could be a recorder, lab
analysis record, etc.  The commenter proposed the following language for
Sections 60.563(e)(l)(11), (e)(2)(ii), and (e)(3)(i):  "An organic
monitoring device used to indicate the concentration level of organic
compounds based on a detection principle such as infrared,
photoionization, thermal conductivity, or other equally effective device
or monitoring scheme."
Response:
      The Agency has revised the wording of the paragraphs referred to
by the commenter so that detection principles such as infrared,
photoionization, and thermal  conductivity are identified specifically.
These paragraphs have been reduced in the final  rule to a single
paragraph -- 60.563(a)(4).  The Agency agrees that these types of
monitors provide equivalent information on the status of the carbon bed.
Each of the monitors would be still  required to be equipped with a
continuous recorder.  The Agency has not included "or other equally
effective device or monitoring scheme" in Section 60.563(a}(4) because
the General  Provisions allow for such to be used upon approval by the
Administrator and it thus is unnecessary to include such language in
this subpart.   Should they be demonstrated to be equally effective,  the
                                  6-12

-------
 use of bagged grab samples or time or flow proportional  sampling could
 be used if shown to be equivalent as provided in the General  Provisions.

 6.6  COSTS

 Comment:
       Two commenters (IV-D-6,  IV-D-8)  expressed concerns  over flow
 monitoring costs.   Commenter IV-D-6 stated that the  cost  justification
 of providing vent  stream monitoring is unclear.  Commenter  IV-D-8 stated
 that while the cost of an individual  flow  monitor may be  small  compared
 to the overall  cost of controlling emissions,  the cost of monitoring  the
 tens of vent streams that can  feed an  incinerator is  not  insignificant.
 This commenter felt that the cost of monitoring flows on  each vent
 stream was not adequately addressed in the economic  analysis  of the
 proposed  rule because flow monitoring  is not  necessary for  efficient
 operation of an  incinerator and  flow indication is not used for
 compliance purposes.
       Commenter  IV-D-8 also believes that  the  cost of installing  flow
 monitors  on  every  vent stream  entering boilers  or process heaters  is too
 high to justify  the inclusion  of  this  requirement into the  standard and
 suggested that Section 60.563(d)(l)  of the  September  30,   1987,  Federal
 Register  be  deleted.   This  commenter concurred  with Section 60.563(d)(3)
 in  that notice in  which  the  owner/operator  can  use his or her own
 discretion on  choosing  the  most efficient monitoring methods  for boilers
 or  process heaters  with  150  million Btu/hr  or greater heat input design
 capacities.
 Response;
      Owners or operators are no longer required to monitor flow on each
 vent stream to a control device; an engineering report is being required
 instead.  Under the final rule, an owner or operator may  elect to
monitor the vent system through the use of flow indicators,  a  car-seal
monitoring program, or a combination of the two (see Section 60.563(d)
of the final rule).  If flow indicators are used, they are now to be
placed downstream of each valve that could  divert the emission flow away
from the control  device to the atmosphere,  either directly or
indirectly.  In the Agency's estimation, the cost of installing such
                                  6-13

-------
 flow  indicators  is  reasonable  in  light  of  the  information  it  provides
 (periods  of  time when  flow  is  being diverted from  the  control  device).
 An  owner  or  operator can elect to use an alternative method subject  to
 the approval of  the Administrator if it is shown to be equivalent  under
 the General  Provisions.
       In  establishing  the monitoring requirements;  for  this (or any
 other) NSPS, EPA had to reconcile the need to ensure the effective
 operation of control and product  recovery devices  and  compliance with
 the rule with the degree of complexity and economic burden of  the
 monitoring systems  on  the industry.  The Agency believes that  it has
 chosen the most  reliable and fair methods of monitoring.  Considering
 the complexity of the  processes involved and the burden on both EPA  and
 industry, the monitoring and reporting methods specified are believed to
 be  the least burdensome way of providing EPA with  reliable information
 about changes in combustion device operation or diversion of flows that
 could lead to increased VOC emissions and to ensure the proper operation
 and maintenance  of  product recovery devices.

 6.7  MODIFICATION/RECONSTRUCTION

 Comment;
      Two commenters (IV-D-44,  IV-D-50)  stated that language in the
January 10,  1989, Federal  Register notice appears to require monitoring
of  existing continuous emission streams  prior to any modifications or
 reconstructions.   The commenters believe that a monitoring requirement
would raise the following  concerns or questions:

         Determination of emissions prior  to modification by  sampling
          is not  appropriate because emissions  may  vary with  product
          runs,  and the worse case product  may  not  be  available for
          monitoring within  a reasonable time.   These  emissions can be
          calculated with  reasonable accuracy.   (IV-D-44)
         What is the economic justification/basis for requiring testing
          of existing streams as opposed to calculating?  (IV-D-44, IV-

         What test method,  duration,  frequency and monitoring are
          contemplated?  (IV-D-50)
                                  6-14

-------
          For processes that make a wide variety of products, what
           product line emissions (different hydrocarbon constituents,
           product densities, etc.)  should be measured?  (IV-D-50)
           What is the environmental benefit of requiring testing of
           existing streams?  (IV-D-50)
       The commenters pointed out that Table 4 "Procedure for Determining
 Control and Applicable Standard for Continuous Emission Streams from
 Modified or Reconstructed Polypropylene and Polyethylene Affected
 Facilities" (Ref:  54 FR 908)  specifies in  Step 3 that calculations of
 VOC concentration in the applicable weight  percent range should be made
 before and after any modification or reconstruction.   The commenters
 requested clarification of this issue (monitoring vs.  calculation)  and
 recommended engineering calculations as specified in  Table 4 of emis-
 sions  before and after any modification or  reconstruction be used in the
 procedural  step  in determining  control  requirements.   The word  "measure"
 on  Page 54 FR 895 should be changed to  "calculate," according to
 commenter IV-D-44.
 Response:
       The language in the preamble  to the January 10,  1989,  Federal
 Register notice  did  not intend  to imply that  monitoring  of existing
 continuous  emissions  was  being  required, although the  language  was  not
 as  precise  as  it  should have been.   In  that notice, the  Agency  intended
 that measurements rather  than calculations  be  used to  obtain  the  VOC
 concentrations of each  applicable VOC stream.  Measurement  of the
 applicable  stream would occur after  a modification or  reconstruction
 determination has  been  made by  the  appropriate enforcement  Agency, but
 before  any  actual  changes  have  been  undertaken.   This  clarification
 narrows  the  language  in the preamble  from "any changes to  an  existing
 process  section that  could  conceivably  be a modification or reconstruc-
 tion" to only those that  are determined to be a modification  or
 reconstruction.
      In the final rule,  the requirement to measure the VOC concentra-
 tion or the annual emission rate rather than calculate these values is
 applied to only those individual streams that an owner or operator seek
to exempt from control through either the VOC weight percent exemption
or the low annual emissions exemption (see Section 60.564(d) of the
                                  6-15

-------
 final  rule).   As  provided  in  the  General  Provisions,  if an owner or
 operator believes that  an  alternative  procedure  is  as  accurate as a
 measurement,  then the owner or  operator may  still petition the
 Administrator for approval.
       While the Agency  would  prefer  actual test  data,  the  final  rule
 allows an owner or operator to  submit  calculations  calculating the
 weight percent and annual  emissions  of each  nonexempt  vent stream in
 lieu of .actual test data,  provided such calculations can be demonstrated
 to  be  sufficiently accurate as  to preclude the necessity of a  test.
       In testing  or calculating the  weight percent  and  annual  emissions
 of  a vent stream,  an owner or operator is required  to  evaluate the
 stream under  representative conditions of normal operating conditions.
 This may require  an owner  or operator  to make assumptions  or estimations
 of  how the affected facility will be operated or how emission  streams
 will vary during  production of  various products.  The period during
 which  testing of  a stream  occurs, thus, need not be a  "worst case"
 product,  but  preferably a  representative product.  Where affected
 facilities are used to produce  a wide  variety of products,  then  an owner
 or  operator would  calculate (or measure)  the emission streams  that would
 occur  during  the  course of a year for  each of the products.  The
 resulting data would be combined to identify composite streams and their
 weighted  average  VOC concentrations and total annual emissions.   Each
 composite stream's  VOC concentration would then be used to  calculate the
 threshold emission  rate and a control/no  control  determination would be
 made by  comparing  the calculated (or measured)  annual  emissions with the
 threshold emission  rate.
       Where an owner or operator tests an  emission stream,   the final
 rule requires the  use of Test  Method 18 to determine the VOC concentra-
 tion and Test Method 2,  2A, 2C,  or 2D,  as  appropriate,  to determine the
 volumetric flow rate.  Each test shall  consist  of three 1-hour runs in
which  either an integrated sample or four  grab  samples  shall be taken.
       In determining whether a test or calculation is  to be  required,
the Agency considers a number  of factors  such as  the use of  the
 information,  the  relative cost of conducting  the  tests, and the
availability of alternative procedures. Because  the individual stream
exemptions allow  an individual  stream to be  exempt from control under
this NSPS, the Agency believes this decision  needs to  be made based on
                                  6-16

-------
test data.  For nonexempt streams, a no control decision may change to a
control decision as more facilities at a plant are constructed,
modified, or reconstructed.  Thus, while test data are preferable for
determining the VOC weight percent concentrations, the Agency has
decided that calculations showing the VOC concentrations can be an
acceptable alternative to testing, and at greatly reduced costs.
                                6-17

-------

-------
                     7.0  TEST METHODS AND PROCEDURES

  7.1    GENERAL

  Comment:
        One commenter  (IV-D-50) referred  to the  proposed  standard  of "98
  weight percent, or a concentration of 20 ppmv  expressed  as  the sum of
  the actual compounds, not carbon equivalents,  corrected  to  3  percent
  oxYqen, whichever is less stringent" and stated that the correction for
  oxygen content may be appropriate for concentrated hydrocarbon streams,
  but is inappropriate for dilute hydrocarbon streams.  According  to  the
  commenter, this type of correction is normally used to keep a source
  from compliance by means of excess air to dilute the emissions.  The
 commenter then stated that when the hydrocarbon is already  in an air
 stream, no additional combustion air is needed or used.
       To illustrate the problems created by using the oxygen correction
 factor, the commenter gave the following example:   assume that a vent
 stream, diluted with  ambient air to 1000 ppmv,  is  sent  to an
 incinerator.   No combustion  air is  required  as  this stream is 20  percent
 oxygen.  The  vent  from the incinerator  will  be  19.9 percent oxygen;
 therefore,  the  hydrocarbon content  must  be  20  ppm  x (3/19.9%)  or  3  ppm.
 This  equates  to a  required destruction  efficiency  of (1000-3/1000)  x 100
 - 99.7 percent.  The  commenter then  recommended that  the phrase
 "corrected  to 3 percent  oxygen"  be  removed from this  section for  dilute
 hydrocarbon streams because  it creates  a more stringent  and  less
 realistic destruction efficiency than the 98 percent  previously
 required.
 Response:
       The Agency agrees with the commenter that correction to  3 percent
 oxygen  is unnecessary for emission streams that already contain
 sufficient air so that additional combustion air is not needed or used,
 and could be  a potential problem for those seeking to comply with the 20
 ppmv limit.   Therefore, the Agency has revised the rule so that
 correction to 3 percent oxygen is required only when supplemental
 combustion air is used.
      [The Agency points out  that the correction to 3 percent oxygen
does not affect those  individuals seeking to comply with the 98 weight
                                  7-1

-------
percent destruction standard because the percent destruction  is  on  a
weight basis  ("Ibs  in and Ibs out).  If the percent destruction was
calculated on a volume basis, as used in the example  by  the commenter,
then the correction to 3 percent oxygen would be a problem.   However,
this is not the case.]

7.2  FLARES

Comment;
      One commenter (IV-D-8) suggested that the maximum  permitted
velocity equation  for nonassisted flares with pilots  be  revised  based  on
the test performed by EER at Irvine, CA.  The commenter  proposed that
Section 60.564(c)(5) of the proposed standards be rewritten as follows:

      The maximum  permitted velocity,  V|nax, for flares complying  with
      Section 60.562-l(a)(ii )(E)(3) shall be determined  using the
      following equation for steam-assisted flares:
               (v)  ' (Ht +  28.8)/31.7

      and the following equation for nonassisted flares
      where:  HT = net heating value of the sample
Response;
      The maximum permitted velocity equation suggested in this comment
describes a line that starts at the point 60 ft/sec, 150 Btu/scf.  This
point is below the 60 ft/sec, 180 Btu/scf required to assure 98 percent
combustion efficiency, as demonstrated in the EER studies referred to  in
Docket Items IV-A-1 through IV-A-3.  The maximum permitted velocity,
VIMX' ^or nonassisted  flares shall continue to be determined using the
maximum velocity equation as proposed (i.e., Log10  (VMX)  =  (HT  + 28.8J/31.7)
                                   7-2

-------
 7.3  BOILERS

 Comment:
       One commenter (IV-D-12) suggested that alternative methods of
 compliance for boilers or process heaters with a heat input of less than
 150 million Btu's be provided rather than requiring just compliance
 testing as implied in Section 60.564(a).   The commenter indicated that
 complete  combustion for a boiler of less  than 150 million Btu's can be
 determined if one knows the boiler temperature,  the organics in the gas
 stream, and their combustion temperature.
 Response:
       The Agency has determined that boilers and process heaters with  a
 heat input capacity of 150 million Btu's  or  greater have sufficient
 combustion temperatures,  residence time,  and mixing characteristics that
 98 percent reduction is achieved without  the need for conducting a
 compliance test.   However,  for boilers  and process  heaters  with a heat
 input capacity of less than 150 million Btu's, the  Agency has  determined
 that sufficient uncertainty exists that compliance  tests for such
 boilers and process heaters are necessary to show 98  percent
 destruction.   The commenter recommended that boiler temperature,  gas
 stream organics,  and their combustion temperature be  used to calculate
 destruction efficiency for purposes  of determining  compliance.   The
 Agency does not believe that  such  data by themselves  are sufficient as
 residence time and  mixing  (both  of which were not mentioned  by  the
 commenter)  are important contributors to destruction  efficiency.  If an
 owner or  operator wished to try  to use a method  other  than testing  to
 show compliance,  the owner  or  operator may seek  to  do  so  under  the
 General Provisions.  However,  for  purposes of this  rule,  compliance
 testing is  being  retained  for  boilers and process heaters with  heat
 input  capacities  of  less than  150 million Btu's.

 7.4  CONDENSERS

Comment:
     One commenter  (IV-D-12) pointed out that where a condenser is the
final unit  in  a system, a temperature recording device or an organic
monitoring device is required to indicate  the level  of organic
                                   7-3

-------
compounds, yet the test methods for determining compliance  (Section
60.564(d) of the September 30, 1987, Federal Register notice) only
addresses the use of a temperature recording device.  The commenter
recommended that a test method for determining compliance where an
organic monitoring device is selected under Section 60.563(e)(2) of the
September 30, 1987, Federal Register notice should be included in
Section 60.564(d).
Response:
      A test method for determining compliance where an organic
monitoring device is selected under proposed Section 60.563(e)(2)
(Section 60.563(b)(5) of the final rule) is not necessary because the
organic monitoring device is for purposes of monitoring the performance
of the condenser and not for determining compliance.  The standards for
polystyrene are either:  (1) meeting an emission limit in terms of kg
TOC/Mg product; (2) maintaining a maximum outlet gas stream temperature;
or (3) controlling emissions by 98 percent or to 20 ppmv.  If an owner
or operator elects to comply with one of the emission limits (i.e., the
TOC/Mg product, 98 percent reduction, or 20 ppmv},  then a compliance
test is required to show that the emission limits are being met by using
the procedures specified in Section 60.564, and the monitoring equipment
once compliance is demonstrated is either a temperature recording device
or an organic monitoring device.   Since these two devices are for
monitoring purposes only,  test methods need not be  developed for them.
If, on the other hand,  an  owner or operator elects  to comply with the
temperature requirement,  then the compliance test consists of showing
that the temperature is at or below the required temperature on a
continuous basis.   In this instance,  a test method  using a temperature-
monitoring device is required.
Comment:
      One commenter (IV-D-12)  stated that the test  method in Section
60.564(d) of the September 30,  1987,  Federal  Register notice should be
modified to include other  methods of showing compliance such as material
balances or emission calculation  based on exhaust rate and saturation
levels of organic compounds.
Response:
      An owner or operator electing to comply with  the emission limit
requirement for condensers may seek to use any alternative method to
                                   7.0.

-------
 demonstrate compli.ance as  provided under the General  Provisions.   For
 the  purposes of this rule,  the Agency has retained the use of a
 compliance  test (see Section  60.564(i)  of the final  rule).
 Comment:
      One  commenter (IV-D-6) stated that compliance with  the 0.0036 kg
 TOC/Mg  standard for the material  recovery section  at  polystyrene  plants
 (Section  60.562-l(b)(l)(1)  of the September  30,  1987,  Federal  Register
 notice) may in  some cases  be  impossible to measure and certify.   The
 commenter pointed  out  that  the volume of vent stream  from  the  condensers
 in the  material  recovery section  can  be very small  (near zero  in  some
 cases), making  proper  sampling very difficult if not  impossible.   The
 commenter stated that  they  have had encountered difficulty  in  getting
 monitoring  contractors  who  can accurately measure  flow and  VOC
 concentration both  of  which are necessary to  determine mass  quantity  of
 emissions)  at the  current emission standard  level  of 0.12 kg TOC/Mg
 product.  The commenter  concluded  that  it should not be the  intent of
 the standard to  impose  emission levels  so small that the mass emissions
 can not be directly measured  and quantified.
 Response:
      The Agency disagrees with the commenter's contention that emission
 streams in compliance with the emission limit for condensers is so small
 that it is impossible to measure and certify.  Method 2C is suitable  for
measuring such low flows.
                                  7-5

-------

-------
                  8.0   ECONOMIC  IMPACTS, COSTS, AND  COST
                         EFFECTIVENESS CALCULATIONS
 8.1  ECONOMIC  IMPACTS

 Comment:
       One commenter (IV-D-8) stated that "pass through of costs for add-
 on control technology, though estimated by the Agency to range from 0.13
 percent to 0.44 percent overall, can be devastating to commodity product
 groups with profitability that is already extremely small due to
 increased market competition."  This commenter offered that the "polymer
 and plastics industry is highly competitive and profit margins are
 generally minimal  for commodity type plastics."  Later,  the commenter
 said that the commodity markets of most concern are bottling and
 insulating,  and that the polymers of most concern are polystyrene and
 PET (see Docket Item IV-E-14).   The commenter also noted that foreign
 competition  and the potential  new (but  unspecified) health-related
 Federal  regulations will  contribute to  problems in these commodity
 markets.   Of the two polymers,  polystyrene  is of most concern to the
 commenter because,  it  is  claimed,  EPA has  significantly  underestimated
 the costs for refrigerated  condensers,  and  because the polystyrene
 industry  in  particular, will soon.be facing several  new  health  and
 environmental  regulations  (still  unspecified)  that could  have a
 cumulative significant  adverse  effect on competitiveness.
 Response:
      As  a result of this comment,  EPA  performed  an  extensive analysis
 of  the economic  impacts of the  new  standards  on the  polystyrene  and  PET
 industries (see  Docket  Item IV-B-19).   New  control  costs were estimated
 in  response to other comments,   and  condenser  system  costs were increased
 to  reflect the above comment that EPA had originally underestimated the
 costs for refrigerated condensers.
      With the new control costs, the maximum price  impacts range from 0
 to 0.09 percent.  With price impacts of this magnitude, the EPA believes
 that the NSPS will  not adversely affect  profit margins or delay
capacity expansions, especially considering the growing demand faced by
                               8-1

-------
polystyrene and PET producers.  PET bottling sales have been  strong,
increasing by 16 percent from 290 gigagrams (Gg) in 1986 to 336 Gg  in
1987.  The range of maximum price increases for PET is 0 to 0.06
percent, which should not have a significant impact on the PET market.
Sales of polystyrene insulating materials are not known.  However,  the
range of maximum price impacts -- 0 to 0.09 percent -- suggests that
control costs can easily be recovered through price increases.
Moreover, imports of both polystyrene and PET resins are negligible.  In
1987, for example, imports of polystyrene accounted for only  1.2 percent
of domestic consumption.
      Foreign competition is a factor in the PET-polyester fiber market.
In addition, this market has had declining sales since 1981,  and, as a
result, has had to reduce capacity.  However,  as mentioned, the maximum
PET price impact is only 0.06 percent, which is not significant.  A
price increase of this magnitude should not be hindered by foreign
competition or declining demand.  Furthermore,  it is unlikely that
capacity expansion is on the drawing board anyway because of  the recent
restructuring of the textile industry, which is the primary consumer of
polyester fibers.
      The EPA agrees that other regulations may have adverse  impacts on
profitability in the regulated industries.  For example,  regulation of
chlorofluorocarbons would negatively impact the polystyrene insulation
industry.  However, other regulations will have to be evaluated
individually for their costs and benefits.  This applies  to the current
NSPS as well.   The individual  impact of this NSPS,  as represented by a
maximum price impact of only 0.09 percent, is  minimal.
Comment;
      One commenter (IV-D-8) stated that the imposition of the standard
in its present form on existing sources (modification and
reconstruction)  is an unreasonable economic burden.   The  commenter
clarified "unreasonable economic burden" to mean that the imposition of
the standard on  modified and reconstructed sources;  would  result in
control costs  above $l,000/Mg  (see Docket Item  IV-D-35).   The commenter
pointed out that regulatory alternatives with  costs below $l,000/Mg were
selected by the  Agency,  while  those with costs  above $l,000/Mg were
rejected.
                               8-2

-------
       Commenter IV-D-10 also stated that the economic impact of
 individual  modifications requiring retrofit should be evaluated against
 the Agency's accepted guideline of $l,000/Mg of TOC controlled.
 Response:
       The Agency disagrees with the commenters that these standards
 impose unreasonable costs.  In developing an NSPS, the Agency considers
 a number of factors including but not  limited to the cost effectiveness
 of a standard.   In the past,  regulatory alternatives have been selected
 that encompass  a range of cost effectiveness values.  Typically,  these
 values fall  between $1,000 and $2,000/Mg.   The Agency has considered the
 types of controls (e.g.,  flares,  incinerators,  condensers,  distillation
 columns) that are likely to be used in  the  polymer manufacturing
 industry.   For  most of these  control techniques,  the Agency has
 concluded that  there will  be  little to  no  increase in costs due to  a
 retrofit in  comparison to  a new installation.

 8.2    COST  IMPACTS

 Comment:
       One commenter (IV-0-39)  stated that the  figures  given for
 annualized cost  benefits of the  "new approach"  versus  the model plant
 approach do  not  include  the engineering  and  technical  costs incurred for
 calculating,  estimating, and  verifying on a  periodic  basis  the  actual
 emissions, and the  figures  may  therefore be  faulty  in  stating  that  costs
 will  decrease with  the  "new approach."
 Response:
       The comparison of cost  impacts in Docket  Item  IV-B-13  involve only
 expected costs for  the control devices required to meet the  standards.
 These  are the costs that are used to determine which alternative is
 selected as the basis of a  standard and in such calculations as cost
 effectiveness.  Thus, on this basis, the costs associated with  the new
 approach were estimated to decrease.
      Costs associated with monitoring, reporting, and recordkeeping are
estimated in the Supporting Statement to SF-83.  The Agency agrees that
the new approach's monitoring requirements impose greater costs on
affected owners  and operators than those associated with the model  plant
                               8-3

-------
 approach.  The Agency has determined that the  increased  costs  are
 reasonable and are,  in a sense, the cost of providing  industry with  a
 standard that avoids very expensive control of not cost  effective
 streams that could have occurred under the model plant analysis due  to
 the changes in the industry, as pointed out by the industry.
 Comment;
      One commenter  (IV-D-43) referred to EPA's statement  "Some process
 sections were assumed to have their own control device..."  (54 FR 893),
 and stated that each process section does not necessarily  have its own
 control device and sometimes, several process trains share  a common
 control device.  The commenter then stated that as a result the economic
 and emissions impacts, as presented in the Federal Register, do not  take
 into account real-life situations.  The commenter recommended  that EPA
 re-evaluate the application of this faulty assumption and  alter
 appropriately the language of the final rule.
 Response:
      The statement referred to by the commenter refers to the
 regulatory analysis procedure used in the model plant approach  for
 determining which process emission streams would be controlled.  At  that
 time, the Agency recognized that each process section does not
 necessarily have its own control device and that several  process
 sections and lines may share a common control  device.  The assumption
 that each process section would have its own control  device was  used as
 an initial conservative approach for identifying the various cost
 impacts.  The model plant approach did examine potential  sharing of
 control devices as a secondary analysis for determining which  process
 emissions should be controlled.   Under the new approach,  the Agency
 recognized the sharing of control  devices in a much more  direct manner
 that is a more accurate representation of the manner in which  industry
 is actually likely to install control  equipment.
      The purpose of the discussion of the emission impacts and cost
 impacts sections "Impacts of New Approach" (54 FR 904)  was to  compare
 how the emission reduction costs incurred would differ if the  new
 approach was used,  rather than the model  plant approach,  to determine
which process emissions would be controlled.   These comparisons were
made by considering which emission streams would be or would not be
                               8-4

-------
 controlled by the new approach vs.  the model  plant approach.  The
 changes in emission reduction represent those that are estimated to
 incur as a result of assuming a more realistic application of control
 equipment.  The cost impacts presented refer  to the costs associated
 with controlling the additional  emissions on  an incremental  cost basis
 that reflected whether the emissions were going to a common  control
 device or a separate control  device.  Therefore,  the emission and
 economic impacts presented in the January 10,  1989,  Federal  Register do
 reflect real-life situations  and no reanalysis is necessary.
 Comment:
       One commenter (IV-D-47) stated that EPA proposed to control  in the
 January 10,  1989,  Federal  Register  notice some emissions  which  in  the
 September 30,  1987,  Federal  Register notice it did not consider cost
 effective to  control  using the same 1980-based economics.  The  commenter
 questioned how the same  1980  economics  can be  used to  support these
 widely divergent positions.
       In  a similar comment, Commenter  IV-D-50  stated that  if controls
 for  the product  finishing  and product  storage  sections at  new LDPE,  high
 pressure  process plants  were  not  cost  effective to control under the
 model  plant approach,  they are still not  cost  effective.   This  commenter
 recommended that EPA  rescind  all  control  requirements  for  process  sec-
 tions  determined not  cost  effective  to  control under the  September 30,
 1987,  proposal,  until  cost estimates and  economic  justifications are
 redone.
 Response:
       A number of  emission  streams  and  process sections were  identified
 as potentially being  controlled when the  "new" approach, which  forms  the
 basis  of the January  10, 1989, Federal  Register notice, is used to
determine which  process emissions are to be controlled, but would  be
 uncontrolled under the original approach based on  a model   plant
analysis, which was the basis of the September 30, 1987, Federal
Register.  These emission  streams and process  sections were considered
 "not cost effective" to control under the model plant approach.   These
results were presented in Docket Item IV-B-13.  The commenter questions
how these emission streams and process sections can now be cost
                               8-5

-------
 effective  to control when the  same  set of costs  in  1980$  is  used  in  the
 new  approach.
      The  basic  answer to the  commenter's concern lies  in  the
 assumptions and  procedures made  in  the model plant  analysis  and the  need
 to base control/no control decision for all affected facilities on the
 basis of a model plant's emission characteristics.  Three  types of
 "conflicting" situations arose under the approach presented  in the
 January 10, 1989, Federal Register  notice.
      First, control of low flow streams.  The model plant analysis  did
 not  consider control of individual  streams within a process  section.
 Rather, all continuous streams were grouped together and a separate
 control device was assumed to control the emissions.  If control  of  the
 entire process section was determined not to be cost effective, then no
 control of any stream was considered.  Under the new approach,
 individual streams with flows equal to or less than 8 scfm and emissions
 greater than or  equal to 1.6 Mg/yr are required to be controlled.  These
 streams are cost effective to control, but the model plant analysis did
 not  consider such streams on an individual basis.
      Second, low VOC concentration streams.  The new approach requires
 control of low VOC concentration streams if emissions are greater than
 specified  levels, which vary according to the VOC concentration.  This
 leads to control of certain product finishing and product storage
 process sections under the new approach.   The particular combination
 (i.e., process line only) considered under the model plant analysis,
 however, led to a rule that control would not be required.  This
 assumption of the model plant analysis was overly conservative in that
 it did not reflect using a single control  device for multiple process
 sections at a new plant or the modification or reconstruction of
multiple process sections.   In these instances,  control  is cost
effective  if sufficient emissions are present.   The new approach  is
simply more flexible in identifying such  situations than was the model
plant approach.
      Third,  intermittent emissions.  Two instances were projected
whereby intermittent emissions were projected to be controlled under the
new  approach but not under the model plant approach.  In both instances,
these emissions were considered part of baseline control and, as such,
                               8-6

-------
  would  be  controlled in the absence of these  standards.   Thus,  these
  emissions and  their process sections  were  identified as  affected
  facilities under the model  plant  approach.   An  uncontrolled threshold
  emission  rate  was  calculated  for  each affected  facility.   The  procedure
  used for  calculating that  rate  was based on  a projection  of future
  growth  -- process  section,  process line, and new  plant.   It turns out
  that the  procedure resulted in  an  uncontrolled  threshold  emission rate
  greater than the model  plant's  emission rate for  those emissions.   Thus,
  for purposes of  impact  analysis only,  those emissions were  shown as not
  being controlled under  the  model plant analysis in Docket  Item  IV-B-13.
  The model  plant analysis,  in fact,  showed a similar  range of cost-
  effectiveness  values as did the new approach.  The new approach,
  however,  does  not  have  a threshold  level for these emissions, because
  (as explained  in more detail elsewhere) of the difficulty and expense of
  identifying actual emission levels from intermittent emissions.
       For these intermittent emissions, the Agency recognizes that there
 is a trade-off between the two approaches.   Under the model plant
 approach,  some cost effective intermittent  emissions are  not controlled,
 because of the analytical  procedure used.   Under the generic approach,
 some small intermittent streams  may be controlled  that are truly not
 cost effective to control,  because of  the  inability to accurately
 measure intermittent emissions.   Under either approach,  individual
 facilities may spend more  or less  money or  control  than at another
 facility.   The  Agency has  selected  the new  approach as the overall
 better  way to  determine which  emissions to  control  from polypropylene
 and  polyethylene  plants.
 Comment:
      One  commenter (IV-D-43/IV-D-50)  pointed out  that the  EPA  has not
 updated  control costs  from  the September 30,  1987,  proposal  (52  FR
 36678) and that the only economic factor that was  updated was the price
 of fuel  gas  (that decreased), while  other costs  (not  updated) rose.  The
 commenter  then  stated that EPA should  reevaluate its  projected economic
 impacts.
      Another commenter  (IV-D-47) pointed out that the information used
by the Agency to provide the basis for the current proposals dates back
to the early 1980's and that the  Agency continues to use 1980-based
                               8-7

-------
economics while abandoning the model plant approach and retaining
process section rationale.
Response:
      The Agency reviewed the costs and costing procedure used in BID
Vol. I for flares, thermal incinerators, and catalytic incinerators.
The review indicated that the values were generally accurate and
consistent with other EPA projects, and the procedures for sizing and
costing control equipment were reasonable and consistent with EPA
procedures.  The Agency decided to retain costs in terms of June 1980S,
      While the Agency found the costs and costing procedures to be, in
general, accurate, reasonable, and consistent with other EPA projects,
there were a few exceptions.  As noted in Docket Item IV-B-10, BID Vol.
I procedures and costs for incinerators have been modified somewhat to
be consistent to those found in EPA's EAB Control  Cost Manual. February
1987.  This manual was released after the initial  cost work was
completed.  Items affected were natural gas requirements calculations
and costs; heat recovery efficiency; incinerator costs,  including the
use of pretreatment air filters for particulate removal; piping lengths;
and installation factors.  Additional review discovered that the
electricity price of $0.049 per kilowatt hour (kWh) in 1980 dollars was
too high for estimating the cost of electricity to the polymer industry.
The 1980 industrial sector price, in 1980 dollars, for electricity was
$0.0369 per kWh, with an average industrial sector price for 1984
through 1988 of $0.0373 per kWh (adjusted to 1980 dollars using the
gross national product implicit price deflators).   Thus, the cost
analysis was also revised using an electricity price of $0.037 per kWh
(see Docket Items IV-B-14 and IV-B-15).  For flares, the only
modifications deemed necessary were to the calculations for estimating
natural gas requirements, the cost of natural  gas, and flare piping
length.
      The effect of these changes in terms of estimating costs varied
from one cost item to another and from one control devTce to another.
The overall effect generally was to decrease marginally the estimated
annualized cost of a flare for a given emission stream and to increase
the estimated annualized cost of a catalytic incinerator for a given
emission stream.
                               8-8

-------
 Comment:
       One commenter  (IV-D-43/IV-D-50) stated that the economic  analyses
 in the revision to the proposal  (54 FR 890) are marred by the fact  that
 EPA did not update emissions data.  The commenter noted that since  the
 time of the September 30, 1987,  proposal (52 FR 36678), the concept of
 zero emissions from  product storage sections has changed.  For  reasons
 described by the commenter, the  commenter indicated that when EPA
 collected its initial data, the  polymer industry reported no emissions
 from storage silos in catalytic  processes (any low pressure process),
 but recent investigations showed that storage silo emissions were not
 completely eliminated, but rather were at levels approaching the 0.1
 weight percent level  that EPA is now considering for exemption from
 control.   According to the commenter,  the cost  of controlling these
 dilute streams far outweighs any benefits and should be reconsidered.
 The commenter did note that it appears that  the proposed  1.6 Mg/yr
 exemption (in the January 10,  1989,  Federal  Register notice)  will  likely
 exempt most  of the vent streams from storage silos  currently operating,
 although  the emission rate from storage  silos  is  largely  dependent on
 the design capacity of the production  facility  and  the  product  turnover
 rate in the  silos.  According  to the commenter,  the  trend  in  the
 industry  is  to build  larger  process  units and the exemption  of  1.6 Mg/yr
 may soon  become  inadequate.  The commenter then  stated  that  this concern
 will  be eliminated  if new sources were able  to  use the  calculated
 threshold emissions (CTE)  formulae proposed  for modified and
 reconstructed  facilities.  The  commenter also recommended that EPA
 reconsider emissions  from  product storage sections to project a  more
 realistic economic  impact  than  that  contained in the reproposal.
 Response:
      As noted in an  earlier response, the Agency has decided to extend
 the 0.1 weight percent exemption  and the various CTE equations for low
 VOC weight percent streams that were proposed for modified and
 reconstructed affected facilities to new facilities.  As the commenter
did not provide any new emissions data on product storage emissions, the
Agency evaluated the economic impact of the  new approach based on model
plant emissions and the extension of the  0.1  weight  percent exemption to
new facilities.
                               8-9

-------
 Comment:
       One  commenter  (IV-D-46)  stated  that  Docket  Item  IV-B-13,  titled
 "Impact of New Approach  Being  Considered on  Process  Sections  Subject  to
 Control" does not  include the  cost to capture  and  send  emergency
 releases to a flare.  The commenter noted  that, in the  proposed model
 plant  approach (Table 1  of 52  PR 36688), all five  polyethylene  processes
 and the polypropylene gas phase process were exempt  from  having to  send
 emergency  releases ("emergency vent streams")  to a flare, while under
 the proposed generic approach, these  six processes would  now  be required
 to capture and send emergency  releases other than decomposition
 emissions  to a flare.  According to the commenter, this exclusion of  the
 cost to capture and send emergency emissions to a flare is also
 reflected  in the written material of  Docket Item IV-B-13; for example,
 on page 2,  it states for low density  polyethylene, high pressure, "The
 primary difference is the potential control of product finishing and
 product storage emissions... ."  The  commenter stated that the  revision
 to the proposed regulations (54 FR 905, January 10,  1989) did not take
 into consideration the cost to capture on-decomposition emergency
 releases to a flare (a minimum of $15,000/Mg) and that the $1200/Mg
 abatement  cost on page 905 is  significantly understated because this  is
 the cost for abatement of emissions from Product Finishing and  Product
 Storage process sections only, and does not consider the cost of control
 of emergency releases which would be  approximately $15,000/Mg according
 to Table 2 of Docket Item IV-B-12 in  addition to the $1200/Mg.
 Response:
      The Agency believed that the definition of decomposition  and
 emergency emissions would cover the same set of emissions.  Thus,  the
 impacts presented in Docket IV-B-13 are correct given that belief.
However,  as pointed out elsewhere,  the Agency now understands this to be
 incorrect.   The final  rule has been revised to exclude control of such
emissions  so that the impacts presented in  Docket  Item IV-B-13 remain
unchanged.
Comment:
      One  commenter (IV-D-50)  stated that  in the cost analyses for the
Federal Register  notice reopening the public comment period,  EPA only
considered the cost of controls in  close proximity to the emissions
                               8-10

-------
  sources.   The  commenter referred  to  Docket  Item IV-B-4,  Table 3,  Piping
  Cost  for  Flares,  in  which  70  feet of piping  for a  source leg  and  20  feet
  of  piping for  the  compressor  leg  are costed.  The  commenter then  stated
  that  this is an  insufficient  amount  of  pipe  (90 feet  total) if the
  emission  source  and  the existing  control device are on opposite sides  of
  a polyethylene or  polypropylene plant.  According  to  the commenter 300
  to  400 feet of piping  is required normally to place the  flare a safe
  distance  from the  process.  The commenter then  concluded that EPA
  underestimated the cost  of piping  requirements,  and therefore,  the costs
  for flares.
  Response:
       The Agency reexamined Docket IV-B-4,  Table 3, discovered  that
  footnote  a was incorrect (in fact  140 feet of piping was costed, but the
 wrong costs were used), and reevaluated flare costs based on  a  total  of
 400 feet of piping.  This length of pipe was selected based on  a similar
 comment and resulting data collected in connection with the SOCMI-
 Distmation project.  The net effect was a small increase in  the  cost
 (see Docket Item IV-B-16).   The final rule  reflects this  change in
 costs.
 Comment:
       One  commenter (IV-D-50)  stated  that  EPA incorrectly calculated  the
 cost effectiveness  of controlling  product  finishing sections at new
 HOPE,  solution  process  plants  and  based  on  the corrected  incremental
 cost effectiveness,  EPA should not control  the product finishing section
 of a HOPE  process line,  which  was  the same  conclusion  reached  by EPA  in
 the  September 30,  1987,  Federal  Register notice.  The  commenter stated
 that EPA used incorrect  emission reductions in Docket  Item IV-B-13
 ("Impact of New Approach  Being  Considered in  Process Sections  Subject to
 Control"),  which  addresses the  impact of the  new  approach on various
 process sections.   The commenter states  that  the  correct  incremental
 cost effectiveness  is $8,850/Mg, not  $3,000/Mg.
 Response:
       In the September 30, 1987, Federal Register notice  the cost
 effectiveness of control was based on an emission reduction of 44.1
Mg/yr for a single process section.  The control  device being used  was  a
catalytic incinerator.  In the course of costing  catalytic incinerators
                               8-11

-------
            instance, one company uses a solvent, that has a boiling
            point of 99 'C.  This reduces the practicality of using  a
            stripper for this company because the temperature of
            atmospheric steam is 100 *C could approach the melting point
            of the copolymer charge causing agglomeration.
            A company with a higher boiling solvent could go to a vacuum
            system.  However, for a large scale plant the cost to go to
            a vacuum system will be very high due to the added
            construction requirements of the stripper vessel.  The EPA
            has failed to account for the added costs associated with
            designing such a large vessel under moderate to low vacuum
            conditions.
Response:
      The commenter is incorrect in stating that the annual emission
reduction is 44.1 Mg rather than 132.3 Mg.   As noted in the above
response, there was an error in the original calculation, which led the
commenter to believe that the annual emission reduction was 44.1 Mg.
      With regard to the commenter's statement that "EPA failed to
consider other important factors in its cost analysis by suggesting that
"all" HOPE solution plants should use a stripper",  the commenter
Incorrectly suggests that EPA is requiring  companies to install
strippers in order to comply with the standards.  The model  plant
presented in the BID was based on a U.S.  HOPE solution plant using a
stripper (see Docket Item II-D-30).   The  presence of the stripper in the
model plant in no way implies that all  companies are expected to use the
same process equipment, but that a typical  company  would have such a
piece of equipment.  To the extent that other companies do not use
strippers, as stated by the commenter,  then the model  plant is not
representative of those facilities.   For  the "new1"  approach for
determining which process streams to control,  this  is not a problem.
The applicable standards require 98 percent destruction or 20 ppmv.
They do not require that a company use a  stripper as part of the
process.  Thus,  the commenters suggestions  for further cost
considerations are irrelevant to the development of these standards.
                               8-13

-------
               9.0  REPORTING AND RECORD.KEEPING REQUIREMENTS

 9.1  GENERAL

 Comment:
       One commenter (IV-D-7) stated that the potential amount of record-
 keeping required to determine if modifications occur is enormous.  The
 commenter noted that,  while the designation of process section as the
 affected facility allows industry to keep the process sections of
 process lines separate, a company with an existing polymer plant must
 divide the plant into  process lines and into process sections, with a
 large polymer plant having as many as 10 to 15 affected facilities.  The
 commenter pointed out  that over the course of a year, many small  changes
 are normally made to polymer plants and stated that in order to evaluate
 whether a modification has occurred,  a company will have  to set up  a
 completely different accounting system to track the changes to those 10
 to 15 facilities.   The commenter believes that this will  result in  an
 unreasonable amount of recordkeeping.
 Response:
       The Polymers  NSPS does not require that  a separate  accounting
 system be kept  for  modifications to each process  section.   Records  do
 need  to be kept in  order to  determine  whether  alterations  are  classified
 as modifications  or reconstruction.  This  requirement is  part  of  the
 General  Provisions  of  40 CFR Part  60.   It  is at  industry's  discretion  to
 keep  these records  organized as  they see  fit.   For  the  purposes of
 determining  reconstruction,  the  affected  facility will  continue to  be
 the process  section.
 Comment:
      One  commenter  (IV-D-5) believes that Section  60.565(f)(i) of  the
 September  30, 1987,  Federal  Register notice, which  requires maintaining
 records of any process changes,  should be deleted because it is directly
 redundant with the General Provision requirements.  The commenter stated
 that the NSPS General Provisions address the topics of modification and
 reconstruction and require sources to consider these areas when making
process changes or additions.
                               9-1

-------

-------
 Response:
       The paragraph referred to by the commenter supplements the General
 Provisions [Section 60.7(a)(4)] by identifying specific items that are
 to be recorded.  The paragraph does not replace the General Provisions
 requirements.   Thus, the final rule retains this paragraph, which has
 been redesignated as Section 60.565(g)(l).

 9.2   FLOW RATES

 Comment:
       One commenter (IV-D-8)  stated that  Section 60.565(j)(2)  of the
 September 30,  1987,  Federal  Register notice (which  requires recording
 periods when the vent  stream has  been diverted from the control  device
 or has no flow rate) is  ambiguous  based on  the definition  of a vent
 stream found in Section  60.561,  "Definitions."
 Response:
       As  proposed,  the definition  of vent stream was  "any  gas  stream
 released  to the atmosphere from any polymers  and  resins  process  line."
 The Agency agrees with the commenter that this definition  of vent  stream
 could  make interpretation of  Section  60.565(j)(2) ambiguous, because  it
 may Imply that  only  vent streams emitted directly to  the atmosphere are
 being  defined,  and  since the  stream referred  to  in  Section  60.565(j)(2)
 is  being  sent to a control device  (rather than  to the atmosphere)  it
 does not meet the definition  of "vent stream."  This  interpretation,
 while  not without some logic  based  on the wording in the proposed rule,
 is  not correct.  The intent of Section 60.565(j)(2), which  has been
 redesignated as 60.565(k)(2)   in the final  rule, is  simply to record
 those periods of time when a gas stream being  controlled by a flare or
 other air pollution control  device  is diverted from that control device
or has no flow rate  (e.g.,  during process  shut-downs).  In order to
remove this ambiguity,  the Agency has revised the definition of vent
stream as  follows:
      "Vent stream" means any gas stream released to the atmosphere
      directly  from an  emission source or  indirectly either through
      another piece of process equipment or  a material recovery device
      that constitutes  part of the  normal  material recovery operations
      in a polymer process  line where potential emissions are recovered
                               9-2

-------
      for recycle or resale, and any gas stream directed to  an  air
      pollution control device.  The emissions released from an  air
      pollution control device are not considered a vent stream  unless,
      as noted above, the control device is part of the normal material
      recovery operations in a polymer process line where potential
      emissions are recovered for recycle or resale.
Comment:
      Two commenters (IV-D-5, IV-D-8) recommended that the provisions
requiring recordkeeping of flared vent stream flow rates (Sections
60.565(d)(l) and (2) of the September 30, 1987, Federal Register notice)
be deleted.  One commenter stated that as long as the vent streams in
question have no physical possibility of being discharged to  the
atmosphere prior to being flared, there is no reason to be concerned
over flow rates or lack of flow for any particular vent stream.
Response:
      As noted in a response to an earlier comment,  engineering piping
reports are now being required to describe the piping arrangement used
to route vent streams to the flare (or other control devices).  If
valves are present in the piping, then those valves  that direct the vent
stream to the control device will be required to be  kept car-sealed
opened; valves that would allow the vent stream to bypass the control
device and to be emitted to the atmosphere either directly or indirectly
through another piece of equipment are to be car-sealed closed.  The
final rule identifies two alternative monitoring methods (see Sections
60.563(d)(l) and (2) of the final rule).   One method calls for the
monthly inspection of all car seals.   For this method,  records are to be
kept of all maintenance performed on car-sealed valves, of all times
when the car seals are broken,  and of all  times when the valve position
is changed (i.e.,  from opened to closed for valves in the vent piping to
the control device,  and from closed to opened for valves venting
directly to the atmosphere).   The other monitoring method calls for the
installation of flow indicators downstream of all  valves that are
required to be car-sealed shut.   At least hourly recordings of flow are
required to be recorded and reported.   The Agency never intended that
flow rates be recorded, only that flow/no flow readings be obtained.
                               9-3

-------
 Comment;
       One commenter  (IV-D-5) recommended that the reporting of no flow
 conditions (Section  60.565(j)(2) of the September 30, 1987, Federal
 Register notice) be  deleted.  This commenter based this recommendation
 on their previous comments concerning the lack of a need for vent stream
 flow indicators.
 Response:
       The engineering report now required does not need periods of no
 flow to be reported  because periods of flows being diverted are being
 monitored.  Thus, the reporting of no flow conditions has been deleted.
 The requirement to record the period when the vent stream has been
 diverted from the control device is retained.

 9.3  PILOT LIGHT FLAME

 Comment:
       Two commenters  (IV-D-5,  IV-D-8)  recommended  that the  requirement
 for submitting continuous records  of the  flare  pilot  light  flame  heat-
 sensing monitoring  (Sections  6C.565(a)(3)(ii)  and  (a)(4)(ii)  of the
 September 30,  1987,  Federal  Register notice)  be deleted.  Commenter  IV-
 D-5 believes  that this requirement  is  inefficient  and  is  redundant with
 the requirements  of proposed  Sections  60.565(a)(3)(iii) and  (a)(4)(iii),
 which  require  records of'all  periods of operations during which the
 pilot  flame is absent.   This  commenter noted  that  a thermocouple, used
 in  this  type of monitoring, senses  the presence or lack of a flame and
 sounds  an  alarm if the  flame  is  extinguished, and  as such, continuous
 records  would  not typically exist.  The commenter  pointed out that the
 alarms would also be  triggered by a thermocouple failure.  Commenter  IV-
 D-5 stated that the periods for  which  the alarms are on are noted by the
 operators.  The commenter believes this arrangement satisfies the
 objective of determining when the flare pilots are operating without the
 paperwork and  expense of maintaining a continuous recorder.
      Commenter IV-D-8 stated that these sections in the proposed
 standards contain ambiguous instructions.   The commenter asked "How can
one record the flare  pilot light flame heat-sensing monitor?"  This
commenter suggested that these sections read that records  of checks on
the monitor should be kept.
                               9-4

-------
Response:
      The EPA considers it very important that records be kept and
reported which identify those periods when the pilot flame is absent
from the flare and unburned emissions are being released to the
environment.  The Agency believes that continuous records of pilot flame
heat-sensing monitoring, which can be obtained using a continuous
recorder that tracks the heat level, are necessary for enforcement
purposes.  Such records allow enforcement personnel  to confirm that
there is a flame present during periods that are not reported as flame
absent periods.  The commenter's arrangement would not provide the
"positive" assurance that a flame is present during  periods not reported
as flame absent periods.  In addition,  proposed Sections
60.565(a)(3)(ii)  and (a)(4)(ii) require that the pilot flame monitoring
data recorded during the initial and any subsequent  performance test be
submitted.  These are the only times the standards require these records
to be submitted.   Other sections of the standards require these records
to be kept for at least 2 years.  For these reasons,  the Agency has
retained these requirements (see Sections 60.565(a)(3)and (a)(5)  of the
final rule).
                               9-5

-------
                            10.0  MISCELLANEOUS

 10.1   DEFINITIONS

 Comment:
       One commenter (IV-D-1)  suggested  using  the phrase "each continuous
 vent  stream"  in  lieu of "vent stream that  emits  continuous  emissions"  in
 Section  60.562-l(a)(l)  of the September 30,  1987,  Federal Register
 notice.   The  commenter  later  clarified  that  this suggestion was  made
 only  to  simplify the expression.
 Response:
       The Agency agrees that  the phrase suggested  by  the commenter is
 less  cumbersome  than the one  used  in the paragraph  referred to by  the
 commenter,  for that single  usage.   However, phrases used elsewhere in
 the rule  are  better understood  if  the terms  "continuous emissions" or
 "emissions" are  retained than if the term  "vent  stream" is  used.
 Therefore,  the Agency prefers to retain the definitions for "vent
 stream,"  "continuous emissions," and "intermittent  emissions," and has
 retained  the  original phrasing  in  Section  60.562-l(a)(l).
 Comment:
       One  commenter (IV-D-2)  requested  that the  term  "thermoplastic" be
 added  to  the  definition  of  polypropylene to make all  polymer definitions
 consistent with  one another.
 Response:
       The Agency  has  inserted the  term  "thermoplastic" into- the
 definition of polypropylene,  because it is the intent of these standards
 to cover those production processes  that produce thermoplastic
 polypropylene and thermoplastic polypropylene copolymers.
 Comment:
       Several  commenters  (IV-D-39,   IV-D-44, IV-D-50) expressed concern
 over the definition  of  "concurrent"  in the January 10, 1989, Federal
 Register notice and the concept of concurrently constructed, modified,
 and reconstructed affected facilities.  Commenter IV-D-39  suggested that
the word  "concurrent" be deleted.   This  commenter stated that its
purpose is unclear and that it appears to require that modifications to
existing  facilities occurring  within two years of each other would be
                               10-1

-------

-------
 treated as new facilities with the stricter low VOC concentration
 requirements being applied.
       Commenter IV-0-44 suggested that the two-year time frame in the
 definition of "concurrent" be replaced with a six-month period.  This
 commenter stated that the definition of concurrent as proposed could
 impose retroactive additional control measures and costs to projects
 already completed or near completion.  Commenter IV-D-44 believes that
 no additional requirements should ever be imposed on projects already
 completed or which have initiated construction.   This commenter
 illustrated their concern by stating that major  projects could be
 vulnerable for additional  control requirements and incurred costs up to
 five years (three years from commencement to completion plus two-year
 concurrent period) from date of commencement.  Commenter IV-D-44  stated
 that this was "totally unreasonable and not cost effective."  The
 commenter believes that all  control  decisions  for concurrent projects
 should be made during their common planning/design phase and that
 projects  should not be considered concurrent unless they have a common
 planning  time frame.   According  to this commenter,  two  years is too  long
 a  time frame  for the  definition  of concurrent  and is  longer than  most
 planning  cycles.
       Commenter IV-D-44 also  stated  that  the proposed use of concurrent
 does not  appear to exempt  projects under  construction,  modification,  or
 reconstruction  prior  to January  10,  1989,  from additional control
 requirements.   The commenter  recommended  that projects  started  prior  to
 January 10, 1989,  be  excluded from being  considered concurrent  with
 other projects  begun  after January 10,  1989.
       Commenter  IV-D-50  stated that the wording of  the  definition of
 concurrent  is confusing  (54 FR 895).  According to  this  commenter, the
 definition  as stated  now would lead to the following scenario:  when  a
plant  commences a  project "B" within two years of the commencement of  a
 previous project "A"  at the same plant, and  it commences a project "C"
within two years of the commencement of project "B," then "B" can  be
concurrent with "A" and "C" can be concurrent with "B" while "C" may not
be concurrent with "A."  Commenter IV-D-50 then stated that this
situation could foreseeably go on for years and makes the completion
date of each successive project irrelevant.  Commenter IV-D-50
recommended that the EPA eliminate the concept of concurrent
                               10-2

-------
 control,  but that if the EPA feels this definition is absolutely
 necessary,  suggested as an alternative to the statement in the revised
 proposal  (54 FR 895) the following language:
       "... construction, modification,  or reconstruction within
       a process unit (replacing,  "of affected facilities") which
       has commenced in the two year period prior to the commencement
       date  of the construction,  modification, or reconstruction of
       an  affected facility."
 Response:
       Under the generic approach  for determining which  process emission
 streams are to be controlled  from polypropylene  and polyethylene
 facilities,  the Agency proposed  in the January 10,  1989,  Federal
 Register  notice that emissions from all concurrently constructed,
 modified, and reconstructed  affected facilities  be  combined (according
 to  the procedures outlined in  that notice)  for purposes  of determining
 which  emission streams would be controlled.   When  a new  plant  is  built,
 all  of the  process sections  are  (obviously) concurrent,  and the generic
 approach  requires combining emission streams  in  the same  weight percent
 range  across all  process  sections.   This  procedure  formed  the  basis  for
 the  development of the calculated  threshold emission  levels proposed  in
 the  January  10,  1989,  Federal  Register  notice.
       The Agency  extended  this concept  of concurrent  construction  to
 modified  and reconstructed affected  facilities.   If two process sections
 are  modified at the  same time, the Agency knows  of  no reason not to
 combine streams across  the two process  sections  for control
 determinations.   In  fact,  the generic approach is specifically  designed
 to reach  more  reasonable control/no  control determinations  when this  is
 done then when  each  process section  is considered individually.
 Further,  the Agency  believes that reasonable control decisions  can be
 made even for  affected  facilities that are not "concurrent," as defined
 in the January  10, 1989, Federal Register notice.
      The Agency  has decided that the term "concurrent" is  unnecessary
 to implement the  generic approach and has eliminated it from the final
 rule.  However, the Agency has replaced it with a different and more
 expansive procedure.  This new procedure requires uncontrolled  emission
 streams from an affected facility to be examined for pollutant  control
whenever a process section at the plant site is constructed, modified,
 or reconstructed  regardless of the time interval  between the

-------
 commencement  or completion  dates  of the  affected  facilities.   Once an
 emission  stream is  controlled  as  a  result  of these  standards,  it is
 never  again considered  for  determining the control  of other emission
 streams.
       In  implementing this  new procedure,  the Agency  disagrees with the
 commenters that it  is unreasonable  to require control  of  emissions from
 an affected facility that has  begun  operation (i.e.,  after  it  has  been
 completed).   The generic approach was designed to identify  that  level of
 annual emissions for a  given weight  percent  of VOC  in  a single or
 combined  emission stream above  which control  is deemed to be reasonable,
 regardless of the number of emission streams, the period of time when
 they became subject to  the standards, or the  planning phases or periods .
 at a plant site.  In addition,  the Agency disagrees that there  is  a  need
 to distinguish  between  those process sections that became affected
 facilities before January 10, 1989,  and those that become affected'
 facilities after January 10, 1989.  Reasonable control determinations
can be made regardless of an affected facility's applicability date.
      To illustrate how this new procedure works,  the following example
is provided.
      Example:  At a polypropylene plant,  Process  Section  A is
      reconstructed.  There  are three continuous  emission  streams (1,2,
      and  3),  one in each  of the three weight percent  ranges.   Stream 3,
      which  is in the 20 to  100 weight percent VOC range,  has  emissions
      greater  than  the calculated  threshold emissions  (CTE)  and,  thus,
      control  is required.   Emissions from  Streams 1 and 2  are  below
      their respective CTEs  and, thus, no control  is required".   Process
      Section  B  is modified,  and has  two emission  streams,  4 and  5.
      Emission Stream 4  is in the  same weight percent  range  as  Emission
     Stream 1,  and  Emission Stream  5 is in  the same weight  percent
     range as Emission  Stream  2.  These emission  streams would now be
     combined (4 with 1 and 5 with 2) to determine whether emissions  in
     each weight percent range  are greater  than their respective CTEs.
     Suppose  the total  emissions from Emission Streams 5 and 2 are
     greater  than the CTE for their weight percent.   These two streams
     would now  be controlled.   Suppose the other two streams (4 and  1)
     remain uncontrolled (i.e., their total annual emissions are less
     than the CTE for their weight percent).  Finally, Process Section
                              10-4

-------
       C is constructed at the plant site.   Process Section C has two
       emission streams,  6 and 7.   Emission Stream 6 is in the same
       weight percent range as Emission Streams 4 and 1.   Since the
       latter two streams are still  uncontrolled,, they would be combined
       with Emission Stream 6 from Process  Section C to determine whether
       controlled is required of all  three  streams.   Suppose the total
       emissions now exceed the CTE  for the combined weight percent,  then
       all  three streams  would be  controlled.   Emission Stream 7,  which
       is  in  the 20  to  100 weight  percent range,  would be  evaluated on
       its  own,  since there are no uncontrolled streams from an affected
       facility  in this weight percent  range.   If the annual  emissions  of
       Emission  Stream  7  are  less  than  the  CTE,  no control  is required.
       (Note:   In all cases where  control is not  required,  streams with
       flow less than 8 scfm  are still  required  to be controlled.)
Comment:
       One  commenter (IV-D-47)  stated that  the  term  "material  recovery"
may actually  provide a disincentive for good emissions control  practice
and that the  requirement  that  emissions control  be  required  at  a  98
percent effectiveness level may possibly encourage  less waste
minimization  to allow the  98  percent control to  be  achieved  at  low or  no
cost.
       In a follow-up conversation with the commenter  (Docket  Item  IV-E-
66), the commenter  clarified  these comments.  According to the
commenter, these comments were meant to convey the  idea that an owner or
operator may design  a plant upfront in a less-than-ideal manner (e.g.,
practice less material  recovery) so that uncontrolled emissions may be
higher than if  the  plant was designed in a more rigorous manner as the
control requirement of 98 percent reduction in the former situation
would be less costly to achieve than in the latter situation.
Response:
      The Agency believes that it is difficult to project the exact
effect these standards  will have on the amount of material recovery that
an owner or operator will practice now that certain emission standards
have to be met.  If an  owner or operator were  to reduce material
recovery (that would otherwise occur in the absence of these standards),
the resulting emission  stream would  be larger,  thereby requiring a
larger 'and more costly  control device.   On  the other hand, if the
                               10-5

-------
 material recovery were  to reduce only VOC concentrations,  the  cost  of
 control could  increase  if natural gas requirements  increased or  if  a
 different and  more expensive control device were required.   If flow was
 also reduced,  there may be no cost difference.  These possible outcomes
 are not unique to this  standard.  The Agency would  encourage an  owner or
 operator to minimize generation of waste materials  (i.e.,  air
 emissions).  (As a point of note, the provision of  the 0.1 weight
 percent VOC exemption would provide an incentive for material  recovery
 in certain situations.)  For purposes of these standards,  the  Agency has
 retained the use of the term "material  recovery" and the 98 percent
 reduction level.

 10.2  OTHER

 Comment
       One commenter (IV-0-46)  stated that the  column titled "Model  Plant
 Based  Approach/Intermittent  Emissions,"  in Table 3  of Docket  Item IV-B-
 13 is  Incorrect for the types  of processes listed on Table 1  of 52  FR
 36688  (September 30,  1987).
 Response:
       The Agency understands the commenter's confusion.   The  confusion
 arises  in part  because  of the different purposes of  the  two tables.
 Table  1  of the  September 30, 1987, Federal Register  notice  identifies
 those  process sections  for continuous and  intermittent emissions  that
 would  be subject  to the  standards as proposed on September  30,  1987.
 Table 3 of Docket Item  IV-B-13 identifies  those process sections  and
 emissions that  are projected to  be controlled using  the stream
 characteristics found in the model plants  and applying the  September 30,
 1987 proposed standards, including the uncontrolled  threshold emission
 rates.  Some process sections and emissions appear in the "Model  Plant"  '
 column of Table 3 of Docket Item IV-B-13 because emissions  occur  from
 these sections but would not be controlled because they were not
 identified as an affected facility in the September  30, 1987,  Federal
Recn'ster notice.  In these instances, their appearance does not imply
that they were identified as  affected facilities in  the September 30,
1987, Federal  Register notice.   They  are  included in  Table 3 of Docket
                               10-6

-------
 Item IV-B-13  because  under  the  new  approach  all  process  sections for
 both continuous  and  intermittent  emissions are  affected  facilities.
       Some  confusion  may  also occur because  "no  control"  is  indicated
 for  some  emissions from certain process  sections  that  were  identified as
 affected  facilities  in the  September 30,  1987,  Federal Register notice.
 This occurs because these emissions  were  included  in the  baseline  (would
 be controlled  in  the  absence of the  standards),  but their emission  rates
 were less than the uncontrolled emission  rate threshold  level  thereby
 potentially excluding them  from control.
 Comment:
       One commenter  (IV-D-39) stated  that the overall  impact of "new
 approach" versus  the original NSPS  proposed  rules;  is much more  effort on
 industry to determine what  is correct in  implementing  the "new  approach"
 rules  at their individual facilities.  The commenter believes that  it
 will be very difficult for  a proposed new facility in  the design phase
 to provide the specific unit-by-unit data required and that the data  can
 only be theoretical at that point or based on pilot plant data, which
 have inherent inaccuracies  in scaling up.
 Response:
      The Agency  agrees that the new approach requires more effort on
 the  industry's part to determine which process emissions are to be
 controlled versus the approach reported in the September 30, 1987,
 Federal Register  notice.   The Agency also feels that the new approach,
which was developed in response to specific industry comments on the
September 30,  1987, Federal  Register notice,  is more equitable and
appropriate in the end for controlling process emissions  that should  be
controlled.   Thus, there  is  a definite trade-off to the two approaches.
Comment:
      One commenter (IV-D-8) pointed out a typographical  error in
Section 60.564(f)'--  reference  should be to Section 60.564 not'Section
60.546.
Response:
      This typographical  error  has been  corrected in  the  final  rule.
                               10-7

-------
             APPENDIX A

FLOW DIAGRAMS ILLUSTRATING PROCEDURES
 FOR DETERMINING WHICH POLYPROPYLENE
  AND POLYETHYLENE EMISSION STREAMS
        ARE TO  BE CONTROLLED
                A-l

-------

-------
       The promulgated standards for polypropylene and polyethylene
 contain fairly complex procedures for determining which process
 emissions are subject to the standards.  This is the result of adopting
 a generic approach.  Commenters requested that the Agency clarify this
 procedure and the standards that are to be met.  To meet this request,
 the Agency developed a series of five flow diagrams, which are presented
 as Figures 1 through 3.   The purpose of these figures is to provide only
 an overview of the determination procedure for polypropylene and
 polyethylene process emissions,  and do not contain specific details
 found in the final rule.   The following paragraphs summarize the purpose
 of each figure.
       Figure 1 initiates  the determination procedure for each  process
 section.  Through this  figure,  affected facilities  are  identified and
 separated  according to  their applicability date  (between September 30,
 1987 and on  or before January 10,  1989,  and after  January  10,  1989).
 This figure  also  includes  the exemption  step  provided to affected
 facilities with an applicability date  between September  30,  1987,  and
 January  10,  1989,  and identifies how these  emissions can become subject
 to  the rule  at a  later date  (see Block  1.6).   For process  sections that
 are  identified affected facilities subject  to the standard,  Figure 1
 directs  the  user  to  Figure 2A for continuous  emissions and to Figure 3
 for  intermittent  emissions.
      Figure 2A is the first of'three flow diagrams applicable to
 continuous emissions.  The first step in Figure 2A separates those
 continuous emissions that are uncontrolled from those that are
 controlled in an existing control  device.  This is necessary as the
determination procedure is different depending on whether the emissions
                                  A-2

-------
 are already being controlled.   If they are,  the flow diagram directs the
 user to  Figure 2C.   For  uncontrolled  continuous emissions,  Figure 2A
 continues  by showing the exemptions provided for individual  emission
 streams  (see Block  2A.5).   Note  that  an  individual  stream that is
 exempted based on its annual emissions or weight; percent  can become
 subject to the standards at a  later date if  its annual  emissions  become
 1.6 Mg/yr  or greater (if exempted  based on the  annual emissions
 exemption)  or its VOC concentration becomes  0.1  weight  percent or
 greater  (.if exempted based on  the  VOC weight  percent  exemption (see
 Block 2A.6).   Once  qualifying  individual emission streams  are  exempted,
 the user is  directed to  Figure 2B.
      Figure 2B outlines  the procedures for combining nonexempt
 uncontrolled continuous  emissions  and determining which emissions  are  to
 be  controlled.  This  figure corresponds to the  steps detailed  in  Table  3
 1n  the final  rule.   An important feature of the  rule is the  "loop"
 provided between  Blocks  2B.12 and  2B.3.  In the  final rule,  uncontrolled
 emissions  that remain uncontrolled after passing through this  determina-
 tion procedure are  still  subject to control  in the future as new process
 sections become affected facilities.
      Figure  2C outlines the procedure for handling emissions  that  are
 already being controlled.  Note that  for these emissions there are  no
 individual  stream exemptions as fdr uncontrolled emissions.  The stream
characteristics of the inlet stream to the control device are  used  first
to calculate the calculated threshold  emission (CTE) level and second to
compare with the CTE level (Block 2C.2).   Also note that uncontrolled
emissions are combined with the controlled emissions in  one of two ways.
First, if the controlled  stream is to  meet the standards the next time
                                   A-3

-------
 the  control  device  is  modified,  reconstructed,  or replaced  (Block 2C.4),
 any  uncontrolled  emissions  in  the  same  weight  percent  range as  the
 controlled  stream are  also  to  be controlled  to  meet  the  standards.
 Second,  if  the  controlled stream's  emissions are  less  than  the  CTE
 level, any  uncontrolled emissions  in  the  same weight percent  range are
 combined with the controlled stream (Block 2C.8)  if  and  when  the  control
 device is modified, reconstructed,  or replaced  (Block  2C.7).
      Lastly, Figure 3 outlines  the determination procedure for
 intermittent emissions.  This procedure is much simpler  than  for
 continuous emissions as it  is based on stream type rather than stream
 characteristics.  This figure shows the exemption for emergency vent
 streams and the timing for when the standards are to be met, which
depends on whether the intermittent streams are uncontrolled or
controlled in an existing control device.
                                  A-4

-------
 BE6IH HERE EACH TIME A PROCESS
SECTION IS CONSTRUCTED, MODIFIED,
        OR RECONSTRUCTED.
 WAS PROCESS SECTION CONSTRUCTED
   MODIFIED, OR  RECONSTRUCTED
   AFTER SEPTEMBER 30, 1987?
                        t.i
    NO
               YES
PROCESS SECTION
NOT AN AFFECTED
FACILITY
                                                                        1.2
 WAS PROCESS SECTION CONSTRUCTED,
    MODIFIED, OR RECONSTRUCTED
     AFTER JANUARY 10, 1989?
                           1.3
 NO

1
               YES
                                           IS  PROCESS  SECTION  (AND
                                          ITS  EMISSIONS) IDENTIFIED
                                          IN TABLE  1 OF  THE  RULE AS
                                           AN AFFECTED  FACILITY?
                                                               1.4
                                                YES
                               YES
                                        ARE UNCONTROLLED EMISSIONS
                                        GREATER THAN THE UNCONTROLLED
                                          THRESHOLD EMISSION RATE
                                          IN TABLE 2 OF THE RULE?
                                                                 1.5
 GO TO FIGURE 2A, BLOCK 2A.I,  FOR
   CONTINUOUS EMISSIONS AND TO
     FIGURE 3, BLOCK 3.1 FOR
     INTERMITTENT EMISSIONS
                              1.7
                                                      NO
    NO CONTROL  IS REQUIRED AT
    THIS TIME.   IF AT A LATER
    DATE, EMISSIONS EXCEED THE
    THRESHOLD RATE OR IF MODIFIED
    OR RECONSTRUCTED AFTER
    JANUARY 10,  1989, THEN PROCEED
    TO FIGURE 2A, BLOCK 2A.1.
                             1.6
         Figure 1.   Initial Oecisionmaking  for Determining Which
           Polypropylene and  Polyethylene  Process Sections Are
              Affected  Facilities Subject to  the Standards
                                     A-5

-------
       FROM FIGURE 1
                   2A.1
ARE  EMISSIONS CONTROLLED
IN AN  EXISTING CONTROL
       DEVICE?
                  2A.2
       NO
YES
      GO TO
    FIGURE 2C.
         2A.3
MEASURE/CALCULATE WEIGHT
PERCENT AND  ANNUAL EMISSIONS
   OF  EACH STREAM.
                    2A.4
CONSIDER EACH  STREAM:
IS VOC WEIGHT  PERCENT
LESS THAN 0.1  OR ARE
ACTUAL EMISSIONS LESS
   THAN 1.6 MG/YR?
             2A.5
YES
            NO
NO CONTROL IS REQUIRED
AT THIS TIME FOR THESE
INDIVIDUAL STREAMS.
IF AT A LATER DATE,
EMISSIONS BECOME OR
EXCEED 1.6 MG/YR OR
CONCENTRATION BECOMES
OR EXCEEDS 0.10 WEIGHT
PERCENT VOC, THEN
PROCEED TO FIGURE 28,
BLOCK 28.1.
                 2A.6
   GO TO FIGURE 2B.
            2A.7 .
                  us  Emissions - Separation  of Controlled from
                Emissions  and Individual Stream Exemptions
                              A-6

-------
      FROM  FIGURE 2A,
      BLOCK 2A.7

                               ACCORDING T0
                                    20> 20 T0 100)
     .ADD IN ANY UNCONTROLLED  EMISSION STREAMS
         IN THE SAME WEIGHT PERCENT RANGE
        FROM PREVIOUS AFFECTED FACILITIES.
                                      a.3
     CALCULATE TOTAL ANNUAL
     EMISSIONS FOR EACH
     WEIGHT PERCENT RANGE
     ACCORDING TO THE
     PROCEDURES IN TABLE 3.
                        a.*
     20 TO 100 WEIGHT
     PERCENT
                     CONTROL 98%,  TO 20 PPMV,  IN  A
                 CONTROL DEVICE THAT MEETS SPECIFIED
                     OPERATING CONDITIONS,  OR  IN
                      AN EXISTING CONTROL DEVICE
                                                 28.7
 ARE EMISSIONS EQUAL TO OR
GREATER THAN  THE CALCULATED
   THRESHOLD  EMISSIONS?
                                                     ZI.S
     5.5 TO 20 WEIGHT
     PERCENT

     0.1 TO 5.5 WEIGHT
     PERCENT
                                                                YES
 ARE EMISSIONS EQUAL TO OR
GREATER THAN THE CALCULATED
   THRESHOLD EMISSIONS?
                                                     28 .A
                              SPLIT STREAMS INTO
                              >8 SCFM AND THOSE
                                  <8 SCFN.
                                           lt.9
                             CONTROL 98%. TO 20 PPMV, JR
                            IN A CONTROL DEVICE THAT MEETS
                            SPECIFIED OPERATING CONDITIONS
                                                    za. a
                      CONTROL AT THIS TIME.   RETURN TO DECISIONMAKING
                     PROCESS NEXT TIME A PROCESS  SECTION BECOMES AN
                   AFFECTED FACILITY OR A CONTROL DEVICE IS MODIFIED,
                 RECONSTRUCTED, OR REPLACED  (SEE  FIGURE 2C, BLOCK 2C.8).
                                                                  23.12
  Figure 2B.   Decisionmaking  Process  for Uncontrolled Continuous
Emissions from Polypropylene and Polyethylene Affected Facilities
                                     A-7

-------
                                 FROM FIGURE 2A,
                                   BLOCK 2A.3
                                           2C.1
                              ARE  EMISSIONS EQUAL TO  OR
                                GREATER THAN THE
                           CALCULATED THRESHOLD EMISSIONS?
    STANDARD DOES
    NOT NEED TO BE
   MET AT THIS TIME.
   	       ZC.4
NO
" YES
    CONTROL DEVICE IS MODIFIED
    RECONSTRUCTED, OR REPLACED.
              	     2C.7
                 DOES EXISTING CONTROL
                 DEVICE REDUCE EMISSIONS
                 BY 98 PERCENT OR TO
                 20 PPMV OR MEET NECESSARY
                 OPERATING REQUIREMENTS?
                                   2C.3
    ADO IN UNCONTROLLED STREAMS
    IN SAME WEIGHT PERCENT
    RANGE FROM PREVIOUS
    AFFECTED FACILITIES.
    .	               2C.
    ARE EMISSIONS NOW
    EQUAL TO OR GREATER THAN
    THE CALCUALTED THRESHOLD
      EMISSIONS?
         NO
                     YES
                           NO
                                       \
                                YES
     STANDARD IS TO BE MET NEXT
     TIME THE CONTROL DEVICE IS
     MODIFIED, RECONSTRUCTED, OR
     REPLACED.  ADO IN ANY UNCON-
     TROLLED EMISSIONS IN SAME
     WEIGHT PERCENT RANGE FROM
     ANY AFFECTED FACILITY.
                                 NO  FURTHER
                                 CONTROL IS
                                 REQUIRED.
                                       2C.S
                                     CONTROL  BY 98 PERCENT, TO 20  PPMV,
                                     OR IN A  CONTROL DEVICE THAT MEETS
                                      SPECIFIED OPERATING CONDITIONS.
                                                                 2C.10
  NOTE:
                                      SSES"FOR EMISSIONS AUEADY
Figure 2C    Decisionmaking Process  for Continuous  Emissions  Already
  Controlled  at Polypropylene and  Polyethylene Affected Facilities
                                      A-8

-------
     FROM FIGURE 1
                3.1
     EXEMPT ANY  EMERGENCY STREAMS
                            3.2
ARE EMISSIONS CONTROLLED IN
AN EXISTING CONTROL DEVICE?
3.3
NO

STANDARDS ARE
TO BE MET.
3.4
           YES
DOES CONTROL DEVICE
MEET STANDARDS?
3.3


YES

NO FURTHER
CONTROL IS
REQUIRED.
3.6
           NO
     STANDARDS TO BE MET NEXT
     TIME THE CONTROL DEVICE IS
     MODIFIED, RECONSTRUCTED,
     OR REPLACED.
                        3.7
Figure 3.   Decisionmaking Process for Intermittent Emissions  from
        Polypropylene and Polyethylene Affected  Facilities
                                 A-9

-------
                                      TECHNICAL REPORT DATA
                              (flense read instructions on the reverse before completing
       EPA-45Q/3-83-mQh
                                                               3. RECIPIENT'S ACCESSION
          IOSU8T1TU5
      Polymer Manufacturing Industry - Background
      Information for  Promulgated Standards"
                                                               5. REPORT DATE
                            October  1990
                i. PERFORMING ORGANIZATION CODE
                                                               8. PERFORMING ORGANIZATION REPORT N.
 9. PERFORMING ORGANIZATION NAME AND ADDRESS
                                                               10. PROGRAM ELEMENT
     Office of Air Quality  Planning and Standards
     Environmental Protection  Agency
     Research Triangle Park, NC   27711
                11 CONTRACT/GRANT NO
              iGENCY NAME AND ADDRESS
    DAA  for Air Quality Planning  and Standards
    Office  of Air and Radiation
    U. S. Environmental Protection  Agency
    Research Triangle Park. NC  27711	
 15. SUPPLEM6NTARY NOTES
               13. TYPE OF REPORT AND PERIOD COVEREC
                       Final
               14. SPONSORING AGENCY CODE
                     EPA/200/04
     ttalvr               f
-------

-------