United States      Office of Air Quality       EPA-450/3-82-011 b
           Environmental Protection  Planning and Standards     March 1984
           Agency         Research Triangle Park NC 27711

           Air    I                   ~~~~~
vvEPA     Synthetic Fiber      Final
           Production           EIS
           Facilities—
           Background
           Information for
           Promulgated
           Standards

-------
                                  EPA-450/3-82-011b
Synthetic Fiber Production Facilities-
       Background Information for
         Promulgated Standards
             Emission Standards and Engineering Division
            U.S. ENVIRONMENTAL PROTECTION AGENCY
                 Office of Air and Radiation
            Office of Air Quality Planning and Standards
            Research Triangle Park, North Carolina 27711

                    March 1 984

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

-------
                    ENVIRONMENTAL PROTECTION AGENCY

                         Background Information
                  Final Environmental Impact Statement
                 Synthetic Fiber Production Facilities
                              Prepared by:
Jac/R. Farmer
Director, Emission Standards and Engineering Division
U.S. Environmental Protection Agency
Research Triangle Park, North Carolina  27711

1.   The promulgated standards of performance will limit emissions of
     volatile organic compounds (VOC) from new and reconstructed
     synthetic fiber production facilities.  Section 111 of the Clean
     Air Act (42 U.S.C. 7411), as amended, directs the Administrator to
     establish standards of performance for any category of new stationary
     source of air pollution that ". . . causes or contributes
     significantly to air pollution which may reasonably be anticipated
     to endanger public health or welfare."  EPA Regions I, II, III,
     and IV are particularly affected, since most synthetic fiber
     production facilities are located in these regions.

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

3.  For additional  information contact:

    Mr. Robert L. Ajax
    Standards Development Branch (MD-13)
    U.S. Environmental  Protection Agency
    Research Triangle Park, NC  27711
    telephone:  (919)  541-5578

4.  Copies of this  document may  be obtained from:

    U.S. EPA Library  (MD-35)
    Research Triangle Park,  NC  27711

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

-------
AL

-------
                         TABLE  OF  CONTENTS

 Chapter                                '                           Page
 1.0   SUMMARY	1-1
      1.1   Summary  of  Changes Since  Proposal   	  1-1
      1-.2   Summary  of  Impacts of the Promulgated  Action	1-1
           1.2.1  Alternatives  to  Promulgated  Action   	  1-1
           1.2.2  Environmental Impacts of  Promulgated Action  .  .  1-1
           1.2.3  Energy  and Economic  Impacts  of  Promulgated
                 Action	]_1
           1.2.4  Irreversible  and Irretrievable  Commitment
                 of Resources	1-2
           1.2.5  Environmental and  Energy  Impacts of Delayed
                 Standards	1-2
           1.2.6  Corrections and Clarifications	1-2
 2.0   SUMMARY OF PUBLIC COMMENTS	2-1
      2.1   Selection of Source  Category 	  2-1
      2.2   Selection of Best Demonstrated Technology	2-4
      2.3   Selection of Format  of the Standards	2-14
      2.4   Environmental  Impact 	  2-16
      2.5   Costs and Economic Impacts 	  2-20
      2.6   General	2-33
APPENDIX A  Calculation of Enclosure Capture Efficiency	  A-l
APPENDIX B  Safety Concerns With the Use of Enclosures 	  B-l
APPENDIX C  Revised Text and Table From BID Chapter Eight. .  . .  C-l

                        LIST OF TABLES
Number                                                            Page
2-1  List of Commenters on the  Proposed Standards of Performance
     for Synthetic  Fiber Production  Facilities 	  2-2
2-2  Summary of Growth Projections Considered  	  2-22

-------
LA

-------
                              1.0  SUMMARY
     On  November 23,  1982, the Environmental Protection Agency (EPA)
 proposed standards of performance for synthetic fiber production
 facilities  (47 FR 52932) under authority of Section 111 of Clean Air
 Act.   Public comments were requested on the proposal in the Federal
 Register.   There were 6 commenters composed mainly of industry and
 industry association  representatives.  Also commenting was one State
 environmental agency.  The comments that were submitted, along with
 responses to these comments, are summarized in this document.
 1.1  SUMMARY OF CHANGES SINCE PROPOSAL
     There  have been no changes, other than editorial or typographical,
 made to  the regulation since proposal.
 1.2    SUMMARY OF IMPACTS OF PROMULGATED ACTION
 1.2.1  Alternatives to Promulgated Action
     The regulatory alternatives are discussed in Chapter 6 of the
 proposal BID.  These regulatory alternatives reflect the different
 levels of the emission control from which one is selected that represents
 the best demonstrated technology, considering costs, nonair quality
 health,  and environmental and economic impacts on the synthetic fibers
 industry.  These alternatives remain the same and are described in
 Chapter  6 of the proposal BID.
 1.2.2  Environmental  Impacts of Promulgated Action
     The environmental impacts of the promulgated standards are described
 in Chapter 7 of the proposal  BID, and are essentially unchanged.
     The analysis of environmental  impacts in the BID Volume I,
with the changes  noted in BID Volume II,  now becomes the final
Environmental  Impact  Statement for the promulgated standards.
 1.2.3  Energy  and Economic  Impacts  of Promulgated Action
     The energy impacts  of  the promulgated action are described in
Chapter  7 of the  proposal  BID,  and  remain unchanged since proposal.
                                  1-1

-------
     Economic impacts are described  in  Chapter  8  and  9.  The  values  for
cost-effectiveness of the standards  at  facilities  [projected in  1987
have been adjusted for accuracy,  but still  indicate the  standards will
achieve emission reductions  at  reasonable cost.
1.2.4  Irreversible and Irretrievable Commitment  of Resources
     These impacts are discussed  in  Chapter 7 of  the  proposal BID, and
remain unchanged since proposal.
1.2.5  Environmental and Energy Impacts of  Delayed Standards
     Table 1-1 in the proposal  BID provides a summary  of the  impacts
associated with the proposed standards. Delay  in  implementation of  the
standards could result in additional VOC emissions, as described in
Chapter 7 of the proposal BID.
1.2.6  Corrections and Clarifications
     Comments received from tie public  following  proposal  included
notation of several minor typographical and mathematical errors in the
proposal BID and preamble.  These are discussed in detail  in  Chapter 2
of this document and docket  item II-B-2. None  of  the  errors  will
result in changes in the environmental  or economic impacts of the
standards.
                                  1-2

-------
                2.0  SUMMARY OF PUBLIC COMMENTS

     The list of commenters and their affiliations is shown in Table 2-1.
Six comment letters were received.  A summary of the comments and EPA's
responses to them are presented in this chapter under the following
headings:
     1.   Selection of Source Category
     2.   Selection of Best Demonstrated Technology
     3.   Selection of Format of the Standards
     4.   Environmental Impacts
     5.   Costs and Economic Impacts
     6.   General
2.1 SELECTION OF SOURCE CATEGORY
2.1.1  Comment (IV-D-1, IV-D-2, IV-D-3, IV-D-4, IV-D-6):
     Five commenters said that the proposed NSPS is not  needed because
there will be no capacity additions in the solvent-spun  synthetic fibers
industry in the next 5 years.
Response:
     EPA does not agree with the commenters that there will  be no
capacity additions in the next 5 years (see responses to  Comments 2.5.1,
2.5.2, 2.5.8, 2.5.9, 2.5.14).  Even if the commenters' projections
are correct, however, EPA believes the NSPS would still  be warranted.
If growth is projected to occur, whether within 5 years  or beyond 5
years, the issue is whether best control  technology should be a  factor
in that growth.  The 5-year period has no special significance in the
decisions as to whether or not an NSPS is warranted.   The 5-year period
is often a reasonable indicator of growth.  However,  such factors as
cyclic growth, the current economic downturn,  etc., can  result in
situations where projected growth in the next  5 years is  not  indicative
necessarily of long-term trends.  Since an NSPS is intended to achieve
long-term benefits, it is important to look beyond 5  years.   Even if
growth is not certain, however, it is still not unreasonable  to  promulgate
the NSPS.  EPA has been developing it for over 3 years, and most  of

                                  2-1

-------
         Table 2-1.  LIST OF PERSONS SUBMITTING COMMENTS ON THE
                           PROPOSED STANDARDS


Docket Entry Number3                    Commenter and Affiliation


IV-D-1                              James C. Pullen
                                    Manager, Environmental  Activities
                                    Celanese Fibers Company
                                    Charlotte, North Carolina  32414

IV-D-2                              Charles W. Jones., President
                                    Man-made Fiber Produces Association, Inc
                                    Washington, D.C.  20036

IV-D-3                              (transmittal  letter)
                                    Robert R. Romano, Ph.D
                                    Manager, Air Programs
                                    Chemical Manufacturers  Association
                                    Washington, D.C.  20037

                                    (body of comment)
                                    Geraldine V.  Cox, Ph.D
                                    Vice President and Technical  Director
                                    Chemical Manufacturers  Association
                                    Washington, D.C.  20037

IV-D-4                              (transmittal  letter)
                                    Robert L. Stoots, Jr.
                                    Coordinator Agency Relations
                                    Tennessee Eastman Company
                                    Kingsport, Tennessee 37662

                                    (body of comment)
                                    James C. Edwards
                                    Manager, Clean Environment Program
                                    Tennessee Eastman Company
                                    Kingsport, Tennessee 37662

IV-D-5                              Daniel  J. Goodwin
                                    Manager, Division of Air Pollution
                                      Control
                                    Illinois Environmental  Protection
                                      Agency
                                    Springfield,  Illinois  62706

IV-D-6                              David T. Modi
                                    Attorney, Environment Division
                                    E.I. duPont de Nemours  and Company
                                    Wilmington, Delaware 19898

aThe docket number for this project is A-80-7.Dockets are on file at
 EPA Headquarters in Washington,  D.C., and at the  Office of Air Quality
 Planning and Standards in Durham,  N.C.

                                   2-2

-------
 the Agency  resources  for  this  project  have  already  been  spent.   Since  no
 extra  effort  is  required,  and  since early promulgation of these  standards
 would  enhance the  ability  of  facility  owners to  plan  for whatever  future
 growth  will be necessary,  establishment  of  the standards at this time  is
 reasonable.   Consequently,  EPA believes  the NSPS will be beneficial  in limiting
 VOC emissions  from new  or  reconstructed  synthetic fiber  production facilities
 when they are built,  regardless of whether  it is within  5 years  or beyond.
 2.1.2  Comment  (IV-D-2):
     One commenter claimed  that the proposed regulation  is unnecessary
 because no  new or  significantly modified facilities that produce acrylic,
 modacrylic, and  cellulose  acetate fibers will be built or required in  the
 next 5 years.  This commenter  referred to past communications for support
 of this position,  but did  not  cite the pertinent supporting portions.
 He also claims the  background  information document  (BID) for the proposed
 standard contains  significant  errors of  an  economic and technical nature,
 but no specific  errors  were noted.  He summarizes by  requesting that
 EPA "---discontinue this needless activity."
 Response:
     As stated in  the response to Comment 2.1.1, EPA  believes that the
 NSPS will be beneficial  in  limiting VOC emissions from affected facilities
 when they are  built, regardless of whether  it is within 5 years or beyond,
 and that there will be  significant growth in this industry in the future.
 The communications that have been provided  by others  have been carefully
 considered, and  it is EPA's conclusion that  none affects EPA's assessment
 of the benefits  of the  standards.  The communications referred to in the
 above comment overall  do not show, in EPA's  opinion, that the standards
 are "needless." The commenter has provided no specific support for his
 comment.  Absent  this information, EPA must  conclude that it  has made a
 reasoned judgment based  on the best information available,  and that
 there is not sufficient  justification for discontinuing  development of
 the standard;  the preponderance of information  available shows that the
 standards will result  in emission reductions at  reasonable  cost  from a
category of sources that contributes  significantly  to ozone pollution.
2.1.3  Comment (IV-D-6):
     One commenter questioned  the significance  of VOC emissions  from
synthetic fiber plants by  pointing out  the small  percentage of VOC

                                  2-3

-------
emissions^ contributed by the synthetic fiber industry to total VOC
emissions from all sources.  He estimated that the fiber industry
baseline emissions are about 0.2 percent of emissions from all sources
and that the emission reduction achieved by the standard would amount
to 0.05 percent, or less, of total  VOC emissions.
Response:
     EPA agrees that VOC emissions  from the synthetic fibers industry are
a small percentage of total VOC emissions; however, most VOC emissions
come from a large number of relatively small sources (when compared,
for example, with sources of particulate matter or sulfur dioxide,
which can be much larger).  There are no relatively large individual
sources of VOC emissions; rather, emissions from all these industries
combined create the ozone danger Congress intended standards of perfor-
mance to address.  Since these emissions can be reduced only by controlling
each type of contributor, most of these individual contributors must be
viewed as significant, even though  emissions from each may seem small
when compared to the total.  More specifically for this industry, EPA
considers synthetic fiber plants to be significant sources regardless
of the percentage of VOC emissions  they contribute to the total.  For
example, a typical dry spinning acrylic plant controlled to baseline
levels would emit about 1,900 megagrams of VOC per year.  The NSPS
would reduce those emissions by about 1,100 megagrams per year.  Section
302(j) of the Clean Air Act defines a major stationary source as one
that emits 100 tons (91 megagrains)  or more of an air pollutant.  For
these reasons, EPA believes that there is no reason to alter its con-
clusion, established by rulemaking  at 40 CFR 60.16, that the synthetic
fiber production industry is a significant contributor of VOC emissions
and should be listed for regulation by an NSPS.
2.2  SELECTION OF BEST DEMONSTRATED TECHNOLOGY
2.2.1 Comment (IV-D-3, IV-D-4):
     Two commenters said that it is not appropriate to transfer enclosure
technology used in some acrylic fiber production to filter tow production
because of fundamental differences  in the types of solvents, raw
materials used, curing rates, aad resulting emission rates.  The two
commenters provided conflicting views, however, concerning the weight
                                  2-4

-------
and  resulting curing rates of acrylic fibers as compared to filter tow.
Commenter IV-D-3 indicated that filter tow fibers are lighter and thus
would have a faster curing rate while commenter IV-D-4 indicated the
opposite would be true.
Response:
     Neither commenter explained why the difference in emission rates
they believe to exist between acrylic and filter tow fiber spinning
would cause the use of enclosures for filter tow to be inappropriate.
Commenter IV-D-3 did provide a comparison between the variables causing
different emission rates for modacrylic and filter tow fiber production
but  did not explain how these variables would affect the effectiveness
of enclosures.  All the variables discussed by the commenters were
carefully considered by EPA during development of the proposed NSPS, as
indicated in Chapter 4 and Appendix C of the proposal BID.  Unfortunately,
most of the data regarding solvents, raw materials, fiber size, and
curing rates are claimed to be confidential by the companies from which
they were obtained and, therefore, could not be discussed in detail
either in the BID or in this document.  After consideration of these
variables, however, EPA concluded that properly designed and operated
enclosures are the most effective means of capturing VOC emissions from
all spinning facilities (with the exception, perhaps, of acetate filament
yarn - see 47 FR 52937, the preamble to the proposed NSPS).
     A thorough investigation revealed no design,  operational, or
safety problems associated with enclosures. In addition,  in correspon-
dence to and meetings with EPA personnel  during development of the
proposed NSPS, representatives of commenter IV-D-4 made the following
statements:  "The design and installation  of an enclosure system does not
pose any formidable problems." (Attachment 1 to docket item II-E-92),
and  ". . .  technically,  enclosures could be designed to effectively
capture VOC  emissions from the spin cell  area and  provide worker access."
(Page 2 of docket item II-D-68).
     One representative noted that his company has studied the possibility
of recovering fugitive acetone emissions  from their filter tow process.
He noted that enclosures  are the  most viable capture system to pursue
but that it  is  not  now economically attractive.  He estimated that  an
                                  2-5

-------
enclosure system would have an approximate 5-year return on investment
(docket item II-D-68).
     Finally, enclosures are being successfully used on a filter tow
spinning facility in Japan, confirming EPA's position that they are
appropriate as control technology on new and reconstructed filter tow
facilities that would be subject to the NSPS.
2.2.2  Comment (IV-D-3,  IV-D-4):
     Two commenters claimed that enclosures do not meet the requirements
for best demonstrated technology for the cigarette filter tow industry
because they have not been demonstrated for domestic facilities.
Commenter IV-D-3 also stated that enclosure technology has not been
demonstrated for many acrylic fiber production areas and is not
representative of technology employed by this  category.
Response:
     The commenters' use of the term "best demonstrated technology" is
a reference to Section lll(a)(l)(C) of the Clean Air Act, which specifies
that a standard of performance "... reflects the degree of emission
reduction achievable through the application of the best system of
continuous emission reduction which (taking into consideration the cost
of achieving such emission reduction, and any  nonair quality health and
environmental impact and energy requirements)  the Administrator determines
has been adequately demonstrated for that category of sources." EPA
normally refers to this  system of continuous emission reduction as
"best demonstrated technology" or BDT.
     The commenters have interpreted the Section 111 requirement that
the system be adequately demonstrated as meaning that it be in actual
use at each type of existing facility in the category of sources being
regulated and that it be achieving the level of the NSPS for which it
is the basis.  EPA interprets the requirement  more broadly.  Control
technology can be considered BDT if it can be  shown to be the best
system demonstrated for  the category of sources, not necessarily on the
category of sources.  This means that a system used in an entirely
different industry using a different process from the one being regulated
can be BDT if its performance would not be affected by the differences
in the sources.  Similarly, a system used in some segments of the
industry being regulated, or in some parts of  the process, but not
                                  2-6

-------
 others,  can  be  considered BDT  for  all  segments or all parts of the
 process  if investigation of the  relevant  variables reveals no reason that
 it  cannot be designed,  installed,  and  operated so that  it achieves the
 same  emission control under all  the conditions in which it would be
 applied.
      This interpretation of the  term "demonstrated" has been upheld in
 the courts.  As stated  by the  D.C. Circuit in Portland Cement Association
 v. Ruckelshaus, 486 F.2d 375,  391  (1973):
      "We begin  by  rejecting the  suggestion of the cement manufacuturers
      that the Act's requirement  that emission limitations be 'adequately
      demonstrated' necessarily implies that any cement plant now in
      existence  be  able  to meet the proposed standards.  Section 111
      looks toward  what  may fairly  be projected for the regulated future,
      rather  than the State of the  art at  present, since it is addressed
      to  standards  for new plants - old stationary source pollution
      being controlled through  other regulatory authority."
      EPA believes  that  enclosure technology as a means of capturing VOC
 emissions from  sources  in synthetic fiber processing facilities meets
 these criteria  (as explained in the response to comment 2.2.1) and
 therefore represents BDT for the industry.
 2.2.3  Comment  (IV-D-3, IV-D-4):
    Two commenters stated that the 90 percent capture efficiency for
 enclosures that is part of the basis for the NSPS is not supported by
 adequate documentation  or data.  They also contended that since enclosures
 must  be opened  frequently to allow worker access to equipment, the 90
 percent capture efficiency is not supportable. Commenter IV-D-3 estimated
 that enclosures could capture no more than 86 percent of VOC emissions.
 He apparently based this efficiency on an estimate that enclosures
 would be open, rendering the system ineffective,  for 8 to 19 percent  of
 the time and that  overall  out-of-service time for the enclosures would
 be about 14  percent.
 Response:
     As indicated  in the preamble to the proposed NSPS (47 FR 52936),
 the 90 percent capture efficiency used  by EPA to  calculate achievable
 emission reductions was based on solvent use and  emission data collected
 for fiber production facilities of several companies.   Emission  tests
were conducted and solvent mass balance data collected at acrylic  fiber
 plants that  use enclosures (docket items II-A-19,  II-A-20,  II-B-99).

                                  2-7

-------
Although these data indicated that the tested plants were achieving
about 95 percent capture with their enclosures, EPA selected 90 percent
to account for the more frequent opening of enclosures to allow worker
access that is necessary at filter tow facilities (estimated at 14
percent based on information provided to EPA by commenter IV-D-4 (docket
item II-D-68)).
     In order to respond to commenter IV-D-3's concern that the opening
of enclosures for as much as 19 percent of the time would further
reduce the capture efficiency of the enclosures to 86 percent,  EPA
recalculated the efficiency.  (For the detailed calculation, see Appendix A.)
The calculation indicates that under these worst-case conditions, the
lowest capture efficiency would be over 91 percent, not the 86  percent
claimed by the commenter.  Therefore, EPA continues to believe  that
enclosures on fiber processing facilities can consistently achieve
greater than 90 percent capture.
2.2.4  Comment (IV-D-3):
     One commenter criticized EPA for not having a standardized method
for measuring capture efficiency and referred to a memorandum written
by EPA's contractor that indicated the lack of a test method to measure
capture efficiency created a problem in determining emission rates.
The commenter also cited a memorandum written by EPA's Office of General
Counsel  regarding capture efficiency as it relates to the NSPS  for
rubber tire manufacturing.  The memo discussed the need to specify an
acceptable method for measuring capture efficiency since it was an
integral part of the rubber tire standard.
Response
     Although an acceptable test method for directly measuring
capture efficiency of enclosures at synthetic fiber plants would have
been helpful during development of the proposed NSPS, it was not
essential  and is not needed to show compliance with the standard.  EPA
was able to determine capture efficiencies of enclosures through use  of
solvent mass balances, emissions data, and other industry-provided
information (see response to comment 2.2.3)  to determine the level  of
VOC emission control reflected in the standard.
     The standard itself does no_t require determination of capture
efficiency.  Rather, it requires a calculation involving plant  records

                                  2-8

-------
of  solvent flow and/or polymer flow Into and out of the affected facility.
While the OGC opinion may be appropriate for the rubber tire NSPS, it
is  not relevant to  the synthetic fibers NSPS because it requires a
different means of  demonstrating compliance that does not involve
determination of capture efficiency.
2.2.5  Comment (IV-D-3. IV-D-4):
     Commenter IV-D-3 asserted that capture efficiency of an enclosure
is  decreased by the efficiency of the ultimate control device (e.g.,
carbon beds).  He noted that if a carbon bed is estimated to be 92
percent efficient,  then the total capture efficiency of the enclosure
could be no more than 80 percent (0.86 x 0.92 = 0.79).  Commenter IV-D-4
said, "...  because vapors captured by the enclosure must be routed
subsequently to other control devices (e.g., carbon beds), a 90 percent
enclosure efficiency is'not supportable or realistic."
Response:
     Both commenters are apparently referring to overall control efficiency
rather than capture efficiency.  Capture efficiency is the amount of
a substance collected or captured by the enclosure compared to the
amount released by  the source enclosed.  Thus, any downstream treatment
by  a control device has no bearing on capture efficiency.  It should
also be noted that  in'the BID the only application for which a carbon
bed was specified to be 92 percent efficient at synthetic fiber production
facilities is one treating filter tow dryer emissions which have unusually
high relative humidity.  All other carbon beds operated in the synthetic
fibers industry can achieve over 95 percent efficiency, as discussed in
the BID and preamble.
2.2.6  Comment (IV-D-3, IV-D-4):
     Two commenters claimed that the use of enclosure technology on
filter tow facilities would create a safety hazard because acetone
concentrations within the enclosures could build up to explosive levels
within 30 seconds.   They acknowledged that a foreign producer uses
enclosures in filter tow facilities but believed that the fundamental
differences (spinning line speeds)  in the foreign and domestic  operations
make transfer of this technology to domestic operations inappropriate.
They expressed concern that the manual  activation of safety  systems
                                  2-9

-------
used by the foreign producer would not be fast enough to rely on to
prevent an explosion within the 30 seconds it would take for a potential
explosion to occur.
Response:
     The safety of enclosures on filter tow facilities was a subject of
several meetings and communications between EPA and commenter IV-D-4
during development of the proposed NSPS (docket items II-B-76, II-B-92,
II-E-87, II-E-89, II-E-91, II-E-92, II-E-94, and others).   All the
points regarding safety raised in their comment on the proposed NSPS
were considered by EPA prior to proposal.  As discussed in Chapter 4 of
the proposal BID, EPA is aware that domestic producers would not rely
solely on manual activation of safety features to avoid explosions
within enclosures.  The system envisioned by EPA would be  designed such
that the exhaust fans would be interlocked mechanically, electrically,
or otherwise with the spin cell extrusion pumps.  Should the exhaust
fans fail, fiber would no longer be produced, additional solvent would
not be released into the enclosure, and the enclosure doors would open
automatically.  This would allow dilution and diffusion of the solvent
vapor into the room air.
     EPA does believe that manually opening the enclosure  doors is a
reasonable and dependable backup or failsafe method of preventing the
buildup of explosive vapor concentration.  It should be noted that
workers currently must observe the machines constantly to  respond
immediately to spinning machine malfunctions and "roll breaks," or
"feed wheel wraps."  (These terms, used by two different fiber producers,
both refer to the malfunction in which fiber exiting the spin cell is
wrapped around the godet roll.  The wrap will become larger as more
fiber is wound, and will cause more serious problems if not cut and
removed quickly.)  Thus, workers are always available to open the
enclosure doors should the automatic opening system fail to operate
when needed.
     EPA believes that due to the automatic safety features
that would be designed into an enclosure system, the occurrence of an
exhaust fan shutdown with the simultaneous continued release of solvent
into a closed enclosure is very unlikely.  Should such a situation
                                  2-10

-------
occur, however, EPA agrees with the commenters that workers would need
enough time to respond to avoid an explosion.  The commenters suggest
that the workers may have no more than 30 seconds to manually open the
enclosure doors but do not provide any supporting data.
     To evaluate this claim, EPA calculated the amount of time required
for solvent concentrations to reach the lower explosive limit (LEL)
within an enclosure under worst case conditions.  Data collected prior
to proposal of the NSPS (primarily from commeter IV-D-4) were used to
make the calculations.
     Two situations were evaluated: one is a pilot enclosure system
designed by commenter IV-D-4 and the other is a system for a hypothetical
50 million pound per year plant.  For both situations, it is assumed
that fiber spinning  continues after the exhaust fan stops suddenly and
the enclosure doors remain closed.  EPA's calculations indicate that it
would take 2.4 minutes to reach the LEL at the pilot system and 5.1
minutes to reach the LEL at the hypothetical plant.  There would be an
adequate amount of time in either situation for a worker tending the
machines to manually open the enclosure doors if the automatic door
opener fails.  (See Appendix B for the detailed calculations).
     Thus, EPA believes that the enclosure systems that represent BDT
for the proposed NSPS do not pose any risk of explosion that cannot be
alleviated by proper design and operation.
2.2.7  Comment (IV-D-3, IV-D-4):
     One Commenter (IV-D-4) claimed that the use of enclosures on
spinning machines would negate current fire protection measures.   He
claimed that "enclosure systems connected to a control device would
have unlimited oxygen supply and ready ignition sources." Another
commenter (IV-D-3) noted that the spinning cabinets in use at domestic
filter tow plants are isolated from each other, and cabinet fires are
prevented from flashing over to other cabinets.  He claimed the tow
line enclosure would provide a connection between all  the cabinets
along a spinning line, so that a fire in one cabinet along a spinning
line could ignite all the cabinets.  This commenter also noted the
danger of an "unlimited supply of oxygen," where enclosures are
used.
                                  2-11

-------
Response:
     The comments that an enclosure system would create an unlimited
oxygen supply, ready ignition sources, a.nd a convenient route for fire to
spread from one to several  or all  the spinning cabinets on a line reflect
an incorrect appraisal of a properly designed and fail-safe enclosure
system.
     Before responding, it  is first important to make clear that  fire can
potentially propogate in two ways  in a situation such as this.  In one,
the flame would propogate along the surface of the fiber, the potential
for which would be the same with and without enclosures.  In the  other,
the flame would propogate through  the vapor space, a phenomenon described
here as flashover.
     The commenters in this case appear to be referring to flashover.
This can only occur when there is  a limited supply of oxygen, which causes
vapor concentrations to be  in the  "explosive range."  When there  is too
little oxygen, the vapor exceeds the upper explosive limit (UEL).  When
there is an excess of oxygen, the  vapor concentration is below the lower
explosive limit (LEL).  In  neither case is fire or explosion of the vapor
possible.
     The comments about an  unlimited oxygen supply apparently refer to
the continuous flow of air  drawn into the enclosure by its exhaust system.
The volume of air drawn into the enclosure, however, is established by
the design of the enclosure and exhaust system so that the solvent vapor
concentration is maintained well below the LEL during normal  spinning
operation.  In other words, the vapor in the enclosure could neither
ignite nor support a flame, and, as a consequence, a fire in one  cabinet
could not spread via the enclosure to other cabinets.  The reference to
an unlimited oxygen supply  being a hazard or safety concern is, therefore,
inappropriate.
     The only upset condition germane to the discussion of oxygen supply
or the flashover of fire from one  to other spinning cabinets is the
malfunction of the enclosure's primary exhaust system concurrent  with
continued spinning.  In this situation, the air (oxygen) supply is no
longer continuous nor unlimited but is fixed by the volume of the enclosure,
                                  2-12

-------
If no other safety features alleviate the malfunction, the solvent vapor
concentration might approach explosive limits,  (see the response to
Comment 2.2.6 and Appendix B for a determination of the time required).
However, the enclosure should be designed and equipped with secondary
exhaust systems, alarms, automatically opening doors, and line shutdown
interlocks with the spinning pumps.  In addition to these mechanical
features, operators are always standing by to correct this and other
malfunctions.  At worst, then, the malfunction would cause the safety
mechanisms to create conditions identical to current operating conditions
without an enclosure, i.e., the spun yarn and solvent would be exposed to
the spinning room atmosphere.
     With respect to the commenter's claim concerning an increased number
of ignition sources, no further information was provided.  No system
designs considered to reflect BDT would affect the type or number of
ignition sources already available.  Note, however, that with the enclosure
doors closed, worker access is prevented, and the solvent vapor is physically
separated from the workers, from sparks or flames caused by tool malfunctions,
and from any other ignition sources.
     EPA concludes that enclosures can be designed and  installed to create
an effective solvent vapor capture system while maintaining operating
conditions that  pose no greater risk for fires or explosions than current
operating conditions without enclosures.
2.2.8   Comment  (IV-D-1):
     One commenter requested that  enclosures  for crimpers at filter
tow facilities  should be considered an emission control option to be
used only if necessary  to  meet the NSPS.  He  discussed  difficult operating
and maintenance  problems that would be caused by crimper enclosures.
He also  indicated that  the amount  of residual solvent  remaining on  the
tow by  the time  it reaches the crimper is miniscule.
Response:
     EPA  recognizes  that the  installation of  enclosures on  crimpers
at existing  plants would present  unreasonably difficult operating and
maintenance  problems.   However, EPA believes  enclosures could be
appropriately  designed  and operated for  new  facilities  subject  to the
NSPS so as  to  present no particularly  difficult operational problems.
                                   2-13

-------
     Information provided in Table 6-8 of the BID, which reflects
data received from several manufacturers, indicates that residual
solvent at the crimper would not be "miniscule" but would amount to a
significant percentage of total  VOC emissions from the entire facility.
However, if it is miniscule, control  equipment would not be needed.
It should be noted that the NSPS is an emission limit, not an equipment
standard, and does not itself require enclosures for crimpers.
2.2.9  Comment (II-D-3, II-D-4):
     Two commenters claimed that the use of enclosures would create
worker exposure problems whenever worker access is necessary because
the enclosures would concentrate acetone vapor to levels higher than
permitted by OSHA and internal  guidelines.   The commenters further
claimed that because of "rapid production rates and need for immediate
worker access," it would not be  possible to purge the'enclosure prior
to opening the access doors.
Response:
     The commenters are correct  in stating  that the solvent vapor would
be more concentrated with enclosures  than without.  However, testing at
a plant that uses enclosures (docket item I1-8-99) revealed that when
a door is opened for worker access, substantial  flow of air into the
enclosure causes considerable dilution.   Secondly, the tested plant
uses a spinning solvent that is  far more toxic than acetone (the TLV is
10 ppm vs. 1,000 ppm for acetone) and has effectively dealt with the
personnel exposure problem through the use  of work practices, as described
below.  It is therefore apparent that plants using acetone in spinning
should as well  be able to effectively control  their solvent vapor so as to
avoid personnel  exposure problems.
     One Japanese manufacturer has installed enclosures at filter tow
spinning facilities not only for solvent recovery but also to limit
worker exposure to acetone.   In  that  country,  the exposure limit is 200
ppm, compared to the OSHA limit  in the U.S. of 1,000 ppm.   The enclosures
are always under negative pressure created  by  the exhaust fan.  Tests
showed (docket item II-B-99) that when an enclosure door is opened
there is a large inflow of room  air past the worker and into the enclosure,
                                  2-14

-------
 flushing the  immediate area.  Thus, there is is no need for a special
 system purge  prior to opening an access door, and there would be no
 resulting production delay.
     Where many doors are opened at once, as in the case of a "roll
 break," the concentrated vapor within the enclosure is immediately
 diluted and diffused into room air.  This is the procedure followed by
 the Japanese  acetate filter tow manufacturer, who also reports no
 problems with worker exposure.  This is all the more significant in
 light of the  lower personnel exposure limit.  Therefore, EPA believes
 that worker exposure to acetone vapors can be effectively controlled to
 safe levels.
 2.2.10  Comment (IV-D-1, IV-D-3, IV-D-4):
     Three commenters said that although air management can be an
 effective tool to control solvent concentration pockets and reduce
 worker exposures, its use as an air emissions control technique is
 limited. Therefore, it should be considered only an optional  and not
 mandatory method to be used if the affected facility cannot otherwise
 achieve the NSPS.  Commenters IV-D-1 and IV-D-3 said that the costs and
 energy impacts of treating the large volumes of air in such a system
 should be evaluated.  Commenters IV-D-3 and IV-D-4 said that air
 management has not been successfully demonstrated for domestic fiber
 production areas.
 Response:
     The air management control option to which the commenters refer
 was used by EPA as the basis for the VOC emission level achievable at
 cellulose acetate filament yarn processing facilities.  Although they
 claimed the use of air management is "limited," the commenters did
 not identify or explain its limitations as a control  technique.   The
 standards do not require that air management be used, but this option
 is the best system that EPA found for controlling emissions at filament
yarn processing facilities.   It is currently being used at  a  filament
yarn plant operated by commenter IV-D-1.  Thus, it has been demonstrated
 for the type of facility for which EPA selected it as BDT.   The  increased
costs and energy impacts associated with treating the large air  volumes
                                  2-15

-------
associated with this system were analyzed by EPA (see Chapters 7 and 8
of the BID) and were found to be reasonable.
2.3  SELECTION OF FORMAT OF THE STANDARDS
2.3.1  Comment (IV-D-1, IV-D-3, IV-D-4):
     Two commenters suggested that the format of the standards should be
changed to kg VOC per Mg fiber extruded (the proposed format was kg VOC
per Mg solvent used).  Commenter IV-D-1 suggested that the standards
should be expressed in both formats.   All  the commenters said that many
plants already keep records of solvent loss compared to the amount of
fiber extruded.  Commenters IV-D-3 and IV-D-4 also indicated that the
proposed format would require unjustified expensive monitoring of
multiple process streams and burdensome recordkeepirig because the costs
would be several times higher than the $5,000 estimated by EPA.   They
further claimed that the proposed format  would allow VOC emissions to
increase with increased solvent usage, and commenter IV-D-3 said that
the proposed format would encourage such  an increase.
Response:
     The format for the proposed standards was selected by EPA to
provide maximum flexibility to owners and operators of affected facilities
in determining compliance.  After a careful review of the commenters1
points, EPA concludes that the format should not be changed.  Even
though the format is expressed in units of VOC emissions per unit
solvent used, the procedures for demonstrating compliance in 40 CFR
60.603 allow the option of determining VOC emissions per unit of fiber
extruded using existing plant records, as recommended by the commenters.
To convert the result to units of the standards as expressed in
40 CFR 60.602, a simple multiplication of the solvent-to-polymer ratio
for the affected facility is all that is  required.  Thus, the regulation
accomplishes what the commenters suggested while allowing owners and
operators to choose the procedure that is best suited to the situation.
It also reflects the use of BDT at each facility, whereas an emissions
per unit fiber format would not.  A format of that type would have to
be based on an assumed solvent-to-polymer ratio, which varies from plant
to plant as discussed below.
                                  2-16

-------
     JPA  disagrees  with  the  commenters'  statement  that  the  format  of
 the  standards  would encourage  increased  use  of  solvent  and  higher  VOC
 emissions.   The  commenter  is apparently  assuming that a company would
 change  its  solvent-to-polymer  ratio  to reduce the  ratio of  VOC emissions
 to solvent  used.   Information  provided to  EPA by all of the  fiber
 producers indicates that a company very  carefully  selects a  solvent to
 polymer ratio  for  process  reasons, and that  it  usually  has  a  small
 variability tolerance.   To adjust the ratio  in  order to achieve apparent
 VOC  emission reductions  would  likely adversely  affect fiber  quality.
 EPA  believes that the marginal changes in  apparent VOC  emission reduc-
 tions following  a  ratio  adjustment would not justify such radical
 process changes.
     EPA  does  not believe that the costs of  solvent flow meters and
 associated  recordkeeping are unreasonably  expensive.  Information
 (docket item II-B-38) submitted by several vendors of flowmetering
 equipment shows that totalizing flow meters with an accuracy of
 ± 1  percent over the operating range typically  cost no  more than
 $5,000  each.   Some  equipment of the type required costs as little as
 $2,000  per  meter, including  peripheral equipment.  To be certain that
 the  cost  of metering equipment was not understated, new information
 was  gathered after  receiving the comment (docket item number IV-D-7
 and  IV-D-8).
     This new  information confirms that the totalizing  flow meters cost
 less than $5,000 per unit.   The number of units required at an affected
 facility depends on the  facility's choice of format for reporting, the
 particular  layout of pipes and tanks, and the process itself. However,
 EPA estimates that no more than 12 meters would be needed at a new
 facility to comply with the monitoring requirements (assuming the owner
 or operator chooses to meter all  solvent flows rather than use the
 plant records procedure described above).  At a maximum cost of $5,000
 each, the total cost for the meters  would be $60,000. Recording costs
 would be very small  since only monthly totals would be  recorded.
     Therefore, the maximum cost  for monitoring would be about 1  percent
 of the capital  cost for the VOC emission  control equipment,  which  in
turn  would be about 8 percent or  less of  the capital  cost for new
                                  2-17

-------
facilities.  EPA considers these monitoring costs to be reasonable in
light of the usefulness of accurate monitoring to assure the emission
reductions intended under the standards.
2.4  ENVIRONMENTAL IMPACT
2.4.1  Comment (IV-D-6):
     One commenter questioned the accuracy of the baseline emission
rate and resulting emission reduction values for Model  Plant 2 (dry-
spun acrylic facility).  He claims the only domestic dry-spun acrylic
manufacturing facilities in the United States currently emit at lower
rates than estimated for baseline (32 vs.  45 kg/1000 kg fiber) by
controlling emissions from spin cell  exits.  As a result,  emission
reductions and solvent recovery credits are overstated, he claims.
Response:
     The issue raised by the commenter is  whether EPA should have
assumed that a new dry-spun acrylic plant  would install the same VOC
emission control  equipment currently used  at existing plants.  DuPont
is currently the only domestic producer of dry-spun acrylic fibers.
However, another company could enter the market in the  future using a
process different from DuPont1s.  The main reason that  DuPont captures
the dimethylformamide solvent (DMF) emitted at the spin cell  exits is
to protect worker health.  DMF is highly toxic, with a  threshold limit
value (TLV) of 10 ppm.  DuPont's capture and recovery of DMF results
in a lower emission rate than would occur  at a plant using a less
toxic solvent that could be removed from a work area through room
ventilation and exhausted to the atmosphere.  Therefore, EPA believes
it is appropriate to evaluate the impacts  of the NSPS against the
baseline assumptions that were presented in the BID.
     To respond to the commenter's concerns, however, EPA calculated
the emission reduction that would occur if new dry-spun acrylic facilities
were identical to the DuPont plants.   The  VOC emission  reduction at
acrylic plants that is attributable to the NSPS would be about 1,500 Mg
rather than about 2,000 Mg estimated in Table 8-14 of the  BID.  Even if
the lower emission reduction were accurate, the cost per megagram of
emission reduction would be as much as $266/megagram, instead of the
$200/megagram projected by EPA.  These costs are considered reasonable.
                                  2-18

-------
2.4.2  Comment (IV-D-6):
     The commenter claims that in Table 6-6 of the BID, the Alternative
III emission rate for Model Plant 2 should be 20 kg instead of 18 kg.
This error, combined with the overstatement of the baseline emission
rate (discussed in comment 2.4.1) results in an overestimated emission
reduction potential.
Response:
     The commenter does not indicate the basis for the 20 kg emission
rate; however, confidential information provided earlier by this com-
menter (docket item II-D-94) was apparently the basis of this claim.
Examination of solvent use and emission figures for two plants in the
document reveal only minor variance from the EPA-developed values given
in the BID, even when compared item for item within the totals.   Further,
emission value variation between the two plants described in the confi-
dential document was slightly greater than the variation between the
values for the plants and the EPA values.  Therefore, EPA believes that
stated emission values in the BID reflect the conditions expected at
facilities operating under Alternative III controls.
2.4.3 Comment (IV-D-6):
     One commenter noted an apparent mathematical  error on the model
plant representing Alternative III for dry-spun acrylics (page 6-23,
Table 6-6).  The make-up rate for Regulatory Alternative III should be
43 kg he claims, not 42 as shown, since the constituents are 18  kg
(emissions), 20 kg (nongaseous losses) and 5 kg (residual  in the fiber)
for a total  of 43 kg.
Response:
     The correct value for total  make up at the model  plant is 42 kg
per 1,000 kg fiber as presented in the BID.   The apparent  error
noted by the commenter results from rounding up of the make-up rate
constituent values to whole numbers for presentation  in the BID.   The
value of 43 kg is the total  of the rounded values.
2.4.4  Comment (IV-D-6):
     The commenter stated that the EPA emission reduction  estimate
projected for 1987 is not accurate because there will  be no new  facilities
subject to the NSPS  through  1985.
                                  2-19

-------
Response:
     EPA's growth projections are discussed in Section 2.5 of this
document.  EPA continues to believe that growth will  occur and that
significant emission reductions will result from the NSPS.  The projected
1987 emission reductions are estimates and may be more or less than the
actual reductions that will occur.
2.4.5  Commenter (IV-D-6):
     One commenter noted that emissions for model plants (47 FR 52938,
3rd column, 1st paragraph) and 1987 projected facilities were not
computed with the same capacity utilization; emissions for the former
were computed based on 95 percent utilization, while the 1987 projected
acrylic dry-spun affected facilities were based on 81 percent utilization.
He thus claimed that EPA was "looking for the figure that is largest."
Response:
     The 95 percent utilization rate was assumed for model plants to
reflect normal operating conditions under a full market demand situation.
EPA would be remiss if this condition were not evaluated, since it
will be experienced by some or all new plants at least some of the
time.  On the other hand, there are other periods when plants are not
fully utilized.  A correct overall, long-term perspective on an annual
basis must consider both.  Therefore, a utilization rate of 81 percent
was assumed as an annual average for new plants for a typical year.
EPA was not "looking for the figure that is largest," but was trying to
estimate emissions and costs under high production as well as nominal
conditions.  Using 95 percent instead of 81 percent to estimate emissions
for individual plants did not affect EPA's decisions regarding the
significance of the sources.  Either utilization rate would cause
emissions of more than 1,000 tons per year from an individual plant
operating under baseline conditions.
2.4.6   Comment (IV-D-3):
     One commenter felt that Regulatory Alternative II for acrylic/
modacrylic production facilities  (Model Plants 1 and 3) failed to
include emission reduction requirements for polymer solutioning, or
cutting and baling areas.  This commenter claimed that "solvent losses
                                  2-20

-------
from these areas will typically account for approximately 33 percent
recovery of total solvent lost."
Response:
     The commenter is correct in saying that EPA did not include
emission reduction requirements for these areas.  Model Plants 1 and 3
were developed to characterize wet-spun acrylic and dry-spun modacrylics,
respectively.  These model plants were reviewed on several  occasions by
industry representatives, and appropriate revisions were made to
subsequent drafts; the parameters given in the BID reflect  the most
accurate information available to EPA.
     Emission values for each 1,000 kg of acrylic fiber, as listed on
pages 6-16 and 6-28 of the BID show that filtering/dissolving
("solutioning") amount to only 1 kg of the 40 kg emissions, and that
total solvent lost (which equals makeup) equals 70 kg.  Likewise, the
emission value shown for the cutting/baling area is 1 kg of the 40 kg
emission losses or 70 kg total losses.  Emission values for each 1,000
kg of modacrylic fiber show that emissions from solutioning amount to 5
kg of the 140 kg emissions or 155 kg total solvent loss.  Industry
reports that the solvent residual  in the fiber after drying is very low
(as low as 0.5 percent by weight).   Thus, it is not possible that
substantial amounts of solvent, relative to total  solvent loss, could
be released after the drying stage; the negligible solvent  released in
the cutting and baling area is not sufficiently concentrated to permit
economic recovery.  None of these emission values  approaches 33 percent
of total solvent loss or total solvent emissions.
     EPA did not include emission reduction requirements for these areas
because the emissions are small relative to either total emissions or
total solvent loss, and it would not be technically or economically
feasible to attempt recovery.  However, should there be a new process
development that causes more solvent to be emitted at these points, they
would probably need to be controlled to achieve compliance  with the
NSPS.
2.5  COSTS AND ECONOMIC IMPACTS
2.5.1 Comment Summary (IV-D-6, IV-D-3, IV-D-4);
     One commenter (IV-D-6)  claimed that the 1982  and 1987  production
forecasts for acrylics as given in  Table 7-1 are overstated.   He  further
stated that current capacity will  be adequate through 1987.   Two
                                  2-21

-------
commenters (IV-D-3, IV-D-4) claimed that the in-house regression equations
used by EPA to project growth in the industry are oversimplified, and
do not yield results consistent with traditional  and historically
proven industry forecasting authorities (Stanford Research Institute,
Textile Industries, and the Chemical Economics Handbook).  Citing a
current domestic production capacity utilization  rate of 74 percent
(Textile Organon. July, 1982), commenter IV-D-3 stated that reliable
industry forecasts predict that no new production capacity will  be
needed during the next five years.
Response:
     EPA made its 1983-87 growth projections for  acrylic and modacrylic
fibers and cellulose acetate arid triacetate fibers using both published
sources referenced by the commenters and in-house regression analyses.
Published projections definitely were not ignored.  In fact, the pro-
jections sections of the Synthetic Fibers Production Facilities  BID
(pp. 9-36 to 9-44, 9-60 to 9-69) discuss nearly all  of the sources
suggested by the commenters.  EPA found that many of the projections
reviewed did not cover the time period (1983-1987) and/or variable of
interest.   Where appropriate, EPA did use the published projections as
well as its regression results in performing its  analysis.
     EPA also found that there was quite a difference in the growth
rates forecast by the published projections.  EPA addressed these
variations by selecting a projection range bracketed by a high and low
production projection for all three solvent spun  synthetic fiber
commodities projected in Chapter 9.  These commodities were: acrylics
and modacrylics, acetate textiles, and acetate filter tow.  For  two of
these commodities, acrylics/modacrylics and acetate textiles, the low
growth projection embraced the contention of some of the commenters
that there would be no new plant construction in  the projection  period
(BID, p. 7-2).
     Table 2-2 provides a summary of the projections considered. Both
Textile Industries and Textile Organon have published projections for
domestic consumption of acrylic and modacrylic fibers, but neither
source specifically addresses production.  Projections have appeared in
Chemical and Engineering News but the bases and sources of these projections
are not specified, and the period of "long-term"  1 percent growth is not

                                  2-22

-------
           Table 2-2.   SUMMARY  OF  GROWTH  PROJECTIONS  CONSIDERED
Fiber products
projected
Acrylic fiber
production
Acrylic fiber
demand
Acrylic fiber
-domestic
mill con-
sumption
Acrylic fiber
-domestic
consumption
Acrylic fiber
production
Acetate textile
fibers
Cigarette filtra-
tion tow
Acetate and
rayon domes-
tic fiber
consumption
Acetate textile
fiber produc-
tion
Cigarette filter
tow production
Source*/
date
DRI
Winter 1979
C&EN
12/1/80
Texti 1 e
Industries
2/79
Texti 1 e
Organon
1/81
Regression
analysis,
1982
CEH
11/76
Textile
Industries
2/79
Regression
analysis,
Regression
analysis,
1982
Period of
projection
1978-90
1982-85
"long-term"
1982-87
1979-85
1981-87
1975-81
1975-81
1977-87
1981-87
1981-87
Projected annual
growth rate
(%)
2.3
5.6
< 1.0
1.6
2.7
2.8
to 2.4
-3.5
4.8
-2.0
1.0
to 0.9
4.7
to 3.6
or 7.2




C81)
C87)


('81)
C87)
('81)
C87)
*Full references contained in Chapter 9 of BID
                                  2-23

-------
dated.  Data Resources, Inc.  (DRI), in its publication Chemical  Review
(Winter, 1979), provides production projections through 1987 based on
data through 1978.  DRI's compound growth projection of 2.3 percent from
1978-1990 compares favorably with EPA's linear regression.   The Textile
Industries projection was ultimately used by EPA for the lower bound
projection of acetate fiber production.
     Published growth projections for cellulose acetate and triacetate
textile fibers in Textile Industries do not separate acetate from rayon
fibers and rayon production greatly exceeds acetate fiber production.
Thus, rayon dominates the classification and the projection.  The
acetate textile projections available from the Chemical  Economics
Handbook (CEH) were based on 1975 data.  Newer projections  are now
available from CEH, but these projections still  cover only  the first
part of the projection period.   Currently, CEH projects very modest
growth in total acetate and triacetate fiber production through 1984.
Acetate textile fiber production is projected to decline, but increases
in cigarette filtration tow production are expected to slightly outweigh
this decline.  The growth projections EPA ultimately used,  in conjunction
with current capacity utilization, resulted in forecasts of no new
acetate textile fiber plants under either the low or high growth  estimates
and two to four new cigarette filtration tow plants under the low and
high growth estimates, respectively.
     The regression estimates employed by EPA to supplement published
projections have been constructed in accordance with accepted methodology.
The regression methodology employed is discussed in Chapter 9 of  the
BID, and, for the purposes of projecting emission potential, is as
complete and sophisticated as is possible given data and resource
availability.
2.5.2 Comment (IV-D-1):
     One commenter contended that, while the demand for domestically
produced cigarette filter tow is likely to grow over the projection
period, additional production of that commodity could be met by shifting
acetate textile fiber capacity  to filter tow production. This would be
feasible because the commenter  projects a decline in acetate textile
fiber demand and production.
                                  2-24

-------
 Response:
      While EPA does  not  project  any  new  acetate  textile  fiber  plants,
 its  examination of recent  production  levels  (1975-1979)  indicated  a
 stabilization  in production  of the product after dramatic  production
 declines  in  the 1970-1975  period.  By  1980,  most of the  excess  capacity
 in the acetate textile fiber sector  had  been eliminated.   (Capacity
 utilization  in 1980  was  over ninety-two  percent.)  Moreover, the time
 trend in  production  over the 1975-1979 period showed  a slight  increase
 (one  percent per year).  As  a result,  EPA concluded that during the
 projection period, excess  acetate textile fiber  capacity would  not be
 available  to augment the current cigarette filter tow capacity.
 Consequently,  any substantial increase in the demand  for production of
 cigarette  filter tow would necessitate new plant construction.
 2.5.3 Comment  (IV-D-6):
      One  commenter claimed that the credit for recovered solvent was
 overstated because the full  market price was  used in  calculating the
 credits. The commenter explained that  a  fiber producer who also is a
 manufacturer of  the  solvent  in question will   not get  the credit for
 full  savings,  but only for the actual  costs  of manufacturing and
 transportation.
 Response:
      The decision to use the market prices of  solvents to calculate
 recovery credits was based on generally accepted economic principles.
 In a  competitive  setting, market prices represent the opportunity cost
 of resources.   Internal  accounting valuations often reflect sunk costs
 or market  structure phenomena which,  while real enough to the firm, do
 not best reflect the value of a good  to the economy as a  whole.  The
 economic concept of opportunity cost  as reflected in market prices
 should be the basis for an economic analysis  of a standard.  To argue,
 as these commenters have, that the internal  opportunity cost of their
 solvent production equipment  is  zero  is in  conflict with  the premium
 (as embodied in price)  that the external  market appears to  place on
that  equipment.
2.5.4  Comment  (IV-D-1,  IV-D-3,  IV-D-4):
     The commenters noted that capital  costs  for  model plants  did not
include costs for utilities and  polymer manufacturing. They indicated

                                  2-25

-------
that  if  present domestic acetate fiber plants run at full capacity,
there would likewise be no excess polymer capacity available to supply
a  new filament or cigarette tow spinning facility.  Commenter IV-D-1
cited the high cost of polymer manufacturing equipment (stainless steel,
etc.) and utilities such as coal-fired steam boilers, which when added
to the cost of a new spinning facility, would drive the product cost
beyond the competitive range with substitute fibers or imports.  Another
commenter (IV-D-4) said the actual  cost to a manufacturer for increased
fiber production should include the cost of new polymer manufacturing
capacity.
Response:
      The commenters are correct that the model  plants for cellulose
acetate fiber production do not include polymer manufacturing capacity
(e.g., acetate flake capacity), and the capital  costs of polymer
production facilities are not included in new plant costs.  This does
not mean that polymer costs are ignored.   In the analysis, market prices
are used to estimate polymer costs  to new plants.   In a competitive
setting, market price represents the economic opportunity cost of any
resource.  In particular, market price covers all  the costs of production,
including payments on principal  and interest, and  therefore embodies
the capital  costs associated with the production of the polymer.
Estimating future polymer prices therefore involves more than just
forecasting the cost of future plants.   If the  polymer industry is
operating near capacity (as is claimed),  then the  current market price
used in the analysis may actually reflect a price  premium corresponding
to the additional  cost of bringing  new capacity  on line.
2.5.5  Comment (IV-D-3, IV-D-4):
     Two commenters claimed that the baseline capital  cost for a new
filter tow plant should be $104 million instead  of $67.1  million shown
in the proposal  BID.  Commenter IV-D-4 said that,  as the builder of the
last filter tow plant in the world,  his company  is uniquely qualified
to provide cost data to EPA and had provided actual  cost  information
documenting a $104 million  estimate.
Response:
     We agree with commenter IV-D-4 that  his company is well  qualified
to estimate capital  cost for new filter tow plants.   The  information
                                  2-26

-------
 supporting the $104 million  estimate was  carefully  considered  by  EPA  as
 was  other information  provided  by  his company  and other  companies.  In
 response to an earlier similar  claim, EPA explained in a letter to  com-
 menter IV-D-4 (docket  item II-C-138), how the  $67.1 million was derived
 and  why it was considered  to be the  best  estimate after  evaluating  all
 available information.   The  commenters have  not  provided any additional
 cost data to support their comment.   Therefore,  EPA believes the  capital
 cost estimates for  new  filter tow  plants  should  not be changed.
 2.5.6  Comment (IV-D-3):
      One commenter  said that, based  on a  survey  of  actual costs,  the
 additional  capital  cost required to  achieve  Alternative  II control  at
 acrylic/modacrylic  production facilities  (Model  Plant 1)  is understated.
 Where the Agency estimated the  Alternative II  incremental capital  cost
 at $3.8 million, the commenter  estimated  it  to be $6.3 million.
      For Model  Plant 4,  the  commenter disputed the  incremental  capital
 cost estimates contained in  the BID  associated with  the  increased
 control  from Alternative II  to  Alternative III.  Specifically, the BID
 estimates an incremental capital cost of  $1.2 million; the commenter
 estimates  these same costs at about  $7.4  million, and claims EPA has
 understated  these costs  by 516  percent.
 Response:
      Because some of the text and a  related table provided by this
 commenter  made confusing and inappropriate comparisons,  EPA contacted
 this  commenter and asked for clarification.  However, the commenter did
 not  believe  any revisions were necessary.
      The  commenter's claim concerning incremental capital costs for
 Model  Plant  1, Alternative II were not supported by  any  information or
 data.   For Model Plant 4, the commenter has inappropriately compared
 incremental costs to arrive at the 516 percent value above.   He correctly
 shows  the BID-reported incremental  capital costs between  Alternative II
 and Alternative III  at $1.2 million.   He  claims this value should  be
 $7.4 million, but the table of costs  provided by the commenter  shows
that $7.4 million is his estimate of  the  additional  cost  required  above
baseline to meet Alternative  III control.   The commenter  has  thus
                                  2-27

-------
 incorrectly compared a  baseline to Alternative III  incremental cost
 ($7.4  million) with the EPA Alternative II to III incremental cost
 ($1.2  million).
     Further  it  should  be noted that the commenter  has reported 1981-based
 costs, and is comparing these costs directly with 1980-based costs.
 Application of an inflation factor of about 10 percent for that period
 would  significantly narrow the difference between the BID values and
 the commenter's  stated  values.  The commenter's 1981 value of $7.4
 million, if reduced by  about 1C) percent for inflation, would be $6.7
 million for comparable  1980 dollars.  EPA reports (BID Table 8-3) the
 incremental cost of control above baseline to Alternative III at $4.2
 million.  The appropriate comparison is therefore $4.2 with $6.7 million,
 a difference of  37 percent, not: 516 percent.
     EPA has received capital cost information from six plants that
 manufacture acrylic and/or modacrylic fibers.  The  reported costs for
 relevant process areas  and items of control  equipment at the various
 plants did not cluster  about a single value, but rather showed a range
 of values.  EPA  recognizes that; there are legitimate reasons for the
 variations, such as non-process related options, materials and quality
 of construction, and variations; in the processes themselves.  The values
 presented in the BID will  therefore not necessarily reflect any given
 plant's costs exactly;  EPA believes that the BID-presented costs are
 reasonably valid representations of costs at new or reconstructed
 facilities.
 2.5.7  Comment (IV-D-3, IV-D-4):
     Two commenters claimed that because the baseline capital  cost for
a new filter tow plant was underestimated by EPA,  a re-evaluation
of the cost-effectiveness  using $104 million as the baseline capital
cost would show the standards to be unduly burdensome.   Commenter
 IV-D-3 also suggested that the same would be true  for acrylic/modacrylic
fiber facilities.
Response:
     The baseline capital  cost referred to by the  commenters is the
 amount needed to build a new facility  in the absence of the NSPS.   It
 is used as a "baseline" against, which  to compare the costs associated
with control  equipment to  achieve the  NSPS.   Thus,  it does not affect

                                  2-28

-------
 the cost-effectiveness of the NSPS.   Cost-effectiveness is  a comparison
 of  the increase in annualized costs (beyond  baseline)  to achieve  the
 NSPS to the increase  in the amount  of emission reduction  (beyond
 baseline).   Thus,  while the amount of baseline  capital  cost  may  affect
 a company's decision to build a new  plant,  with or  without  an  NSPS in
 effect, it  does  not affect the cost-effectiveness of  the NSPS.   A
 higher baseline  facility cost does result  in  air pollution  control
 costs becoming a smaller percentage  of the  total  cost of producing a
 product.
 2.5.8  Comment  (IV-D-3, IV-D-4);
      Two  commenters said that filter tow producers  could not afford to
 build a new filter tow production facility  because  the  high  capital
 cost of a new plant would require a  price  increase  of 50 percent if
 EPA's capital cost estimate for a baseline  plant  of $67.1 million  is
 used, or  75 percent if the commenters1  estimate of  $104 million  is
 used. The commenters claimed  that there would be  no new plants built,
 therefore,  particularly since current  prices of domestically produced
 filter tow  are under serious  pressure  from  foreign  competition.
      Commenter IV-D-3  suggested  that  substitutes  for  filter tow might be
 used  by cigarette  manufacturers  if the  price of  filter  tow increased
 significantly.   He also recommended  that EPA recalculate the growth
 forecast  for filter tow by  incorporating the 40  percent  export market,
 considering the  effect  of  price  increases on the export market, and
 considering the  excess  capacity  currently available.  He also took
 issue with  the EPA position that the demand for cigarettes is highly
 inelastic.  The  commenter  felt that this assertion was  disputed by
 recent  significant  decreases  in cigarette consumption in both the
 United  Kingdom and  West Germany after the imposition of excise taxes.
 In addition, the commenter cited the American Tobacco Institute estimation
 of a  3  to 8 percent  decline in 1983 domestic consumption of  cigarettes
 following the imposition of an $0.08 per pack excise tax.
 Response:
     Acetate filter  tow manufacturers have  a legitimate concern about
the international competitiveness of  new facilities.  Foreign exports
make  up about 40 percent of their market.   It should be  noted,  however,
that the high implicit prices for filter tow from new  facilities
                                  2-29

-------
originates almost exclusively from the cost of the facility, not from
the cost of air pollution control.  To the extent that foreign producers
share the conditions that have resulted in such an increase in the
baseline cost of cigarette filter tow from a new facility,  the export
market for domestic tow producers will probably remain strong.
     Interaction between domestic and international  makers  of cigarette
filter tow also depends on differing demand trends across markets.  In
particular, market penetration of filtered cigarettes  is lower abroad
than in the United States and will probably continue to increase outside
the United States.  Increased consumption  of filtered  cigarettes will
result in increased demand for filtration  tow.  Thus the world market
for tow will probably continue to grow.
     Since available data and resources did not permit detailed analysis
of the demand and supply relationship discussed above, estimates of the
net effect of potential positive or negative influences on  the compet-
itiveness of domestic tow producers in the international  market could
not be made.  EPA therefore selected the neutral  assumption that domestic
manufacturers would maintain their market  share.
     EPA's position that price increases for cigarette filter tow will
have little effect on the export market derives in part from an assumption
that the price elasticity of export demand is similar  to the price
elasticity of domestic demand.  The estimate of the  price elasticity of
domestic demand for cigarette filter tow was derived from the price
elasticity for cigarettes themselves.  Cigarette demand is  generally
found to be highly price inelastic (see Source 90, BID).  The demand
for cigarette filter tow is estimated to be even more  inelastic because
it is only a small part of a cigarette cost.  Accordingly,  an increase
in cigarette filter tow price is not expected to have  a major impact on
secular growth in acetate filter tow consumption.
     The EPA analysis does not ignore the  possibility  of substitution
in cigarette filter production, but it does judge such substitution to
be of limited potential.  It is believed that cigarette manufacturers
experiment with alternate materials principally for  longrun purposes,
especially for incorporation in new brands.  Since filter design affects
                                  2-30

-------
 the  taste  and  therefore  the  desirability  of  a  cigarette, manufacturers
 are  extremely  hesitant to  substitute  filter  materials  in existing
 brands  (p.  9-91  of  BID).
      It  is  recognized that the  U.S. position in  the world  filter tow
 market  is not  what  it once was.  However, the  bulk of  the  decline in
 U.S.  market share occurred some years  ago.   Throughout the 1970's the
 United  States  was responsible for 54  percent of  world  filter tow produc-
 tion.   According to the  most recently  available  data,  the  U.S. share
 appears  to  have  been sufficiently stable  over  the last decade to
 justify  an  assumption of continued maintenance of its  world market
 share.
 2.5.9  Comment  (IV-D-3,  IV-D-4, IV-D-6):
     One commenter  (IV-D-6) said that  EPA's capital cost estimate of
 $70 million for  a baseline acrylic/modacrylic  plant was too low.
 Commenters  IV-D-3 and IV-D-4 said that the cost  estimate for the
 cigarette filter tow plant was too low.  All the commenters concluded
 that the underestimated capital  costs  resulted in implicit price
 estimates for  the fibers that were too low.  They suggested that higher
 capital cost estimates would result in correspondingly higher implicit
 fiber prices that would remove any incentive to  build  new  plants.
 Response:
     EPA does  not believe  that larger  new plant  capital costs would
 dramatically change the results  of the implicit price analysis.   This
 is because  so much  of the  product cost is related to the cost of
 operating inputs, and not to capital  equipment  (pp. 9-72 and 9-85 of
 the BID).   In order to test this contention, EPA ran the implicit price
 model with the capital  cost values suggested by the commenters.  It  was
 found that the implicit price of acrylics increased only four to nine
 cents over the baseline price of $2.45/kg; the  implicit price of cigarette
 filter tow increased fourteen cents  over the baseline analysis  price of
 $3.47.
 2.5.10  Comment (IV-D-3):
     While granting  that  it may  be possible to  design adequate  safety
controls and monitors for enclosure  systems  at  filter tow  facilities,
one commenter stated that the Agency  has failed to  adequately account
 for the incremental  cost  of such safety equipment systems.   He referenced
                                      2-31

-------
the December 8, 1981, memorandum titled, "Discussion of Potential Safety
(Explosion) Problem Associated with the Use of Enclosures (Synthetic
Fibers NSPS, ESED Project 80/15), (docket item II-B-86) and noted in
his comment that equipment costs associated with the prevention of
excessive risks were, "substantially greater than the total  direct cost
of $3.2 million estimated by EPA for the Alternative II control system
stated in the BID, Page 8-10."  He quoted the memorandum to read, "the
equipment costs associated with the prevention of excessive risk are
...5 percent of the total capital cost for a new spinning process."
Response:
     The equipment cost referred to in the December 8,  1981, memorandum
was estimated to "amount to less than 5 percent of the  total capital
costs for a new spinning process."  The wording was incorrect.  It
should have read "amount to less than 5 percent of the  total control
equipment capital  costs above the baseline level  for a  new spinning
process", not the total  cost of a baseline plant.  The  commenter, in
applying the 5 percent to the total  capital  cost for a  new spinning
process EPA baseline plant, arrived at a safety equipment cost of $3.4
million (.05 x $67.1 = 3.4).
     Based on this, the commenter concluded  that the $3.4 million
safety equipment cost represented a "substantially greater"  cost than
the total direct cost of $3.2 million estimated by the  EPA for the
Alternative III entire control  system, as stated in the BID, Page 8-10.
The incremental capital  costs above the baseline level  for an acetate
filter tow manufacturing plant (EPA model plant 4) are  estimated to be
$3.0 million, which represents 4.5 percent of the total  capital cost
associated with a  baseline plant (Table 8-3  of BID).  Consequently,
the safety equipment costs are $0.15 million (0.05 x 3.0 = 0.15) and
not $3.4 million,  as was previously interpreted from the cited memorandum.
2.5.11  Comment (IV-D-3);
     One commenter noted that the capital investment required to achieve
emission reductions at polymer solutioning and cutting/baling areas in
acrylic or modacrylic fiber plants would be  $8.8 million.   He noted
that EPA had failed to include emission reduction requirements for
these areas.
                                  2-32

-------
 Response:
      As shown in the response to comment number 2.4.6,  the standards  are
 not based  on control of emissions from these areas  because they are
 relatively small compared to total  solvent emissions,  and  more difficult
 or expensive to capture relative to other emission  points  in  the same
 plants.  Because these emissions were not considered to be controllable,
 no capital  or other costs were associated with  these emission points.
 2.5.12  Comment (IV-D-3):
      One commenter indicated that Agency-generated  operating  costs
 associated with the regulatory alternatives  are underestimated.   This
 was supported,  in the commenter's opinion,  by the Agency not  including
 the following two factors in its  cost  estimates:
     (1) the additional  cost of the water required  to strip the  solvent
         from the scrubber system and  to  maintain the solvent  level in
         the fiber at or within product  or  process  specifications,  and
     (2) the cost  of the additional  horsepower  required to operate  the
         large  scrubber  system.
 Response:
      Although  not  specifically  delineated, the  incremental  costs  for
 the  items mentioned  were included in the  Regulatory Alternative II  and
 III  costs of increased control  in Tables  8-4 through 8-8 in the BID.
 Docket  items  II-B-32,  II-I-3,  II-B-41, II-D-66, II-D-68, and II-D-22
 provide complete  information on the  development of control  costs for the
 factors  noted by .the  commenter.
 2.5.13   Comment  (IV-D-6):
     One commenter claimed that the  operating conditions for scrubber #2
 at a dry-spun acrylic fiber plant were incorrect.  This scrubber is
 used to  control and  recover solvent emissions from the  steaming/drying
 portion  of the process,  and requires 2,000 kg/h  of demineralized water
 to achieve 98 percent scrubber efficiency, according to Table  6-6 in
 the BID.  However, the commenter claims that engineering calculations
 performed for a similar scrubber indicate 3.5 times  as  much water would
 be required to achieve 98 percent efficiency.
Response:
     The amount of water estimated for the scrubber  was  based  on
information from several  manufacturing plants relevant  to gas  flow,
scrubbing liquid flow, scrubbing efficiency,  and other  parameters.   As

                                  2-33

-------
well, manufacturers of scrubbing equipment were contacted  for their
assistance in determining correct parameters,  and  appropriate informa-
tion was given to a firm that generates  computer-aided  scrubber parameters
to obtain its estimate.   All  the above were considered  in  determining
the model plant parameters the commenter questions.   Assuming the
commenter is correct, however, the increase in water  usage would have a
minimal effect on costs  for utilities (an increase of about $2,900,  or
less than 1 percent of the $319,200 utilities  cost in Table 8-5).
2.5.14  Comment (IV-D-6):
     One commenter emphasized that a large portion of the  current
acrylic fiber production was  and is being exported (29  percent in  1979,
25 percent in 1980, 28 percent in 1981).  He claims that should the
domestic demand increase to the point where total  demand approaches
total capacity, the industry  would not consider new capacity, but
instead simply reduce the export percentage to accommodate the domestic
demand.  This is asserted because, he claims,  the  domestic market  is
"more attractive."  Thus, the commenter  concludes, there is no need for
new capacity, even though total  production may soon approach existing
capacity.
Response:
     EPA performed an analysis using historic  data on domestic and
export acrylic shipments to determine whether  the  export market was  a
residual market as the commenter contended.  This  analysis is described
in docket item II-B-84.   It shows that the data suggested  that domestic
shipments and export shipments tend to move in the same direction,
increasing and decreasing together.  In  addition,  there is no correlation
between the changes in the annual observations of  the variables contained
in the two time series.   These facts suggest that  the export market is
not simply a residual market  for the acrylic producers, that will  be
curtailed when domestic  consumption of acrylics increases.
2.6 GENERAL
2.6.1 Comment(IV-D-6):
     This commenter noted that Table 1-1 in the proposal BID expressed
the expected Alternative II and III emission reductions as being a
"moderate" impact, while an earlier draft of the BID  (July 1981)
                                  2-34

-------
 expresses  a larger reduction but  characterized  this  as  being  a  "small"
 impact.   "There  is an  obvious  inconsistency here  in  the assessment
 process,"  this commenter  claims.
 Response:
      The  first draft  (July  1981)  was  in  error;  an 8,500 Mg per year
 emission  reduction is  not a small  impact.
 2.6.2 Comment  (IV-D-6):
      One  commenter pointed  out what he felt to  be an  inconsistency
 between the Regulatory Alternative II 1987 projected  emission reduction
 shown on  page 6-12 and 7-2  (31 percent to 44 percent) and the Alternative
 II  emission reduction  (31 percent to  47  percent) stated on page 52936,
 3rd column, last  paragraph  of  the Federal Register notice.  These
 reductions  are expressed  for dry-spun acrylic and cellulose acetate
 filter tow  production  facilities, respectively.
 Response:
      The 31 percent to 44 percent range  noted refers  specifically to
 reductions  achieved at the two type plants projected to be constructed
 by  1987, acrylic/modacrylic and cellulose acetate filter tow.  Emission
 reductions  for Alternative  II shown in table 6-2 are for ^1_I_ the model
 plants developed,  and  not only for the projected facilities.  The
 correct emission  reduction range associated with all  the model plants
 developed for Alternative II is 31 percent to 55 percent.  This percentage
 reduction range as shown on Table 6-2 is consistent with table 6-3,
 "Summary of Control Options and Regulatory Alternatives."
      The emission  reduction range noted by the commenter on page 52936
 of  the Federal Register notice should reflect the emission reductions
 achievable  through the use of enclosures when applied to certain process
 stages of the model plant types that can utilize enclosures (that is,
 all but Model Plant 5, for acetate filament.)  The correct range at
 these plants as derived from Table 6-2,  Table 7-1, or as noted the
 table on page 52938 of the Federal Register proposal  notice is
 31-44 percent.  Thus,   the range shown  on page 52936,  31-47 percent, is
 indeed in error.
 2.6.3  Comment (IV-D-6):
     In reference to Chapter 8.0,  Cost Analysis, one  commenter claimed
that the annual  cost figures for two  acrylic/modacrylic  fiber plants
 (baseline,  Alternative II, and  Alternative III)  presented  in  Table 8-14,

                                  2-35

-------
Projected 1987 Cost Effectiveness of Regulatory Alternatives," page 8-21,
do not correspond to the Model  Plant estimates in Table 8-11, "Regulatory
Alternative II Cost Effectiveness Compared to Baseline Costs," page 8-17,
and Table 8-12, "Regulatory Alternative III Cost Effectiveness Compared
to Baseline Costs," page 8-18.
Response:
     The appropriate projected  1987 Model  Plant annual  cost figures for the
baseline, Alternative II, and Alternative III estimates found in Table
8-14, were not derived directly from Table 8-11 and Table 8-12.
For each selected model plant,  the value for "Total Annual  Expenses"
from Tables 9-34 and 9-36, less polymer costs and adjusted  for
81 percent capacity utilization,  was multiplied by the projected number
of new affected facilities for  the years 1982 through 1987, to reflect
the associated growth scenario.
     Thus, the values in Table  8-14 reflect expected 1987 utilization
rates and resultant costs; the  values in Tables 8-11 and 8-12 reflect
the model plants and their associated costs, all of which were based on
95 percent capacity utilization.   (See Response 2.4.5.)
     Because of rounding of solvent recovery credits, cost-effectiveness
estimates for Alternative II and  Alternative III and the totals associated
with the high growth scenario have been slightly affected.   These minor
modifications are reflected in  the revised Table 8-14 and associated
text.  (See Appendix C.)
2.6.4  Comment (IV-D-6):
     One commenter noted that the summary  paragraph on capital  and
annual costs (47 FR 52939) was  not found in the BID, and that it was
difficult to trace the development of the various figures.
Response:
     Several reported values on Tables 8-14 involve compounded rounding.
Direct comparison of this table in the BID to the specific  values in
the Federal Register notice is  not possible.  The values in each are,
however, correct and comparable when rounding is accounted  for.  Appendix
C includes for convenience a revised Table 8-14 and the accompanying
text.  Because of the difficulty  this commenter had in tracing some of
the figures in the Federal Register notice, the derivations and corrected
values are provided below:

                                  2-36

-------
      Capital  costs of installed  control  equipment for Regulatory
 Alternatives  II  and III  are  $18.0  and  $27.8  million  respectively.   From
 Table 8-3,  these values  are  derived  by adding  twice  the  Model  Plant 2
 capital  costs  and four times the Model  Plant 4 capital costs,  for  each
 alternative.
      The capital  costs of  implementing Alternative II at individual
 projected plants is shown  on BID Table 8-3 as  $3.0 million;  for Alterna-
 tive  III the  costs range from $4.2 to  $5.5 million.
      "Annualized costs"  for  operating  control  equipment  at all projected
 plants should  instead read "annual operating costs,"  and are derived
 from  BID Table 8-3 by adding twice the annual  operating  costs  for  Model
 Plant 2  and four times the annual  operating  costs for model  plant  4.
 The corrected  values  thus  determined are  $3.8  million  for Alternative
 II and $5.4 million for  Alternative  III.
      The values  for "additional  solvent recovered" read  for  Alternatives
 II and III  respectively, $3.8 million  and 6.2  million  (in the  FR notice);
 the correct values  should  be $3.89 million and  $6.20  million.  These
 values are  also  found on the revised BID Table  8-14,  in  Appendix C.  In
 the Federal Register  notice,  the "Net Annualized  Costs"  of $1.2 million
 to the industry  for either alternative, and  individual plant costs
 of as  much as  $0.2  million are found on the  revised and  original
 Table  8-14.
     The  phrase  "Net  Annualized Costs" as used  in the table  and the
 Federal  Register  notice means  "increase in net  annualized costs above
 baseline."
 2.6.5  Comment (IV-D-3,  IV-D-4):
     Two  commenters agreed with EPA's decision  to exempt modified
 facilities from the NSPS.
 Response:
     No  response  necessary.
 2.6.6  Comment (IV-D-2):
     The commenter  supports EPA's decision to delay development of an
NSPS for viscose rayon.
Response:
     No response necessary.
                                  2-37

-------
2.6.7  Comment(IV-D-3):
      One commenter requested that information provided during development
of the NSPS by three fiber producers be incorporated into the docket.
This information from Celanese Corporation,  E.I.  duPont de Nemours and
Company, and Eastman Kodak Company, is dated 7/3/80, 9/22/81, 11/30/81,
12/2/81, 12/3/81, 12/23/81, and 4/23/82, according to the commenter.
Response:
     These comments were included in the rulemaking docket prior to
proposal of the NSPS and were carefully considered by EPA in  developing
the proposed NSPS.   The issues they raise are fully discussed in the
BID, the preamble to the proposed standards, and  the comment  summaries
and responses in this document.  (See docket entries II-D-52, II-B-89,
II-D-70, II-D-71, II-E-93, II-D-76, and II-D-97.)
2.6.8  Comment (IV-D-6):
     One commenter noted an apparent inconsistency in values  given for
recovered solvent (BID Tables 8-4, 8-5, 8-6, 8-7, 8-8)  and the supposed
origin of these values in the model plant parameters given in Chapter  6.
He notes that if the emission rates and indicated reductions  as shown
for the model plants are multiplied by the expected production rates,
then the resulting  emission figures do not agree  with the recovered
solvent figures on  the Chapter 8 tables.
Response:
     The commenter has erroneously compared  rounded values with values
calculated from more precise basic data.  Tables  8-4 through  8-8 show
recovered solvent amounts, and these precise values are used  for
computing the economic value of the solvent, as reported in Chapter 8.
The derivation of these amounts; is shown in  docket Item II-B-90.  In
other places in Chapter 8, however, these values  were rounded for
simplicity from thousands of kilograms to gigagrams,, then the rounded
values used in subsequent calculations (see  Tables 8-11, 8-12, 8-13,
and 8-14).  The latter (rounded) values were used for descriptive purposes,
not exact evaluations.  Where economic impacts were considered, however,
the more precise values were used (Tables 8-4 through 8-8).
     Also, it should be noted that the emission values  given  for the
various model plants in Chapter 6 have themselves been  rounded to whole
numbers.  Any minor variation in these values, when multiplied by large
                                  2-38

-------
production amounts, would result in apparent real differences.
Emission reductions thus computed may not be as precise as the emission
reductions shown on Tables 8-4 through 8-8 and on docket item II-B-90.
     Thus, it can be seen that the variations are not errors of fact
but rather the result of compounded rounding errors.  No changes to the
figures are considered necessary.
2.6.9  Comment (IV-B-5, IV-B-6):
     A number of editorial  comments were made by two commenters, and
these are summarized in docket item IV-B-2.
Response:
     None of the editorial  comments or changes affect the standards,
and no further response is  necessary.
                                 2-39

-------
2-40

-------
                               APPENDIX A

CALCULATION OF ENCLOSURE CAPTURE EFFICIENCY

    Although not  required  by the standards, EPA has assumed in developing
model  plants for  the industry that enclosures would achieve at least
90 percent capture efficiency.  This efficiency was in part based on
solvent mass balance data  collected at plants where enclosures are now
in use.   In addition, both concentration and flow values at the enclosures
were measured at  one of the plants.  These measurements and the information
obtained  from several fiber producers are the basis for EPA's calculation
of enclosure capture efficiency, as described below.
     EPA's test data indicate that while all enclosure doors are closed,
the continuous exhaust creates negative pressure within the enclosure;
consequently, vapor leakage from the enclosures does not occur.  In
fact,  there is significant inflow of room air into the enclosures
through all available openings, cracks, sheet metal joints, etc.  Thus,
while  operating normally and with doors closed, the enclosure would
exhibit complete  capture and exhaust of any VOC released within the
enclosure.  For this reason, 100 percent capture during periods of
normal operation was assumed for the time periods when the doors are
closed.
    During a portion of production time, one door is open,  and capture
efficiency is then dependent on face velocity into the enclosure through
the opening, whether negative pressure still exists within the enclosure,
and the degree of turbulence at the edges of the door opening.  Testing
performed on enclosures revealed that when one or two doors are opened,
there  is significant flow of room air into the enclosure.   This flow is
great  enough to prevent diffusion of vapor into the room,   (docket
items  II-A-15 and II-B-99)  The only VOC losses at the door would be
very low amounts due to turbulence.  Room air concentration measurements
made 4-5 feet from an opened door did not show any increase beyond
background levels, although the VOC concentration within the enclosure
was significantly greater than  background levels.   As  well, measured
exhaust flows from the entire spin cell  enclosure were greater than  the

                                  A-l

-------
air flow into the enclosure via one or two open doors.   Therefore,  the
remainder still  comes in via the aforementioned cracks,  leaks,  and
other openings.   Since negative pressure is maintained  even with  one or
two doors open,  then it is seen that capture is still essentially
complete.  For the purposes of calculating overall  efficiency,  99
percent will  be  assumed.
     About half  the time enclosures would need to  be  open,  access to
the entire line  is necessary to repair "roll  breaks"  (tow  line  breaks).
Because of this, all doors would be opened initially, then  closed one
by one as fiber  from each spin cell is rethreaded  and adjusted,  (docket
item II-B-76)  [During this period, all  doors would not  necessarily be
left open for the entire repair operation.  All the doors would be
opened initially, then closed one by one as repairs are  made,  until
only one door remains open as the last spin cell  is tended.  It is  then
seen that the average area of opening (number of  doors  open) is half
the entire possible area (all doors open), during  the time  access of
any kind is required.  This concept will not directly be used  in
computations, but is made to further support later  developed values.]
    Following the opening of all the enclosure doors, the concentration
within the enclosures, initially at about 5,200 ppm or  20 percent LEL
as reported by industry personnel (II-C-125 and II-E-86), would quickly
be reduced due to diffusion and dillution, and approach  room air  VOC
concentrations.   The equilibrium concentration reached within  the
enclosure will be at least as great as the room air concentration.
OSHA limits the  room air concentration of acetone  to  1000 ppm,  and
industry personnel report typical room air concentrations at about
800 ppm.
    While the doors are opened and spinning repairs made, the  enclosure
exhausts will continue to operate.  A finite amount of  solvent  VOC  is
captured as long as the exhausts continue to operate, regardless  of
the position of  any of the enclosure doors.  The  exhausted  VOC  concentration
will of course be lower than under normal operation,  but will  be  at
least as great as the room air concentration, shown above as about
800 ppm.  The important point is that the enclosure exhibits some
degree of capture, even with all doors opened.  The amount  of  capture
                                  A-2

-------
can be expressed as the ratio of vapor concentration with the doors
closed to the vapor concentration with the doors  opened.   This ratio
provides an indication of the efficiency of the entire enclosure.
    Concentration of captured vapor with doors opened x 100
    Concentration of captured vapor with doors closed
          PPm   x 100 = 15%
      5200 ppm
     Capture efficiency was also based on information provided by  the
industry concerning the amount of time a hypothetical enclosure would
be opened for required worker access:   "...worker access  to the
individual spin cells is required approximately 14 percent  of  total
production time."  He noted that... during the time when  worker access
is necessary the enclosures would be less effective in capturing VOC
emissions.  During approximately half  of the time access  is required,
it would be necessary to open at least one of the enclosure doors.  The
remaining time would require all of the enclosure doors to  be  open.
[He] further estimated that approximately two-thirds of the time
worker access is required, the spinning operation continues.   The
remaining one-third of the time requires a process shutdown.
    One industry representative (IV-D-3) has indicated that during as
much as 19 percent of total production time, access to the  spinning
machines is required.  (This is significant, since one or more doors
to a hypothetical enclosure would be opened during this period,  and
enclosure capture efficiency would conceivably be reduced.) Combining
this information with the information  in the preceeding paragraph, it
is estimated that all doors would be opened (at least initially) during
9.5 percent (about one-half) of the total  required access time,  and one
or two doors would be opened during the other 9.5 percent of the time.
In addition, of the time access is required (19 percent), spinning
continues for about 12.7 percent (two-thirds) of the time,  and the
process is shut down for the remaining 6.5 percent (one third).
Information is not now available to reveal  how these simultaneous
events overlap (that is, for example,  whether the process shutdown
occur with one door open or with all doors open,  on average).   For the
                                A-3

-------
purpose of computing capture efficiency,  the  two  extreme  cases  will  be
considered; actual  capture efficiency  would fall  between  the  two
extremes.
A.  The first situation would yield the highest capture efficiency:
19%
9.5% One door open
9.5% All doors open
12.7% Spinning continues
6.3% Process
shutdown
81%
/^Doors closed
74Spinning
YAcontinues
81% of the time doors are closed and spinning  continues  ....100%  capture
9.5% of the time one door is open and spinning continues....  99%  capture
3.2% of the time all doors are open and spinning continues..  15%  capture
6.3% of the time process is shut down	
     These values could be used at this  point  to  calculate overall
capture efficiency, except for a further consideration.   Since the
process is shut down for a portion of the total time,  it  is not appropriate
to consider capture or lack of capture during  this  period.  Thus, the
period of shutdown should be eliminated  from consideration, or factored
out in some appropriate manner.  The spinning  process  continues for
about 93.7 percent of total time (100 -  6.3);  this  should be considered
as the available time, during which capture efficiency may be considered:

         93.7 percent of total time = 100 percent available time
         100% = LQ67
    This is the factor by which all other reported vaues  should be
    multiplied to provide corrected values.
    Thus, correcting for process shutdown:

        1.067 x 81.0 percent x = 86.4 percent available time
        1.067 x  9.5 percent x = 10.1 percent available time
        1.067 x  3.2 percent x =  3.4 percent available time
                                  A-4

-------
 These  values  are  then  multiplied  by the  appropriate  capture efficiency
 to determine  overall capture:


     86.4  percent  available  time x 100  percent  capture  =  8640
     10.1  percent  available  time x  99  percent  capture  =  1010
     3.4  percent  available  time x  15  percent  capture  =  	51_
                                                         9701

                                           9701 4  100   = 97%
                                             overall  capture efficiency
 B.   The  second  case  and most conservative value  is determined  as  follows

                                                 19%  81%
9.5% One door open
6.3% Process
shutdown
9.5% All doors open
12.7% Spinning continues
J7/Doors closed
£< Spinning
//continues
    81% of the time doors are closed and spinning continues	100% capture
     9.5% of the time all doors are open and spinning
          conti nues	 15% capture
     3.2% of the time one door is open and spinning
          conti nues	99% capture
     6.3% of the time the process is shutdown	


    Determine the factor for the period of process shutdown:

    93.7 percent of total time = 100 percent of available time

    100%  = 1.067
    93.7

Again, correcting for process shutdown:

1.067 x 81.0 percent total  time = 86.4% available time
1.067 x  9.5 percent total  time = 10.1% available time
1.067 x  3.2 percent total  time =  3.4% available time


The corrected values are again multiplied by the appropriate

capture efficiency to determine overall efficiency:
                                  A-5

-------
86.4 percent available time x 100 percent capture    = 8640
10.1 percent available time x  15 percent capture    =  151
 3.4 percent available time x  99 percent capture    =  337
                                                       9T28
                                  9128 * 100 = 91.3%
This value represents the worst case, most conservative conditions.
For the purposes of model plant parameters and  calcuations relative to
solvent recovery, this value is rounded down,  again  conservatively, to
90 %.
                                  A-6

-------
                                APPENDIX B

 SAFETY CONCERNS WITH THE USE OF ENCLOSURES

     In developing control  options for the standards,  EPA has envisioned
 enclosures  that would capture solvent vapor evolved  from the spun
 fiber, and  vapor pumped or pulled along with the  fiber as it leaves  the
 spin  cell.   In  the absence of enclosures, this  vapor  would  be diluted
 with  room air and exhausted to the atmosphere.  By limiting dilution
 air and controlling the vapor concentration, the  enclosures would  actually
 permit the  solvent vapor within  the enclosure to  be concentrated sufficiently
 to make the solvent economically recoverable using carbon adsorption or
 scrubbing,  but  not so concentrated as to  present  a hazardous condition.
    The amount  of solvent  released into an  enclosure  is  dependent  on
 two sets  of parameters:  those related  to the fiber and  its spinning
 step,  and those related  to  the design and operation of the  enclosure.
 Fiber-related parameters include extrusion  rate,  fiber denier (size)
 and shape,  type of solvent,  the  size  of the  opening at the  base of the
 spin  cell through which  the  extruded  fiber exits, and the pressure
 maintained  within  the spin  cell.   The enclosure-related  parameters are
 air temperature  and  the  relative  pressures in the spin cell  and the
 enclosure.
    To  insure high  quality,  a  fiber manufacturer would attempt to keep
 all of  these parameters constant  for  a  given  product; under  steady-
 state conditions  the  amount  of solvent  released into an enclosure would
 be constant.  For  this reason, it  should  be  possible to design the
 exhaust rate of the enclosure  such that the  solvent vapor concentration
 can be maintained at  any desired level.  One company has claimed that
 the exhaust flow  rate for a  new 50 million pound per year facility
 would  be about 12,000 cfm,  and the concentration would be about 5200 ppm
 or 20  percent of the LEL for acetone.  (II-D-71, II-D-81, II-E-86).
 The obvious  conclusion is that during normal  operation, explosive
conditions would not exist  nor could the solvent vapor in the enclosure
 support a flame.
                                  B-l

-------
    The discussion reduces to the safety of the enclosure only during
upset conditions (for example, when an enclosure exhaust fan fails).
Based on information provided by the industry, the following conditions
occuring simultaneously could possibly create unsafe situations:
(1) the enclosure's primary anc backup exhaust systems fail, (2) the
evolution of solvent vapors into the enclosure continues, and (3) the
enclosure doors remain closed at least until  the lower explosive limit
is reached.  EPA believes preventive measures can be included in the
design of the exhausts, enclosures, spinning  equipment, and safety
mechanisms such that explosive conditions would be avoided.  Obviously,
there could be many variations of this system, but as a minimum the
following are significant design considerations:
    •    The primary exhaust system would be  connected to a warning
         (horn, light, etc.) system.
    •    A backup exhaust system would also be connected to the
         primary system for automatic emergency activation.
    •    The primary and/or backup exhaust systems would be interlocked
         such that failure of one or both would cause the spinning
         machine to shut down, and the enclosures doors to open.
    •    The enclosure doors would be spring-loaded or otherwise set
         to be opened, and pneumatically or mechanically
         held shut under normal  operation.
    o    Employees that constantly tend the machines would recognize a
         malfunction (fan failure) and could  open the enclosure doors
         manually if the automatic system failed.  Estimates show
         they would have a minimum of 2 1/2 minutes before flammable
         conditions developed,  (see page B-5)
    t    Once the doors are opened, any vapor present would quickly
         diffuse into the room.,  preventing any further buildup
         toward the LEL, i.e., the concentration in the area of the
         enclosure would quickly approach that of the spinning room.
    Regardless of the exact design, the general  approach is feasible,
and in fact, representatives of two domestic  companies have noted that
their firms either operate such  systems or have agreed that such designs
are feasible.  In summary, the design of the  enclosure, exhaust fans,
                                  B-2

-------
 spinning pumps,  and safety interlocks  would prevent the vapor concen-
 tration within  the enclosures  from reaching an  explosive or flammable
 level.
     Nevertheless,  some industry  officials  have  questioned the use  of
 enclosures,  and  have claimed that  the  enclosures  would  provide "ready
 ignition sources",  an "unlimited oxygen  supply",  and  would "negate
 current fire protection measures."
     With respect to "ready ignition  sources,"  (see  docket items  IV-D-3
 and  IV-D-4)  no  further information was provided by  the  commenters  to
 identify these  sources.   No changes  are  being considered that would
 affect  the number  or type of ignition  sources already existing.  The
 enclosures do not  increase the risk  associated with additional  ignition
 sources.  In  fact,  since  the enclosure vapor is not flammable,  if  there
 is a source  of  ignition,  then the  flame  could not be  transmitted to
 other spin cells through  the enclosure.
    The existing ignition sources are:   (1) the fiber itself,  (2)  cabinet
 fires,  and (3) static  electricity.   Although these  sources  do  currently
 exist,  the presence of enclosures would  not affect  the  level  of risk
 from these ignition  sources, based on the discussion of  fail  safe
 mechanisms.   Note that with doors closed, worker  access  is  prevented,
 and possible  ignition  sources such as nonapproved tools  are separated
 from the  vapor.
    The commenter's  claim of unlimited oxygen supply may  be a  reference
 to the  continuous flow of  air into the enclosures that is generated by
 the exhaust  fans.   There  appears to  be a concern that in  the event of a
 fire within an enclosure,* air would be continuously provided and would
 thus promote  a still larger fire.  Although the exhaust  fans continuously
 draw room air into the enclosures,  it is difficult to recognize how
 this situation would somehow provide an unlimited oxygen supply.  The
 only entrances for air are via  the  small  opening for spun fiber at  the
 base of the spin cabinet, the cracks between doors and other places,  or
through designed vents.  Because  these are restrictive openings, they
would not provide an "unlimited"  supply of oxygen, but as discussed
earlier, the  continuous flow of air  (oxygen) prevents  the vapor concen-
tration from  approaching explosive  or flammable  levels.   If the exhaust
*Such a fire is assumed to be possible only for the sake of discussion.

                                  B-3

-------
were to fail, the failsafe mechanisms would, by opening the enclosure
doors, preclude the development of explosive conditions and duplicate
currently existing conditions.
Determination of Minimum Time to Achieve LEL
    To insure that the worst-case conditions are examined, however, a
situation in which the above design features nevertheless fail  will be
considered.  The most important factor then becomes how long it will
take for an explosive concentration to develop within the enclosure,
and the length of time it will take for a nearby operator to recognize
the condition and manually open the enclosure door;;.
    To determine the explosion risk under abnormal  conditions,  EPA
has collected relevant information from a number of plants, and has
calculated the minimum time to reach the LEL, given the data received
from the industry.  Because the information describes both real  and
hypothetical cases, two separate calculations were  made.
A.  Determination of Time to LEL Based on Pilot System
    To determine minimum time to reach the LEL, it  is necessary to
determine:  (1)  the mass of vapor released during a given period
(2) the size of the particular enclosure (3) the LEL  concentration
(4) the density of the solvent vapor (5) initial  concentration  within
the enclosure:
    1)   An industry representative reported in earlier communications
         that the dope extrusion rate at a pilot enclosure was  5.2 Ibs/min
         (2.36 Kg/min).   He also reported that 5 percent of the weight of the
         extruded dope would evaporate as solvent between the  spin cell  exit
         and the crimper (the area enclosed).  Thus,  0.05 x 2.36 Ibs/ min =
         0.118 kg/min =  evolution rate of solvent.
    2)   It was also reported in the same communication that the
         enclosure volume was 200 ft3.  (5.66 m3).
    3)   The LEL for acetone, used at this facility,  is about  2.6 percent
         by volume (minor variations would result from temperature,
         changes and levels of 03 and other gasses.)   This also may be
         expressed as 26,000 ppm.
                                  8-4

-------
    4)   The density of the solvent  (acetone) vapor at 80 °F is
         about 0.067 kg/ft3.  (2.37  kg/m3).
    5)   The steady-state concentration is 20 percent LEL = 5200 ppm
    Using the above, minimum time to LEL can be determined for the
enclosure, assuming the exhaust fan  stops suddenly and completely and
that all enclosure doors remain closed.
    Calculate weight of acetone within the enclosure initially:
5.66m3 x 5200 ppm x 2.37 kg/m3 = Q.Q70 kg
    Calculate weight of acetone within the enclosure at the LEL:
5.66m3 x 26,000 ppm x 2.37 kg/m3 = 0.349 kg
    The difference is the additional  acetone required to achieve LEL:
0.349 kg - 0.070 kg = 0.279 kg
    We know the rate of evolution is 0.118 kg/min, therefore it will
take 0.279 •* 0.118 = 2.4 minutes to achieve the lower explosive limit
within this enclosure.
B.  Determination of Time to LEL Based on Hypothetical  Facility
    One company has provided information on enclosures  at a hypothetical
50 million pound/year plant (22.7 gigagrams).  This company has suggested
that although their pilot enclosure has operated at about 8-10  percent LEL,
the optimum concentration would  be about 20 percent LEL.   As  well,
such enclosures would exhaust  enclosure air and  solvent  vapor at about
12,000 cfm (339.8m3).
    1)    These two parameters  determine the mass of solvent  that
         would be released  into  the enclosure,  since  mass x  flow
         rate = concentration.   The vapor at 20  percent  LEL  (5200 ppm)
         exhausted from the  enclosure would  contain the  following mass:
    exhaust  rate  x     concentration     x  density (80°F)  = mass  exhausted
      339.m3  x  5,2000 parts  acetone     x   2.37   kg       = 4.19  kg
       min       1000,000 parts air               iiiT          rnTn

    2)    It  is  also  necessary to determine the volume of  a hypothetical
         enclosure  as a preliminary step to  determining time  required
         to achieve  LEL.   Information provided earlier by one company
         noted  that  a pilot  system  that could extrude 2.36 kg of
         dope/minute  was enclosed,  and the volume  of this enclosure
         was  about  5.66m3.   A plant that produces  50 million

                                 B-5

-------
         pounds per year would extrude 200 million pounds (90.7 gigagrams)
         of dope, assuming a typical 3:1 solvent-polymer ratio.  We can
         further assume about 95 percent utilization of such a plant,
         or 8400 hours/year operation.  The extrusion rates can then be
         expressed as follows:
         90.7 gigagrams/yr = 10,780 Kg/hr = 180 Kg/min

    We can assume that the pilot enclosure volume and extrusion rate
would be scaled up to the 22.7 gigagram million pound/year plant thusly:
    extrusion rate at large plant = enclosure volume at large plant
    extrusion rate at pilot plant   enclosure volume at pilot plant
                       180 kg/min _   x m3
                       2.36 kg/rnin~  5.66 m3
                   x = 432 m3 = total  enclosure volume for a
                                22.7 gigagram/year plant
    3)   Now the time to achieve the lower explosive limit can be
         determined for the full size  plant, given a sudden and
         complete loss of exhaust and  also given that all  enclosure
         doors remain closed.
The initial concentration within was already set at 20 percent of
the LEL; the resulting mass is then determined:
         5200 ppm x 424.5m3 = 2.36m3 pure acetone vapor
At the lower explosive limit of 26,000 ppm, the enclosure would contain:
            26,000 ppm x 424.5m3 =  H.lm3 pure acetone vapor

    Thus, the difference is the additional  vapor required to reach
the LEL:
                         Il.lm3 - 2.2m3 = 8.9m3
The density of acetone at 80°F is:
                               2.37kg/m3
Therefore,  the weight of additional  acetone would be:
                      2.37kg/m3 x 8.9m3 = 21.1 kg
    We have already determined that the evolution rate of acetone
into this enclosure is 4.18 kg/min. Therefore, the time to reach LEL =
21.1 kg •* 4.18 kg/min = 5.1 minutes, or about 5 min.  6 sec.
                                  B-6

-------
 Summary
     The  times  calculated  above,  2.4 minutes  and  5.1 minutes,  both
 show that  there  is  adequate  response  time  for  nearby  operating  personnel
 to  open  the  enclosure  doors,  shut  down  the spinning machine,  or otherwise
 respond  to the malfunctions.
     One  domestic manufacturer of acetate filter  tow has  begun a pilot
 system to  study  the  use of enclosures.  They claim that  because the
 system is  fairly new,  conclusions  should not be  made  with  regard to
 long-term  use  or benefits.   However,  they do report that thus far no
 explosive  situations have developed during the operation of this
 system.
     One  Japanese manufacturer of acetate filter  tow that uses enclosures
 as  envisioned  by EPA was questioned concerning the safety  of their
 enclosures.  Their response  is self-explanatory:
    "Electric power of  the exhaust  fan comes from the  same  power  line
 of  spinning machines.  If power were  to fail, extrusion of acetate
 dope  in  the spinning machines would stop at the  same  time when
 the  exhaust fan  is down.  Therefore there would  not be the development
 of  explosive conditions in the enclosures with no further generation
 of  acetone vapor.
      If  only an exhaust fan were to fail, alarm bell   and light  in the
 control  room tell an emergency in the exhaust fan and a room
 attendant  inform promptly operators in the spinning area of the
 exhaust  fan emergency.  Then operators rush to open all the
 enclosure doors.  To make doors wide open, there will  be no
 development of explosive condition.
     There have never been any operational  and safety problems such as
 fires or explosions resulting from  a build up of the  solvent vapor
within the enclosures since we ran  for more than two  years  with  the
use of enclosures."   [Now almost four years; this correspondence
was received  in April  1982.]
                                B-7

-------
                  APPENDIX C

           REVISED TEXT AND TABLE  FROM
               BID CHAPTER EIGHT
The following pages are revised versions of pages 8-19
through 8-21 of the October 1982 Draft Background
Information Document for Synthetic Fiber Production
Facilities.
                      C-l

-------
                            (BID page 8-19)
 plants  1 through  5 are presented in Tables 8-11, 8-12, and 8-13.  All
 costs,  production levels, emission reductions, etc., are based on
 model plant  parameters described in Chapters 6 and 8.
     Under Alternative II, model plant 1 shows a zero cost effectiveness
 when compared to  baseline, Alternative I; the costs of controls for
 emission reduction are offset by the value of the recovered solvent.
 Model plant  3 shows a net gaii of $182/Mg under Alternative II; the
 additional solvent that could be recovered beyond baseline would
 result  in decreased annualized costs.  Model plants 2, 4, and 5 show
 positive cost effectiveness of $166, $166, and $588 per Mg VOC reduction,
 respectively, when compared to Alternative I.
     Under Alternative III, model plants 1 and 3 show a net gain or
 annual  savings of $350 and $193 per Mg of emission reduction, respectively,
 when compared to  the baseline.  Model plant 2 would experience a zero
 cost effectiveness under this alternative.  However,  model  plants 4
 and 5 would incur positive increases in annual  costs  of $120 and $442
 per Mg  of emission reduction, respectively, above the baseline case.
     Compared to Alternative II, the application of Alternative III to
 the model plants would result in decreased annualized costs of control
 and thus in decreasing cost per Mg of emission  reduction, as presented
 in Table 8-13.
 8.4.2   Projected  1987 Cost Effectiveness
     The projected capacity shortfalls  as presented in Tables 9-20 and
 9-33 of Chapter 9 lead to the following conclusions  concerning likely
 capacity additions by synthetic  fiber producers  by  1987:
     (1) The projected capacity  shortfall arising from the  high growth
 projection  for acrylic and modacrylic fibers  would  support  additional
 plant capacity.   For this analysis,  it  is assumed that capacity is
constructed in increments of model  plant  capacity,  and that plants of
 model plant 2 type would be built.   Two  plants,  each  with 45.36 Gg
capacity, would  be constructed by 1987  since  there  would  be significant
capacity shortfall if only one were  constructed.  These two plants
would each  operate at 81  percent capacity utilization in  1987.
 (Table 9-11 in this BID  indicates that  this capacity  utilization rate
 is well  within the range  of historical  values.)

                                  C-2

-------
                             (BID  page  8-20)
                                                         «
      (2)  The  projected  capacity shortfall  arising  from the  low  growth
 projection  for  acrylic  and  modacrylic  fibers would not support  additional
 capacity.   This  shortfall would likely  be  met  by debottlenecking
 current  production  processes.
      (3)  The  projected  capacity shortfall  arising  from the  high growth
 projection  for  cigarette filtration tow would  support additional
 capacity  by 1987.   Again, it  is assumed that capacity is constructed
 in  increments of model  plant  capacity.  Four plants, each with  22.7 Gg
 capacity, would  be  constructed.   These  four plants would each operate
 at  95  percent capacity  utilization.  A  capacity shortfall of 4.3 Gg
 would  still exist,  but  this  shortfall would not support an  additional
 plant.
      (4)  The  projected  capacity shortfall  arising  from the  low  growth
 projection  for cigarette filtration tow would  also support  additional
 plant  capacity  by 1987  (capacity  that would be constructed  in increments
 of  model  plant capacity).  Two plants, each with 22.7 Gg capacity,
 would  be  constructed.   These two  plants would  each operate  at 95 percent
 capacity  utilization.   Excess capacity of  1.7 Gg would exist.
      (5)  The projected  capacity shortfalls arising from either the
 high or low growth  projections for cellulose acetate textile yarn
 would  not support additional capacity by 1987.
     Based on the above conclusions, comparisons of annualized costs
 per megagram of  emission reduction were made for those plants that are
 most likely to be built in the next 5 years.  All  three regulatory
 alternatives were examined.  Compared to the baseline,  Alternatives II
 and III result in emission reductions of as much as 5.5 and 8.5 Gg/year,
 respectively,  by 1987.  Annualized costs per megagram of  emission
 reduction for typical plants would be as much  as $412 and $200 respectively,
 for Alternatives II  and III.
     Because the Alternative III  increased costs of control  over Alternative
 II are ofsett  by increased solvent recovery, Thus,  there  is  net  cost
 per megagram of emission reduction to the industry  in implementing
Alternative III  over Alternative  II.  Table 8-14 presents  the projected
 1987 cost effectiveness of the regulatory  alternatives.

                                  C-3

-------
                                     (SID  page G-21)

              Table  8-14.    PROJECTED  1987 COST  EFFECTIVENESS
                           OF  REGULATORY ALTERNATIVES
Cel lulose- ^estate
Acryl ic/Moaacryl ic filter ~ow 'otals
Srowtn
Scenario
1982- L987
Numoer of
Affected .
Facilities3
1987
Alternative II
Solvent Recovery
Credit3
(10S5)
Alternative III
Solvent Recovery
Credit8
(10°$)
3asel ine
Annual Cost
(105S/yr)
Alternative II
Net Annual Cost
(105S/yr)
Alternative III
Net Annual Costc
(10SS/vr
Alternative II
Increase in Net
Annual Cost Over
Baseline
(106$/yr)
Alternative III
Increase in Net
Annual Cost Over
Sasellne
(I05S/vr)
Alternative II
Emission
Reduction (Mg)
Alternative III
Emission .
Reduction (Mg)
Alternative II
Cost
Effectiveness
(S/Mg)
Alternative III
Cost
Effectiveness
(S/Mg)
nl9" low high low ngn low
2042 52
I-06 0.0 2.33 1.41 3.39 1.41

2-18 0-2 1.02 2.01 5.20 2.01
108.0 0.0 12.0 64.0 236.0 64.0
1°8'* 0.0 128.3 54.4 237.2 64.4
I08-* 0.0 128. 3 54.4 237.2 S4.4
0.4 ' 0.0 0.3 0.4 1.2 0.4
0.4 0.0 0.3 0.4 1.2 0.4
972 — 4560 2280 5532 2280
2000 — 5480 3240 3480 3240
412 — 175 175 217 175
200 — 123 123 142 ;23
 All  values projected assume 31  percent cauacity utilization  for acryl ic/moaacryl ic
 facilities and 95  percent for acetate filiar  tow facilities.

'Amount of solvent  recovered multiplied by solvent cost,  S1.39AQ 3MF ana
 >0.a2/kg acetone.

 These values inciuae solvent recovery creaits, but do not  include polymer
 See  also Tables  9-34 and 9-37.

 Emission reduction  is from Tables 3-4 cnrougn 3-3, Line  10 "Recovered  Solvent "

                                        C-4

-------
TECHNICAL REPORT DATA
(Please read Instructions on the reverse before completing)
1 REPORT NO. 2.
EPA- 450/3-82-011 b
4. TITLE AND SUBTITLE
Synthetic Fiber Production Facilities-
Background Information for Promulgated Standards
7 AUTHOH(S)
9. PERFORMING ORGANIZATION NAME ANO AOOH6SS
Office of Air Quality Planning and Standards
Environmental Protection Agency
Research Triangle Park, North Carolina 27711
12. SPONSORING AGENCY NAME ANO ADDRESS
DAA for Air Quality Planning and Standards
Office of Air, ;Joise, and Radiation
U.S. Environmental Protection Agency
Research Triangle Park, Nortii Carolina 27711
3. RECIPIENT'S ACCESSION»NO.
5. REPORT DATE
March 1984
6. PERFORMING ORGANIZATION CODE
8. PERFORMING ORGANIZATION REPORT NO.
10. PROGRAM ELEMENT NO.
11. CONTRACT/GRANT NO.
68-02-3060
13. TYPE OF REPORT ANO PERIOD COVERED
14. SPONSORING AGENCY CODE
EPA/ 200/04
15. SUPPLEMENTARY NOTES
        Standards of  performance to control emissions  of volatile organic
   compounds (VOC)  from  new and reconstructed  synthetic fiber production facilities
   are being promulgated under the authority of  Section 111  of the Clean Air  Act.
   This document contains a detailed summary of  the  public comments on the  proposed
   standards (47 FR 52932), responses to these comments, and a summary of changes  to
   the proposed standards.
 7.
                                
-------