>EPA
          National Air Pollution
           Control Techniques
          Advisory Committee

           Minutes of Meeting
        March  17 and 18, 1981
        U.S. ENVIRONMENTAL PROTECTION AGENCY
             Office of Air, Noise, and Radiation
          Office of Air Quality Planning and Standards
          Emission Standards and Engineering Division
         Research Triangle Park, North Carolina 27711

-------
     National Air Pollution
       Control Techniques
      Advisory Committee

       Minutes of Meeting
    March 17 and 18, 1981
U.S. ENVIRONMENTAL PROTECTION AGENCY
     Office of Air, Noise, and Radiation
  Office of Air Quality Planning and Standards
  Emission Standards and Engineering Division
 Research Triangle Park, North Carolina 27711

            April 17, 1981

-------
                         U.  S.  ENVIRONMENTAL PROTECTION AGENCY

             NATIONAL AIR POLLUTION CONTROL TECHNIQUES ADVISORY COMMITTEE

                           Chairman and Executive Secretary
                  Mr. Don R. Goodwin
                  Director,  Emission Standards and Engineering Division
                  Office of Air Quality Planning and Standards (MD-13)
                  Research Triangle Park,  North Carolina  27711
                                                                                 March 1981
                                        Members
Mr. Carl G. Beard II
Director, West Virginia
 Air Pollution Control Commission
1558 Washington Street, East
Charleston, West Virginia  25311

Dr. Eugene M. Bentley III
President, ECO-Labs, Inc.
1836 Euclid Avenue-Room 608
Cleveland, Ohio  44115

Mr. Russell 0. Blosser
Technical Director
National Council of the Paper Industry
 for Air & Stream Improvement, Inc.
260 Madison Avenue
New York, New York  10016

Mr. Robert J. Castelli
Director of Environmental Quality
Ideal Basic Industries
Cement  Division
Post Office Box 8789
Denver, Colorado  80201

Mrs. Janet Chalupnik
Director of Env. Health Programs
Washington Lung Association
216 Broadway East
Seattle, Washington  98102

Ms. Frances Dubrowski
Senior  Project Attorney
Natural Resources Defense Council, Inc.
1725 I  Street, N. W.-Suite 600
Washington, D. C.  20006

Dr. Robert W. Dunlap
Executive Vice President
Environmental Research & Technology, Inc.
696 Virginia Road
Concord, Massachusetts  01742

Ms. Elizabeth H. Haskell
P.O. Box 3903
Martinsville, Virginia  24112
(Member, Commonwealth of Virginia  State Air
Pollution Control Board, Richmond, Virginia)
Mr. Eric E. Lemke
Chief Deputy Executive Officer
South Coast Air Quality Mgmt District
9150 East Flair Drive
El Monte, California  91731

Dr. James M. Lents
Director, Air Pollution Control Division
Colorado Department of Health
4210 East llth Avenue
Denver, Colorado  80220

Mr. Robert A. Moon, Jr.
General Manager, Synthetic Fuels Department
Brown and Root, Inc.
P.O. Box 3
Houston, Texas  77001

Mr. Venkataraman Ramadass
Chief, Engineering Services Division
Department of Environmental Services
Bureau of Air and Water Quality
District of Columbia
5010 Overlook Avenue, S. W.~2nd floor
Washington, D. C.  20032

Mr. William Reilly
Assistant Health Commissioner
 for Air Management Services
Philadelphia Dept. of Public Health
801 Arch Street—6th floor
Philadelphia, Pennsylvania  19107

Mr. William M. Reiter
Director, Pollution Control
Corporate Environmental Affairs
Allied Chemical
Post Office Box 2332R
Morristown, New Jersey  07960

Dr. Claibourne D. Smith
Manager Applied Technology
F & F Department
E. I. du Pont de Nemours & Company
Clayton Building, Concord Plaza
Wilmington, Delaware  19898

Mr. Bruce A. Steiner
Supervising Project Engineer
Armco Steel Corporation
P. 0. Box 600
Middletown, Ohio  45043
                                        11

-------
                               CONTENTS


  I.   INTRODUCTION AND COMMENTS 	Don Goodwin  1-1

 II.   NATIONAL EMISSION STANDARDS FOR ARSENIC FROM COPPER SMELTERS

      A.   EPA Presentations
          1.   Technical Background	Al  Vervaert  II-l
                  Emission Standards and Engineering Division-EPA
          2.   Regulation Development 	Graham Fitzsimons  11-19
                  Emission Standards and Engineering Division-EPA

      B.   Industry Presentation
          1.   ASARCO*, Incorporated 	Edwin Godsey  11-27

      C.   Discussion 	  II-29

      D.   Correspondence
          1.   NAPCTAC Member  William Reiter	  11-33

III.   CONTROL TECHNIQUES GUIDELINES

      A.   EPA Presentation 	Fred Porter  III-l
                  Emission Standards and Engineering Division-EPA

      B.   Correspondence
          1.   NAPCTAC Member  William Reiter 	  III-8

 IV.   CONTROL TECHNIQUES GUIDELINE DOCUMENT FOR VOLATILE  ORGANIC
      COMPOUND EMISSIONS FROM PETROLEUM DRY CLEANERS

      A.   EPA Presentation 	Steven J.  Plaisance  IV-1
                                                TRW, Incorporated

      B.   Industry Presentations
          1.   Patton, Boggs and Blow	Timothy Vanderver, Jr.  IV-16
          2.   Institute of Industrial Launderers  	/	  IV-24
                                              Mervyn Sluizer, Jr.
          3.   International Fabricare Institute ...William Fisher  IV-31
          4.   Van Dyne Crotty, Inc	Duane  E. Early  IV-44

      C.   Discussion 	  IV-45

      D.   Correspondence
          1.   Illinois Environmental Protection Agency 	  IV-50

  V.   CONTROL TECHNIQUES GUIDELINE DOCUMENT FOR VOLATILE  ORGANIC
      LIQUID  STORAGE VESSELS

      A.   EPA Presentation 	Rebecca Sommer  V-l
                                                              GCA
                                   11 i

-------
     B.  Industry Presentations
         1.  Chemical Manufacturers  Association  ..Bruce  C.  Davis   V-13
         2.  Texas Chemical  Council  	Andrew Nickolaus   V-20
         3.  GATX Terminals  Corporation  	R.  w.  Bogan   V-35
                                            (Read by Fred Porter)

     C.  Discussion                                                V-37

     D.  Correspondence
         1.  Chemical Manufacturers  Association  	   V-40
         2.  American Petroleum  Institute 	   V-51
         3.  Illinois Environmental  Protection Agency 	   V-56
         4.  NAPCTAC Member   William Reiter	   V-58

VI.  CONTROL TECHNIQUES  GUIDELINE  DOCUMENT FOR CONTROL OF FUGITIVE
     VOC EMISSIONS FROM  SYNTHETIC  ORGANIC CHEMICAL  AND POLYMER AND
     RESIN  MANUFACTURING EQUIPMENT

     A.  EPA Presentation 	Samuel  Duletsky   VI-1

     B.  Industry  Presentations
         1.  Chemical Manufacturers  Association  . ...J. D.  Martin   VI-9
         2.  Analytical  Instrument Development,  Inc	   VI-18
                                                  F. J. Debbrecht
         3.  Carolina Machinery  and  Supply Company	   VI-26
                                                    Pat Patterson
         4.  E.  I. du Pont de Nemours and Company 	   VI-30
                                                 Thomas Kittleman
         5.  Texas Chemical  Council  	Andrew Nickolaus   VI-59
         6.  Tennessee  Eastman Company 	J. D.  Thomas   VI-69
          7.  Monsanto Company 	J. M.  Schroy   VI-75
         8.  The Fertilizer  Institute	R.  G.  Wells   VI-125

      C.  Discussion	   VI-127

      D.  Correspondence
          1.   Illinois  Environmental  Protection Agency 	   VI-129
          2.  Monsanto Company		   VI-132
          3.   E.  I.  du Pont de Nemours and Company 	   VI-134
          4.   Chemical Manufacturers  Association  	   VI-194
          5.   NAPCTAC Member   William Reiter	   VI-232

VII.  Appendix:   Record  of Attendance		   A-l

-------
                     I.  INTRODUCTION AND COMMENTS

                       Don R. Goodwin, Chairman
              National Air Pollution Control Techniques
                          Advisory Committee

     The National Air Pollution Control Techniques Advisory Committee (NAPCTAC)
held its first meeting of 1981 on March 17 and 18, 1981, at the Royal Villa Hotel
in Raleigh, North Carolina.  Chairman Don Goodwin called the first session
to order on March 17 at 9:00 a.m.  NAPCTAC members in attendance for both
days of the meeting were:

     Mr. Carl G. Beard II                    Ms.  Elizabeth H.  Haskell
     Dr. Eugene M. Bentley III               Mr.  Eric E. Lemke
     Mr. Russell 0. Blosser                  Dr.  James M.  Lents
     Mr. Robert J, Castelli                  Mr.  William Reilly
     Mrs. Janet Chalupnik                    Dr.  Claibourne D. Smith
     Ms. Frances Dubrowski                   Mr.  Bruce A.  Steiner

     Messrs. Robert A. Moon and Venkataraman Ramadass and  Dr.  Robert Dunlap
were unable to attend the meeting.  Mr. William M. Reiter, although present
for the first day, was unable to attend on the second day  of the meeting.

     The following EPA staff members were present for all  or part of the
meeting:

Robert Ajax         Fred Dimmick          Bob Kolbinsky       Dave Stonefield
James Bain          George Duggan         Randy McDonald      Bruce Tichenor
Doug Bell           James Durham          Bill Polglase       Bill Tippitt
Frank Bunyard       Jack Farmer           Fred Porter         Ronald Turner
Mary Jane Clark     Graham Fitzsimons     Roy Rathbun         Bill Vatavuk
George Crane        Don Goodwin           Janet Scheid        Al Vervaert
James Crowder       K. C. Hustvedt        Stephen Shedd       Tom Williams
Stanley Cuffe       Dick Jenkins          Gene Smith

     A copy of the registration sheets of the meeting, which include the
names and addresses of all those in attendance representing the private sector,
is included in the Appendix.

     Media involvement included the announcement of the time, place, and agenda
for the meeting in the February 12, 1981, Federal Register.

     Mr. Goodwin explained the ground rules of the meeting and the order of
the presentations, and requested members of the audience to state their names
and affiliations when asking questions or volunteering  information from the
floor.  He explained that the official record of the meeting is in the form of
a tape recording and that the tapes can be duplicated and made available from
his office.  Contributions to the minutes of the meeting would be accepted
until April 1, 1981, and the minutes will be available  about 30 days after the
meeting.
                                   1-1

-------
     Chairman Goodwin introduced his EPA colleagues at the speakers' table,
Messrs. James Bain, Stanley Cuffe, Robert Ajax, Graham Fitzsimons, James Crowder,
and Al Vervaert, who opened the technical portion of the meeting with a
presentation on a proposed NESHAP regulation for arsenic from copper smelters.

     Because of a national trade show for the dry cleaning industry being held
on the same day as the NAPCTAC meeting, representatives of the dry cleaners
could attend the meeting for only a brief time.  To accommodate the dry
cleaners, the program agenda was modified to permit the dry  cleaning portion
of the program to be held on the afternoon of the first day  of the meeting.
Mr. Goodwin thanked Mr. Edwin Godsey of ASARCO whose presentation was deferred
until after the dry cleaning discussion.

     The Chairman called an early morning session on April 18 to discuss
the arsenic problem at Tacoma, Washington, with the Committee.  Conceivably
the atmospheric arsenic problem at the ASARCO copper smelter, the Nation's
principal arsenic source, could be resolved without EPA spending any more
funds for the development of a standard because the company  has offered to
install equipment to limit emissions to those that would be  allowed by the
EPA regulation, when and if it came into being.

      After  a  lengthy discussion, the Committee concensus was that EPA should:

      1.  Urge the local agency to require that ASARCO enter  into an agreement
that  would  result in the installation of the control equipment within a stipulated
time  frame.

      2.  The  agreement  should be a legal document signed by  corporate officers
of ASARCO so  that the  corporation can be held responsible for failure to
honor the agreement.

      3.   EPA  should  continue to develop a NESHAP for arsenic from copper smelting
as is required  by the  Clean Air Act and the list of hazardous pollutants.

      Mr.  Goodwin  stated  that he would get back in touch with the Tacoma air
pollution  control officer  to recommend a course of action.

      During the two-day meeting, a total of 24 speakers representing both EPA
and  the industries  being considered for regulation appeared  before the Committee.

      At the conclusion of  the final session on March 18, Chairman Goodwin
 announced that  because of  the  lateness of the hour and the need for Committee
members and other participants  to  leave to make travel connections, the review
 of projects in  progress planned for the afternoon would be postponed until
 the next NAPCTAC  meeting.   Mr.  Goodwin declared the meeting  adjourned at
 5:05 p.m.
                                   1-2

-------
                II,  NATIONAL EMISSION STANDARDS FOR HAZARDOUS
                AIR POLLUTANTS FOR INORGANIC ARSENIC EMISSIONS
                         FROM PRIMARY COPPER SMELTERS
                            A, EPA PRESENTATIONS

1.   Technical  Background of Standard Development Project

                             Mr, Alfred Vervaert
                          Industrial Studies Branch
                 Emission Standards and Engineering Division
                    U. S, Environmental Protection Agency
                Research Triangle Park, North Carolina  27711
      Selection of Pollutant

      EPA has been concerned over the public  health  implications  of
 exposure to inorganic arsenic  for several  years.  -During  1976,  the
 Office of Air Quality Planning and Standards prepared  a draft Air Pollutant
 Assessment Report on  Arsenic which concluded that regulation of  arsenic
 might be necessary due to its  potential  as a carcinogen,  but that
 available scientific  data were not sufficient to support  such action.

      Following the promulgation of new source performance standards  for
 nonferrous smelters in 1976, the Natural  Resources  Defense Council
 petitioned EPA for review of the standards because  the standards did not
 include an arsenic emission limit and the  absence of  such a limit precluded
 the establishment of  arsenic emission limits for existing copper smelters
 under Section lll(d)  of the Clean Air Act.   In response to NRDC's concerns
 and because ambient arsenic concentrations around copper  smelters were
 observed to be many times higher than general  urban levels and  because
 copper smelters were  estimated by EPA, based on preliminary estimates,
 to  account for over 60 percent of the total  nationwide emission  of
 inorganic arsenic, EPA initiated work in the fall of  1976 to assess
 arsenic emissions from existing primary copper smelters and to  evaluate
 applicable control technology.
                                 II-l

-------
     In the interim, amendments to the Clean Air Act, adopted in August
1977, required EPA to evaluate whether ambient concentrations of arsenic
observed in the community endangered the public health.  In response, EPA
undertook several studies to evaluate the public health implications of
arsenic exposure.  These included a health assessment, an exposure
study, and a risk assessment.

     Investigations pertaining to the control  of process sources of
arsenic at primary copper smelters were completed in 1978 and our findings
were presented to this committee in July of that year.  At that time, we
indicated that we would return at some later date to report on our
findings pertaining to the significance and control  of fugitive sources
of arsenic at primary copper smelters and that, depending on the findings,
recommendations would be made for the regulation of process sources
alone, fugitive sources alone, or a combination of process and fugitive
sources.

     On June 5, 1980, inorganic arsenic was added to the list of hazardous
air pollutants  (Slide #1)  The listing was based on the Agency's
conclusions that (1) there is a high probability that inorganic arsenic
causes cancer in humans and that (2) there is  significant public exposure
to inorganic arsenic emitted into the air by stationary sources.  These
conclusions were based on documentation developed by EPA on the health
effects associated with human exposure to low levels of inorganic
arsenic and on the results of EPA's exposure study which identified
multiple stationary sources of arsenic and showed that large numbers of
people are exposed to ambient concentrations many times the national
average.  Of the stationary sources of arsenic identified, primary
copper smelters, especially those which process materials of high
arsenic concentration, were found to be the most predominant source.

     Selection of Source Category

     Currently, there are 15 primary copper smelters operating in the
United States.  Combined these smelters have an annual charge capacity
of over eight million tons and are capable of producing nearly two
million tons of copper per year.

      (Slide #2)  Shown on the screen is a listing of the 15 existing
smelters and estimates of the arsenic throughput at each under peak or
full-smelt conditions.  The estimates are based on information submitted
to us by each of the 15 smelters in 1978 on the arsenic content of
copper-bearing feed materials processed at each smelter.  As one can
see, the arsenic content of feed materials processed varies widely from
smelter to smelter ranging from a few PPM to several percent.  Of special
interest is the fact that the ASARCO smelter at Tacoma is essentially in
a class by itself.  As the data indicate, the Tacoma smelter processes
feed materials which contain nearly 4 percent As and is capable of
                                 II-2

-------
                                          A-1

   BASIS FOR LISTING

   •CAUSES CANCER IN  HUMANS
   •SIGNIFICANT  PUBLIC  EXPOSURE
 ARSENIC INPUT  AT PRIMARY COPPER SMELTERS

                  CONTENT      THROUGHPUT
 SMELTER            (%)            (Ib/hr)
A-TACOMA
K-GARFIELD
PD-AJO
A-EL  PASO
A-HAYDEN
PD-HIDALGO
PO-OOUGLAS
PD-MORENC!
K-HAYDEN
K-McG!LL
INSPIRATION
MAGMA
COPPER RANGE
K-HURLEY
CITIES SERVICE
TOTAL
KEY

 A = ASARCO

 K = KENNECOTT

 PD = PHELPS DODGE
                     II-3

-------
processing about 4500 pounds of As per hour.  In comparison, the other
14 primary copper smelters process feed materials of much lower arsenic
content with the next highest smelter in terms of arsenic concentration
in the feed at 0.3 percent, and the next highest smelter in terms of
arsenic throughput rate at about 300 Ib/hr.  In fact, the A-Tacoma
smelter processes more arsenic than the other 14 smelters combined.

     In addition to processing more arsenic than the other 14 smelters,
the ASARCO-Tacoma smelter emits more inorganic arsenic to the atmosphere.
This is illustrated on the next slide.  (Slide #3)  Shown here are
estimates of current arsenic emission rates under full-smelt conditions
from process and fugitive emission sources at each domestic smelter now
operating.  "High level" emissions represent process emissions and
captured fugitive emissions which are discharged to the atmosphere
through a stack.  "Low level" emissions represent fugitive emissions
discharged to the atmosphere at or near ground-level.

     The distinction between "high level" and "low level" emissions is
an important one.  It  is made here because "low level" emissions have a
significantly greater impact on ambient concentrations in the near-field
(i.e., close-in to the smelter), and thus pose a greater health risk
than do "high level" or stack emissions which result in more dilute
concentrations more distant from the smelter due to dispersion effects.

     As you can see, the total emission of inorganic arsenic from all
smelters is estimated to be about 370 Ibs/hr.  Emissions at A-Tacoma
account for approximately 155 Ibs/hr or 42% of this total.   The next
highest emitter accounts for about 85 Ibs/hr or only 23% of the total.

     Of particular interest is the large difference between smelters in
terms of their  "low level" or uncaptured fugitive emissions.  We estimate
that total fugitive emissions from all smelters amount to about 115
Ibs/hr. Of this total, the A-Tacoma smelter is estimated to emit about
90 Ibs/hr or nearly 80%.  In contrast, the next highest emitting smelter
is estimated to emit less than 5 Ibs/hr or slightly more than 4% of the
"low level" emissions from all smelters.

     Due to the large disparity between smelters in terms of the quantity
of arsenic processed and their emissions, especially low-level emissions,
it was decided, for the purpose of regulation, to address only high
arsenic input smelters at this time.

     The source category is defined as primary copper smelters which
process feed materials with an annual average arsenic content of 0.4
percent or more.  Thus, the source category as it is presently defined,
includes only one existing copper smelter:  the ASARCO smelter at
Tacoma, Washington.  The selection of this 0.4 percent cut-off is based
                                  II-4

-------
                       ARSENIC  EMISSIONS

                    FROM  COPPER SMELTERS
en
z
g

(/i

i
LU

(J

Z
LU
W1
l±.
<
     400
     350
     300
     250
A-3
                                       HIGH LEVEL
                                       LOW LEVEL
                                                     ROASTER I1%l


                                                     1 CHARGING
                           COPPER SMELTING  PROCESS





                              ORE  CONCENTRATES

                                    (Tkl/
                                                                                                    A-4
            2 CALCINE DISCHARGE

             AND TRANSFER
FURNACE  H6%l


3 CHARGING


4 SLAG TAPPING


5 MATTE TAPPING


6 CONVERTER SLAG

 RETURN
                                                   CONVERTER I83%l


                                                   7 CHARGING


                                                   8 SLAG SKIMMING


                                                   9 PRIMARY HOOD LEAKS



                                                  10 BLISTER POURING
                                                                            ROASTER
                                                                                         OFFGASES  TO

                                                                                         CONTROL

                                                                                         DEVICE
                                                                            CAL
                                      CINES
                                                                            SMELTING

                                                                            FURNACE
                                                       t
                               MATTE
                               <•>
                               J<
                                                                                    SLAG
                                                   OFFGASES TO

                                                 —CONTROL

                                                   DEVICE


                                                —C-SLAG TO  DUMP
                                                                           CONVERTER
                                                   .OFFGASES TO

                                                 9) ACID  PLANT




                                                   BLISTER COPPER

                                                 10) TO REFINING
                    SMELTER

-------
primarily on our desire to limit the quantity of arsenic which non-
affected smelters could process and not be subject to regulation.  A decision
on whether to proceed with standards development for the remaining 14
copper smelters will be made later.

     Process Description

     For the benefit of those who may not be familiar with copper
smelting, I would like at this time to briefly review the processes
involved and identify the pertinent emission points.

     Simply, primary copper smelters use pyrometallurgical processes to
separate the copper contained in the copper-bearing materials processed
from the Fe, S, and other materials also present.  Eighty-five to
ninety-five percent of the materials processed consist of copper ore
concentrates obtained from the mining, milling, and beneficiation of
low-grade sulfide ores, which typically contain less than one percent
Cu.

     These concentrates usually contain from 15 to 30 percent Cu, and
comparatively large amounts of Fe and S.  They also contain quantities
of earthy matter, as well as a host of minor constituents including As.
In addition to concentrates, most smelters also process quantities of
copper precipitates obtained from the acid leaching of oxide ore deposits
and mine wastes as well as quantities of recovered flue dusts and other
smelter by-products.

     (Slide #4)  Illustrated on the screen are the three basic pyrometallurgical
operations employed:  roasting, smelting, and converting.

     If the concentrate is low in Cu relative to S, Fe, and other impurities,
the concentrate is generally roasted prior to smelting.  During roasting,
the concentrates are heated in an oxidizing atmosphere to a high temperature
but below the melting point of the constituents.  This results in the
elimination of a portion of the sulfur contained and the elimination
of some of the volatile impurities present.

     Two types of roasters are used:  multi-hearth roasters and fluid-
bed roasters.  The A-Tacoma smelter uses the former.

     The roaster product (called calcine) or raw, unroasted concentrates
(if roasting is not required) is then charged to the smelting furnace
where it is melted with fluxing materials.  The lighter impurities
combine with the flux and float to the top as a slag which is periodically
drawn off at a furnace tapping location and discarded at a dump.  The
copper, iron, and most of the sulfur, contained in the furnace form a
product known as matte which collects in the lower part of the furnace.
The molten matte, which typically contains from 40 to 45 percent Cu, is
periodically tapped and transferred by ladle to a converter.
                                II-6

-------
      Unlike the roasting and smelting processes which are continuous,
the converter process is a batch operation.  After a sufficient quantity
of matte has been added, air is blown through the molten bath.  This
results in the removal of S as SCL and the formation of an Fe rich slag
which is periodically skimmed and returned to the smelting furnace for
reprocessing.  This phase of the converter operation is called the
"slag" blow.

      Once most of the Fe has been eliminated by slagging, the second
phase of the converter operation is begun.  This phase is called the
"finishing" or "copper" blow.  Again, air is blown through the molten
bath and the remaining S is oxidized to SCL and eliminated.  The result
of this second blowing phase is the production of "blister" copper
which is 98 to 99 percent pure copper.  The "blister"copper is then
poured into ladles and, although not shown here, is transported to an
anode furnace for further refining and then cast into copper anodes
prior to electrolytic refining.

     Arsenic present in the feed materials is eliminated during the
smelting process by one of two mechanisms.  It is either volatized and
carried off as a metallic oxide in the process off-gases or removed by
slagging.  The relative proportions volatized and slagged, vary widely
from one smelter to another.  This is due to many factors including the
quantity of As present in the feed and differences in smelting configurations
and equipment used,

     The primary mechanism is, however, volatilization.  Depending on
the arsenic input, anywhere from 55 to 90 percent of the input As may be
volatized and 10 to 45 percent eliminated by slagging.   This is due to
the very high temperatures associated with copper smelting and the
inherent volatility of arsenic and its prevalent oxide, arsenic trioxide,
which is the predominant As compound found in smelter off-gases.

     Emission Sources

     Inorganic arsenic emissions from the smelting process can be categorized
as either process or fugitive.  Process emissions include emissions from
smelting equipment which are confined in process flow streams.  Fugitive
emissions include emissions which escape from process flow streams and
equipment due to leakage and emissions produced by the handling and
transfer of materials.

     Major process sources of inorganic arsenic at high arsenic throughput
smelters include all roasters, smelting furnaces, and converters.  If
left uncontrolled at Tacoma, these sources could potentially emit about
4000 Ibs of As per hour to the atmosphere under full-smelt conditions.

     (Slide  5)  Potential fugitive sources of inorganic arsenic are
shown here in red and listed in the margin.  As you can see, fugitive emissions
                                  II-7

-------
are associated with essentially each step in the smelting process.  The
more significant sources include calcine discharge operations, furnace
matte and slag tapping operations, and converter operations.

      Of these sources, fugitive emissions associated with converter
operations are by far the most significant, accounting for over 80% of
the total emission of As from fugitive sources at Tacoma.  (Slide #5)
Shown here is a side view of a conventional copper converter.  The
converter is a cylindrical vessel which typically measures about 30 feet
in length and 13 feet in diameter.  It is equipped with a large movable
primary hood which is used to capture the process off-gases generated
during slag and copper blowing.  Materials are added and recovered from
the converter through an opening in the center called the "mouth," which
measures about 5 feet in diameter.

     Substantial fugitive emissions are produced during each phase of
operation.  During converter charging the primary hood is retracted to
its highest position.  The converter is tilted by a drive mechanism
until the mouth of the converter is approximately 45 degrees from the
vertical.  Fugitive emissions during this operation result when matte or
other materials are poured from a ladle into the converter mouth.
Although primary converter hoods are relatively close-fitting, emissions
nonetheless occur during blowing operations.  These emissions consist of
leaks which escape through openings between the primary hood and converter
shell.  During slag skimming and blister pouring, the mouth of the
converter is rotated to a position between 65 to 125 degrees from the
vertical, depending upon the bath level.  Fugitive emissions during this
operation result as the molten material (slag or blister copper) is
poured from the converter into a ladle.

     Selection of Affected Facilities

     Listed on this slide (Slide #6) are the smelting facilities selected
for regulation.  As indicated, these include both process and fugitive
sources.  The process sources to be regulated include roasters, smelting
furnaces, and converters.  Although these sources are currently well-
controlled at the ASARCO-Tacoma smelter with 98 to 99 percent control,
emissions in the absence of controls would be extremely significant.

     The fugitive sources selected for regulation include multi-hearth
roaster calcine discharge operations, furnace matte and slag tapping
operations, and all converter operations (charging, blowing, skimming,
and pouring).

     Converter operations were selected for regulation because they
represent the single most significant source of fugitive arsenic emissions
at Tacoma, and are currently uncontrolled.  The other sources, selected
                                    II-8

-------
                               A-5.
  CHARGING
BLOWING
SKIMMING
       COPPER CONVERTER OPERATIONS
  SOURCES SELECTED AS  AFFECTED  FACILITIES

PROCESS
    : ROASTERS
    :SMELTING  FURNACES
    -.CONVERTERS

FUGITIVE
    :MULTI-HEARTH  ROASTER  CALCINE  DISCHARGE
    :SMELTING  FURNACE  MATTE TAPPING
    :SMELTING  FURNACE  SLAG TAPPING
    CONVERTER OPERATIONS
                  II-9

-------
for regulation, although controlled at Tacoma, were selected because
they also represent potentially significant sources of inorganic arsenic.

     Alternative Control Techniques

     Alternative techniques for the control of inorganic arsenic emissions
from the process and fugitive sources selected for regulation were
identified and their performance capabilities evaluated.

     For process sources, both hot and cold control devices were assessed.
Because the off-gases from the smelting processes under investigation
are typically high in temperature (ranging from 350 to 700°F) and because
arsenic trioxide (the predominant compound found in smelter off-gases) has an
appreciable vapor pressure even at moderate temperatures, it was suspected
that the operating temperature of a control device would have a significant
effect on the quantity of arsenic which could potentially be collected.

     To substantiate this premise, arsenic emission measurements were
conducted across a hot ESP used to control particulate emissions from a
reverberatory smelting furnace.  The precipitator was operated at 600°F
or higher, and had a demonstrated particulate removal efficiency, measured
at its operating temperature, of 97 percent.

     In contrast, the arsenic collection efficiency, based on three
sample runs conducted across the ESP, averaged less than 27 percent.
It was concluded that although the subject ESP was reasonably effective
in removing material which existed as particulate at its operating temperature,
any material such as arsenic trioxide which existed in the vapor state
passed through with little or no removal.  This clearly demonstrates a
need to cool the gas stream sufficiently to condense the arsenic present
in the vapor state prior to entering a control device if efficient
arsenic collection is to be achieved.

     As a result, the alternative control techniques for arsenic evaluated
all employed preceding as an integral part of the overall control
system.   (Slide #7)  Shown here are the control devices sampled, their
operating or outlet temperature, and process sources which they served.

     The  alternative control techniques evaluated included an ESP and
baghouse  (both of which used a spray chamber for gas cooling and were
operated  between 200 and 230°F), a baghouse which used air dilution for
cooling and was operated at 180-190°F, and a venturi scrubber operated
at about  150°F.  Although not considered an alternative control technique
because of economic considerations and because its application is limited
to strong SOo off-gas streams, we also conducted arsenic emission measurements
on a contact sulfuric acid plant and ancillary gas cleaning system to
assess its performance capabilities on arsenic emissions.  Simultaneous
inlet and outlet arsenic measurements were conducted across each control
device sampled.
                                  11-10

-------
               PROCESS EMISSION  CONTROL DEVICES EVALUATED

                  OPERATING           PROCESS  SOURCES
 CONTROL  DEVICE   TEMPERATURE  (°F)    CONTROLLED
SPRAY CHAMBER/        220-230
   ESP

BAGHOUSE              180-190

SPRAY CHAMBER/        220-230
   BAGHOUSE

VENTURI SCRUBBER        150

DOUBLE  CONTACT         150
   ACID PLANT
MULTI-HEARTH ROASTERS  AND
REVERBERATORY FURNACE

MULTI-HEARTH ROASTERS

FLUID-BED  ROASTERS,  ELECTRIC
FURNACE  AND CONVERTERS

FLUID-BED  ROASTER

CONVERTERS
A-7
 SMELTER

  A-EL PASO

  A-TACOMA
  ANACONDA

  K-HAYDEN
  A-EL PASO
                 ARSENIC COLLECTION EFFICIENCIES-PROCESS EMISSIONS
            100
             98
          S  36
          2
          UJ
          g
          £
             92

                                                    A-a
                   ESP   SAGHOUSE  BAGHOUSE  SCRUBBER ACID PLANT

                               CONTROL DEVICE
                                 11-11

-------
     (Slide #8)   This slide shows the results of these tests, in terms
of the arsenic collection efficiencies recorded.  For the five control
devices sampled, the average efficiencies measured ranged from a low of
about 98 percent for the SC/ESP to over 99 percent for one of the
baghouses and the acid plant.  The lowest efficiency measured for a
single sample run was 96.5 percent measured across the SC/ESP.  Based on
these results, it is our conclusion that each of the control devices
sampled are capable of essentially equivalent performance for the collection
of inorganic arsenic.

     For the fugitive sources selected for standards development, both
capture systems and a collection device were evaluated.  (Slide #9). The
capture systems evaluated for calcine discharge, matte tapping, and slag
tapping consisted of the application of conventional local  ventilation
techniques.

     Alternative capture systems evaluated for converter fugitive
emissions included both local ventilation and general  ventilation
techniques.  The local ventilation techniques evaluated included the use
of fixed hoods, applied at a number of U.S. smelters,  and a fixed enclosure/air
curtain system applied at a Japanese smelter.  The fixed hood systems
evaluated at U.S. smelters were judged to be only marginally effective
and were thus dismissed from further consideration.   The general  ventilation
technique evaluated consisted of enclosing and evacuating the building
housing the converters.

     (Slide #10)  Shown on this slide are the results  of visual observations
made on the calcine discharge, matte tapping, and slag tapping operations
at the ASARCO-Tacoma smelter.  Emissions from all three sources are
captured using close-fitting hoods and ventilation.   The visual observations
were made using EPA Method 22 which, rather than opacity, simply records
the presence of visible emissions.

     Calcine Transfer - Thirteen calcine transfer operations, each
averaging about two minutes in duration were observed.  The visual
observations were made at the opening of the tunnel-like structure
houses the roaster calcine hoppers and larry cars during the calcine
discharge operation.  As indicated, no visible emissions were observed
at any time.

     Matte Tapping -  In the case of furnace matte tapping, simultaneous
but separate observations were made at both the furnace tap port and at the
launder-to-ladle transfer point.  Sixteen taps, averaging approximately
5.5 minutes in duration were observed at the tap port.  For the 16
observations made at the matte tap port, visible emissions  were observed
to be present only 0.2 percent of the time on average.  The single
highest reading recorded for an individual tap showed visible emissions
present only 3 percent of the time.  No visible emissions were observed
100 percent of the time from the launder to matte ladle transfer point
during all 15 observations made at that location.
                                 II-12

-------
     FUGITIVE EMISSION CAPTURE TECHNIQUES
    SOURCE
    TECHNIQUE
CALCINE TRANSFER

MATTE TAPPING

SLAG TAPPING
LOCAL VENTILATION
CONVERTER
LOCAL VENTILATION
   Fixed Hood
   Fixed Enclosure/
     Air Curtain

GENERAL VENTILATION
   BuiJdirtg Evacuation
           VISIBLE EMISSIONS  OBSERVATIONS
                 AT ASARCO-TACOMA
                  (EPA METHOD 22)
                     A-10



SOURCE
CALCINE DISCHARGE
MATTE TAPPING
TAP PORT
LADLE
SLAG TAPPING
TAP PORT
SLAG POT


MIIMRFR OF
PIUIVIDI^IX \Jt
OBSERVATIONS
13

16
15

8
11
PERCENT OF TIME
VISIBLE EMISSIONS OBSERVED

AVERAGE
0

0.2
0

5.3
88
MAXIMUM
0

3.0
0

15
99
                      11-13

-------
     Slag Tapping -  As with matte tapping, separate observations were
made for slag tapping at the furnace tap port location and at the slag
launder to slag pot transfer point.  Results obtained for 8 observations
at the slag tap port showed that visible emissions were observed about 5
percent of the time on average.  The highest single observation showed
the presence of visible emissions 15 percent of the time.

     In contrast to the other sources, visual observations made at the
slag launder-to-pot transfer point indicated poor performance.  Visible
emissions observed at this location showed emissions present 88 percent
of the time over 11 slag taps.  The highest single observation recorded
showed visible emissions present 99 percent of the time.  Additional
data obtained using EPA Method 9, not presented here, showed that the
emissions observed were significant with opacities as high as 50 percent.

     Conversations with smelter personnel revealed that the ventilation
hood at the slag launder discharge point had been damaged when hit by a
truck.  Although an inspection of the ventilation hood and ancillary
ductwork showed no apparent damage, ventilation at this location was
concluded to be inadequate to handle the volume of emissions and fume
generated.

     Conclusions -  Based on these data, it was concluded that a properly
designed and operated low ventilation system applied to calcine discharge
operation and matte tapping operations could readily achieve a minimum
capture efficiency of 90 percent.  Based on the results of the visual
observations obtained on the slag tapping operations, especially at the
launder to slag pot transfer point, and the fact that the capture system
had been reportedly damaged, it was concluded that the slag tapping
ventilation system observed at Tacoma, as it was operating at the time,
should not be considered representative of a best system of emission
reduction.

     Although slag tapping operations do represent a somewhat
more difficult control situation than matte tapping, the outstanding
performance demonstrated by the matte tapping controls at Tacoma
strongly suggest that a properly designed and operated ventilation
system applied to slag tapping operations should be capable of achieving
a capture efficiency equivalent to that observed for matte tapping.

      (Slide #11)  Shown here is a simple illustration of the fixed
enclosure/ air curtain system evaluated for the control of converter
fugitive emissions. The system consists essentially of a complete enclosure
equipped with front-doors and a movable roof.  During blowing and slag
skimming and blister pouring operations, the doors and roof are closed
and the enclosure is ventilated.

      (Slide #12)  During charging operations, the doors and roof are opened.
An air curtain jet stream is blown horizontally across the top.  The air
curtain is formed by blowing ambient air through a slot located along one
                                    11-14

-------
                                  MOVABLE ROOF
                                  IVIUVMDL.C nuur      .   . i
                                  A	   A-ll
  FRONT DOORS •
                                            CONVERTER
      CONVERTER  FIXED  ENCLOSURE/AIR CURTAIN  (CLOSED)
                             AIR CURTAIN STREAM
AIR FROM SLOWER
                                           CONVERTER
       CONVERTER FIXED ENCLOSURE/AIR CURTAIN (OPEN)

                       11-15

-------
side of the enclosure at the top.  Emissions entrained by the air curtain
stream are swept across the opening and collected at a capture hood
located along the opposite sidewall.

     During a brief visit to the Mitsui smelter in Tamano, Japan, our
contractor was able to make a limited number of visual observations on
the fixed enclosure/air curtain system operated at that smelter to assess
the performance of the system during the various modes of converter
operation.  (Slide #13)

     Three matte charges were observed using both EPA Methods 22 and 9
simultaneously and one matte charge with EPA Method 9 alone.  Each charge
lasted from 1 to 1 and 3/4 minutes in duration.

     Although the Method 22 results indicate that emissions were present
a good portion of the time (60 percent on average), the opacity results
indicate that the emissions observed were slight, averaging less than 3
percent opacity for 4 charges with the maximum opacity observed
at any time being 25 percent.  When present, the emissions appeared as
small puffs penetrating the air curtain stream.

     Only two slag skims were observed.  For the first, which lasted 11
minutes, no visible emissions were observed at any time.  In contrast,
during the second, which lasted for 9 minutes, visible emissions were
observed 100 percent of the time.  Again however, the emissions observed
were slight, appearing as small puffs which escaped from the enclosure
through a narrow opening between the front doors and the enclosure roof.

     Two blister pours were observed using EPA Method 22 and 3 using EPA
Method 9.  Again, the Method 22 results indicate that visible emissions
were generally continuous.  The Method 9 results indicate that the
emissions observed were somewhat more substantial than those observed
during either matte charging or slag skimming, averaging nearly 9 percent
opacity with the maximum single observation recorded being 35 percent.

     As with slag skimming, the emissions were observed above the narrow
opening between the front doors and roof.

     Visual observations were also made for both slag and copper blowing
using EPA Method 9.  Each was observed for about 1/2 hour.  As the data
show, no visible emissions were observed at any time during the observation
periods.

     It is our conclusion, based on these observations that the fixed
enclosure/air curtain system is an effective capture device for the
control of fugitive emissions from converter operations.  When compared
to fugitive emissions typically observed from an uncontrolled converter,
an overall capture efficiency of 90 percent is judged to be achievable.

     As noted previously, the other alternative considered for the
control of fugitive emissions from converter operations is building
                                 11-16

-------
SUMMARY OF VISIBLE EMISSIONS OBSERVATIONS-
       FIXED  ENCLOSURE/AIR  CURTAIN
A-13
OPERATION
MATTE CHARGING
SLAG SKIMMING
BLISTER POURING
SLAG BLOW
COPPER BLOW
EPA METHOD 22
NUMBER OF
OBSERVATIONS
3
2
2
—
—
AVE.
60
45
84
—
—
MAX.
77
100
100
—

EPA METHOD 9
NUMBER OF
OBSERVATIONS
4
1
3
(30min)
(27 min)
AVE.
2.8
0
8.5
0
0
MAX.
25
0
35.
0
0
   PERFORMANCE  DATA FOR CONVERTER BUILDING
         BAGHOUSE AT ASARCO-EL PASO

              ARSENIC  EMISSIONS
   A-I+
SAMPLE
RUN
1
2
3
AVG.
IN
mg/Nm3
6.21
2.09
1.53
3.27
LET
kg/hr
5.51
1.96
1.31
2.92
GUI
mg/Nm3
0.39
0.017
0.0015
0.137
["LET
kg/hr
0.310
0.017
0.012
0.113
EFFICIENCY,
PERCENT
94.5
99.1
99.1
96.2
              PARTICULATE EMISSIONS
SAMPLE
RUN
1
2
3
AVG.
INLET
mg/Nm3
60.3
53.3
70.5
61.3
kg/hr
44.7
46.3
61.2
50.7
OUTLET
mg/Nm3
11.6
2.5
1.1
5.1
kg/hr
10.40
0.92
0.40
3.90
EFFICIENCY,
PERCENT
76.7
98.0
99.3
91.3
                  11-17

-------
evacuation.  Simply, building evacuation consists of enclosing the
building housing the converter aisle, allowing openings only for make-up
air and access, and ventilating the fugitive emissions discharge into
the building through ventilation points located in the building roof.
This approach is currently being used at one domestic smelter, the
ASARCO smelter at El Paso, Texas.  At this smelter, the entire volume of
the converter building is ventilated at a rate of about 16 changes per
hour.

     Our conclusions regarding the effectiveness of building evacuation
applied to converter fugitive emissions are based primarily on engineering
judgment and informal observations of the system used at the ASARCO-
El Paso smelter.  Providing the building is properly enclosed and adequate
ventilation rates are applied, essentially 100 percent capture should be
possible.  However, owing to the need for openings in the building for
access and makeup air, a more conservative estimate of 95 percent capture
is considered more reasonable.

     For the collection of fugitive emissions, only one control device was
evaluated, a baghouse used to collect the fugitive emissions captured by the
converter  building evacuation system at the ASARCO-E1 Paso smelter.
(Slide #14)  Data obtained are presented on this slide.

     As you can see, test results are reported for both arsenic and
particulate matter.  In both cases, simultaneous inlet and outlet
measurements were made.

     As shown here, the baghouse achieved an average control efficiency
for arsenic of 96 percent and an average control efficiency for total
particulates of 91 percent.  Although at first glance, the average
efficiencies measured appear to be somewhat low, it should be noted that
the inlet  loadings were very low, averaging 3.3 mg/m  for the As samples
and 61 mg/m  for the particulate samples.

     The outlet concentrations measured were somewhat variable,3ranging
from 0.14  mg/m  to 0.39 mg/m  for As and 1.1 mg/m  to 11.6 mg/m  for
total particulate.

     It should also be noted that we expect that the inlet concentration
associated with the application of the fixed enclosure/air curtain system
on converters to be about 3 times higher than that measured here because
of the lower flow rate requirements for local ventilation systems in
comparison to general ventilation systems.
                                 11-18

-------
2.  Regulatory Approach
                              Mr.  Graham  Fitzsimons
                           Standards  Development  Branch
                   Emission Standards  and Engineering  Division
                      U.  S. Environmental  Protection Agency
                  Research  Triangle  Park, North Carolina   27711
     Following the presentation given by Mr.  Vervaert,  a  summary  of  the
regulatory approach taken by EPA was  described  by Mr, Graham  Fitzsimons
of the Standards Development Branch.   An outline  of his presentation  follows.
The preamble to the NESHAP for inorganic arsenic  emissions  from primary
copper smelters contains a more detailed discussion of  the  regulatory
approach.

     I.   NESHAP DEVELOPMENT PROCESS UNDER PROPOSED "CARCINOGEN POLICY" IS
         DESCRIBED. (Slide G-l)


    II.   REGULATORY ALTERNATIVES

         A.   Baseline Alternative (No additional  regulation)

             1.  Importance of Baseline

             2.  Other Regulations Analyzed

                 a.  Summary of economic and technical  impacts (Slide G-2)

                 b.  More detailed description of the  impacts on Asarco-
                     Tacoma due to S02 NAAQS and OSHA  regulation for arsenic

             3.  Baseline as Characterized by Controls at Asarco-Tacoma
                 (Slide G-3)

         B.   Description of Three, More Stringent Alternatives in Terms of
             Control Technology on Process and Fugitive Sources  (Slide G-3)


   III.   ANALYSIS OF ENVIRONMENTAL, ECONOMIC AND ENERGY IMPACTS  OF THE
         ALTERNATIVES

         A.   The Impacts of Each Alternative are Presented Beginning with
             the Most Stringent Alternative (Slide G-4).

         B.   Based on this Analysis, Alternative 2 Selected  as "BAT"


    IV.   CONSIDERATION OF "BEYOND BAT" ALTERNATIVE

         A.   Preliminary Estimates of Further Reduction  in Risk  and  Deaths
             Associated with Going from Alternative 2  to Alternative 3

                                    11-19

-------
    B.  Discussion of  Impacts of Alternative  3  in  Terms  of Loss  of Jobs
        and Increase in Copper  Imports

    C.  Conclusion:  It is  not  Reasonable  to  go Beyond  BAT (Impacts of
        Going Beyond BAT Disproportionate  to  Reduction  in Risks  and
        Deaths to be Gained)

    D.  Review of BAT  (Slide G-5)


V.  FORMATS  FOR  STANDARDS

    A.  Generally Acceptable  Formats for Standard  Under Section 112
        of the Clean Air Act  (Slide G-6)

    B.  Formats  Selected for  Recommended Standards (Slide G-7)

        1.  Process  Sources (Roaster, Furnace and  Converter) - Efficiency
             Standard  for Arsenic

        2.  Collection of  Fugitive Emissions  -  Concentration Standard
             for  Particulate

         3.  Capture  of Fugitive Emissions

             a.   Calcine discharge, matte and  slag  tapping - visible
                  emission  standard (EPA Method 22)

             b.   Converter  operations - equipment standard


VI.   SELECTION OF NUMERICAL EMISSION LIMITS
     (for  sources where emission limits are feasible)

     A.   Collection of Process Emissions - 96  Percent Removal Efficiency
         for Arsenic Based  on ESP,  Baghouse, Scrubber and Acid Plant
         Performance (Slide G-8)

     B.   Collection of Fugitive Emissions - 11.6 mg Particulate/m
         Based on Data from Asarco-El Paso (Slide G-9)

     C.   Capture of Fugitive Emissions (Slide G-1Q).

         1.  Calcine Discharge - No Visible Emissions (NVE) 100 Percent
             of the Observation Time Based on Observations at Asarco-Tacoma

         2.  Matte and Slag Tapping - Standards Based on Observations
             of Matte Tapping at Asarco-Tacoma; Slag Tapping Data Not
             Representative of BAT

             a.   Tap Port - NVE 95 percent of the observation time

             b.   Transfer point -  NVE 100 percent of the observation time
                               11-20

-------
     EQUIPMENT STANDARD FOR CAPTURE  OF CONVERTER  FUGITIVES -  General
     Description of What Would be  Required; Technical  Details Not Yet
     Complete
VI.  SUMMARY OF RECOMMENDED STANDARDS (Slide G-ll)
                       NESHAP DEVELOPMENT  FOR HIGH
                       ARSENIC  THROUGHPUT  SMELTERS
                            IDENTIFY  ALTERNATIVES
                     ANALYZE ENVIRONMENTAL, ECONOMIC,
                   AND ENERGY IMPACTS OF  ALTERNATIVES
                                      1
                 SELECT "BEST AVAILABLE TECHNOLOGY"  (BAT)
                                      1
                        EXAMINE  RESIDUAL RISK AFTER
                     APPLICATION  OF BAT vs.  THE IMPACTS
                           OF  GOING "BEYOND BAT"
                        DEVELOP REGULATION BASED ON
                      APPROPRIATE CONTROL TECHNOLOGY
                SUMMARY OF EFFECTS OF REGULATORY BASELINE
                                           G-Z
    REGULATION
                  COMPLIANCE
                     DATE
                       EFFECTS ON
                    ARSENIC EMISSIONS
                             ECONOMIC
                              EFFECTS
    OSHA
    SO2-NAAQS
     FUGITIVE
    RCRA
    CWA
    S02 -NAAQS
      PROCESS
                    JULY  1982
   MARCH 1983
                NOVEMBER  1983
                    JULY  1984
. JANUARY 1988
                 CAPTURE AND DISPERSION OF
                 ROASTER AND SMELTING
                 FURNACE FUGITIVE  EMISSIONS
CAPTURE AND DISPERSION  OF
CONVERTER FUGITIVE EMISSIONS
                                NONE
                                 NONE
                 NONE
                                                          VERY SIGNIFICANT
SIGNIFICANT
                                                          MODEST
                                                          MODEST
                                                           POTENTIALLY
                                                           VERY SIGNIFICANT
                                11-21

-------
             REGULATORY ALTERNATIVES
c?-3
REGULATORY
ALTERNATIVE
1
2
3
4
FUGITIVE
CAPTURE
CALCINE DISCHARGE,
MATTE AND SLAG TAPPING
LOCAL HOODS
AND
VENTILATION
LOCAL HOODS
AND
VENTILATION
LOCAL HOODS
AND
VENTILATION
CONVERTER
FIXED ENCLOSURE/
AIR CURTAIN
FIXED ENCLOSURE/
AIR CURTAIN
BUILDING
EVACUATION
COLLECTION
ALL SOURCES
NONE
BAGHOUSE OR
EQUIVALENT
TECHNOLOGY
BAGHOUSE OR
EQUIVALENT
TECHNOLOGY
PROCESS
COLLECTION
ALL SOURCES
BAGHOUSE OR
EQUIVALENT
TECHNOLOGY
BAGHOUSE OR
EQUIVALENT
TECHNOLOGY
BAGHOUSE OR
EQUIVALENT
TECHNOLOGY
FURTHER REDUCTION
ON ARSENIC EMISSIONS
(RESULTING IN PLANT SHUTDOWN)
SUMMARY OF ENVIRONMENTAL, ECONOMIC, AND ENERGY
INCREMENTAL IMPACTS OF REGULATORY ALTERNATIVES
        G-i
REGULATORY
ALTERNATIVE
1
1
3
4
TOTAL ARSENIC
EMISSION REDUCTION
(TONS/YR)
—
260
275
666
ENERGY
IMPACT
(x 106kWh)
— -
+1.5
+23.6
— -
COST($x106)
CAPITAL
....
+ 1.5
+8.1
NA
ANNUAL
....
+0.43
+2.8
NA
ECONOMIC
IMPACT
AFFORDABLE
AFFORDABLE
PLANT CLOSURE
PLANT CLOSURE
              11-22

-------
                  SUMMARY OF BAT FOR PROPOSED NESHAP
          SOURCES
                              CAPTURE
                           G-5
                                                  COLLECTION
PROCESS

ROASTER

FURNACE

CONVERTER
      NA
BA6HOUSE OR
EQUIVALENT WITH
PRE-COOLING TO 250°F
OR BELOW
FUGITIVE

ROASTER: CALCINE DISCHARGE

FURNACE: MATTE AND SLAG TAPPING

CONVERTER OPERATIONS
  (charging, blowing, skimming
  holding, and pouring)
LOCAL VENTILATION

LOCAL VENTILATION

FIXED ENCLOSURE/
 AIR CURTAIN
BAGHOUSEOR
EQUIVALENT
 NA=not applicable
              FORMATS CONSIDERED FOR STANDARDS
                • DIRECT MEASUREMENT
                     -MASS RATE  (kg/hr)
                     -CONCENTRATION (mg/m3)
                     -PROCESS WEIGHT (kg/Mg)
                     -COLLECTION EFFICIENCY
                         ( % removal efficiency)
                 'INDIRECT MEASUREMENT
                     -VISIBLE EMISSIONS
                •EQUIPMENT SPECIFICATION
                                11-23

-------
              FORMATS SELECTED FOR STANDARDS
                       C-7
• COLLECTION OF PROCESS EMISSIONS
-EFFICIENCY (ARSENIC)
• COLLECTION OF  FUGITIVE EMISSIONS     - CONCENTRATION (PARTICULATE)
 •CAPTURE OF FUGITIVE EMISSIONS:
      CALCINE DISCHARGE, MATTE AND
      SLAG TAPPING
      CONVERTER OPERATIONS
-VISIBLE EMISSIONS

- EQUIPMENT STANDARD
   (for Fixed Enclosure/
   Air Curtain System)
          ARSENIC COLLECTION EFFICIENCIES-PROCESS EMISSIONS
    100
     98
  8
  c
  UJ
  § 96
  LU.
   g
   S! 94
     92
     90

            ESP    BAGHGUSE  3AGHGUSE  SCRUBBER  ACID PLANT

                        CONTROL DEVICE
                                   11-24

-------
PERFORMANCE DATA FOR CONVERTER BUILDING
      BAGHOUSE AT ASARCO-EL PASO  •
          PARTICULATE EMISSIONS
SAMPLE
RUN
1
2
3
AVE.
INLET
mg/Nm3
60.3
53.3
70.5
61.3
kg/hr
44.7
46.3
61.2
50.7
OUTLET
mg/Nm3
11.6
2.5
1.1
5.1
kg/hr
10.40
0.92
0.40
3.90
          VISIBLE EMISSIONS OBSERVATIONS
                AT ASARCO-TACOMA
                 (EPA METHOD 22)
£'10



SOURCE
CALCINE DISCHARGE
MATTE TAPPING
TAP PORT
LADLE
SLAG TAPPING
TAP PORT
SLAG POT


MHMRFR OF
I "I \J ITIUL. Ix *_/!
OBSERVATIONS
13

16
15

8
11
PERCENT OF TIME
VISIBLE EMISSIONS OBSERVED

AVERAGE
0

0.2
0

5.3
88
MAXIMUM
0

3.0
0

15
99
                    11-25

-------
                  SUMMARY OF  PROPOSED STANDARDS
           SOURCE
                                 CAPTURE
                                                         COLLECTION
PROCESS

ROASTER
FURNACE
CONVERTER
         NA
96% ARSENIC
COLLECTION
EFFICIENCY '
FUGITIVE

CALCINE  DISCHARGE

MATTE AND  SLAG TAPPING
       TAP  PORT AND
       LAUNDER

       TRANSFER POINT



CONVERTER  Ore RATIONS
0 VISIBLE EMISSIONS
 100% OF TIME

0 VISIBLE EMISSIONS
AT  LEAST 95% OF TIME

0 VISIBLE EMISSIONS
  100% OF TIME


 FIXED  ENCLOSURE/AJR
 CURTAIN EQUIPMENT
 STANDARD
11.6 MILLIGRAMS
PARTICULATE MATTEF
PER STANDARD
CU81C M€TER
    1  This efficiency would  not in  general  apply  to process  emissions treated  in
       sulfuric  acid  plants.
                                   11-26

-------
                       B,  INDUSTRY PRESENTATION


1. ASARCO. Incorporated

                         Mr.  Edwin  S,  Godsey
                       ASARCO,  Incorporated
                       500  Crandall Building
                    Salt  Lake City, Utah  84101
     Mr.  Edward S.  Godsey  presented  a  description  of ASARCo's planned
converter  fugitive  emission capture  and  collection  system.   Three
converters  are operated  at the Tacoma  facility.  ASARCo  has  decided
and appropriated  sufficient funds  to design,  install,  and  operate
fixed  enclosure/air curtain systems  (FE/AC) at  all  three converters.
An existing  electrostatic  precipitator (No. 2 Cottrell)  will  be used
to collect  the particulate and arsenic captured  by  the converter
FE/AC.

     A two-phase  design  and construction  effort  is  planned over a
36-month  schedule.   Phase  I involves the  design  and  installation of an
FE/AC  at  the No.  4  converter.  Necessary  ductwork  for all  three converter
FE/AC  at  Tacoma will  be  installed  during  the  first  phase.  The ductwork
will tie  into an  existing  brick  flue that leads  to  an electrostatic
precipitator, the No.  2  Cottrell.  Only  the FE/AC  for the  No.  4 converter
will be  installed during Phase I.  This  capture  system will  be tested
for approximately three  months and modified as  necessary to  maximize
the capture  efficiency.  Following the test program,  Phase II  will  be
initiated.   Phase II  involves the  redesign and  installation  of the
FE/AC  for  the remaining  two converters at Tacoma.

     On  October 29,  1980,  the ASARCo Board of Directors  approved a
$4.45  x  10   budget  to  complete the Tacoma converter  fugitive emission
capture  and  collection system.   Design and equipment ordering  was
begun  in  January  of 1981.  By January  of 1984 (36  months)  the entire
system should be  operational.  To  date,  the main fan (1250 horsepower)
has been  ordered, as  well  as many  of the  electrical  system items.

     The  planned  ASARCo-Tacoma fixed enclosure/air  curtain system is
similar  to  the Tamano  FE/AC and  the  system described in  Regulatory
Alternative  II in the  BID.  One  difference is that  ASARCo  does not
plan to  install doors  on the  front of  the fixed  enclosure.   Probable
damage to  the doors by cranes or ladles  and the  resulting  maintenance
problems  are cited  as  the  reasons  for  Tacoma's  decision  to utilize  a
fixed  enclosure without  doors.   ASARCo visited  several smelters in
Japan  which  operate FE/AC  systems  without doors.  The ASARCo plan and
design for  the fixed  enclosure allows  for the addition of  doors if
necessary.   The costs  of doors were  included  in  the  cost estimate and
budget for  the planned FE/AC  system.

     The  Tacoma air curtain will be  designed  to  operate  at 18,000 acfm
at a 30-inch static pressure.  Fans  and  ductwork designed  to handle
100,000  acfm per  converter when  the  air curtain  is in use  and
60,000 acfm  per converter  when the air curtain  is  not in use are
planned.
                              11-27

-------
     The air curtain will be operated  during  converter  turn  in and
turn out, skimming, and charging.  The air  curtain  will  not  be operated
during blowing or holding.  The  primary hood  will  be  in  place
during blowing and holding.

     An existing electrostatic precipitator (ESP),  the  No.  2 Cottrell,
will be used to collect particulate  and arsenic  captured by  the FE/AC
system.  The No. 2 Cottrell was  first  installed  in  1924, and has been
modified several times.  This ESP will  treat  anode  furnace  process
emissions, as well as the converter  FE/AC gas stream.   A total design
gas volume of 250,000 acfm will  be routed to  the No.  2  C^ttrell.  The
future specific collection area  will be reduced  to  59  ft /100 cfm,
compared to a 62 ft /cfm present specific collection  area.   A new
electrical substation will be added  to allow  ASARCo to  increase the
power input to the No. 2 Cottrell.   Fewer electrical  sections in the ESP
will be controlled by each control unit.  Following the  installation
of  the FE/AC system, the grain load  is predicted to be  0.1  gr/ft  at
the inlet and 0.003 gr/ft  at the outlet.   This  is  equivalent to a
97  percent collection efficiency.
                                 11-28

-------
                            C.  DISCUSSION

     Following the EPA presentation, Mr. Don  Goodwin  opened  the floor
to questions and comments from the NAPCTAC members.   EPA  staff and
contractor personnel were on hand to respond  to  questions  and  discuss
issues of concern to the NAPCTAC members.  Industry representatives
from ASARCo then made a presentation (see Section  B of  these minutes)
which was followed by discussion.  For clarity,  discussions  are grouped
by subject matter rather than being placed in chronological  sequence.

     The major concern of the NAPCTAC members was  towards  the  proposed
regulation.  The members questioned the  reasonableness  of  establishing
a regulation only for the ASARCo-Tacoma  smelter.   They  felt  that for
an efficient utilization of resources EPA should also reach  conclusions
(in terms of regulation) for all other smelters  since so much  data are
available for them.  Don Goodwin told the members  that  the ASARCo-Tacoma
and Anaconda smelters, of all the copper smelters, were found  to be
the most significant sources of arsenic  emissions, and  that  there was
an immediate need to regulate these two  high  arsenic  throughput smelters.
When the Anaconda smelter was shut down, it was  eliminated from the
regulation, and the regulation would only apply  to the  ASARCo-Tacoma
smelter.  EPA has not ruled out a regulation  for the  other smelters.
Mr. Stan Cuffe added that another study  is being done by EPA which
assesses all sources of arsenic emissions including the copper smelters.
The emissions from the remaining copper  smelters will be compared with
the emissions from the sources such as zinc and  lead  smelters,  pesticide
manufacturing, cotton gin, and glass manufacturing.  Then  these emis-
sion sources will be prioritized for regulation  based on their arsenic
emissions.

     Don Goodwin told the NAPCTAC members that ASARCo is planning to
install a fixed enclosure/air curtain system  (FE/AC) on the  converters
at the Tacoma smelter.  (See Section B of these minutes for  a  description
of ASARCo's planned FE/AC system.)  The  EPA staff  has examined  the
details of the system and is convinced that ASARCo's  proposed  controls
are similar to those which would be required  by  the proposed standard.
As a result, EPA does not plan to proceed with the standard-setting
process beyond this point since it will  have  no  impact  on  the  ASARCo-
Tacoma smelter.  EPA will monitor ASARCo's progress.

     Many of the members strongly felt that EPA  should  continue with
the proposed arsenic NESHAP.  They noted that no enforceable agreement
exists between ASARCo and EPA to install an FE/AC  system at  the Tacoma
smelter and wondered if the State or regional agency would be  able to
enforce such an agreement in the absence of a Federal regulation.

     During the discussions the committee agreed that it would be
inappropriate to set a national standard for  only  one smelter.   However,

                                11-29

-------
the committee's feelings varied  on  EPA's  future  action  for the ASARCo-
Tacoma smelter.  Some committee  members  felt  that it is acceptable for
EPA not to establish a national  standard  if a legally binding agreement
is obtained with ASARCo committing  them  to  install  the  planned controls.
Since it is a one plant situation,  such  a legal  agreement should bind
ASARCo with the local agency  (but not  with  the Federal  EPA).   Mr.  Reiter
said that the agreement should be signed  by a Chief Executive Officer
of ASARCo, not a vice president.

     Other members of the committee were  of the  opinion that it is
EPA's duty to develop the arsenic NESHAP.   They  were concerned over
the possibility of new smelters  or  existing smelters treating high
arsenic feeds.  They said, referring to  the National Cancer Institute's
findings, that arsenic emissions at some  other smelters also pose
serious health risks.  For this  reason,  and the  fact that considerable
time and money have been spent and  sufficient data  exist, EPA should
proceed to develop the NESHAP for all  copper  smelters.   In regards to
ASARCo's planned controls for ASARCo-Tacoma,  one committee member
commented that based on the past history  of legal  battles between  the
local agency and ASARCo for controlling  S02 and  arsenic,  the local
agency would require an enforceable requirement  for arsenic control.

     The committee members wondered if the  plan  to  install an FE/AC
system at the Tacoma smelter was a  voluntary  action by  ASARCo or if it
was a part of a permit application  to  a  local, State, or  Federal
agency.  Mr. Goodwin explained that the  possible increase in production
rate may be the main incentive for  ASARCo to  install an FE/AC system
at the Tacoma smelter.  The smelter has  been  operating  a  supplementary
control system (SCS) in conjunction with  the  existing S0« control
system.   (The SCS reduces emissions when  ambient air quality monitoring
data and meteorology information indicate the possibility of a NAAQS
violation by limiting the emissions of a  pollutant  through production
curtailments.)  Since the FE/AC  system will collect S02 in addition to
arsenic, it would enable the  smelter to  increase its annual  production
of copper by reducing production curtailments.   Mr.  Steve Bundi,  an
ASARCo representative, added  that the  installation  of an  FE/AC system
on converters would contribute in achieving OSHA standards for arsenic
and SO-.  As a condition of using the  SCS, ASARCo  obtained a variance
for the Tacoma smelter from PSAPCA  which  provided  extensions  from
compliance from several PSAPCA regulations.   In  the variance procedure,
the smelter is required to list  all  the  changes  to  be made to the
control systems.  ASARCo believes that when applying for  variance
the next time, it will submit to the local agency  the plan for
installation of an FE/AC system  at  the ASARCo-Tacoma smelter.

     Mr. W. Reiter questioned the practicality of  the proposed visible
emission standard of zero emissions for  100 percent of  the time for
calcine discharge and dust transfer points.   He  was concerned that
even a wisp of smoke will cause  a source  to violate this  stringent
level of the standard.  Mr. C. Beard questioned  the visible emission
limit of zero emissions for 95 percent of the time  for  matte and  slag
                                   11-30

-------
tapping operations.  He wondered why  EPA  will  limit the emissions to
zero for 95 percent of the time and allow unlimited emissions for the
remaining 5 percent of the time.   Mr.  Al  Vervaert explained that the
zero visible emission limit  for the calcine  discharge was based on no
emissions observed for 13 discharges.   The standard for the matte
tapping operation was based  on observations  made  for 16 taps.  Regarding
the concerns about the 95 percent  visible emission limit, he argued
that a source is required to  install  the  best  controls to achieve the
limit.  If the control system limits  the  emissions to zero for 95 percent
of the time, it is very unlikely that  it  would allow a big cloud of
emissions to escape during the remaining  5 percent of the time.

     Ms. J. Chalupnik inquired of  any  available method which will
ensure the proposed converter control  at  ASARCo-Tacoma to operate
continuously without violations.   Mr.  Vervaert stated that EPA methods
for visible emissions would  be used as  a  check for emissions from the
source.  EPA has recommended  equipment  specifications for the converter
FE/AC system.  After controls are  installed  at the ASARCo-Tacoma
smelter, visual observations  should be  made  using an appropriate
method (Method 22 or Method  9) during  each mode of converter operation,
and the observed values should be  used  as an indication of proper
maintenance and operation.   Also,  the  company  should be required to
obtain periodic readings and  report them  to  the Agency.   Mr.  G.  Fitzsimons
added that the operation and  maintenance  requirements for the control
systems, though not included  in the current  draft preamble given to
the committee members, would  be added  to  the preamble if EPA proceeded
with the regulation.

     Ms. J. Chalupnik expressed concern that the  slag dump area  was
considered by EPA to be a negligible  source  of arsenic emissions and
not regulated, while the slag tapping  operation was found to emit only
0.54 Ib arsenic per hour and  considered for  regulation.   Mr.  Vervaert
explained that EPA did not consider the 0.54 Ib of arsenic per hour
from the slag tapping operation to be  significant.   The main reason
for regulating the operation  is that  it is relatively easy and inexpensive
to control.  EPA obtained samples  of  the  slag  at  the slag tapping
operation and at the slag dump location.   The  analysis of samples
showed no change in the percent arsenic contained in the cooled  slag
at the dump and that contained in  the  slag at  the tap location.
Similar results have been obtained by  ASARCo as documented in a
confidential report provided  to EPA by  the company.

     Nonetheless, Mr. Vervaert stated  that EPA will  contact PSAPCA for
available arsenic emission data for the slag dump area at the
ASARCo-Tacoma smelter.  EPA will change its  position if the data prove
the slag dump area to be a significant  source  of  arsenic emissions.

     Ms. Chalupnik wanted to  know  why  control  measures were not  proposed
for the miscellaneous emission sources  such  as dust transfer and
handling, reverberatory furnace roof  tops, and others.  Mr. Vervaert
stated that the isolated actions,  such  as pulling of flue dust once a
year, do not result in continuous  emissions, and  they can be handled
by OSHA requirements such as  good  housekeeping measures.
                                   11-31

-------
     Ms. J. Chalupnik  inquired  of  any  requirement in the SIP that
would impose particulate emission  limits  for the FE/AC control  system
once installed.  Al Vervaert  stated  that  he was  not aware of a SIP
requirement for fugitive particulate emissions  from converters.   His
opinion was that the smelter  would control  the  fugitive particulate
emissions from the FE/AC,  since it currently controls particulate
emissions from existing fugitive emission sources.

     Several questions were raised about  the planned design and
operation of ASARCo-Tacoma's  fixed hood/air curtain (FE/AC) system and
planned modifications  to the  existing  No. 2 ESP  collection system.  A
concern was expressed  whether the  modified ESP  could achieve the
96 percent collection  efficiency for arsenic emissions.  ASARCo indicated
that based on the company's considerable  experience with ESPs it was
confident of achieving 96  percent  collection.

     Ms. E. Haskel inquired of  the reason for  the Tacoma smelter to
use a high arsenic feed material and of the possibility of other
smelters increasing the arsenic content in their feed materials.
Mr. Vervaert explained the reason  for  the Tacoma smelter to process
high arsenic feed materials.  The  Tacoma  smelter is a custom smelter,
and it  has the only arsenic plant  in the  United  States.  It produces
arsenic as a by-product from  the recovered flue  dust containing  arsenic
and markets it.

     Mr. Vervaert also explained that  the only  other custom smelters
are ASARCo's smelters  at Hayden and  El  Paso.  ASARCo purchases  ore
concentrates and other material  containing copper and relatively high
arsenic on the world market.  It treats higher arsenic material  at the
Tacoma  smelter.

     Several questions about  the health/risk associated with the
standard were raised by Messrs.  W. Reiter and C.  Smith and Ms.  E.  Haskel.
They asked for the details of the  health/risk assessment mentioned in
the EPA presentation,  including information on the  dispersion modeling
and the respiratory cancer study.

     During the discussion, Mr.  Vervaert  told the committee members
that the dispersion estimates for  the  Tacoma smelter were obtained
using the  Industrial Source Complex  Model,  and the  population exposure
estimates were obtained using a population distribution model available
to EPA.  The modeling  results for  ASARCo-Tacoma  extended for a radius
of 60 km and included  about 1.8 million people as far away as the
Seattle area.  The data were  obtained  very recently and had not  yet
been compared with the actual measurements  made  by  ASARCo and obtained
by PSAPCA.

     Mr. Vervaert also stated that the analysis  of  the ambient arsenic
concentration data for the copper  smelters  shows a  significant risk
associated with the ASARCo-Tacoma  smelter.   He  indicated that the
documentation of health effects is contained in  the Health Effects
Document,  and he did not know if the document has data for the Tacoma
smelter.   EPA will have copies  of  document available upon request.


                                  11-32

-------
                                 D. CORRESPONDENCE

  1.  NAPCTAC Member William Reiter
         emical
      Corporate Environmental Affairs
      P.O. Box2332R
      Morristown, New Jersey 07960                       i_  -,->   inQ<
                                         March  2.3,  lyoi
Mr. Don Goodwin
Director,  Emission Standards  & Engineering Division
Office of  Air Quality Planning & Standards
U.S. Environmental Protection Agency
Research Triangle Park, NC  27711

       EDITOR'S NOTE:

                   NAPCTAC member William  Reiter was unable to attend the
       meeting on March 18, 1981,   So that he could  contribute his views to EPA
       and  fellow Committee members, Mr, Reiter wrote a lengthy letter to the
       chairman.  The contents of  that letter have been divided by subject and
       are  included in the relevant sections of the  minutes,  The portion of the
       letter that applies to this section follows.
1)  NESHAP  Arsenic Emissions—Primary Copper  Smelters

    I  support your proposal to discontinue  the  development  of  the
    NESHAP.   Since the NESHAP  has become specific to the Arsarco
    plant  in Tacoma, Washington,  it has become  site-specific and
    does  not represent a  national standard.   I  concur with
    your  position that it  is  not  efficient  to spend the money  to
    complete the document.

    I  believe,  however, that  some assurance must be obtained that
    Arsarco  will complete  the  installation  of the facilities that
    they  are now proposing  to  install.  I suggest a sit-down
    discussion  between your staff and the Washington State
    agency,  and the transferral of the draft  NESHAP as a interim
    or  preliminary guidance document to the State.   I would
    suggest  to  Arsarco that they  sign a Consent Agreement with
    the local agency to cover  the installation  as it would  have
    been  dictated by the  NESHAP.

    I  cannot support the  discontinuation of the NESHAP develop-
    ment  without a local  agreement as specified above.  The
    commitment  that the company has made, based on the infor-
    mation  you  have provided,  appears insufficient to insure
    compliance.
                                   11-33

-------
                       III.  CONTROL TECHNIQUES  GUIDELINES
                                Mr, Fred Porter
                   Emission Standards and Engineering Division
                       U,S Environmental Protection Agency
                   Research Triangle Park, North Carolina 27711
     The 1977 Clean Air Act Amendments required each State in which there
were areas exceeding the national ambient air quality standard for ozone
to submit revised State implementation plans to EPA by January 1, 1979.
States which were unable to demonstrate attainment with the national ambient
air quality standard for ozone by December 31, 1982, could request extensions
for attainment with the standard.  Twenty-three States and the District of
Columbia requested, and were granted, such extensions as shown in Figure 1.
The Clean Air Act now requires these States to submit revised State imple-
mentation plans to EPA by July 1, 1982.

     In order to meet the requirements of the Clean Air Act, revised State
implementation plans must provide for implementation of reasonably
available control technology, or RACT, on stationary sources of air
pollution.  As shown in Figure 2, RACT has been defined as the lowest
emission limitation that a particular source is capable of meeting by
application of control technology that is reasonably available considering
technical and economic feasibility.
                                 III-l

-------
     To aid States in revising their State implementation plans, the
Clean Air Act directs EPA to develop and issue various guidance materials.
Among these guidance materials are the Control Techniques Guideline
documents, or CTG's.

     The purpose of CTG's is to provide States and local air pollution
control agencies with an initial information base for proceeding with
development and adoption of regulations which reflect RACT for specific
stationary sources of volatile organic compound, or VOC, emissions.  Thus,
CTG's review existing information concerning the performance and cost of
various VOC emission control techniques applicable to a particular
stationary source category.  In addition, CTG's identify emission control
techniques and emission limitations which are considered the "presumptive
norm" broadly representative of RACT for the particular stationary source
category examined by a CTG.

     CTG documents also include "model" regulations.  These "model"
regulations constitute the RACT recommendation of each CTG document,
and  are included solely to assist and guide State and local agencies
in development of their own RACT regulations for specific stationary
sources.

     CTG documents are developed in a manner generally analogous to that
employed for development of new source performance standard background
information documents.  The industry in question is surveyed to characterize
VOC  emissions and to identify various emission control techniques that
                                III-2

-------
                                    3
are, or could be, applied to reduce emissions, the performance of these
emission control techniques, and their costs.   Through the use of "model
plants," generally representative of existing  plants within the industry,
an assessment is then made of the performance  and cost of emission
control.  This assessment serves as the basis  for the selection of RACT.

     CT6 documents, therefore, are, of necessity, general in nature and
do not fully account for the complete range of variations which may be
found within a stationary source category.   The selection of RACT contained
in a CTG document is only a recommendation  and a number of reasons may
exist for regulations developed by States to deviate from this recommendation.

     On the other hand, CTG documents are considered part of the State
rulemaking record which EPA considers in reviewing revised State implementation
plans.  The RACT recommendation along with  the other information contained
in a CTG document, therefore, is highly relevant to EPA's decision to
approve or disapprove a State implementation plan revision.

     Where a State adopts regulations significantly different from
the RACT recommendation in a CTG document it must include documentation
with the State implementation plan revision to support and justify these
RACT regulations.  Regulations which the State can demonstrate satisfy
the definition of RACT, will be approved by EPA.

     The first group of CTG documents, which were published in 1977,
covered fifteen stationary source categories of VOC emissions.  The
second group of CTG documents, published in 1978, covered an additional

                                 III-3

-------
                                   4
eight stationary source categories.  As Shown in Figure 35 the third
group of CTG documents, which will be published this year, will cover
eight stationary source categories.
     The first two groups of CTG documents were not brought before the
National Air Pollution Control Techniques Advisory Committee for discussion;
although they were distributed to some extent to various industrial contacts
and trade associations in draft form for comment prior to being published.  In
an effort to increase public participation and gain greater review and
comment on the third group of CTG documents, they are being brought before
this Committee for discussion.

     Today and tomorrow, the Committee will have the opportunity to discuss
draft CTG documents for petroleum solvent dry cleaning, volatile organic
liquid  storage vessels, and fugitive emissions from synthetic organic
chemical manufacturing plants.  As shown in Figure 3, we plan on bringing
CTG documents for manufacture of styrene-butadiene copolymers, and
fugitive emissions from natural gas and natural gasoline processing
plants  to a National Air Pollution Control Techniques Advisory Committee
meeting on April 29 and 30.  We plan on bringing the remaining three CTG
documents in group three - offset lithography printing, air oxidation
processes within synthetic organic chemical manufacturing plants, and
manufacture of polymers and resins - to a National Air Pollution Control
Techniques Advisory Committee meeting tentatively scheduled for June 3 and 4.
                                     III-4

-------
              FIGURE 1
STATES WITH EXTENSION  N 0 N - A T T A I N PI E N T  AREAS
     CALIFORNIA



     COLORADO




     CONNECTICUT




     DELAWARE



     ILLINOIS




     INDIANA



     KENTUCKY



     MARYLAND



     MASSACHUSETTS




     MICHIGAN



     MISSOURI
NEW JERSEY




NEW YORK




OHIO




OREGON




PENNSYLVANIA




RHODE ISLAND



TENNESSEE




TEXAS



UTAH




VIRGINIA



W A S H I N 6 T 0 N
         DISTRICT OF  COLUMBIA
                  III-5

-------
                             FIGURE    2
                             -  R  A   C  T  -
LOWEST    EMISSION    LIMITATION   THAT
A    PARTICULAR    SOURCE    IS    CAPABLE    OF
MEETING    BY    THE   APPLICATION   OF
EMISSION   CONTROL    TECHNOLOGY    THAT
 IS    REASONABLY    AVAILABLE   CONSIDERING
TECHNICAL    AND    ECONOMIC    FEASIBILITY,
                                    III-6

-------
              FIGURE 3
SOURCE
  N A P C T A C
PET,  SOL,  DRY  CLEANING
VOL STORAGE VESSELS
SOCHI-FUGITIVE
MARCH 17-18
STYRENE-BUTADIENE COPOLYMERSAPRIL 29-30
NAT, GAS/GASOLINE  FUGITIVE
SOCMI - AIR OXIDATION
POLYMERS/RESIN
HEATSET OFFSET  LITHOGRAPHY
JUNE
                  III-7

-------
                             D. tUKKtiKUINL'tlNLt
     1.  NAPCTAC Member Vim 1am Reiter
      Allied.
      Chemical
      Corporate Environmental Affairs
      P.O. Box 2332R
      Morristown, New Jersey 07960                   ..   ,   __   ,«n-
                                         March  23,  1981
Mr. Don Goodwin
Director,  Emission Standards  & Engineering Division
Office of  Air Quality Planning & Standards
U.S. Environmental Protection Agency
Research Triangle Park, NC  27711

       EDITOR'S NOTE:

                   NAPCTAC member William Reiter was unable to attend the
       meeting on March 18, 1981,   So that he could contribute his  views to EPA
       and  fellow Committee members, Mr. Reiter wrote a lengthy letter to the
       chairman.  The contents of  that letter have been divided by  subject and
       are  included in the relevant sections of the minutes,  The portion of the
       letter that applies to this section follows,


2)  Control Technology Guidelines - General

    As I said on March 17,  I  must compliment the  EPA on ini-
    tiating Control Technology Guideline document  review.  I
    believe that your efforts in development of  CTG's  can be made
    more efficient by the  following:
       a)    The CTG should  not be a re-statement  of the New Source
             Performance Standard  (NSPS).   I believe the requirements
             are not identical and I think it is a  very significant
             technical and administrative  error to  simply  re-publish
             portions of the NSPS  documentation as  a CTG.   Should the
             States desire to  assess RACT  as BACT,  that should be
             their perogative.  It should  not be dictated  by EPA.

       b)    I would suggest that  distribution of  the  draft CTG be
             limited until after the NAPCTAC review or  if  you desire
             (or are actually  obtaining)  comments  from  the Regions
             and the States  that the caveats limiting  use  of the
             draft CTG be enlarged.  I  believe there should be a
             clear statement on the cover and on the title page that
             this is a working document,  not to be  used in developing
             the SIP.

             In addition, I  would  suggest  that the  Introduction (1.1)
             be modified to  remove the  ambiguity as to  the use  of the
             document.  The  last paragraph of the  Introduction states
             that "the CTG is  a working draft".  It says,  however,
             that it has as  its objective the proving of the oppor-
             tunity for public review and comment  and to "provide as
             much assistance and lead time as possible  to  State and
             local agencies  preparing RACT regulations	".

                                   III-8

-------
         I believe this direction to be wrong.  The  draft  docu-
         ment should be provided with peer and  public  review
      .   before it is used by the states in preparing  RACT  regu-
         lations.  The limitations that you have  identified in
         the past are insufficient to prevent the  draft  document
         from being used as an Agency dictate.  I  believe  the
         comments by Carl Beard, Eric Lemke, and  Jim Lents
         clearly indicated that the Regions are forcing  the
         States to adopt the CTG as you issue it.  In  some  cases,
         this includes the draft document (refer  the past
         Region V problem) .

    c)   I submit that EPA has the responsibility  for  technically
         establishing RACT.  However, I do not  believe,  and
         strongly recommend against, the position  that the  model
         regulation is to be adopted as identified in  the  docu-
         ment.  Referring to the Introduction again, and I  quote,
         "the CTG document, however, is a part  of  the  rule-making
         record which EPA considers in reviewing  revised SIP's
         and the information and data contained in the document
         is highly relevant to EPA's decision to  approve or
         disapprove SIP revision".  That statement forces accep-
         tance of the model regulation.

          I  believe  the  model  regulation  should  be general  in
          nature  and  should  provide  a  range  of  control  levels that
          may  be  used  by  the State.   That  range  should  cover areas
          more  stringent  and more  liberal  than  EPA's  opinion.

          In  addition,  I  strongly  recommend  that recordkeeping  and
          enforcement  be  at  the  States' discretion.  The  develop-
          ment  of a  large mass of  records  sent  to  the States may
          be worthless  in that they  do  not have  the staff to
          review  them.  This places  a  cost burden  on  industry and
          the States.

          I  suggest  that  a self-monitoring approach be  applied  to
          industry where  sufficient  records  to  satisfy  State
          enforcement  are maintained at the  plant,  available for
          State or EPA  review and  that  a  simple  document  certified
          by the  plant manager be  forwarded  to  the State  reporting
         on compliance.
                                        Best regards,
                                        W. M. Reiter
                                        Corporate Director
                                        Pollution Control
cc:  NAPCTAC Members
                                 III-9

-------
                   IV, CONTROL TECHNIQUES GUIDELINE'FOR
      VOLATILE ORGANIC COMPOUND EMISSIONS FROM PETROLEUM DRY CLEANERS

                          A, EPA PRESENTATION

                        Mr, Steven 0, Plaisance
                           TRW Incorporated
                 3200 E''. Chapel Hill Road/Nelson Hwy.
             Research Triangle Park, North Carolina 27709
     A control techniques guideline document, or CTG, has been developed
for the regulation of volatile organic compound emissions, or VOC, from
petroleum dry cleaners.  This presentation will briefly characterize the
petroleum dry cleaning industry, typical  dry cleaning process procedures,
and the VOC emissions associated with particular items of dry cleaning
equipment.  In addition, the reasons for  which this source category was
selected for development of a CTG will be outlined, and the emissions
control equipment selected as representing reasonably available control
technology, or RACT, will be described.  Finally, the basis for the
selection of the provisions of the CTG model regulation will  be delineated,
focusing, in particular, on the savings in annnualized operating costs
and VOC emissions reductions resulting from implementation of the model
regulation based on RACT in a typical large industrial plant.  (Figure 1)

     The petroleum solvent dry cleaning industry is a part of the overall
dry cleaning industry which also includes facilities using both perchloroethylene
and trichlorotrifluoroethane solvents. While the chemical composition
of petroleum solvent varies between manufacturers and individual production
runs, it typically comprises a range of hydrocarbons with chemical
properties similar to those of kerosene.   As a result of petroleum
solvent's relatively flammable nature, self-service or coin-op facilities
are not permitted under fire regulations, and existing plants are limited
to commercial or industrial operations.

     Approximately 30 percent of the total dry cleaning industry throughput
of 750,000 Mg of articles cleaned per year is processed in petroleum
solvent.  Moreover, about 30 percent of this petroleum solvent industry
throughput is processed in commercial plants which cater, almost exclusively,
to the cleaning of personal clothing.  Typically, these plants clean
less than 100,000 kilograms of articles per year.

     Industrial facilities process approximately 70 percent of the total
petroleum solvent industry throughput, and clean large volumes of articles
ranging from dust mops to rental uniforms.  Throughputs of over 100,000 kilograms
of articles cleaned per year are characteristic of this segment of the
industry.
                                      IV-1

-------
                                                               TYPICAL DRY CLEANING SYSTEM
              DRY CLEANING  INDUSTRY
      CLOTHING THROUGHPUT AND EMISSIONS
                                                                 OMy
I
1X3
                      Total Dry Cleaning Industry

                         (750,000 Mg/yr)
                         etroleum Solven
                        Dry Cleaning Industry
                       •0% OF TOTAL INDUSTRY
                          THROUGHPUT
                    Commercial'    Industrial
                    THROUGHPUT"/  7O* °" PE™OLEUM
                    THROUGHPUT iNDUSTRY THROUGHP
                       ISSION9 AND EMISSIONS
 WASHER
  o
                                                                              ArVcht
                                                                         Sotod
                                                                         Solvent
         FLTER
                                                                        I	>
STORAGE

 TANK
             Purtltod So(v«nt
                        0.4% OF NATIONWIDE
                   STATIONARY SOURCE VOC EMISSIONS
                           Figure 1
          Figure 2

-------
     The petroleum solvent dry cleaning industry produces approximately
0.4 percent of the nationwide stationary source VOC emissions, thereby
contributing significantly to ambient photochemical oxidant levels.
Petroleum dry cleaning facilities are typically found in populated areas
that frequently coincide with National Ambient Air Quality Standards
non-attainment areas.  Furthermore, it has been determined that the
reduction of VOC emissions from larger facilities can be accomplished
with actual reductions in annualized costs to the individual plants.
Thus, the petroleum solvent dry cleaning industry was chosen as one of
the source categories for which the EPA will  write a control techniques
guideline document.  (Figure 2)

     A typical petroleum solvent dry cleaning facility contains one or
more washers, dryers, solvent filters or settling tanks, and solvent
stills.  This equipment serves the dual purpose of cleaning the dirty
articles and rejuvenating the cleaning solvent.  While the washer and
dryer clean and dry the processed articles, the filter or settling tank,
and vacuum still remove dirt and other contaminants from the solvent
prior to its re-use.  (Figure 2-A)

     Petroleum solvent washing is similar to home water-washing, with
articles being agitated in a bath of solvent, rinsed in new or rejuvenated
solvent, and then spun at high speed to remove the excess solvent.  VOC
emissions from washing operations are limited to fugitive or unspecified
emissions from washer gaskets, pumps, and access covers, as well as from
the transfer of loads of solvent-laden articles.

     After washing, solvent-laden articles are transferred to a standard
dryer which operates in a manner similar to that of a home clothes
dryer.  The drying articles are tumbled in a performated drum, and
ambient air is heated by steam coils in a steam chest, drawn into the
dryer, and then circulated through the tumbling articles.  Solvent
evaporated from the articles enters the air stream in the tumbler and is
continuously exhausted to the atmosphere.

     Dryer VOC emissions result, almost exclusively, from this atmospheric
exhaust.  EPA tests have found these dryer emissions to range from 10 kg
to 28 kg of VOC per 100 kg dry weight of articles dry cleaned.  (Hereafter
these emissions will be referred to simply as "kilograms.").  Finally,
dryer fugitive VOC emissions result from the eventual evaporation of
solvent remaining in the articles after drying.  (Figure 2-B)

     In many commercial and industrial facilities, the process of solvent
rejuvenation begins when soiled solvent from the washer is pumped to a
diatomite filter, where solid contaminents such as lint and dirt, as
well as other non-solvent soluble contaminents, are removed by entrapment
in the porous surface of filtration tubes coated with diatomaceous
earth.  Typically, waste material containing dirt, diatomaceous earth,
and solvent is removed from the filter daily.  Industry tests have found
that this disposed waste can contain from 5 kg to 10 kg of solvent.
Also, fugitive VOC emissions can result from the filter piping system
and transfer of this waste.
                                     IV-3

-------
 TYPICAL DRY CLEANING SYSTEM
       AND VOC EMISSIONS
         WASHER AND DRYER
    (Emissions In kg VOC p«r 100 kg Article* Cl«»n«d)
      Fugitive
    Dryer
i Atmospheric
S  Exhaust
e 10 kg- 28 kg

Dirty ^
Articles W
New and

WASHER
O

i
rugiuve
0 Washed f\
Articles V

DRYER
O


i Fugitive
f\ Dry
L/ Articles

Purified Solvent Soiled Solvent
from Storage to Fitter or
Settling Tank
              Figure 2-A
     TYPICAL DRY CLEANING SYSTEM
           AND VOC EMISSIONS

          FILTER AND SETTLING TANK
        (Eml**lon* In kg VOC p«r 100 kg Article* Cl«an«d)
                 Fugitive
 From
Storage
                        Fiftration Waste
                         5 kg-10 kg
                SETTLING TANK

                (Large Industrial)
                 Figure 2-B
                                    To
                                 Vacuum Still
                  IV-4

-------
     Large industrial facilities with heavy soil loadings usually omit_
filtration, because removal of filtration wastes would have to be carried
out on a nearly continuous basis.  Instead, these facilities employ one
or more solvent settling tanks which partially remove solids from the
process solvent stream.  Fugitive VOC emissions typically result from
the atmospheric venting of these tanks.  (Figure 2-C)

     Following filtration or settling, the solvent is usually piped to a
vacuum still where it is boiled under a vacuum and then condensed,
leaving behind a solvent-laden liquid containing solvent soluble contaminants
such as grease and oil.  Periodically, this residual  liquid is boiled at
higher temperatures to evaporate additional solvent,  with the remaining
waste liquid being removed for disposal.  EPA and industry tests have
found that this waste liquid can contain from 1 kg to 7 kg of solvent.
Furthermore, the still mechanical components and waste transfer operations
can contribute to the plant's overall fugitive emissions.

     Finally, the purified solvent is pumped to a holding tank where new
solvent is added on an as-needed basis.  This tank, like the settling
tank, usually produces fugitive VOC emissions from its atmospheric vent.
(Figure 2-D)

     Fugitive VOC emissions, as previously described, result from practically
all of the equipment and transfer operations in a typical petroleum dry
cleaning facility.  While these emissions are difficult to pin-point and
quantify individually, industry and EPA tests have found their total to
range from about 0.5 kg to 3 kg, depending on the plant size, throughput,
and effectiveness of maintenance.  (Figure 3)

     The total uncontrolled VOC emissions in a typical petroleum solvent
dry cleaning facility ranges from about 16.5 kg to 48 kg, and represents
a summation of emissions from dryers, filters, stills, and fugitive
sources.

     Dryer emissions account for about 65 percent of  the total plant VOC
emissions, with a nominal value of 18 kg.  Filtration system emissions
contribute about 28 percent of the total plant VOC emissions in facilities
utilizing filtration, with the nominal uncontrolled value being 8 kg.

     Distillation and fugitive source emissions each  account for 3.5 percent
of the remaining total emissions in plants employing  solvent filtration,
with a nominal value of 1 kg for distillation and 1 kg for fugitive
sources.  In a large industrial plant that does not employ solvent
filtration, however, the nominal value for distillation emissions would
increase to 7 kg, or 25 percent of the total plant emissions, and that
of fugitive emissions would increase to 3 kg, or 10 percent of the total
plant emissions.
                                    IV-5

-------
                TYPICAL DRY CLEANING SYSTEM

                      AND VOC EMISSIONS

                   VACUUM STILL AND STORAGE
                   (Eml**lon* In kg VOC per 10O kg Article* Cleened)
                      WASHER          | r\ I DRYER
I
cr>
        Fugitive.
                    STORAGE

                     TANK
Purified
Solvent
                                                 .Fugitive
                                            VACUUM

                                             STILL
                           Figure 2-C
           > Still Waste
           ?1 kg -7 kg
                                                               TYPICAL DRY CLEANING SYSTEM
                                                                     AND VOC EMISSIONS
                                                               	FUGITIVE EMISSIONS	
                                                                   (Emlnlon* In kg VOC per 1OO kg Article* Cleened)
                                              Total Fugitive
                                              0.5 kg-3 kg

                                                                           Figure 2-D

-------
     Thus, the nominal value for the total uncontrolled emissions from
both commercial and industrial plants is 28 kg.  This value, when applied
to a typical large industrial plant that annually cleans 435,000 kg of
articles, results in total plant VOC emissions of about 122,000 kg per
year.  (Figure 4)

     The selection of the equipment representing reasonably available
control technology is based on the reduction of VOC emissions from the
two most significant sources in a typical petroleum solvent dry cleaning
plant: the dryer, and the solvent filter.  Dryer VOC emissions would be	,
reduced with the installation of a solvent recovery "dryer, and solvent
ftttratlon emissions would be reduced witn the installation of a cartridge
filtration system.  (Figure 5)

     The domestically manufactured solvent recovery dryer functions like
a standard petroleum solvent dryer that has been fitted with a water-
cooled condenser to liquify solvent vapors by condensation.  The currently-
marketed unit has two operating cycles, recovery and exhaust, which
automatically dry articles while reclaiming the dispelled solvent.

     When the dryer is operating in its solvent recovery cycle, an
automatic damper diverts heated air through the dryer tumbler where it
evaporates solvent from the tumbling articles.  This mixture of heated
air and solvent vapor passes through a lint filter and blower, and is
diverted to the condenser by a second automatic damper.  The water-
chilled surface of the condenser lowers the temperature of the incoming
vapor stream to the point at which solvent and water vapors condense.
The condensed liquid is collected and piped to a gravimetric separator
where the liquified solvent and water are separated due to the difference
in their densities.  Finally, the vapor stream exits the condenser and
is re-heated by steam coils in the steam chest, and the cycle of evaporation
and condensation is resumed.

     The period of recovery drver operation in which Pvarnra*QH g"1"°nt
is reclaimed is followed by the exhaust cycle.  In this cycle, an automatic
damper allows ambient air to be drawn into the tumbler, circulated~
around the tumbling articles, and then^xhaust.Pd dirprtlv to the atmosphere
through the second automatic damper, without passing through the condenser
or steam chest.  This typically brief cycle permits the rapid removal of
residual solvent prior to the removal of the articles from the dryer.

     Recovery dryer atmospheric emissions occur during the exhaust
cycle, and the quantity of VOC emitted to the atmosphere is a function
of the solvent content of the articles at the end of the recovery cycle.
Recovery dryer VOC emissions were evaluated in two EPA tests, and average
emissions were found to range from about 1 kg in a large industrial
facility cleaning work gloves to about 4 kg in a commercial plant
cleaning synthetics.  (Figure 6)
                                     IV-7

-------
       UNCONTROLLED EMISSIONS
                                                              BASIS FOR  RACT
  Source
                 Nominal
          Range   Value
Dryer      10 - 28   18
Filtration    5 - 10(0)* 8(0)"
Distillation   1-7     1(7)*
Fugitives  0.5-3     1(3)"
TOTAL
                            Percent of Total
                            Plant Emissions
                                  65
                                  28 (0)*
                                  3.5 (25)"
                                  3.5(10)*
          16.5-48   28
     * lndu»ul«l Plant
                                  100
  Typical Large Industrial Plant -

    435 QQOk9 cleaned  x 28 kg VOC emitted
       '      year          100 kg cleaned
           , 22,000
                   k9
                          •mltted
EMISSION SOURCE

  •  Dryer
  •  Filter
RACT EQUIPMENT

 • Recovery dryer
 • Cartridge filter
                        y Gal
                  Figure 3
                                                                      Figure 4
SOLVENT RECOVERY DRYER
MOSPHERIC AIM INTAKE
• _
rr-i HEATED S
D""E"Krj 	 "A« 	 c
Tumbler
i;
Lint
Filter
!i 	 £.
	 g
F

team '•
:hest " 	 i
i
LIGEHD
— — MECOVEMT CTCLI
• mm EIHAU1T CrCLf
Condenser
t
i
i i—

Recovered
WBt«f
DAMPEM) s«D*r«t«r
AIM t »olvtMT_[J"J
) i"o!( nr
--< T
fjQf ATUOSPHEMIC
EIHAUIT
:igure 5 §uojm
»
T

«ATEM
                                                 Paper filter-*
                                                                CARTRIDGE FILTER
                                                                          Activated
                                                                            carbon
                                                    Carbon-Core Cartridge        All-Carbon Cartridge

                                                                        Figure 6
                                                  IV-8

-------
     The cartridge filter decreases filtration system VOC emissions by
drastically reducing both the solvent content of disposed filtration
waste and the frequency of waste disposal.   A typical cartridge filtration
system employs a two-stage filtration process to prolong the service
life of the filter cartridges.  Soil-laden  solvent is pumped first
through carbon core cartridges that entrap  solids in layers of filtration
paper and partially remove solvent soluble  contaminants such as fugitive
dyes in their central cores of activated carbon.  Then, the solvent is
pumped through one or more all-carbon cartridges in which additional
solvent-soluble contaminants are removed.

     VOC emissions from the cartridge filter occur when the cartridge
elements are replaced.  Typically, the used cartridges are heavily
coated with solvent-laden dirt and lint, and atmospheric VOC emissions
result from the disposal of these cartridges.  EPA and industry tests
have found these emissions to be 1 kg or less.  In addition, results of
an EPA test indicate that drainage of the used cartridges in their
sealjdjhousings for at least'8'hours prior  to disposal wmil-l pn?rli,ico—
fuftfier fedTjctTons in vui emissions of as much as 40 percent.., based on
decreasing tne solvent content of these disposed cartridges.  (Figure 7)

     The control techniques guideline document contains a model regulation
based on RACT that would apply only to large petroleum dry cleaning
facilities that annually consume at least 123,000 liters of petroleum
solvent, as indicated by their solvent purchases.

     Dryer VOC emissions would be reduced by the installation of a
solvent recovery dryer which would be operated with a condenser vapor
outlet temperature not in excess of 34°C, and with a recovery cycle
duration sufficient to attain a final recovered solvent flow rate not^in
excess of 0,02 liters per minute.

     Emission control devices other than the recovery dryer could be
used, provided that their maximum VOC emissions do not exceed 2.4 kg.

     VOC emissions from filtration wastes in facilities using solvent
filtration would be reduced with the installation of a cartridge filtration
system.  The filtration cartridges should be drained in their closed
housing for at least 8 hours prior to disposal.

     Alternative filtration emission control devices could be used,
provided that their maximum VOC emissions do not exceed 1.0 kg.

     Vacuum still VOC emissions would be reduced by storing all vacuum
still wastes in a manner that minimizes VOC emissions to the atmosphere
in the plant.

     Fugitive VOC emissions from miscellaneous sources would be reduced
by promptly repairing solvent liquid and vapor leaks.
                                     IV-9

-------
        MODEL REGULATION
     (EmUiloiw Iff kf VOC »«r 1OO kf Arltcl** Cl«an«4>


• ^123,000 liters Solvent per year for Regulation

• Solvent Recovery Dryer with Condenser Vapor
  Outlet Temperature < 34°C and Final Recovered
  Solvent Flow Rate S 0.02 llters/mln.

• Dryer Alternative Control Emissions < 2.4 kg

• Cartridge Filter and Cartridge Drainage  -8 hours

• Filter Alternative Control Emissions £1.0 kg

• Minimize Still Waste Emissions  by Proper Storage

• Minimize Fugitive Emissions by Improved
  Maintenance and Operating Procedures

• Compliance in 19 Months or 3 Months after
  Control Equipment Delivery
                                                          DERIVATION OF THE
                                                          MODEL REGULATION
                                                               k« VOC p«r 1OO k| Article* Cl««n«d)
                                                      BASIS                   STANDARD

                                                          Plant Size to be Regulated
                                                  • Control of Sources
                                                   with Emissions
                                                   6100 tons per year


                                                   Control of up to
                                                   70% of Total Petroleum
                                                   Industry Emissions
                     - Annual Plant Solvent
                      Consumption
                      £123,000 liters
                      per year for
                      Regulation
                  Figure 7
                                                                      Figure 8
         DERIVATION OF THE
         MODEL REGULATION
  (Emu. 10". k> k« VOC »« 100 >

BASIS
                           clM CL.n.dl

                             STANDARD
           Dryer Emissions

 Minimize Emissions         - Recovery Dryer

' Minimize Control Costs

       Recovery Dryer Compliance

• Minimize Emissions        - Condenser Vapor
                           Outlet Temperature
 Maximize Solvent          < 34°c
 Recovery
                          - Final Recovered
 Maximize Safety           Solvent Flow Rate
                           <0.02 liters/minute

   Alternative Dryer Emission Control Device

• Emissions Rate Equivalent    - Emissions Rate
 to the Recovery Dryer        < 2.4 kg
    DERIVATION OF THE
    MODEL REGULATION
  (EmUilon* k* k« VOC »«r 1OO k« A/uclM Cl*an«
-------
     Final compliance with the regulation would be achieved no later
than 19 months after implementation of the regulation.  Delays in control
equipment delivery could postpone final compliance to 3 months after
delivery of the control equipment.  (Figure 8)

     The selection of the provisions of the CTG model regulation is
based on reducing VOC emissions from the four primary sources in a
petroleum solvent dry cleaning facility - dryers, solvent filters,
vacuum stills, and fugitive sources.  The model regulation based on RACT
would be implemented in facilities that consume at least 123,000 liters
of solvent per year.  This minimum solvent consumption would include all
facilities emitting 100 metric tons or more of VOC annually, and would
represent up to 70 percent of the total uncontrolled VOC emissions from
the petroleum dry cleaning industry.  (Figure 9)

     Dryer emission control equipment evaluated for RACT included both
solvent recovery dryers and carbon adsorbers.  In contrast to the chilled
condensation process of recovery dryers, carbon adsorption reduces VOC
emissions from existing, standard dryers by entrapping and condensing
solvent vapors in vessels containing activated carbon.  While both
devices have been shown to be capable of producing significant reductions
in dryer VOC emissions, the_recoverv drver has proved to HP mnrp rnst
effective, with lower annuaTTzed operating costs and lower overal 1
emissions tnan tne carbon adsorber."          "               '

     Minimization of VOC emissions from the recovery dryer requires the
optimization of recovery cycle thermodynamic conditions, as well  as
maintaining a recovery cycle duration sufficient to insure adequate
solvent recovery.  The results of an EPA test of a recovery dryer have
shown that the condenser vapor outlet temperature and the recovered
solvent flow rate at the end of the recovery cycle can be indicative of
optimum operating conditions.  Moreover, the test results indicate that
the safety and solvent recovery efficiency of the dryer are enhanced at
condenser vapor outlet temperatures not exceeding 34°C.  Also, VOC
emissions were minimized when the recovery cycle was continued until the
recovered solvent flow rate had fallen to 0.02 liter  per minute.

     Dryer emission control devices other than the recovery dryer should
be capable of producing emissions reductions at least equal to those of
the recovery dryer.  Accordingly, the maximum alternative control device
emission of 2.4 kg was established as representing the average recovery
dryer emission rate determined in two EPA field tests.  (Figure 10)

     Cartridge filters were the primary solvent filtration emissions
control equipment considered for RACT.  Based on the results of industry
and EPA tests, it was found that cartridge filtration systems could
reduce filtration VOC emissions to 1 kg or less, while simultaneously
decreasing annual filtration system operating costs.  Test results
further indicated that the final solvent content of disposed cartridges
could be reduced by as much as 40 percent by allowing the used cartridges
to drain in their sealed housing for 8 hours or more.
                                     IV-11

-------
     Alternative emission control devices that might be used instead of
cartridge filters should be capable of producing VOC emissions no greater
than those attributed to the cartridge filtration system.  The maximum
RACT emission limit of 1.0 kg for filtration emission control devices
other than the cartridge filter represents the maximum emission rate
derived from industry and EPA tests of cartridge filtration systems.
(Figure 11)

     The reduction of vacuum still waste emissions is based on the
contrr~of emissions from still waste that is stored ..AjLtne^ary rTEan_j ng
               itc r^wwaT  Proper storage of this waste in containers
that will minimize its exposure to the atmosphere would be a cost effective
method of reducing still waste emissions in dry cleaning facilities,
without changing vacuum still operating parameters or procedures.
(Figure 12)

     Control of fugitive source emissions is based on the prompt identification
ajid repair- nf snlypnt liquid nr vnrnr IrnK   Proper maintenance and
operating procedures could reduce VOC emissions from these sources
without incurring significant additional costs or requiring any equipment
modifications.  (Figure 13)

     Compliance with the model regulation requires prompt implementation
of its provisions, while allowing for delays in control equipment delivery.
Typically, a duration of 19 months from adoption of the regulation to
final compliance would be sufficient.  If, however, delays in final
compliance result from control equipment backorders or delivery postponement,
the deadline for final compliance would be delayed until 3 months after
delivery of the control equipment.   (Figure 14)

     Implementation of the model regulation based on RACT would reduce
total plant VOC emissions from about 28 kg to less than 4.4 kg.  Dryer
emissions would be reduced from 18 kg to 2.4 kg.  In plants with existing
diatomite filters, filtration emissions would be reduced from 8 kg to
1 kg or less.  On-site VOC emissions from solvent stills would be
nearly eliminated with the proper storage of still wastes.  Finally,
fugitive emissions in plants using solvent filtration would be reduced
from 1 kg to less than 1 kg, while those in plants using settling tanks
would be reduced from 3 kg to less than 2 kg.

     A large industrial plant that cleans approximately 435,000 kg of
articles per year and emits about 122,000 kg of VOC. a'nnually, would have
total annual controlled emissions of approximately 20,000 kg, as a
result of the installation of RACT equipment and the implementation of
the maintenance and operating guidelines contained in the model regulation.
Thus, the total annual emission reduction would be 102,000 kg, or 84 percent
of the total uncontrolled emissions from the plant.  (Figure 15)
                                     IV-12

-------
       DERIVATION OF THE
       MODEL REGULATION
     «m(»l»«> to k| VOC »•' tOO kf «tllcl.. Cl.«n.a)

   BASIS                   STANDARD

        Vacuum Still Emissions
                                    DERIVATION OF THE
                                    MODEL REGULATION
                                        kl VOC M> <«° kl »«««l«« Cl»i»
Source
Dryer
Filtration
Distillation
Fugitives
TOTAL ,
Lwg» todmntal Ptonl
Typical Large
Uncontrolled RACT
18 2.4
8 (0)" <1.0 (0)*
1 (7)" 0
• 1 (3)* <1.0 (<2.0
28 <4.4

Industrial Plant -
Uncontrolled Emissions
                                                       435,000*^x28510^
                                RACT Emissions

                                435,000 >*,tL"ri x 4-4 'iM^'^d ;
                                                                                          kg VOC •mmad
                 Figure 13
                                                                     Figure 14
                                              IV-13

-------
     The total, retro-fit capital cost of RACT equipment in a large"
industrial facility that cleans about 435,000 kg of articles per year
would be approximately $79,000, and would vary based on whether the
particular facility utilizes a solvent filtration system.  The annualized
operating cost associated with RACT would be about $33,000 per year.
Based on a solvent cost of 41 cents per liter, solvent recovery would
produce an annual credit of about $36,000, resulting in a net savings in
total annualized operating costs of approximately $3,000 per year.  This
annualized operating cost is based on recovery dryer cooling water being
supplied by a cooling tower.  The alternative installation of refrigerated
chillers to supply cooling water in warmer climates would result in a
RACT equipment net annualized operating cost savings of approximately
$1,700 per year.  (Figure 16)

     The estimated annualized operating cost of existing, non-RACT
equipment in a large industrial facility is about $31,000 per year.  The
installation and operation of RACT equipment would produce a savings of
about $34,000 per year in relation to the annualized operating cost of
existing equipment.  As previously stated, the installation and operation
of RACT equipment in a large industrial plant would result in VOC emissions
reductions of approximately 100 megagrams per year.

     The cost effectiveness of RACT is defined as the difference in
existing and RACT equipment annualized operating costs per megagram of
emission reduction, and results in a savings of approximately $340 per
megagram of emission reduction.  This value would decrease to a savings
of approximately $320 per megagram of emission reduction with the installation
of refrigerated chillers instead of cooling towers.

     In summation, implementation of the CTG model regulation in a large
industrial petroleum solvent dry cleaning facility that annually consumes
more than 123,000 liters of solvent could result in a plant VOC emissions
reduction of approximately 84 percent.  In a plant cleaning about
435,000 kg of articles per year, this 102 Mg per year VOC emission
reduction could be realized with a total annual cost effectiveness
ranging from $320 saved to $340 saved per megagram of emission reduction.

     Nationwide, annual VOC emissions from the petroleum solvent dry
cleaning  industry could be reducejd by as much as 40,000 Mg as a result
of the implementation of the CTG model regulation.
                                      IV-14

-------
RACT COSTS IN LARGE INDUSTRIAL PLANT
         Cleaning 435,000 kg /year
                   (Savings)

  Capital Cost of RACT Equipment
$78,800
  RACT Equipment Annualized
    Operating Cost                   $32,800

  Annual RACT Solvent Recovery
    Credit                         ($35,800)

  Net RACT Annualized Operating
    Cost                           ($3,000)

  Net RACT Annualized Operating'
    Cost with Refrigerated Chillers
    instead of Cooling Towers          ($1,700)
                         RACT COST EFFECTIVENESS
                        IN A LARGE INDUSTRIAL PLANT
                            Cleaning 435,000 kg per year
                 (Savings)

Net Annualized Operating Cost
  of Existing (non-RACT) Equipment
                                                                                        $31,400
                    Difference Between Existing and
                      RACT Equipment Annualized Costs
                                ($34,400)
                    Total Annual VOC Emission Reduction
                      due to RACT Equipment and
                      Procedures
                                102 Mg/year
                    RACT Equipment Cost Effectiveness   ($340/Mg)
                    RACT Equipment Cost Effectiveness
                      with Refrigerated Chillers instead
                      of Cooling Towers
                                                                                       ($320/Mg)
                   Figure 15
                                                                       Figure 16
                                                IV-15

-------
                          B.   INDUSTRY PRESENTATIONS
1.  Patton, Boggs & Blow
                      Mr. Timothy A, Vanderver, Jr.
                          Patton, Boggs & Blow
                          2550 M Street, N, W.
                        Washington, D.C. 20037
                COMMENTS TO THE NATIONAL AIR POLLUTION
                CONTROL TECHNIQUES  ADVISORY COMMITTEE ON
          THE PRELIMINARY DRAFT CTG FOR "CONTROL OF VOLATILE
            ORGANIC EMISSIONS FROM  PETROLEUM DRYCLEANERS"
          Good afternoon.  My name  is  Timothy A.  Vanderver, jr.

     I am a member of the Washington,  D'.C.  law firm of Patton,

     Boggs & Blow, and have  been  asked to speak on behalf of

     the dry cleaning industry  on EPA1s draft guideline document

     on "Control of Volatile Organic Emissions from Petroleum

     Dry Cleaners."

          Bill Fisher of the International  Fabricare Institute, a

     trade association which represents primarily commercial

     cleaners, and Bud Sluizer  of the  Institute of Industrial

     Launderers, a trade association which  represents primarily

     industrial dry cleaners and  launderers,  will also be submitting

     presentations today.  We have  tried to coordinate our

     presentations in order  to  present a readily understandable

     industry-wide position.

          With one significant  exception, we believe that the

     proposed model regulation  included in  the draft CTG will

     prove to be an acceptable  and  effective means of reducing
                                IV-16

-------
PATTQN, BOGGS & BLOW




                                     _2_




       VOC  emissions  from petroleum dry cleaning operations.   I



       will discuss the one major problem area -- the mandating of



       recovery dryers  — last.   First, I would like to out-



       line those parts of the  regulation that are fully appropriate



       and  certain technical changes to the proposed model regulation,



       Bill Fisher and  Bud Sluizer will .discuss several of these



       same points in more detail and will outline several



       deficiencies in  the discussion sections of the draft



       CTG.



            At the'outset, I note that we do not agree with the



       assertion, at  page 6-1 of the draft CTG, that "the model



       regulation ... is not  to be construed as rulemaking  by



       -EPA."  However,  we hope  to be able to resolve all out-



       standing issues  without  facing the question of whether the



       CTG  constitutes  Agency rulemaking.



            As indicated earlier, there are a number of ways  in



       which the proposed model regulation imposes an appropriate



       system of controlling emissions from petroleum dry cleaning



       establishments.   Some of these are:



            —The requirement that each owner or operator of  a



       petroleum solvent filtration system install and operate



       a  cartridge filtration system is clearly justified.



       Cartridge filters are already in wide use.  Requiring



       them is economically feasible and  will achieve significant



       emission reductions.
                                  IV-17

-------
PATTO.N, BOGGS & BLOW





                                     -3-



            —The regulatory requirements concerning maintenance,



       inspection,  housekeeping,  and repairs are fully acceptable.



       They require no more than good business practices, and there



       can be no objection to imposing such requirements.



            —The provisions concerning the storage of still wastes



       are properly limited to on-site storage.   However, these



       provisions must be given a common sense interpretation, and



       clarification of them is in order.  The discussion



       of the proposed model regulation should explain that



       this provision's objective is an achievable reduction in



       the emissions from still wastes, not an absolute ban on such



       emissions.



            —The proposed model regulation permits the use of



       alternative emission control devices if they achieve



       emission reductions equivalent to the recommended technology -



       Although we have some problems with the details of the



       equivalency provisions, we believe that the concept is sound



       and should be included in the model regulation.



            --Finally, a small plant exemption,  such as that set



       forth in the proposed model regulation, is necessary.  We



       understand that EPA established the exemption level at a



       point which was designed to include the plants which produce



       the most emissions.  We urge that this cutoff level be accepted



       in the final model regulation.
                                   IV-18

-------
PATTON, BOGGS & BLOW




                                     -4-



            We  also  have several suggestions with respect to



       the  language  of  the proposed model regulation:



            In  paragraph XX.010(C)  of the model regulation, we



       suggest  that  the geographic  areas that are to be covered be



       specified in  brackets  for ease of understanding.  Following



       the  colon at  the end of the  paragraph, the phrase "[those



       areas that will  not be in attainment for hydrocarbon



       emissions after  December 31, 1982]" should be inserted.



       This phrase will clarify the geographic areas to which



       the  proposed  model regulation should be applied.



            Second,  we  suggest that the words "frequently called



       'Stoddard1 solvent" be struck at the end of the definition



      .of "petroleum solvent" in paragraph XX.020(A).   This wording



       is confusing  and adds  nothing to the definition.



            In  the definition of "vacuum still" -we suggest that



       word "passed" be substituted for the word "pumped."



            We  also  have several concerns with respect to



       the  model regulation provisions which permit the use of



       alternative technology.  Bill Fisher and Bud Sluizer will



       also comment  on  these  equivalency requirements.



            Paragraphs  XX.040(A)  and (C)  permit "the owner or



       operator" to  conduct testing which establishes equivalency-



       In some  circumstances, the manufacturer of equivalent



       equipment or  some other party may be the most logical one
                                  IV-19

-------
PATTON, BOGGS & BLOW





                                     -5-




       to  conduct this testing and the cited language should




       be  changed to permit any party to conduct the required




       testing.  This change parallels language in the proposed




       New Source Performance Standard for perc dry cleaners.




            In addition, the equivalency factor for the recovery




       dryer (2.4 kilograms of solvent emissions per 100 kilograms




       of  clothes cleaned)  is too low.  The EPA-sponsored test of
       the recovery dryer in Lakeland, Florida (which results



       are set forth in an Appendix to the CTG) shows far higher



       emissions for the recovery dryer.  Indeed, average



       emissions are over 50 percent higher than the 2.4 kilogram



       level posited in the model regulation.  Obviously,



       equivalency should mean equivalency -- not a more stringent



       standard than the recommended technology can achieve.



       Accordingly, we recommend tha t the equivalency factor for



       recovery dryers be revised upward to an appropriate level.



       Both Mr. Fisher and Mr. Sluizer will comment on this point.



            Further, we believe that there are acceptable alternatives



       to testing with the flame ionization analyzer procedure set



       forth in section XX.040(A).  Quite simply, the FIA method



       is not a reliable test method in the circumstances; other,



       more reliable methods should be acceptable alternative



       means of demonstrating alternate technology.  These include



       mass balance, manufacturers' certification, and so forth.



       Mr. Fisher and Mr. Sluizer will speak in more detail on this point.
                                  IV-20

-------
PATTO.N, 9OGGS  & BLOW



                                     -6-



            My final  point is our major concern with the proposed




       model regulation — the requirement that would mandate the




       installation of recovery dryers.   We do not feel that recovery




       dryers should  be mandated at this time.   In essence,  our




       position  is that such a mandate is not yet appropriate in




       view of the paucity of data made available by EPA.




            The  Hoyt  Manufacturing Company is to be commended for its




       efforts in developing and marketing its recovery dryer.   Efforts




       to reduce emissions and to improve the economics of dry



       cleaning operations deserve commendation, and we would like



       to extend such commendations to Hoyt.



            Nonetheless, there are significant safety concerns




       which are not  adequately answered by the draft CTG.  Earlier




       versions of the recovery dryer experienced safety problems



       and, although  these problems may have been resolved,  this
       is not clear from the draft CTG.   Further,  there are other
       safety concerns.   For example,  the test data concerning



       the LEL is totally useless;  EPA has conceded that the testing



       device failed in  one of the  two tests that it relied on,  so



       that we simply do not know whether the recovery dryer can be



       operated below this limit.  In these circumstances, we suggest



       that a recovery dryer .does not satisfy the statutory definition



       of "reasonably available control technology."  Unless all



       doubts concerning safety can be dispelled by complete and adequate



       data,  the recovery dryer is  not technologically feasible and



       thus does not qualify as reasonably available control technology.
                                 IV-21

-------
PATTON, BOGGS & BLOW




                                     -7-



            The dry cleaning industry is willing to work with EPA



       in order to demonstrate whether or not the recovery dryer



       is a safe means of emission control.  We believe that a



       joint EPA/industry test on the recovery dryer could



       promptly demonstrate its safety characteristics,  unlike



       the previous tests conducted by EPA's contractors.  In the



       meantime, we recommend that EPA either withhold publication



       of the final CTG or issue the CTG without a provision mandating



       the installation and operation of recovery dryers.  Once a



       joint test is completed, a reasoned decision can then be



       made with respect to recovery dryers.



            Both Mr. Sluizer and Mr. Fisher will discuss the recovery



      .dryer issue in more detail.  In conclusion, it should be



       noted that if the economics of the recovery dryer are as



       pictured in the draft CTG, there should be a speedy, voluntary



       conversion by the industry to recovery dryers should the



       safety issue not prove substantial.  The economic feasibility



       of recovery dryers, as outlined in the CTG, is so substantial



       as to serve as an extremely strong incentive to potentially



       affected petroleum dry cleaner owners/operators to install



       such equipment.   Accordingly, delaying publication of the



       finaly CTG or issuing the CTG without provision mandating the



       installation of recovery dryers should have minimal environmental



       effects if the economic analysis in the CTG is correct and if



       such devices can be safely operated.
                                  IV-22

-------
PATTQN, BOGGS  & BLOW





                                     -8-



            I  appreciate  the  opportunity  to make this presentation.




       If  there  are  any questions,  I  would  be happy to answer them,

-------
2.  Institute of Industrial Launderers
                        Mr. Mervyn Sluizer, Jr,
                   Institute of Industrial Launderers
                          601 Fox Croft Road
                    Philadelphia, Pennsylvania 19117
      MY NAME IS MERVYN SLUIZER, JR., TECHNICAL DIRECTOR OF THE
 INSTITUTE OF INDUSTRIAL LAUNDERERS (III)

      BEFORE STARTING, MAY I FIRST THANK THE ADVISORY  COMMITTEE
 FOR THE OPPORTUNITY TO PRESENT THESE REMARKS AND  FOR  THE SPECIAL
 CONSIDERATION OF SCHEDULING THE PETROLEUM DRYCLEANERS AT A SPECIFIC
 TIME IN YOUR AGENDA,  FOR THOSE UNAWARE OF THE REASON FOR THIS
 SECOND REMARK, MAY I STATE A WORSE TIME FOR THIS  HEARING COULD NOT
 HAVE BEEN SCHEDULED AS FAR AS WE ARE CONCERNED.   THE  ENTIRE
 DRYCLEANING INDUSTRY IS IN THE MIDST OF THEIR BIANNUAL WORLD EDUCATION
 SHOW IN ATLANTA,  THIS ACCOUNTS FOR THE ABSENCE OF  MANY PERSONS
 VITALLY INTERESTED IN THIS MATTER,  OUR PRESENCE  IS MANDATED BY
 THE FACT THAT WE, THE INDUSTRIAL LAUNDERERS, ARE  THE  SEGMENT OF THE
 INDUSTRY, BEYOND A DOUBT MOST IMPACTED BY THESE DRAFT GUIDELINES,

      THE INSTITUTE OF INDUSTRIAL LAUNDERERS, IS THE ONLY TRADE
 ASSOCIATION REPRESENTING SOLELY THE INDUSTRIAL LAUNDRY SEGMENT OF
 THE INDUSTRY,  OUR MEMBERSHIP ACCOUNTS FOR ALMOST THE ENTIRE
 VOLUME OF INDUSTRIAL LAUNDRY, CERTAINLY OVER 80%,   THERE ARE A
 GREAT MANY SMALL INDUSTRIAL LAUNDERERS  IN OUR MEMBERSHIP AND
                               -1-
                                IV-24

-------
SLUIZER SPEECH - PAGE 2

MANY OTHER FIRMS WHO ARE NOT AFFILIATED WITH US DOING SMALL
AMOUNTS OF INDUSTRIAL LAUNDERING WORK,  OUR BASIC BUSINESS IS
THE RENTAL OF WORK CLOTHING, SHOP TOWELS, GLOVES, MOPS AND
MATS TO INDUSTRY, COMMERCE AND TRANSPORTATION WITH FREQUENT
DELIVERY OF CLEAN ITEMS, PICK UP OF SOILED FOR PROCESSING AND
RETURN,
     IN A SURVEY A FEW YEARS BACK, DRYCLEANING IN OUR MEMBERSHIP
REPRESENTED LESS THAN 20% OF THE VOLUME AND OF THIS APPROXIMATELY
ONE HALF WAS PETROLEUM DRYCLEANING,  THIS PICTURE HAS NOT CHANGED
MATERIALLY DUE TO THE UNCERTAINTY OF ENVIRONMENTAL AND WORKPLACE
REGULATIONS WHICH WOULD BE PROMULGATED.  THE HUGE  SHIFT TO
DRYCLEANING IN THE EARLY 70's WAS DUE TO WASTEWATER REGULATORY
ACTIVITY AND THIS WAS BROUGHT TO A SCREECHING HALT WITH THE
INAUGURATION OF AIR EMISSION CONTROL EFFORTS AND THE ALLEGED
POTENTIAL OF SOME MATERIALS BEING CARCINOGENIC.

     TODAY WE ARE FACED WITH AN EVALUATION OF A SET OF GUIDELINES
PROPOSED TO CONTROL THE VOLATILE ORGANIC EMISSIONS FROM PETROLEUM
DRYCLEANERS,   MUCH WORK HAS GONE INTO THIS DOCUMENT AND THOSE
WHO HAVE DEVELOPED IT ARE TO BE COMMENDED,  FIRST FOR THEIR
INITIATIVE FOP. ALLOWING INDUSTRY TO PROVIDE THEM WITH INFORMATION
AND THEIR DEDICATION TO ASSIMILATE  AND EVALUATE THIS MATERIAL
AND SECOND, FOR THEIR PATIENCE AND WILLINGNESS TO ASSESS AND
RE-ASSESS THEIR WORK AS IT PROGRESSED,  AS A RESULT, GOVERNMENT
AND INDUSTRY WORKING TOGETHER AS A TEAM HAS ALMOST RESOLVED, mi .APPEARED
TO BE AT THE OUTSET, AN IMPOSSIBLE TASK,

                              -2-
                             IV-25

-------
SLUIZER SPEECH - PAGE 3

     THE INAUGURATION AND COMPLETION OF THE CARBON, ADSORBER
PROJECT PROVIDED MUCH VALUABLE BACKGROUND EVEN THOUGH IT MAY
NOT PROVE TO BE THE FINAL SOLUTION TO THE PROBLEM,  THE ADVENT
OF THE CLOSED RECOVERY DRYER APPEARS TO BE A MORE SUITABLE
SOLUTION,     THIS WILL BE DISCUSSED LATER,
     IN THE INTEREST OF TIME CONSERVATION, YOURS AND MINE, A
NUMBER OF US  HAVE  COORDINATED OUR-REMARKS SO THAT MINOR POINTS
WILL BE PRESENTED ONLY ONCE AND MAJOR POINTS WILL BE REINFORCED
BY REPETITION, BUT ONLY TOKEN-WISE EXCEPT FOR THE PERSON GIVING
THIS POINT FOR MAJOR CONSIDERATION,  MY ASSOCIATES IN THIS EFFORT
ARE MR, TIM VANDERVER, ESQ, AND MR, WILLIAM FISHER,'IN ABSENTIA,
FURTHER TO CONSERVE TIME, A GREAT DEAL OF SMALL DETAILS IN
BACKGROUND INFORMATION,INCLUDING SOME MISQUOTES ATTRIBUTED TO ME,
WILL BE HANDLED ON A PAGE AND LINE BASIS DIRECTLY WITH THOSE
RESPONSIBLE FOR THE WRITING OF THE DOCUMENT,

     WE ARE CONFUSED BY THE DESCRIPTION OF A SMALL INDUSTRIAL
MODEL PLANT,   SIZE IS THE ONLY DIFFERENCE BETWEEN A LARGE AND
SMALL INDUSTRIAL PLANT, NOT THE THROUGHPUT,  IN ADDITION, THE
EFFORT TO DESIGNATE A MODEL PLANT IS       MOST DIFFICULT, SINCE
MOST PLANTS DIFFER CONSIDERABLY DUE TO THEIR INDIVIDUAL  OPERATION
AND ITS NEEDS,  THERE HAS BEEN MUCH DIFFICULTY IN SETTING UP
THE TWO UNITS DESIGNATED,  WE FEEL THE SMALL MODEL DESCRIBED
IS LARGER. THAN IT SHOULD BE IN SOME OF THE FIGURES PRESENTED,
AND THE LARGER MODEL SMALLER THAN IT SHOULD BE.

     THE NUMBER OF EXISTING PLANTS CITED AS BEING INDUSTRIAL,
MAY BE GREATER, THE FIGURES PRESENTED REPRESENT THE III ESTIMATE
OF ITS MEMBERSHIP AND IT IS FELT THIS DOES NOT REPRESENT THE TOTAL,
                              -3-
                              IV-26

-------
     FURTHER, THE ALLOWABLE VOC EMISSION OF 2.4 KILOGRAMS PER'
100 KILOGRAMS DRY WEIGHT OF ARTICLES DRYCLEANED IS AN AVERAGE
AMD REPRESENTS A FIGURE FROM ONLY TWO TESTS,  AN AVERAGE OF ,96
AND 3,85,  THIS SHOULD BE CORRECTED TO AVOID AN ENFORCEMENT
ERROR,  THE FIGURE SHOULD BE ADJUSTED TO COVER EACH ITEM BEING
PROCESSED/ BE SPECIFIC~FOR THE ITEM AND BE EQUIVALENT TO THE AVERAGE
PERFORMANCE OF MY OPERATIONS WITH THE PROPOSED RACT EQUIPMENT
FOR THE SPECIFIC ITEM INVOLVED,
    ^ALTHOUGH THE CARBON ADSORPTION EQUIPMENT PERFORMANCE WAS
A WEECOMTADDITION TO PETO! FtlM RECOVERY TECHNOLOGY'.'  ITS ECONOMICS '
MAKE IT A QUESTIONABLE SOLUTION TO THE PROBLEM.   THE  COMMENDABLE
WORK DONE BY THE MANUFACTURER IN DEVELOPING~THE RECOVERY DRYER
HAS MADE IT AN APPARENT SOLUTION TO THE EMISSIONS RECOVERY PROBLEM,
HOWEVER, THE EVIDENCE PRESENTED IN THE CTG DOCUMENT FAILS TO CONVINCE
THAT THIS IS AN ACCEPTABLE SOLUTION EITHER FROM THE POINT OF RECOVERY
POTENTIAL OR SAFETY,  "[HLSINGLE EXAMPLE PRESENTED FOR ACCEPTABLE
RECOVERY IS UNDER CONDITIONS HIGHLY FAVORABLE TO GOOD RECOVERY.
SUFFICIENT TESTS TO SUBSTANTIATE THE ACCEPTABLE PERFORMANCE OF THIS
EQUIPMENT UNDER ALL CONDITIONS OF LOADING,  SOLVENT CONCENTRATION AND
VARYING ITEMS, SHOULD BE PRESENTED BEFORE THIS ONE PROCEDURE IS
ACCEPTED AS A SUITABLE EXAMPLE OF RACT FOR THIS INDUSTRY,  THIS
ALSO PERTAINS TO THE ECONOMICS OF THIS EQUIPMENT,  THE POINT BEING
MADE DOES MOT SUGGEST THE EQUIPMENT IS OR IS MOT SUITALBE, BUT
STRONGLY INDICATES SUFFICIENT EVIDENCE HAS NOT BEEN PRESENTED IN THE
DOCUMENT TO JUSTIFY THE CONCLUSION MADE.
                               IV-27

-------
SLUIZER SPEECH - PAGE 5

     IN ADDITION, THERE WAS ORIGINALLY CONSIDERABLE CONCERN
ABOUT THE SAFETY OF THIS EQUIPMENT''WITH REGARD TO FIRE AND EXPLOSION.
IN NON-RECOVERY DRYERS CURRENTLY BEING USED, EVERY EFFORT HAS
BEEN MADE TO STAY CONSIDERABLY BELOW THE LOWER EXPLOSION LEVEL (LED
BY INTRODUCING EXCESS AIR INTO THE SYSTEM.   NOTWITHSTANDING THIS,
THERE IS A WELL-RECOGNIZED RISK OF FIRES AND EXPLOSIONS IN THIS
EQUIPMENT,  IT IS NOT UNREASONABLE TO PREDICT, THEREFORE, THAT THE
PROPOSED RACT EQUIPMENT. WHICH ADMITTEDLY CLOSELY APPROACHESJHE^
LELJJNDER OPTIMUM OPERATING CONDITIONS .MAY EXCEED THE LEL MY MORE
TIMES THAN EXISTING MnN-RFOWFRY FQHTPMFNT.
     	—	•	•   "
     THE DOCUMENT EXPLAINS THIS SITUATION AND EFFORTS MADE TO
MINIMIZE PERSONNEL AND PROPERTY DAMAGE ON PAGE 3-5, AND DESCRIBES
MUCH OF THE GOOD WORK THAT HAS BEEN DONE TO MINIMIZE THIS DANGER,
AT A TIME WHEN GOVERNMENT AGENCIES ARE SEEKING "ZERO" EFFECTS FROM
CARCINOGENS, TOXIC POLLUTANTS AND OTHER RISKS TO SAFETY AND HEALTH
IN THE WORK PLACE,  HAS THIS PROPOSED EQUIPMENT BEEN ADEQUATELY
TESTED TO SATISFY THIS PERFORMANCE LEVEL?  IF SO, THE DOCUMENT
SHOULD SO STATE,  WORKER ABILITY IN THE DRYCLEAMING INDUSTRY MAY
NOT BE ADEQUATELY QUALIFIED TO "MAINTAIN SOLVENT CONCENTRATIONS
IN THE CIRCULATING VAPOR STREAMS AT LESS THAN THE RECOGNIZED LEL"
OR BE PROPERLY RESPONSIBLE FOR A "THOROUGH INSPECTION OF ARTICLES
FOR RANDOM IGNITION SOURCES BEFORE DRYING",   THE DOCUMENT SHOULD
COVER THIS MATTER MORE THOROUGHLY AND ASSURE THE LEL RISK AND
POTENTIAL HAZARDS ARE NO MORE THAN EXISTING NON-RECOVERY EQUIPMENT,
                                -5-
                               IV-28

-------
SLUIZER SPEECH - PAGE 6
     WE ALSOJEQUEST DATA BE OBTAINED TO VALIDATE THE PERFORMANCE
OF THIS EQUIPMENT WITH 14QQ FLASHPOINT PETROLEUM SOLVENT.  INASMUCH
AS SOME PLANTS UTILIZE THIS MATERIAL IN THEIR PROCESS,   THIS
INCLUDE CONTROL TEMPERATURES,  COSTS,  ETC,   WHICH WILL PROBABLY
VARY FROM THOSE PRESENTED IN THE DOCUMENT,

     THE PROPOSED EXEMPTION IS OUR GREATEST CONCERN WITH THE MODEL
REGULATION ASIDE FROM THE MANDATING OF THE  RECOVERY DRYER.
     WE AGREE 123,000 LITERS PER YEAR IS A  PROPER LEVEL IN  ALMOST
ALL CASES,  HOWEVER, SOME PROVISION SHOULD  BE MADE TO EXEMPT PLANTS
USING  MORE THAN 123,000 LITERS PER YEAR WHERE THERE  ARE EXTENUATING
CIRCUMSTANCES.  THESE CIRCUMSTANCES COULD  INCLUDE SITUATIONS
WHERE IT IS NOT POSSIBLE TO INSTALL RECOVERY DRYERS DUE TO  LACK  OF
SPACE, FIRE REGULATIONS, EXCESSIVE ECONOMIC BURDEN, ETC,  THIS HAS
BEEN DONE IN OTHER INSTANCES AND IS REQUESTED TO ALSO BE INCLUDED
IN THESE GUIDELINES,
     ALSO WE RECOMMEND FOR CONSIDERATION FOR INCLUSION  IN THE
GUIDELINES, AN ALLOWANCE FOR THOSE PLANTS COMPLYING WITH THE EMISSIONS
REDUCTION PROGRAM, WHETHER BY REGULATORY REQUIREMENT  OR VOLUNTARILY,
THE OPPORTUNITY TO RE-INTRODUCE UP TO 8% TOTAL AROMATIC COMPOUNDS
INJHE CLEANING SOLVENT.  THIS ENHANCES THE CLEANING  PROPERTIES  OF
THE PETROLEUM SOLVENTS,   THE FORMER LOS ANGELES  RULE  56 ALLOWED
THIS, BUT AS A RESULT OF DETERMINATIONS THAT EMISSIONS  OF AROMATICS
WERE HARMFUL THE AMOUNT WAS REDUCED IN MANY PLACES TO  2%.   INASMUCH
AS COMPLIANCE WILL REDUCE EMISSIONS OVER 80%,  THE TOTAL AROMATICS
EMITTED WILL BE EQUAL TO AT MOST OR LESS THAN CURRENT EMISSIONS  AND
THE INDUSTRY WILL BE BENEFITTED BY BETTER QUALITY, ECONOMICS,  SHORTER
CYCLES AND SOME FURTHER REDUCTIONS IN TOTAL EMISSIONS DUE TO LESS
OVERALL EXPOSURE IN AREAS OTHER THAN DRYING,
                              IV-29

-------
SLUIZER SPEECH - PAGE 7

     IN THE PROPOSED MODEL REGULATION, SECTION XX,040 (TESTING
AND MONITORING) ON PAGE 6-4  SHOULD BE WRITTEN TO CLEARLY STATE
THAT THIS IS FOR EVALUATING ALTERNATE METHODS ONLY. ALSO, IT SHOULD
NOT BE THE FUNCTION OF THE OWNER OR OPERATOR SOLELY,  THE OWNER
OR OPERATOR MAY BE INVOLVED, BUT THERE SHOULD ALSO BE PROVISIONS
FOR MANUFACTURER CERTIFICATION INSTEAD OF EACH PLANT HAVING TO
TEST A PIECE OF EQUIPMENT INDIVIDUALLY,  A BETTER PROCEDURE SHOULD
BE PROVIDED,  THE PROGRAM DESCRIBED IS IMPRACTICAL FOR AN OWNER
OR OPERATOR, BUT REQUIRES TECHNICAL HELP,  A SINGLE SERIES OF TESTS
OF A PIECE OF EQUIPMENT SHOULD BE SUFFICIENT FOR ALL SIMILAR PIECES,
JUST AS WITH THE CURRENTLY RECOMMENDED DRYER,  IT IS RECOMMENDED
THIS PORTION (A) OF'THE SECTION BE REVISED TO AVOID LOCAL ENFORCEMENT
FROM BECOMING CONFUSED.
     THE COMMENTS PRESENTED REPRESENT THOSE MOST IMPORTANT IN THE
CTG PRELIMINARY DRAFT,  AS STATED PREVIOUSLY, THE MANY DETAILS,
COMMENTS AND DATA REQUIRING CLARIFICATION WILL BE HANDLED SEPARATELY,
AS THEY ARE NOT SIGNIFICANT TO THE OVERALL PROGRAM PRESENTED,
     AGAIN, THE INSTITUTE OF INDUSTRIAL LAUNDEREP.S WISHES TO THANK
THE ADVISORY COMMITTEE FOR THIS OPPORTUNITY TO COMMENT,  IT IS HOPED
THESE COMMENTS WILL BE HELPFUL IN THE EFFORT TO IMPROVE, IN A PRACTICAL
AND REASONABLE MANNER, THE ENVIRONMENT SO FAR AS INDUSTRIAL PETROLEUM
DRYCLEANERS ARE INVOLVED,

     THE III IS WILLING, AS ALWAYS, TO COOPERATE WITH THE EPA.  IN
THEIR WORK AND TO FIND THE ANSWERS TO THE QUESTIONS RAISED BY THESE
COMMENTS,  ARE THERE ANY QUESTIONS?
                                -7-

                               IV-30

-------
3.  International Fabricare Institute
                            Mr. William Fisher
                     International Fabricare  Institute
                             122^1 Tech Road
                       Silver Spring, Maryland  20904
  COMMENTS TO THE NATIONAL  AIR POLLUTION CONTROL TECHNIQUES  ADVISORY
   COMMITTEE ON THE PRELIMINARY DRAFT CTG FOR "CONTROL  OF  VOLATILE
           ORGANIC EMISSIONS  FROM PETROLEUM DRYCLEANERS"
        My name  is William Fisher of the International  Fabricare

  Institute^-the national  and international trade association for

  retail drycleaners.   Because of the conflict with  the bi-annual

  CLEAN  '81 industry-wide  Convention and Exhibit, the  following

  comments are being presented on our behalf.


        On a national  basis,  close to 40% of the retail drycleaning

  establishments in the United States are direct members  of IFI,

  while many others are reached through affiliated state  and local

  organizations.  Of the approximately 25,000 commercial  drycleaning

  businesses in  the Nation, an estimated 6,000 use petroleum solvent,

  It is on behalf of these plants that we are making our comments.


        I.  Appropriateness of the 123,000 Liter Exemption Level

            IFI  believes that the designated exemption level for

  annual solvent consumption of 123,000 liters is appropriate and

  reasonable for a CTG recommendation for control of emissions from

  existing petroleum drycleaning facilities,

                                                           (more)

 	RESEARCH AND EDUCATION CENTER FOR THE PROFESSIONAL FABRICARE INDUSTRY	

                        IFI WESTERN LABORATORY
                1515 E. Chevy Chw Dr., GUndale, Calif. 91206, (213) 244-1331)

                               IV-31

-------
                            -2-

          Specifically,  we believe that the chosen exemption level

is one which strikes a suitable balance between emission reductions

and economic cbnsiderations across the spectrum of petroleum plant

sizes.  We are certain that EPA gave careful consideration to the

setting of this exemption level, and we would like to commend their

effort and the appropriateness of their decision.
      II.   The Model Regulation As It Applies to Non-Exempt
           Plants


           With the exception of certain aspects of the requirement

for dryer VOC emission control—which I will comment on separ-

ately--IFI believes that the model regulation is relatively

appropriate and well written.  Our specific comments on these areas

are as follows:

           A.  The requirement for the use of cartridge filtration

    (in all facilities using filtration) is suitable, as is the

    equivalency for alternative filtration system emissions of

    1.0 kg VOC per 100 kg of articles cleaned.


           B.  On-site storage of vacuum still  wastes in closed

    containers may be of concern in some smaller commercial plants,

    but on the whole we do not believe that this requirement will

    be burdensome.  However, we believe that it would be appropriate

    for EPA to specify a maximum solvent concentration on  a

    weight/weight basis in the still waste, below which it would

    not be necessary to containerize wastes.  We suggest that a 60%


                                                         (more)


                             IV-32

-------
                            -3-
    concentration level~-as is specified in the perchloroethylene
    CTG and NSPS^-would be appropriate.

           C.  The requirements for control of fugitive emissions
    is appropriate for the drycleaning industry and is identical to
    the specifics set in the perchloroethylene CTG and NSPS.
   Ill,   Questions on the Control of Dryer Emissions

         Petroleum drycleaners in all segments of our industry are
concerned over the rising costs of Stoddard and other petroleum
solvents, creating great interest in potential methods of
recovering petroleum solvent during drying.

           The use of carbon adsorption as a control device for
petroleum dryers is not attractive and/or feasible for a number
of reasons; these include the high initial costs and operating
costs, the large space requirements, and the technical knowledge
to properly operate and maintain  the system.

           Just as carbon adsorption for petroleum has not been
considered as generally feasible in our industry--which the draft
CTG recognizes~-development of a petroleum recovery dryer has been
recognized to have its own pitfalls.  For example, in an Institute
Technical Bulletin (T^422) published 15 years  ago, we said:

                 "Since we have recovery tumblers for perchloro-
           ethylene, the question is naturally asked,'Why not have
           one for Stoddard?'  The major difficulty facing the
           designer of a simple condenser type is the explosion

                                                           (more)
                               IV-33

-------
                             -4-

           hazard.  The system would have to be as tight as a
           vacuum still, and loaded with 'fail safe' controls."


    Still, the concept of a petroleum recovery dryer has continued

to be of interest because of its close correspondence to existing
petroleum dryers in terms of space requirements and loading
capacities, as well as the fact that operation would be identical
in all the essential aspects to existing perchloroethylene recovery
tumblers.  As .you know from this draft CTG, the Hoyt Manufacturing

Company has in the recent past developed a petroleum recovery
tumbler,

    As an association, we are glad to see the availability of a
system which is of potentially great benefit to many petroleum
drydeaners throughout the U.S.  However, we at the same time
feel that there are questions which are not fully resolved about
the operation of a recovery system.  Specifically, we have not
seen substantive data which deals with the question of crossing
over the lower explosive limit (LEL) or with the emission
reduction efficiency under "normal" conditions.  Our detailed
comments on these two points are as follows:

           A. The Question of Exceeding the Lower Explosive Limit

              I'd like to start with three quotes from the draft

    CTG.  The first is from Chapter 3, Emmission Control Techniques
    on Page 3-5:
              "Safety from fire and explosion is a key factor in
    recovery design based on the obvious dangers of vaporizing a
    combustible liquid that is heated in an enclosed space.

                                                       (more)
                            IV-34

-------
       Preventive measures focus on maintaining solvent concentrations

       in the circulating vapor stream at less than the recognized

       lower explosive limit for petroleum solvent, which is

       typically 1% by volume of vapor (Ashland, 1980)."




       The second is from Page 3-6:

             "Based on these assumptions, the recovery dryer

       performance parameters of primary importance are VOC emission

       reduction, recovery, and the ability of the dryer to operate

       with solvent vapor concentrations less than the LEL of 1%

       solvent by volume."




       The third quote is from Appendix B, Page B-l:


             "Maintaining solvent concentrations not  greater than

       95% of solvent LEL would allow for small variations in the

       LEL based on variations in the solvent's chemical composition,

       thereby eliminating the hazard of explosion that might exist

       if solvent concentrations exceeded the LEL in  the presence

       of an ignition source,"
                                               \


     Obviously, this point is clearly covered in the  draft CTG.   As

a matter of information, it might be helpful to mention that standard

non-recovery petroleum dryers are designed so that peak vapor

concentrations never exceed approximately 30-40% of the LEL.


     To date,  information is only available from EPA  on two test sites.

Summaries of the results from the Pico Rivera, California?and the


                                                            (more)
                             IV-35

-------
Lakeland, Florida, facilities are contained in the CTG.  What comments



are made about these plants?  In Appendix A, we have the following



on Page A"4:




      "One problem that was not resolved during this test was whether



the recovery dryer operated above the lower explosive limit (LEL) of



solvent  (.1% by volume or 10,000 parts per million)."  (As is later



described, the flame-ionization analyzer became saturated at a range



of 90-93% of the LEL and the actual concentrations of vapor were



higher than that range).




      What of the Lakeland, Florida, test?  On Page A-5, the CTG



states that:



      "The solvent concentration at the condenser gas inlet never



exceeded 95% of the solvent's lower explosive limit (LEL) during



the portion of the test in which the condenser water inlet temperature



was varied."





      Of course, the use here of the phrase "never exceeded 95%"



means that 95% was reached.  And what were the test conditions under



which these concentrations were found?  Further information on



Page A-5 shows that load weights averaged 55 pounds of synthetic



fabrics.





      I think I can summarize conditions of the Lakeland test as



follows:  In a test where the tumbler was loaded to only 52% of



rated capacity with fabric/fiber types which are known to have the



lowest solvent retentions after extraction, the vapor concentrations



in the tumbler reached 95% of the LEL.




                                                      (more)
                             IV-36

-------
                                -7-





      So, the data from one plant cannot answer whether the LEL was



exceeded and the data from the other plant raises questions as to



where the concentration would have peaked if the dryer had been



loaded to capacity,  Where does this leave us?  IFI believes that



three questions must be answered and resolved before a petroleum



recovery tumbler can be designated as RACT:




          1,  Under normal operating conditions, is there a potential



      for the LEL to be exceeded?
          .2,  If the LEL is exceeded and an ignition source is present-



      and an explosion occurs--are there adequate safety features  to



      handle an explosion?




          3.  If No. 2 immediately preceding is  true,  is a petroleum



      recovery tumbler acceptable to insurance companies and state



      or local fire marshals?
      Again, we applaud the work of Hoyt  Manufacturing in developing



this system, which we believe can be of significant  benefit  to  our



industry — and to the environment.   However,  as  I  stated in a letter



to TRW in March, 1980 while this draft  standard was  being developed:




          "Frankly, such acceptability  (to  insurance companies, fire



      marshals and national underwriters)  is  paramount to the use of



      recovery devices in our industry;  for  this  reason,  information



      of this nature must be developed  to  support any draft  CTG or



      NSPS."
                                                       (more)
                             IV-37

-------
     B,   Recovery Tumbler Efficiency and Equivalency Factor

         We believe that the format chosen by EPA for the model
requlation in stating that the owner of an affected facility
shall install and operate a recovery dryer or limit emissions
to a specified level is proper,   However, we do have some dis-
agreement with the specific equivalency emission factor of 2.4 kg
of solvent/100 kg of articles chosen by EPA.
    In reviewing the data as presented in the CTG and in the
individual test reports for the Pico Rivera and the Lakeland
plants,  we do not believe that there is any justification for
the 2,4  kg emission factor at this time.   Specifically, we note
the following:
        1.  The test report for Pico Rivera states that an average
emission factor of 1,0 kg/100 kg was found--based on FIA deter-
minations of emissions.  We have difficulty with accepting the
accuracy of FIA measurements of a fluctuating exhaust vapor
concentration for purposes of determining an overall emission
factor,   While it may not be as  elegant,  a well-performed mass
balance  determination is simple and highly accurate.  (That is,
a comparison of the recovered solvent vs. the wet minus dry
weights  of individual loads put into the recovery tumbler).
        ,In the contractor's test report for Pico Rivera, Table
C-8 reports "Solvent Reclaim Data" for all runs made.  Using the
data in that table — and compensating for water content of the
load--the recovery efficiency by mass balance is 90% on a weight
                                                    (more)
                         IV-38

-------
                             . 9. .





      basis, or a 10% loss.  Since this report also states that an



      average solvent retention of 30 kg/100 kg was found in the



      tested loads, we can easily calculate that the average loss



      was closer to 3 kg/100 kg,  not 1 kg/100 kg.   I would be happy



      to review my calculations with EPA at a later date.




          2.  In the Lakeland test, dryer emissions (as measured at



      the dryer exhaust by the FIA) varied from 2.34 kg to 9.45 kg per



      100 kg of articles cleaned.  Thus, emission factors (by FIA)



      varied by a factor of 4.  An average of 3.85 was calculated



      from all test runs made.  Using the data in the test report,



      we calculate an average emission factor of 3.58. -- relatively



      close to EPA's average.




          3,  In both test sites, load conditions  were not represen-



      tative of the industry as a whole.  In Pica Rivera, only leather



      gloves were dried during the testing.  Leathers have a solvent



      retention after extraction which is exactly double that of the



      average for normal textiles-^wools, cottons  and so forth.   In



      the Lakeland test, the dryer was underloaded to only 52% of



      its rated capacity.   Thus,  the solvent available for recovery--



      which has a significant bearing on the calculation of recovery



      efficiency—varied from one-half normal to double normal.








      In summary, we believe that EPA's method of taking an arithmetical



average from two tests-~that is,  averaging emission factors of 1.0 and



3.85--to derive an equivalency factor of 2.4 kg/100 kg is not supported



by the available data.   If anything, we believe that the true average




                                                         (more)




                              IV-39

-------
                              -10-





would be in the range of 3,0 — 4.0 kg/100 kg,  Additional— and truly



representative—test data is needed before a determination of the



correct number can be made.
      IV.  Alternative Test Method for Equivalency





           The Section XX.040 of the draft recommendation specifies



the procedure for an owner or operator to prove compliance with the



equivalency factor which is an alternative to use of a recovery



dryer,



          This equivalency method requires the use of a flame



ionization analyzer.  In many respects, it is similar to the compliance



test method proposed by EPA at one time in the draft perchloroethylene



NSPS standard.  Our objections to this test method remain   the same:



extreme  cost.  Previous estimates of $5,000 and up for this test



would apply to the draft petroleum CTG as well.




          As I have already commented, a properly run mass-balance



determination is a) simple  b) accurate  c) very inexpensive.  The



only equipment required is an accurate weighing device and a graduated



cylinder.




          We strongly believe that a mass-balance test method be



incorporated in the model regulation as a replacement to or an



alternative for the FIA determination for DOC emissions from the



dryer.
                                                      (more)
                              IV-40

-------
     V,  Analysis of Economic Impact




         In reviewing Chapter 5, "Control Cost Analyses of RACT,"



we believe that some of the basic assumptions, calculations and



analyses are questionable as they apply to commercial plants.   And,



while many of the smaller commercial plants might be below the



suggested exemption level, we feel that accuracy in this section  is



of importance because of state reviews which will take place after



release of the final CTG.




          As a full analysis of the questionable areas is not



appropriate at this time, I would instead like to touch on a few



high points and offer our cooperation to EPA and TRW in sitting down



to review and correct this section.  My brief comments, which  are



primarily directed towards the summarized costs in Table 5-4,  are



as follows:




              A.   While it may be desirable from EPA's viewpoint  to



     maintain consistency among CTG documents by using a standardized



     interest rate when calculating capital charges, the specifics



     of our industry's situation with respect to the availability of



     financing is quite different.  In the CTG, EPA calculates capital



     recovery charges on the basis of a 101 loan with an assumed



     equipment life of 30 years.  In contacting five banks in  IFI's



     local area,  we find that a 20 to 21% interest rate would  be



     typical.   Additionally, an assumption of a 20-year life for  a



     recovery dryer would be more accurate than the 30 years assumed



     in the CTG.



                                                        (more)
                              IV-41

-------
                         -12-


       Taking as an example the '-'small commercial" model plant


calculations, revision of the cost analyses based on a 20% loan

        7*
and a 204 life for capital recovery-^and allowing for a correct


credit for recovered solvent—the total annual operating costs


would increase from the $5300 annually projected by EPA to


between $7600 and $7950, depending on whether a cartridge filter


was already present or not.    Compare  this annual cost for


20 years against the approximately $2200 net profit (before


Federal taxes) for a plant of this size.  (Based on the annual


cost surveys run by IFI).



     B.  In Table 5-4, the "bottom line" in the major part of


table--the "difference from existing equipment annual costs"


shows a credit for all plant sizes from small commercial to


large industrial.  This line is based on subtracting the total


annual operating costs for RACT equipment from the total annual


operating costs for existing equipment.  This is predicated upon


existing petroleum plants purchasing new standard dryers and new


standard diatomite filters for their operations --a proposition


which is hardly true in the case of petroleum drycleaners.


EPA is certainly aware of this, and this is, in fact, referenced


in CTG,  This particular calculation might be appropriate in


an NSPS document but for the CTG, where existing petroleum


plants are absolutely not purchasing standard dryers or


diatomite systems, the presence of this "calculation" is totally


inappropriate and misleading.



                                                    (more)
                         IV-42

-------
                             -13-



           Specifically, this calculation could be extremely mis-



     leading to state officials who could likely presume from the



     presence of the calculation that replacement of existing dryers



     is actually occurring, when it is, in fact, not.









     Again, we offer our cooperation to EPA and TRW in sitting down



to review and correct the calculations on economic feasibility.
     In closing, let me again state that IFI believes that EPA has



done a commendable job in developing the preliminary draft CTG.



With the resolution and correction of the specific areas  which we



have commented on, we believe that the Guideline for Petroleum



Drycleaners will be realistic and in a form which will be readily



usable by the States in revising their SIP(s).   Thank you.
                                # $ #
                              IV-43

-------
4.  Van Dyne Grotty. Inc.
                               Mr, Duane E. Early
                              Van Dyne Crotty, Inc
                               903 Brandt Street
                               Dayton, Ohio 45401
            Mr. Duane Earley, Corporate Technical Manager for Van Dyne Crotty,
       spoke in support of the recovery dryer.  His firm is a very large industrial
       operation, and would be regulated under the CTG model regulation.  He said
       that the industry has blown the safety issue out of proportion.  The
       petrochemical industry, he noted, has been using condensation recovery
       of flammable liquids for decades.  He recommends that an engineering study
       of the  recovery dryer be undertaken to verify the safety of the unit.

            Mr. Earley's company has recently purchased 16 recovery dryers.  In
       his evaluation of the dryer's performance and safety prior to purchasing
       them, he consulted with owners of the unit who, while having no problems
       with local approval or insurability, had recovery efficiencies of as much
       as 90 percent.  Mr- Earley concluded his remarks by reiterating his
       recommendation of an engineering analysis of the recovery dryer's safety,
       and verified the CTG estimates of the dryer's performance.
                                     IV-44

-------
                          C.  DISCUSSION
     Following the EPA preservation, Mr.  Don Goodwin opened the floor to
questions and comments from the NAPCTAC members.  Four presentations by
industry representatives were made, and each was followed by a period of
question and discussion.  These discussions are summarized below in
chronological sequence.

     Mr. Lemke inquired about the rationale for the 123,000 liter per year
size cutoff for regulation, particularly since all plant sizes would
benefit from operating cost savings due to recovery.  Mr. Plaisance
explained that, based on the latest survey data on a typical non-attain-
ment area, approximately 70 percent of the uncontrolled emissions could
be controlled by regulating only 30 percent of the existing plants.
And while all plants could realize operating cost savings from solvent
recovery, UitJ lawyer plants would experience Lhe yr'BdlesL savings dtie to
their higher LhrOugtrputsT                                            '

     Mr. Lemke asked Mr. Beard if all states would be required to adopt
the CTG, and Mr. Beard replied that only those with non-attainment areas
would be required to implement RACT.

     Mr. Beard asked about the test procedure used to verify compliance
of the recovery dryer.  Mr. Plaisance explained that, when the dryer is
first installed, the operator would frequently monitor the condenser
vapor outlet temperature and the final recovered solvent flow rate in
order to familiarize himself with the operation and performance of the
recovery dryer.  After this brief training period, however, the monitoring
of these compliance parameters could be limited to one dryer load per
day.  Mr. Shedd added that the model regulation is an equipment and work
practices standard, with no specific emissions limits for compliance of
the recovery dryer.  Mr. Durham explained that the only equipment
necessary for recovery dryer compliance testing is a graduated cylinder.

     Mr. Beard said that local standards without detailed test procedures
might not be approved by EPA, but Mr. Goodwin stated that the purpose of
the meeting was to discuss the technology and not the implementation of
the regulation.

     Mr. Reiter asked who  receives  the draft CTG for review, and voiced
concern about states adopting the draft regulation.  Mr. Porter explained
that states with non-attainment areas, EPA regional offices, industry
trade associations, and individual  plant operators all receive the
document for review.  Mr.  Goodwin added that the CTG was sent to 42
                               IV-45

-------
industrial sources and 45 state and local agencies.  Mr. Farmer noted that
the draft CTG will not be finalized until the available comments are con-
sidered.  Mr. Reiter said that the statement in the CTG advising states
against adopting the regulation in its present form is too weak.  Mr. Porter
explained that, while EPA does not advocate adoption of the draft model
regulation, the existance of the model regulation indicates EPA's intent
to issue final guidance on these sources, and states should begin to
compile data on them.

     A discussion between Mr. Reiter and Mr. Porter ensued in which
Mr. Reiter questioned the propriety of EPA specifying a regulatory size
cutoff.  Mr. Porter noted that such a cutoff is part of the RACT assessment,
and the states have the right to set their own cutoff limits.  Mr. Beard
voiced his opposition to Mr. Reiter1s position, saying that a "floating"
or undefined cutoff would result in regulatory inequities from state to
state.  Mr. Smith added that, while some states have defined and adopted
their own versions of RACT-, only those states without SIP's would have to
adopt the RACT in the CTG.  Mr. Lents noted that the Clean Air Act
requires  states whose SIP's have not been approved by 1982 to adopt RACT
for 1987  compliance.

     Mr.  Reiter stated that the introduction of the CTG implies easy
acceptance for any SIP that incorporates, the CTG model regulation.
Mr. Porter replied that states wishing to define RACT in a manner that
differs substantially from that contained in the CTG would have to
demonstrate equivalence prior to approval.  Mr. Reiter suggested that
the individual states be allowed to formulate the details of the regulation,
with the  EPA  providing guidance only in  the selection of the control
technology.   Mr.  Beard warned that some  states~~or regions might write
their  regulations based on nothing more  than the draft CTG.

     Mr.  Steiner  questioned the 100 ton  per year emissions used to justify
the 123,000  liter per year cutoff.  Mr.'  Porter explained that this is a
criterion for defining a major source that will be used in the review of
1982 SIP's.   And  while states are free to set their own cutoffs, it is
EPA's  position that  the 100 ton figure is reasonable because it is based
on controlling 70 percent of the emissions by regulating 30 percent of
the sources.

     Mr.  Goodwin  asked the committee if  the states and regions want specific
cutoff numbers.   Mr.  Beard responded that guidance is needed, but without
exact  numbers.  Mr.  Lents said that exact cutoff numbers should be avoided,
because the  values are usually somewhat  arbitrary, and the states have
the burden  of substantiating them.

     Mr.  Goodwin  asked the committee if  a range of values would be prefer-
able  to exact cutoff numbers.  Mr.  Reilly said that, while the states want
EPA to define RACT,  a  range of values should be given to emphasize the
variability  of the cutoff.  Mr. Lents agreed that RACT should be clearly
defined and  that  a cutoff range should be included.  Mr. Beard reiterated
his previous  warning about  the regions adopting the model regulation in its
current form.  Ms. Dubrowski spoke  in favor of retaining exact cutoff
                                    2

                                 IV-46

-------
numbers to reflect actual  industry operating conditions.   Mr. Castelli
agreed, explaining that industry would prefer to have consistent, reliable
values.

     Ms. Haskell asked about the lack of record-keeping in the recovery
dryer compliance testing.   Mr.  Porter explained that it is EPA's position
that the individual states should determine the format and extent of
record-keeping.  Mr. Goodwin added that the CTG's purpose is to provide
guidance on control technology.

     Mr. Reiter inquired about the rationale for selecting the implemen-
tation-to-final-compliance duration of 19 months.  Mr. Plaisance explained
that this duration was sufficient to allow for selection, ordering,
delivery, installation, and testing of the control  equipment.  Mr.  Reiter
responded that this compliance period might be inadequate for smaller
plants.  Mr. Farmer noted that the extension of final complaince to 3
months after delivery of the control equipment would account for delivery
delays.  Mr. Porter added that the 19 month duration was  derived from
consultations with control equipment vendors and operators of plants who
had previously installed the equipment.

     Mr. Reilly questioned the detection of solvent leaks, as well  as
the reporting of the FIA calibration factors.  Mr.  Goodwin re-emphasized
the fact that the model regulation is an equipment and work practices
standard.  Mr. Durham added that the FIA is not involved  in the recovery
dryer compliance testing, and that leak detection is a part of general
plant housekeeping which requires the prompt repair of known leaks.

     Mr. Lents asked if the annualized operating cost calculations
include capital recovery, and Mr. Plaisance responded affirmatively.

     Following Mr. Vanderver's presentation, Mr. Reiter asked Mr. Vanderver
if he felt that the economic analysis contained in the CTG was correct.
Mr. Vanderver said that Mr. Fisher's presentation would deal in detail
with the accuracy of the CTG economic analysis, adding that his prime
concern was with the safety of the recovery dryer.

     Ms. Dubrowski questioned Mr. Vanderver's proposal of a joint
industry-EPA test of the recovery dryer, asking Mr. Vanderver if he or
the industry had their own test data indicating safety problems.
Mr. Vanderver replied that it was inappropriate for EPA to mandate
equipment whose safety was unsubstantiated.  Furthermore, the two EPA
tests of the recovery dryer were, in his opinion, unreliable.  In the
Pico Rivera test the dryer loading was atypical (overloaded) and the FIA
"failed" in monitoring the LEL.  The Lakeland test data,  he added,  were
also unreliable due to underloading of the dryer with atypical articles
(synthetics).  He reiterated his suggestion of a joint industry-EPA test
of the  recovery dryer.

     Mr. Durham explained that the proximity of the dryer concentration
to the  solvent LEL was a function of the fabric, load weight, drying time,
and other dryer operating parameters.  Although problems were encountered
in the  first two recovery dryer tests, he added, a third test has been


                                   3
                                 IV-47

-------
completed, and industry should examine the results of all three tests
before advocating further testing.  Mr. Vanderver concurred with
Mr. Durham's recommendation, adding that the industry is not opposed to
the recovery dryer, but rather wishes the safety issue to be resolved
before the equipment is mandated.  And if the economics of the recovery
dryer are as reported, he explained, the industry will adopt them
without regulatory inducement.

     Mr. Beard noted that the recovery dryer is designed to minimize the
effects of an explosion and that the existing units are insured, and asked
Mr. Vanderver to explain his criteria for verifying the safety of the dryer.
Mr. Vanderver replied that, while there have been reports of Factory Mutual
approval of the dryer, there are still major questions to be answered as
to approval by local fire marshals and overall insurability.  Mr. Goodwin
asked Mr. Vanderver if the resolution of these questions would result in
industry acceptance of the recovery dryer.  Mr. Vanderver replied that,
until the questions of safety are adequately resolved, the industry will
oppose mandatory installation of the dryers.

     Ms. Dubrowski asked Mr. Vanderver if the recovery dryer safety
problems had been documented in industry tests of the dryer.  Mr. Vanderver
responded that solvent recovery dryers have been extensively tested in
the perc dry cleaning industry, adding that he understood that one of
the first petroleum solvent recovery dryers installed "blew up."
Mr- Plaisance noted that the National Fire Protection Association code
for dryers makes no distinction between recovery or non-recovery dryers,
nor does it stipulate a maximum internal concentration, but rather
specifies a minimum explosion damper area of one-third square foot per
30 cubic feet of dryer volume.  Mr- Vanderver added that the currently-
produced recovery dryer is the first in the petroleum solvent dry cleaning
industry.  Mr. Goodwin asked Mr. Vanderver if the resolution of the
questions of approval and insurability would be a step in the right
direction, and Mr. Vanderver agreed.

     Mr.  Porter noted that EPA's investigations into the operation of
both recovery and non-recovery dryers have indicated that, if not properly
operated, both dryers can exceed the solvent LEL and can operate unsafely.

     Mr.  Plaisance explained that, due to the newness of the solvent
recovery  technology as applied to domestic petroleum solvent dryers, TRW
investigated the safety and performance of recovery and non-recovery dryers
in Japan where identical solvent recovery technology has been applied to
petroleum  solvent dry cleaning for about five years.  Currently, he added,
there  are about 6,000 non-recovery and 1,700 recovery dryers operating  in
Japan,  and  the explosion-fire rate for both is about one per 100 dryers.

     Mr.  Vanderver stated that non-recovery dryers operate at 30-40
percent of  the solvent LEL, while the recovery dryer may exceed the LEL.
He added  that he had  no knowledge of the report on the Japanese industry,
and would withhold comment until after distribution of the report.

      Following Mr. Sluizer's presentation, Ms; Dubrowski asked about the
number of  recovery dryers .in  use.  Mr. Sluizer estimated about 100 units,


                                   4
                                  IV-48

-------
adding that the rate of installation has steadily increased since the firsl
units were installed approximately \h years ago.   Ms.  Dubrowski  asked if
any of the current owners of recovery dryers had  experienced problems in
obtaining insurance, and Mr. Sluizer replied that, to  his knowledge, there
had-been no such problems.   Ms.  Dubrowski  next asked if Mr. Sluizer knew
the rate of accidents with the recovery dryer, and he  replied that he did
not know.  He added that, although the recovery dryer  is gaining acceptance
in the industry, there are still  concerns  about its safety which have to
be resolved.   Mr.  Sluizer said that he was not aware of the study on the
Japanese industry, and that his  segment of the industry frequently
employs personnel  who do not adapt well to new technology.  Ms.  Dubrowski
asked if the economic savings in the existing units are equivalent to
those portrayed in the CTG.  Mr.  Sluizer replied  that,  neglecting some
of the more extreme variations in individual plant equipment, operations,
and economics, the savings attributed to the recovery  dryer in the CTG
are correct.  The industrial sector, he added, has the  high throughput
that results in substantial savings, and at least half  of the existing
recovery dryers were installed for economic reasons, particularly with
solvent prices rising from 20-25 cents to  $1.30-$1.75  per gallon.

     Ms. Dubrowski inquired about the prospects for a  large capacity
recovery dryer, and Mr. Sluizer explained  that a  conference had  been held
with a major manufacturer of non-recovery  dryers  who declined to pursue
development of a 400 pound-capacity recovery dryer due  to the high costs
of development.  Mr. Lemke asked if there  was a plant  size below which
a savings would not result from recovery dryer installation.  Mr. Sluizer
responded that, although Mr. Fisher's presentation would focus on the
CTG's impact on smaller plants,  it was his opinion that smaller  plants
would experience reduced savings as a result of their  lower throughputs.

     Mr. Reiter asked about the aromatics  content of the solvent that
Mr. Sluizer had referred to in his presentation,  and Mr. Sluizer replied
that, while he could not specify their chemical identify, the quantity
of aromatics contained in the solvent directly related  to the rate and
efficiency of cleaning.

     Following Mr. Fisher's in absentia presentation by Mr. Vanderver,
Mr. Beard asked if industry was consulted  during  the development of the
CTG.  Mr. Goodwin replied that this current CTG has been developed like
previous CTG's with industry input, and the only change in development
of this document has been the addition of the NAPCTAC  evaluation.

     A discussion ensued in which industry's opportunities to comment on
the draft CTG and supporting test reports  were clarified.  Mr. Goodwin
and Mr. Porter explained that, while industry was consulted during the
development of the CTG, the present NAPCTAC meeting represented the first
opportunity for industry-wide comment on the draft CTG and the supporting
test reports.

     Mr. Reiter asked if there was an NSPS for petroleum dry cleaning,
and Mr. Goodwin replied that it was being developed.
                                   5

                                  IV-49

-------
                               D.   CORRESPONDENCE
1.   Illinois Environmental  Protection Agency
 March 13, 1981
 National Air Pollution Control
   Techniques Advisory Committee
 U.S. Environmental Protection Agency
 Office of Air Quality Planning and Standards
 Research Triangle Park, North Carolina   27711

 Gentlemen:

 For your information and record, the  Illinois Environmental Protection
 Agency submits the following comments:
 Attachment 1 -*-  Comments on Preliminary Draft "Control of Volatile
     ---- ""/  Organic Emissions from Petroleum Dry Cleaners";

 Attachment 2 —  Comments on Preliminary Draft "Control of Volatile
                  Organic Emissions from Volatile Organic Liquid Storage  in
                  Floating and Fixed Roof Tanks";

 Attachment 3 —  Comments on Preliminary Draft "Control of Volatile
                  Organic Fugitive Emissions from Synthetic Organic
                  Chemical, Polymer and Resin Manufacturing Equipment".

 Your consideration of these coitments  is most appreciated.

 Sincerely yours,
 John C. Reed, Ph.D., P.E.
 Supervisor, Technical Support Unit
 Air Quality Planning Section
 Division of Air Pollution Control

 JCR:jab/2852H/24
                                   IV-50

-------
                                Attachment 1

                        Comments on Preliminary Draft

     "Control  of Volatile Organic Emissions  from Petroleum Dry Cleaners"


1.  123,000 liter exemption - This level appears without  documentation.
    Since the thrust of the conclusions is to control only the  large
    industrial sources, the CTG should be totally rewritten to  reflect
    that only large industrial^are to be considered and regulated.

2.  Limit emissions to 2.4 Kg from dryer or install solvent recovery
    dryer.  This is the average of two series of tests (one commercial,
    one industrial) and has not been demonstrated to be RACT.   The
    industrial facility averaged 0.96 Kg (industrial RACT?) while the
    commercial averaged 3.85 Kg (commercial RACT?).  If CTG is  to apply
    only to large industrial sources, the use of the commercial factor  is
    inappropriate as it skews the average.

3.  Solvent recovery operating  parameters do not ensure emissions of less
    than 2.4 kg/100 Kg clothes  cleaned.  Based  on the information in Table
    A-2, the tests at average temperature of 94° F emitted 2.90 Kg which
    is 20% more than proposed as limit.  The only test that emitted <2.4
    Kg was at 95° F but two other tests at 95°  F emitted  more than 2.4
    Kg/100 Kg clothes cleaned.  The operating parameters  should be shown
    for these tests and figures similar to Figure A-l supplied  for all
    tests at 94 or 95° F.

4.  Documentation on waste storage requirements should be provided and  the
    requirement for decreased number of boildowns should  be specified (if
    this is to be part of RACT).  Reduced boildowns do not have an
    analysis of the optional efficiency.  No testing has  been documented
    on the level of emissions from improper storage.  Eliminate specific
    control for stills and include it as fugitives as far as controls go.

5.  RACT requirement "to minimize" emissions from solvent stills is not
    defined and enforcement would depend on administrative rulemaking by
    state agency.

6.  The repair/replace schedule was not documented as being technically
    and economically reasonable.  Why three days?  What parts are not
    normally on hand and what portion of fugitive emissions is  due to
    parts that cannot be replaced within the required period?

7.  The dryer testing requirements for alternative devices are
    undocumented.  A facility should only be required to  duplicate its
    range of variations.

8.  The solvent recovery dryer  testing requirements are undocumented.
                                 IV-51

-------
Page 2


9.  The solvent filtration system RACT was  determined  on  the  basis  of  one
    sample but the proposed rule requires five  samples  in  the testing
    procedure.

10. No documentation on the occurrence of leaks to  show weekly  inspection
    of containers and other equipment is reasonably available considering
    technical and economic feasibility.  "Other"  equipment has  not  been
    defined  and the proposed  language would not be  unenforceable  as such.

11. Since there is only one supplier of RACT  devices,  the  question  of
    reasonable further progress due to nationwide demand  for  these  devices
    should be addressed.

12. One time test procedure undermines continuing compliance  efforts.
    Alternative dryer devices  are to be tested  once but monitoring  the
    operating parameters  of the solvent recovery  device requires  ongoing
    costs to facility choosing that technology.   Additionally,  while
    design changes may not take place, changes  in the  nature  of cleaning
    items done for customers  could effect emissions as  shown  in tests  of
    differing types of loads.  USEPA shouljjjnclude language  clarifying
    its position that recordkeeping ano^compliance  programs will  have  to
    be developed by the individual states and  included  in  the RACT  control
    regulation.


JCR:ct/2820H,6-7
                                  IV-52

-------
             V.   CONTROL TECHNIQUES  GUIDELINES  FOR  VOC  EMISSIONS
                  FOR VOLATILE  ORGANIC  LIQUID STORAGE TANKS

                            A.   EPA PRESENTATION

                             Ms.  Rebecca Sommer
                          GCA/Technology Division
                             500 Eastowne Drive
                     Chapel  Hill, North Carolina  27514


INTRODUCTION

     This presentation discusses the development of the volatile  organic
liquid (VOL) storage tank control techniques guideline  (CTG)  document.   This
presentation consists of an  overview of the source  category,  a discussion of
the reasons why the source category was selected for development  of a CTG
document, a brief outline of the reasonably available control  technology
(RACT) for the source category, and a discussion of the major decisions made
in selecting RACT.  (overhead #1)

OVERVIEW OF SOURCE CATEGORY

     The source category for this document is volatile  organic liquid storage
tanks.   A volatile organic  liquid is defined as any organic  liquid that
produces volatile organic compounds (VOC) as vapors and is not considered a
petroleum liquid.  Petroleum liquids are defined to be  petroleum, condensate,
and any finished or intermediate products manufactured  in a  petroleum refinery.

     There are three basic types of VOL storage tanks:   (1)  fixed-roof tanks,
(2) external floating-roof tanks, and (3) internal  floating-roof  tanks.  A
typical fixed-roof tank consists of a cylindrical steel shell  with a cone-
or dome-shaped roof that is  permanently affixed to  the  tank  shell,  (overhead
#2)

     An external floating-roof tank consists of a cylindrical  steel shell
equipped with a deck or roof that floats on the surface of the stored liquid
and rises and falls with the liquid level.  The liquid  surface is covered by
the floating roof except for a small annular space  between the floating roof
and the tank.  This annular  vapor space is generally enclosed by  a primary
seal or a primary and secondary seal combination.  VOC  emissions  from an
external floating-roof tank  are generally less  than the emissions from a
comparable fixed-roof tank,   (overhead #3)
                                     V-l

-------
         OUTLINE OF THE PRESENTATION
     I. Overview of Source Category
     II. Selection of Source Category
    III. Discussion of Reasonably Available
       Control Technology
    IV. Selection of Reasonably Available
       Control Technology
                FIXED  ROOF  TANK
PRESSURE-VACUUM
      VALVE
GAUGE HATCH
      MANHOLE
              MANHOLE
                                         NOZZLE
                                         IFOR SUBMERGED FILL
                                         OR DRAINAGE)
                          V-2

-------
     The third type of tank used to store VOL's is the internal floating-roof
tank.  This type of tank has a permanently affixed roof, and either a contact
internal floating roof, which floats on the liquid surface, or a non-contact
internal floating roof, which is supported by pontoons and rests several
inches above the liquid surface.  Circulation vents and an open vent at the
top of the fixed roof are provided to minimize the possibility of hydrocarbon
vapors accumulating in concentrations above the lower explosive limit.  The
annular space between the tank wall and the.floating roof is the primary
source of emissions from this type of tank; however, fittings such as column
wells are additional sources of emissions,  (overhead #4)

     A non-contact internal floating roof reduces liquid evaporation by
confining the vapors to a small space above the liquid surface.  A metal rim
plate projecting downward from the floating roof into the liquid forms a seal
that encloses the vapor space directly below the floating roof.  A contact
internal floating roof, which floats directly on the liquid surface, eliminates
evaporation by restricting vapor formation,  (overhead #5)

     Regardless of tank design, a floating roof requires a closure device to
seal the gap between the tank wall and the roof perimeter.  Primary seals,
the lower seal of a two-seal system, can be made from a variety of materials
suitable for organic liquids.  The basic designs available are:  (1) mechanical-shoe
seals, (2) liquid-filled seals, and (3) resilient-foam-log seals.

     A shoe seal is characterized by a metallic sheet known as the "shoe,"
which is held against the tank wall.  A flexible coated fabric (the "envelope")
is suspended from the shoe to the floating roof to form a gas-tight cover
over the annular space,  (overhead #6)

     A liquid-filled seal is typically a flexible polymeric tube filled with
a liquid and sheathed with a tough fabric scuff band.  The liquid is commonly
a petroleum distillate or other liquid that would not contaminate the stored
product if the tube ruptured.  Liquid-filled seals are mounted on the surface
with no vapor space below the seal,  (overhead #7)

     A resilient-fcam-filled seal is a tough fabric band filled with a resilient
foam log.  The resiliency of the foam log permits the seal to adapt itself to
some imperfections in tank dimensions or in the tank shell.  The foam log may
be mounted above the liquid surface (vapor mounted) or on the liquid surface
(liquid mounted),  (overhead # 8)

     A liquid-mounted seal provides no space under the seal for the formation
of vapors.  In cases where gaps exist between the seal and tank wall, only
the exposed liquid surface is available for evaporation and, thus, for emissions.
In contrast, a vapor-mounted seal allows vapors to form in the volume enclosed
by the seal.  In cases where gaps exist, the entire annular space is available
for vapor formation.  Therefore, emissions will be much higher with a vapor-mounted
seal than with a liquid-mounted seal with similar gaps.
                                     V-3

-------
   EXTERNAL FLOATING ROOF TANK
FLOATING ROOF
                              .PRIMARY SEAL
                                          4
    INTERNAL FLOATING  ROOF TANK
                     CENTER
                      VENT
    ROOF VENT
     INTERNAL
     FLOATING
        ROOF
  PRIMARY SEAL

   ACCESS PORT
                           TANK SUPPORT COLUMN

                          COLUMN WELL
                     V-4

-------
            INTERNAL FLOATING ROOFS
TANK WALL
           SECONDARY SEAL
  LIQUID
                                     SECONDARY SEAL
                          PRIMARY SEAL
                          NONCONTACT
                          INTERNAL FLOATING
                PRIMARY SEAL
                IMMERSED IN LIQUID
                            TANK WALL-*
CONTACT INTERNAL
FLOATING ROOF
                                             PONTOON
           CONTACT
                   NONCONTACT
           MECHANICAL  SHOE SEAL
             TANK WALL
                SHOE-
                             ENVELOPE
                              EXTERNAL FLOATING ROOF
                              ANNULAR VAPOR SPACE
                             V-5

-------
                LIQUID FILLED SEAL
              TANK WALL-*-



              SCUFF BAND
                    FLOATING
                       ROOF
                          ^^Sk- LIQUID FILLED TUBE
                FOAM  FILLED  SEALS
                                                       8
TANK WALL-*-
VAPOR SPACE- -
!*•,"»•;* i*/iit
&ft«l
'?/•# '••
     FLOATING
        ===*
         ROOF

        •SEAL FABRIC
                         TANK WALL-*-
                 •RESILIENT FOAM
           VAPOR MOUNTED
                                           FLOATING
                                              ROOF

                                             SEAL FABRIC
                                           RESILIENT FOAM
                        LIQUID MOUNTED
                             V-6

-------
     A secondary seal can be mounted above the primary seal to minimize the
effects of air currents, which sweep vapors out of the annular space.  A
rim-mounted secondary seal is a continuous seal which extends from the floating
roof to the tank wall, covering the entire primary seal.  A rim-mounted
secondary seal is typically a wiper seal  or a resilient-foam-filled seal.
Another type of secondary seal is a shoe-mounted secondary seal.   A shoe-mounted
seal extends from the top of the shoe to  the tank wall.   These seals do not
provide protection against VOC leakage through the envelope.   Holes, gaps,
tears, or other defects in the envelope can allow the VOC vapors  under the
envelope to be emitted to the atmosphere  (overhead #9).

SELECTION OF CATEGORY

     Factors leading to the selection of  VOL storage tanks for development
of a CTG document include the significant reductions in  VOC emissions that
are achievable with the application of available emission control  technology.
The application of available control technology can reduce VOC emissions
from a tank by 90 percent or more.

     In 1977 there were about 32,000 VOL  storage tanks in use at  chemical
plants and liquid bulk storage terminals.  It is estimated that the total
1977 VOC emissions from these storage tanks were 71,500  megagrams.   (overhead
#10)  The list of areas requesting an extension beyond 1982 for compliance
with the national ambient air quality standard for ozone includes  areas where
many storage tanks are located.  Consequently, the application of  RACT to
storage tanks in these areas will result  in the reduction of VOC  emissions.

DISCUSSION OF RACT

     Reasonably available control technology is a contact internal  floating
roof with a liquid-mounted or metallic-shoe primary seal and a continuous
secondary seal,  (overhead #11)  It is recommended in the draft model regulation
that RACT be installed on all 40,000 gallon or larger tanks that  store a
volatile organic liquid with a vapor pressure greater than 1.5 psia.  Implementation
of RACT would require various types of equipment to be retrofitted to the
different types of existing storage tanks.  Fixed-roof tanks would be retrofitted
with a contact internal floating roof with a liquid-mounted or metallic-shoe
primary seal and a continuous secondary seal,  (overhead #12)  External
floating-roof tanks would require a continuous secondary seal and  a permanently
affixed roof.  Tanks that are currently equipped with an internal  floating
roof are exempt.  Pressure vessels designed to operate without emissions to
the atmosphere except under emergency conditions are also exempt.   The draft
model regulation does not suggest that existing primary  seals that are not
RACT be removed and replaced by a liquid-mounted or metallic-shoe  primary
seal, or that existing secondary seals that are not RACT be removed and
replaced with a continuous secondary seal.  However, when a primary seal is
replaced, it is recommended that it be replaced with a liquid-mounted or
metallic-shoe primary seal; and when a secondary seal is replaced, it is
recommended that it be replaced with a continuous secondary seal.
                                     V-7

-------
                 SECONDARY  SEALS
TANK WALL-*-
               RIM-MOUNTED
               SECONDARY SEAL
                (WIPER SEAL)
                            TANK WALL
                    FLOATING        \

                       ROOF          ^

                     SEAL FABRIC
                   RESILIENT FOAM
            RIM-MOUNTED
            SECONDARY SEAL
   'SHOE-MOUNTED
   SECONDARY SEAL
    (WIPER SEAL)
     -ENVELOPE
        FLOATING

           ROOF
                                               VAPOR SPACE
 SHOE-MOUNTED
 SECONDARY SEAL
                                                    10
         VOL STORAGE TANKS AND EMISSIONS*
         Number of Existing Tanks

         VOC Emissions
32,000

71,500   megagrams
         "Estimated annual emissions for 1977
                             V-8

-------
                                                   11
SUMMARY OF REASONABLY AVAILABLE
	CONTROL TECHNOLOGY	


       •Contact  internal floating roof

       • Liquid mounted or metallic shoe primary seal

       •Continuous secondary seal

       •Tanks >40,000 gallons storing VOL with
        vapor pressure > 1.5 psia
   RACT RETROFIT REQUIREMENTS
     Fixed Roof Tanks
     • Install contact internal floating roof
      with liquid mounted or metallic shoe
      primary seal and continuous secondary seal

     External Floating Roof Tanks
     •Install a fixed roof and a continuous
      secondary seal

     Internal Floating Roof Tanks
     • Exempted from requirements of RACT
                           V-9

-------
     Provisions are made in the draft model regulation for an owner or
operator of a storage tank to apply for an alternative control device other
than floating-roof control technology.  Alternative control devices must
reduce emissions by at least 90 percent.  The emission reduction efficiency
would be calculated by comparing emissions resulting from the use of the
control device with emissions from a fixed-roof storage tank fitted with a
conservation vent.

SELECTION OF RACT

     During the development of this CTG document several major decisions were
made.  Among them were the values for the capacity and vapor pressure cutoffs,
the control equipment which was to be defined as RACT, and the emission
reduction efficiency of acceptable alternative control devices.

     An emission   analysis indicated that the emission reduction obtained
through installation of RACT decreased with decreasing tank size and vapor
pressure.  Small tanks are used at research laboratories, retail outlets, and
some other small facilities.  As a result, it was determined that a lower
cutoff limit based on tank capacity and vapor pressure should be established
in this CTG document.

     The criteria used for the selection of the tank size and vapor pressure
cutoffs were:

     1.   Cost of RACT.

     2.   Consistency with previous EPA documents.
     Based on an analysis of these factors, it was decided to exempt all
tanks  smaller than 40,000 gallons and all tanks storing liquids with vapor
pressures less than 1.5 psia.  (overhead #13)  The annualized cost of retrofitting
a fixed-roof tank at these cutoff points is $2928, and the cost per megagram
of VOL saved represents "worst case" costs.  These cutoff points are consistent
with the capacity and vapor pressure cutoffs presented in the petroleum
liquid CTG documents.

     For a fixed-roof tank the contact internal floating roof was chosen as
RACT over the non-contact roof because of the significant increase in emission
reduction.  Removing an existing non-contact internal floating roof to
install a contact internal floating roof is not recommended in the draft
model  regulation  (overhead #14).

     The emissions from an external floating roof tank with primary seals can
be significantly reduced by building a fixed roof over the tank.  This action
would  reduce the affect of winds sweeping VOC vapors out of the annular vapor
space  and would convert the tank to a contact internal floating-roof tank.
The emissions from an external floating-roof tank with primary seals can be
further reduced by retrofitting the floating roof with a secondary seal as
well as with a fixed roof.  For these reasons RACT for an external floating-roof
tank was chosen to be a permanently affixed roof and a primary and secondary
seal combination for the floating roof.
                                    V-10

-------
                                                    13
      EMISSION REDUCTION AND COST
          OF RACT FOR SMALL TANK*

   Annualized Cost                         $2928
   Recovery Credit                         (5640)
   Net Annualized Cost                     52288
   Total Emission Reduction                   1.94 Mg/yr
   Cost Effectiveness                        S1179/Mg
    "Capacity of 40,000 gallons, storing VOL
    with vapor pressure of 1.5 psla
                                                         14

         COST OF IMPLEMENTING RACT

Annualized cost (S)
Recovery credit (S)
Net annuallzed cost (5)
VOL reduction (Mg/yr)
Cost effectiveness (J/Mg)
Average fixed
roof tank"
3894
(2191)
1703
6.64
256
Average floating
roof tank"
9945
(7098)
2847
21.51
132
•Diameter of 26 feet; capacity of 127,000 gallons; vapor pressure of 1.5 psla.
"Diameter of 62 feet; capacity of 920,000 gallons; vapor pressure of 2.2 psla.
                             V-ll

-------
     The annualized cost of implementing RACT for an average fixed-roof tank
is $3894 and the annualized cost for an average floating-roof tank is $9945.
An average external floating-roof tank is generally larger than an average
fixed-roof tank.  A credit of $330/Mg from preventing the VOL from evaporating
reduces the annualized cost even further.  The product recovery credit of
$330/Mg (15
-------
                       B,  INDUSTRY PRESENTATIONS

   1,   Chemical Manufacturer's Association

                           Mr. Bruce C,  Davis
                         Exxon Chemical  Company
                      Florham Park, New Jersey 07932
     MY NAME  is BRUCE C, DAVIS,   I AM A  STAFF  CHEMICAL  ENGINEER FOR
THE CORPORATE ENVIRONMENTAL CONTROL  SECTION WITHIN  THE  CENTRAL
ENGINEERING DIVISION OF THE EXXON CHEMICAL COMPANY,   I  AM  A REGISTERED
PROFESSIONAL  ENGINEER IN THE STATE OF NEW YORK,   I  AM ALSO A MEMBER
OF THE CHEMICAL MANUFACTURER'S ASSOCIATION'S PROCESS  EMISSION REGU-
LATIONS TASK  GROUP AND A MEMBER OF THE VOL STORAGE  WORK GROUP,
                                                                 «
     TODAY I  AM SPEAKING ON BEHALF OF CMA, A NON  PROFIT TRADE ASSOCIATION
HAVING 188 U,S, COMPANY MEMBERS THAT REPRESENT MORE THAN 90 PERCENT
OF THE PRODUCTION CAPACITY OF BASIC  INDUSTRIAL CHEMICALS WITHIN THIS
COUNTRY,  CMA MEMBER COMPANIES HAVE  A DIRECT AND  CRITICAL  INTEREST*-
IN ENSURING THAT EPA DEVELOP CONTROL TECHNIQUE GUIDELINES  WHERE A '
DEMONSTRATED  NEED IS PRESENTED, THAT ARE SCIENTIFICALLY AND TECHNICALLY
SOUND, REASONABLE, PROCEDURALLY WORKABLE, COST EFFECTIVE,  AND CLEARLY
AUTHORIZED BY THE CLEAN AlR ACT,
     CMA PRESENTED COMMENTS ON THE VOL STORAGE NEW  SOURCE  PERFORMANCE
STANDARD AND BACKGROUND INFORMATION  DOCUMENT AT THE DECEMBER 3, 1980
NAPCTAC MEETING,  ALTHOUGH WE AND OTHERS  EXPRESSED CONCERN, PARTICULARLY
ABOUT THE EPA DATA BASE AT THAT TIME, THE AGENCY  HAS  NOT ADDRESSED OUR
CONCERNS IN THE DEVELOPMENT OF THIS  CTG,  ACCORDINGLY WE STILL HAVE
SEVERAL SIGNIFICANT RESERVATIONS AND CONCERNS  WITH  THE  DRAFT CTG,  IN
                                 V-13

-------
 THIS REGARD THE THOUGHTS WE OFFER TODAY AND THE MORE DETAILED WRITTEN
 COMMENTS WE WILL SUBMIT BY MARCH 20, 1981 WILL ADDRESS THESE  ISSUES
 AND OFFER APPROPRIATE RECOMMENDATIONS,
 1,  REASONABLE AVAILABLE CONTROL TECHNOLOGY (RACT)  is DEFINED AS A
    CONTACT INTERNAL FLOATING ROOF WITH PRIMARY AND SECONDARY SEALS.
    CTIA DOES NOT BELIEVE THE DATA EPA.PRESENTS SUPPORT THE CONCLUSION
    THAT THIS DESIGN PROVIDES THE GREATEST REDUCTION IN EMISSIONS OF
    ALL ROOF AND SEAL COMBINATIONS FOR ALL VOLATILE ORGANIC LIQUIDS (Vfll)
    A, THE AGENCY CONTINUES TO EXTRAPOLATE TESTING DATA FROM BENZENE
    TO OTHER ORGANIC LIQUIDS,
      ACTUALLY, THE ONLY VALID CONCLUSION FROM THESE DATA is THAT O_F_
    THE ROOF/SEAL COMBINATIONS TESTED. AND ONLY UNDER  TEST CONDITIONS.
    THIS DESIGN IS THE MOST EFFECTIVE FOR REDUCING EMISSIONS FROM
    BENZENE STORAGE TANK-S.   THE BENZENE TEST RESULTS CANNOT BE APPLIED
    GENERICALLY TO ANY TANK STORING ANY VOL,
      A REVIEW OF THE EPA DATA ON BENZENE EMISSIONS FROM A FLOATING
ROOF TEST TANK SHOWS:
    1, WITHDRAWAL LOSS is A DIRECT FUNCTION OF A "CLINGAGE" FACTOR,--
       THIS SHOULD BE DIFFERENT FOR DIFFERENT CHEMICALS,
    2, BOTH THE SEAL AND FITTING LOSSES ARE DIRECT FUNCTIONS OF
       MOLECULAR WEIGHT,  OBVIOUSLY,  THE MOLECULAR WEIGHT OF OTHER
       VOL'S DIFFERS FROM THAT OF BENZENE.
      THUS, CMA DOES NOT BELIEVE THE TEST RESULTS  FOR  BENZENE SUPPORT
                            • >
ERA'S CHOICE OF CONTROL TECHNOLOGY,   THE AMERICAN  PETROLEUM INSTITUTE,
TEXAS CHEMICAL COUNCIL AND CMA HAVE ALL COMMENTED  IN DETAIL ON THIS
ISSUE IN THE PAST, SO WE WILL NOT ELABORATE FURTHER EXCEPT TO MAKE A
RECOMMENDATION,  WE UNDERSTAND THAT THE AMERICAN PETROLEUM INSTITUTE
IS CURRENTLY TESTING INTERNAL FLOATING ROOFS WITH  VARIOUS SEAL COM~

                                 V-14

-------
BINATIONS ON PURE COMPOUNDS OTHER  THAN BENZENE,   THESE  TESTS  SHOULD BE  ,
COMPLETED THIS  SPRING,  WE RECOMMEND EPA DELAY  ISSUING THE FINAL CTG
UNTIL THIS  STUDY  CAN BE EVALUATED,
    B, THE  AGENCY'S RECOMMENDED EMISSION CONTROL TECHNOLOGY  MAY  BE
       TECHNICALLY  INFEASIBLE  IN  MANY CASES,
      RACT  FOR  AN EXTERNAL FLOATING ROOF TANK  IS INSTALLATION  OF
SECONDARY SEALS AND A  FIXED  ROOF,  THIS REQUIRES THE FIXED ROOF  BE
RETROFITTED ONTO  THE EXISTING  TANK,  WHILE  INSTALLATION MAY  BE READILY
ACCOMPLISHED ON SMALL  DIAMETER TANKS, THERE ARE  SERIOUS TECHNICAL
  AND ECONOMIC  PROBLEMS WITH  INSTALLING A RETROFIT FIXED ROOF  ON LARGE
DIAMETER TANKS,   THE WALLS AND FOUNDATIONS  OF LARGE DIAMETER  EXTERNAL
FLOATING ROOF TANKS  MAY NOT SUPPORT A FIXED ROOF,
2,  THE AGENCY  HAS AN  OBLIGATION UNDER EXECUTIVE ORDER  12291  TO REVIEW
OTHER EMISSION  REDUCTION SCHEMES WHICH MAY  BE MORE  COST EFFECTIVE THAN
AN  INTERNAL FLOATING ROOF WITH PRIMARY AND  SECONDARY SEALS.   FROM ERA'S
EMISSION VS, VOLUME GRAPHS IN THE VOL STORAGE BACKGROUND INFORMATION
DOCUMENT, WE FIND  SEVERAL ALTERNATIVES  WHICH ALSO ACHIEVE 90  PERCENT
REDUCTION FOR LARGE TANKS,  ANY OTHER  STRATEGY AS EFFECTIVE  AS  THAT WHICH
EPA HAS DEFINED AS RACT SHOULD BE ALLOWED,   WE SUBMIT THAT EPA'S  OWfr
DATA SHOW THE EFFECTIVENESS OF EXTERNAL FLOATING ROOF TECHNOLOGY  FOR
LARGE DIAMETER TANKS,   FURTHER, THE API  DATA BASE ON SEAL TECHNOLOGY
WILL FILL THE GAPS IN EPA'S RESEARCH,   WE BELIEVE, AND THESE  DATA  MAY
PROVE,THE EMISSION RATE FROM VOL STORAGE TANKS IS MORE  A FUNCTION OF
SEAL TYPE AND COMBINATION RATHER THAN  ROOF  TYPE,   THE API STUDY ON
CONTROL EFFICIENCIES OF OTHER TECHNOLOGIES  WHICH MAY BE MORE  COST-
EFFECTIVE THAN  THE RECOMMENDED RACT IS  IMMINENT,  THE USE OF  SEAL
TECHNOLOGY MAY  WELL  BE  MORE COST-EFFECTIVE  THAN THE COSTLY RETROFIT
                                 V-15

-------
OF FIXED ROOFS ON EXTERNAL FLOATING ROOF VOL STORAGE TANKS,  THEREFORE,
WE RESTATE OUR CONCERN THAT EPA SHOULD FURTHER EXPLORE THE  USE OF
ALTERNATE TECHNOLOGIES BEFORE  ISSUING THE FINAL CTG,  UNLESS AN ACCURATE
ASSESSMENT OF COST-EFFECTIVENESS OF THE VARIOUS ALTERNATIVE  TECHNOLOGIES
IS INCLUDED IN THE FINAL CTG,  THE STATES WILL NOT BE ABLE TO INCORPORATE
THE MOST COST EFFECTIVE CONTROLS NECESSARY TO ATTAIN THE AMBIENT
STANDARDS,
3.  THE FIVE YEAR INSPECTION REQUIREMENT is UNNECESSARY  •
    CMA SUBMITS THAT THE REQUIREMENT FOR AN INSPECTION EVERY FIVE YEARS
AFTER INSTALLING AN INTERNAL FLOATING ROOF WITH LIQUID MOUNTED PRIMARY
SEAL AND CONTINUOUS SECONDARY  SEAL IS TOO FREQUENT,  SEAL LIFE IS CON-
SIDERABLY MORE RELIABLE THAN EPA ESTIMATED,   WE WOULD RECOMMEND AN
INSPECTION .WHENEVER THE TANK IS EMPTIED  AND  DEGASED,  BUT  AT  A MINIMUM
OF EVERY 10 YEARS,   THE FIVE YEAR  EMPTY  TANK INSPECTION FREQUENCY  REQUIRED
BY EPA WAS NOT DERIVED FROM A STUDY OF SEAL  LIFE  BUT  RATHER  FROM A
STUDY OF WHEN TANKS ARE EMPTIED AND DEGASED,  As API .STATED  IN THEIR
NAPCTAC COMMENTS ON VOC STORAGE TANKS, DIFFERENT  COMPANIES HAVE
DIFFERENT OPERATIONAL AND DESIGN PRACTICES ON  INTERNAL  FLOATING  ROOFS,
THESE INVOLVE THE USE OF CONTACT AND NONCONTACT  ROOFS,  TEMPERATURE
LIMITATIONS,  SIZE LIMITATIONS,  GAS  FREEING PROBLEMS AND OTHERS,  API
       *
MAKES THE POINT, WHICH WE WANT  TO  RE-EMPHASIZE,  THAT  IN THE  ABSENCE  OF
COMPELLING EMISSION DATA, THE CTGs  SHOULD RESPECT  INDIVIDUAL COMPANY,.
PRACTICES,
    WITH REGARD TO  THE ISSUE OF EMPTY TANK  INSPECTION  FREQUENCY, WE
DISCUSS AS AN EXAMPLE ONE MEMBER COMPANY'S SEAL  TECHNOLOGY  EXPERIENCE.
THE COMPANY HAS EXTENSIVE OPERATING EXPERIENCE  WITH  FLOATING ROOF  SEAL
TECHNOLOGY IN EXTERNAL FLOATING ROOF SERVICES,   IN  GENERAL,  CAUSES OF
                                  V-16

-------
PREMATURE SEAL FAILURE HAVE BEEN DUE TO THE FOLLOWING:
    1,  ULTRA VIOLET DEGRADATION OF THE SEAL MATERIAL,
    2,  THE PRESENCE OF DEBRIS WHICH PUNCTURES THE SEAL (USUALLY DUE
       TO POOR MAINTENANCE OF THE WEATHER SHIELDS),
    3,  ABRASION OF THE SEAL DUE TO OUT OF ROUNDNESS,  AND
    4,  PRODUCT INCOMPATIBILITY,
IN SPITE OF THESE PROBLEM% AND EVEN WITH POOR WEATHER SHIELD MAINTENANCE,
A 10 YEAR SEAL LIFE IS NORMAL IF PROPER SEAL MATERIAL IS USED,   HOWEVER,
THE EPA REQUIREMENT OF AN INTERNAL FLOATING ROOF FOR  VOL STORAGE WILL
MINIMIZE THESE CAUSES OF SEAL FAILURE AND EXTEND SEAL LIFE WELL BEYOND
THE NORMAL 10 YEARS FOR THE FOLLOWING REASONS!
    1,   THE PRESENCE OF A ROOF WILL MINIMIZE ULTRA VIOLET DEGRADATION,
    2,   THE ROOF WILL MINIMIZE WEATHER-INDUCED RUST  (THE MAIN SOURCE
        OF DEBRIS CAUSING SEAL DAMAGE),
•  '  3,  THE PRESENCE OF A ROOF WILL IMPROVE THE ROUNDNESS OF THE
        TANK AS COMPARED TO AN EXTERNAL FLOATING ROOF TANK,
WE CONCLUDE EPA SHOULD SIMPLY REQUIRE THE EMPTY TANK  SEAL INSPECTION
WHEN THE TANK  IS EMPTIED AND DEGASED AND THAT THIS OCCUR AT A
MINIMUM OF ONCE EVERY 10 YEARS,
4,  THE 30-DAY NOTIFICATION REQUIREMENT FOR THE FIVE-YEAR INSPECTION IS
ADEQUATE FOR THE CASE OF'THE SCHEDULED INSPECTION BUT IS NOT ADEQUATE
FOR AN UNSCHEDULED INSPECTION.
    THERE ARE OCCASIONS WHERE TANKS ARE EMPTIED AND  INSPECTED BECAUSE
       «
OF UNPLANNED CIRCUMSTANCES,  THESE INCLUDE:
    li   A LOW  INVENTORY DUE TO PLANT SHUTDOWNS, SHIPPING DELAYS OR
        STRIKES,
    2,   A LEAK IN THE TANK OR FLOATING ROOF,
    3,   A CHANGE IN PRODUCT IN THE TANKS, AND OTHER  REASONS,
THE AGENCY SHOULD DROP THE REQUIREMENT FOR A 30-DAY  PRIOR NOTICE FOR

-------
REFILLING A STORAGE VESSEL TO ALLOW THE ADMINISTRATOR THE OPPORTUNITY
TO INSPECT THE STORAGE VESSEL,  HAVING A STORAGE VESSEL.OUT  OF  SERVICE
FOR THIRTY DAYS TO ALLOW THE ADMINISTRATOR TIME TO DECIDE IF HE WILL
INSPECT THE STORAGE VESSEL IS PUNATIVE, .INSTEAD,  WE  SUGGEST  THE AGENCY
REQUIRE THAT A CERTIFICATION OF THE TANK INSPECTION  FROM A COMPANY
REGISTERED PROFESSIONAL ENGINEER BE ON FILE AT THE PLANT,  THIS
APPROACH HAS BEEN USED SUCCESSFULLY TO CERTIFY AND UPDATE OIL SPILL
AND COUNTERMEASURE PLANS AND SHOULD BE SUFFICIENT FOR THIS PROGRAM,
5,    COMPLIANCE SCHEDULE
      THE COMPLIANCE SCHEDULE IN THE MODEL REGULATION IS UNREALISTIC.
PARTICULARLY FOR MULTIPLE TANK INSTALLATIONS.  THE TWO MONTHS ALLOWED
FOR ON SITE CONSTRUCTION IS A PARTICULAR PROBLEM,  AN OPERATOR  WITH
MULTIPLE TANKS 'HAVING TO CYCLE TANKS IN  AND  OUT OF  PRODUCTION  DURING
WINTER TO ACCCmmATE  RETROFITS WOULD NOT BE ABLE'TO  MEET THE TWO MONTH
LIMITATION,  ALSO, OTHER SEGMENTS OF THE SCHEDULE MAY BE IMPOSSIBLE
TO MEET WITHIN THE SPECIFIED TIMES BECAUSE  OF VENDOR DELIVERY SCHEDULES,
WEATHER, LABOR PROBLEMS OR OTHER INDIVIDUAL CIRCUMSTANCES,    $E  RECOMMEND
INSTEAD OF SPECIFING A CONSTRUCTION SCHEDULE  IN THE MODEL REGULATION THAT
EPA  SHOULD SPECIFY THE PERMITTEE TOTAL ELAPSED TIME AND  THE  SOURCE  SHOULD
AGREE WITH THE STATE ON THE SCHEDULE WITHIN THAT TIME,
                        V
      FURTHER, THE 14 MONTH COMPLIANCE  SCHEDULE is INADEQUATE,  A 24
                                                                 «
MONTH TIME IS MORE REALISTIC,
      CMA WILL SUBMIT SEPARATE WRITTEN  COMMENTS  WHICH  WILL DETAIL
OJR  CONCERNS ON THESE AND OTHER ISSUES,   SOME  OF  THE  OTHER  ISSUES  •
                                                                   «.
 INVOLVE A REQUEST TO EXEMPT FROM THIS  CTG THOSE COMPOUNDS WHOSE
VAPOR PRESSURES EXCEED 1,5 PSIA FOR  10 PERCENT OF  THE YEAR
AND  THAT THE RECORDKEEPING REQUIREMENTS LAPSE  AFTER TWO  YEARS FOR
TEMPERATURE AND VAPOR PRESSURE DATA FOR COMPOUNDS  ABOVE  1,0  PSIA,
                                 V-18

-------
CONCLUSION
      WE FEEL EPA SHOULD CONSIDER ALTERNATE TECHNOLOGIES IN DEFINING
RACT IN. THE FINAL CTG,   THIS WOULD ALLOW FOR THE USE OF THE MOST COST
EFFECTIVE EMISSION CONTROL TECHNOLOGY IN ACCORDANCE WITH EXECUTIVE
ORDER 12291,•
      THIS CONCLUDES MY FORMAL STATEMENT, I  WILL  ATTEMPT TO ANSWER
ANY QUESTIONS YOU MAY HAVE CONCERNING MY PRESENTATION,
                                 V-19

-------
    2.  Texas Chemical Council
                              Mr. A.  H. Nickolaus
                       E.  I. du Pont de Nemours & Company
                               P. 0. Box 2626
                            Victoria, Texas  77901
          My name is Andy Nickolaus and I represent  the  Texas Chemical
Council (TCC).  The TCC is an association of  85 chemical companies having
more than 67,000 employees and representing approximately 90% of the
chemical industry in Texas.  About 35% of the tanks  that will be affected
by this guideline are in Texas so its requirements are of vital concern
to us.  Our comments are given below.

          Since the test data upon which the  EPA claims  to be basing
this guideline are for benzene, we have used  benzene to  illustrate our
concerns about VOL storage generally; but we  do not  agree with page 3-1
of the guideline where it states "It is believed that the benzene test
results can be applied to any tank storing VOL".

  I.  Comments On The Technical Basis For The Proposed Guideline

      Before starting our comments we believe clarification of some of
the terms used is necessary-  The EPA is basing their proposed standard
on test work done in a 20 ft. diameter pilot  test tank at the Chicago
Bridge and Iron Company's  (CBI) research facility at Plainfield, Illinois.
These tests followed work done with petroleum mixtures in both external
and "internal" floating roofs.  But there are at least two major
configurations of "internal" floating roof tanks.  One is the type
tested by CBI which is described in Appendix  H of API Standard 650 as
a "Covered Floating Roof"  (CFR).  These roofs are ventilated with
openings around, the top of the tank wall whose purpose is to prevent a
flammable mixture from developing between the fixed  and  floating roof.
The second type is the usual fixed-roof tank  with an internal floating
roof inside it.  Normally these will have only a breathing vent which
may be equipped with a conservation vent and/or flame arrester (see
Figure 1).

      We have used the following definitions  in our  discussion to
distinguish among the three types:

      •  External Floating Roof  (EFR)  -  A floating roof with no cover
         above it.

      •  Covered Floating Roof (CFR)  -  A floating  roof of the type
         tested for the EPA by CBI which is described in Appendix H
         of API Standard 650 and is characterized by having peripheral
         vents whose purpose is to ventilate  the space between the
         floating roof and the fixed roof  (tank cover) .
                                   V-20

-------
      •  Closed Tank Floating Roof (CTFR)  -  A floating roof installed
         inside a fixed-roof tank that has no vents whose purpose  is  to
         ventilate the space between the floating roof and the fixed
         roof.

      A. Is The Proposed Technology Safe?

         The safety of the proposed technology is not discussed in the
guideline.  We think it should be.

         Benzene vapors are flammable between 1.41 and 6.75 vol% in air.
Benzene vapor pressure ranges from 4.9 vol% at its freezing point  to
over 22 vol% at 100°F.  Thus, throughout the range of probable storage
temperatures benzene is capable of generating a flammable mixture.  A
flame or explosion requires three things - fuel, oxygen, and an ignition
source.  At least one of these must be controlled.

         For normal storage in closed, fixed-roof tanks the approach
is to eliminate ignition sources; and this is what the National Fire
Protection Association Code 30 (Ref. 1) is intended to do.  Benzene with
a 12°F flash point and 176°F boiling point is a Class IB flammable
liquid.  As such, it requires for "Normal Venting for Aboveground  Tanks"
(2-2.4.6) that "Tanks and pressure vessels storing Class IB and 1C
liquids shall be equipped with venting devices which shall be normally
closed except when venting under pressure or vacuum conditions, or with
listed flame arresters".

         The Covered Floating Roof (CFR) technology proposed by the EPA
depends on natural circulation from vents (without flame arresters) in
the top of the tank wall to keep the organic vapor concentrations  below
the flammable limit.  Chicago Bridge and Iron Company (CBI) claims in
Bulletin No. 3200 for their CBI Weathermaster CFR tank system  (the one
tested for the EPA) that "Tests have proven that the space between the
floating roof and the fixed roof does not contain flammable mixtures
except for a short time immediately after product is pumped into an
empty tank".  They also state that these meet the requirements of  API
Standard 650  (Sixth Edition), Appendix H.  This publication requires
(H.3.8b) that "Circulation vents or openings shall be located above the
seal of the floating roof when the tank is full.  The maximum spacing
shall be 32 ft., but in no case shall there be less than four equally
spaced vents.  The total open area of these vents shall be equal to,
or greater than, 0.2 sq.ft. per ft. of tank diameter.  This empirical
figure has been established, on the basis of successful practice."
(Underlining added)  In proposing this standard, the EPA apparently
intends to support their limited benzene test data with technology
transferred from the more extensive test and field experience with
petroleum mixtures.  But is this technology applicable to benzene
storage?  We think not.
                                   V-21

-------
                                   - 3 -
         EPA's test results show that benzene emissions are over ten
times greater than those of their prototype petroleum liquid even after
normalizing to the same "True Vapor Pressure" (TVP) and molecular weight
(see Figure 2).   With such poor agreement, even under controlled test
conditions, it is apparent that technology applicable to petroleum
liquids cannot be transferred wholesale to benzene or to VOL storage.
Further, no data are given by the EPA to show that natural circulation
through the vents is sufficient to maintain benzene or VOL concentrations
below explosive limits.

         Also, the proposed pan-type floating roof does not fail
safely, and there have been a number of sinking failures reported.  One
company survey showed that of 138 steel-pan roofs in use, 22 had sunk
at one time or another.  A second company reported that 11 pan-type
floating roofs had experienced 12 sinkings over a five year period.  A
third company reported that over a period of time all 5 pan-type internal
floating roofs at one location had to be replaced because of a high
frequency of sinkings arising from the nature of their operations.  One
company described their experience at one location as follows:

         "We have, since 1972, installed at least 20 internal floating
         roofs in existing or new tanks for the control of vapor emissions
         from a variety of volatile organic liquids.  Of these floating
         roofs 14 pan-type were installed in 1972-75 to meet TACB
         requirements and 8 pontoon types have been installed since then
         to meet various compliance requirements.  Of the 8 pontoon types
         installed 2. were to replace pan-types that could not be repaired.
         These replacements cost, an additional $126,000 beyond the
         original installations.'

         'Of the 14 pan-type installed, 8 have sunk at least once, several
         as many as three different times.  Two have been replaced, as
         mentioned above, and the ones that could be repaired were
         strengthened with auxiliary pontoons to provide stiffness
         and bouyancy before returning to service.  ...  We have had
         no  (sinking) problem with pontoon-type roofs ...."

         These examples are sufficient to show that roof sinkings are
not a remote possibility but a real probability that must be considered.
When such a sinking occurs- a free liquid surface is exposed and the
vessel becomes equivalent to the ordinary cone-roof tank covered by
National Fire Protection Association Code 30.  But it does not meet
this code for it is now a vessel with vents that cannot be closed and
which do not have flame arresters.  This problem cannot be dismissed
with a statement that proper operation and maintenance will prevent roof
sinkings.  Safety must be concerned with the consequences of inadvertant
mis-operation and undetected deterioration so as to minimize the hazard
from these also.
                                    V-22

-------
                                   - 4 -
         With the higher emissions from benzene  (and presumably vols
generally),  without supporting safety data from  the EPA, and with  the
reported tendency of pan-type roofs to sink, the TCC questions the safety
of the proposed technology.  The TCC is not saying the proposed technology
is unsafe;  we don't know.  But we do believe it  is incumbent on the EPA
to address and affirmatively answer this question before proceeding further)

         The justification for this guideline is to help attain the
National Ambient Air Quality Standard for ozone.  This standard was set
"allowing an adequate margin of safety ... to protect the public health".
We believe that any technology imposed in trying to get there should also
contain an "adequate margin of safety" to protect the health of those of
us who must apply it.

      B. Is The Proposed Guideline Adequately Supported By Test Data
         Or Field Experience?

         Because of the large difference between the predicted and
observed benzene emissions, we concluded above that the data and field
experience from petroleum liquids is not generally applicable to VOL
storage.  This means the whole standard rests on a few experiments
carried out in a small test tank under conditions which may not reflect
field conditions.

         Further, the TCC does not believe that the data they do have
clearly support what they have proposed.  The EPA has specified an
"internal", contact floating roof with a liquid and vapor mounted  seal
based, they say, on test results and engineering judgement showing
this to be the most effective configuration.  We disagree.  We do  agree
with the November 29, 1979 comments by the American Petroleum Institute
(Ref.2) which stated:

         (1) "The emission test data do not support the conclusion in
         the draft report that contact-type floating roofs significantly
         reduce emissions as compared to non-contact type internal
         floating roofs.'

         (2) 'The data utilized do not support the general conclusion
         in the draft that the addition of secondary seals over primary
         seals provides significant emission reduction in internal
         floating roof tanks (IPRTs).'

         (3) 'The data do not support the conclusion that secondary
         seals in external floating roof tanks provide only a small ...
         emission reduction over primary seals."

The TCC recommends the EPA review this API statement for a fuller
discussion of these, the determination of the ks, n, kf, and m
parameters, and other important points.
                                    V-23

-------
                                   - 5 -
         This is the third or fourth time this information has been
presented to the EPA, so far to no avail.  In fact the EPA seems to be
going away from the data.  The people at CBI who did the test work
thought most of the differences between emissions of contact and non-
contact roofs was due to the differences in the seals (Ref. 3, page 3) .
In the BID for the NSPS the EPA incorrectly summarizes the test results
by ascribing the differences to the type of roof rather than the seals
(Ref. 4, page 4-12).  Now, in this guideline, they dogmatically state
that "A contact internal floating roof with primary and secondary seals
provides the greatest emission reduction over all other roof and seal
combinations" (see page 3-2).

      C. Why Are There No Closed Tank Floating Roof (CTFR) Data?

         Earlier we distinguished between two types of "internal" floating
roofs; the Covered Floating Roof (CFR) and the Closed Tank Floating Roof
(CTFR).  What the EPA tested was the CFR configuration,  and this is the
one they mean when they specify an internal floating roof - see paragraph
3.3.1 in the guideline.  No tests were run under conditions equivalent to
those of a CTFR.  We are puzzled by this since the CTFR would be less
expensive in most cases, is easier to retrofit, fails safely, and would
appear to have lower emissions.  Extrapolation of the benzene test data
to 1 mph wind speed as an approximation of the zero wind conditions inside
a CTFR shows lower emissions in all cases (see Figures 3-2, 4, 6 of Ref. 3)
The TCC doesn't want to make too much of these extrapolations since data
under CTFR conditions are what is needed, but they certainly don't support
the need for a pan-type roof.  In fact, even the shingle seals on the
pontoon roof appear to be equivalent to the double seals on the pan-type
roof under these circumstances (see Figure 3 attached).

         Perhaps the EPA's obsession with "enforceability" has driven
them from considering this internal roof configuration since the hatches
cannot be opened for inspection without degassing the tank as a flammable
mixture may exist or be created near the opening.  Safety for Closed Tank
Floating Roofs is provided by flame arresters on the vent and making the
internal roof spark-proof by grounding, fabrication from non-sparking
materials, etc., instead of trying to keep vapor concentrations below
flammable limits.

      D. Summary Of Comments On The Technical Basis For The Guideline

         For the reasons discussed above, we conclude:

          (1) Data and field experience from petroleum liquid storage
         cannot be transferred wholesale to VOL storage.

          (2) The EPA has not addressed the safety of the technology
         they are imposing.  They should.  Also, the proposed system
         fails unsafely and there is a history of failures.
                                   V-24

-------
                                   - 6 -
         (3)  The test data are limited to less than a dozen runs/ all
         gathered in a small (20 ft. diameter) pilot test tank, and the
         EPA has misinterpreted and mis-used these.

         (4)  The EPA has ignored other floating roof configurations
         that may be better and less expensive.

         From all this the TCC believes that the data are woefully
inadequate to support the proposed guideline.

 II.  Comparison With The Petroleum Liquid NSPS

      At the December 2-3, 1980 NAPCTAC meeting, the TCC questioned why
the proposed NSPS for VOL was so much more stringent than the NSPS for
Petroleum Liquid Storage Vessels promulgated April 4, 1980.  Today, we
are quentioning why the RACT guideline for VOL storage is more stringent
than this same NSPS for Petroleum Liquids.  Some of the major differences
are:

                                   PETROLEUM         VOL STORAGE
                                  LIQUID NSPS       RACT GUIDELINE

Size Cut-Off, Gal.                  40,000              40,000

Vapor Pressure Cut-Off, psia         1.5                 1.5

Allowable Configurations:  External                 Not
                           Floating                 Allowed
                           Roof
                             Or

                           Internal Floating        Internal
                           Roof                     Floating Roof
                             Or                          Or

                           Vapor Recovery           90% Reduction
                           System (95%)             System

Internal Roof Features:

      Seals                Single                   Double

      Type Roof            Any                      Pan-Type Only

III.  Comments On The Model Regulation

      A. Applicability

         Petroleum liquids are to be exempted from this regulation.  We
presume this means that cyclohexane made at a refinery (Ref. 5) and
stored at a chemical plant could be stored in a fixed-roof tank fitted
                              V-25

-------
                                   - 7 -
with any sort of internal floating roof and seal  (Ref. 6) or even in a
single seal external floating roof (Ref. 7).  In contrast, cyclohexane
made by benzene hydrogenation would have to be stored in a CFR tank
having a pan-type internal roof and double seals.  In our comment on
the VOL NSPS  (Ref. 8) we didn't think regulating chemicals based on their
ancestry made any sense.  We still don't.

      B. Definitions

         (1)  "Actual Vapor Pressure" - Why was this term introduced when
         "True Vapor Pressure" was used for the VOL NSPS BID and proposed
         draft regulation?  What is the difference between the two terms?
         "Entities should not be multiplied beyond necessity." (Occam's
         razor).

         (2)  "Internal Floating Roof" - This definition is incomplete.
         What the EPA has in mind is the ventilated CFR configuation
         (see paragraph 3.3.1) and the definition should say so.

      C. Standards

         For  reasons already discussed we have concluded that the EPA
does not have the technical and safety data needed to support the
standard they are proposing.  In fact, the TCC does not believe a
standard is necessary at all.  Adequate controls have already been
imposed by the states under their Implementation Plans.  However, if
we must have one, we propose something similar to the Petroleum Liquid
NSPS where several options were allowed.  We recommend:

         "The owner or operator of each storage tank to which this
         regulation applies shall equip each storage tank with one
         of the following:

         (1) An external floating roof with double seals,

         (2) A ventilated internal Covered Floating Roof  (CFR) with a
         liquid-mounted primary seal and a secondary seal  (assuming
         the EPA is able to certify that this configuration is safe
         for VOL storage),

         (3) A closed-tank floating roof (CTFR) with either a single
         or double seal, or

         (4) An alternate control technology which achieves an overall
         emission reduction equivalent to that calculated for option 2
         above or an 80% reduction calculated for a fixed roof storage
         tank, whichever is less.  The equations used shall be as
         specified in the attached method  (attached equations from
         part 2 of the proposed guideline)."
                                   V-26

-------
                                   - 8 -
         We consider the proposal to require retrofitting existing
external floating roof  (EFR) tanks with fixed roofs to be completely
unreasonable and far beyond any definition of RACT.  The walls  of many
EFR tanks will not support a fixed roof, and for those that will, the
foundations probably won't.  Most tanks in Texas are on concrete slabs
which "float" on gumbo.  If adding a fixed roof doesn't buckle  the  tank
walls, it will probably crack the foundation along the tank perimeter.
Further, we along with the API question the emission reductions to  be
gained by this.

         No upper limit is given for the vapor pressure of volatile
organic liquids to be stored in CFR tanks and we presume it is  up to
76.6 KPA (11.1 psia), the same as the draft NSPS and we are still
concerned about the safety of these.  The benzene data do appear to
support the benefits of a double-seal for this configuation and we
accept this.  We do believe unsinkable types are the preferred  floating
roof for these ventilated tanks.  The pontoon type is more suitable for
retrofitting as it can be taken into the tank through a manhole and
installed within a week.  Some of the other types involve removing  the
fixed-tank roof — an expensive and time consuming operation.

         Although the EPA has no data taken at Closed Tank Floating
Roof  (CTFR) conditions, extension of the data they do have and  engineering
judgement indicates that a floating roof, pontoon or pan-type,  with
either single or double seals would be equal to the CFR double  seal
configuration.

         For alternate control technology we believe the basic  comparison
should be to the proposed technology, not to some other basis.  We
presume the EPA is not proposing to foist off the three cases in Table 4-1
as representative of the reductions attainable by their technology  for
all VOLs under all conditions of storage and use.  Thus, we have
recommended that the basic comparison be against the recommended technolo^
For those, who for one reason or another, cannot use the prescribed
technology we believe their liability should be limited to an 80% reduction
from a fixed-roof tank emission under the same conditions.  The TCC
believes that 90% is a more realistic level for Best Available  Technology
 (BAT) and that 80% is more representative of Reasonably Available Control
Technology.  Without data on each specific VOL covered by this  proposed
regulation we do not believe a figure higher than 80% can be justified.
In proposing these broad, industry-wide standards the TCC believes  the
EPA must provide several options of approximately equal cost to the owner/
operator or be willing to concede in specific cases that their  technology
is not RACT and that the standard does not apply.

         We take exception to the use of a conservation vent on the base
case for fixed-roof storage emissions.  Although there is mention  (Ref. 8)
of a 0.86 inches of water provision to simulate a conservation  vent,
there is no indication that the breathing losses reported in References
                                    V-27

-------
                                   - 9 -
8 and 4 were from tanks with conservation vents.  Besides, properly
used, the lowly conservation vent can achieve significant percentage
reductions in emissions.  In dead storage (zero turnover) it is better
than EPA's proposed technology for the example storage tank in Appendix B
of the guideline (see Attachment 1).  Comparable emissions are:

         Conservation Vent:       1.8 Mg/Yr

         Covered Floating Roof:   2.6 Mg/Yr

         We presume that conditions appropriate to the location can be
used in calculating equivalence and emission reductions.  For eample,
a 20°F diurnal temperature change was used to calculate breathing
losses.  We believe this is too high for most cases.  VOL producers and
terminals are concentrated in areas where the mean average daily
temperature change is nearer 15°F.  See average daily temperature changes
on Page 8 of CBI Bulletin 533 for the Gulf Coast, California Coast, the
East Coast from Norfolk to New York, and around Lake Michigan.

      C. Inspection

         Some plants only inspect tanks storing certain materials every
ten years.  We request that XX.040(B) be changed to say "at least once
every 10 years after installing the control equipment".

         The discussion of the regulation implies that an inspector can
come in, "equipped with an explosion proof flashlight", and open a hatch
and inspect the floating roof; all in short order.  It's not that simple.
The CTFR hatches normally cannot be opened without degassing the tank.
For both CFR and CTFR tanks making explosibility measurements prior to
opening will be required in many cases.  Unless the VOC concentrations
are below permissible exposure limits, respiratory protection will be
necessary.  Also it is highly probable that the proposed inspection will
fall under OSHA's vessel-entry regulations..  This means that in addition
to a supplied-air, full-face mask and a rescue rope on the inspectior, a
similarly equipped stand-by man is needed.  All this takes coordination
of several plant groups, time, etc.  Also, some materials are sensitive
to oxygen and/or water and are stored in nitrogen padded tanks so that
the hatches on these cannot be opened while the tank is in use.

      D. Recordkeeping

         Why should the owner/operator be required to keep records on
materials whose vapor pressure is greater than 1.0 psia?  Does the
regulation apply to VOLs with vapor pressures 1.5 psia and greater, which
is what it says, or not?  What useful purpose is served by these records?
Are they necessary for compliance, or are they for ease of enforcement?
The former is lawful, the latter is not.
                                    V-28

-------
                                    - 10  -
      E. Compliance Schedule

         The model compliance schedule is 14 months.  As noted in the
discussion in the guideline this is for a single installation  and when
State regulations become effective there may be a  surge in  demand that
.will extend this time.  We think 24 months is a more realistic estimate
and one that allows scheduling tanks out for retrofitting without
unreasonable production penalties.

 IV.  Miscellaneous

      A. Need

         The TCC doubts that this guideline is needed for reasons set
forth in our comments to NAPCTAC on the VOL Storage NSPS (Ref.  9).   So
far all our comments on need have been ignored so  we see no use in
elaborating on them here.  The TCC believes that laws and regulations
should be made only where there is a definite need and where a dis-
cernable benefit will result; and also that regulations should be
reasonable, fair, and soundly based technically.   This proposed regu-
lation fails on all counts.

      B. Costs

         After providing detailed costs and analysis on the Fugitive
Emission Monitoring draft BID only to be told at a meeting with the EPA
last summer that it wouldn't matter if the BID costs were off  by  a factor
of two or more, the TCC has become discouraged in  this area and we have
not carefully reviewed the cost analysis in Chapter V.  It's not  that
costs aren't important to us.  They are; but, we don't believe  this is
a useful forum to discuss them in.

      C. Rulemaking

         We believe EPA's insistance that these model RACT regulations
be incorporated essentially verbatum into the SIPs is tantamount  to
rulemaking despite their disclaimers and legal technicalities  that
avoid this fact.


  V..  Endorsement of CMA Comments

      The Chemical Manufacturers Association (CMA)  comments cover  some
items we have not and compliments our's on others.   The TCC agrees with
and endorses the CMA comments.
AHN/rtg
3-9-81

Attachments

                                    V-29

-------
REFERENCES:
      (1)  National Fire Protection Association, 1980 Fire Codes,
           Vol. 2, NFPA30-1977.

      (2)  American Petroleum Institute (API): November 29, 1979,
           Comments on the Chapters 3, 4,  and 5 of the Draft BID.
           Contained as Attachment C in the API's comments of December 1,
           1980, in the Minutes of the December 2&3, 1980 NAPCTAC Meeting.

      (3)  Laverman and Cherniwchan, "Measurement of Benzene Emissions
           From a Floating Roof Test Tank" Draft Final, May, 1979-

      (4)  Preliminary Draft BID, "VOC Emissions from Volatile Organic
           Liquid Storage Tanks - Background Information for Proposed
           Standards", EPA, November, 1980.

      (5)  Petroleum Liquid Storage Vessels; Standards of Performance
           for New Stationary Sources.  FR45 P- 23374  (April 4, 1980)
           identified cyclohexane as an example material that would be
           covered under "petroleum liquids".

      (6)  EPA 450/2-77-036 Guideline for Control of Volatile Organic
           Emissions from Storage of Petroleum Liquids in Fixed-Roof
           Tanks.

      (7)  EPA 450/2-78-047 Guideline for Control of Volatile Organic
           Emissions from Petroleum Liquid Storage in External Floating
           Roof Tanks.

      (8)  EMB Report 78-OCM-5, February,  1979, "Emission Test Report -
           Breathing Loss Emissions from Fixed-Roof Petrochemical Storage
           Tanks".

      (9)  TCC Comments on the BID and Proposed VOL Storage NSPS given
           at the December 2&3, 1980 NAPCTAC Meeting.

                                   V-30

-------
                                        ATTACHMENT 1
                         EMISSIONS FROM DEAD STORAGE IN EXAMPLE TANK

EXAMPLE CALCULATION STORAGE TANK - APPENDIX B, PRELIMINARY DRAFT CTG  -
 DEAD STORAGE, FULL TANK CASE
                                                          CHANGES
MATERIAL:  METHYL EHTYL KETONE   Mw » 98.96                FROM
                                                          EXAMPLE
TEMP-:                   77°F
ANNUAL TURNOVERS:        0 (DEAD STORAGE)  VS 15       *
AVG.AT, °F(GULF COAST): 15°F  VS.  20°F              —»	
TANK CAPACITY:           1,697,933 GAL.
TANK DIA. :               85'
TANK HEIGHT:             40'
VAPOR SPACE:              5f (~90% FULL)  VS 20'     -«	
A.  EMISSIONS - NO CONTROL
    Lt = Lb + Lw = 9.15 X 10"6(98.96)(0.2733)(85)1'72(5)>51(15)*5(1)(1) + 0
         4.74 + 0 « 4.74 Mg/Yr
B.  EMISSIONS WITH CONSERVATION VENT, 6" H2P PRESS -I- 1" H20 VAC.
    AT% -  Av% =  AP%
    AT        15
     T   =  460 + 77  *
0.02796 OR 2.796%
    .  .Ap  -  2.796% OR  .02796 ATM X 407" H20/ATM  =  11.4" H20
       ASSUME PORTION  AP CONTAINED  -  PORTION  AV CONTAINED.

       6" * f  •  61.4% AP CONTAINED  -  61.4%   AV CONTAINED.
         X<1> * *T

    .  .EMISSIONS  »  4.74 - (4.74 X 61.4%) -1.83 Mg/Yr

C.  EMISSIONS WITH CONTACT INTERNAL FLOATING ROOF

    CALCULATION SAME AS EXAMPLE EXCEPT Lw = WORKING LOSSES  = 0 SINCE TURNOVERS * 0
     *
    .  . Lt  »  Lw + Ls -I- Lf - 0 + 2.19 + 0.41  -  2.60 Mg/Yr
AHN/rtg
3-13-81
                                           V-31

-------
                  INTERNAL  FLOATING  ROOF TANKS
           COVERED FLOATING ROOF


                  (CFR)
CLOSED-TANK FLOATING ROOF


            (CTFR)
I
CO
PO
                          Circulation vents
                  Pan-type
                                                                    Flame arrester
         Pontoon-type
                                                                              i
2

5
c:
:o
rn
      Flammability protection by natural


      circulation to keep vapor concentrations


      below explosive limits.
  Flammability protection by


  eliminating ignition sources.

-------
         BENZENE VS. PETROLEUM MIXTURE EMISSIONS
 10
  a

  N.
  a
 0.1
 m
 e
 to
 "35
0.01
                                            l«nz*n*

                                            T«s« IPA-12 Q
                                            Prep«n«/eceaae
                                           SR-3 flexible fsan aririary s«ai
                                           vithout gaps, flapper secar.dary
                                           s«al without ?aps, deck  fittings
                                           sealed, results normalised so
                                           1.75 psia vapor pressure and
                                           73.1 vapor ealeeular weight
                          10
                                          4O m ph
                 Wind Speed
Ret. 2 j Figure 3-5
                      EMISSIONS VS. W1SD SP5ID  FOR AN
                      ISITESNAL PAN TY?E FLCA/TING ROOF
                      WITH PSIMAP.Y AND  SZCCIIDASY S2AL ,
                      COMPARISON ?-7IT2 P30PAHZ/OCTANZ T2ST DATA
                            V-33

-------
                           FIGURE 3
   100
    10
X

g
\
CO
CO
H
i
    0.1
                                           TEST 18 Q  PONTOON TYPE

                                           TEST 11 &  PAN TYPE

                                         "&TEST 14 0  PAN TYPE
                                            DATA FROM REF. 2,

                                            FIGURES 3-2,4,6
                             10

                       WIND  SPEED
                 40 MPH
           COVERED FLOATING  ROOF TANK EMISSIONS
        (EMISSIONS EXTRAPOLATED TO  1 MPH WIND SPEED)
       EPA TEST  18   -


       EPA TEST  11   -

       EPA TEST  14   -
BOLTED COVER (PONTOON)  TYPE INTERNAL FLOATING
ROOF WITH PRIMARY & SECONDARY SEALS, BOTH
GAPPED (FIG. 3-6)
PAN-TYPE FLOATING ROOF WITH PRIMARY SEAL ONLY,
GAPPED (FIG. 3-2)
PAN-TYPE FLOATING ROOF WITH PRIMARY & SECONDARY
SEALS, BOTH GAPPED (FIG. 3-4)
                                 V-34

-------
3.  GATX Terminals Corporation
                                Mr. R. W.  Bogan
                           GATX Terminals Corporation
                            120 South Riverside Plaza
                            Chicago, Illinois  60606

                 (Letter read at NAPCTAC Meeting by Mr. Fred Porter)
    Mr. Don R. Goodwin,  Director
    Emission Standards and
      Engineering  Division
    U.S. ENVIRONMENTAL PROTECTION AGENCY
    Office of Air  Quality Planning
      and Standards
    Research Triangle Park,  North Carolina 27711

    Dear Mr. Goodwin:

         Thank you for your  Notice of February 10,  1981,  concerning
    the NAPCTAC meeting  of March 18,  1981, concerning the Draft
    Guidelines for RACT  for  Control of Emissions from Volatile Organic
    Liquid Storage in Floating  and Fixed Roof Tanks.   It does not
    appear now that we will  be  able to attend the meeting, so we offer
    these written  comments which we ask be included in the review.

         First, since every  State now has as a part of their SIP a
    rule requiring the use of floating roofs or equivalent control for
    tanks storing  organic liquids with a vapor pressure of 1.5 psia
    or less, the guideline is misleading.  Model tanks should not be
    fixed roof tanks for this service since none exist under present
    regulation.  The model should be a single seal floating roof for
    all sizes.  The calculation of emission reduction in Appendix B
    is inappropriate since single seal floating roof reductions have
    already been made in compliance with existing SIPs.

         The number of turnovers used in the model calculations is
    far too high.  Page  2-12 states that "A storage tank at a chemical
    manufacturing  plant  usually has a higher annual turnover rate than
    a tank at a storage  terminal".   This is true.  At a storage
    terminal the average non petroleum tank turns over 4.5 times.  A
    tank described as a  small fixed roof tank which is probably truck
    or tank car supplied will turn over eight to ten times.  If a
    small tank at  a chemical manufacturing plant turns over an average
    of 50 times  (the highest I  have ever heard of is 86 times) and
    recognizing that there are  at least two distribution system tanks
    for every manufacturing  plant tank, the average number of turnovers
    used in the model is not representative of actual practice.
                                    V-35

-------
Mr. Don R. Goodwin            -2-                 March 4, 1981


     External floating roof tanks are also used as a model tank.
To my knowledge there are no external floating roof tanks used
for non-petroleum storage because the commodities stored are water
sensitive.  Advising States that emission reductions can be made
by this modification is misleading.

     An aluminum contact floating roof is used as the basis for
calculating costs for retrofit.  Aluminum is not an acceptable
material of construction for chlorindated organic liquids and some
other volatile organic commodities.  Likewise standard Buna N or
other inexpensive elastomer coated seal materials are not suitable
for use with ketones, chlorinated hydrocarbons and other high
solvent power organic liquids.  Hence the cost assumptions made
in the guideline are inaccurate; low .by a factor of two on the
average, we believe, since lined steel construction with teflon on
glass seal materials is required in many cases.

     Finally, no guideline recommendation is made that the State
inventory the community of affected tanks to determine the net
benefit that may accrue in non-attainment areas where the rule
would be applicable  (the Model Rule text in 6.0 should apply only
in non-attainment areas as specified on Page 1-1.  Without such
an inventory, a State has no basis for calculating the reduction
in emissions which might be attributable to such a rule and thus
no ability to determine how the costs involved would contribute
to reaching attainment.

     We believe the guideline document proposed has been drafted
as an extension of the floating roof seal efficiency studies
which are incomplete as evidenced by the comments on the earlier
documents based on the CBI test tank preliminary data.  Failure
by those who drafted the Guideline Document to understand the
community of tankage used for non petroleum storage, the materials
of construction limitation and the commercial use of such tankage
has led to an erroneous theory of the cost-benefit ratio, the net
emission reduction possible and the need for such a guideline.
We suggest the guideline be withdrawn for correction of the basic
assumptions.  We believe that when this is done, there will be
little justification for such a Guideline to be issued.

                                   Very .trolly ypurs,

                                      C^U^-^~—
                                     *  '         I
                                   R. W. Bogan, Vice President
                                   International and Environmental

RWB:sl
                                V-36

-------
                               C.  DISCUSSION

     Following the EPA presentation, Mr. Don Goodwin opened the floor to
questions and comments from the NAPCTAC members.   EPA staff and contractor
personnel were on hand to respond to questions and discuss issues of concern
to the NAPCTAC.members.  Industry representatives then made presentations,
each of which was followed by discussion.  For clarity, discussions are
grouped by subject matter rather than being placed in chronological sequence.

     Mr. Steiner and Mr. Nickolaus asked why the recordkeeping requirements
applied to volatile organic liquids with a vapor pressure of 1.0 psia or
greater when the recommended model regulation applied to VOL's with a vapor
pressure of 1.5 psia or greater.  Ms. Sommer responded that this action would
aid enforcement officials in determining whether a given tank was suppose to
have an internal floating roof.  Some of the tanks would be storing liquids
whose vapor pressure exceeded 1.5 psia for only a portion of the year.  The
extra records would identify these tanks throughout the entire year.
Mr. Porter then stated that the recordkeeping portion of the model  regulation
was being re-evaluated.  The decisions on recordkeeping requirements may be
left to the states.

     Mr. Davis and Mr. Nickolaus requested that the 5-year tank inspection be
changed to a 10-year inspection.  Mr. Davis said that the 5-year inspection
was based on a study of when tanks are emptied and degassed, not on seal life.
He stated that seal life is generally 10 years and that installing  a fixed
roof over an external floating roof tank would extend the seal life beyond
10 years by reducing the effects of the weather on the seal.  Mr. Goodwin
asked Mr. Davis to clarify the CMA's position on seal life.  They recommend
that fixed roofs for external floating roof tanks be deleted from the model
regulation and that the inspection time be increased because of the extended
seal life obtained through the installation of the fixed roof.  Mr. Davis
stated that a 90% emission reduction is achievable for large external floating
roof tanks by using seal technology alone and requested that the recommendation
of a permanently affixed roof be deleted from the model regulation.

     Mr. Steiner asked what the equipment lifetimes used in the annualized
costs are,  Ms. Sommer responded that the roof is annualized over 20 years
and the seals are annualized over 10 years.  Mr. Steiner asked if the
annualized cost presented in the CTG document included maintenance costs.
Ms. Sommer responded that 4% of the installed capital cost was included in
the annualized cost for maintenance purposes.  (It was later determined that
5% of the installed capital costs was used in the economic analyses for
maintenance costs).  Mr. Porter stated that the annualized cost also included
costs for. degassing, certifying, and inspecting a tank.
                                    V-37

-------
     Several questions were raised concerning the selection of reasonably
available control technology (RACT).  Mr. Lemke pointed out that the EPA
benzene test results for tests 12 and 27 suggest that the contact internal
floating roof tank is more sensitive to windspeeds than the external floating
roof tank.  He pointed out that an increase in windspeeds from 5 mph to
15 mph doubled the emissions from the contact internal floating roof tank
while the emissions from the external floating roof tank increased approximately
10%.  He felt these test results should lead to the selection of an external
floating roof tank as RACT instead of the contact internal floating roof
tank.  Mr. Moody stated that a comparison of just two test results was
misleading because the emission equations were developed through a statistical
analysis of a group of tests.  Mr. Moody further stated that the two floating
roofs had different types of seals.  A committee member pointed out that both
tests were conducted with no gaps in the seals.  Seals with no gaps is not a
real world occurrence; and as more gaps are introduced into the tests, the
difference in sensitivity between the two tank types decreases.  Mr. Porter
said that even though the emissions from a contact internal floating roof
tank double with the increase in windspeeds, the emissions at 15 mph are
still one-fifth of the emissions from an external floating roof tank.

     In their presentations, Mr. Davis and Mr. Nickolaus requested that the
publication of the CTG document be delayed until the testing API is currently
-sonducting on storage tanks is completed.  Mr. Blosser asked when the new
data would be available.  Mr. Porter said the testing should be completed
soon and  the data made available on April 1, 1981.  (API recently stated that
the new data will be available in May, 1981).  Once the data are available,
EPA will  re-evaluate the equipment specifications in the model regulation.

     Mr.  Davis pointed out in his presentation that the conclusions reached
from the  benzene test results apply to benzene and not necessarily all VOL's.
He noted  that the withdrawal loss is a direct function of a "clingage" factor
that would differ for different materials.  Mr. Reilly asked how dependent
the working losses are on the clingage factor.  Mr. Moody responded that the
working losses for a floating roof tank are about 5 percent of the total
emissions.  The total emissions are primarily dependent on seal losses, which
are being retested by API.

     Mr.  Nickolaus stated in his presentation that the EPA didn't consider
floating  roof configurations that have not yet been tested; in particular a
closed-tank floating roof outfitted with a flame arrester.  Mr. Moody responded
that this type of tank would be considered if Mr. Nickolaus could supply some
data on the VOC emissions from this type of tank.  Mr. Nickolaus stated that
it  is not the purpose of the TCC to supply emissions data to the EPA.

     Ms.  Chalupnik asked if there was a specific cutoff value for cost
effectiveness used in the selection of RACT.  Mr. Porter responded that the
cost effectiveness of reasonably available control technology varied for each
CTG published.  He said that generally speaking a value between $1,000-$2,000 per
                                      V-38

-------
megagram of VOC is considered reasonable.  Ms. Chalupnik asked what percentage
of the VOL tank population is attributable to small tanks.  Ms. Sommer responded
that there is a large population of small tanks in the synthetic organic
chemical manufacturing industry (SOCMI).

     Several questions about existing and proposed regulations for storage
tanks were raised.  Ms. Chalupnik asked what was required in the proposed
benzene NESHAP.  Mr. Porter responded that the equipment specifications were
the same but the capacity cutoff was lower.  Mr. Davis stated that the
capacity cutoff for benzene storage tanks is 1050 gallons.  Mr. Smith asked
if the RACT recommendations in the CTG document were more stringent than the
petroleum liquid NSPS.  Mr. Porter stated that RACT recommendations are more
stringent.  He pointed out that more data are available now than when the
standard was written for petroleum liquids and the more recent data were used
in the selection of RACT.  Mr. Reilly asked if the toxicity of VOL's was
considered when the more stringent RACT recommendations were developed.
Mr. Porter responded that the toxicity was not considered.  The recommendation
in the CTG is designed to eliminate ozone formation.

     Mr. Steiner asked what differences existed between the recommended model
regulation of the CTG document and existing regulations in Texas for VOL
storage tanks.  Mr. Nickolaus responded that the recordkeeping and inspection
requirements in the model regulation are more stringent.  Also Texas does not
require a secondary seal on internal floating roofs and does not specify that
the internal floating roofs be pan-type roofs.  Mr. Steiner asked if the
model regulation specified a pan-type roof.  Mr. Porter responded that the
recommendation in the model regulation was for a contact internal floating
roof for fixed-roof tanks, not a pan-type roof.

     Mr. Nickolaus requested in his presentation that the model regulation in
the CTG document be written so as to recommend as one control option an
internal floating roof without specifying a contact or a non-contact internal
floating roof.  This raised several questions from committee members about
the differences between a contact and a non-contact internal floating roof.
Mr. Reilly asked if the data really supported the contact-type roof over the
non-contact-type roof.  Mr. Moody responded the difference in emissions is
indeed significant.  Ms. Chalupnik asked how the cost effectiveness of the
two types of roofs compared.  Mr. Nickolaus responded that the cost effective-
ness between the two types of roofs is not that different.  He then re-emphasized
that the lower initial capital costs, the lower maintenance costs, and the
safety aspects of the non-contact type of internal floating roof were the
reasons why the TCC wants this type of roof to be considered reasonably
available control technology in the CTG document.
                                     V-39

-------
                           D.  CORRESPONDENCE

1.  Chemical Manufacturers Association
 Mr. Fred  Porter
 Emissions Standards and Engineering
   Division
 Environmental Protection Agency
 Research  Triangle Park, NC 27711

 Re:  Comments on Control Technique Guidelines Document
      Volatile Organic Liquid Storage

 Dear Mr.  Porter:

      The  Chemical Manufacturers Association submits this letter
 and the enclosed materials in response  to the Environmental
 Protection Agency's February 12,  1981 solicitation for public
 comment.   The subject comments include  our testimony from the
 March  17, 1981 National Air Pollution Control Technique Advi-
 sory Committee and additional specific  comments.

      The  Chemical Manufacturers Association (CMA) is a nonprofit
 trade  association made up of approximately 188 member companies
 in the United States representing more  than 90 percent of the
 domestic  production capacity for  basic  industrial chemicals.

      CMA  member companies have a  direct and critical interest
 in ensuring that EPA develop control technique guidelines, when
 a demonstrated need is presented, that  are scientifically and
 technically sound, reasonable, procedurally workable, cost effective,
 and  clearly authorized by the clean Air Act.

      Should the Agency require further  information or wish to
 discuss any of the issues raised  in our comments, you may contact
 me at  202/887-1179.

                                  Sincerely,
                                   Janet S.  Matey
                                   Manager,  Air Programs
  Enclosure

                                 V-40

           Formerly Manufacturing Chemists Association—Serving the Chemical Industry Since 1872.
        2501 M Street, NW • Washington, DC 20037 • Telephone 202/887-1100 • Telex 89617 (CMA WSH)

-------
SPECIFIC COMMENTS






          Section 2.2.1  (Types of Storage Tanks)


     The Agency makes no mention of the use of flame arresters or


      atmospheric blanketing in Section 2.2.1.  The Control Tech-


nique Guideline  (CTG) defines reasonable available control technol-


ogy (RACT)  as the ventilated internal floating roof storage tank.


This tank design is the predominant one used by the refining indus-


try.  Within the Synthetic Organic Chemicals Manufacturing Industry


(SOCMI), however, various companies use flame arrested or inert


blanketing designs with internal floating roofs.  This point was


made by the Texas Chemical Council (TCC) at the December 2-3, 1980


National Air Pollution Control Techniques Advisory Committee


(NAPCTAC) meeting.  At that time, the Agency stated that the test


work only applied to the neutral internal roof design and not to a


blanketed system.  We submit that the test work also does not apply


to a flame arrested system.  We recommend that EPA broaden the CTG


to include the inert blanketed and flame arrested systems.  The Agency


could apply zero or low "windage" factors to estimate the emissions


from the flame arrested configuration.


          Table 2-1 (page 2-13) and Table A-6 (page A-19)


     We cannot determine the source of the KS values reported in


Tables 2-1 and A-6 of the CTG.  A contact internal floating roof


with only primary .seals is listed in the CTG as having a Kg factor


of 26.7  (Table A-6).  Yet, when the benzene test work (Table 4-1,


page 6-8)  is examined for roofs with primary seals with and without


gaps,  the listed K  values are 12.2 and 13.6 respectively.  These
                  s
                                V- 41

-------
                              - 2 -






K  factors are a direct function of emissions.  We recommend  that
 s


EPA evaluate the CTG emission reduction calculations using  the



American Petroleum Institute's  (API) test work on internal  floating



roofs which will be available shortly.



          Section XX.050  (Recordkeeping)



     CMA recommends that EPA change the statement "These records



shall be kept for two years..." to read "Recordkeeping on a tank with



a continuous product type is not required after two years unless



the product type is changed."  A two year history should be suffi-



cient to determine the vapor pressure history for a given volatile



organic liquid  (VOL) in the same tank.  The vapor pressure  history



should not change after two years, thus additional data and record-



keeping are unnecessary.



          Section 5.0  (Control Cost Analysis of RACT)



     CMA has reconstructed several tables (Attachments A -  p)



to illustrate one member company's estimated costs compared to



EPA's costs.  The industry data presented is based on recent  indus-



try experience for the installation of internal stainless steel



floating roofs.  Comparisons were based on stainless steel  because it



is  a preferred material of construction for many uses within the



chemical industry.  Corrosion characteristics limit the use of alumi-



num to those instances where it is required, while stainless  steel



provides greater flexibility in the range of products for which



it is 'a suitable material.  These comparisons are also based  on



pontoon rather than contact type roofs.  EPA's cost figures are based



on pan type floating roofs.  These generally are not used in  our



industry because they offer no reserve buoyancy in the event  of a



leak and are unstable under operating upsets.
                                 V-42

-------
                              - 3 -


     There are six basic designs for surface contact roofs  as

listed below:

          1)  metallic pan roofs with a peripheral rim  (contact
              type)

          2)  metallic pan roofs with open top compartments  (con-
              tact type)

          3)  metallic pontoon roofs with closed top compartments
              (contact type)

          4)  metallic double deck roofs  (contact type)

          5)  metallic roofs on floats (noncontact type)

          6)  metallic and plastic sandwich panel roofs  (contact
              type)

     The sinking problems reported during verbal testimony at the

NAPCTAC meeting are experienced by roof types 1 and 2 above.  A

common engineering practice to minimize sinking tendency is to

require 100 percent excess buoyancy and that the excess buoyancy

be positively maintained (i.e. sealed and not open topped compart-

ments) .  This is possible for roof types 3 through 6 above, but

not for types 1 and 2.  However, there are other liquid surface

contact roof designs which satisfy our buoyancy requirements  (types

3, 4 and 6 above).  Thus the recommendation against the use of pan

type floating roofs (types 1 and 2) for new installation does not

preclude the use of liquid surface type contact internal floating

roofs.  We recommend that EPA state the costs for a range of options

to reflect designs in use by industry with different construction

materials.

          Section 6.0   (Model Regulation and Discussion)

     CMA agrees with the minimum capacity cutoff of 40,000  gallons

and vapor pressure greater than 1.5 psia for applicability.  We
                              V-43

-------
                              _ 4 -






also agree with the provision to exempt VOL storage tanks with an



internal floating roof.  However, we recommend that EPA include an



exemption for VOL storage tanks whose minimum vapor pressure only



exceeds 1.5 psia for 10 percent of the year.  This would exempt



tanks located in areas where the vapor pressure would exceed 1.5



psia only on a few high temperature days during the year.  We



believe this exemption is justified as cost-effective and cite the



following example:



     On page B-2 of the CTG, a 1.7 million gallon tank operating



at a vapor pressure of 1.4 psia all year will emit 46.15 Mg per



year.  A tank which operates at 1.4 psia for 90 percent of the year



and 1.9 psia for 10 percent of the year will emit 47.72 Mg per year



or an increase in emissions of 3.3 percent.  We believe this incre-



mental decrease in emissions is not cost-effective to implement RACT.



Again, we recommend that EPA include an exemption for VOL storage



tanks whose minimum vapor pressure only exceeds 1.5 psia for 10 per-



cent of the year.
                                 V-44

-------
                                                            ATTACHMENT A
                    COST OF IK'STALDKG AN INTERNAL

             FLOAT IHG ROOF IK AK EXISTING FIXED ROOF


                                    Installed Capital Cost &>
Tank Diameter               EPA - Aluminum               CKA Stainless-Steel
  (Keters)                   Contact Roof                   Pontoon Roof

    8                          $\0,700
    5.1                                                       $41,200
   12.2                                                       $$6,700
   13                          $17,600
   \k                          $20r200
   lit.6                                                       $66,500
   18.3                                                       $83,600
   2\                          $37,300
   27                          $56,300
d Costs do not  Include:  taking tank out of service, cleaning, degassing,
  and certHicat Jon.

* See Figure A
TABLE  1-2

         COST  OF  CLEAHfHC,  &EGASSWG,  CERTIFICATtOK OF A

             Tank  Capacity                 Cost  (2Q80) £stl«aat«
               (Liters)                     EPA        " CRft *
               602,100
             2,000,000                     2300
             2,610,000                     2875
             k, 160, 000                     3310
             8,590,000                     7070
      s estimate based on ? tanks, ranging in capacity  from 800,000
  to 3*201,000 liters,  in this ran$e, cost is not a function  of
  capacity.
                                     V-45

-------
m - r "i~T\ rm 1 " "«" ' fT f'H ff " 7 • n •
i Ml *
it i ! ; .
fe' ,,4 i {
•h ? ' » A , , . . .! - 1 . « J. .
•M I 1 !
»?"")- 1 — i 	 j i 1 	 1 \y-
i 	 /-
ii i j f i
4-1 ! I' 1
. V J ! ! 1 - . . *< £ ! ! * .1 • J
i, >•
1 • y \
\\ i * ti.t* / »
flS:5ljni) 	 	 1- . . / J . . . . ...
>*lif t 1 j , 1 X (
TT ' ' t w J 	 i. .... .1 	 	 ......
Ji r
S i
K ....!.. i I
p ! V
V. v\ > 3
•-•*•' 	 • ^ • • 	 • • }
: i I--
1:^4 • ' 	 - i 	 I'*- \
I i +••
. : , ..•••'
' \ ll*" '
. > . -«-• r i
-/«<« 	 f- -J f" 	 ' 	
1
i If. 1 I \
:' 2 I 1 •
( ii i '
i i'li'i i '
to • • { • I
i f j iir t | i I
! 4 f 1 ,1. ., r ,
1! i i [' { ' '"
iii iii i * i
> 1. . .......] U. .. ... .1 .1 . .. '*. . . . ^ • l ' •
r r • T » rn if r :
Hi
I/ i ] 111 4
/r i i i '
i> i t i i i
v; ' * ! ! !
/ 1 | ill 1
1 i i !
i i ( ' f H
1 ( I ) 4 •
M l 1 1 1 1 1
! ! 	 1 ; i J
tt :
! !
i iii
• ii »
i i i
i \ \ i »
Jtt: . : 3 : . ! - I !
;i
! ' ^1 '
! 1
r 1
>^ j
• ,-''' \
\,<'' \
	 . . . -.- . *. - - • • • 	 - .... . M l
{••''' \ ' 1
>•"' 	 !
,--'"' ito l\\.,cM^:[.^.'4. \
' ' » (
/ j kit
v i 1 1
• i
1 j J
it. . ! . . i
: : i
! !
• iiii
* i ' i ' *
\ li ^ .-i rll: stti '
1 ' ' 1 1 1
i ; )
i ' i i S
-_.- --i- ' 1 n * r- . U.I . . L J., J , ^ L f Ij '
f l> ' t:;' ^ f ^
.. 9 , .! . | I
••H 1 1 . if .1

-------
                                     INSTALLED CAPITAL COST5
Model Storage Tank
Cost Item
Instil led Floating
Roof
Installed Secondary
Seals
Cleaning. Degassing,
Smal 1 Tank3
EPA - Aluminum CHA - SS
Contact Roof Pontoon Roof
$ 6,120 $21,500
M10 3,3)0
1,110 1M5
Average Fl
EPA - Aluminum
Contact Roof
$10.000
'<,530
1.3*0
xcd Roof
CMA - SS
Pontoon Roof
$35,500
4,530
2,425
    Certification
       TOTAL
12,5*40
27,235
16,670
42t<»55
^Installed capital costs for a small tank are based on retrofitting o fixed roof tank
 with a diameter of 5 meters (17 feet).
i^
 Installed capital costs are based on retrofitting a fixed roof tank wKh a diameter of
 8 meters (26 ft).
                                                                       If

                                                                       o

-------
                                                                 ATTACHMENT
                           AKKUAU2ED  CONTROL COSTS FOR.

                                HDDEL  STORAGE TAKKS
                                                Mode?  Storage Tank
Cost  Parameter
INSTALLED CAPITAL COSTS

AKKUAUZED COSTS

 .  ANNUAL CAPITAL CHARGES

    Cap Us V recovery
    Taxes,
    and administration

    Subtotal
  . &t*ECT  OPERATtKG COSTS

    Maintenance

     inspection

    Subtotal


 TOTAL  ANNUAL UED  COST

 AKKUAM. RECOVERY CREDIT

 KET ANNUAL 1 2 ED COST
                                     Small
                             Contact Ro&f   Pontoon Roof
$12.51*0
    502
    C.27

    ns
    752


   2,528

    (WO)

   2,288
527.255
    372
    (Wo)

   5,5^3
                              Average  Fixed Roof Tank
                               EPA-At.CKA-SS~"~
                             Contact Roof   Pontoon Roof
$16,670
                                2,226
                   66?

                 2,893
    167
                 1,001
 (2,191)

  1,703
                                              7,002
                               2,123
 9,550

(2J9U

 7,359
 8Annua)Ized control  costs  for a  small  tank are based on retrofitting a fixed  roof
  tank with a capacity of 151,^16 liters and a diameter of 5 meters.


  Annuall?ed control  costs  are based  on retrofitting a tank with a capacity  of
  ^60,747 liters and  a diameter of 8  meters.
                                       V-48

-------
                                                                  ATTACHMENT fi:
                           COST EFFECTIVENESS FOR KODEL

                            STORAGE TANKS UNDER RACT
Cost Parameter
                                                Model  Storaoe Tank
                                      Small Tank*
Total Annual!zed Cost
	    Average Fixed  Roof Tafek"
  EPA-A1.CKA-SS        EPA-A1.          CKA-SS
Contact Roof   Pontoon Roof   Contact  Rocf    Pontoon ftoof
   2928
389**
                                            9550
Totsl Annual
 Recovery Credit
Set Annoalired Cost
(6*0)


2286
                                 1703
               7555
Total VOl Redact'ion
Cost Effectiveness
  (Annual  $/Hg VOL)
    H79
 256
                                             11D8
 The cost  effectiveness  for  a  small  tank is  based on a  fixed roof tank with a
 capacity  of  l£l»M6  Uters  (40,000  gallons)  and 3  diameter of 5 raters- ()? feet)


 The cost  effectiveness  of this  tank !s  based on a  capacity of ABO,7**? ^
  (I27»000  gallons)  and a dlsraier  of 8 meters (2&  feet).
                                      V-49

-------
                                                       ATTACHMENT
1.   All costs are in second quarter 1980 dollars.





2.   The EPA data as well as the basis for estimating annualized



    cost/cost-effectiveness are from Control of Volatile Organic



    Liquid Storage in. Floating and Fixed Roof Tanks (Preliminary



    Draft) , January 1981.





3.   CMA estimates for cleaning, degassing and certification include



    costs for labor and location overhead.  All other CMA estimates



    are contract costs based on lowest vendor quotes received and



    do not include costs of process engineering, overhead, etc.
                               V-50

-------
2.  American Petroleum Institute
  March 18, 1981

                                                    Re:  CELM

  Mr. Fred Porter
  Emissions Standards Engineering Div. (MD-13)
  Environmental Protection Agency
  Research Triangle Park, NC 27711

  Dear Mr. Porter:

  The American Petroleum Institute (API)  is pleased to offer
  comments to the Environmental Protection Agency (EPA) and the
  National Air Pollution Control Techniques Advisory Committee
  on the EPA draft document "Control of Volatile Organic Emissions
  from Volatile Organic Liquid Storage in Floating and Fixed
  Roof Tanks" (January 1981).

  API is a voluntary, not-for-profit trade association registered
  in the District of Columbia, composed of approximately 270
  company and about 7,000 individual members and representing
  virtually all facets of the petroleum industry.

  API's comments are directed to those portions of the proposed
  guideline which deal with the calculations of emissions from
  tankage and proposed emission control techniques.   Specifically,
  API wishes to make three points:  (1) The proposed CTG is
  premature in light of insufficient data for evaluating alternative
  regulatory procedures; (2) There is an API test program underway
  which addresses the insufficiency of data; and (3) The equivalency
  provisions of the guidelines are too limited in scope and are
  inconsistent.   These points are reviewed in detail below.

  (1)  The proposed CTG is premature in view of the insufficiency
  of test data.

  API has commented to EPA several times in the past year regarding
  its deep concern about EPA's attempt to extrapolate emission
  levels from limited testing with only benzene to all organic
  liquids, and to suggest regulatory alternatives which are
  restricted to only those systems which were included in the
  limited test program.  These concerns are included in the
  attachment (J. K. Walters' December 4,  1980 letter to D. R.
  Goodwin) and will not be repeated again.  Please note that
  paragraph three of the December 4, 1980 letter lists API's
  previous submittals on this subject.
                                   V-51

-------
Mr. Fred Porter                                          p. 2


(2)  There is an API test program underway which addresses
the data insufficiency.

As EPA knows, API's current test program on internal floating
roof tanks will provide a broad and more thorough data base
on pure compounds other than benzene which will be relevant
to this subject.  Further, the API program has encountered
and successfully addressed technical problems that were  encountered
in the previously mentioned EPA test program, with the result
that API's data should offer a more valid and defensible
basis for calculating emissions and evaluating regulatory
alternatives.  API expects its testing to be completed by May
1981, with the results available as soon as possible thereafter.

(3)  The equivalency provisions in the CTG are too limited in
scope and internally inconsistent.

Paragraph 6.2.4, lines 1-8, Equivalency, specifically excludes
the use of floating roof control equipment other than that
specified in the paragraph ("contact internal floating roof
with a liquid-mounted on metallic-shoe primary seal and  a
continuous secondary seal").   This requirement appears to
arise out of the definition of "alternative control technology"
found on page 6-2 which excludes floating roof equipment.

The result is to unfairly limit an owner's option in selecting
alternative equivalent control technology which may meet the
90 percent emission reduction criteria.  Since this technology
could include floating roof equipment other than that noted
above (i.e., contact type floating roofs), the definition of
alternative control technology should be expanded to allow
for the use of other types of floating roof control equipment.

In addition to these three main points, API expresses concern
that Section 6.3 "References for Chapter 6" (page 6-10), the
basis for 6.25 "Compliance Schedule," is a series of telephone
conversations which includes only tank manufacturers.  API's
concern is that there may be a vast difference between the
information provided and actual site-specific installation
times.  API asks that memos-to-file of these calls be provided
for review by all parties and made a part of the administrative
record.

API hopes that these comments are helpful.  API has kept and
will continue to keep EPA informed of the progress of its
test program.  API believes that the results of its program
must be considered before this draft CTG is published.   There-
fore, API requests that the comment period on this CTG be
kept open until API's test results are available.  API will
be happy to answer any questions.
V/
                                V-52

-------
                                                    ATTACHMENT
  American Petroleum Institute
  2101 L Street, Northwest
  Washington, D.C. 20037
  202-457-7055
  J. K. Walters
  f.'eas-'Sf-.ent Coordinator

  December 4, 1980
                                       RE.-2517

 Mr. D. R. Goodwin
 Director
 Emissions Standards and Engineering
   Division
 U.S. Environmental Protection Agency
 Office of Air Quality Planning  & Standards
 Research Triangle Park, NC  27711

 Dear Mr. Goodwin:

 API is pleased to offer comments to the EPA  and the National
 Air Pollution Control Techniques Advisory Committee on the EPA
 draft document "Standards of Performance for New Stationary
 Sources — Volatile Organic Liquid Storage Vessels" and the
 associated Background Information Document.

 API is a voluntary, not-for-profit trade association registered
 in the District of Columbia/ composed of approximately 350
 company and 8,000 individual .members and representing virtually
 all facets of the petroleum industry.

 API had previously commented on an earlier partial draft of this
 background document (J.K.  Walters to J.R.  Farmer,  June 30, 1980),
 as well as on other related documents on benzene storage which
 were based on the same technical data base (J.G. Zabaga for API
 at April  16-17,  1980,  NAPCTAC meeting; G.T. Patton to J.R.
 Farmer, November 29,  1979; and J.K.  Walters to R.K. Burr, Sep- ' •
 tember  14,  1979).

 In these previous  submittals,  API expressed concern about the
 sufficiency  and  adequacy of emissions data being used by EPA to
 evaluate various  regulatory alternatives.   These previous com-
ments  (Attachments  A,  B, C and  D)  are applicable to this draft
NSPS and are  attached  for  your  further consideration.   Therefore,
they will not  be redeveloped in  detail here,  although they are
summarized in  the following  comments.

This draft NSPS for VOL storage  is also based on the same in-
sufficient data and, therefore,  it does not  allow  for a valid
evaluation of all regulatory alternatives  for all  VOLs.  API
                             V-53

-------
 Mr. D. R. Goodwin
 believes that in view of the insufficiency of this data base and
 the fact that API is currently engaged in a test program to
 produce new data, the draft NSPS is premature.

 API's primary concern with the draft NSPS is that the emissions,
 data used to evaluate the benefits of the various regulatory
 alternatives may not be representative of field conditions
 because of temperature effects peculiar to the pilot-tank in
 which the tests were conducted.  Further, the data base is
 insufficient to properly distinguish among all possible regu-
 latory alternatives.  Specifically, the data do not allow for
 a clear understanding of the contribution of emissions from
 different types of roofs, relative to the emission contribution
 from different seal types.   Therefore, the regulatory alterna-
 tives only address the specific roof/seal combinations xtfhich were
 tested,  as opposed to other possible and  potentially equivalent
 roof/seal combinations.

 As EPA is aware,  API is  currently conducting pilot tank .tests
 which are expected to significantly expand our understanding of
 the important parameters affecting VOL storage emissions,  rela-
 tive to  both equipment and  ambient factors.   Specifically,  the
 program  includes  the study  of:   (1)  3  different  internal  floating
 roofs, both contact and  non-contact types,  and both  liquid--
 mounted  and vapor-mounted seal  types;  (2)  pure components, as
 well as  multi-component  liquids;  and (3)  secondary seals,  as well
 as only  primary seals.   It  is expected  that  this  study  will
 provide  a better  understanding  of  the  separate effects  of  roof
 types and seal  types on  emissions.   The results of  this work
 should permit  an  adequate evaluation of various regulatory
 alternatives.

 Test work on  the  first roof to  be  tested  is now underway.  The
 test program  is expected  to be  completed  by the summer  of  1981.
 During this  time  the API  will continue its current practice  of
 inviting  EPA attendance at our  meetings to review  interim results
 of the test program.

 It  is API's position that no one type of  equipment should be
 specified to the  exclusion of other viable options, as  is done  in
 this  current draft NSPS,  unless conclusive data•are available to
 support any claim of significantly lower emissions potential
 relative  to other equipment types.  Also,  operational aspects of
 different applications may warrant selection of a specific type
 of  equipment, other than  the one-chosen regulatory alternative,
 in  order  to achieve safe and reliable operations.  This opera-
 tional flexibility should not be precluded by overly restrictive
 equipment specifications.

For example, some companies have strong reservations about the
use of steel contact roofs of the type tested due to a belief


                             V-54

-------
  Mr. D. R. Goodwin
  that they have  an  increased potential  to  sink relative  to non-
  contact roofs.  Other types of  internal floating  roofs  have other
  safety and operational restrictions  in some companies,  such as
  product temperature limitations and  problems associated with.
  gas-freeing tanks due to vapor permeation into the roof.  Dif-
  ferences in companies' safety and operational assessments should
  be respected in the absence of compelling data to support the
  use of only one type of equipment.

  The draft NSPS, although requiring only one type of roof and
  seal system, does allow for a demonstration of "equivalence^"
  However/  it is much more cost-effective arid efficie'nt^fb'r^
  both industry and EPA to also permit a 1*1 now known equivalent
  equipment alternatives specifically in the regulation,  as
  opposed  to individual testing  and  submission of  equivalency
  requests,  with EPA  being  required  to review  these  requests
  on a case-by-case basis.

  API  requests that the term  volatile  organic  liquid be defined in
  this document  so that  the reader can  distinguish petroleum
  liquids defined  herein from those  defined  in other "VOL  related
  documents."  For example, VOLs as  defined  in the-existing NSPS
  for  tankage  (Federal Register, Vol. 45, (Friday, April 4, 1980)
  p. 23379) apparently are not the same VOLs defined  in this
  draft NSPS.

  In summary, API  submits that the draft^KSPS_^pr._Vp_L_storage is
 based on an incomplete data base which does notTallow for a valid
 evaluation of all relevant reglatory .alternatives.   In recogni-
 tion of API's current test program, the draft NSPS  is premature.
 If the regulatory process must proceed at this time despite these
 concerns,  this draft should qualify the emission estimates and
 the resultant evaluation of the various regulatory  alternatives
 as being  preliminary and clearly provide for a re-evaluation  by
 EPA as additional data become  available.   API will  be pleased to
 cooperate  with EPA in such a re-evaluation as our test program
 progresses.

 API will be  happy to meet  with  the  EPA to discuss our comments or
 to  answer  any questions you  may have.

 Very  truly yours,
 / K. Walters

JKWrracm
                             V-55

-------
        3.  ITh'rvoi-s environmental^ Protection Agency
ILLINOIS
Environmental   Protection  Agenq
2200  Churchill  Road, Springfield, Illinois 62706
        217/782-2113
        March 13,  1981
         National Air Pollution Control
          Techniques Advisory Committee
         U.S. Environmental Protection Agency
         Office of Air Quality Planning and Standards
         Research Triangle Park, North Carolina   27711

         Gentlemen:

         For your information and record,  the Illinois Environmental Protection
         Agency submits the following comments:

         Attachment 1 —  Corments on Preliminary Draft "Control of Volatile
                        Organic Emissions from Petroleum Dry Cleaners";
          -""*"—*~~	
         Attachment T~::=^Comments on .Preliminary Draft "Control of Volatile
                   •—-"^Organic Emissions from Volatile Organic Liquid Storage in
                        Floating and Fixed Roof Tanks";

         Attachment 3 —  Comments on Preliminary Draft "Control of Volatile
                        Organic Fugitive Emissions from Synthetic Organic
                        Chemical,  Polymer  and Resin  Manufacturing Equipment".

         Your  consideration of these comments is most  appreciated.

         Sincerely yours,
                      \A     f)
                   L ^L2-*-A/
                      W     S~*''\
         John  C. Reed,  Ph.D., P.E.
         Supervisor, Technical Support Unit
         Air Quality Planning Section
         Division of Air  Pollution  Control

         JCR:jab/2852H/24
                                             V-56

-------
                                Attachment 2

COMMENTS ON PRELIMINARY DRAFT "CONTROL OF VOLATILE ORGANIC EMISSIONS FROM
VOLATILE ORGANIC LIQUID STORAGE IN FLOATING AND FIXED ROOF TANKS.

1.  Page 1-1, paragraph 2, Section 172(a) and (b)(B) of the-Clean Air Act
    requires "... all reasonably available measures ... "not" ...
    reasonably available control technology (RACT) ... ."

2.  Page 2-8, paragraph 2.2.3.2, are references 7 and 8 reversed?

3.  Page 5-11, paragraph 5.2.3, the VOL value should also be stated in
    familiar english units (e.g., Ib/gal) in parenthesis.

4.  Page 6-2, paragraph 5, better definition of "Metallic shoe" needed,
    especially the term "... but not limited ..." should be elaborated.
    What other types are there?

5.  Page 6-3, article XX.030(A)(1)(C), what does "... no visible gap
    ... ." mean?  How can this be enforced?

6.  Page 6-4, article XX.030(B)(3) what does "... no visible gap ... ."
    mean?

7.  Page 6-5, article XX.050(A)(2) change "...  the lesser ... ." to "...
    the greater  ... ." since the methods given in XX.050(A)(2)(a), (b) are
    approximations and the time vapor pressure could be greater than the
    value given.

8.  Page 6-6, article XX.060, the compliance schedule only allows 14
    months to achieve compliance.  This should be more completely
    documented and explained.

9.  Page 6-8, paragraph 3, using the actual vapor pressure at the highest
    anticipated average monthly temperature will result in a lower overall
    control efficiency than using the yearly average of the monthly
    temperature.

10. Page A-21, paragraph A.3.1.2 and page A-22, page A-23, Table A-7,-the
    variance of the data gives a 95% confidence interval of the mean *s
    .15 to .42.  Therefore the factor between calculated and actual may be
    only approximately two rather than four.

11. Page A-24, paragraph A.3.2.2 and page A-25, Table A-8, the variance of
    the data gives a 9556 confidence interval of the mean breathing loss to
    be .05 to .36.  Therefore the calculated/actual factor may be three
    rather than two.
JCR:ct/2760H,47
                                      V-57

-------
   .4.  NAPCTAC Member  Hi Hi am Reiter

       ~*AJBed.  ,
         Chemical
         Corporate Environmental Affairs
         P.O.Box2332R
         Morristown, New Jersey 07960                   March  23  1981
    Mr.  Don Goodwin
    Director, Emission Standards & Engineering  Division
    Office of Air Quality  Planning & Standards
    U.S.  Environmental Protection Agency
    Research Triangle Park,  NC  27711

    EDITOR'S NOTE:

                NAPCTAC member William Reiter was unable to attend  the
    meeting on March 18, 1981,  So that he could contribute his views to EPA
    and fellow Committee members, Mr,  Reiter wrote a lengthy letter to the
    chairman. The contents  of that letter have been divided by subject and
    are included in the relevant sections of the minutes,  The portion of the
    letter that applies to this section follows,


3)  CTG -  Control of Volatile Organic Emissions from Volatile
    Liquid  Storage Tanks	M	

    a)    Referring to Page 1-2 of  the draft CTG,  I raise the
          question of why EPA, in its model regulation,  cites  with
          firmness that the affected  facility is to be  storage
          tanks  with a capacity £  40,000 gallons.  The  State  is
          responsible for its  SIP and for meeting  the air quality
          deadlines set in the Act.   I believe the  States,  there-
          fore,  should be responsible for setting  the applicable
          tank  size.  In California,  it  may be advisable to set
          the level at 20,000  gallons; however,  in  Arkansas,  it
          may be feasible to  set  the  applicable tank size at
          80,000 gallons.  This  should be left to  the States'
          discretion.

    b)    In reviewing the draft  document,  I find  no Justification
          for the specification  of  40,000 gallons  as the minimum
          size  for control.

    c)    In the recent past,  NAPCTAC reviewed the  NSPS  proposed
          for VOL storage tanks.   Considerable comments  were  given
          during that meeting  covering a proposed  API study,
          sinking of various  types  of roofs, safety of  inspection,
          etc.   It appears that  none  of  those comments  were con-
          sidered in developing  the CTG.  I believe NAPCTAC con-
          sidered the NSPS in  early December.  The  draft document
          appears to have  been printed  in January.  I find  that  to
          be poor administration  on EPA's part.

                                  V-58

-------
Mr. Don Goodwin
March 23, 1981
Page 4 -


    d)   Has EPA considered the API test  results?   If this has
         not been done, I believe  technically  we are  in error to
         proceed without that consideration.

    e)   It appears that EPA has not  provided  data  on non-contact
         floating roofs.  Does EPA have data?   Could  you please
         advise me.

    f)   During the December NAPCTAC  meeting,  a suggestion was
         raised that there should  be  an exemption where the vapor
         pressure of the VOL exceeds  the  limit  only during the summer
         months.  The suggestion,  I believe, was made that the
         exemption should exist where the  vapor was exceeded for
         less than 10? of the year.   I believe  such an exemption
         is merited and should have been  included.

    g)   Refer Page 1-3, the third paragraph.   Recordkeeping
         appears to be quite extensive and I believe  the EPA owes
         the public an explanation of why  the  records are
         required.  Are the States going  to use these data;  what
         is the benefit of this recordkeeping?  Normally,  a plant
         keeps a running inventory.   Why  specify additional
         requirements?  If enforcement is  required,  plant inven-
         tory records which identify  storage conditions and the
         material stored can be examined.  I reiterate my earlier
         suggestion that the only  submission to be  made is a
         certification by the plant manager that the  plant is in
         compliance.  (I made this comment during the NSPS review.)

         Page 1-4 - statement is made that "alternate control devices
         must reduce emissions by  at  least 90?".  The compliance
         level should be formulated by the State, governed by
         localized conditions and  not dictated  by Durham lacking
         evaluation of localized conditions.

         Please refer to Table 2-4.   In examining the data pre-
         sented in that Table, there  does  not  appear  to be any
         change in the rate of emission with tank size.  The
         small tank emission rate  calculates to 0.57  MG/yr/10,000
         gallons.  The same factor is obtained  for  the larger
         tank size (127,000 gallons).
                                V-59

-------
Mr. Don Goodwin
March 23,  1981
Page 5 -

         Equation  2-1  indicates  the  tank  diameter  has  an effect
         with  other  factors  being  kept  constant.   I  would appre-
         ciate  some  explanation,  especially  in  light of the state-
         ment  on Page  3-1.

         Last  paragraph,  which  states,  "the  VOC  emissions from
         storage tanks increase  with increasing  tank
         capacity....".   There  appears  to  be an  inconsistency.

    h)   On  Page 4-1,  RACT  is  identified  as  a contact  internal
         floating  roof with  secondary seals  and  a  fixed roof
         tank,  and for an external floating  roof  tank,  RACT is
         said  to be  secondary  seals  and a  fixed  roof.   There does
         not appear  to be any  justification  or  evaluation of the
         basis  for establishing  these requirements as  RACT.   The
         potential problems  and  limitations  as  stated  in the
         December  NAPCTAC meeting  on the  NSPS apparently have not
         been  considered.  For  instance,  it  is  believed that the
         application of secondary  seals for  an  external floating
         roof  tank is  sufficient.   What is the  purpose  of
         installing  a  fixed  roof over that system, especially
         since  EPA does not  appear to have obtained  data to
         establish the loss  rate?

         Again,  the  prohibition  of non-liquid surface  contact
         roof  designs  has not  been supported with  test  data.  I
         urge  you  to do considerable work  on this  document  before
         proceeding.

     i)   Refer  to  Table 5-9  on  Page  5-16.  A comparison of  the
         cost  effectiveness  and  its  variation with tank size has
         not been  clearly identified.  Crudely  calculating  that
         value  and its variation with tank volume, it  appears
         that  the  breakpoint for cost effectiveness  should  be in
         the range of  100,000  to 127,000  gallons.  I suggest you
         review the  selection  of the cut-off size.

                                       Best regards


                                       <&£&>
                                       W.  M.  Reiter
                                       Corporate Director
                                       Pollution Control

cc:  NAPCTAC Members
                                V-60

-------
                  VI,  CONTROL TECHNIQUES GUIDELINE DOCUMENT
                  FOR CONTROL OF FUGITIVE VOC EMISSIONS FROM
                  SYNTHETIC ORGANIC CHEMICAL AND POLYMER AND
                         RESIN MANUFACTURING EQUIPMENT

                            A.  EPA PRESENTATION

                            Mr.  Samuel  Duletsky
                          GCA/Technology Division
                             500 Eastowne Drive
                     Chapel Hill, North  Carolina  27514
Introduction
     This presentation discusses the results of the development of the control
techniques guideline (CTG) document for fugitive emissions  of volatile
organic compounds (VOC) in the synthetic organic chemical manufacturing
industry (SOCMI) and in the polymer and resin manufacturing industry.

     This presentation will consist of an overview of the source category,  a
discussion of the reasons why the source category was selected for development
of a CTG, a brief outline of the reasonably available control technology
(RACT) for the source category, and a discussion outlining  the major decisions
made in selecting RACT.  (Figure 1)

OVERVIEW OF SOURCE CATEGORY

     The source category for this CTG consists of fugitive  emission sources
in process units in the synthetic organic chemical manufacturing industry and
in the polymer and resin manufacturing industry.

     For this category, the SOCMI is defined as process units that produce
any of 378 organic chemical compounds that are produced as  intermediates or
final products.  These 378 chemicals are generally manufactured from
ten petroleum-derived feedstocks.  The 378 chemicals are generally used as
feedstocks for manufacturing of plastics, fibers, surfactants, Pharmaceuticals,
elastomers, dyes, pesticides, and specialty organics.  However, many of these
chemicals are used as final products.  (Figure 2)  A list  of these 378 chemicals
appears in Appendix B of the CTG document.

     The polymer and resin manufacturing industry is defined as process units
which produce any of sixteen polymers and resins as intermediates or final
products.  As end products, these polymers and resins are  plastics, fibers,
and elastomers.  (Figure 3)  A list of these polymers and  resins appears in
Appendix B of the CTG document.
                                    VI-1

-------
             FIGURE 1

  OUTLINE OF PRESENTATION

 ». OVERVIEW OF SOURCE CATEGORY

 II. SELECTION OF SOURCE CATEGORY

III. DISCUSSION OF RACT

IV. SELECTION OF RACT
               FIGURE 2

THE SYNTHETIC ORGANIC CHEMICAL
    MANUFACTURING INDUSTRY
      REFINERIES, NATURAL GAS
          PROCESSING, ETC.
            FEEDSTOCKS
         FINAL PRODUCTS OR
           INTERMEDIATES
                VI-2

-------
     Fugitive emissions can be described as leaks of process fluid, either
liquid or gas, from equipment components in chemical processing units.  Leaks
may be the result of the effects of age, lack of maintenance, or externally
caused damage.  For this project, the fugitive emission sources in process
units which were considered are:  pumps, compressors, valves, open-ended
lines, and pressure relief valves.  (Figure 4)

     Both the SOCMI and polymer and resin manufacturing industries have the
same types of components in process units.  Therefore, fugitive emissions
occur from leaks in the same types of equipment in process units in both
industries.

SELECTION OF CATEGORY

     The SOCMI was listed number 1 of 59 industries on the final priority
list for new source performance standards published in the Federal Register
on August 21, 1979.  Fugitive emissions were specifically included as part of
this listing.  (Figure 5)  The VOC emissions from existing SOCMI plants are
estimated to be one million megagrams per year.  Fugitive emissions of VOC in
the industry contribute approximately 400,000 megagrams per year,  or
forty percent of the total.  These emissions can be controlled by the
implementation of reasonably available control technology.  The implementation
of RACT will significantly reduce fugitive emissions.

     The list of areas requesting an extension beyond 1982 for compliance
with the National Ambient Air Quality Standard for ozone includes  areas where
SOCMI and polymer and resin process units are located.  These areas contain
many process units.  Consequently, the application of RACT to fugitive
emissions sources in these areas will result in the reduction of VOC emissions
and contribute toward attainment with the NAAQS for ozone.

DISCUSSION OF RACT

     Leak detection methods provide a means to identify sources that are
leaking significant amounts of VOC.  The most rigorous and universally
applicable leak detection method is the individual component survey using a
portable VOC detector to check for VOC leakage at each source.  The detector
probe is placed at each potential leak area and the maximum VOC concentration
is noted.

     The RACT selected for control of fugitive VOC emissions in SOCMI and
polymer and resin process units is a leak detection and repair program.  A
model regulation which incorporates the recommendations of RACT has been
included in the CTG document.  The model regulation recommends that certain
components in contact with process fluid containing at least ten percent VOC
by weight should be monitored with a VOC detection instrument once every
three months.  These components are:  pumps in light liquid service,
compressors, valves in light liquid service, valves in gas service, and
pressure relief valves in gas service.  (Figure 6)  Pumps in light liquid
service should be checked visually each week for indications of leaks.
Open-ended lines should be capped with a second valve, a cap, a blind flange,
or a plug.
                                    VI-3

-------
          FIGURE 3
 THE POLYMER AND RESIN
MANUFACTURING INDUSTRY

        FEEDSTOCKS
             \
    POLYMER AND RESIN
      MANUFACTURING
             T
          SIXTEEN
         POLYMERS
         	L
           FINAL
         PRODUCTS
           FIGURE 4
    POTENTIAL FUGITIVE
     EMISSION SOURCES
   • PUMPS
   •COMPRESSORS
   •VALVES
   • OPEN-ENDED LINES
   •PRESSURE RELIEF VALVES
           VI-4.

-------
                 FIGURE 5

 NATIONAL SO CM I VOC EMISSIONS


•TOTAL VOC EMISSIONS =
    1,000,000 Mg/yr

•FUGITIVE VOC EMISSIONS =
    400,000 Mg/yr

• FUGITIVE VOC EMISSIONS-
    40% OF TOTAL
                 FIGURE 6

     REASONABLY AVAILABLE
  CONTROL TECHNOLOGY (RACT)

  I. Quarterly Monitoring of:
    • Pumps in Light Liquid Service
    • Compressors
    • Valves in Light Liquid Service
    • Valves In Gas Service
    • Pressure Relief Valves In Gas Service

  II. Weekly Visual Inspection of Pumps In
    Light Liquid Service
 III. Caps on ail Open-Ended Lines

 IV. Leaking Components:
    • Measured VOC Concentration > 10,000 PPMV
    • Should be Repaired Within 15 Days or
     at Next Unit Turnaround
    • Tag Affixed Until Repaired
                  VI-5

-------
     Components which have VOC concentrations at or above 10,000 ppmv would
be considered leaking components.  Leaking components should be repaired
within fifteen days of the date of detection.  If a leaking component cannot
be repaired within fifteen days of the date the leak is detected, repair may
be delayed until the next unit turnaround.  A tag should be affixed to a
component when a leak is detected and should remain in place until the leak
is repaired.

SELECTION OF RACT

     A leak detection and repair program was selected as RACT for several
reasons.  Leak detection and repair has been proven to find leaks, result  in
the repair of leaks, and effectively reduce emissions from leaking components.
EPA has confirmed from data collected in SOCMI process units that directed
maintenance will result in leaks of VOC being found and repaired.

     Implementation of RACT is shown to result in a net savings for all
levels of model process unit complexity (Figure 7).  The value of VOC saved
annually as the result of repairing leaks in SOCMI process units is greater,
than the annual cost of detecting and repairing leaks.  This results in a
net annual savings for the implementation of RACT.

     Recommending a leak detection'and repairing program is consistent with
control technology published in other EPA documents.

     The components selected for control by RACT are the component types
within process units that contribute most to the total of fugitive emissions
from process units.  These components as a group are responsible for about
ninety percent of the emissions from model process units.  (Figure 8)  By
selecting these components for control by RACT, a large portion of the
fugitive emissions can be affected without having to monitor large numbers
of components such as flanges which leak very little or not at all.

     A quarterly monitoring interval was chosen for this CTG document for
the following reasons.  Quarterly monitoring of the components affected by
RACT would result in an estimated emission reduction of about 65 percent in
model units.  By comparison, a monthly monitoring interval would result in
an estimated emission reduction of about 70 percent.  Quarterly monitoring
will achieve an emission reduction nearly as great as monthly monitoring but
would cost about one-third as much as monthly monitoring.

     A quarterly monitoring interval is consistent with guidance given in  a
previously issued CTG document.  The CTG for control of VOC leaks from
equipment in petroleum refineries recommends a quarterly monitoring interval
for compressor seals, valves in gas service, and pressure relief valves in
gas service.
                                     VI-6

-------
                    FIGURE 7
 MET COST OF RACT FOR MODEL UNITS
                                 MODEL UNIT
	ABC
ANNUALIZED COST BEFORE        15    27    65
  CREDIT (S1000)
ANNUAL RECOVERY CREDIT        18    70    215
  (S1000)
NET ANNUALIZED COST'          (3.0)   (43)  (150)
  ($1000)	
a (XXX) = NET CREDIT
                     FIGURE 8
   PERCENT OF TOTAL FUGITIVE EMISSIONS FROM
 	SPECIFIC COMPONENT TYPES	
                               PERCENT OF TOTAL
                                 UNCONTROLLED
 COMPONENT TYPE	EMISSIONS
 PUMPS IN LIGHT LIQUID SERVICE            12
 VALVES IN LIGHT LIQUID SERVICE           26
 VALVES IN GAS SERVICE                  11
 PRESSURE RELIEF VALVES IN
   GAS SERVICE                         23
 COMPRESSORS                         4
 OPEN-ENDED LINES                      14
                          TOTAL       90%
                       VI-7

-------
     The list of sixteen polymers and resins have been included in the
applicability of the model regulation so that the control of fugitive
emissions from this industry could be addressed in one document on VOC
fugitive emissions.  Process units in the polymer and resin manufacturing
industry have the same types of components and same chemicals-as process
units in the SOCMI.  Therefore, the control technology applicable to control
of fugitive VOC emissions in the SOCMI can be applied to the control of
fugitive emissions in the polymer and resin manufacturing industry.

     An equivalency provision is included in the RACT model regulation in the
CTG.  The purpose of the equivalency provision is to allow a company to
develop an equally effective leak detection and repair program which is
specific to the plant.  For equivalency a plant must demonstrate that
fugitive emissions expected under an alternative fugitive emission control
program are less than or equal to fugitive emissions from leaks using
component monitoring and leak repair as recommended by the model regulation
in the CTG.

     Industry has expressed a desire to have the latitude to try their own
programs for control of fugitive emissions in process units.  The equivalency
provision in the model regulation in the CTG gives industry the flexibility
needed to develop and test alternative fugitive emission control programs for
individual plants.

     This concludes the presentation for this CTG document.
                                    VI-8

-------
                       B,  INDUSTRY PRESENTATIONS
   1,  Chemical  Manufacturers Association

                           Mr, J. D. Martin
                         Union Carbide Company
                               Box 186
                        Port Lavaca, Texas  77979
     MY NAME is J, DANIEL MARTIN,  I AM A REGULATORY  MANAGER FOR
UNION CARBIDE CORPORATION'S POLYOLEFINS DIVISION,   I  AM  A  REGISTERED
PROFESSIONAL ENGINEER IN BOTH TEXAS AND LOUISIANA,   IN ADDITION,
I HAVE THIRTY YEARS OF CHEMICAL PLANT EXPERIENCE WITH UNION  CARBIDE
CORPORATION,  I AM ALSO A MEMBER OF THE CHEMICAL MANUFACTURERS
ASSOCIATION'S PROCESS EMISSION REGULATIONS TASK GROUP, AND I AM
LEADER OF ITS FUGITIVE EMISSIONS WORK GROUP,  TODAY  I AM SPEAKING
ON BEHALF OF CMA, A NONPROFIT TRADE ASSOCIATION HAVING 188 UNITED
STATES COMPANY MEMBERS THAT REPRESENT MORE THAN 90  PERCENT OF THE
PRODUCTION CAPACITY OF BASIC INDUSTRIAL CHEMICALS WITHIN THIS COUNTRY,
CMA MEMBER COMPANIES HAVE A DIRECT AND CRITICAL INTEREST IN  ENSURING
THAT EPA DEVELOPS CONTROL TECHNIQUE GUIDELINES (CTG)  WHERE A DEMONSTRATED
NEED IS PRESENTED, THAT ARE SCIENTIFICALLY AND TECHNICALLY SOUND,
REASONABLE, PROCEDURALLY WORKABLE, AND COST-EFFECTIVE,
     CMA HAS ACTIVELY WORKED WITH EPA OVER THE PAST FEW  MONTHS TO
DEVELOP A CTG FOR VOLATILE ORGANIC COMPOUND  (VOC)  FUGITIVE EMISSIONS
FROM THE SYNTHETIC ORGANIC CHEMICALS MANUFACTURING  INDUSTRY (SOCMI),
IN THIS REGARD, WE HAVE REVIEWED, COMMENTED  ON AND  MET  ONCE WITH
REPRESENTATIVES OF EPA'S OFFICE OF AlR QUALITY PLANNING  AND STANDARDS
(OAQPS) TO DISCUSS OUR CONCERNS WITH THE AGENCY'S  DRAFT  OF A CTG
FOR VOC FUGITIVE  EMISSION SOURCES,  WE HAVE  SEVERAL SIGNIFICANT
RESERVATIONS AND  CONCERNS WITH THE PROPOSED  DRAFT,  INCLUDING THE
NEED FOR THIS CTG,   IN THIS REGARD, THE THOUGHTS  WE OFFER  TODAY
                               VI-9

-------
   .AND THE MORE DETAILED WRITTEN COMMENTS WE WILL SUBMIT IN CONJUNCTION
    WITH THOSE OF THE TEXAS CHEMICAL COUNCIL (TCC) BY MARCH 20, 1981
    WILL ADDRESS THESE ISSUES, PROVIDE ILLUSTRATIVE DATA, INFORMATION
    AND RATIONALES, AND OFFER APPROPRIATE RECOMMENDATIONS,
1,   THE AGENCY HAS PUBLISHED THE DRAFT CTG WITHOUT THE BENEFIT OF THE
    RESULTS FROM SEVERAL ONGOING STUDIES.
         CMA RECENTLY REVIEWED AN UNPUBLISHED DRAFT CONTRACTOR REPORT
    ENTITLED "EVALUATION OF MAINTENANCE FOR FUGITIVE VOC EMISSION
    CONTROL,"  THIS FINAL STUDY WILL BE, PUBLISHED AS AN EPA REPORT IN
    APPROXIMATELY ONE MONTH,  THE STUDY CONTAINS DATA WHICH MUST BE
    REVIEWED WITH RESPECT TO THEIR EFFECT ON THE CTG,  ACCORDING TO
    OUR PRELIMINARY REVIEW THESE DATA MAY SIGNIFICANTLY AFFECT THE
    FOLLOWING ISSUES, AMONG OTHERS:
              1,  ON-LINE MAINTENANCE EFFECTIVENESS
              2,  THE COST-EFFECTIVE CHOICE OF MONITORING AND
                  MAINTENANCE INTERVAL
              3,  EMISSIONS REDUCTIONS RESULTING FROM THE PROGRAM
              4,  THE ADEQUACY OF THE REFINERY/SOCMI COMPARISON
              5.  THE 10,000 PPM LEAK DEFINITION
              6,  TIME TO CONDUCT ACTIVE MAINTENANCE AND MONITORING
    CMA FEELS STRONGLY THAT THESE DATA SHOULD BE INCLUDED IN THE CTG,
         EPA IS CURRENTLY ANALYZING THE DATA FROM THE MAINTENANCE STUDY
    AND FROM TWO OTHER STUDIES AND THEY WILL ISSUE A REPORT ON THE
    RESULTS,  ALSO, THE AGENCY IS REVIEWING THE RESULTS OF A STUDY BY
    ALLIED CHEMICAL CORPORATION ON LEAK OCCURRENCE AND RECURRENCE,
         WE REQUEST THAT EPA DELAY ISSUING THE FINAL CTG UNTIL THESE
    STUDIES ARE PROPERLY EVALUATED,  WE BELIEVE THESE STUDIES WILL
    JUSTIFY WITHDRAWAL OF THE CTG,
                                   VI-10

-------
                                  - 3 -

2,  IN THIS CTG THE AGENCY REQUIRES A STATE IMPLEMENTATION PLAN (SIP)
    REVISION FOR EVERY ALTERNATIVE WORK PRACTICE AND EVERY PERFORMANCE
    STANDARD VARIANCE.  CMA BELIEVES THE STATES SHOULD HAVE THE
    AUTHORITY TO MAKE DECISIONS ON ALTERNATIVE METHODS WITHOUT A SIP
    REVISION AND WITHOUT EPA APPROVAL.
         SECTIONS 101 AND 107 OF THE CLEAN AIR ACT EXPRESSLY PLACE THE
    PRIMARY RESPONSIBILITY FOR PREVENTING AND CONTROLLING AIR POLLUTION
    AT ITS SOURCE ON THE STATES,  SECTION 110 REQUIRES THE STATES TO
    SUBMIT TO THE AGENCY SIPs WHICH PROVIDE FOR IMPLEMENTATION, MAINTENANCE,
    AND ENFORCEMENT OF NATIONAL AMBIENT AIR QUALITY STANDARDS SET BY
    EPA,  ONCE THE BROAD AND GENERAL SIP is APPROVED BY EPA, CMA BELIEVES
    THAT THE STATES SHOULD HAVE THE AUTHORITY TO MANAGE THE AIR POLLUTION
    PROGRAMS DESCRIBED IN THEIR SIPs ON A DAY-TO-DAY, CASE-BY-CASE BASIS,
    WITHOUT UNDUE INTERFERENCE FROM EPA,  THIS IS NOT PRESENTLY THE
    CASE, AND THE CTG FOR VOC FUGITIVE EMISSIONS EXEMPLIFIES THE PROBLEM,
         THE AGENCY HAS INSISTED THAT IT BE PERMITTED TO SECOND-GUESS,
    BY MEANS OF AN INDIVIDUAL SIP REVISION, EACH AND EVERY STATE EXERCISE
    OF DISCRETION WITH REGARD TO EMISSION LIMITS ON INDIVIDUAL SOURCES,
    EPA HAS PROPOSED TO ADD 40  C.F.R,  SECTION  51,9  TO  REQUIRE
    SIP REVISION (APPROVED BY THE AGENCY) EVERY TIME THE STATE GRANTS
    ANY VARIANCE, EXTENSION OF TIME, REVISION OR WAIVER OF AN INDIVIDUAL
    SOURCE'S EMISSION LIMITS,
         CMA RECOGNIZES  THE NEED FOR  EPA TO REVIEW  THE  STATE  PLANS  TO  ENSURE
    REASONABLE FURTHER PROGRESS TOWARD ATTAINMENT IN NONATTAINMENT
    AREAS,  AND,  THEREFORE, THE  NEED FOR A FORMAL SIP REVISION IN CASES
    WHERE ATTAINMENT OR MAINTENANCE OF THE NAAQS MAY BE JEOPARDIZED,
    HOWEVER, WHERE THAT is NOT THE CASE, THE  STATES SHOULD BE ABLE TO
    ESTABLISH OR REVISE INDIVIDUAL SOURCE EMISSION LIMITATIONS ON A
    CASE-BY-CASE BASIS WITHOUT A SIP REVISION AND WITHOUT EPA APPROVAL,
                                  vi-n

-------
THIS FLEXIBILITY SHOULD INCLUDE THE AUTHORITY TO GRANT VARIANCES.,
EXEMPTION,  TIME EXTENSIONS, AND WAIVERS FOR SUCH REASONS AS
TECHNOLOGICAL FEASIBILITY, ECONOMIC HARDSHIP, ENERGY  CONSIDERATIONS,
OR  IMPRACTICALITY, SO LONG AS ATTAINMENT OR REASONABLE  FURTHER
PROGRESS TOWARD ATTAINMENT WILL BE MAINTAINED,
     THIS  IS PARTICULARY TRUE WITH REGARD TO THE GRANT  OF PERMISSION
BY  THE STATES TO APPROVE ALTERNATE PROGRAMS FOR VOC EMISSION  CONTROL,
SINCE A SOURCE'S USE OF THE ALTERNATE  PROGRAM BY DEFINITION MEANS
THAT ITS NET EMISSIONS WILL NOT EXCEED LEVELS ALLOWED BY THE  SIP,
OR  IN ANY  WAY ENDANGER A STATE'S REASONABLE FURTHER PROGRESS  TOWARD
ATTAINMENT,  IT MAKES NO SENSE FOR THE  AGENCY TO REQUIRE A SEPARATE
SIP REVISION EVERY TIME A STATE PERMITS THE USE OF AN ALTERNATE
PROGRAM FOR VOC FUGITIVE EMISSION CONTROL,  YET EPA HAS INSISTED
ON  MAINTAINING SUCH CONTROL OVER THE STATES,  CP1A SUBMITS THAT  EPA'S
POSITION  IS NOT SUPPORTED BY THE LANGUAGE OF THE ACT, AND IS  DESIGNED
SOLELY TO  PERMIT THE AGENCY TO SECOND-GUESS THAT STATE'S DECISIONS  IN
ACHIEVING  THE NATIONAL AMBIENT STANDARDS, AND TO PLACE  ENFORCEMENT
OF  THOSE REQUIREMENTS IN THE HANDS OF  EPA RATHER THAN THE STATES,
     ALTHOUGH THIS REQUIREMENT, AT FACE VALUE, MAY NOT  SEEM UNDULY
BURDENSOME,  ITS PRACTICAL CONSEQUENCES  ARE SEVERE,  THE SIP REVISION
REQUIRES:

           1,  ONE OR MORE PUBLIC HEARINGS, PRECEDED BY  AT LEAST 30
              DAYS' PRIOR NOTICE TO THE PUBLIC,
           2,  SUBMITTAL BY THE STATE OF THE PROPOSED  REVISION (ONCE
              THE STATE HAS APPROVED IT) TO EPA FOR REVIEW,
           3,  FULL REVIEW BY EPA, AND
           4,  A DECISION BY THE ADMINISTRATOR TO APPROVE THE  REVISION,
                                 VI-12

-------
                                  -  5  -

    IN  ADDITION,  ERA'S  INTERFERENCE  CREATES  A  CREDIBILITY PROBLEM THAT
    UNDERMINES  THE  STATE'S  ABILITY TO  IMPLEMENT  AND  ENFORCE ITS SIP,
    THE DUPLICATION OF  EFFORT  INVOLVED IN  EPA'S  SECOND-GUESSING OF THE
    STATES  INVOLVES A GREAT WASTE  OF STATE RESOURCES AND  AN ADDED COST
    BURDEN  TO  INDIVIDUAL  SOURCES THAT  IS WHOLLY  UNJUSTIFIED,
         THE INDIVIDUAL STATE  DIRECTOR SHOULD  EVALUATE  THE DATA SUBMITTED
    IN  REQUESTING A PERFORMANCE STANDARD OR  ALTERNATIVE PROGRAM,
    PERSUANT TO A GENERIC PROCEDURE  APPROVED BY  EPA,   (IF THE GENERIC
    PROCEDURE  WERE  FOLLOWED, THE EQUIVALENT  PROGRAM  WOULD BE ENFORCEABLE
    BY  BOTH THE STATE AND FEDERAL  EPA,)  THEN  THE  DIRECTOR SHOULD
    DETERMINE  IF  THESE  DATA ARE SUFFICIENT TO  SUPPORT THAT PERFORMANCE
    STANDARD OR ALTERNATIVE PROGRAM,   IF THEY  ARE  SUFFICIENT HE SHOULD
    DECLARE THE PROGRAMS  EQUIVALENT  TO THE STATE REGULATION,   II SHOULD
    NOT BE  NECESSARY THEN FOR  THE  DIRECTOR TO  SUBMIT A  SIP REVISION TO EPA,
3,   IN  DEVELOPING THE CTG THE  AGENCY SHOULD  USE  LEAK FREQUENCY DATA
    DEVELOPED  FOR THE SYNTHETIC ORGANIC CHEMICALS  MANUFACTURING INDUSTRY -
    (SOCMI).
         IN DEVELOPING  LEAK FREQUENCY  DATA THE AGENCY HAS PLACED EXTENSIVE
    RELIANCE ON DATA FROM THE  REFINING INDUSTRY, RATHER THAN THE SOCMI,
    FROM THIS  DATA  BASE EPA ESTIMATED  THE  EMISSIONS  REDUCTIONS FROM
    LEAKING COMPONENTS  IN SOCMI AND  THE EFFECT ON  AMBIENT AIR QUALITY,
    EPA ASSUMES THAT THE  REFINERY  AND  SOCMI  DATA ARE  SIMILAR,   HOWEVER,
    DATA FROM TABLE A-7 SHOW THAT  FUGITIVE EMISSION  RATES FROM THESE
    TWO INDUSTRIES  ARE  IN  FACT NOT SIMILAR,  OUR WRITTEN  COMMENTS GIVE
    GREATER DETAIL  AND  DISCUSSION  ON THE ISSUE,
         FURTHER, CMA HAS  COMPARED THE SOCMI/REFINERY DATA USING  THE  EPA/
    RADIAN  DATA CONTAINED  IN THE AGENCY'S  SOCMI  MAINTENANCE STUDY,   OUR

                                  VI-13

-------
                              - 6 -

PRELIMINARY ANALYSIS OF THIS UNPUBLISHED CONTRACTOR STUDY,
RELEASED FEBRUARY 17, 1981, FURTHER DEMONSTRATES THE PROBLEM
OF BASING THE CTG ON REFINERY DATA,  CMA RECOMMENDS A SCREENING
VALUE RANGE OF 40,000 TO 100,000 PPMV AND ANNUAL OR PRE
MAINTENANCE SHUTDOWN MONITORING,   THE MAINTENANCE SCREENING
STUDY SUPPORTS THE LONG STANDING C1WTCC POSITION THAT
SOCMI FUGITIVE EMISSIONS ARE SIGNIFICANTLY LOWER THAN THE VALUES
REFLECTED IN 'THE REFINERY DATA,  OUR FIRST ANALYSIS INDICATES THAT
ON A MASS EMISSION BASIS OVER 84 PERCENT REDUCTION IS ACHIEVED AT
THE 100,000 PPM LEVEL AND THAT ONLY 3 PERCENT MORE EMISSION REDUCTION
IS ACHIEVED WITH A 10,000 PPM SCREENING VALUE.  AT 40,000 PPM
THERE IS AN 86 PERCENT EMISSION REDUCTION,  THE INCREMENTAL EMISSION
REDUCTION ACHIEVED WITH A 10,000 PPM SCREENING VALUE IS NOT COST-
EFFECTIVE,  FURTHERMORE, THE FAILURE OF ON-LINE REPAIR TECHNIQUES
AFTER THE FIRST ATTEMPT TO YIELD ANY FURTHER SIGNIFICANT REDUCTION
SHOULD REQUIRE ONLY PRE-SHUTDOWN MONITORING AND STARTUP CHECKOUT,
A DIRECTED MAINTENANCE PROGRAM DURING A SHUTDOWN WILL YIELD OPTIMUM
RESULTS AT MINIMUM COSTS,
     THESE DIFFERENCES DO NOT ACCOUNT FOR THE FACT THAT THE CHEMICAL
INDUSTRY STUDIES WERE CONDUCTED USING A CENTURY OVA-108 INSTRUMENT
CALIBRATED WITH METHANE WHILE THE REFINING STUDIES WERE CONDUCTED
USING A BACHARACH TLV INSTRUMENT CALIBRATED WITH HEXANE.  STUDIES
BY EXXON CHEMICAL ON BOTH  INSTRUMENTS USING BOTH CALIBRATION GASES
SHOW THAT 29  PERCENT MORE LEAKS ARE FOUND USING THE CENTURY CALIBRATED
WITH METHANE  AS COMPARED TO THE BACHARACH CALIBRATED WITH HEXANE,
THUS THE SOCMI LEAK FREQUENCY IS PROBABLY EVEN LOWER THAN THE REFINING
DATA,
     FINALLY  AN ANALYSIS OF THE LEAK DATA FOR INDIVIDUAL SOCMI
PROCESSES FROM THE EPA/RADIAN 24 PLANT STUDY SHOWS THAT THOSE
                                 VI-14

-------
                              - 7 -
PROCESSES THAT EXCEED THE AVERAGE OF THE INDUSTRY ARE THOSE WHICH
ARE VERY SIMILAR TO REFINERY PROCESSES (l.E. THOSE INVOLVING ETHYLENE,
PROPANE OR PROPYLENE AS EITHER A PRODUCT OR RAW MATERIAL),
CONVERSELY, THOSE PROCESSES INVOLVING SPECIFIC CHEMICAL REACTIONS
RATHER THAN CRACKING OR FRACTIONATION, AS IN REFINERIES, SHOW A VERY
LOW FREQUENCY OF FUGITIVE EMISSIONS,  ACCORDINGLY, CMA BELIEVES
THIS STUDY SIGNIFICANTLY AFFECTS THE RACT ANALYSIS AND WE RECOMMEND
THE CTG BE REVISED TO INCLUDE THE SOCMI DATA FROM THE VARIOUS STUDIES,
IT MAKES NO SENSE TO PROVIDE THE STATES WITH CTG DOCUMENTS THAT WILL
RESULT IN UNJUST OVERCONTROL OF SEGMENTS OF THE CHEMICAL INDUSTRY,
THE PROPOSED CTG FAILS TO ADDRESS THE UNIQUE PROBLEM OF INACCESSIBLE
VALVES,
      IN OUR REVIEW OF THE PROPOSED REGULATIONS, WE IDENTIFIED AN
 ISSUE WHICH EPA APPARENTLY OVERLOOKED ~ VALVES THAT ARE INACCESSIBLE
 FOR SAFETY REASONS OR BECAUSE OF ELEVATION AND/OR CONFIGURATION,
 IN THIS REGARD, CERTAIN CHEMICAL PROCESSES ARE CARRIED OUT AT SUCH
 EXTREME CONDITIONS OF TEMPERATURE OR PRESSURE, OR THE CHEMICALS
 THEMSELVES ARE SO UNSTABLE OR HAZARDOUS THAT THE OPERATION IS DONE
 BEHIND BARRICADES AND THE LIKE AND, FOR SAFETY REASONS, PERSONNEL
 ARE NOT ALLOWED IN THESE AREAS WHILE THE UNIT IS IN OPERATION,  IN
 ADDITION, MANY VALVES ARE NOT ROUTINELY ACCESSIBLE IN EXISTING
 FACILITIES BECAUSE OF ELEVATION OR BECAUSE ACCESS TO THE VALVE
 BONNET IS RESTRICTED,  MANY OF THESE VALVES CAN BE ELIMINATED IN
 AN ENTIRELY NEW PLANT,  IN AN OLDER PLANT THAT WOULD BECOME SUBJECT
 TO THIS CTG THESE VALVES BECOME A PROBLEM,   As A RESULT, THE
 PROPOSED VALVE MONITORING REQUIREMENTS ARE INAPPROPRIATE FOR
 INACCESSIBLE VALVES,   WE PROPOSE THAT AN ALTERNATIVE REQUIREMENT
 BE PROVIDED FOR SUCH INACCESSIBLE VALVES INSTEAD OF THE REQUIREMENTS
 THAT WOULD OTHERWISE BE APPLICABLE UNDER THE CTG,  WE WILL DETAIL
 THE PROPOSAL FOR AN ALTERNATIVE REQUIREMENT IN OUR WRITTEN COMMENTS,

                               VI-15

-------
                                  - 8 -

5,  THE DRAFT CTG FAILS TO PROVIDE FOR EXTENSIONS TO COMPLETE REPLACEMENT
    AND/OR REPAIR OF A LEAK FOR CIRCUMSTANCES BEYOND THE NORMAL CONTROL
    OF THE OWNER OR OPERATOR,
         THE DRAFT CTG DOCUMENT REQUIRES THAT LEAKS BE REPAIRED AS SOON
    AS PRACTICABLE, BUT NOT LATER THAN 15 DAYS AFTER THE LEAK  IS DETECTED,
    EXCEPT WHERE THE REPAIR IS TECHNICALLY INFEASIBLE WITHOUT  A COMPLETE
    OR PARTIAL PROCESS UNIT SHUTDOWN,  THE EXTENSION, HOWEVER, CANNOT
    EXCEED THE PROCESS UNIT SHUTDOWN,  WE CONCUR THAT MANY REPAIR ACTIONS
    WHICH CANNOT BE TECHNICALLY OR SAFELY CONDUCTED WHILE THE  PROCESS
    IS IN OPERATION WILL BE REMEDIED DURING A PROCESS SHUTDOWN,
    NEVERTHELESS, MOST SCHEDULED SHUTDOWNS ARE ON AN ANNUAL BASIS OR
    BASED UPON OPERATING PERFORMANCE OF THE PROCESS UNIT,  As  A RESULT,
    THERE MAY BE SOME LIMITED INSTANCES WHERE REPAIR OR REPLACEMENT PARTS FOR
    LEAKING EQUIPMENT MAY NOT BE AVAILABLE UNTIL AFTER THE SHUTDOWN IS
    COMPLETED,  SUCH INSTANCES INCLUDE, BUT ARE NOT LIMITED TO, THE
    FOLLOWING:
         1,  ABNORMAL DEMANDS FOR REPLACEMENT PARTS THAT EXCEED
             THE QUANTITY OF REPLACEMENT PARTS THAT ARE NORMALLY -
             MAINTAINED IN STOCK AND CANNOT BE REPLACED ON SHORT NOTICE,
         2,  THE REPLACEMENT PARTS AND/OR EQUIPMENT THAT ARE NOT "OFF
             THE SHELF" ITEMS AND WHICH REQUIRE A LONG LEAD TIME FOR
             DELIVERY, AND/OR
         3,  (JNFORSEEN MANUFACTURERS AND/OR DELIVERY DELAYS,
    ANY ONE OR A COMBINATION OF THESE SCENARIOS WOULD RESULT IN THE
    NECESSARY REPAIR AND/OR REPLACEMENT PART(s) NOT BEING AVAILABLE
    UNTIL AFTER THE NEXT SCHEDULED SHUTDOWN,
                                     VI-16

-------
                              - 9 -

     SINCE THE PROPOSED CTG, IF INCORPORATED INTO A SIP WOULD MAKE
CONTINUED OPERATION AFTER SUCH A SHUTDOWN A VIOLATION OF THE CLEAN
AlR ACT, WE STRONGLY RECOMMEND THAT A LIMITED EXTENSION PROVISION
BE INCORPORATED BY EPA INTO THE FINAL CTG,  WE ENVISION PLACING THE
BURDEN OF REQUESTING SUCH AN EXTENSION UPON INDUSTRY BY REQUIRING A
WRITTEN REQUEST FOR EXTENSION BE SUBMITTED TO THE STATE, IN WHICH
THE SOURCE WOULD HAVE TO JUSTIFY THE NEED FOR THE FURTHER DELAY IN
REPAIR AND THE PROJECTED TIME FRAME FOR ACHIEVING COMPLIANCE,  (WE
WILL RECOMMEND APPROPRIATE LANGUAGE IN OUR WRITTEN COMMENTS,)
     THE CONSEQUENCES OF NOT INCLUDING SUCH A PROVISION IN THE FINAL
CTG COULD RESULT IN UNANTICIPATED AND COSTLY CONTINUANCES OF A
SHUTDOWN UNTIL THE REPAIR PARTS ARE OBTAINED, OR IN EXPOSING A
PERSON TO POTENTIALLY SIGNIFICANT CRIMINAL AND CIVIL PENALTIES FOR
RESUMING OPERATION WITHOUT REPAIRING ALL LEAKS.  WE BELIEVE THAT
THE AGENCY SHOULD PROVIDE IN THE MODEL REGULATION FOR A LIMITED
EXTENSION OF TIME, WHERE A SOURCE HAS ACTED IN GOOD FAITH, TO REPAIR
ALL REMAINING LEAKS AT THE NEXT SCHEDULED SHUTDOWN,
     CMA IS CONCERNED ABOUT THE USE OF REFINERY DATA IN A RUSH
TO ISSUE A DRAFT SOCMI CTG DOCUMENT,  CMA REQUESTS THAT THE CTG
NOT BE FINALIZED UNTIL AT LEAST 90 DAYS AFTER THE FINAL REPORTS
OF THE SOCMI MAINTENANCE SCREENING STUDIES AND THE FIVE IERL PROJECTS
TO ANALYSE THE STUDY ARE COMPLETE AND PUBLISHED,  THERE WILL STILL
BE ADEQUATE TIME TO REWRITE THE GUIDELINES BY OCTOBER, 1981 IF SUCH
A DOCUMENT IS STILL JUSTIFIABLE,  CMA IS AVAILABLE TO ASSIST IN THIS
REVIEW AND EVALUATION,
     THIS CONCLUDES MY FORMAL STATEMENT,  I WILL ATTEMPT TO ANSWER
ANY QUESTIONS YOU MAY HAVE CONCERNING MY PRESENTATION,
                              VI-17

-------
2.  Analytical  Instrument Development,  Inc.
                  COMPARATIVE METHODS FOR THE DETECTION OF
                     VOLATILE ORGANIC FUGITIVE EMISSIONS

                  Frederick J. Debbrecht, Jim D. Mitchell
                  Analytical Instrument Development, Inc.
                         Route 41 and Newark Road
                       Avondale, Pennsylvania  19311
INTRODUCTION
In EPA's Method 21 for the measurement of Fugitive Emissions Of Volatile Organic
Compounds (VOC), there are four detectors that are mentioned as possible detectors
for these organic materials.  The method is not limited to the use of these
four detectors, but in general the instrumentation available today uses one
of the four mentioned.  For the measurement of fugitive emissions, Method 21
also calls for the measuring device to be completely portable such that the
source of these emissions can be determined and, to some extent, the leak rate
actually determined.

The detection techniques suggested by Method 21 are Catalytic Oxidation,
Infrared Adsorption, Flame lonization and Photoionization.  In attempting to
quantitate the rate of fugitive emissions, it si important that the responses
of these various detectors be understood and their differences, one from
another, understood.  Even with a given detection system, the response for
different materials causes a good bit of confusion in the quantisation of the
results.  The thrust of this presentation is to point out the differences in
response factors on each of these different detectors for different materials.

The two most prevalent techniques that are used for fugitive emissions are
the Flame lonization System and the Photoionization System.  These -two techniques
will  be discussed in a bit more detail.  Also a little-recognized difference
between two different methods of Flame lonization will be presented and their
differing response factors discussed.
                                     VI-18

-------
CATALYTIC OXIDATION
The technique of Catalytic Oxidation brings the sample continuously across
the heated fialment coated with a catalyst.  The combination of the heat of
the filament and the catalyst brings about the oxidation of the volatile organic
compounds to carbon dioxide and water.  In the process of the combustion, if
you will, heat is generated which in turn changes the temperature of the filament,
and thus, its resistance.  The filament is generally in a Wheatstone bridge
such that the resistance change provides an unbalance of the bridge that is
the output of the catalytic oxidation detectors.  These types of detectors
are generally referred to as LEL Meters, or lower-explosive limit meters.  In
general, they are the most simple, smallest and most economical meter to use.
However, they are also the least sensitive meter.  Their range generally includes
something above 20,000 ppm by volume or 2% of the organic material down to
                       a
a minimum detectable in the range of 200 ppm.  The response of these devices
is significantly dependent on the heat of the combustion of the materials being
measured.

In general, within the hydrocarbon series of organic materials, the response
increases as the number of carbons in the molecule increases.   However, this
is certainly not a direct proportion increase to the number of carbons.  For
instance, ethane having two carbons has a greater response than does methane
with only one carbon; however, it is not double the response of methane.  Due
to their extreme stability, some of the chlorinated solvents have almost no
response in a catalytic oxidation detection system, passing over the filament
unchanged.

INFRARED ADSORPTION
The Infrared Adsorption technique measures the decrease of the infrared radiation
as it passes through the sample.  By selection of the proper wavelength of
this radiation, the instrument can be made to respond to virtually all  organic
materials.  In general, for this type of measurement, the stretching frequency
of the CHp group is used.  This occurs around 3.5 micrometers wavelength.  Most
all organic materials will have at leastoneCHp grouping that can provide some
response.
                                    -  2 -

                                    VI-19

-------
The response of the infrared detector in the use of the measurement of volatile
organic compounds is dependent then on the number of CH2 groups present per
molecule.  Methane in this instance provides a very poor to zero response.
It, being the first member of the hydrocarbon series, is somewhat unique. .Certainly
if one wanted to measure methane, there is a wavelength that could be used
that would provide excellent sensitivity for methane.  However, the 3.5 micro-
meter band is not the one for methane.

With other hydrocarbons such as butane, which has two CH,, groups in it, and
hexane, which as four ChL groups in it, we would find approximately double
sensitivity for hexane; thus, 100 ppm of hexane would provide approximately
the same output as 200 ppm butane.  Certainly the infrared adsorption technique
can be used for concentrations as high as 10,000 to 20,000 ppm.  By using an
extremely long path length for the infrared adsorption, sensitivities down
at the ppm region can be attained.  Thus, Infrared Adsorption is significantly
more sensitive than Catalytic Oxidation.  The instrumentation required is more
sophisticated and more delicate than the LEL-type meters.  However, there is
a fully portable infrared adsorption unit in the marketplace.

PHOTOIONIZATION
With Photionization Detectors, the sample is pulled continuously into a very
small chamber.  In this ionization chamber, the sample is continuously radiated
with a very short wavelength ultraviolet source.  The energy of this ultraviolet
source is sufficient to cause the ionization of the volatile organic compounds
in the stream within the chamber at that instant.  There is an electrical  field
present in the chamber that causes the ions to move towards one of the two
electrodes present.  On arriving at these electrodes, the positive ion or the
electron that was knocked out of the molecule by the energy of the ultraviolet
source becomes neutralized causing a very small  ion current to flow.  This
ion current is amplified by an electrometer-type amplifier and presented  to
a meter for the instantaneous readout of the total  organic content in the sample.
Generally provisions are made for a recorder to be used for continuous readout
as well -

The sensitivity of the Photionization Detector depends to some extent on the
ease of ionization of the individual organic molecule.  It also depends very

                                    - 3 -
                                    VI-20

-------
strongly on the energy available from the light source providing the ultraviolet
radiation.  In general, the low molecular weight hydrocarbons such as methane,
ethane and propane do not respond at all to this type of detector because sufficient
energy is not available to ionize the molecules.  In addition, some of the
highly chlorinated small molecules such as the freons are not ionized, again
due to the insufficient energy provided by the ultraviolet source.  Comparison
of the various sensitivities on a Photoionization Detector will  be given shortly,
following a discussion of Flame lonization.  The two techniques  of Flame lonization
and Photoionization are quite similar.

FLAME IONIZATION
Flame lonization Detection brings about the ionization of the volatile organic
compounds by use of the energy in a hydrogen flame.   The hydrogen flame causes
combustion of the organic compounds to carbon dioxide and water.   In the process,
there are some carbon ions formed and these are measured in an electrical field
much the same as in the Photoionization technique.   When these ions, or charged
carbon particles, arrive at one of the two electrodes and become  neutralized
at that electrode, a small ion current flows.  This  is amplified  by an electrometer
amplifier and presented to a meter and made available for recording purposes.
The Flame lonization Detector responds to virtually  all  organic  compounds.
Since its response is due to carbon particles being  charged,  any  organic-bound
carbon will show some response in a Flame lonization Detector.

There are two types of Flame lonization Detectors in use today for the measurement
of total organic materials.  The one most familiar  is an adaptation of a Gas
Chromatographic Flame lonization Detector.  This is  a detector that is referred
to as a three-gas system.  In operation, a small flow of the  sample containing
the volatile organic compound is mixed with hydrogen prior to the flame..  An
external source of air provides the necessary oxygen for the  combustion of
the hydrogen as well as the organic materials present in the  sample.  In general,
the flow rates are such that the sample flow and hydrogen flow are approximately
equal while the combustion air flow is approximately 5 to 10  times that of
the hydrogen and sample flow.  The three gases referred to in describing this
type of detector then become the sample flow, the hydrogen flow  and the air
flow.  It is this technique that is used on most fixed total  hydrocarbon monitors.
                                    - 4 -
                                     VI-21

-------
Most of the portable total hydrocarbon monitors in the marketplace today use
what is referred to as a two-gas system.  In this case, pure hydrogen, or perhaps
a mixture of hydrogen and nitrogen, is burned at the jet.  The sample flow
is brought in around the jet.  The sample, of course, has the organic materials
to be measured, but it also provides the oxygen necessary for the combustion
that must occur in the detector.  The two gases in this system then are the
fuel gas, hydrogen, and the sample gas, which also contains the oxygen for
support.  A little recognized fact is that these two modes of Flame lonization
provide significantly different responses for different organic materials.

The advantages, of course, of the two-gas system over the three-gas system
in a portable piece of equipment is that only one gas has to be supplied in
the two-gas system; namely, the fuel gas.  In the three-gas system, the fuel
gas, hydrogen, as well as the combustion air must be supplied.   A good bit
of the weight in a portable system is consumed in the gas flow systems and
the gas containers.  Thus, the two-gas system can indeed be made Tighter and,
therefore, more portable than the three-gas system.

In general, if one is simply measuring hydrocarbons, the response of the three-
gas Flame lonization Detector is pretty much proportional to the weight of
the hydrocarbon.  Thus, ethane will have about twice the response of methane
for the same concentration.  In the case of the two-gas system, we find that
the response is more on a molar or volume basis.  Thus, the response for ethane
will be about the same as the response for methane at the same concentration.

The Table shows response factors for various organic materials in both of the
Flame lonization modes as well as the Photoionization mode.  The simple interpreta-
tion of the flame response on either a weight basis  for the three-gas system
or a mole base for the two-gas system, of course, falls apart completely as
soon as one gets into organic compounds other than the simple hydrocarbons.
In general, oxygen introduced into the molecule or halogens introduced into
the molecule will depress the response obtained on the1 Flame Detection Systems.

CONCLUSION
When discussing, and most importantly, when using any of these techniques to
measure total  organic compounds in fugitive emissions, it is important to understand

                                    *• 5 -
                                    VI-22

-------
the differences in the different detector systems and how these different  systems
respond to the various organic materials.  The fact that these systems  all  respond
differently to different organic materials does not negate  the value  of these
detection systems in any way.   It  is  important that one understands these  different
response factors to properly  interpret  the data one is generating.   Even though
response factors can vary widely in a given  detection system,  it  should be
pointed out that if all of  these systems  are calibrated on  a known  benzene
standard and  one uses  them  to measure benzene in  an unknown sample  that contains
only  benzene, all of  them should give the proper, accurate  value  for the benzene
concentration in  the  unknown  sample.  The real  problem  comes in  interpreting
the data  on  samples  that  contain mixtures of organic  compounds and in some
cases, mixtures of  unknown  organic compounds.  Certainly even on  an unknown
organic mixture,  these instruments can  be used to give  meaningful data at  least
 on a relative basis such  that it can be determined where the emissions are
 coming from even though the quantisation of the leak rate of these emissions
 may be difficult to come by using these simple continuous monitors.
INLCI UUILEI UCICl^tUn
/"PORTS LENS A
U PUMP i— 1
17 -
BATTERY
_. 	 /]
— ... 	 — — \
HCTCD _ Uinoui 1
METEH -\ BOARDS

V ^
m FOCUS

                 7
                  CELL
                          CELL WINDOWS
FOCUS
 LENS
  FILTER -^ * SOunCE
             SCHEMATIC DIAGRAM OF PORTABLE INFRARED SPECTROMETER
                                     VI-23

-------
COLLECTOR
ELECTRODE
                                                        COMPRESSION  SPRING
                                                        LAMP HOUSING
                                                        LAMP
                                                        HIGH VOLTAGE CONTACT
                                                        LAMP WINDOW
                                                        IONIZATION CHAMBER
                                                      J  DETECTOR  EXIT
                                                              BIAS VOLTAGE
                                                           CROSS  SECTION
                                                   PHOTOIONIZATION DETECTOR
                                    SAMPLE  INLET
                                       7-4
                                      VI-24

-------
SCHEMATIC DIAGRAM OF THE FLAME IONIZATIQN DETECTOR
             RELATIVE RESPONSES
Material
Methane
Propane
Hexane
Benzene
0-Xylene
Acetone
Isopropanol
Chloroform
Ethanol
cs2
FID
2-gas 3-gas
0.50
0.42
0.55
1.00
0.70
0.30
0.65
0.35
0.30
0
0.18
0.49
1.01
1.00
1.21
0.33
0.36

0.24

PID
0
0
0.20
1.00
0.84
0.51


0.14
0.70
IR










CAT OX










                     VI-25

-------
3,- 'CarcH-iiia^Machrnery^ Sh Sypply Cpffipany
                         Mr, Pat Patterson
                          P. 0, Box 30187
                          Highway 64 East
                   Raleigh, North Carolina  27612
          The purpose of  this  presentation is to take a closer
     look at pump sealing, with  special emphasis on double mech-
     anical seals.

          Single seals, although reliable and efficient have a
     calculated life expectancy,  after which time product leaks
     into the atmosphere. Thus,  in the case of dangerous fluids,
     it is important to use  a  back up seal for protection.

          There are three types  of double seals:
                1. Back-to-back
                2. Tandem
                3. Face-to-face

          I. Back-to-back seals  can be indentified as the
             faces face in opposite directions.
         II. Tandem seals are  installed one behind the other
             with rotating faces facing in the same direction.
        III. Face-to-face seals  have a single stationary with
             rotating faces  facing toward one another.

          Of the three, only the Tandem and Face-to-face arrange-
     ment can provide the protection needed when sealing dangerous
     fluids. The Back-to-back, although a double seal, can only
     seal as effectively  as  a  single seal, and in some cases much
     less.

          There are three reasons why the Back-to-back seal design
     is subject to pre-mature  failure.

          1. The barrier  fluid between the seals has to be kept
     at 15 PSI greater than  the  pump pressure. Any fluctuation in
     this pressure can cause the internal seal to blow open and
     fail.
          2. The exterior seal sees the greatest pressure of all,
     because of the higher barrier fluid pressure. Thus the exter-
     nal seal will fail first, leaving no protection between the
     dangerous fluid and the environment.

          3. Because many fluids have high concentrations of solids
     or they may crystalize, the inboard seal is prematurely worn
     as the centrifugal force  generated by the rotation of the pump
     shaft throws the solids into the seal faces. Again the result
     is the same,  premature  seal failure.
A~-\
                               VI-26
                          A Division of
                       1  ismond Hill Supply Company, !r>c.

-------
                                          Page 2
     The other two designs, the Tandem and the Face-to-face,
work equally well. The barrier fluid in these seals is at a
lower pressure than the pump pressure, insuring longer life
for the all-important backup seal.  Also,  unlike the Back-
to-back design, these seals are independent of each other,
thus insuring maximum safety to the environment.

     A lot of time, research,  and money has been devoted to
setting up the standards for handling dangerous fluids and
protecting the environment. I  trust the point is clear, just
because a pump has two seals installed it doesn't necessarily
offer the protection needed. In setting up industry standards
for handling dangerous fluids  don't stop short of your intend-
ed goal	a safe environment.
                               Pat Patterson
                        VI-27

-------
  BACK TO BACK SEAL

   A.Inboard Seal
   B.Outboard Seal
                                    Barrier _Fluid
                                      (45 PSD-
Pump Pressure
  (30 PSI)
    Solid Particals  or
    Crystalized  Product
                     3
                                          VI-28

-------
        TANDEM SEAL

         A.Inboard Seal
         B. Outboard Seal
                                                          Barrier Fluid
Pump
Pressure
Atmosphere
       FACE TO FACE SEAL

        A.Inboard Seal
        B.Outboard Seal
                                                   Barrier  Fluid
                                                                                        Atmosphere
                                           VI-29

-------
4.   E. I.  du Pont de Nemours & Company
                          Mr, Thomas Kittleman
                    E, I. du  Pont de Nemours & Company
                       Wilmington, Delaware 19898
  We are pleased to offer our comments on the proposed Control
  Techniques Guideline (CTG) for volatile organic emissions.
  Later this week,  we will submit more detailed written comments
  regarding the items which I plan to discuss today as well as
  other issues relating to EPA development and use of this CTG.

  Today, I plan to  highlight Du Font's concerns regarding the
  proposed monitoring requirements.   Briefly, it is our position
  that the draft monitoring requirements ignore effective
  scientific sampling principals.  As I believe my remarks and
  supportive material will show, the proposed requirements are:

  o not cost effective for either industry or control agencies
    that will be responsible for their implementation.
  o put more monitoring burden on clean plants than on dirty ones.
  o discourage innovative approaches to fugitive emission control.

  The alternative inspection requirement we suggest would provide
  essentially equivalent control at a lower cost, place the
  monitoring burden on the dirty plants and encourage innovation.

  Last April I presented Du Font's views on monitoring in the
  draft fugitive emissions NSPS to this Committee.  We proposed
  the use of statistical inspection plans.  These sampling plans
                              VI-30

-------
                                -2-
are widely used throughout industry to monitor the quality of
manufactured products and man-*' other control parameter.  We see
no valid reason why such plans should not be used to monitor the
performance of values in chemical processes.

We have done additional studies to answer concerns raised at
your April 1980 meeting.  We have submitted these studies to the
EPA and visited with them to discuss our results.  EPA has not
responded to the issues raised, and we see no evidence that
qualified statisticians have reviewed them.  The draft model
regulation rejects sound principles of statistical sampling
without explanation.

At the April NAPTAC meeting we recommended that skip-period
inspection plans and their equivalents be used to determine how
much monitoring a new plant would be required to do.  We used a
skip-period inspection plan to demonstrate how one statistical
plan works.  We also believe the CTG should allow skip plans and
their statistical equivalents.  Adopting a skip- period plan and
the option to use its equivalents will result in good leak
protection at a far more reasonable cost.  If a plant's low-leak
performance deteriorates, these plans require increased
monitoring.  This approach allows small emission increases at
plants where emissions are low.  As a result, monitoring costs
                              VI-31

-------
                                -3-
will be reduced where inspections will not achieve any
significant environmental benefits.   High leak rate plants would
be required to do the most monitoring.  These plants could,
however, reduce monitoring costs by reducing leaks enough to
demonstrate good performance.   This insures that the most
intensive inspection is required where it will have the greatest
environmental benefit.   It also gives a dirty plant an incentive
to identify and correct the cause of leaks.

As an example of how the draft CTG model regulations would work,
consider two plants:

  Plant A initially has about  0.1% leaks.  Its overall
  uncontrolled valve leak rate is 0.35 Ib/hr.  After the CTG model
  regulation is applied, the leak rate is reduced to 0.04 Ib/hr.
  This leaves no room for additional reductions to allow an
  equivalency determination.  Therefore, Plant A is stuck with the
  most intensive inspection requirement.

  Plant B has about 22% leaks  on the initial screening.  Its
  overall uncontrolled valve leak rate is 49 Ib/hr.  After the CTG
  model regulation is applied, the leak rate is reduced to 4.9
  Ib/hr.  The plant has the option of further reducing this loss
  to compensate for a less demanding inspection requirement.
                               VI-32

-------
                                -4-
The result is that Plant B emits 100 times more and does less
monitoring than Plant A.

Consider what would happen with a skip-period approach to
monitoring.

Plant A could monitor once instead of four times a year.  Its
initial emissions would be reduced from 0.35 Ib/hr to 0.07 Ib/hr
instead of 0.04 Ib/hr under the CTG example.

Plant B, on the other hand, would be required to reduce valve
emissions from 49 Ib/hr down to4.9lb/hr as was required by the
CTG example.  To reduce its monitoring burden, this plant would
have to improve its leak performance until it met a good
performance criterion.  We believe a good performance level of
4% leaks can be justified for existing plants.  If the plant
cleans up and meets the good performance level, which we believe
it would be more likely to do than to go through the hassle of
proving equivalence that was postulated for the CTG example,
resulting emissions would be 4.4 Ib/hr.  Plant B would have an
additional 10% emission reduction compared to the CTG example.

The end result of these comparisons is that the skip-period
approach can yield greater emission reductions at much lower
                                VI-33

-------
OJ
             BEFORE
                   '7'
                          \
                          B
BEFORE
          50
                                               AFTER
                                    V
                                     (SKIP PERIOD  INSPEC
                   ION)
      AFTER
                                           AFTER
                                 (CTG
REQUIREMENT)
                                                                           BEFORE
                                                                      A
                                                                               EFORE
r-r
AFTER
                                                                          0
                                            EMISSION RATE #/HOUR

-------
                                -5-
cost and also greatly reduce the cost and hassles of equivalency



determination and SIP revision.







In short, we believe the CTG monitoring provision has the



following disadvantages:







o It attempts to inspect good leak performance into a plant rather



  than setting a realistic goal  and encouraging plants to maintain



  a low leak operation.  (Quality cannot be inspected in, it must



  be built in.)







o It is not cost effective because the same amount of inspection



  is required regardless of the  plant's leakage.







o There is little incentive to reduce emissions through equipment



  or work practice improvements  because the equivalency



  requirements are so demanding.







o The incentive may be to allow equipment and work practices to



  deteriorate.  The model regulation's requirements could supplant



  existing practices which may or may not more effectively reduce



  emissions.
                            VI-35

-------
                               -6-
o Low-leak plants will  be  least  able  to  take  advantage  of  reduced
  inspection alternatives.   The  better a plant  is  engineered  to
  reduce leaks,  the less opportunity  will exist for  reducing
  emissions to permit  reduced  inspection frequency.   For example,
  a plant that had no  leaks in the  initial baseline  could  never
  reduce inspection frequency  because it is impossible  to  have
  equivalent performance.

o Getting an alternative is too  complicated and costly  to  justify
  for most plants.  Significantly more time and money will  be
  required to get an alternative considered.  In addition,  there
  are uncertainties such as the  possibility that a given
  alternative wouldn't  be  accepted  or that the  time  that an
  alternative could be  used would be  short due  to  plant
  modification,  etc.  These factors reduce the  likelihood  that the
  alternatives would even  be considered.

o The draft CTG model  regulation focuses on local  emission
  reductions and, as a  result, ignores overall  or  national
  emission reduction.   We  believe that the model regulation should
  focus upon more demanding inspection procedures  with  respect to
  high leak rate plants and should  provide a  degree  of  flexibility
  for well constructed  plants  that  demonstrate  low leak
                            VI-36

-------
                                -7-
  performance.   This would not be the case with EPA's draft CTG



  model regulation   In fact, EPA's alternative could allow less



  demanding inspection of high-leak plants.







In the documentation we are submitting with our presentation



today, we have reviewed many aspects of the draft CTG model



regulation's approach and the skip-period approach to regulation.



We conclude that a 4% level of good performance is justifiable for



existing plants.  For a plant to continue to use skip-period



inspection, this 4%  criterion will require a group of valves to



have average leak frequencies less than 2%.







We further conclude  that even in the worst case, that is,  if



nobody improved performance to reduce their monitoring costs,  a 4%



level of good performance would still achieve 98% of the emission



reduction that EPA claims for the draft CTG.  This 98% reduction



would cost about half of what the CTG model regulation approach



would cost.  Based on EPA data in the CTG, their regulation will



cost the country about 50 million dollars a year.  Therefore,  a



skip-period approach could save about 25 million dollars a year



and in the worst case obtain nearly equivalent emission reductions.







In summary, based on the extensive experience of the quality



control profession with statistical sampling plans, we believe the
                            VI-37

-------
                                -8-
skip-period approach to regulation will result in greater emission



reductions than will be obtained by the draft CTG approach.   The



skip-period approach has the added benefit of achieving the



reductions in a much more cost-effective way.







With me today is Dr. Ron Snee.   The two of us have worked together



to develop these comparisons and would be happy to answer your



questions.
                            VI-38

-------
                                DRAFT OF CTG
                      GUIDANCE DOCUMENT FOR CONTROL OF
                     VOLATILE ORGANIC FUGITIVE EMISSIONS
                        COMMENTS ON REGULATION BASIS
We believe the CTG- format is inappropriate and should be revised.  For instance,
there is extensive discussion of control techniques that go beyond RACT.


We believe the applicability of the draft model regulation is too broad and
would cause inconsistency and waste.  For instance:

•  To avoid inspections of values that by definition cannot leak, the
   applicable vapor pressure should be 1.0 Kilo Pascal at process conditions.

•  Applying the monitoring equipment to mixtures that contain 10% or more VOC
   can also require equipment monitoring that cannot possibly have 10,000 ppm
   screening values.

•  The sensitivity of the specified analytical method varies greatly for many
   of the compounds that would be monitored.  So at the 10,000 ppm screening value,
   leak rates will differ greatly.  (For instance, chlorosulfonic acid would not
   be detected but would plug or corrode the instrument.)

•  Several of the compounds covered are not volatile organics (i.e., chlorosulphonic
   acid and hydrogen cyanide).

•  Some of compounds covered are self-detecting through means such as hydrolysis
   and plume formation in air.  This is probably a more sensitive test of leakage
   than required by the CTG.
                                 VI-39

-------
                                     -2-
The model regulation (Chapter 6.0) proposed to define a "leak" to mean a VOC




concentration greater than or equal to 10,000 ppmv as shown by monitoring or




   dripping of process fir id.  It further proposes that monitoring be conducted




at given intervals, the monitoring to be done with the aid of a portable analyzer




of the types briefly described in Appendix A.








As envisioned, the operator would standardize the instrument using "clean" air




and a calibrated source of methane (10,000 ppmv).  If he works in a methanol




plant, he immediately faces a dilemma.  Methane leaks around the reformer can




be monitored with reasonable accuracy.  But, how should he interpret a leak in




the methanol system if the instrument shows a reading 'equivalent to say 6,000 ppmv?




The operating manual for the OVA gives a response factor for methanol of 15.  We




would interpret that to mean that a reading of 1500 ppmv could actually be a leak




equivalent to 10,000 ppmv of VOC.  Therefore, the operator's finding of an indi-




cated 6,000 ppmv might actually be a major leak of 40,000 ppmv.








It could be argued that this was an extreme example to make a point.  But, now con-




sider a gas leak emitting a reformer gas containing hydrogen, methane, carbon




monoxide and carbon dioxide.  What is the response factor, bearing in mind that




each of these components can vary independently.  The tip of an iceberg is emerging.




The OVA manual lists nearly 80 common organic compounds (see Attachment), 7 of which




have a response factor of 100 (equivalent to methane).  Thirteen have response fac-




tors greater than 100 and would erroneously show a "leak" of the pure component at




some value less than 10,000 ppmv.  The rest have response factors less than 100.




For these, as sighted in our example, actual leaks greater than 10,000 ppmv could




be accepted as being in compliance.
                                VI-40

-------
                                     -3-
It is possible to go through Appendix B and site numerable examples where




materials being measured do not give the same instrument response  as methane.




Apparently, no consideration has been given to this very important aspect  of




the whole fugitive emission exercise.  Much is said about the  fractional sampling




of components but not one word covers the variability of the method.  What good




is it to develop a 98% or 99% precision in selecting components only to use a




method that has a response range from methanol at 15 to acetylene  at 225.








•  Does EPA expect the operator to restandardize the instrument for each compound




   in the process being tested in process?






•  Will EPA rely on the response factors as a correction for the meter reading?




   If so, where are the factors for all the compounds which are raw materials,




   intermediates and finished products in Appendix B?






•  For a complex system containing a mixture of compounds, how is  the operator




   to interpret the meter reading?








The  regulations are intended to be applicable to the SOCMI as  listed in Table I




in Appendix B.  Specific monitoring requirements may vary depending on whether




the  component services gases, light liquids or heavy liquids (vapor pressure less




than 0.3 kilo Pascal).  But, for each, the inventorying and categorizing of com-




ponents and recordkeeping is required.  Yet, many compounds on the list involve




organics and solids which, because of their extremely low vapor pressure,  cannot




and  will not be a measurable factor in the total scheme of fugitive emissions




from organic sources.  The list should be subdivided and those processes utilizing




heavy liquids, such as aniline manufacture by the hydrogenation of nitrobenzene,
                                 VI-41

-------
                                     -4-
should be excluded.  To require their compliance with the proposed fugitive




emissions regulations is a wanton waste of industrial resources.  The CTG




makes a strong case for the exclusion of heavy liquids from the regulation.









A significant number of the compounds listed in Table 1 are regulated by OSHA




as toxics or carcinogens.   Operating personnel exposures are regulated to eight-




hour averages of less than 1 ppmv in some cases.  A 10,000 ppmv leak in HCN or




dimethyl sulfate manufacture would be intolerable if not fatal.  Why should




such processes be loaded with these proposed added redundant administrative




and monitoring requirements.  Exclusion from the regulation on the basis of




equivalency should be allowed without the added administrative burden necessary




to obtain a SIP variance.
                                 VI-42

-------
                  EPA'S DRAFT CTG MODEL REGULATION
                 FOR SOCMI  FUGITIVE  EMISSION  SOURCES
                  Comments On Inspection Frequency ..

Introduction

EPA's draft CTG model regulation would require that all valves,
pumps, compressors, etc. to be inspected quarterly.  We agree that
quarterly inspection of some units may be needed.  It seems,
however, that inspection of all valves in all quarters is overly
conservative and would serve to unnecessarily impose an onerous
inspection program, which would not be required if the draft CTG
model regulation focused on emission reduction.

We have made extensive studies of the draft CTG model regulation
and some realistic alternatives.   (Snee and Kittleman 1980a, Snee
and Kittleman 1980b,  see Appendix B.)  These studies have been
submitted to and reviewed with EPA.   In the following comments we
identify the weaknesses in the draft CTG model regulation,
describe our alternative skip-period inspection plan, and suggest
ways in which EPA can revise its  model regulation to take
advantage of both its fixed-period inspection plan and our
skip-period inspection plan.
                            VI-43

-------
                                 -2-
                 Probleras with the Draft CTG Model
               Inspection Scheme  and  Its Alternatives

The draft CTG model regulation would  allow alternatives designed
to. guarantee leak frequency levels equivalent to that of the basic
plan.  Use of an alternative requires the collection of inspection
data for at least one year and demonstration of equivalency for
another year.  Provided equivalency is demonstrated, a State
Implementation Plan (SIP) revision would ultimately be required.
We have the following concerns with this approach to the
regulation of fugitive emissions.

o It attemps to inspect quality into  a product  rather than setting
  a realistic goal  and encouraging plants to maintain a low leak
  operation.

o It is not cost effective because the same amount of inspection
  is required regardless of the plant's leakage.
o
There is little incentive to reduce emissions  through  equipment
or work practice improvements because the  equivalency
requirements are so demanding.
o The incentive may be to allow equipment and work practices  to
  deteriorate.   The model regulations requirements could surplant
  existing practices which may or may not more effectively reduce
  emissions.
                           VI-44

-------
                                 -3-
o Low-leak plants will be least able to take advantage of reduced
  inspection alternatives.  The better a plant is engineered to
  reduce leaks, the less opportunity will exist for reducing
  emissions to permit reduced inspection frequency.  For example,
  a plant that had no leaks in the initial baseline could never
  reduce inspection frequency because it is impossible to have
  equivalent performance.

o The alternatives are too complicated and costly to justify their
  use at most plants.  Significantly more time and money will be
  required to get an alternative considered.  In addition,  there
  are uncertainties such as the possibility that a given
  alternative wouldn't be accepted or that the time that an
  alternative could be used would be short due to plant
  modification, etc.  These factors reduce the likelihood that the
  alternatives would even be considered.

o The draft CTG model regulation focuses on local emission
  reductions and, as a result, ignores overall or national
  emission reduction.  We believe that the model regulation should
  focus upon more demanding inspection procedures with respect to
  high leak rate plants and should provide a degree of flexibility
  for well constructed plants that demonstrate low leak
  performance.  This would not be the case with EPA's draft CTG
                           VI-45

-------
                                 -4-
  model regulation.  In fact, EPA's alternative could allow, less
  demanding ^nspection o£ high-leak plants.

            Skip-Period  Inspection  Can Reduce  Emissions

At the April 1980 NAPCTAC meeting we proposed skip-period
inspection for monitoring fugitive emissions (Snee and Kittleman
1980a).  We support this concept and encourage its inclusion in
the draft CTG model regulation.   Then,  as now,  we use skip-period
inspection to exemplify one quality control  approach to valve
monitoring.  If a regulation were written to specify the
appropriate parameters, other techniques  could give similar
results.

Skip-period inspection is based on skip-lot  sampling procedures
which have been used in the quality control  field for more than 25
years (Dodge 1955).  When a group of valves  has been found to be
within a specified good performance level (eg,  less than or equal
to 4% leaks) for five successive quarters,  the group of valves is
inspected once per year.   If the good performance level is
exceeded, then quarterly inspection is reinstituted.
                           VI-46

-------
                                 -5-
We believe that skip-period inspection answers the concerns with
the draft CTG model regulation.  For the following reasons,
skip-period inspection is a viable inspection plan which can be
used when the plant leak frequency is low.

o Through incentives to do less inspection, greater emission
  reductions may be realized than from fixed-period inspection.
  Plants will be encouraged to achieve and maintain good
  performance.

o Low-leak plants will be able to lessen the inspection burden.
  High-leak plants will be forced to continue the most intensive
  monitoring, insuring that the greatest emphasis is placed on
  plants that have the highest frequency of leaking valves.

o Well-defined rules for reduced inspection can be specified in
  advance.  This will encourage sources to use the option
  enhancing the likelihood of greater emission reductions.   It
  will also eliminate the consumption of industry and government
  time and money to judge the effectiveness of a variety of
  individual alternatives.
                            VI-47

-------
                                 -6-
      Justification of Skip-Period Inspection Good Performance
      Level of Less Than or Equal to 4 Percent Leak Frequency

At first exposure to the concept of skip-period inspection,  the
biggest single concern is the definition of the good performance
level (GPL).   We have made an extensive study of this issue  as it
pertained to  new plants (see Snee and Kittleman 1980b).   We  have
also concluded from cost and emission reduction effectiveness
studies (see  Appendix A) that a higher good performance  level is
justified for existing plants.   The results of our studies as they
apply to existing plants are summarized below.

o A 4% Good Performance Level would result  in fewer leaks and
  lower emissions than the draft CTG model  regulation when it
  encourages  plants to achieve and maintain good performance.  In
  order to use skip-period inspection on a  regular basis, the
  valves would have to be maintained at average leak rate of less
  than 2% in  order not to exceed a 4% GPL.

o Any plant that is designed and maintained to normally  perform at
  a leak frequency below 4% and uses skip-period inspection, will
  have a maximum emission increase of only  6.8% when compared to
  the model regulation.
                            VI-48

-------
                                 -7-
o Using the EPA chemical industry data and assuming that no plant
  improves performance, in the worst case we estimate that, with a
  \\ good performance level, leak frequency would increase by only
  1.6%.  (Appendix A)  If the two highest leak plants improved
  performance to take advantage of skip-period inspection, that
  reduction would more than compensate for reduced inspection at
  all other plants.

o Over the range of practical good performance levels,  (ie, levels
  that will provide an incentive to use skip-period inspection and
  still effectively reduce emission) there is little difference,
  for any given plant, between the emissions after fixed-period
  and skip-period inspection.

  We conclude that a 4% good performance level is attainable at
  many well designed and maintained plants, would encourage use of
  quality control options, and, hence, could achieve greater
  emission reductions at lower cost than the EPA draft  CTG model
  regulation.  Consequently, we believe that skip-period
  inspection with a 4% good performance level represents
  reasonably available control technology (RACT).
                           VI-49

-------
                                -8-
               Useful Changes to the Model Regulation

As stated earlier,  inclusion of a skip-period inspection program
would be a more reasonable approach to the regulatory framework.
This can be done by allowing for alternatives with the following
properties.

o Annual inspection should be permitted when the leak frequency is
  less than the specified good performance level for five
  consecutive quarters.

o The base inspection requirement in the model regulation would be
  reinstituted whenever  the good performance level is not
  achieved.  Increased inspection frequency would be the plant's
  penalty if an excessive number of leaks occurred.

o A level of good performance for purposes of defining the
  required inspection frequency should be specified  in the CTG
  model regulation.

o Alternative inspection plans that have the same statistical
  characteristics should be allowed.
                           VI-50

-------
                                 -9-
                              Summary

The draft CTG model regulation unfairly puts the most onerous
burden on the cleanest plants.  The approach violates sound
quality control principles in that the draft CTG model regulation
attempts to reduce emissions by plugging leaks rather than
encouraging good operation and maintenance.   It is not cost
effective because 98.4% of the emission reduction could be
obtained for approximately half the cost.

A valid quality control approach will provide a greater incentive
to reduce leaks, is easier to implement, and insures that plants
with the most leaks do the most inspections.  We support adopting
a skip-period inspection provision as a means of introducing
viable quality control options in this model regulation.
                             VI-51

-------
                                -10-
                             REFERENCES







Dodge, H. F. (1955).  Skip-Lot Sampling Plan, Industrial Quality



  Control. 11,  No. 5, 3-5.







Snee, R. D. and Kittleman, T. A. (1980a).  Statistical Inspection



  Plans for Monitoring Fugitive Emissions from Leaking Valves.



  Statement before the National Air Pollution Control Techniques



  Advisory Board, Hilton Hotel, Raleigh, NC, April 16-17, 1980.







Snee, R. D. and Kittleman, T. A. (1980b).  Skip-Period Fugitive



  Emission Inspection Plans:  Choosing A Level of Good



  Performance.   Report sent to the Environmental Protection



  Agency, November 1980.
                           VI-52

-------
                             APPENDIX A
        Comparison  of  Fixed-Period  and  Skip-Period  Inspection
          Using  EPA Chemical  Plant  Fugitive  Emissions  Data

The reductions in emissions (Ibs/hr) due to fixed-period and
skip-period inspection were compared using the leak frequency data
collected in the EPA Chemical Plant Study, the relationship
between before inspection emissions and leak frequency developed
by Snee and Kittleman (1980b)

  Before Inspection Emission (Ibs/hr) = 0.233 + 0.549 (% Leaks),

and EPA estimates of the effectiveness of fixed-period and
skip-period inspection
                                              Reduction In
           Inspection Plan            Leak Frequency and Emission
            Fixed-Period                           90%
            Skip-Period                            80%

which were summarized in the EPA Fugitive Emmissions Background
Information Document (BID).
It is shown in Figures 1A, 2A, and 3A that, at a good performance
level of 4% leaks, skip-period inspection will produce an
emissions reduction which is approximately 98% of fixed-period
inspection at a cost which is only 55% of fixed-period
                            VI-53

-------
                                A-2
inspection.  It is also clear that there is little difference



between the emission reductions using good performance levels of



2% and 4%.  Beyond a good performance level of 4% the emission



reductions drop off sharply.







Figures 1A and IB were developed from the EPA Chemical Plant



Study,  these curves show skip-period emission reductions (ibs/hr)



and inspection costs (No. valves inspected) as a percent of



fixed-period.  These two inspection performance measures are shown



as a function of good performance level, with GPL=0.4 (initial



leak frequency) as detailed in the BID.   Each point on these



curves was developed by fixing the GPL and computing the emission



reduction and inspection costs for those plants less than the GPL



(80% reduction, skip-period) and those plants greater than the GPL



(90% reduction, fixed-period).  Figure 3A shows a plot of



inspection cost curve in Figure 2A versus the emission reduction



curve in Figure 3A.  The calculations used to develop



Figures 1A-3A are described in Tables 1A and 2A and Exhibit A.







                              Reference







Blacksmith, J. R., Harris, G. E., and Langley, G. L. (1980),



  Problem Oriented Report:  Frequency of Leak Occurrence for.



  Fittings in Synthetic Organic Chemicals Process Units.  EPA



  Report, Contract No. 68-02-3171, Task 001, September 1.980.
                           VI-54

-------
                                         A-3





                                      TABLE 1A



                     EPA CHEMICAL PLANT FUGITIVE EMISSIONS STUDY
Plant
No.
21
60
34
12
22
20
61
66
1
29
3
33
65
32
28
31
5
11
6
2
4
%
Leaks
0
0.06
0.08
0.33
0.61
0.93
1.1
1.2
1.2
1.5
1.5
1.5
3.0
4.4
5.8
8.0
9.4
12.6
13.1
18.8
22.1
Valves
749
1648
1248
1826
162
1074
373
1096
1065
1903
2020
729
794
476
548
402
436
3166
811
3344
3960
Valves
x 4
2996
6592
4992
7304
648
4296
1492
4384
4260
7612
8080
2916
3176
1904
2192
1608
1744
12664
3244
13376
15840
GPL
For
Skip
0
0.02
0.03
0.13
0.24
0.37
0.44
0.48
0.48
0.6
0.6
0.6
1.2
1.76
2.3
3.2
3.76
5.04
5.24
7.52
8.84
EMISSIONS/PLANT U/HR)
Before
Inspection
0.17
0.44
0.35
0.75
0.09
0.79
0.31
0.98
0.95
2.02
2.14
0.77
1.49
1.26
1.87
1.86
2.35
22.64
6.02
35.28
48.99
After
Fixed
0.017
0.044
0.035
0.075
0.009
0.079
0.031
0.098
0.095
0.202
0.214
0.077
0.149
0.126
0.187
0.186
0.235
2.264
0.602
3.528
4.899
After
Skip
0.034
0.089
0.070
0.150
0.018
0.158
0.063
0.195
0.190
0.404
0.428
0.155
0.300
0.252
0.374
0.372
0.470
4.528
1.204
7.056
9.798
Total
27830
111320
131.52
13.154
26.308
                                   VI-55

-------
                                 A-4
                              TABLE 2A
             COMPARISON OF FIXED-PERIOD AND SKIP-PERIOD
                EMISSION REDUCTION § INSPECTION COST
                                   Skip-Period As A Percent
                               	Of Fixed Period	
   Good
Performance                                           Emissions
 Level (%)                     Cost                   Reduction
   0.1                         90.2                      99.9
   0.5                         75.1                      99.6
   1.0                         62.6                      99.2
   2.0                         59.1                      98.9
   3.0                         57.7                      98.8
   4.0                         55.4                      98.4
   6.0                         44.7                      96.0
   8.0                         35.7                      93.0
   9.0                         25.0                      88.9
                           VI-56

-------
                                       A-5
                                    EXHIBIT A

              COMPARISON OF SKIP-PERIOD AND FIXED-PERIOD INSPECTION
                COST AND EMISSION REDUCTION CALCULATION PROCEDURE
                               Emission Reductions

A = Total emissions from plants using fixed-period sampling  (ie,  % leaks  GPL)

B = Total emissions from plants using skip-period sampling (ie, % leaks  GPL)

C = Total emissions when all plants used fixed-period  inspection  (ie, model
    regulation)

«. ian^<^~r,.» D^,,^^.-i«r,   irtf>rTotal Emission Before Inspection-A-B1
% Emissions Reduction = 100[ Total ^issiou Before Inspection-C ]


Example:  Good Performance Level =4.0% Leaks


» D^,,,~H «   inn r 131.52 - 11.239 + 1.722,
\ Reduction = 100 [	m.52  - 15.154	]

            = 98.4
Inspection Cost

A = Number of valve checks per year for plants on fixed-period inspection (ie,
    % leaks  GPL)

B = Number of valve checks per year for plants on skip-period inspection (ie,
    \ leaks  GPL)

C = Number of valve checks per year when all- plants use fixed-period inspection
    (ie, model regulation)


% COST = 100 (A-1-^)
Example:  Good Performance Level =4.0% Leaks


% Cost = 100(16>55'124) - 55.4
                                    VI-57

-------
100%
                                                                                            75Z
 90% ~
                           FIGUKE 1A
SKIP PERIOD EMISSION  REDUCTIONS AS I OF FIXED PfRIOD [(EDUCTION
                              VS.
             GOOD PERFORMANCE  LEVEL FOR SKIP PERIOD
              (COMPOSIT  DATA FROM EPA 2'l PLANT STUDY)
                                                                                           251
 80Z
                            31      ill      51      6Z
                                 GOOD PERFORMANCE LEVEL
                                                                   8Z
                                                                                                     FIGURE 2A
                                                                                     SKIP PERIOD COST AS Z OF FIXED PERIOD COST
                                                                                                        VS.
                                                                                       GOOD PERFORMANCE LEVEL FOR SKIP PERIOD
                                                                                       (COMPOSIT DATA FROM EPA V\ PLANT STUDY)
                                                                                                                    51
                                                                                                                            6Z
                                                                                                                                            8Z
                                                                                                                          GOOD PERFORMANCE  LEVEL
                                      lOOE
                                       75%
                                       50Z
                                       25Z
                                                                           FIGURE 3A
                                                      NORMALIZED COST VS.  NORMALIZED EMISSION REDUCTION
                                                                   (FIXED  PERIOD =•  100Z)
                                         80Z
                                                                                90Z
                                                                       NORMALIZED REDUCTION
                                                                                                                       100Z
                                                                     VI-58

-------
    5.  Texas Chemical  Council
                              Mr, A, H, Nickolaus
                        E, I.  du Pont de Nemours & Company
                                P. 0, Box 2626
                             Victoria, Texas 77901
          My name is A. H. Nickolaus and I represent the Texas  Chemical
Council (TCC).   The TCC is an association of 85 chemical companies
having more than 67,000 employees and representing approximately  90%
of the chemical industry in Texas.  Since a significant portion of  the
nation's petrochemicals are produced by member companies operating  in
Texas, the proposed guideline is of vital concern to us.

          This Control Technique Guideline (CTG) is based on information
taken either directly or from the same sources as the Background  Information
Document (BID)  developed to support the Fugitive Monitoring New Source
Performance Standard (NSPS).  The TCC has provided comment and  information
to the EPA at several stages during the development of this document on
the need for,  cost of,  and technical basis for a regulation.  A chronology
of our comments is given in Table 1.  Many of them are still current and
apply equally well to this guideline.  Since the information is already
available to the EPA, we will not review it here.  In general we  concluded,
and still do,  that a regulation is unnecessary, and that costs  are  greatly
underestimated.

  I.  Technical Basis For The Guideline

      A.  Use Of Refinery Data

         From the very start the TCC had serious reservations about
the use of refinery data and some of the assumptions made in the  BID.
Recent data for the Synthetic Organic Chemical Manufacturing Industry
(SOCMI) indicate that our fears were well grounded and that there are
substantial differences between petroleum refineries and SOCMI.   Some
of these are:

         (1) EPA data  (Ref. 1)  on 23 SOCMI process units indicate
         considerable difference in the frequency of leaks from
         various equipment pieces when compared to petroleum refin-
         eries (see Table 2).

         (2) Refinery emissions were measured with a Bacharach  TLV
         instrument calibrated with hexane.  Similar SOCMI data were
         gathered using a Century OVA-108 calibrated on methane.
         Studies by Exxon (Ref. 2) on both instruments using both
         calibration gases show 29 percent more leaks are found using
         the Century calibrated on methane than with the Bacharach
         calibrated on hexane.

         (3) In a recent maintenance study (Ref. 3) significant
         differences were found in measured leak rates at the same
         screening value between valves in SOCMI and refining service.


                                VI-59

-------
                                   - 2 -
      B. Variability In Screening Data

         In addition to differences between the SOCMI and petroleum
refining data, recent studies show wide variability in the screening
process.

         (1)  Repeat measurements  (Ref. 1) show variations up to
         eight-fold in repeat screening values for the same source.

         (2)  There is about a two-fold difference in the confidence
         interval for repeat readings by the Century OVA of a constant,
         known, VOC concentration (Ref. 4).

         (3)  There is also a wide variation in the response of both
         the Century and Bacharach instruments to different chemicals
         (Ref. 4).  For example butane had a response factor of 0.38,
         n-butanol 1.43, n-butyl ether 2.70, and sec-butyl ether  0.26.

      C. Maintenance Data

         We are still studying the recent, complex maintenance report
(Ref. 3) and will comment on it in more detail later.  However, the
summary in Section 2 of it raises some disturbing questions.  A basic
premise of the EPA's inspection program has been the assumption that
once a leak had been identified, it could then be fixed promptly.  In
the maintenance study only 28.9% of the leaking sources could be  repaired
on-line with directed maintenance.  Further, although the data are
limited, a significant percent of those begin to leak again within a
few days after repair.  It is doubtful that these could be "repaired"
on-line a second time.  Thus, there is some question as to whether more
frequent inspection would reduce emissions or just rediscover the same
leakers that cannot be repaired.  We believe the EPA must make a  detailed
evaluation of this maintenance study and its implications.

      D. Conclusion And Recommendation

         The TCC believes (1) that sufficient information has now been
gathered to show that the petroleum refinery data are not representative
of SOCMI, (2) that there are serious problems with the screening  concept
because of variability within samples and among chemicals, and (3) that
the basic approach to maintenance may be faulty.  For these reasons we
recommend that the guideline be redone using current information  and
both the need for, and approach to fugitive emissions monitoring  in the
SOCMI be reappraised.

 II.  Comments On The Model Regulation-

      Our comments below reflect both long-standing concerns and  our
recommendations above.
                                VI-60

-------
                             - 3 -
A. Applicability

   (1) The specification of components exempted should be modified
   to include "pressure relief devices which are connected to an
   operating flare header,, vapor recovery system/ incinerators, or
   other equivalent system."  (Underlined portion added)

   (2) The EPA has apparently overlooked inaccessible valves in
   the model regulation.  These fall into two general categories,
   valves inaccessible for safety reasons and valves inaccessible
   because of elevation and/or configuration.  Certain chemical
   processes are carried out at such extreme conditions of
   temperature or pressure, or the chemicals themselves are so
   unstable or hazardous that the operation is done behind
   barricades and the like, and, for safety reasons, personnel
   are not allowed in these areas while the unit is in operation

   As the Agency is well aware,  in existing facilities many
   valves are not routinely accessible because of elevation or
   because access to the valve bonnet is restricted, etc.  Most,
   if not all of these, could be eliminated in an entirely new
   plant, but they are a problem in an older plant.

   To correct these problems, we propose valves that are in-
   accessible for safety and other reasons be excluded from the
   monitoring requirement.

B. Definitions

   (1) The guideline defines 10,000 ppmv as the definition of a
   leak.  No justification is given for this level in the guideline,
   It is the same level as the NSPS.  In addition, it is also the
   same action level as set forth in the refinery CTG,  and we
   understand from a July 17, 1980, EPA meeting with the TCC
   that this particular level was partly chosen because it is the
   top of the scale on the Century Volatile Organic Analyzer.
   Apparently the EPA felt 10,000 ppmv would make compliance
   easier for the .chemical industry.  We have some significant
   disagreements with these underlying assumptions.  We believe
   a higher trigger level would achieve essentially the same
   control with much improved maintenance efficiency.

   The scale on the Century GC is a minor consideration.   Readings
   higher than 10,000 can be obtained readily with a dilution
   apparatus.  In addition, we are confident that equipment
   manufacturers will be able to supply instruments with direct
   reading scales to whatever level we require.

   As the EPA has correctly pointed out,  most of the leakage comes
   from only a few valves and the problem is to locate  and repair
                         VI-61

-------
                          - 4 -
these.  Based on the refinery data,.about 98% of the emissions
from valves in gas/vapor service will be from those having leak
concentrations greater than 10,000 ppmv.  Similarly, about 97%
of the emissions will come from valves with screening values
above 20,000 ppmv (Fig. 4-7A, EPA 600/2-79-044, "Emission
Factors and Frequency of Leak Occurrence For Fittings in
Refinery Process Units").  In a previous letter to the Agency
(Ref. 5), the TCC has shown that using a 20,000 ppmv action
level versus 10,000 results in only 1% more emissions but
reduces maintenance costs by 30%.  This factor is of major
significance since Section 111 requires the Administrator to
take into consideration the cost of achieving continuous emission
reductions.

Methane at 10,000 ppmv is specified as the calibration gas.
This requirement differs from the refinery data on which most
of the technical support is based where hexane was used for
this purpose.  A study of the relative response of various
gases with respect to methane and hexane indicates that a
methane calibration will, in effect, lower the trigger point
to about 8,000 ppm with the Century GC and even lower with
The Bacharach TLV meter.  This means more maintenance effort
will be spent on valves with inconsequential leak rates.

Recent data  (Ref. 1) from EPA's studies on leak occurrence
and recurrence in the SOCMI show a wide variability in repeat
screening values for the same source.  For example in Figure
4-3 of the Problem-Oriented Report, on the second day, values
of approximately 2,000, 6,000, and 15,000 were obtained from
repeat measurements on the same valve.  It is important that
maintenance efforts be spent on the large leakers and not on
the small ones since some data indicate that attempts to
repair these only made matters worse.  Thus the trigger point
should be set high enough to insure isolating only the bad
leakers.

Still more recent data from the maintenance study (Ref. 3)
indicate that practically all of the on-line maintenance
reductions obtained  (on a mass basis) were from valves with
screening values of 50,000 ppmv or above.  Also leak rate
versus screening value were different and substantially less
in the SOCMI study than for refinery valves.

Based on maintenance efficiency, the effect of methane
calibration, and measurement variability, the TCC had
recommended a trigger point of 20,000 ppmv for the NSPS.
However, considering the maintenance study results (Ref. 3)
                           Vl-62

-------
                             - 5 -
   and the response factor variability with different chemicals
   (ref.  4),  the TCC now believes a leak definition of approxi^
   mately 50,000 ppmv is necessary to achieve the same results
   as we  had  originally expected from a 20,000 ppm cut-off.

   We recommend that the definition of a leak be changed to
   50,000 ppm for both the NSPS and this guideline.

   (2) The definition of "repair" should be changed to mean
   reducing a leaking component to below 50,000 ppmv as shown
   by monitoring.

   (3) The present definition of "Volatile Organic Compound"
   includes some materials such as methane, ethane, etc., which
   the EPA has stated are not photochemically active but which
   are measured by the applicable test method.  The definition
   should be changed to eliminate this contradiction

C. Standards

   (1) Monitoring Frequency:   No justification is given for a
   quarterly monitoring frequency.  It is simply assumed along
   with the leak occurrence,  recurrence, etc., numbers used to
   calculate  emission reductions and costs.  We are frankly
   mystified as to how the EPA can use refinery data and come
   up with a monitoring schedule for the SOCMI that is more
   stringent than the one for petroleum refineries; expecially
   when SOCMI data show lower leak rates (see Table 2).  We
   believe monitoring frequency should be determined from a
   rational analysis of fugitive emission reduction (mass)
   versus cost.  In the absence of this, we recommend that the
   monitoring frequencies be changed to the same as the petroleum
   refining CTG.

   (2) Relief valves are to be monitored within 24 hours after
   venting etc.  (XX.030B).  The Petroleum Refining CTG (Ref. 6)
   allows 15  days for this and even the proposed SOCMI fugitive
   monitoring NSPS (Ref. 7) allows five days.  Surely the
   proposal in the guideline is an error.  We recommend it be
   changed to 15 days, the same as the petroleum refining CTG.

   (3) We presume the EPA is being facetious when they say that
   from the date (?)  a leak is detected the owner/operator shall
   "Affix within 1 hour" a tag, etc.  To seriously propose this
   would be unconscionable.  There is no such equivalent or even
   similar requirement in either the refinery CTG or the pro-
   posed SOCMI NSPS.   Obviously the whole phrase must be deleted.
                          VI-63

-------
                             - 6 -
   (4)  Again,  the repair requirement for pumps is more stringent
   than the SOCMI NSPS.   It should not be.  Based-on the refinery
   CTG and the SOCMI NSPS,  we recommend the following for.
   paragraph XX.030 (D):

   "Each owner/operator shall visually inspect each pump weekly
   for evidence of a leak.   When a potential leak is found, the
   seal area should be monitored within five days to determine if
   a leak is present,  i.e.  a concentration over 50,000 ppm.  If
   so it should be repaired within 15 days.  Delay of repair
   beyond 15 days is allowed if the repair is technically in-
   feasible without a complete or partial shutdown of the process
   unit.  Delay beyond a process or unit shutdown is allowed only
   when repair or replacement equipment cannot be obtained until
   after the shutdown is complete."  Note;  See the Chemical
   Manufacturers Association comments on this at the March 3, 1981
   public hearing on the SOCMI NSPS.

   (5)  The discussion of paragraph XX.030(G) on alternate
   emission control systems specified at least one year's
   data.  We don't know that this would be necessary in all
   cases and we recommend this language be deleted.  Also on
   page 6-11 we note that if the State Director decides some-
   thing is equivalent he must run to the EPA and see if he
   can get an SIP revision for it.  This sounds like EPA
   rulemaking to us.

D. Reporting

   (1)  The reporting period should be geared to the inspection
   schedule whatever it turns out to be after the EPA has re-
   appriased this program per our above recommendations.

   (2)  The content of the report should be no more than:

        (a) Process unit identification.

        (b) Number and type of leaking components not repaired
       within 15 days.

        (c) Reason for non-repair within 15 days.


   We do not believe the EPA has authority to require more than
   this  (see discussion in Ref. 8 and joint CMA-TCC written
   comments to be submitted by April 6, 1981 concerning the
   SOCMI Fugitive NSPS)  and we don't see how they can propose
   more for the States.
                              VI-64

-------
                                   - 7 -
III.  Rulemaking?

      The TCC believes that EPA's insistance that these guidelines be
incorporated essentially verbatum into the SIPs is tantamount to rule-
making despite their denials and legal technicalities that avoid this
fact.

 IV.  Endorsement of CMA Comments

      The Chemical Manufacturing Association (CMA) comment covers some
items we did not and some other aspects of items we did discuss.  The
TCC agrees with and endorses the CMA comments.
AHN/rtg
3-9-81

Attachments
                                 VI-65

-------
REFERENCES:
      (1) EPA Problem-Oriented Report,  "Frequency of Leak  Occurrence
          for Fittings in Synthetic Organic Chemical Process  Units",
          September, 1980.

      (2) Letter:  B. C. Davis (Exxon)  to D. W. Carroll  (CMA),
          "Analysis and Comment Regarding SOCMI Leak Frequencies  ..."
          February 4, 1981.

      (3) Langley & Wetherold, "Evaluation of Maintenance  For Fugitive
          VOC Emissions Control" Radian Corp. Report, February  17,  1981.

      (4) "Response Factors of VOC Analyzers Calibrated  with  Methane
          For Selected Organic Chemicals" Prepared  By Radian  Corp.,
          September 30, 1980.

      (5) Letter: H. H. McClure (TCC) to Jack R. Farmer  (EPA) Feb.  1,
          1980 - TCC Comments on the  Draft BID.

      (6) CTG for "Control of Volatile  Organic Compound  Leaks from
          Petroleum Refinery Equipment" EPA-450/2-78-036,  June  1973.

      (7) Proposed NSPS for VOC Fugitive Emissions  from  the SOCMI,
          FR 46, page 1154, January 5,  1981.

      (8) Chemical Manufacturers Association (CMA)  statement  at the
          March 3, 1981 public hearing  on Proposed  Rulemaking of  New
          Source Performance Standards  for Volatile Organic Compound
          Fugitive Emission Sources.
                                VI-66

-------
                                   TABLE 1

                    TEXAS CHEMICAL COUNCIL (TCC) COMMENTS
                      TO THE EPA DURING THE DEVELOPMENT
              OF VOC FUGITIVE EMISSIONS MONITORING REGULATIONS
May 17, 1979
Letter From H. H. McClure  (TCC) to David  R.
Patrick (EPA).  Comments on the March 1979
Hydroscience Report on Fugitive Loss Control
Option.
Feb. 1, 1980
Letter from H. H. McClure  (TCC) to Jack R.
Farmer (EPA).  Comments on the Draft Background
Information Document.
June 30, 1980
Letter from TCC to EPA.  Comments on the Draft
BID and recommended SOCMI standard.
July 28, 1980
Letter from H. H. McClure  (TCC) to Walter Barber
(EPA),  "Texas Chemical Council Data On Capital
1 Creep'".
July 30, 1980
March 3, 1981
Letter from H. H. McClure  (TCC) to Walter Barber
(EPA), "TCC/EPA Conference on Proposed SOCMI
Fugitive Emission NSPS."

TCC Testimonry at the public hearing on the SOCMI
Fugitive Emissions Monitoring NSPS.
                                    VI-67

-------
                                         TABLE 2

                           COMPARISON  OF SOCMI AND PETROLEUM
                 REFINERY LEAK  RATES AND RACT MONITORING REQUIREMENTS
      COMPONENT
    SOCMI LEAK
  RATE AS % OF
PETROLEUM REFINING
                                                  MONITORING REQUIREMENTS
  SOCMI
PETROLEUM REFINING
VALVES
 •GAS SERVICE
 •LIGHT LIQUID SERVICE
     SIMILAR
       50%
QUARTERLY
QUARTERLY
    QUARTERLY
    ANNUALLY
PUMPS
 •LIGHT LIQUID
 •HEAVY LIQUID
       33%
     SIMILAR
QUARTERLY
    ANNUALLY
VISUAL WEEKLY    VISUAL WEEKLY
COMPRESSORS
RELIEF VALVES
       21%

       45%
QUARTERLY
QUARTERLY
  .  QUARTERLY
    QUARTERLY
PROCESS DRAINS
                    CAP
                 ANNUALLY
PETROLEUM REFINERY LEAK RATES FROM TABLE  4-2  NSPS BID
SOCMI LEAK RATES FROM REF.  1

PETROLEUM REFINING RACT MONITORING FROM THE CTG FOR PETROLEUM REFINING
  EPA 450/2-78-036, JUNE  1978
                                       VI-68

-------
6.  Tienn-essee €astfnan  Company


                          Mr,  J. D, Thomas
                            P.O. Box 511
                     Kingsport, Tennessee 37601


  Good morning, ladies and gentlemen.  I  am  J.  D.  Thomas,  Civil

  Engineer, of the Clean Environment Program staff of Tennessee

  Eastman Company.  Tennessee Eastman Company is  a producer of

  chemicals, fibers, and plastics.  Our facilities located in

  Kingsport, Tennessee, provide employment for  approximatley 12,000

  men and women.   Because many of the materials listed in  Appendix B

  are produced at  Tennessee Eastman Company,  this  proposed Control

  Technique Guidelines (CTG) will have a  significant  impact on our

  operations.  Tennessee Eastman Company  is  a member  of the Chemical

  Manufacturers Associaton  (CMA).  We support the  comments made by CMA

  today.  The comments that we are submitting today make certain

  recommendations  that attempt to clarify, and/or  modify the agency's

  proposed regulatory program to make the program  technically more

  sound, to simplify procedures, recordkeeping, and reporting.

  Specifically, we wish to comment on five aspects of the  proposed

  Control Technique Guideline.



  1.  The data for this Control Technique Guideline are based on

      information gathered from the petroleum refining industry.

      These data were previously used to  establish Control Technique

      Guidelines for "Emission Factors and Frequency  of Leak

      Occurrence for Fittings in Refinery Process  Units."   The

      Agency states on Page 2-19 that "data  characterizing

      uncontrolled levels of  fugitive emissions in the Synthetic

      Organic Chemicals Manufacturing Industry  (SOCMI) are presently

      unavailable."  Appendix A, however, lists data  for a total

                             VI-69

-------
Page 2





    of 34 surveyed SOCMI units.  Table A-7-on Page  A-13  compares



    leak frequency for  fugitive emission  sources  in SOCMI  units"



    and petroleum refineries.  These data  clearly demonstrate that



    leak frequency for  SOCMI  is substantially less  than  that  found



    for the refining  industry.








    These differences do not  account for  the  fact that  the



    chemical industry studies  used  a Century OVA-108 instrument



    calibrated on methane while the refining  studies used  a



    Bacharach TLV instrument  calibrated on  hexane.   The  Radian



    Corporation report  "Response Factors  of VOC Analyzers  at  a



    Meter Reading of  10,000 PPM for Selected Organic Chemicals"



    finds that more leaks are  "found" using the Century  calibrated



    on methane as compared with the Bacharach calibrated on



    hexane.  Clearly, the SOCMI data available shows fundamental



    differences in leak frequency for the  SOCMI industry.   The CTG



    should be revised to reflect the SOCMI  data now available.







2.  The agency should rework  Chapter 5 "Control Cost Analysis of



    RACT" to properly reflect  cost data.   All model plant  cost



    calculations are  based on  leak frequency data experienced by



    the refining industry.  As pointed out  in No. 1  above, measured



    frequency of leaks  for SOCMI differs  fundamentally  from that



    of refineries.  Thus, initial leaks are overestimated  and a



    falsely high estimate of  material recovery is reported in the



    CTG.
                           Vl-70

-------
Page 3







    The agency has also seriously  underestimated  the true cost- of



    the RACT program.  For example,  the  CTG  estimates labor costs



    at $18 per hour.  Means  1980 Standard  Cost  Index places labor



    costs at $22.00 per hour.  The agency  estimates  the cost of



    all capped lines on the  basis  of the cost of  a one-inch valve



    plus one hour of labor.  No data are presented to support the



    supposition that such a  one-inch line  is an accurate average



    value.







    Finally, the CTG assumes a fixed 10% interest rate.  At a time



    when the prime rate is approximately 18%, this assumption does



    not reflect a reasonable cost  of money.







3.  In requiring that every  valve  be monitored  once  a quarter,  the



    agency has apparently overlooked the fact that many valves  are



    not routinely accessible.  Reasons for inaccessibility vary



    due to safety reasons, configuration,  and elevation



    constraints.  In addition, many  valves may  be inaccessible



    because of existing insulation or other  valve coverings.



    While many inaccessible  valves could be  eliminated in an



    entirely new plant design, It  would  be unduly burdensome on



    the regulated community  to require monitoring of these valves



    in an RACT standard.  The agency should  provide  regulatory



    relief for inaccessible  valves in finalizing  the CTG.
                         VI-71

-------
Page 4

4.  The model  regulation  at Page  6-3  requires repair of the
    leaking component within  15 days;  or repair of the leaking
    component  at or  before the next  scheduled unit turnaround if
    unable to  do so  within 15 days.   This proposal ignores the
    instances  where  replacement parts  for leaking equipment may
    not be available until after  the  next scheduled turnaround is
    completed.  In many cases, the demand for replacement parts
    may exceed  the quantity of parts  normally stocked.  In these
    cases, it  may not be  possible on  short notice to obtain
    adequate replacement  parts.   The  proposal also ignores
    unforeseen manufacturing  and/or  delivery delays due to
    strikes, raw material delays, etc.   The effect of the proposal
    would be to require expensive duplicative spares for all
    valves, pumps and other regulated  equipment.

5.  The agency states on  Page 6-5 that  "the purpose of the
    regulations is to have owners and  operators of plants
    implement  a leak detection and repair program.  A plant is in
    violation  of the regulation if they are not trying to find
    leaks and  repair them."  The  agency makes no attempt to equate
    such a program with actual reduction in VOC fugitive
    emissions.  In fact,  a facility  could be in compliance even
    though each affected  piece of equipment were found to "leak"
    at each subsequent  inspection so  long as the inspection is
    performed  and there is an attempted repair.  There appears to
    be a little impetus under the proposed regulation to actually
    reduce fugitive  emissions of  VOCs.
                           VI-72

-------
Page 5
    The agency has provided for development of  alternative



    equivalent programs including a performance  standard.   We



    believe that if a performance standard could  be  judged



    equivalent, it should have been substituted  for  the  work



    practice standards chosen by the agency in  the proposal.



    Unfortunately, the performance standard outlined  in  the



    equivalence section (Page 6-10-6-12) would  tend  to penalize



    plants with good performance records.  Because the equivalent



    performance standard is based upon the last  two  quarters'



    monitoring data under the work practice, a  plant  with a large



    percentage of "leaking" process components  during the



    equivalence determination period would not  be required  to



    achieve a level of performance consistent with a  plant  having



    few "leaking" process components during the  same  equivalence



    period.  For example, a plant with 12% "leakers"  (the refinery



    average) during the equivalence determination period would



    have a performance standard of 88%; a plant  with  7%  "leakers"



    (the SOCMI average) during the equivalence  determination



    period would have a performance standard of  93%.







    Finally, we do not believe that determination of  an  equivalent



    performance standard should require independent  SIP  revision



    and EPA approval.  We believe that such equivalence



    determinations are within the scope of the  discretionary



    authority granted to the state within the State  Implementation



    Plan.  So long as an equal or greater reduction  in VOC



    fugitive emissions occurrs under the equivalent  standard,  no



    separate SIP revision should be required.
                              VI-73

-------
Page 6





We appreciate the opportunity  to address  the Committee concerning



this draft Control Technique Guideline.   We  would be happy to



attempt to answer any  questions  that  you  may have or to di-scuss



any of our comments  further.
                           VI-74

-------
 7.  Monsanto Company
                           Mr, Jerry M, Schroy
                            Monsanto Company
                       800 N. Lindbergh Boulevard
                       St, Louis, Missouri  63166
I.   Introduction

     Good afternoon.  My name is Jerry Schroy and I am a Monsanto

     Fellow in Monsanto's Corporate  Engineering Department.  I am

     here to present Monsanto's comments  on the preliminary draft

     control technique guideline dealing  with volatile organic

     fugitive emissions.   Based on  limited time available I will

     attempt to summarize the comments contained in the document

     you have before you.  A copy  of Monsanto's detailed comments

     been provided for each member of the committee and the EFA

     staff present today.



     Understandably my comments will center on those issues of

     concern to Monsanto although  they may be discussed by others today

     The areas of concern I will cover are:

             * The need for additional fugitive emission

              regulations.

             •The technology for control  of fugitive

              emissions.

             •Model regulation questions.

             •Validity of the data base.
                                 - 1 -


                                  V1-75

-------
II.   Need for Regulations



     The primary issue of concern to Monsanto is the real need  for



     additional regulations to control fugitive emissions.  This



     concern is based on the EPA's use of petroleum refinery  data



     to characterize the performance of synthetic organic chemical,



     polymer and resins manufacturing industries  (SOCMI).   The  EPA's




     lack of understanding of regulations that impact SOCMI facilities is



     illustrated by a statement made in the background  information



     document  (BID), dealing with a proposed new source performance



     standard  for this topic  (USEPA  (1980).  The statement  is:








        "There  are presently no federal regulations that



        specifically reduce emissions from synthetic organic chemical



        manufacturing plants.  However, some fugitive emission



        reduction is achieved by operating practices currently



        followed by industry and applicable state or local



        regulations."



     This statement and further discussions of Safety and Health



     Regulations in the BID are of concern because they are in  error.



     The Occupational Safety and Health Administration  (OSHA) regulations



     for control of contaminant levels in the workplace are in



     fact fugitive emission control standards.  EPA's limited knowledge



     and lack  of understanding of OSHA regulations is of concern because



     the draft CTG does not address this topic.








     OSHA's  regulations stipulate the application of engineering



     controls  and work practices to  reduce workplace exposure.   The



     engineering controls and work practices suggested  by OSHA  are






                                 VI-76~

-------
in fact identical to those discussed in the draft CTG.  A



more detailed discussion of the impact of OSHA is given in



Section I of Appendix B.








We believe that this misconception of OSHA's impact is a



result of the transfer of technology from the petroleum



refining industry.  The differences in the chemical, physical



and biological impact of the chemicals handled by the petroleum



refining and SOCMI industries do not allow direct transfer



of the fugitive emission concerns.  A comparison of several



petroleum refining and SOCMI chemicals is given in Table 1.



With materials handled by the petroleum refining industry, exclusion



of oxygen and explosion hazards dictate operating and design



criteria.   While in SOCMI facilities, the toxicity of the



chemicals handled often controls design and operating decisions.








An example of how OSHA regulations impact a SOCMI plant is



illustrated in a case study presented in Appendix B.  The



results are shown on the following two tables (Table 2 and 3) .








We believe the CTG should be modified to include a discussion



of OSHA's role in control of fugitive emissions.  Since the



CTG is to be used (1) by state and local environmental control



agencies in preparation of their state implementation plans,



and (2) the EPA in their review of the state and local programs,



all parties should clearly understand the impact of OSHA's workplace



exposure limits.  The explanation of OSHA's role would assist



in transfer of the technology concerns from the petroleum



refining industry to the SOCMI facilities.




                           VI

-------
      In many instances  where OSHA regulations are strict  because



      of the nature of the chemicals involved, regulations by state



      or local organizations would be unnecessary.  (eg. HCN,



      vinyl chloride,  and acrylonitrile)








III.   Technology for Control of Fugitive Emissions



      A second area of concern to Monsanto is EPA's understanding



      of available fugitive emission control technology.  Two examples



      of this limited knowledge are illustrated by the descriptions



      of (1) the function and performance of pump seals and  (2) the



      definition of sampling system technology.








      While these points are not major they are symptoms of the



      over simplification of SOCMI fugitive emission controls.  The



      performance of fugitive emission control elements and



      practices in the petroleum refining industry do not properly



      address the concerns and practices in the synthetic organic



      chemical, polymer and resin manufacturing industries.  Some



      of these concerns were addressed earlier in the discussion



      of OSHA's impact on SOCMI plants.  However, the technology



      descriptions in the CTG will not allow innovative responses



      to fugitive emission control practices and could lead to



      unduly restrictive specification standards at the state and



      local levels.








      For instance, the CTG indicates that in a pump seal"...the



      seal faces must be lubricated to remove frictional heat...".



      The explanation also indicates that much less lubricant is






                                 VI-78~

-------
needed for a mechanical seal than for a packed seal because

of the seals'  construction.  This simple, explanation does

not properly characterize the relationship of seal face

cooling and seal emission.



While the fluid pumped into the seal, from either the

pump impeller cavity or an external source, may lubricate

the seal faces, its primary function is to remove heat.

The lubrication properties of the fluid,  which reduce heat

generation, are only secondary benefits.   The key factors

which impact heat generation (eg. seal face area,  face

velocity of the mating surfaces,  pump operating temperature

and pressure,  and seal mechanism design)  coupled with the

physical/chemical properties of the seal  cavity fluid dictate

the type and rate of seal emissions.



To illustrate this point several chemicals covering a

wide range of physical and chemical properties were examined

in a simulation of a seal operating conditions (Table 4).  The

results of this ambient condition analysis indicate the

importance of the heat of vaporization and vapor pressure

to emission rate.  A summary of the simulation results is given

in Table 5.  This simulation is consistent with a study

performed in England to characterize the performance of packed

and mechanical seals  (Summerfield  (1980)}.  These data  are also

presented for comparison.
                           - 5 -
                           VI-79

-------
The data clearly illustrate that the physical/chemical

properties of the material contained by the sealing  component

will have a direct impact on seal emission rate.


The reason for this comparison is to emphasize the function

of basic knowledge to the innovation process.   An understanding
of the behavior of fluids in seal emission performance

would allow innovative corrective action.   For instance

when product purity specifications prevent use of a  "heavy

liquid" flush fluid,  the use of auxiliary cooling to reduce

the process fluids'  vapor pressure and promote sensible

cooling of the seal faces is a realistic method of controlling

seal emissions.  A detail discussion of this practice is given

in Section II of Appendix B.


Another example of a simplistic discussion of control technology is

the description of sampling connections.   The application of

limited volume, low pressure sampling systems as described

by Bruce Lovelace (Lovelace (1979))  is more representative

of technology used in SOCMI plants,  than the discussion

of sampling practices in the petroleum refining industry

presently included in the CTG.   An example of this type of

sampler is shown in Figure 1.


We believe these two points should be corrected prior to

publication of the CTG for use by state arid local agencies.
                            - 6 -
                            VI-80

-------
IV.    Model Regulations
      The third area of concern involves  the  proposed model
      regulation included in the CTG.   Specifically, the model
      regulations include definitions  that are technically in error.
      The three definitions of concern to Monsanto are those for
      "light liquid service," 'Volatile organic compound," and
      the definition of applicability  of  the  standard^ (§xx.010)	 	
      A detail discussion of those topics  are given in Appendix
      B.  A summary of our concerns is  presented on Table 6.   I
      won't read them for the sake of time.   However,  we believe
      the definitions should be changed as follows to rectify
      these concerns.
      Applicability
        (b)   This regulation applies to sealing components
             that can be characterized  as  being "In Light Liquid
             service" in the synthetic  organic...
        "In Light Liquid Service"  means that the component contacts
        a liquid with a concentration greater than 20 percent by
        weight of a volatile organic compound having a vapor pressure
        greater than 760 mmHg at component operating temperature.

        "Volatile Organic Compound" means  any orgsnic compound
        which participates in atmospheric  photochemical reactions
        and is measurable by the applicable test methods described
        in Reference method 21 which can be calibrated fay a
        saturated straight chain hydrocarbon or equivalent state
        method.
                                - 7 -
                                VI-8T

-------
        A discussion of the basis for these changes is given in

        Section III and IV of Appendix B.




V.    Validity of the Data Base

      The fourth area of concern is the validity of the emission

      rate data base.  Specifically, the data in the open

      literature is inconsistent, for an as yet unidentified

      reason.   This inconsistency is complicated by the EPA's

      rejection of some of the data collected specifically on

      SOCMI facilities.




      The EPA has rejected data which they indicated was not

      comparable to the results of other studies.   However,

      the data rejected is consistent with data developed by

      groups other than the EPA.  A discussion of these studies

      is included in Section V of Appendix B.




      I might point out that this is not a gored ox comment,

      even though the work rejected was completed by the

      Monsanto Research Corporation.  The comment is made

      because no one should be permitted to discard data, regardless

      of its impact on the project study,  test etc.




      The incompatibility of the data calls for a more

      reasoned analysis of the results to identify the cause

      of this incompatibility.  Data incompatibility due to

      physical/chemical differences in properties discussed

      earlier and differences in physical properties will cause

      differences in fugitive emission rates.


                                - 8 -
                                VI-82

-------
Another cause is the definition of the VOC.  If the
definition is based on different calibration gases the

rate will vary.  This was shown in Table  3.  Use  of  different
detectors will lead to different results  for identical
calibration gases  (Table 7).

Another reason may be the method used to  convert  field
emission rate test results  (i.e. bagging  tests) into
emission rate data.  Several EPA contractors have reported
their sampling and data conversion procedures in  a variety
of documents  (Smith  (1979), Tierney, et.  al.  (1978), Radian
Corporation  (1979), Hughes et. al.  (1978)).  Although the
procedures for data  conversion are described in each report,
they are not  consistent.

A detail discussion of this question is included  in  Section V
of Appendix B.  A  summary of our analysis is given in Figure  2
and Table 8.


Based on our  calculation, the conversion  factors  used by
Radian were a factor of approximately 5 in error  on  some
of their studies.  This would result in reported  emission
rates that were a  factor of 5 above the actual field results.
Since the field results for tests performed by Radian
are not available  we can not make any specific comments.

However, a comparison of several studies  may clafify our
concerns.
                            -  9  -
                            VI-83

-------
Several studies from the public literature are summarized

on Table 9.  These data indicate a sharp division between

the results report in Table 2-2 of the CTG for light liquid

pump seals and data from the open literature.  This includes

the data discarded by the EPA.

 Based on  this  analysis  we  suggest  that an  independent

 contractor review Radian's test results  and  determine

 if in fact the data  base  is valid.



 This  concludes my remarks.   If you have  any  questions

 or comments,  I will  be  happy to answer them  now  or following

 the presentations of the  others scheduled  to speak today.
                           - 10 -
                           VI-84

-------
                  APPENDIX A:  TABLES AND  FIGURES
Areas of Concern
 • The need for additional fugitive emission  regulations,
 •The technology for control of fugitive emissions.
 • Model Regulation Questions
 'Validity of the data base.
   "There are presently no federal  regulations  that



   specifically reduce emissions  from  synthetic organic



   chemical manufacturing plants.   However,  some fugitive



   emission reduction is achieved by operating  practices



   currently followed by industry and  applicable state



   or local regulations."



                             USEPA  (1980)
                           VI-85

-------
              Table 1
                   Comparison of Relative Hazards  for API  and  SOCMI  Chemicals
oo
01
Chemical
Name	
methane
ethane
propane
butane
pentane
n-hexane
nonane
decane
dodecane
acetic acid
acrylonitrile
chlorobenzene
epichlorohydrin
ethylene
ethyl ether
vinyl acetate
vinyl chloride
                                         TLVR,
                                         ppm (v)
                                   (AGCIH 1980)  or OSHA
                                              -asphyxiant
                                              -asphyxiant
                                              -asphyxiant
                                              -asphyxiant
"inert" gas-
"inert" gas-
"inert" gas-
"inert" gas-
TWA 8 = 600
TWA 8 =  50
TWA 8 = 200
No criteria
No criteria
TWA 8 =  10
TWA 8 =   2 OSHA
TWA 8 =  75
TWA 8 =   2
"inert11 gas-asphyxiant
TWA 8 = 400
TWA 8 =  10
TWA 8 =   1 OSHA
Equilibrium Vapor Composition
with Pure Liquid @  75°F  in  Air
 (@ Specific Chemical), ppm  (v)
         1,000,000
         1,000,000
         1,000,000
         1,000,000
           644,350
           189,290
             5,325
             1,664
                168
            19,010
           133,830
            14,580
            19,995
         1,000,000
           673,780
           142,950
         1,000,000

-------
Table 2    Monsanto Survey
                        Percent of Fittings  Found to Leak
                              Pumps              Valves
Source                     (Light Liquid)     (Light Liquid)

Radian Refinery Study           23.0              10.0
 (Hexane calibration)

Radian SOCMI Study               8.8               6.4
 (Methane calibration)

Radian AN Data                   8.2               1.9

Monsanto Study                   5.9               0.9
                     VI-87

-------
Table 3   VOC Emission Rate Comparison


Device
valve
valve
valve
pump
pump
flange
i — i
m CTG
co
Pump
Valves
Flanges


Process
Chemical
Acrylonitrile
Acrylonitrile
Acrylonitrile
Acrylonitrile
Acrylonitrile
Acrylonitrile

Light liquid
Light Liquid
Light Liquid


Screening
Value
6,000
10,000
1,000
2,600
600
3,000






Line
Temperature
100
130
150
130
115
130






Line
Pressure
(psi)
60
90
50
90
60
90






Spec. Chem
Spec. Chem
1.9
9.1
0.2
0.2
11
0.3






VOC
(methane)
1.8
8.7
0.2
0.2
10.7
0.3






VOC
(hexane)
0.6
3.1
0.1
0.1
3.8
0.1

120
10
0.3
Process
Chemical
Vapor
Pressure
mmHg
183
348
516
348
155
348

>2.25
>2.25
>2.25

-------
Control Technology








Pump Seals






   Function of seal face lubrication








Sampling Systems






 • Limited Volume,  Low Pressure Liquid



   Sampling




  •Flow through Sampling Systems
    VI-89

-------
              Table 4    Physical/Chemical Characteristics of SOCMI and API Chemicals
10
o
                                                                   Carbon
Vapor Pressure
Chemical
Name
References
methane
ethane
propane
butane
pentane
hexane
vinyl acetate
aery lonit rile
crylohexane
toluene
chlorobenzene
x-xylene
nonane
dccane
dodecane
Molecular Boiling Point Heat Values Atoms @75°F (2;
Weight °C K BTU/lb mole per Molecule mmHg
(1) (2)
16.04
30.04
44.10
58.12
72.12
86.18
86.09
53.06
84.16
92.14
112.99
106.17
128.26
142.29
170.34
(1) (2)
-161.5
- 88.6
- 42.1
- 0.5
36.1
68.7
72.7
77.3
80.7
110.6
132.3
139.1
149.1
172.0
214.8
Combustion
(3) (4)
345.2
614.3
879.3
1143.7
1407.7
1672.1
—
757.4
1587.0
1622.7
1337.5
1881.9
2465.5
2729.9
3258.9
Vaporization
(3) (4)
3.52
4.06
6.62
7.06
11.36
15.57
15.90
14.00
13.98
16.09
18.10
18.40
19.97
22.10
22.66

1
2
3
4
5
6
4
3
6
7
6
8
9
10
12
(1) (2)
184,800
28,670
6,870
1,760
490
144
109
102
93
27
11
8
4
1.3
0.13
).9~C)
psia

3,570
554
133
34
9.5
2.8
2.1
2.0
1.8
0.5
0.21
0.15
0.08
0.03
0.003
     (1)  Dean  (1973)
     (2)  Dean  (1979)
     (3)  Perry and  Chilton (1973)
     (4)  Weast  (1975)

-------
              Table 5.  Comparison of Simulated Pump Seal Emissions Rates
Chemical
Name Emission Rate,
as specific
chemical
(1)
methane —
ethane 137
propane 71
butane 24
pentane 9
hexane 3
vinyl acetate 3
acrylonitrile 1
cyclohexane 2
toluene 0.5
chlorobenzene 0.25
m-xylene 0.2
nonane 0 . 1
decane 0.03
dodecane 0.004
grams per hour @ 75°F
as VOC w/flame
ionization
detector (methane)
(2)
—
200
105
56
19
8
2
1
5
1
0.7
0.6
0.06
0.3
0.05
seal cavity temperature
as VOC w/catalytic Summerfield (1980)
oxidation detector Boiling Pt VOC(hexa
(methane) °C Avg.
(2)
--
193
167 • -55 17+
37
14 50 2.0
4 60 2.8
0.6
3
3 80 1.6
0.2
0.4
0.04
0.01
0.2
180 0.5
(1)  Emission rate for a 100mm diameter shaft seal.   Simulated as a pool of liquid 5.sq. in. in
    surface area for identical heat input conditions (Wu and Schroy (1979)) .
(2)  Based on data presented in Table 7.
    times actual organic rate.
Average values for VOC (methane) per actual organics

-------
                  Figure 1
          OPERATION  OF  ,AN  *I.SOLOK*
SAMPLE.
CONTAINER
                        VI-92

-------
Table 6

Applicability
     - should be consistent with definition of light liquid service

     - consistent with data base.
Light Liquid Service

     - 20% by weight is consistent with emission rate study

       (Wetherold et. al. (1979)).

     - 20% by weight is not consistant with the definition of

       applicability.

     - Limitation of the number of available safe pump seal buffer

       fluids.

     - Emission rate impact of luids with low vapor pressure at

       component operating temperature (Table 4).

         • Seals do not stop vapor movement when pressure is- above

          atmospheric pressure.


Volatile Organic Compound

     - Reference Method 21 deficiencies

        • Inability to calibrate all detectors defined in Reference

         Method 21.

        •Inability to measure all photoreactive compounds

        •Consistency with data base developed by EPA contractors.
                               VI-93

-------
Table 7
Ratio of Observe vs. Actual Organic Concentration
                  VOC @ Methane eq.
                  Flame lonization
                                    VOC 8 Methane Eq.
                                   Catalytic Oxidation
Chemical
Methane
Ethane
Propane
Butane
Pentane
Hexane
Vinyl Acetate
Acrylonitrile
Cyclohexane
Toluene
Chlorobenzene
m-xylene
Nonane
Decane
Dodecane
Theory
1
2
3
4
5
6
4
3
6
7
6
8
9
10
12
.0
.0
.0
.0
.0
.0
.0
.0
.0
.0
.0
.0
.0
.0
.0
Ref 1
1.0
1.18
1.82
2.0
1.92
2.44
0.79
0.97
2.13 -
2.56
2.63
2.50
0.65
11.1
-
Ref 2
1
1
1
2
2
3
0
0
2
3
2
3
0
.0
.75
.14
.63
.38
.23
.76
.96
.78
.03
.78
.33
.62













00

Observed


Methane
Ethane
Propane
Butane
Pentane
Hexane
Vinyl Acetate
Acrylonitrile
Cyclohexane
Toluene
Chlorobenzene
m-xylene
Nonane
Decane
Dodecane
(1) Brown et.

















al.





•











(1980)
Flame
Ref 1
1
0.59
0.61
0.5
0.38
0.41
0.20
0.32
0.36
0.37
0.44
0.31
0.07
1.11

(2)
-

Methane
Theory
1.
1.
2.
3.
4.
4.
-
2.
4.
4.
3.
5.
7.
7.
9.
0
86
55
31
08
84

19
60
70
87
45
14
91
44
Ref
1.
1.
1.
1.
1.
1.
0.
3.
1.
0.
1.
0.
0.
6.
-
0
45
67
59
59
45
17
49
43
37
14
17
09
25

ea . /Theoreti cal
lonization
Ref 2













1
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.

88
38
66
48
54
19
32
46
43
46
42
07













00



Dubose and

Harris
Catalytic
Ref 1

0.78
0.65
0.48
0.39
0.30
_
1.59
0.31
0.08
0.29
0.03
0.01
0.79































1 Ref 2
1.0-
1.37
3.03
1.47
1.61
1.39
0.25
2.70
1.39
0.43
1.14
0.28
0.18
5.0
-
Methane ea
Oxidation
Ref 2

0.74
1.19
0.44
0.39
0.29
—
1.23
0.31
0.09
0.29
0.05
0.03
0.80

(1981)
                               VI-94

-------
Figure 2
Determination of Units conversion factor
(a) )Q,aefjnf x P,
            T,9Rr
     x 10~6  IbsAHf    60
     where  10.73 cf x psia ~   gas constant
                   OR-lb moles

     k =  10"6  x 60    5.592 x  10~6
            10-73


(b)
VQ,.a
-------
           Table  8   Emission  Rate Calculation Factors-
Author
 (Contractor)
 Smith  (1979)

 Langley  and
 Wetherold (1981)

 Radian (1979)

 MRC (private
 communication)
                                                Units Correction Factor
                                                     k           k

psia
in Hg
psia •
in Hg

ppm(w)
ppm(w)
ppm (w)
ppm(v)
Reported
2.74x10-5
2.7xlO~6
2.74x10-5
2.7xlO-6
Correct
Value
5.592xlCT6
2.746xlO-6
5.59x10-6
2.746xlO~6
            Table 9   Emission Rate  for  Single  Mechanical
                       Seal in "Light Liquid" Service(D
                         VOC  (hexane) grams  per hour
Date
Source
USEPA  (1980)
Summerfield  (1980)

Monsanto Study  (Section II)
  Acrylonitrile

Bierd et. al.  (1977)
  Benzene

Tierney et. al.  (1978)
  Monochlorobenzene

Hughes et. al.  (1979)
  Monochlorobenzene
 120
   4
10.6
  29
    (2)
 9.7 to 29
          (2)
Avg. of values except USEPA  (1980)   11.9
(1)   Based on definition proposed by EPA

(2)   Corrected from chemical to VOC  (hexane) based  en  Table  7
                            VI-96

-------
                  APPENDIX B:  DETAIL TECHNICAL COMMENTS
                                            Monsanto Co.
                                            March 18, 1981
Technical Comments on Proposed VOC Fugitive Emission

The following comments are submitted by Monsanto Company on the
proposed regulation based on a technical review of the issues
concerning control of fugitive emissions.  The comments are
intended to assist the EPA in clarifying and/or correcting those
points of concern to Monsanto-  The comments are based primarily
on information available in the public literature or on data
submitted to various US EPA groups in the past.  Those
references used as a basis for these comments are cited in the
text and listed at the end of the discussion section.

The primary goal of those reviewing the proposed regulation for
Monsanto are to protect the health and safety of the workplace
and the environment surrounding Monsanto's plants.  Those groups
in Monsanto which deal with worker health, personnel safety and
property protection, and environmental concerns work closely
together to insure a proper assessment of risk and benefit for
each project or process change.  The proposed regulation dealing
v/ith fugitive emissions was reviewed as a process modification.
Interaction by these groups are believed necessary to avoid
creation of one or more problems while attempting to solve a
preceived problem (Schroy, 1980).

Monsanto's commentors feel qualified to provide technical input
on the subject of fugitive emissions control based on the
Monsanto material which has been published in the open
literature over the past few years (Crocker (1979a), Crocker
(1979b), Freeman (1979), Schroy (1979), Schroy (1980), Wu and
Schroy (1979)).  Those commenting have also contributed to input
given to the EPA from industry and trade associations (Beale
(1980) , Schroy (1978)) .
                             VI-97

-------
                                          Monsanto Co.
                                          March 18, 1981
I.  Impact of Other Regulations

    A statement is made in the Background Information Document
    (USEPA 1980) that is central to the issue of the need for
    VOC Fugitive Emission regulations.  The statement involves
    a key issue because it is incorrect.  The statement is:

        "There are presently no federal regulations that
        specifically reduce emissions from synthetic organic
        chemical manufacturing plants.  However, some fugitive
        emission reduction is achieved by operating practices
        currently followed by industry and applicable state or
        local regulations." (USEPA (1980)- 3.3 Baseline
        Control)

    This statement and further discussion of Health and Safety
    Regulations clearly illustrate the EPA's misunderstanding
    of Occupational Safety and Health Administration (OSHA)
    regulations.   The OSHA regulation for control of
    contaminant levels in the workplace are in fact fugitive
    emission control standards.  The standards are designed to
    protect the worker from exposure to specific contaminants
    and they are performance standards.  However, the
    regulations do stipulate how the ambient levels should be
    achieved.

    The EPA appears to believe that the "...workers can be
    protected from high ambient VOC levels by:  1)   a
    reduction in the fugitive VOC emissions or 2) the use of
    special equipment (such as personal respTFators)  to
    isolate the worker from the emissions."  The error in this
    statement is the word or.  The OSHA regulations require
    the application of engineering controls and work practices
    to limit exposure.  Only when engineering controls and
    work practices are not available or are found to be
    economically impractable can personal protective equipment
    be used as the primary exposure control procedure.   (e.g.
    OSHA  (1971), OSHA (1974), OSHA (1978)).  The engineering
    controls considered by OSHA are identical to those
    described by the EPA in the Background Information
    Document (US EPA  (1980)).

    The engineering controls available to SOCHI facilities
    have in fact been one of NIOSH's primary concerns over the
    last few years.  As the technical organization that
    supports OSHA, the National Institute of Occupational
    Safety and Health's (NIOSH) concern for cor.trol technology
    should be considered by the EPA in its evaluation of the
    proposed VOC fugitive emission standards.   A few of the
    technical survey's made by NIOSH recently are as follows:
                           VT -98

-------
                                      Monsanto Co.
                                      March 18, 1981

    Engineering Control Technology Assessment for the
    Plastics and Resins Industry (Enviro Control Inc
    (1978))

    Symposium on Control of Workplace Hazards .in the
    Chemical Manufacturing Industry (NIOSH and CMA (1981)

    Control Technology Assessment of the Pesticides
    Manufacturing and Formulating Industry (SRI
    International (1980))

These studies have been summarized in both written reports
(Enviro Control Inc. (1978), NIOSH (1979), SRI
International (1980)) and through the use of Symposia for
Industry and Regulatory Personnel (NIOSH (1979), NIOSH and
CMA (1981),  SRI International (1980)).   The symposia have
been cosponsored with NIOSH by both industry and/or the
EPA.

The EPA's attempt to transfer technology and regulatory
concerns from the petroleum refining industry to the SOCMI
may in fact be the cause for not understanding the impact
of OSHA.  While the regulations established by OSHA have
little impact on fugitive emissions in the petroleum
refining industry they have a direct impact on the SCCMI
plants.   The reason for this difference is the
differences in the chemical, physical and biological
impact of the chemicals handled by the two industries.
SOCMI plants handle many more toxic and hazardous
chemicals than do petroleum refineries and are therefore
more frequently impacted by OSHA.  A comparison of several
petroleum refinery industry and SOCMI chemicals is given
in Table 1.   With chemicals the exclusion of oxygen and
explosive concerns dictate operating and design
practices.  While in SOCMI facilities, the toxicity of the
chemicals often control design and operating practices.

If fugitive emission concerns were analyzed for SOCMI
facilities instead of transferring concerns identified in
the petroleum refining industry, the EPA would in fact
identify the impact OSHA has on fugitive emissions in
SOCMI facilities.

An example of how OSHA regulations impact a SOCMI plant is
illustrated by the following case study.  The study was
carried out at a Monsanto plant to define not only the
number of sources that would be indicated as "leaking",
based on a reading of 10,000 ppm or over on an OVA-128
calibrated with methane in air, but also to obtain an
emission rate for the valves and pumps that showed
screening values form 600 to over 10,000 ppm.  This study
was completed for a chemical which is specifically
                      VI-99

-------
             Tc'ible 1
                   Comparison of Relative Hazards for API and SOCMI Chemicals
•o
o
Chemical
Name	
methane
ethane
propane
butane
pentane
n-hexane
nonane
decane
dodecane
acetic acid
acrylonitrile
chJ orobenzene
epichlorohydrin
ethylene
ethyl ether
vinyl acetate
vinyl chloride
                                        TLVR,
                                        ppm (v)
                                  (AGCIH 1980) or OSHA
                                              asphyxiant
                                              asphyxiant
                                              asphyxiant
                                              asphyxiant
"inert" gas
"inert" gas
"inert" gas
"inert" gas
TWA 8 = 600
TWA 8 =  50
TWA 8 = 200
No criteria
No criteria
TWA 8 =  10
TWA 8 =   2 OSHA
TWA 8 =  75
TWA 8 =   2
"inert" gas-asphyxiant
TWA 8 = 400
TWA 8 a  10
TWA 8 =   1 OSHA
Equilibrium Vapor Composition
with Pure Liquid @ 75°F in Air
(@ Specific Chemical), ppm  (v)
         1,000,000
         1,000,000
         1,000,000
         1,000,000
           644,350
           189,290
             5,325
             1,664
               168
            19,010
           133,830
            14,580
            19,995
         1,000,000
           673,780
           142,950
         1,000,000

-------
                                      Monsanto Co.
                                      March 18, 1981
regulated by OSHA.  A total of 17 pumps, 309 flanges and
221 valves were screened.  Of this number,! pump and 2
block valves all in light liquid service, were found to
leak, based on a 10,000 ppm reading on the instrument.
Based on the Radian report of instrument response factors
(DeBose and Harris (1981)) the concentration of material
would be about 9,700 ppm in order to provide a" 10,000 ppm
VOC indication.  Data taken from the SOCMI fugitive
emission study and this test effort shows the following
comparisons:

Location                Percent of Fittings Found to Leak

                               Pumps          Valves
                         (light liquid)    (light liquid)

Radian Refinery Study          23.0            10.0

Radian SOCMI Study              8.8             6.4

Radian AN data                  8.2             1.9

Monsanto Study                  5.9             0.9
This data supports very clearly the conclusion that the
chemical industry has fewer leaking components than a
refinery, when materials in terms oif workplace exposure
concentrations are regulated by OSHA .   The leak frequency
is being kept to a low level by existing maintenance
programs, dictated by existing federal regulations and
company programs.

The Monsanto program was also concerned with the
measurement of leakage rates.  Those fittings that were
found to produce an indication of a response on the
OVA-128 were tested using the leak rate procedures
detailed by Radian in many of their reports and work
plans.   The following data was obtained:
                       VI-101

-------
                                      Monsanto Co.
                                      March 18, 1981
                 VOC Emission Rate Comparison
Device
Valve
Valve
Valve
Pump
Pump
Flange
Process
Chemical
Acrylo-
nitr ile
Acrylo-
nitrile
Acrylo-
nitrile
Acrylo-
nitrile
Acrylo-
nitrile
Acrylo-
nitrile
Screening
Value
6,000
10,000
1,000
2,600
600
3,000
Line
Temp
(°P)
100
130
150
130
115
130
Line
Pressure
(psi)
60
90
50
90
60
90
VOC
Emission
Rate
grams/hr
1.8
8.7
0.2
0.2
11
0.3
Process
Chemical
Vapor
Pressure
mm Hg
183
348
516
348
155
348
At each fiting, duplicate samples were collected in
sampling bags.   The samples were analyzed by GC-FID for
the potential components in the streams.  Only one organic
compound was found and the analysis was done in terms of
ppm of this compound.  The specific chemical emission rate
has been converted to VOC (methane) emission rate based on
data in Table 4.  Calibration was done in the range of the
samples by recording the GC-FID response of samples
prepared using a diffision tube apparatus.  The results in
general show reasonable agreement and also provide an
indication of the precision that can be expected by the
material.

For comparions purposes, data in the CTG document (USEPA
(1981)) concerned with control of VOC fugitive emissions
in SOCMI, show that, based on the refinery study done
earlier the uncontrolled emission factor for light liquid
pumps is indicated to be 120 grams/hr and for light liquid
valves, 10 grams/hr.  These data clearly support the
judgment that has been made by the chemical industry --
that is, the refinery and chemical industry are not
similar and that the emissions in SOCMI, both form a
frequency point of view and by emission rate are lower
than expected by application of refinery data.

Based on our understanding of OSHA's impact on fugitive
emissions, we believe the proposed EPA regulation is
unnecessary.   If in fact, the EPA does not agree with our
assessment the proposed regulation should be modified to
that of a performance standard to make it compatible with
                      VI-102

-------
                                      Monsanto Co.
                                      March 18, 1981
OSHA regulatory actions.   This change would at least
minimize the negative impact of duplicate regulatory
actions.

A performance standard would also provide the flexibility
of control strategies implied by the "bubble concept" of
emission control.   A specification standard, by its rigid
emission control requirements, would not.
                      VI-103

-------
                                           Monsanto Co.
                                           March 18, 1981
II.  Equipment Specification vs Performance Standards

     The proposed standard is based on the requirement to
     utilize specific engineering controls for reduction of
     fugitive emissions from pump seals, sampling systems and
     compressor seals.  An engineering specification is also
     made for relief valves  That topic was discuss'ed earlier
     as a safety concern.  The specification standards proposed
     are essentially based on the impact assessment made for
     the petroleum refining industry and do not recognize the
     impact of this specific stipulation on the SOCMI plants.
     The limited knowledge of the chemical industry is
     illustrated by the report prepared by Hydroscience
     (Erikson and Kalcevic (1979)).  This simplified review is
     but one example of limited knowledge of an industry as
     broad as the SOCMI.

     To specify a single solution for the preceived fugitive
     emission concerns in an industry as complex as SOCMI will
     lead to significant problems.  Unlike the petroleum
     refining industry, SOCMI facilities handle a wide variety
     of chemicals with a -wide variety of chemical, physical and
     biological properties.  A single solution for a group of
     chemicals as different as those presented in Appendix E,
     of the proposed standard, is not sound technically.  The
     behavior of chemicals as acutely toxic as hydrogen
     cyanide, as chronically toxic as vinyl chloride and
     corrosive as acetic acid call for different control
     equipment.  In fact, the wide variety of pump seals and
     sampling systems are in part due to the need for different
     solutions to different concerns.  While control of
     fugitive emissions is necessary to protect the workplace
     and ex-plant boundary areas, some understanding is
     necessary of the impact of the chemical on the control
     device prior to its application.  This does not mean to
     say that an added regulation is needed, in fact, this
     evaluation of concerns can not be made for a list of
     chemicals , as different as those given in Appendix E in a
     regulation.  The concerns must be handled on a case by
     case basis by the system designers.  To deal with the
     varied behavior of different chemicals, if a standard is
     issued, a performance standard is needed such that the
     designers of new facilities can deal with fugitive
     emissions in a technically sound manner.

     The regulation, as proposed, calls for control of
     fugitives by application of solutions already practiced in
     SOCMI facilities for acutely and chronically toxic
     chemicals.  The types of controls available today are in
     fact more sophisticated and innovative than those proposed
     by the EPA.  A comparison of the proposed and available
     solutions is summarized below.
                            VI-104

-------
                                  Monsanto Co.
                                  March 18, 1981
Sampling Systems

Flow through sampling systems were developed as gas
sampling systems to insure collection of
representative samples.  While this technique will
work for some liquid services it is not the most cost
effective or safe sampling procedure for all liquids.
Loss of liquid from between valve connections and
pressure at the valves are two concerns which are not
considered significant when sampling gases.

A better method is the collection of the desired
volume of liquid from a sampler that reduces the
sample pressure to atmospheric pressure as it is
collected.  The available technology for this type of
sampling has been discussed by Bruce Lovelace of Dow
several times (Lovelace (1979), NIOSH (1979), NIOSH
and CMA (1981)) .

Under the proposed regulation these limited volume low
emission  (zero for some units where displaced vapors
are returned to 'the process line)  sampling techniques
could not be permitted.  Based on a performance
standard they could be utilized.

Pump Seal

The proposed specification standard, calling for
double mechanical seals and buffer fluids, is another
example of limited understanding of chemical plant
safety, health and environmental concerns.  With the
proposed definition of light liquid service, the
number of buffer fluids available to the chemical
industry is extremely limited.  Concerns for product
quality are such that the buffer fluid must be
compatible with the process regardless of the seal
design.  Precautions are taken to prevent product and
buffer fluid mixing.  Product purity specifications,
chemical reactivity and catalyst decativation are but
a few of the basic design considerations that must be
answered when specifying a pump seal.

Chemical reactivity in the seal cavity and on the
mating seal faces are also a design concern.  For
example monomers, such as acrylonitrile, will
polymerize on the hot seal face surfaces, causing
failure of the seal and excessive emissions.  These
seal failures also result in excess maintenance
requirements and the attendant vapor emissions during
repair procedures.  To minimize vapor losses from
these seals, flush fluids are used with single
mechanical seals incorporating a throttle bushing.
The flush fluids are often process fluids to allow
                   VI-105

-------
                                  Monsanto Co.
                                  March 18, 1981
injection back into the process and eliminate product
contamination concerns.   The flush fluid is cooled to
very low temperatures to allow sensible cooling of the
seal.  The low temperature seal operation reduces
polymerization/maintenance concerns and vapor
emissions by reduction in fluid vapor pressure.

This type of innovative emission control is not
compatible with the specification standard even though
it is one method Monsanto uses to comply with the
2 ppm OSHA eight hour exposure regulation.  This is
just one more reason why we feel a performance
standard should be used, if in fact a new standard is
found to be necessary.

Zero Leak Equipment

The regulation as proposed allows use of zero leak
equipment as an alternative to the engineering
controls established in the specification standard.
The performance of these equipment are cited as 100%,
when in fact they normally contain static seals which
will not permit 100% control.  For instance, bellows
valves are suggested for use in SOCMI facilities.
While these values have been used in blocking valve
service in the nuclear field, they have not been
widely applied in the chemical industry.  Concerns
with corrosion and mechanical failure have yet to be
resolved for many of the chemicals in the Appendix E
list.  Comments and concerns for leak free control
technology were submitted to the EPA by CMA (Beale
(1980)).  Although these comments, submitted to K. C.
Hustuedt, dealt with benzene,they could apply to all
VOC's listed in Appendix E.  These comments should be
taken into consideration during review and revision of
this standard.
                   VI-106

-------
                                            Monsanto Co.
                                            March 18, 1981
III.   Definition of Light Liquid Service

      The proposed definitions of light liquid service, as
      described in paragraph 60.481 and 60.485(c), are not
      consistent with testwork referenced in the Background
      Information Document (US EPA (1980)).  Specifically,
      Wetherold et al (1979)  reported light liquids to be
      streams that contained the following components as liquids
      (as determined by the stream conditions within the process
      lines):

      - C2 through Cg hydrocarbons

      - Naphtha

      - Light Distilate

      - Low molecular weight aromatic hydrocarbons

      - Miscellaneous streams

      The streams were classified as a light liquid based on the
      identity of the most volatile stream component present at
      a concentration of 20% or more.

      In defining light liquid Wetherold et. al. also defined
      heavy liquid as any compound with a vapor  pressure less or
      equal to that of kerosene (eg.  0.3 k?a or  0.0435 psia or
      2.25 mm Hg).  As proposed in the NSPS the  definition of
      light liquid is given in terms of materials that are not
      heavy liquids.  This change does not properly recognize
      the basic differences in the types of feed stocks handled
      by SOCMI versus those handled by the petroleum
      industry.

      The petroleum refining industry uses and produces
      materials that are mixtures with an understandable
      variation in physical,  chemical and biological
      properties.  Some SOCMI processes also utilize feed stocks
      that are  mixtures, however, the majority  of SOCMI
      processes utilize "pure" components to produce "pure"
      products.  A summary of specific chemicals handled by the
      two industries is given in Table 2. The difference is that
      the petroleum industry handles mixtures of some of these
      components.

      Based on our review of the fugitive emissions survev work,
      available in the literature, which deals with specific
      chemicals (Blacksmith et. al (1980), Tierney et. al. (1978
      a, b &  c), Summerfieid (1980),  Fernandes (1973)  and Bierd
      et.al.  (1977)), it is our judgment that only chemicals
      with very high vapor pressure (i.e. low boiling points)
      are of  concern when evaluating fugitive emissions based on
      emission rate data.  An analysis of these  data is
      discussed below.
                             VI-107

-------
              Table  2   Physical/Chemical Characteristics of SOCMI  and API  Chemicals
o
c»
                                                                   Carbon
Vapor Pressure
Chemical
Name
References
methane
ethane
propane
butane
pentane
hexane
vinyl acetate
acrylonitrile
crylohexane
toluene
chlorobenzene
x-xylene
nonane
decane
dodecane
Molecular Boiling Point Heat Values Atoms @75°F (2:
Weight °C K BTU/lb mole per Molecule mmHg
(1) (2)
16.04
30.04
44.10
58.12
72.12
86.18
86.09
53.06
84.16
92.14
112.99
106.17
128.26
142.29
170.34
(1) (2)
-161.5
- 88,6
- 42.1
- 0.5
36.1
68.7
72.7
• 77.3
80.7
110.6
132.3
139.1
149.1
172.0
214.8
Combustion
(3) (4)
345.2
614.3
879.3
1143.7
1407.7
1672.1
—
757.4
1587.0
1622.7
1337.5
1881.9
2465.5
2729.9
3258.9
Vaporization
(3) (4)
3.52
4.06
6.62
7.06
11.36
15.57
15.90
14.00
13.98
16.09
18.10
18.40
19.97
22.10
22.66

1
2
3
4
5
6
4
3
6
7
6
8
9
10
12
(1) (2)
184,800
28,670
6,870
1,760
490
144
109
102
93
27
11
8
4
1.3
0.13
J.9UC)
psia

3,570
554
133-
34
9.5
2.8
2.1
2.0
1.8
0.5
0.21
0.15
0.08
0.03
0.003
     (1)   Dean  (1973)
     (2)   Dean  (1979)
     (3)   Perry and Chilton (1973)
     (4)   Weast (1975)

-------
                                      Monsanto Co.
                                      March 18, 1981

The study of pump seal emissions reported by Summerfield
is the only study involving pure components and a specific
pump equipped with a single mechanical seal, or packed
gland.  The data from Summerfield's study, -run at ambient
temperature, illustrates the impact of volatility on
fugitive emission rate.  For chemicals with boiling points
above 20°C there was little in the way of emissions.  -The
average emission rate for twenty tests was 2.8 grams per
hour.  Ninety percent of the pumps, tested, with chemicals
that boil at a temperature above 20°C, had emission rates
of less than 4 grams per hour.   These results were
consistent with the results reported by Bierd et. al.
(1977) for benzene.  Unlike the data reported by the EPA
for mixtures (Wetherold et. al. (1979)) a direct
correlation of emission rate and the physical properties
of the material being pumped was identified.   These data
are significant in that the 4 gram per hour emission rate
is one third of that reported by the EPA for alternative A
(i.e. 0.13 x 120 grams per hour = 15 grams per hour)

To better understand- the relationship of volatility and
emission rate several chemicals were modelled using an
evaporation model developed for evaluation of spills (Wu
and Schroy  (1979)).   This comparison of emission rates,
given in Table 3, shows the importance of the cause of
emissions from pump seals (Schroy (1978))  (i.e. cooling of
seal faces).  A review of the results shows the
significant difference between chemicals of concern in the
petroleum refining industry and SOCMI.

Based on the data analysis presented in Table 3,  the data
developed by Summerfield and Bierd et. al. and the data
prepated by the EPA, we feel the definition of light
liquid service should be changed as follows:

(c) A fugitive emission source is in light liquid service
    if the following conditions apply:

    (1)  The vapor pressure of one or more components,
        present in concentrations greater than 20% by
        weight, is greater than 760 mm Hg at operating
        temperature.  Vapor pressures may be obtained from
        standard reference texts or may be determined by
        ASTM Method D-2879.

    (2)  The fluid is a liquid at operating conditions.
                       VI-109

-------
                  Table 3.   Comparison of Simulated Pump Seal Emissions Rates
    Chemical
    Name
                Emission Rate, grams per hour @ 75°F Seal Cavity Temperature
o
methane
ethane
propane
butane
pentane
hexane
vinyl acetate
aeryIonitrile
cyclohexane
toluene
chlorobenzene
m-xylene
nonane
decane
dodecane
as specific
chemical
(1)

137
71
24
9
3
3
1
2
0.5
0.25
0.2
0.1
0.03
0.004
as VOC w/flame
ionization
detector (methane)
(2)
200
105
56
19
8
2
1
5
1
0.7
0.6
0.06
0.3
0.05
as VOC w/catalytic
oxidation detector
(methane)
(2)
193
167
37
14
4
0.6
3
3
0.2
0.4
0.04
0.01
0.2
_
as VOC w/flame
ionization
detector (hexane)

70
37
20
7
3
0.7
1
2
0.4
0.2
0.2
0.02
0.1
0.02
as VOC w/catalytic
oxidation detector
(hexane)

136
118
26
10
3
0.4
2
2
0.1
0.3
0.03
0.01
0.1
__
    (1)  Emission rate for a 100mm diameter shaft seal.  Simulated as a pool of liquid 5 sq. in. in
        surface area for identical heat input conditions (Wu and Schroy (1979)).
    (2)  Based on data presented in Table 4.  Average values for VOC (methane) per actual organics
        times actual organic rate.
    (3)  Corrected by ratio of hexane to methane value from Table 4

-------
                                           Monsanto Co.
                                           March 18, 1981
IV.   Definition of Volatile Organic Compounds (VOC)

     The proposed definition for  VOC's is unclear.   As
     originally proposed in the draft regulations the
     definition was too  broad.   However,  the changes in the
     proposed definitions,  utilizing an analytical  procedure,
     do not reflect the  capabilities of the reference method
     protocols to measure contaminant levels.

     Specifically, Reference Method 21 allows use of several
     different analytical detectors.  One detector, the
     photoionization device, will respond to most chemicals
     listed in Appendix  E,  but  the detector will not respond to
     the calibration gas called for in the standard (i.e.
     methane) .  In our opinion, if the unit cannot  be
     calibrated, it should not  be used for measurement of
     emissions.  The detector functions based o.n the principle
     of electron excitation, and  since saturated straight  chain
     hydrocarbons (eg. methane, ethane, propane  and hexane)  do
     not react to provide free  electrons  under photoionization
     excitation the detector will not register the  presence of
     these materials.

     The definition is also vague in that it is  meaningless for
     those chemicals which participate in atmospheric
     photochemical reactions but  cannot be measured by all
     procedures allowed  under Reference Method 21 (eg. carbon
     tetrachloride cannot be detected by  the flame  ionization
     unit permitted under Reference Method 21) .   We feel the
     regulations should  recognize the technical  limitations of
     the tools to be used for their enforcement.

     We recommend the definition  for VOC's be changed as
     follows:

             "Volatile Organic  Compounds" means  any organic
             compound, which participates in atirospheric
             photochemical reactions and  is measurable by  the
             applicable  test methods described in Reference
             Method 21 which can  be calibrated by a saturated
             straight chain hydrocarbon.

     These changes in the definition are  technically consistent
     with the work completed by the EPA and its  contractors on
     emission monitoring.  This judgment  of consistency of this
     definition and-the  contractors'  work is based  on the  EPA's
     contractors'  use of tne flame ionization (I.e. carbonium
     ion detector)  and catalytic  oxidation units (i.e. heat of
     combustion detector).   (DeBose and Harris <1980),
     Whetherold et.  al.   (1979),  Smith (1979), Brown et. al.
     (1980)).  A summary and comparison of these data are  given
     in Table 4.  It should be  noted that when rhe  Radian  data
     are compared to the theoretical response of the two
                           VI-111

-------
                                      Monsanto Co.
                                      March 18, 1981
detectors a very poor relationship is found.  This is
illustrated by the inconsistent reponse of nonane and
decane.

The change from a hexane to methane calibration gas is
also of concern because of its impact on conversion of
emission data.  The calibration gas for the pe'troleum
refinery studies was hexane (Honerkamp et. al. (1979)r
Wetherod et. al. (1979), Rosebrook (1977), Fernanandes
(1978), Wallace (1979)).  However, although the EPA has
propsoed to transfer control technology from the petroleum
refinery industry to the SOCM industry it has adopted a
new calibration gas.  This one change will result in
reporting leaks, that by the criteria established for the
refinery industry would not be classified as leaks.  (e.g.
a 10,000 ppm methane value would be a 3,520 ppm hexane
value based on the data presented in Table 4.)  This is an
unacceptable change in that the actual organic vapor
emission rate would be acceptable in an API facility and
not in a SOCMI facility.  A single calibration gas should
be used in all industries.  Since the majority of
available data are based on a hexane standard, we
recommend hexane be used for this regulation if a new
regulation is found to be necessary.
                       VI-112

-------
  Table 4
Ratio of Observe vs. Actual Organic Concentration
Chemical
      VOC @ Methane eq.
      Flame lonization
  Theory  Ref .1  Ref 2
     VOC @ Methane Eq.
    Catalytic Oxidation
Theory  Ref 1  Ref 2
Methane
Ethane
Propane
Butane
Pentane
Hexane
Vinyl Acetate
Acrylonitrile
Cyclohexane
Toluene
Chlorobenzene
m-xylene
Nonane
Decane
Dodecane
1.0
2.0
3.0
4.0
5.0
6.0
4.0
3.0
6.0
7.0
6.0
8.0
9.0
10.0
12.0
1
1
1
2
1
2
0
0
2
2
2
2
0
11

.0
.18
.82
.0
.92
.44
.79
.97
.13.-
.56
.63
.50
.65
.1
-
1.
1.
1.
2.
2.
3.
0.
0.
2.
3.
2.
3.
0.
0
75
14
63
38
23
76
96
78
03
78
33
62













00
w
Observed
Flame


Methane
1
1
2
3
4
4

2
4
4
3
5
7
7
9
.0
.86
.55
.31
.08
.84
-.
.19
.60
.70
.87
.45
.14
.91
.44
1
1
1
1
1
1
0
3
1
0
1
0
0
6

.0
.45
.67
.59
.59
.45
.17
.49
.43
.37
.14
.17
.09
.25
-
ea . /Theoretical
lonization
Ref 1
Methane
Ethane
Propane
Butane
Pentane
Hexane
Vinyl Acetate
Acrylonitrile
Cyclohexane
Toluene
Chlorobenzene
m-xylene
Nonane
Decane
Dodecane
(1) Brown et.















al. (1S80)
1
0
0
0
0
0
0
0
0
0
0
0
0
1



.59
.61
.5
.38
.41
.20
.32
.36
.37
.44
.31
.07
.11

(2)















Ref 2
1
0
0
0
0
0
0
0
0
0
0
0
0



.88
.38
.66
.48
.54
.19
.32
.46
.43
.46
.42
.07
00

Dubose and















Harri
Catalv
tic
Ref 1

0.78
0.65
0.48
0.39
0.30
—
1.59
0,31
0.08
0.29
0.03
0.01
0.79

s (195















!1















)
1.0
1.37
3.03
1.47
1.61
1.39
0.25
2.70
1.39
0.43
1.14
0.28
0,18
5.0
-
Methane ea
Oxidation
Ref 2

0.74
1.19
0.44
0.39
0.29
_
1.23
0.31
0.09
0.29
0.05
0.03
0.80


                                vi-.ns

-------
                                          Monsanto Co.
                                          March 18, 1981
V.  Data Inconsistency

    The data available in the literature are inconsistent,  for
    an as yet unidentified reasons.  Some of the possible
    reasons (eg. definition of light liquid fluid) have been
    discussed earlier.  However, these inconsistaacies should
    be identified, explained and only then should the data  be
    used to define emission concerns and/or regulations.

    One major concern in data interpretation is the method
    used to convert field emission rate test results  (i.e.
    bagging tests) into emission rate values.  Several EPA
    contractors have reported their sampling and data
    conversion procedures in a variety of documents (Smith
    (1979), Tierney, et. al.  (1978), Radian Corp. (1979),
    Hughes, et. al. (1978)).  Although the procedures for
    estimating emissions are given in each report they are  not
    consistent.

    The procedures call for a conversion factor k to correct
    units for variables in the emission rate equation.
    Different values of k are reported by Radian in separate
    documents.  These values differ from the value used by
    Monsanto Research Corporation.

    Specifically the formula for calculation of emission rate,
    as given by Smith of Radian Corporation (Smith (1979)  is
    as follows: :

        EH = K(QPMa/T) (Cs-Ca)

        Where:

        EH = hydrocarbon emission rate, Ibs/hr

        Q  = flow rate of gas through sample train,  actual
             cubic feet per minute

        P = sampling system pressure at dry gas meter, pisa

        Ma = molecular weight of air/hydrocarbon mixture,
             effectively the molecular weight of air

        T = Temperature at dry gas meter,  °R {460° + °F)

        Cs= concentration of hydrocarbon in the gas sample
             from the sampling train, ppm (wt)

        Ca= concentration of hydrocarbon in the ambient air,
             ppm  (wt)

        k = units conversion constant
                           VI-114

-------
                                      Monsanto Co.
                                      March 18, 1981

To determine which values of k were correct calculations
were made for the various unit systems.  These
calculations are shown in Figure 1.  A comparison of
reported and calculated values is presented in Table 5.

Based on these calculations it is obvious Radian has used,
in two instances, conversion factors (k)  that were a
factor of five high.  If in fact they used these erroneous
factors in all their calculations, the reported emission
rates are a factor of five higher than actual.

Evaluation of pump seal emission data,  available in the
public literature, suggests that this error exist
throughout the data base.  A comparison of available data
is given in Table 6.
                      VI-115

-------
JFigure 1
Determination of Units conversion factor
(a) lQ,ae£gTVx P,
     x  10"6 Ibg/abB*    60

                     x
                         hr     x 10. 73
     where 10.73 cf x psia ^   gas constant

                   OR-lb moles
                       ««



     k  =  10"6 x 60    5.592 x  10~6
            10.73
(b)  lQ,ja*rf5rx in^H M, £at	\  x}AC,

    \     T ,^K"         I X   I    ib—ffioirsij   (           J
    k                        V            '


     x 10~6 Ibs/Ob-    60 jaJnf   SR* lb-mo±gs"

            S£x*frft)  X    hr     21.85 -ef  i



     where 21.85  cf  in Her abs

                  °R  x Ib moles



     k = 10"6 x 60 _ 2.746 x 10"6

           21.85      —:	


    <                _-?  (         1   (
(c)  JQ,agg;rx P, .i-a-Hay  1M Ib	£   AC,

    1       T, °R     \ X   JJb-mcrr^) x
    »                  J  \.         J   \


     x 10-6  IJa-mcrres/Lb-moTe  „   60 a&n  „   QR x


                  BP



     k = 2.746 x  10
                                           x
                                       hr  A  21.85


                    -6
                          VI-116

-------
           Table 5   Emission Rate Calculation Factors^
Author
(Contractor)
Smith (1979)

Langley and
Wetherold (1981)

Radian (1979)

MRC (private
communication)
Pressure   Concentration
psia

in Hg


psia

in Hg
ppm(w)

ppm (w)


ppm (w)

ppm(v)
Units Correction Factor
     k           k
             Correct
              Value

             5.592xlO~6

             2.746xlO-6
Reported


2.74x10-5

2.7xlO~6
 2.74x10-5   5.59x10-6

 2.7xlO-6    2.746xlO-6
                            VI-117

-------
            Table 6   Emission Rate for Single Mechanical
                       Seal in "Light Liquid" ServiceU)
                         VOC  (hexane) grams per  hour
Date
Source
USEPA (1980)                      120
Summerfield  (1980)                  4

Monsanto Study  (Section II)
  Acrylonitrile                  10.6

Bierd et. al. (1977)
  Benzene                           6

Tierney et. al.  (1978)                , .
  Monochlorobenzene                29 * '

Hughes et. al.  (1979)
  Monochlorobenzene               9.7 to 29

Avg. of values except USEPA  (1980)   11.9


(1)  Based on definition proposed by EPA

(2)  Corrected from chemical to VOC (hexane) based on Table 4
                             VI-118

-------
                                            Monsanto Co.
                                            March 18, 1981
References
Beale (1980)
      Beale,  J.  S.,  Letter to K.  C.  Hustuedt, US EPA, "CMA
      Comments on Leak Free Technology for Control o-f Benzene
      Fugitive Emissions," February  12,  1980, Chemical
      Manufacturers  Association,.  Washington, B.C.

Bierd, et. al. (1977)
      Bierd,  A;  Stoekel,  A; Sinn, R. and Kremer, H., "Leckraten
      Von Dichtelementen," Chemie Igenieur Technik, Vol 49, No.
      2 (1977) pp 80-95.

Blacksmith et. al. (1980)
      Blacksmith, J.  R.,  Harris,  G.  E. and Langley, G. L.,
      Frequency of Leak Occurrence for Fittings in Synthetic
      Organic Chemical Plant Process Units, Radian Corporation
      tor US EPA"!  Industrial Environmental Research Laboratory;
      Office of Environmental Engineering and Technology;
      Research Triangle Park, N.C.,  September, 1980.  (Contract
      No. 68-02-3171-Task 001, EPA Program Element No. 1AB604.)

Brown et al (1980)
      Brown,  G.  E.,  DuBose, D. A., Phillips, W. R. and Harris,
      G. E.,  Project Summary - Response  Factors of VOC Analyzers
      Calibrated with Methane for Selected Organic Chemicals,
      Radian Corporation for US EPA, Office cf Environmental
      Engineering and Technology, Industrial Environmental
      Research Laboratory, Research  Triangle Park, N.C.,
      September 30,  1980  (Contract #68-02-3171-Task 1).

Crocker (1979a)
      Crocker, B. B., "Removal of Hazardous Organic Vapors from
      Vent Gases," APCA/WPCF Speciality  Conference;  Proceedings
      Control of Specific  (Toxic) Pollutants, APCA/WPCF,
      Gainesville, Florida, February 13-16, 1979, pp 360-376.

Crocker (1979b)
      Crocker, B. B., "Capture of Hazardous Emissions",
      APCA/WPCF Speciality Conference;  Proceedings Control of
      Specific (Toxic) Pollutants, APCA/WPCF, Gainesville,
      Florida, February 13-16, 1979, pp 415-433.

Dean  (1973)
      Dean, J. A. (editor), Lange's  Handbook of Chemistry, llth
      Edition, McGraw Hill Book Company, New York, New York
      1973.
Dean
(1979)
 Dean,  J.
 Edition,
 1979.
               A. (editor),  Lange's Handbook of Chemistry, 12th
               McGraw Hill Book Company, New York, New Yock,
                             VI-119

-------
                                            Monsanto Co.
                                            March 18, 1981
OSHA (1974)
      OSHA/  "Exposure to Vinyl Chloride - Occupational Safety
      and Health Standards," Federal Register Vol 39, No. 194
      Friday, October 4, 1974, pp 35890-35d98.

Perry and Chilton (1973)
      Perry, R. H. and Chilton, C. K.  (editors), Chemical
      Engineers' Handbook, 5th Ed., McGraw-Hill Book Company,
      New York, New York, 1973.

Radian Corp. (1979)
      Radian Corp, "Test Plan for Control of Fugitive Emissions
      from the Synthetic Organic Chemical Manufacturing
      Industry," Oct. 23, 1979, prepared for Robert C. Weber, US
      EPA, by Radian Corporation, Austin, Texas (Contract
      168-03-2775-04).

Rosebrook (1977)
      Rosebrook, D. D., "Fugitive Hydrocarbon Emission,"
      Chemical Engineering, 8£ (22) pp 143-149, Oct. 17, 1977.

Schroy  (1978)
      Schroy, J. M., Technical Analysis given to Dale Denny  (US
      EPA) and Leigh Short  (EPA Consultant), "Emissions from
      Pump Seals," Pg 1-7, Nov. 7, 1979, AIChE National Meeting
      Miami, Florida.   (Technical Analysis also given to Bruce
      Tichenor US EPA and Atley Jefcoat US EPA.)

Schroy  (1979)
      Schroy, J. M., "Prediction of Workplace Contaminant
      Levels," Symposium on Control Technology in the Plastics
      and Resins Industry, February 26-28, 1979, Atlanta,
      Georgia.

Schroy  (1980)
      Schroy, J. M. "Conflicts and Constraints:  Technology Gaps
      Between Critical OSHA and EPA Concerns," Professional
      Safety, November 1980, pp 26-29.

Smith  (1979)
      Smith, D. D., "Field Sampling," Sympsoium on Atmospheric
      Emissions from Petroleum Refineries , U.S. Environmental
      Protection Agency,Industrial Environmental Research
      Laboratory, Sheraton Crest Inn, Austin, Texas, Nov. 5 and
      6,  1979  (hosted by Radian Corporation)

SRI  International  (1980)
      SIR International, Control Technology Assessment of the
      Pesticides Manufacturing and Formulating Industry, us
      Department of Health, Education and Welfare; Public Health
      Service;  Center  for Disease Control; National Institute
      for Occupational Safety and Health; Division of Physical
      Sciences and Engineering, Cincinnati, Ohio, May 1980.
                             VI-120

-------
                                            Monsanto Co.
                                            March 18, 1981
      (Contract No.  210-77-0093)  (In printing)  Report
      distributed at Symposium held by NIOSH and EPA in St.
      Louis,  MO, December 2 and 3,  1980.

Summerfield (1980)
      Summerfield,  J., "Extracts From B.H.R.A.  Seal-Emission
      Project Report,  Vapor Emissions from Rotary Seals BHRA
      Meeting , London,  England,  March 13, 1980.

Tierney,  et al  (1978)
      Tierney, D. R.,  Khan, Z. S.,  Hughes, T. W., Source
      Assessment; Fugitive Hydrocarbon Emissions - Testing of
      MonochloroBenzene  Manufacture, US EPA, Office of Energy
      Minerals Industry, Industrial Environmental Research
      Laboratory, Research Triangle Park,  N.C.  (Contract
      #68-02-1874,  June  1978).

Tierney,  et al  (1978)
      Tierney, D. R.,  Mote, L. B.,  and Hughes,  T. W., Source
      Assessment; Fugitive Hydrocarbon Emissions from
      Petrochemical Plants -- Summary Report,  Monsanto Research
      Corporation for  the US EPA Office of Research and
      Development,  EPA-600/4-78-004, April 1978.

Tierney (1978c)
      Tierney, D. R.,  Khan, Z. S.,  and /Hughes, T. W.,
      "Measurement of Fugitive Hydrocarbon Emissions From
      Petrochemical Plants," Third  Symposium on Fugitive
      Emissions;  Measurement and Control, San Francisco,
      California Oct.  23-25, 197b.


USEPA (1978)
      OSEPA,  Guideline Series - Control of Volatile Organic
      Compound Leaks from Petroleum Refinery Equipment, USEPA,
      Emission Standards and Engineering Division,Chemical and
      Petroleum Branch,  Office of Air Quality Planning and
      Standards, Research Triangle  Park,  N.C.,  June 1978 (EPA
      450/2-=8-0=6;  OAQPS No 1.2-111)

USEPA (1980)
      US Environmental Protection Agency,  VOC Fugitive Emissions
      in Synthetic Organic Chemicals Manufacturing Industry -
      Background Information for Proposed Standards, Emission
      Standards and Engineering Division,  USEPA, Office of Air,
      Noise and Radiation, office of Air  Quality Planning and
      Standards, Research Triangle  Park,  North Carolina, March
      1980 (Draft) .
USEPA (1981)
      USEPA,  Guideline Series-Control of Volatile Organic
      Fugitive Emissions from Synthetic Organic Chemical7
      Polymer, and Resin Manufacturing Equipment USEPA,Office
      of Air  Quality Planning and Standards Research Triangle
      Park, NC, Jan. 1981  (Preliminary Draft).
                             VI-121

-------
                                            Monsanto Co.
                                            March 18, 1981
Vincent (1979)
      Vincent, G. C., "Rupture of a Nitroaniline Reactor," Loss
      Prevention, Vol. 5, AIChE 19  .

Wallace (1979)
      Wallace, M. J. "Controlling Fugitive Emissions-," Chemical
      Engineering, August 27, 1979, pp 78-92.

Weast (1975)
      Weast, R. C., (editor), Handbook of Chemistry and Physics,
      56th Edition, CRC Press, Cleveland, Ohio, 1975.

Wetherold, et al  (1979)
      Wetherold, R. G. and Provost, L. P., "Emission Factors and
      Frequency of Leak Occurrence for Fittings in Refinery
      Process Units, Radian Corporation for the US EPA Office of
      Research and Development, EPA 600/2-79-044, February
      1979.
Wu and Schroy (1979)
      Wu, J. M.C. and
Schroy, J. M.,  "Emissions from Spills,
      APCA/WPCF Speciality Conference;  Proceedings Control of
      Specif ic~(Toxic) Pollutants, APCA/WPCF, Gainesville,
      Florida, February 13-16, 1979, pp 377-393.
                            VI-122

-------
                                            Monsanto Co.
                                            March 18, 1981

DuBose and Harris (1980)
      DuBose, D. A., Harris, G. E., Response Factors of VOC
      Analyzers At A Meter Reading of"^10,QOQ ppmv for Selected
      Organic~Chemicals,  Radian Corporation for US EPA, Chemical
      Processes' Branch,  Industrial 3nvironmental Research
      Laboratory, Research Triangle Park, N.C., February 5, 1981
      (Contract #68-02-3171-28).

Ellison (1981)
      Ellison,  E. D., "Statement Made on Behalf of the Synthetic
      Organic Chemical Manufacturers Association at EPA's Public
      Hearing on Proposed New Source Performance Standards for
      VOC Fugitive Emissions Sources in the Synthetic Organic
      Chemicals Manufacturing Industry," Raleigh, North
      Carolina, March 3,  1981

Enviro Control, Inc. (1978)
      Enviro Control Inc., Engineering Control Technology
      Assessment for the  Plastice and Resins Industry, US
      Department of Health,  Educat: ,on and Welfare;Public
      Health Service; Center of Disease Control; National
      Institute for Occupational Safety and Health;   Division of
      Physical Sciences and Engineering, Cincinnati,  Ohio, march
      1978 (Contract No.  210-76-0122).

Erikson and Kalcevic (1979)
      Erikson,  D. G. and  Kalcevic, V. Emission Control Options
      for the Synthetic Organic Chemicals Manufacturing
      Industry, Hydroscience Inc. prepared for US SPA Office of
      Air Quality Planning and Standards, March 1979  (Draft).

Fernandes (1978)
      Fernandes, S. R.  (compiler), Proceedings:
      Symposium/Workshop  on Petroleum Refining Emissions,  Radian
      Cokrp for US EPA, Office of Research and Development,
      Jekyll Island, GA,  April,1978.   (EPA 600/2-78-199,
      September, 1978).Freeman (1979)

Freeman (1979)
      Freeman,  R. A., "Stripping Hazardous Chemicals  from
      Surface Areated Waste Treatment Basins," APCA/WPCF
      Speciality Conference;  Proceedings Control of  Specific
      (Toxic) Pollutants, APCA/WPCF,  Gainesville, Florid?.,
      February 13-16, 1979,  pp 464-481.

Honerkamp, et al (1979)
      Honerkamp, R.  L., Provost, L. P.,  Sawyer, J. W; and
      Wetherold, R.  G. , Valve Screening Sue!  at Six San
      Francisco Bay Area  Petroleum Refineries; Radian
      Corporation for Oil Company CcTnsordium,  February 6,  1979.
                            VI-123

-------
                                            Monsanto Co.
                                            March 18, 1981
Hughes, et. al. (1980)
      Hughes, T. W.,  Tierney, D. R.,  and Khan, Z. S., "Measuring
      Fugitive Emissions from Petrochemical Plants," Chemical
      Engineering Progress, August 1979, pp 35-39.

Langley and Wetherold (1981)
      Langley, G.  J.  and Wetherold, R.  G. Evaluation of
     .Maintenance for Fugitive VOC Emissions Control, Radian
      Corporation for US EPA Industrial Environmental Research
      Laboratory,  Cincinnai,  Ohio, February 17. 1981.

Lovelace (1979)
      Lovelace, B. G.,  "Safer Design  for Manual Sampling of
      Liquid Process  Streams," 86th National Meeting of the
      American Institute of Chemical  Engineers., April 1-5, 1979,
      Houston, Texas.

Martin  (1981)
      Martin, J. D.,  "Statement on Behalf of the Chemical
      Manufacturers Association at EPA's Public Hearing on
      Proposed Rulemaking of New Source Performance Standards
      for Volatile Organic' Compound Fugitive Emissions Sources
      under Section III of the Clean  Air Act, Research Traingle
      Park, N.C.,  March 3, 1981.

NIOSH  (1979)
      NIOSH, Proceedings from the NIOSH Sponsored Symposium on;
      Control Technology in the Plastics and Resins Industry,
      U.S.  Department of Health, Education and Welfare; Public
      Health Service;  Center for Disease Control; National
      Institute for Occupational Safety and Health; Division of
      Physical Sciences and Engineering, Atlanta, GA, Feb.
      27-28, 1979-   (In Printing).

NIOSH and CMA  (1981)
      NIOSH and CMA,  Symposium on Control of Workplace Hazards
      in the Chemical" Manufacturing industry, Chemical
      Manufacturers Association and National Institute for
      Occupational Safety and Health, Philadelphia,
      Pennsylvania, March 11 and 12,  1981.

OSHA (1978)
      Occupational Safety and Health  Administration,
      "Occupational Exposure to Acryionitrile  (Vinyl Cyanide) -
      Final Standards"  Federal Register Vol 43 (192) Tuesdav,
      October 3, 1978,  pp 45761-45819.

OSHA (1971)
      OSHA, "Occupational Safety and  Health Standards Subpart B
      Adoption and Extention of Established Federal Standards,"
      Federal Register  Vol 36 No. 	,  pp 10466
                             VI-124

-------
8.  The Fertilizer  Institute
                           Mr, R. Gordon Wells
                         The Fertilizer Institute
                           1015 18th  Street N.W.
                          Washington, D,C, 20036
         THE FERTILIZER INSTITUTE (TFI) IS A NON-PROFIT TRADE ASSOCIATION

     FOR THE  FERTILIZER  INDUSTRY,  REPRESENTING  ALL PHASES  OF THE

     FERTILIZER INDUSTRY,  INCLUDING DISTRIBUTION AND  MARKETING.  TFI'S

     MEMBERSHIP OF OVER 340 COMPANIES MANUFACTURES MORE THAN 90 PER-

     CENT OF DOMESTICALLY-PRODUCED FERTILIZERS.   A  MAJOR  FERTILIZER

     MATERIAL,  WHICH  IS ALSO A  SYNTHETIC  ORGANIC CHEMICAL,  IS UREA.

     ANNUAL PRODUCTION OF UREA HAS REACHED 6 MILLION TONS WITH OVER

     75 PERCENT OF THAT USED AS A FERTILIZER FOR U.S. FOOD AND FIBER PRO-

     DUCTION.  THUS, TFI'S MEMBER COMPANIES HAVE A VITAL INTEREST IN THE

     DRAFT AIR EMISSION CONTROL TECHNIQUES GUIDELINE IN WHICH EPA HAS

     INCLUDED THE UREA MANUFACTURING INDUSTRY AS A  FUGITIVE EMISSION

     SOURCE OF VOLATILE ORGANIC COMPOUNDS.



         IN THE FEDERAL  REGISTER OF FEBRUARY 12, 1981, THE ENVIRON-

     MENTAL PROTECTION AGENCY (EPA) ANNOUNCED  THE AVAILABILITY OF

     THE PRELIMINARY  DRAFT CONTROL TECHNIQUES GUIDELINE (CTG) DOCU-

     MENT FOR VOLATILE ORGANIC COMPOUNDS (VOC) FUGITIVE EMISSIONS FROM

     THE SYNTHETIC ORGANIC CHEMICALS MANUFACTURING INDUSTRY (SOCMI).

     WHILE IT HAS BEEN OUR PRACTICE TO WORK CLOSELY WITH EPA ON MATTERS
                              VI-125

-------
AFFECTING FERTILIZER PRODUCTION FACILITIES, WE DID NOT PARTICIPATE


IN THE DEVELOPMENT OF THE CTG FOR VOC FUGITIVE EMISSIONS, BECAUSE


NO FERTILIZER MATERIALS ARE VOLATILE ORGANIC COMPOUNDS. INDEED,


NO  PHOTOCHEMICALLY  ACTIVE  VOCs  ARE  USED  TO  MANUFACTURE


FERTILIZERS.  HENCE, TFI CONCLUDED THAT CONTROLS ON VOC EMISSIONS


COULD NOT REASONABLY 'APPLY TO  ANY SEGMENT OF THE FERTILIZER


INDUSTRY.



    TO OUR DISMAY, WE DISCOVERED THAT  EPA'S "GENERIC"  CONTROL


TECHNIQUES GUIDELINE  FOR VOC  FUGITIVE EMISSIONS FROM SOCMI


INCLUDED UREA PRODUCTION FACILITIES.  SYNTHETIC UREA IS PRODUCED


BY REACTING AMMONIA AND CARBON DIOXIDE IN A HEATED, PRESSURIZED


REACTOR TO FORM AMMONIUM CARBAMATE WHICH DECOMPOSES TO UREA


IN WATER.  THE UREA PRODUCED IS VERY PURE CONTAINING ONLY A TRACE


OF THE CO-PRODUCT, BIURET.  NEITHER UREA NOR BIURET ARE VOLATILE


COMPOUNDS.   AT  ELEVATED  TEMPERATURES,  UREA  DECOMPOSES  TO


AMMONIA, BIURET, AND CYANURIC' ACID. BIURET ALSO DECOMPOSES UPON


HEATING.   THUS, TFI SUBMITS THAT THERE IS NO JUSTIFICATION FOR RE-


QUIRING THE UREA MANUFACTURING INDUSTRY TO COMPLY WITH THE VOC


CONTROL  REQUIREMENTS OUTLINED  IN  THE CTG, SINCE THE PROCESS


CLEARLY CONTAINS NO SUCH COMPOUNDS. THEREFORE, TFI RECOMMENDS


THAT EPA DELETE UREA FROM ITS LIST OF MANUFACTURING INDUSTRIES


COVERED BY THE PRELIMINARY DRAFT CONTROL  TECHNIQUES GUIDELINE


DOCUMENT.


                                     Respectfully submitted,


                                     R. Gordon Wells, Director
                                     Environmental Programs
                          VI-126

-------
                               C.   DISCUSSION


     Following the EPA presentation,  Mr.  Don Goodwin opened the floor to
questions and comments from the NAPCTAC members.   EPA staff and contractor
personnel were on hand to respond  to  questions and discuss issues of concern
of the NAPCTAC members.   Industry  representatives then made presentations,
each of which was followed by discussion.   For clarity, discussions are
grouped by subject matter rather than being placed in chronological sequence.

     Mr. Bruce Steiner asked whether  the  CTG model regulation was intended to
apply to coke processing plants.  Other committee members had questions about
what facilities would actually be  covered by the  model regulation in the CTG.
Mr. Russell Blosser suggested a lower limit cutoff based on pounds of
chemicals produced below which a plant would be exempt from the recommendations
of the model regulation.  Mr. Steiner stated that the applicability portion  of
the model regulation should state  more specifically what is covered and what
is not.

     Mr. Fred Porter responded by  stating that the CTG covers only the
manufacture of the chemicals listed in the Appendix B.  If a chemical  was
being manufactured at a coke processing plant or  any type of plant, it would
be covered by the CTG..  Mr. Goodwin said  that the size problem is very
complicated, and that it might not be possible for EPA to identify in a
document of this type which plants should be covered and which plants should
not.  This should probably be left up to  the States.  Mr. Eric Lemke stated
that the South Coast Air Quality Management District regulation for control
of fugitive VOC emissions applies  to  all  facilities which utilize VOC's in
manufacturing, including, for example, adhesive manufacturing.

     Following the presentation by Dan Martin of  the Chemical Manufacturers
Association, one of the committee  members asked if he was recommending a
change in the definition of a leak from 10,000 ppmv to 100,000 ppmv.
Mr. Martin replied that the definition of a leak  should be changed to 100,000  ppmv
following an initial inspection and repair of leaks using a lower definition
of a leak.  Mr. Bruce Davis of Exxon  Corporation, in speaking from the
audience, stated that the data in  the CTG showed  that changing the definition
of a leak from 10,000 ppmv to 40,000  ppmv would result in an emission
increase of only 3 percent while greatly  reducing the amount of time required
to repair leaks.  Ms. Elizabeth Haskell asked Mr. Martin how much time is
currently spent on leak repair and how much the recommendations of the CTG
would increase the time required to maintain leaks.  Mr. Martin replied that
there would not be a big increase  in  the  time required for maintaining leaks.
The model regulation would mainly  require the purchase of a few instruments
                                    VI-127

-------
and more organization in recording information.  Mr. Martin also stated
that his company does not keep records of maintenance at their chemical
plants.  They keep a "punch list" of things to repair at unit turnaround.

     The CMA feels that the CTG does not take inaccessible components into
account.  Ms. Janet Chalupnik asked Mr. Martin to give an example of an-
inaccessible component.  Mr. Martin identified high-pressure polyethylene
pumps as an example of equipment that is barricaded for safety reasons.

     Following Mr. Fred Debbrecht's presentation on the differences between
four types of portable hydrocarbon detectors, Mr. Blosser asked if the
HNu Photoionization detector could detect methane.  Mr. K.C. Hustvedt replied
that it could not and that EPA would probably need to specify a calibration
gas for this instrument.

     Mr. Tom Kittleman of du Pont gave a presentation in which he suggested
that EPA adopt a skip-period monitoring plan using a level of good per-
formance based on a percent of valves leaking.  He suggested 4 percent valves
leaking as a reasonable level of good performance.  A committee member asked
if EPA had looked into Mr. Kittleman's plan.  EPA responsed that it had
looked at the plan but there were some problems with implementing it, such
as defining a level of percent of valves leaking which would represent "good
performance."  Mr. Bill Tippitt asked if du Pont had actually tried to
implement a skip-period monitoring plan.  Mr. Kittleman said they had not
tried  to implement such a plan.

     Mr. Jerry Schroy of Monsanto raised the issue of EPA regulations
overlapping and duplicating OSHA regulations.  He said that EPA has ignored
the fact that OSHA regulations could include the kind of control recommended
in the CTG.  Mr- Blosser asked if he was suggesting that EPA recognize the
impact of OSHA on the chemical industry.  Mr. Schroy replied that he was
suggesting this.  Several questions were asked of Mr. Schroy regarding the
specific OSHA requirements he was referring to, but no specific reference
was given.

     Mr. J.D. Thomas of Tennessee Eastman stated in his presentation that the
interest rate of 10 percent assumed in the CTG was much too low and thus does
not reflect the real cost of money.  Mr. Bill Vatavuk replied that 10 percent
represents the rate of real return and has been recommended by OMB.
                                   VI-128

-------
                                 D,  CORRESPONDENCE

         1.  Illinois  Environmental Protection Agency
ILLINOIS
        217/782-2113
        March 13,  1981
Environmental   Protection   Agency
2200  Churchill Road, Springfield, Illinois 62706
        National  Air Pollution Control
         Techniques Advisory Committee
        U.S. Environmental  Protection Agency
        Office of Air Quality Planning  and Standards
        Research  Triangle Park, North Carolina   27711

        Gentlemen:

        For your  information and record, the Illinois Environmental  Protection
        Agency submits the  following comments:

        Attachment 1 —  Comments on Preliminary Draft "Control of Volatile
                       Organic Emissions from Petroleum Dry Cleaners";

        Attachment 2 —  Comments on Preliminary Draft "Control of Volatile
                       Organic Emissions from Volatile Organic Liquid Storage  in
                       Floating and Fixed Roof Tanks";
      /""^Attachment 3  --^ Comments on Preliminary Draft "Control of Volatile
      ^ ________ _J Organic Fugitive Emissions from Synthetic Organic
                  "~~~"   Chemical, Polymer and Resin Manufacturing Equipment".

        Your consideration of  these comments is most appreciated.

        Sincerely yours,
        John C. Reed, Ph.D.,  P.E.
        Supervisor, Technical  Support Unit
        Air Quality Planning  Section
        Division of Air Pollution Control

        JCR:jab/2852H/24
                                    VI-129

-------
                                Attachment 3

                        Comments on Preliminary Draft

       "Control of Volatile Organic Fugitive Emissions from Synthetic
        Organic Chemical, Polymer  and Resin Manufacturing Equipment"


1.  p. 1-1, 2nd paragraph, the  language of the Clean Air Act  states,
    "...all reasonably available measures..." not reasonably  available
    control technology.

2.  p. 1-2, 4th paragraph, why  not make the model regulation  apply to SIC
    categories?

3.  p. 1-3, 1st paragraph, possible exemption for non-ozone season?

4.  p. 2-19, 2nd paragraph, what is the basis for assuming 1/2 of SOCMI
    liquid sources are in light liquid service?  Why not assume all liquid
    sources since this is the most conservative consumtpion?

5.  p. 3-1, 4th paragraph, define  close proximity.

6.  p. 3-3, 2nd paragraph, sentence structure confusing, "...Most control
    values have a manual bypass loop which allows them to be  isolated
    easily although temporary changes in process operation may allow
    isolation  in some cases..."

7.  p. 3-4, paragraph 2, reference 6 seems poor documentation for such a
    conclusion.  If correct, this  data should be summarized and included
    in the CT6.  Is that the purpose of Table A-10?  If so this should be
    used as reference.  Please  note in Table A-10, the percent of values
    with decreased emissions was greater than that with increased
    emissions, also there was a very small sample.

8.  p. 3-7, paragraph 2, there  should be a reference for the  statement,
    "...The percent of emissions from a component which would be affected
    by the repair interval if all  other contributing factors  were 100
    percent efficient is 97.9 percent..."  The correction factor C on page
    3-9 indicates 97.9% of the  emissions will be captured after an average
    interval of 7.5 days.

9.  p. 4-5, Table 4-4, where do emission factors for open-ended valves
    come from? They are not mentioned in Table 4-3. and p. 3-2, 2nd
    paragraph, states the emissions have not been quantified  for each time
    the cap, plug, flange or second valve is opened?.  In addition, this
    paragraph  states that caps, plugs, etc., for open-ended valves do not
    affect emissions which may  occur during use of the valve.  Apparently,
    the controlled emission factors were taken as those of in-line valves
    and gives  no credit for the caps, plugs, etc. required as RACT.
                                VI-130

-------
Page 2


10. p. 4-4, 2nd paragraph, it states, "...The reduction  in emissions  for
    the model units after RACT is implemented is 66 percent..."   It should
    be emphasized that the RACT reduction would be dependent, upon  tJje
    distribution between components  in the model units.   In  fact,  t/e
    values are only approximately 66 percent since the values in  Table  4-1
    calculate as A = 66.4%, B = 65.2*, and C = 65.4X.

11. p. 5-3, paragraphs 5.1.2 and 5.1.3, the costs for repairing leaks do
    not seem to reflect a reasonable progression.  Why should it  take one
    hour to put on a cap on an open-ended line and only  1.13  hours for
    repairing a valve?  Why should it take 80 hours to replace a  pump seal
    but only 40 hours to replace a compresser seal?  There are certainly
    other devices besides a screw-on type globe valve to  use  for  a cap
    (and much cheaper ones at that!).  These costs should all  be
    completely documented and average values used.  Actual cost data
    should be available from refinery programs that are more  accurate than
    these estimates.

12. p. 6-1, what is the basis for the exemption of less  than  10%  VOC
    containing components in xx.010(B)?

13. p. 6-2, definition of "volatile  organic compound" should  only  refer to
    measurement by applicable test method in xx.020.

14. p. A-ll, paragraph 2, the correlation between maximum observed
    screening value and measured non-methane leak rate should be  given
    plus supporting data (in tabular form).  One possible regulation  may
    include a priority repair schedule based on the observed  screening
    value.

15. p. A-ll, paragraph A.I.5 and Table A-7, statistical  analysis  such as
    Chi-Square or analysis of variance should be used to  demonstrate
    percentage of leaks in SOCMI is  comparable to petroleum refinery
    source.

16. p. A-ll, paragraph A.I.6, states there was no significant difference
    in leak rates between manual and automatic valves and signficant
    trends were observed with changes in vapor pressure.  Table A-8 should
    include data to support those statements.

JCR:sh/2721H/37-38
                                VI-131

-------
 2,  Monsanto Company
Monsanto
                                                         .CORPORATEENGINEERING DEPARTMENT
                           Monsanto Company
                           800 N. Lindbergh Boulevard
                           St. Louis, Missouri 63166
                           Phone: (314) 694-1000

                                         March 19,  1981


U. S. Environmental Protection Agency
Emission Standards and Engineering  Division
Research Triangle Park, North  Carolina   27711

Attention:  Fred Porter/Don  Goodwin

Dear Mr. Porter and Mr. Goodwin:

I wish to express my appreciation for the  opportunity to
express Monsanto's views on  the preliminary draft CTG dealing
with fugitive emissions in SOCMI facilities.   I hope the points
raised during my presentation  will  assist  you in finalizing the
CTG.

Following the meeting a question was raised by K.  C. Hustvedt
that I believe should be addressed  by one  additional comment for
the record.  His question dealt with the response factors of the
flame ionization and catalytic oxidation detectors  and indicates
to me he does not understand how these  devices function.  Based
on our discussion it appears I misunderstood the data presented
by Radian in the two reports used to prepare slide  seven of my
presentation.  The two references provided instrument responses
at different concentrations  of the  specific chemical of concern
and I should not have averaged the  two  values to arrive at a single
actual value.  The point K.  C.  Hustvedt raised however reinforces my
concern in terms of the validity of the data.   The  principal of
performance for the flame ionization detector should cause a
reduction in percent of theory response as the concentration of
contaminant increases in the detector chamber.   This reduced
effectiveness of the detector  is based  on  the inability to create
and maintain 100% of the carbon atoms as carbonium  ions from the
flame to the detector plate.   Just  as the  complexity of the
molecule (e.g. number of carbon-carbon  bonds to a specific carbon,
the number of double bonds,  or the  number  of other  materials,  such
as chlorine or oxygen,  in the  molecule)  impact the  life of a carbonium
ion so does the absolute level  of contaminant molecules in the gas
stream.  This is one of the  reasons that these simple detectors
fail to provide consistent response in  the workplace as spill
detectors.   Maintaining the  units in calibration is extremely
difficult.

A second point was raised by Fred Demmick  that should also be
clarified.   During a discussion following  the meeting,  he indicated
concern over the comments related to the "light liquid service"
definition and Monsanto's views on  seal performance.  The definition
of light liquid service we proposed is  based on the inability of

                             VI-132

-------
Page 2                                              March 19,  1981


a mechanically sound seal to contain vapor loss.  Flashing of
the material occurs during seal face cooling.  This is true for
butane at ambient conditions or dodecane at a temperature above
220°C (428°F).   For chemicals which are processed as liquids
at temperatures where they will exist as a gas at atmospheric
pressure they should be considered as light liquid.  However,  this
does not mean dodecane or any low volatility chemical should be
considered as light liquids for all processes.   A definition
as we proposed would correct this technical concern.  The
requirement to monitor only those potential emission points and
install appropriate corrective measures would be reduced and
far more manageable.  For those services concerned as heavy
liquid service (e.g. dodecane, decane and nonane etc. at ambient
conditions) visual inspections for liquid leaks would identify
seal failures as a reasonable control practice.

Again thank you for the opportunity to present our views.  I
have attached a copy of my comments with typo errors corrected.
                                  Sincerely yours,
/dg

Attachment
                                           Schroy
                                         ,o Fellow
                             VT-133

-------
   3.  E. I. du Pont de Nemours & Company
             ESTABLISHED 1602
E. I. DU PONT DE NEMOURS & COMPANY
             INCORPORATED
     WILMINGTON, DELAWARE oasa


LEGAL DEPARTMENT                          March  20,  1981
     Emission Standards and Engineering
       Division  (MD-13)
     Environmental Protection Agency
     Research Triangle Park, NC  27711

     Attention:  Mr. Fred Park

     Dear Sir:

                           Preliminary Draft
                  Guideline Series  for the  Control  of
               Volatile Organic  Fugitive  Emissions  From
                  Synthetic Organic Chemical,  Polymer
               	and Resin Manufacturing  Equipment

               We have reviewed  the preliminary  draft  of  the
     United States Environmental Protection Agency's  (EPA's)
     Guideline Series for the Control of  Volatile Organic
     Fugitive Emissions From Synthetic Organic Chemical,
     Polymer and Resin Manufacturing Equipment.  Representa-
     tives of the E. I. du Pont  de  Nemours  and Company
     appeared at the March 18, 1981 National Air Pollution
     Control Techniques Advisory Committee  (NAPTAC) meeting
     and provided oral commentary with respect to the  prelim-
     inary draft.  The instant comments are entered as a
     supplement to the comments  set forth at that time.

               The CTG document  "purports to provide state and
     local air pollution control agencies with an information
     base for proceeding with the development  and adoption of
     regulations which reflect RACT for specific stationary
     sources."  They are purported  to provide  a  "review of
     existing information and data  concerning  the technology
     and cost of various control techniques to reduce  emissions."
     In addition, "the CTG documents identify  control  techniques
     and suggest emission limitations which EPA  considers the
     'presumptive norm' broadly  representative of RACT for the
     entire stationary source category covered by the  CTG document.

               The "draft CTG document includes  a model regulation
     based upon the  'presumptive norm' considered broadly repre-
     sentative of RACT for the stationary source category covered
     by that document."  "The sole  purpose  of  the model regulation
     is to assist state and local agencies  in  development and
     adoption of regulations for specific stationary sources."
                               VI-134

-------
Environmental Protection Agency
Page Two
March 20, 1981
          While the EPA makes a specific point that the
model regulation and the CTG itself is not to be considered
rulemaking by the EPA, the Agency is well aware that prepara-
tion and dissemination of the document vests its provisions
with precisely the same force and effect.  It is highly
unlikely that states, faced with the choice of adopting the
guidelines and obtaining EPA approval or rejecting and/or
modifying same and risk EPA disapproval, will opt for the
latter.  In fact, it has been our experience that the
emission control requirements embodied in the referenced
CTG's will be imposed upon industry by nearly every state
virtually intact.  This is a natural consequence of the
fact that few states have the resources to rewrite such
guidelines.  The 1979 round of SIP revisions was character-
ized by the wholesale adoption of the CTG's, under EPA
pressure, in order to avoid the severe penalties set forth
in the Clean Air Act.

          The EPA approach is highly incongruous in the
light of the cloak of rulemaking with which the guidelines
will be vested.  Having obtained all of the benefit of rule-
making status for the guidelines, the EPA has chosen not to
adhere to any of the due process and procedural requirements
established under the Administrative Procedures Act (APA)
established for precisely this type of purpose.  In addition,
the Agency has not adequately explained the basis for the
guidelines -- a procedural safeguard which the Clean Air
Act mandates for all rulemaking.  EPA's failure to submit
the contents of the draft guidelines to the rigors of notice
and comment procedures set forth in the APA and adequately
explain the basis therefor virtually eliminates the ability
of those industries and individuals most impacted by its
broad application from providing meaningful commentary
thereon.

          Based upon these factors, coupled with the fact
that each CTG has an impact fully equivalent to other EPA
actions requiring rulemaking (e.g., Clean Air Act,'Section
lll(d) standards on Existing Source Categories, Clean Water
Act, section 304 Effluent Guidelines) the Du Pont Company
submits that the fundamental principles of due process
compel public notice, comment,  Agency justification and
opportunity for judicial review.
                          VI-135

-------
Environmental Protection Agency
Page Three
March 20, 1981
          In addition, by effectively setting forth a
"presumptive norm" considered broadly representative of
RACT for stationary sources, the EPA is setting forth a
mechanism it has already rejected in the area of water
quality standards.  In its Advanced Notice of Proposed
Rulemaking, EPA had announced a policy of "presumptive
applicability" for section 304(a) (1) Clean Water Act
criteria codified in the "Red Book".  "Presumptive
applicability" meant that a state had to adopt criterion
for a particular water quality parameter at least as
stringent as the recommendation in the Red Book unless
the state was able to justify a less stringent criterion.
Conceding that the policy of presumptive applicability
has proved to be too inflexible in actual practice, EPA
announced the rescission of that policy on November 28,
1980.

          The logic that led EPA to the rescission of its
"presumptive applicability" policy is equally applicable to
the situation herein and should compel reconsideration of
the "presumptive norm" approach.  The states have been
vested with certain discretion and authority under the
provisions of the Clean Air Act which they should be per-
mitted to pursue consistently with the development and
adoption of regulations which reflect RACT for specific
stationary sources.  It is submitted that the establish-
ment of inflexible federal guidelines which go beyond the
identification of control techniques and operate to mandate
state compliance with inflexible limitations unduly usurp
the authority and discretion of the states and are inappropriate.

          Turning to the technical aspects of the proposed
CTG, Du Pont has several concerns regarding the proposed
monitoring requirements.  We attach hereto Du Pont technical
comments on inspection frequency contained in EPA's draft
CTG model regulation, as well as comments on the basis for
the model regulation  (Attachments A and B).  These comments
were submitted for the record on March 18, 1981, but are
included herein for the sake of completeness.  We incorporate
these comments by reference to the instant statement.  Briefly,
it is our position that the draft monitoring requirements
ignore effective scientific sampling principles.  In short,
the proposed requirements are not cost effective for either
industry or the control agencies that will be responsible
for their implementation.  In addition, they impose a greater
monitoring burden on clean plants than on dirty ones, and
finally, serve to discourage innovative approaches to fugitive
emission control.
                          VI-136

-------
Environmental Protection Agency
Page Four
March 20, 1981
          The alternative inspection requirement suggested
by Du Pont would provide essentially equivalent control at
a lower cost, place the monitoring burden on the dirty
plants and encourage innovation.

          In April of 1980, Du Font's views on monitoring
in the draft fugitive emissions NSPS were presented to
NAPTAC.  We proposed the use of statistical inspection
plans.  These sample plans are widely used throughout
industry to monitor the quality of manufactured products
and many other control parameter.  We perceive no valid
reason why such plans should not be used to monitor the
performance of valves in chemical processes.

          We have conducted additional studies to answer
concerns raised at the April 1980 meeting.  These studies
have been submitted to EPA and the results discussed at
meetings with the Agency.  EPA has not responded to the
issues raised, and we see no evidence that qualified
statisticians have reviewed them.  The draft model regula-
tion rejects sound principles of statistical sampling
without explanation.

          At the April 1980 NAPTAC meeting, Du Pont recommended
that skip-period inspection plans or their equivalents be used
to determine how much monitoring a new plant would be required
to do.  A skip-period inspection plan was utilized to demonstrate
how one statistical plan works.  We also believe a CTG should
incorporate such skip plans or their statistical equivalents.
Adopting a skip-period plan and the option to use its equiva-
lents will result in good leak protection at a far more reasona-
ble cost.  If a plant's low-leak performance deteriorates, these
plans require increased monitoring.  This approach allows small
emission increases at plants where emissions are low.  As a
result, monitoring costs will be reduced where inspections
will not achieve any significant environmental benefits.  High
leak rate plants would be required to conduct the most monitor-
ing.  These plants could, however, reduce monitoring costs by
reducing leaks enough to demonstrate good performance.  This
insures that most intensive inspection is required where it
will have the greatest environmental benefit.  It also provides
a dirty plant with an incentive to identify and correct the
cause of leaks.

          As an example of how the draft CTG model regulations
would work, consider two plants:
                          VI-137

-------
Environmental Protection Agency
Page Five
March 20, 1981
          Plant A initially has about 0.1% leaks.  Its
overall uncontrolled valve leak rate is 0.35 Ib/hr.  After
the CTG model regulation is applied, the leak rate is
reduced to 0.04 Ib/hr.  This leaves no room for additional
reductions to allow an equivalency determination.  There-
fore, Plant A is required to comply with the most intensive
inspection requirement.

          Plant B has about 22% leaks on the initial screening.
Its overall uncontrolled valve leak rate is 49 Ib/hr.  After
the CTG model regulation is applied, the leak rate is reduced
to 4.9 Ib/hr.  The plant has the option of further reducing
this loss to compensate for a less demanding inspection
requirement.

          The result is that Plant B emits 100 times more
and is required to perform less monitoring than Plant A.

          Consider what would happen with a skip-period
approach  to monitoring.  Plant A could monitor once instead
of four times a year.  Its initial emissions would be reduced
from 0.35 Ib/hr. to 0.07 Ib/hr. instead of 0.04 Ib/hr. under
the CTG example.

          Plant B, on the other hand, would be required to
reduce valve emissions from 49 Ib/hr. down to 4.9 Ib/hr.
as was required by the CTG example.  To reduce its monitoring
burden, this plant would have to improve its leak performance
until it  met a good performance criterion.  We believe a good
performance level of 4% leaks can be justified for existing
plants.   If the plant cleans up and meets the good performance
level, which we believe it would be more likely to do than to
go through the procedure of proving equivalence that was
postulated for the CTG example, resulting emissions would be
4.4 Ib/hr.  Plant B would have an additional 10% emission
reduction compared to the CTG example.

          The end result of these comparisons is that the
skip-period approach can yield greater emission reductions
at much lower cost and also greatly reduce the cost and
difficulties of equivalency determination and SIP revision.

          In short, we believe the CTG monitoring provision
has the following disadvantages:
                          VI-138

-------
Environmental Protection Agency
Page Six
March 20, 1981
          1.  It attempts to inspect good leak performance
into a plant rather than setting a realistic goal and encourag-
ing plants to maintain a low-leak operation.   (Quality cannot
be inspected in, it must be built in.)

          2.  It is not cost effective because the same amount
of inspection is required irrespective of the plant's leakage.

          3.  There is little incentive to reduce emissions
through equipment or work practice improvements because the
equivalency requirements are so demanding.

          4.  The CTG may create an incentive to allow equipment
and work practices to deteriorate.  The model regulation's
requirements could supplant existing practices which may or may
not more effectively reduce emissions.

          5.  Low-leak plants will be least able to take
advantage of reduced inspection alternatives.  The better
a plant is engineered to reduce leaks,  the less opportunity
will exist for reducing emissions to permit reduced inspec-
tion frequency.  For example, a plant that had no leaks in
the initial baseline could never reduce inspection frequency
because it is impossible to have equivalent performance.

          6.  Selection of an alternative methodology is
too complicated and costly to justify for most plants.
Significantly more time and money will be required to get
an alternative considered.  In addition, there are uncer-
tainties such as the possibility that a given alternative
will not be accepted or that the time that an alternative
could be used would be short due to plant modification,
etc.  These factors reduce the likelihood that the alterna-
tives would even be considered.

          7.  The draft CTG model regulation focuses on local
emission reductions and, as a result, ignores overall or
national emission reduction.  We believe that the model regu-
lation should focus upon more demanding inspection procedures
with respect to high-leak rate plants and should provide a
degree of flexibility for well-constructed plants that demon-
strate low-leak performance.  This would not be the case with
EPA's draft CTG model regulation.  In fact, EPA's alternative
could allow less demanding inspection of high-leak plants.
                          VI-139

-------
Environmental Protection Agency
Page Seven
March 20, 1981
          In the documentation attached to our comments
(in order to provide the basis for conclusions set forth
in our comments on inspection frequency, we attach for
your review, as Attachment C, technical studies in support
thereof.  These include a statement on Statistical Inspection
Plans for Monitoring Fugitive Emissions from Leaking Valves
[April 16-17, 1980], a background document on skip-period
fugitive emission plans choosing a level of good performance,
and a draft "background information" document provided to
Mr. Robert Ajax of EPA on February 18, 1981), Du Pont has
reviewed many aspects of the draft CTG model regulation's
approach and the skip-period approach to regulation.  We
conclude that a 4% level of good performance is justifiable
for existing plants.  For a plant to continue to use skip-
period inspection, this 4% criterion will require a group
of valves to have average leak frequencies less than 2%.

          We further conclude that a 4% level of good
performance would still achieve 98% of the emission reduc-
tion that EPA claims for the draft CTG.  This 98% reduction
would cost about half of what the CTG model regulation
approach would cost.  If we accept the EPA data in the CTG
as accurate, and we do not concede same, it is evident that
the nationwide cost of compliance with the model regulation
will be approximately $50 million per year.  The skip-
period approach which we have set forth would result in
a cost savings of approximately $25 million with an equiva-
lent reduction in emissions, even assuming a worst-case
analysis of emissions.

          Therefore, based on the extensive experience of
the quality control profession with statistical sampling
plans, we believe the skip-period approach to regulation
will result in greater emission reductions than will be
obtained by the draft CTG approach.  The skip-period or
equivalent approach has the added benefit of achieving
the reductions in a far more cost-effective way.  We urge
the Agency's consideration of such methodologies.

          In sum, we submit that in its dissemination of
the preliminary draft of the CTG, EPA has circumvented
fundamental principles of due process by failing to submit
the contents of the draft guidelines to the rigors of notice
and comment procedures set forth in the APA and failing to
adequately explain the basis for the proposal.
                          VI-140

-------
Environmental Protection Agency
Page Eight
March 20, 1981
In addition, insofar  as  the  very approach which constitutes
the basis for EPA's methodology herein — the presumptive
norm — has been rescinded as  too inflexible by EPA in
another regulatory area, we  seriously question its utiliza-
tion for the control  of  fugitive emissions.  Furthermore,
we maintain that the  proposed  requirements are technically
deficient in that they are not cost effective for either
industry or the control  agencies who will be responsible
for their implementation;  they impose a greater monitoring
burden upon clean plants than  on dirty ones, and serve to
discourage the use of innovative approaches to fugitive
emission controls.  We trust our comments provide assistance
to the Agency in its  deliberations on this matter.  We will
be available to supplement or  provide further information
with respect to this  matter  if requested.

          Du Pont is  an  active member of the Chemical
Manufacturers Association  (CMA)  and the Texas Chemical
Council (TCC) and has worked extensively with CMA and TCC
in the preparation of its  comments on this issue.  Du Pont
fully endorses and adopts  CMA's and TCC's comments as our
own and we incorporate them  in our comments by reference.

                                    Sincerely,
                                  Steven A. Tasher
                                Environment Division
SAT:scl
Attachments
                                                                  Qu,
 EDITOR'S NOTE:  Attachment A, "Comments  on Inspection  Frequency," and
              Attachment B, "Comments  on Regulatory  Basis," were received
              as part of Mr. Kittleman's presentation and are included in
              Section IV. B. 4 of these  minutes.
                           VI-141

-------
                  ATTACHMENT C
               E.  I.  du Pont de Nemours  & Co,
                        Statement on
        STATISTICAL INSPECTION PLANS FOR MONITORING

           FUGITIVE EMISSIONS FROM LEAKING VALVES
                        Prepared By

               R. D.  Snee and T. A. Kittleman
                  Engineering Department


                            for

                      Presentation to


National Air Pollution Control Technicues Advisory Committee
                     April 16-17, 1930
                  Hilton Hotel, Raleigh, NC
                         VI-142

-------
        STATISTICAL INSPECTION PLANS FOR MONITORING



          FUGITIVE EMISSIONS FROM LEAKING VALVES





1 .  Introduction






The NAPCTAC is reviewing fugitive emissions monitoring



proposals for benzene and general volatile organic compounds.



Although the logic of the comments we wish to make today



applies to both monitoring plans, our presentation addresses



the requirements for general volatile organic compounds.





EPA's proposal would require that all valves, pumps,  com-



pressors, etc. be checked on a regular basis (i.e. monthly or



quarterly).  While regular checks on some units is certainly



appropriate, it seems that requiring 100X inspection  of all



units at each screening period is overly conservative and



unnecessarily costly particularly when it can be shown that



the leak frequency of some units is very low.  It is'reason-



able to ask, therefore, whether data on unit performance and



statistical sampling procedures can be  used  to  reduce the



costs of screening for leaking units and still  provide an



adequate margin of protection.






Qur comments are aimed aftwo aspects of the sampling



requirement:  1)  The base sampling period 'and 2)  The



desirability of incorporating statistical sampling plans



to  achieve a high degree of protection  whil-e minimizing



costs.   Our following  discussion illustrates an inspection



plan for valves.   The  plan  is general,  however,  and works



equally well for  pumps,  compressors,  etc.
                        VI-143

-------
                           - 2 -





2.  Base Sampling Period





We believe the proposed monthly sampling requirement  is



excessive and question whether the environmental benefits



can justify the high' relative costs of a monthly, as  opposed



to a quarterly sampling frequency.  Ou Pont has reported



costs of $3 to 4 per sample.  In vinyl acetate produc-



tion at our La Porte Plant an EPA contractor recently



sampled 1232 valves.  This sample b'y no means covered the



entire, plant.  'For instance, units associated with this  '



product, such as refining, and "inaccessible" valves  were



not sampled.  Nonetheless, at 53 per sample, a monthly



sampling requirement as opposed to a quarterly one,  would



add about S30M annually to the cost to sample only these



valves.  We question whether these costs could justify what



we beleive would be a small emission reduction.   Unfortunately



emission rate data are not yet available from this sampling



program,, let alone one comparing the effects of monthly



versus quarterly sampling.  However, the data that are



available do clearly show that the percentage of leakers is



much less than reported for petroleum refineries.





3.  Skip-Period Inspection Plans





There are statistical inspection plans which work  for any



base period and provide adequate protection against  leaks



but minimize cost by allowing reduced inspection when good
                          VI-144

-------
                           - 3 -


performance is present.  We believe the Skip-Period  Inspection

Plans have merit for this purpose.  These plans assume  that

a population of N valves will be monitored on a periodic
                                                       •
basis (i.e. monthly or quarterly).  The plans call  for  all N

valves to be inspected in a given period.  If acceptable

performance is found in ± successive periods then one or

more periods may be skipped depending on the desired level

of protection.  This approach rewards the plant for  good

performance (i.e. few leaks) a-nd forces the plant to increase

the level of inspection when an unacceptable number  of  leaks

are found.  It also insures that all valves are inspected

periodically.


The Skip-Period Inspection Plan is based on skip-lot inspec-

tion plans which have been used in the quality cont'rol  field

for more than 25 years (Dodge 1955).  The single-level  plan

has four basic steps:


1.  Inspect a group of N valves of a particular class or

    type for _i_ successive periods.


2.  If performance is acep.table, inspect the group of N

    valves in only f_ fraction of the following periods.


3.  Return to inspection of all valves in all periods when

    unacceptable performance is found.


4.  Fix all leaks found.
                         VI-145

-------
The values of i and f are selected to give the 'desired  level



of protection.  Several alternative plans are gi-ven  in  Table  1



For illustrative purposes we will consider the plan  with  is-2



and f=1/2.  This plan states- that if acceptable performance



is found in 2 successive periods, then the population of



valves can be inspected in every second period (i.e. inspect



in periods 2, 4, 6, ...}.  Qf course, as soon as unacceptable



performance is found the plan requires that inspection  be



done in all periods until acceptable performance has been

                                           •

found once more in 2 successive periods.  If this plan  is



used, on the average, the population of valves will  have



acceptable performance at least 90S of the time.





It is important to recognize that there are other skip^period



inspect ion'plans which provide the same level of protection



and use the presence of good .performance to reduce the



amount of  inspection.  In particular we recommend the



Multi-Level Inspection plans developed by lieberman  and



Solomon (1955) .and a related set of plans described  in



Department of Defense Handbook H-1Q6 (also see Ireson and



Biedenbender 1958).  These plans switch back and forth


                                                2    "5
between different levels of inspection (ie, f, f , f ,...)



depending  on the performance of the valves.  The single-level



skip-period plan (Table 1) uses only one level of inspection



(ie, f) and returns to inspection in all periods if  poor



performance is found in any period.  The single level plan
                          VI-146

-------
                           - 5 -





is illustrated here because of its simplicity.  We feel,



however, that any plan which gives an equivalent level of



protection should be acceptable.





Both the single-level and multi-level inspection plans place



a lower limit on the average percent of the time that the



population of valves will have acceptable performance.  The



protection level is high because all leaks are fixed when



detected and, when performance is poor, all valves are



inspected in all periods.





4.  Definition of Good Performance





The skip-period inspection plan wall work with any- definition



of "good performance".  A maximum fraction of valves leaking



and/or a maximum number of leaks per plant could define good



performance.  We submit that a group of valves is performing



well if less than 2* of them are leaking.  According to the



plan discussed earlier (is2, f=1/2) a group of N values



would be inspected every second period if not more than 2%



of the valves were leaking in two successive period.



(Figure 1 shows how this particular plan would operate.)



Under this plan, on the average, the population of valves



would have 2%, or fewer, leaks at least 90% of the time.



Skip-period plans can also be used in situations where good



performance is defined as a maximum number of leaks, MAXL,



CKMAXl£N.  If the is2, f=1/2 plan were used, then, on the



average, the population of N valves would have MAXL, or



fewer, leaks at least 90S of the time.
                        VI-147

-------
                           - 6 -





5.   Group Definition and Protection Level





As  inspection data are collected it will become apparent



that the leak frequency will vary with the different types



of  valves.  It is certainly appropriate to group the valves



into different categories according to leak frequency (e.g.



leak frequently, leak sometimes, rarely leak).  It is also



possible to have different inspection plans for the different



groups.  In the end, the objective is to focus the inspection



effort where the leaks occur most often.





It  is also appropriate to vary the protection level depending



on  the effects of the leaking chemical.  Certainly hazardous



chemicals may require a higher protection level than chemicals



which are relatively nontoxic.  Plans can also be developed



for protection levels other than 90-percent (Sheesley 1975).





6.   Assumpt ions





The skip-period plan assumes, as does any inspection plan,



that if a valve is not a leaker it will not leak for the



entire period.  It is also assumed that the leak frequencies



of the different periods are independent of each other.



This assumption may not be appropriate; however, major leaks



will carry over and be detected in the next period.
                        VI-148

-------
                          - 7 -


                        References


Dodge, H. F. (1955), "Skip-Lot sampling plan", Industrial

   Quality Control, 11,  No. 5, 3-5.



Ireson, W. G. and Biendenhender,  R. E. (1958), "Review of

   Department of Defense Handbook H-106, 'Multiple-level

   continuous sampling procedures'", Industrial Quality

   Control, j_5, No. 4,  10-15.



leiberman, G. 3. and Solomon, H.  (1955), "Multi-level

   continuous sampling plans", Annals of Mathematical

   Statistics, 2£t 686-704.



"Multi-level continuous sampling  procedures and tables of

    inspection by attributes", Inspection and Quality

    Control Handbook (Interim) H-106, October 1958, Office

    of the Assistant Secretary of- Defense (Supply and

    Logistics), Washington, D.C.



Sheesley, 3. H. (1975),  "A computer program to evaluate

   Dodge's continuous sampling plans", Journal of Quality
                    •
   Technology , _7_, 43-45.
R. D. Snee
4/10/80
                         VI-149

-------
                                 TABLE 1




               ALTERNATIVE  SKIP-PERIOD INSPECTION PLANS
i f Lower Limit On
# Successive Traction of Percent of Time With
Periods Inspected Periods Inspected Acceptable Performance
2
4
5
6
12


1/2 90
1/3 90
1/4 90
1/5 90
1/14 90
FIGURE 1
SKIP-PERIOD INSPECTION PLAN OPERATION
            GOOD PERFORMANCE                    REDUCED INSPECTION
            'LESS THAN 2% \	/INSPECT IN

             LEA-KS FOUND   	\ f PERIODS
              1 (. PERIODS /                      V T PER'ODS
               INSPECT .   \                      /MORE THAN 2jr
START	H    IN ALL      K	4
               PERIODS	/                      VLEAKS  FOUND
           COMPLETE INSPECTION                    BAD PERFORMANCE
                               VI-150

-------
SKIP-PERIOD FUGITIVE EMISSION INSPECTION PLANS
     CHOOSING A. LEVEL OF GOOD PERFORMANCE
    Ronald D. Snee and Thomas A. Kittleman
            Engineering Department
     E. I» du Pont de Nemours and Company
         Wilmington, Delaware   19898
                 VI-151

-------
                                -1-
                              SUMMARY

A fugitive emissions monitoring provision  that  allows  low  leak
rate plants to reduce inspection frequency will encourage
industry to design plants to minimize fugitive  emissions.  The
major step needed to specify such a provision is  to  select a
percentage of leaking valves, "level of good performance", that
would allow a plant to increase the time between  inspections  if a
lower leak occurrence is demonstrated.  This percentage  should be
low enough that good design is required by plants wanting  to
reduce the inspection burden but high enough to be attainable and
attractive for wide industry use.  The objective  of  the  work
reported here was to quantify emission differences between skip-
period incentive inspection plans for a range of  "good
performance levels" and fixed-period inspection plans.   Two
different bases were used to make the comparisons:
 1}  A percent leak reduction approach similar  to  that  used  in
    EPA's Background Information Document,  and
 2)  Plant fugitive emissions simulation  techniques.

 Both comparisons showed that, within a realistic  range,  the
 "level of good performance" chosen will  not greatly  effect
 emissions from a given plant.  However,  new plants designed  to
                          VI-152

-------
                               -ii-
use skip-period inspection with any realistic  "level of  good
performance" will have significantly greater emission reductions
than the average refinery with fixed-period inspection.

The maximum differences between post-inspection .emissions of  the
skip- and fixed-period plans, applied to an existing plant,
             i
varied from 5.2 to 6.3% of emissions before inspection when the
good performance level was varied from 0.5 to  3% leaks.  This
study shows that greater fugitive emissions reductions will
result from encouraging the design of low leak rate plants than
from controlling high leak rate plants through frequent
fixed-period inspection.
                         VI-153

-------
          Skip-Period Fugitive Emission Inspection Plans
               Choosing A Level of Good Performance
Introduction

In an earlier paper  (Snee and Kittleman 1980), we introduced the
Skip-Period inspection plan as an effective -alternative to the
Fixed-Period inspection plan for fugitive emissions proposed by
the EPA.  The skip-period plan permits one to inspect valves,
pumps, compressors,  etc., at a specified reduced frequency when
the group inspected  is performing well and meets a .specified
"level of good performance".  The question remains as to how to
choose the level of  good performance.  We selected 2% leaks as
the level of good performance because we believe that
well-designed plants run with an effective maintenance program
can meet this performance level.  The 2% good performance level
produces a leak frequency equivalent to that cited in EPA's
example emissions calculation (Background Information Document,
March 1980) and produces lower emissions than fixed-period
inspection of the average refinery  (Wetherold et al 1980).

Most fugitive emissions come from high leak frequency plants
(more than 10% leaks).  Encouragement to design low-leak plants
is the most effective strategy to control these emissions.  The
opportunity to use skip-period inspection rewards the low leak
                          VI-154

-------
                                -2-
plants for good performance and provides the high leak  frequency
plants with an incentive to reduce emissions thereby enabling
them to take advantage of skip-period inspection.  The  added
benefit of skip-period inspection is that if the high leak
frequency plants improve their operation so that these  inspection
plans can be used/ greater emission reductions will result  than
under the strict use of the fixed-period inspection plan proposed
by the EPA.

The question remains "what effect does the level of good
performance have on the effectiveness of the skip-period plan?
Is 2% adequate?  Is 1% needed?  Will 3% work just as well?"  We
chose to answer these questions by studying both the leak  .
frequency (% of valves leaking) and emissions  (Ib/hr) remaining
after inspection.  We conclude that the difference between  the
 *
two plans is insensitive to the level of good performance and
that 2% is adequate.

Percent of Valves Leaking After Inspection

The EPA has defined a "leak" as any valve whose screening value
is greater than or equal to 10,000 ppm.  The base period most
recently considered for EPA's fixed-period sampling plan is
quarterly.  They estimate that this plan will be 90% effective  in
                         VI-155

-------
                                -3-
that over a quarter the number of leaking valves due to
reoccurrence of old leaks or the occurrence of new leaks will
average 10% of the valves found leaking at the beginning of  the
period (EPA Background Information Document, BID, March 1980).

After the level of good performance is met in five (5) successive
quarters, our proposed skip-period plan calls for inspection  one
quarter per year  (f=l/4) as long as the good performance level  is
met  (Snee and Kittleman 1980).  All valves would be checked
during each inspection and all leaks would be repaired.  Only
plants that were always below the good performance level could
continue reduced inspection.  One inspection showing a leak
frequency above the good performance level would necessitate  five
more quarters of good performance before a reduced inspection
schedule could be resumed.  This would assure that only
inherently low leak rate plants could reduce inspection frequency
and  insure that later changes (ie, reduced maintenance, altered
process conditions, etc.) don't substantially increase leakage.

The  EPA provided the following emission effectiveness estimates
(BID, March 1980) which enables us to evaluate the performance  of
the  skip-period plan.
                          VI-156

-------
                                -4-
                                       EPA Estimates
                                      BID, March 1980
                              Total                  Average
                         Number Of Leaks          Number Of Leaks
  Time Period            At End Of Period        During The Period
  1 Quarter                    ,2N                      .IN
  1 Year                       .4N                      .2N

N = Number of Valves with initial screening values greater than
    or equal to 10,000 ppm.

During the year, on the average, the number of valves leaking
will be 20% of the valves initially leaking (i.e., 80%
reduction).  The EPA also estimates that one quarter after a
group of valves are inspected'and fixed, the number of leaking
valves will be 20% of those initially leaking.  If no further
inspection is done, the number of valves leaking at the end of
one year will be 40% of the valves initially leaking.

A plot of the average percent leaks after inspection versus
percent leaks initially found for 80% leak frequency reduction
(skip-period) and 90% leak frequency reduction (fixed-period) is
shown in Figure 1.  From this plot it is apparent that using the
2% good performance criterion, (viz, 2% found leaking at end of
period corresponds to 1% average) the skip-period plan can be
used only for plants which initially had a leak frequency less
than or equal to 5%.  Results for other levels of good
performance are:
                          VI-157

-------
                                -5-
                                                  Maximum Average
Maximum Initial          Good Performance         Leak  Frequency
Leak Frequency           	Level             After  Inspection
     1.25%                   .5% Leaks                    .25%
     2.5%                   1.0% Leaks                    .5%
     5.0%                   2.0% Leaks                   1.0%
     7.5%                   3.0% Leaks                   1.5%

These results are based on the EPA estimates which  indicated that
the number of leaks found after a one year period will  be 40%  of
the initial leak frequency and that the average leak  frequency is
20% of the initial leak frequency.
The next step  is to compare the reductions  in  leak  frequencies
which result under fixed-period sampling and skip-period
sampling.  This is done by studying Figure  1 and  the calculations
detailed in Table 1.  These simplistic comparisons  are based  on
emissions from valves with screening values greater than  10,000
ppm only.  As  will become apparent from the plant simulations
•reported later, this is not the most important difference between
skip- and fixed-period inspection plans.  Nonetheless, we believe
there is some  value to making these worst case comparisons.   They
show that:
                         VI-158

-------
                                -6-
o  The reduced inspection incentives provided' by the skip-period
   inspection plan illustrated for a 2% good performance level
   can result in greater leak frequency reductions than provided
   by the fixed-period plan for plants with leak frequencies of
   10% or more.
              /
o  There is little difference between the results of 1% and 2%
   good performance levels.  For example, at the 12% leak
   frequency  found in the EPA Refinery Study (Wetherold et al
   1980) skip-period reduction in leak frequency was 106% and
   102% of fixed-period reduction for 1% and 2% levels of good
   performance, respectively.

o  Of equal importance, it shows that plants using skip-period
   inspection with 2% good performance level will have lower leak
   frequencies than the "average refinery"  (Wetherold et al 19'80)
   using fixed-period sampling.

Hence, in addition to being indicative of a well-designed plant
with an effective maintenance program, the  2% good performance
level results in lower leak frequencies than will result for many
plants using  fixed-period inspection.  Not  allowing rewards for
good performance gives the high leak frequency plants an unfair
advantage and provides no incentives to do  a better job.
                         VI-159

-------
                                -7-
Relationship Between Leak Frequency and Emissions

in interpreting the effect of skip-period  inspection  on  leak,
frequency, it is helpful to know the relationship between  the
percentage of leaking valves and emissions before and after
inspection.  These relationships, developed by simulation
techniques described in the Appendix, are  shown in Figures 2-5.
The effect of 80 and 90% emission reductions  for valves  with
screening values above 10,000 ppm is shown in Figures 3-5.

We conclude  (Figure 2) that total emissions before inspection  are
a linear function of percentage leaks (i.e.,  percent  of  valves
with screening values greater than 10,000  ppm).  Emissions after
inspection  (Figure 3), or equivalently, emission reductions,
(Figure 4) are curvilinear functions of percentage leaks.
Consequently, emissions after inspection are  a curvilinear
function of  emissions before inspection  (Figure 5).   We  note that
the curvature in Figures 3-5 is small.  TO a  first approximation,
both emission reduction and emissions after inspection are linear
functions of percentage leaks.  It is apparent, therefore, that
leak frequency is a meaningful surrogate for  emissions and the
associated conclusions will be directly applicable to emissions.
This conclusion will be further supported  in  the next section.
                          VI-160

-------
                                -8-
Emissions After Inspection

The next step in our analysis was to compare the emission
reductions (Ib/hr) obtainable at existing plants using
skip-period and fixed-period sampling plans and to evaluate  the
effect of the good performance level on emissions after
inspection.  The following discussion will show that  the
skip-period emission reduction is 88.9% of the fixed-period
emission reduction at the same plant.  In the good performance
level range of 0.5% to 3.0%, the maximum difference between
skip-period and fixed-period emission reductions, as  a percentage
of emissions before inspection, varies from approximately  5  to
6%.  We conclude that the level of good performance has little
effect on emissions after inspection.

If we let
     TE- = Total emissions 'before inspection
     TE, = Total emissions after inspection, and
       F = Fraction of total emissions from valves with screening
           values more than 10,000 ppm
                          VI-161

-------
                                -9-
then TE. = (Emissions from Valves  not  Requiring  Repair)
                            Plus
           (Emissions from Valves  Repaired)
         = TEB(1-P) + TEB(E)(F)
         = TEB(1-F+EF)

where E = Inspection effectiveness
        =0.2 for Skip-Period and  0.1  for Fixed-Period.

Emission reduction is therefore given  by
     Reduction =  (TE--TE,)/TE-
                    DA   a
               «  (TEB-[TEB(1-F+EF)])/TEB
               »  (l-E)F
We conclude that  the  skip-period  reduction  is  0.8F  and  the
fixed-period reduction  is  0.9F, hence
         Skip-Period        =   8        Fixed  Period        .,.
      Emissions Reduction   =   J     Emissions Reduction    ^  '
and,  at a given  plant  skip-period  emissions  reduction  is
100(8/9) =  88.9%  of  fixed-period emissions  reductions.
                          VI-162

-------
                                -10-
Now if we let



     STEA = Total emissions after Skip-Period  inspection



     FTEA « Total emissions after Fixed-Period inspection







then from equation  (1) we have



     (TEB-STEA)/TE3 = 8(TEB-FTEA)/(9TEB).







After some algebra



     (STEA-FTEA)/TEB =  (1-FTEA/TEB}/9            (2)







and we obtain an expression for  the difference between  emissions



after skip-period inspection and fixed-period  inspection



expressed as a fraction of total emissions  before  inspection.







Using equation 2 and our simulation results (Appendix),  it  is



shown in Table 2 that, at well performing plants that meet  the



performance level, varying the good performance level from  0.5%



to 3.0% varies the difference between  the skip-period to



fixed-period emissions from 5.2  to 6.3%  of  total emissions  before



inspection.  We see once again that the  benefits of  the



skip-period inspection plan are  not greatly affected by the



chosen level of good performance.
                         VI-163

-------
                               -11-
Effect of Skip-Period incentive on Emission Reductions

In an earlier section we showed that leak  frequencies lower  than
those produced by fixed-period sampling could result because of
the incentives to improve valve performance so  that skip-period
inspection could be used.  Similar calculations  for emissions
reductions are detailed in Table 3.  Here  we see  that,  for the
12% leak frequency observed in the EPA Refinery  Study (Wetherold
et al 1980), the skip-period plan produces an additional  emission
reduction of approximately 20% and 35% for good  performance
levels of 2% and 1%, respectively.

It is also apparent from Figure 3 and Table 3 that using  a
skip-period plan with 2% good performance  level,  which  will  be
used by any plant with an initial leak rate of  5% or less, will
result in emission reductions equivalent to those of plants  with
leak frequencies of 6% or more using fixed-period.  Hence, with
respect to emissions from a given plant, there  is less  difference
between the two plans than shown earlier when leak frequency was
used as the basis for comparison  (Table 4) .  More importantly, if
new plants are designed with low leak valves so  they can  benefit
from skip-period inspection, greater industry-wide emission
control will result.  This study clearly shows  that greater
fugitive emission reductions will result from encouraging the
design of low leak plants rather than by controlling high leak
plants through frequent fixed-period inspections,
                          VI-164

-------
                             APPENDIX

Simulation of Valve Emissions

In order to study the effects of valve' leak frequency,  it was
necessary to determine the relationship between leak  frequency
(i.e. % of valves with screening values greater than  10,000  ppm)
and

o  Total emissions from all' valves
o  Emissions using Fixed-Period inspection, and
o  Emissions using Skip-Period inspection.

These simulations were conducted using the data and models
reported in. EPA's Refinery Study (Wetherold et al 1980) and
supplemented by data that EPA collected at one of Du  Pont's
plants.

It was assumed that the nonzero screening values followed a
lognormal distribution as shown in the EPA Refinery Study
(Wetherold and Provost 1979) .  The two parameters of  the
lognormal distribution are the average.or mean  (X) and  the
coefficient of variation  (CV = X/S, S = standard deviation).  Our
approach to the simulation was to assume a coefficient  of
variation and then determine the average  (X) which would produce
                         VI-165

-------
                                A2
the desired frequency of screening values greater  than  10,000

ppm.  The CV's found in the EPA Refinery and Du  Pont  studies are

shown in Table 5.  Based on these data, it was decided  to use

values of CV = 2 and CV = 3 in the simulation.



The next concern was the frequency of zero screening  values.  In

the EPA Refinery and Du Pont studies it was found  that  the

percent of zero screening values varied from 45% to 97% depending

on  the percentage of screening values greater than 10,000 ppra

(Table 5).  The frequency of zero screening values used in the

simulation  (Table 6) were interpolated from the  observed data in

Table 5.



The next step in the simulation was to compute the emissions from.

each valve given its simulated screening value.  This was done

using the models for gas/vapor valves and light  liquids/two-

phase valves published  in the EPA Refinery Study Report

(Wetherold et al 1980).



Gas/Vapor Valves

            -7   1 23
     E^ » 10   (x) *       x = screening value  (ppm)
    ^
Light Liquid/Two-Phase Valves

     E, = 10-4-9(x)°'8
                          VI-166

-------
The emissions associated with  the  simulated screening value x was
given by the average of E,  and E2/  ie>

Emissions  (Ibs/hr) =  (E1-i-E2)/2.

These equations show that  emissions  are  essentially a linear
function of screening value, hence,  as shown  in  Table 7 and in
Figure 2, emission is essentially  a  linear  function of percent of
screening values greater than  10,000  pom.

One thousand  (1000) screening  values  were generated in each
simulation.  Each screening value  was computed  by converting a
uniform random number to a normal  random deviate by the
procedures of Mueller  (1959) and  the  following  relationships for
the mean  (y) and standard  deviation  (Sy) of the  lognormal
distribution.

     x     * ppm     y = In x
     y     = In  (x/(CV2-i-l)1/2)
     S2y   = In  (CV2+1)
     y     = y-i-Z Sy, Z = random normal deviate
     x     » e^ = simulated screening value
                          Vl-167

-------
                                 A4
The parameters specified by  the  simulation  are the average (jc)
and coefficient of variation  (CV) of  screening values in' pp;m
                                                •
(Table 6).  The following  statistics  were computed for each
simulation:

     Average  (x)
     Coefficient of variation
     Percent  screening values greater than  10,000  pom
     Total emissions
     Emissions from valves with  screening values greater  than
       10,000 ppm  (Ib/hr and % of total  emissions)
     Emissions after fixed-period inspection  (Ib/hr  and % of
       total  emissions)
     Emissions after skip-period inspection (Ib/hr and %  of
       total  emissions)

As estimated  by the EPA./ emissions  after fixed-period and
skip-period inspection were  assumed,  respectively,  to be  10% and
20% of the emissions from  valves with screening values greater
than or equal to 10,000 ppm.

In order  to reduce the variation introduced by the simulation,
the results were smoothed  by least  squares  curve fits.  The
emission  estimates presented in  Table 7  and Figures  3 are the
resulting smoothed values.

0936A/cak
R.D. Snee/T.L. Kittleman
November  17,  1980
                         VI-168

-------
                            REFERENCES

Muller, M. E. (1959) J. Association of Computing Machinery,
   Volume 6, No. 3, p. 376.

Snee, R. D. and Kittleman, T. A.' (1980) Statistical inspection
   Plans for Monitoring Fugitive Emissions from Leaking Valves,
   Statement before the National Air Pollution Control Techniques
   Advisory Board, Hilton' Hotel, Raleigh, NC, April 16-17, 1980.

VOC Fugitive Emissions in Synthetic Organic Chemicals
   Manufacturing Industry - Background Information for Proposed
   Standards, .Preliminary Draft, U. S. Environmental Protection
   Agency, March 1980.

Wetherold, R. G. and Provost, L. P. (1979) .  Emission Factors and
   Frequency of Leak Occurrence for Fitting in Refinery Process
   Units, U. S. Environmental Protection Agency Contract Nos.
   68-02-2147 and 58-02-2665, Report No. SPA-600/2-79-044,
   February 1979.

Wetherold, R. G., Provost, L. P., and Smith, C. D. (1980).
   Assessment of Atmospheric Emissions from Petroleum Refining,
   Volume 3, Appendix B.  U. S. Environmental Protection Agency
   Contract No. 68-02-2147, Report No. EPA-600/2-80-075c, April
   1980.
                         VI-169

-------
                                              TABLE 1
                                  Effect  of Good  Performance Level
                            on Benefits of Skip-Period Inspection Plans
   Good
Performance
   Level

 2% Leaks
 2% Leaks
 2% Leaks

 1% Leaks
 1% Leaks
 1% Leaks
 1% Leaks
Leak Frequency
Initial
10%
12%
20%
5%
10%
12%
20%
Average After
Fixed-Period
1.0%
1.2%
2.0%
.5%
1.0%
1.2%
2.0%
Percent
Reduction
90
90
90
90
90
90
90
Leak Frequency
Average After
Skip-Period*
1.0%
1.0%
1.0%
.5%
.5%
.5%
.5%
 Percent
Reduction

   90.0
   91.7
   95.0

   90.0
   95.0
   95.8
   97.5
Reduction
Skip/Fixed

   100%
   102%
   106%

   100%
   106%
   106%
   106%
*It is assumed that the plant  has improved  the  performance of  their valves,  thereby  lowering  the
leak frequency and permitlng skip-period  sampling  to be used with the stated  (2% or.  1%) good
performance level.

-------
                              TABLE 2

            Skip-Period Versus Fixed-Period Inspection
     Effect  on  Post-Inspection Emissions For an Existing Plant
   Good
Performance
   Level

 .5% Leaks
1.0% Leaks
2.0% Leaks
3.0% Leaks
 Initial
  Leak
Frequency

   1.25%
   2.5%
   5.0%
   7.5%
    Emissions (Ibs/hr)
  Before
Inspection

    .919
   1.61
   2.98
   4.35
   After
Fixed-Period

    .433
    .826
   1.41
   1.88
Skip-Fixed*
  Before

   5.2%
   5.4%
   5.8%
   6.3%
*Frora equation 2.  After inspection emissions, skip-period minus
fixed period expressed as a percent of total emissions before
inspection.
(Skip-Fixed)/Total
(Equation 2)
        (1 - After Fixed-Period/Before)/9
                        VI-171

-------
                                                TABLE 3

                                    Effect  of  Good  Performance Level
                         on Emissions Reductions Due to Skip-Period Inspection
I

ro
        Good      Initial
     Performance   Leak       Initial    Fixed-Period   Percent   Skip-Period   percent   Reduction
        Level     Frequency   Emissions    Emissions    Reduction   Emissions*  Reduction  Skip/Fixed
2% Leaks
2% Leaks
2% Leaks
1% Leaks
1% Leaks
1% Leaks
12%
10%
6%
12%
10%
3%
6.82
5.72
3.53
6.82
5.72
1.88
2.44
2.24
1.62
2.44
2.24 -
• .952
64.2
60.8
54.1
64.2
60.8
49.4
1.60
1.60
1.60
.915
.915
.915
76.5
72.0
54.7
86.6
84.0
48.7
119%
118%
101%
135%
138%
99%
     *It  is  assumed  that  the  plant  has  improved  the performance of its valves,  thereby  lowering
     leak frequency  and emissions and permitting skip-period  inspection  to be used.

-------
                              TABLE 4
      Leak Frequencies at Which Skip-Period  and  Fixed-Period
                 inspection Plans are Equivalent
Good
Performance
Level
2% Leaks
2% Leaks
1% Leaks
1% Leaks
Comparison
Basis
Percent Leaks
Emissions after
Inspection
Percent Leaks
Emissions after
Leak Frequencies Producing
Equal Performance
Skip-Period
5%
5%
2.5%
2.5%
Fixed-Period*
10%
6%
5%
3%
               Inspection
*Skip-period inspection  plans will perform better  than  fixed-
period inspection  plans  when a plant's  leak  frequency is  higher
than the fixed-period  leak  frequencies  shown.
                              TABLE 5

                    Screening Value Statistics

Study
EPA Refinery
Du Pont
Ball Valves
Globe Valves
Gate Valves
No. of
Valves
1476

723
78
379
Coeff. of
Variation
2.16

2.58
3.17
3.03
Zero Screening
Values (%)
45

97
92
64
                                                         Percent
                                                         Leaks*

                                                          11.9

                                                             .55
                                                            1.85
                                                            2.56
*A valve is a "leaker" if its screening value is greater  than  or
eaual to 1Q.QQO nnm
equal to 10,000 ppm
                         VI-173

-------
                         'LIST OF FIGURES
Figure                                   Title

  1                       Leak  frequency after  inspection versus
                          leak  frequency before inspection for
                          80% and  90%  leak  frequency  reductions
                          due to inspection.

  2                       Total emissions before inspection
                           (Ibs/hr)  versus percent of  screening
                          values greater than or equal  to
                          10,000 ppm.

  3                       Emissions after inspection  (Ibs/hr)  for
                          30% and  90%  emission  reduction  versus
                          percent  of screening  values greater
                          than  or  equal  to  10,000 ppm.

  4                       Emission reductions  (Ibs/hr)  due to
                          inspection (80% and 90% effectiveness)
                          versus percent of screening values
                          greater  than or equal to 10,000 ppm.

  5                       Total emissions after inspection
                           (Ibs/hr)  with  80% and 90% emission
                          reduction versus  total emissions before
                          inspection (Ibs/hr).
                          VI-174

-------
                              TABLE 6

               Valve Emission Simulation Parameters*
Percent
 Leaks

  12
  11
  10
   9
   8
   7
   6
   5
   4
   .3
   2
   1
   .5
    Zero Screening
      Values (%)

         45
         47.5
         50
         52.5
         55
         57.5
         60
         65
         70
         75
         30
         87
         93
          Screening Value Average  (pom)
              CV=2             CV=3 '
              8419
              8002
              7703
              7322
              6960
              6532
              5977
              5681
              5332
              5068
              4403
              3738
              3420
                    -9830
                     9251
                     8340
                     8319
                     7829
                     7257
                     6526
                     6141
                     5693
                     5357
                     4534
                     3722
                     3347
*% leaks = % of screening values greater than 10,000 ppm
 Zero screening values = % of screening values = 0 ppm
 CV = coefficient of variation  (average/std dev.)
                              TABLE 7

                 Valve Emission Simulation Results
Percent
 Leaks

    .5
   1
   1.25
   1.5
   2
   2.5
   3
   4
   5
   6
   7
   7.5
   8
   9
  10
  11
  12
   Good
Performance
   Level
 .5% Leaks


1%   Leaks


2%   Leaks


3%   Leaks
                                    Emissions  (lbs/hr)*
Total

 .508
 .782
 .919
1.06
1.33
1.
1.
  ,61
  ,88
2.43
2.98
  ,53
  .08
  ,35
  .62
  ,17
5.72
6.27
6.82
3,
4,
4,
4,
5,
    After
• Fixed-Period

     .272
     .417
     .488
     .558
     .694
     .326
     .952
    1.19
    1.41
    1.62
    1.80
    1.83
    1.96
    2.11
    2.24
    2.35
    2.44
   After
Skip-Period

    .279
    .445
    .526
    .606
    .763
    .915
   1.06
   1.34
   1.60
   1.85
   2.07
   2.18
   2.28
   2.46
   2.63
   2.78
   2.91
*Smoothed values developed from  a  least  squares  curve  fit,
Adjusted R2 statistics were 0.96,  0.96,  and  0.97.
                         VI-175

-------
      AVERAGE X LEAKS BEFORE AND AFTER INSTITUTING A LEAK MONITORING PLAN

                                                T
                  6         8         10        12        14

                      PERCENT LEAKS BEFORE INSPECTION

                                  FIGURE 1
   16
  18        20
    8  -
    7  -
    6  -
sa
**  5
o
on
    4  -
    3 -
    2 -
    1 -
1
*s*
e»
&
i
i i i i i
•
•
• •
• • *
*• •
• *
• •
•
•
••
• •
	 L 	 1 	 1 	 1 	 | 	
                                    6          8         10

                     PERCENT  SCREENING  VALUES > 10, 000 ppm
                                     FIGOTZ 2
12
14
                            VI-176

-------
LO

CO
    3.0
    2.5
    2.0
0.


1  1.5
    i.o
to
(/I
     .5
  IS)
  OQ
  u

  o
  IXI
  o:
  o

  t/1
      1  -
80% REDUCTION
                24        6         8        10

                   PERCENT SCREENING VALVES > 10.000  ppm
                              12
                                 FIGURE 3
                                       90% REDUCTION
              80% REDUCTION
                          _I_L
                                    4—
            _L
                 2         4        6         8        10

                   PERCENT  SCREENING VALUES > 10,000 ppm
                                12
                                  FIGURE 4
                            VI-177

-------
a
c_
    3.0
    2.5
    2.0
    1.5
I   -5
                                   80% REDUCTION
                                            90% REDUCTION
                123         456
                  TOTAL EMISSIONS  BEFORE INSPECTION (LBS/HR)
                                  FIGURE 5
                               VI-178

-------
CN-IOIt
                unniSHiomi
 E. I. DU PONT DE NEMOURS & COMPANY
                INCOKFOIUTXO
      WILMINGTON, DELAWARE 19898
CC:  S. Wyatt   -EPA
     7. Demmick -EPA
     K. Hustuedt-EPA
     G. Wilkins -Radian
     S. Duletsky-GCA
 ENGINEERING DEPARTMENT

 LOUVIERS BUILDING
      February  18,  1981
     Mr. Robert Ajax
     Emission Standards and Engineering Division
     Office  of Air  Quality Planning  & Standards
     Environmental  Protection Agency
     Research Triangle Park, NC   27711

     Dear  Bob:

     We appreciate  your meeting with us to discuss some of our most recent
     studies on control alternatives for  leaking valves.  We intend to submit
     additional material before the  NSFS  comment period closes.  Unfortunately
     many  of the  concerns raised  on  January  20, including questions about equip-
     ment  specification as a means of emission control, were outside  the focus
     of our  most  recent study.  Most of these concerns, however, we have already
     studied and  presented to EPA.   You were not present at the June  nreeting
     when  we presented our earlier studies.  I believe you should know what
     these studies  show.  Since we didn't leave a written text of that presen-
     tation, I have attached one  to  this  letter.  I presented the written text
     and Dr. Snee talked from the diagrams (also attached).  If you have any
     questions, please give me a  call (302)  366-4718.

     Since our meeting I received a  copy  of  the draft "Background Information"
     document for the fugitive emissions  NSPS.  A quick review of this document
     shows that the Du Pont field studies reported many leaks (i.e.,  greater than
     10,000  PPM screening value).  This is incorrect.  The data we reported were
     for screening  values greater than 10 PPM.

     Very  truly yours,

     ENGINEERING  SERVICE DIVISION
     Air Quality  &  Hazards Engineering
                  <. vxjv

      T.  A.  Kittleman
      Senior Engineer

      TAK:cpr
      Attach.
                                    VI-179

-------

D RAF T
6/11/80 - T. A. Kittleman

We wish to address the draft NSPS monitoring requirements for
valves.  Sometime ago B.J. and I talked with K.C., Bob Webber and
others about our concern that these regulations could require
extensive monitoring of equipment that inherently is leak-tight.
We understood EPA's reluctance to promulgate a list of exemptions
which could require much work to update and provide a way to cir-
cumvent the regulations' intent.

Following the EPA meeting we got some statisticians busy reviewing
data and considering statistical techniques that would provide  for
cost effective monitoring and good leak protection.  Our
recommendation of what we believe is the best statistical
monitoring technique was presented at the April 14, 1980 NAPCTAC
meeting in Raleigh.  I believe it was unfortunate that we didn't
have time to sit down with EPA prior to the NAPCTAC meeting to
discuss our work.  The significant features of the possible alternate
monitoring approaches are subtle and I believe this kind of meeting,
with statistician input, is the best way to show you what our
concerns are and consider what approach best meets your concerns
as well.

Du Pont is most concerned about the cost effectiveness of
monitoring requirements as they would apply to clean (i.e., low
leak rate) processes and particularly to modified processes.
                            VI-180

-------
Draft                         - 2 -             T. A. Kittleman

We believe your objective is to write an enforceable regulation
that minimizes emissions.  We believe the alternate approach
suggested by our statisticians provides a better way to meet
both of our objectives (i.e., it is a cost effective approach to
monitoring that could result in lower emissions than any fixed
period monitoring requirement).  Why?  In general, before
getting into some examples, a fixed period monitoring requirement
tries to "inspect" quality into the product whereas the alternate
monitoring requirement gives an incentive to "engineer" quality
into the product.  The history of quality control shows that
quality, in this case a non-leaking process, must be engineered into
the product.  It cannot be inspected into the product.

OK, let's compare the two approaches to inspection and look at some
examples.

Example 1
First, let us consider a new grass roots process.  Even though
this type of situation is not our major concern, the example
can illustrate the desirability of designing quality into the
product.

The Radian screening study data for a process at our La Porte
plant showed the fraction of leaking valves to be as follows:

-------
Draft
    - 3 -
      T. A. Kittleman
      Valve Type

         Globe

         Gate

         Ball
Screening
      PPM
  14.1%

   2.1%

   0.69%
Concentration
? 10,000 PPM

     2.6%

     1.8%

     0.55%
 (More than 60% of the valves in this process are ball valves.)

When all the data from the screening study are plotted, it is

obvious that the three types of valves have different leak

characteristics.  From a practical standpoint, this is an

 important consideration particularly when viewed from the stand-

 point of relative control strategy effectiveness.  To illustrate

 let's consider the valve screening data from our La Porte plant.

 When the number of valves is normalized and emission rates

 computed from petroleum industry calibrations,- some striking

 differences become evident.
Valve Type
Globe
Ball
Gate
All Valves
2.142
0.42
6.167
Valves Screened
€ < 10,000^ PPM
0.427
0.005
0.017
Scr'IlnfHg
Value, PPM
300
35
200
 Ball  valves  have  the  lowest  leak rates at high and  low  screening

 values.   Gate  valves  had  the highest leak rates.  Globe valves,
                           VI-182

-------
Draft                        - 4 -              T. A. Kittleman



although having intermediate overall leak rates, would still emit

as much after the 10,000 PPM valves are repaired as the ball valves

emit total.  Consequently, to achieve the lowest possible leak rate

from this plant the best strategy appears to be to encourage the

maximum practical use of the inherently low leak ball valves.

In other words, design quality into the product.



By contrast, according to some of our valve maintenance and

repair experts, the difficulty to repair a leaking valve would

rank them quite differently.



     Valve Type                   Repair Difficulty

   Gate and Globe      Relatively easy repair while in service.

   Ball                Difficult to repair and may require process

                       shutdown.



Consequently, if meeting the draft NPSP were the only considera-
                                                  \
tion, different monitoring possibilities could push valve

specification in different directions.



Under the fixed monitoring period approach, there is no penalty

for finding numerous leaks and consequently minimal incentive

to prevent their occurrence.  The penalty, in terms of added

reporting, possible process shutdown, and even fines, occurs if
                          VI-183

-------
Draft                        - 5 -              T. A. Kittleman




a leak cannot be quickly repaired.  This might provide some

incentive to install globe or gate valves instead of the lower

leak frequency ball valves.




Under a requirement that allows monitoring frequency to be

reduced if few leaks are found, there is an incentive to prevent

leaks from occurring.  This might encourage use of lower leak

frequency valves such as ball valves.  In either case, of

course, all leaks found would be repaired as quickly as practical




In other words, if a monitoring requirement is written to

encourage leak elimination by design, not by inspection, sub-

stantial further reductions in air emissions may result.  Using

the Radian screening results reported earlier for our plant,

valve emission rates could be cut by a factor of about 3 to 5

by selecting ball valves over globe and gate valves.



                                                   \
As I'understand the draft regulations' equivalency provision,

a new process could possibly get relief from some of the

monitoring requirements by installing cleaner valves if they

are willing to go through the motions (i.e., data collection,

public hearings, etc.).  We believe EPA should allow this flexi-

bility and that there will be places where it can be used.

Once again, our major concern is for plants such as our existing
                            VI-184

-------
Draft                        - 6 -              T. A. Kittleman

plant example where low leak technology was widely used in design
and the plant becomes subject to NSPS through modification.

Example 2
OK, now let's look at a modified plant example.  Processes that
will come under NSPS's thorough modification represent Du Pont's
major concern.  Let's look at two separate cases:  (1) a process
with poor leak performance (i.e., high leak valves and poor
maintenance)  and (2) a process with good leak performance (i.e.,
low leak valves and good maintenance).  In the first case the
option of .replacing some valves with lower leak frequency valves
could allow for reduced monitoring leak frequency under the
equivalency provisions.  That is, if the work and delays
necessary to show equivalency did not rule out this approach.
The clean process, however, would not have this option.  Con-
sequently, the plant that had always had good leak performance
would be forced to do more monitoring than the dirtier plant.
In fact, it is easy to see possibilities where, following mod-
ification, plants with higher emissions would be required to
monitor less frequently than lower emission plants.  As an
example, consider two plants.
                          VI-185

-------
Draft                        - 7 -              T. A. Kittleman

             Al  Premodif ication High Leak) AA result of valve change
             A2  Modified                 )  out and 111 requirements
Emissions
             Bl  Premodification Low Leak ) ^ result of m require-
             B2  Modified             •    )  ments

Plant A, with globe and gate valves, would have the  option of
replacing many with ball valves and using the emission reduction
to obtain a less costly monitoring requirement allowed by an
equivalency determination.  However, Plant B, which like our
plant screened by Radian, may already have made maximum practical
use of ball valves and would have significantly less opportunity
for emission reductions.  As a result, Plant B would be faced with
more costly monitoring requirements than Plant A in spite of the
fact that Plant B's-emissions were lower.

Our major concern is for the modified plants that already demon-
strate good leak performance before modification (i.e., Plant B
in the example).  The equivalency clause will not help a plant
unless physical changes can be made.  Therefore, if the only
means a plant has to modify its monitoring requirement is the
equivalency clause, a premodification clean plant will be at a
disadvantage.   In fact, the cleaner  a plant is before modification
the more demanding would be the monitoring requirements after
                            VI-186

-------
Draft                        - 8 -              T. A. Kittleman

modification.  We don't believe this is appropriate; nor that
it is what EPA intends.  However, as long as the regulation is
in terms of a work practice and the only alternate is a performance
equivalence, this type of discrimination is unavoidable.  The
cleanest plants will pay the highest costs, either through more
stringent monitoring requirements or through the necessity of
making more costly process changes to ease the monitoring burden.
The only way we can see to avoid this inherent feature of the
draft regulation is to put some flexibility into the work practices
part of the regulation.  We believe this can be done so that the
regulation would actually achieve greater emission reductions,
encourage industry innovation, reduce compliance costs and cut
the number of equivalency determinations that would otherwise be
requested.

EPA can accomplish this simply by specifying the monitoring
program effectiveness required.  That is, X% of time a process
                                                   \
will not have more than Y% exceeding a Z screening value.  We
believe a 90% protection level and a 10,000 PPM screening value
are reasonable and that there are a variety of ways that
"good performance", Y, can be defined to insure that the monitor-
ing plan is adequate.  At the NAPCTAC meeting we presented 2% as
an example definition of "good performance".  This is in fact the
definition of good performance you applied in your BID emission
                           VI-187

-------
Draft                       - 9 -              T. A. Kittleraan

reduction example.  In that example you assumed that 10%  of
the valves had screening values ;-10,000 PPM and calculated that
quarterly inspection would result in an 86% emission reduction.
In Tables 4-6 of that document you also estimate that  0.2 of
those leaks, or 2%, would reoccur during each three-month
period.  You have therefore established 2% leaks as a  "good
performance" level for valves.

Other "good performance" levels could be supported and used
effectively-  For instance, the Radian screening study referred
to earlier showed that 1% of the plant's flanges leaked.  Since
flanges are exempt from monitoring, why not exempt valves that
have even lower leak frequencies?  It would even be possible
to define the screening level concentration at other than 10,000
PPM for the purposes of defining the monitoring frequency if you
believe it desirable to do so.

                                                   \
The point I am trying to make is that there are really two issues
to discuss:   (1) is the concept and (.2) is the specific numbers
to insert for X, Y and Z.  These issues need to be discussed
separately.  Don't dismiss the concept because you think  X, Y & Z
sound too high or too low.  Let's discuss the concept  and then
the X, Y, Z's.

                       The Concept, Ron!!

TAK:mmc

                           VI-188

-------
VI-189

-------
                      VALVE LEAKS BEFORE AND AFTER MODIFICATION
                      FOR PLANTS WITH HIGH AND LOW INITIAL LEAKS
EMISSIONS
                                    HIGH LEAK PLANT
                      A1   PREMODIFICATION
                                                                 VALVE CHANGE OUT
                                                                 AND NSPS
     MODIFIED PLANT
         INSPECT ONCE/YEAR - MONITORING REDUCED BY EQUIVALENCY
                                    LOW LEAK PLANT
                           PREMODIFICATION
                           MODIFIED PLANT
                               INSPECT 1 TIMES/YEAR
                                                                      NSPS
                                       FIGURE 1
                         SKIP-PERIOD INSPECTION PLAN OPERATION
                  GOOD PERFORMANCE
                                REDUCED  INSPECTION
       START-
                  /LESS THAN 2%
                    LEAKS FOUND
                  \IN I PERIODS
INSPECT
 IN ALL
PERIODS
                                    INSPECT  IN
                                    f PERIODS
/MORE THAN 2%
\LEAKS  FOUND
                  COMPLETE  INSPECTION
                                   BAD PERFORMANCE
                         THE AVERAGE, THE POPULATION OF VALVES WILL HAVE
                            12% LEAKS AT LEAST 90% OF THE TIME
                                    VI-190

-------
               ALTERNATIVE SKIP-PERIOD INSPECTION PLANS
         I
   I SUCCESSIVE
PERIODS INSPECTED
         5

         6

        12
   FRACTION OF
PERIODS INSPECTED

      1/2

      1/3

      1/1

      1/5

      1/11
                  LOWER LIMIT ON
               PERCENT OF TIME WITH
              ACCEPTABLE PERFORMANCE.

                        90

                        90

                        90

                        90
        COMPARISON OF FIXED PERIOD AND SKIP-PERIOD INSPECTION PLANS
                       FOR MONITORING FUGITIVE EMISSIONS
BID EXAMPLE
10% >10,000 PPM
QUARTER
0
1
2
3
1
5
6
7
8
9
10
11
12
13
TOTAL
10
2
2
2
2
2
2
2
2
2
2
. 2
2
2
AVG
1
1
1
1
1
1
1
1
1
1
1
1
1
BETTER CONTROLLED PLANT
52 >10,000 PPM
i TOTAL
5
1
.; i
1 i
i
; 1
2
2
AVG
.5
.5
.5
.5
.5
1
1
    1 QUARTER
    1 YEAR
    *EPA ESTIMATES
TOTAL*
 ,2N
AVG'
JN~
,2N
N = I >10,000 PPM
                            VI-191

-------
       SKIP-PERIOD INSPECTION DOES NOT SIGNIFICANTLY INCREASE TOTAL EMISSIONS
PLANT
  A
  B
  C
  D
  E
TOTAL

INITIAL

20*
15
10
5
2
52
EDUCTION
FIXED
AFTER
., ^
2.Q(i)
1.5
1.0
.5
.2
5.2

PERIOD
REDUCTION
i
18
13.5
9,0
1.5
1.8
16.8
90%
SKIP-PERIOD
AFTER
1 1 ^
2.Q(i)
1.5
1.0
1.0^2)
.1
5.9

REDUCTION

18
13.5
9.0
1.0
1.6
16.1
89%
*   I SCREENING VALUES >10,000 PPM
(1) REDUCTIONS BASED ON-EPA ESTIMATES
     EMISSION REDUCTIONS FOR FIXED PERIOD AND SKIP-PERIOD INSPECTION PLANS
                                 FIXED PERIOD                   SKIP-PERIOD
PLANT        INITIAL        AFTER        REDUCTION        AFTER        REDUCTION
  1             -*            -                     '        _
  2
  P
TOTAL           A             B             A-B             C             A-C
TOTAL  REDUCTION (%)    ,        '         100 (A-B)/A                  100 (A-C)/A
 •TABLED VALUE  IS ANNUAL EMISSIONS
                                 VI-192

-------
10r
                       GOOD  CONTROL PLUS SKIP-PERIOD INSPECTION
                          WILL PRODUCE LOW EMISSION LEVELS
         BID
        \ EXAMPLE
                                   BETTER CONTROLLED PLANTS
    INITIAL     FIXED
              PERIOD
INITIAL   FIXED
         PERIOD
 SKIP     INITIAL   FIXED      SKIP
PERIOD             PERIOD     PERIOD
 •REDUCTIONS BASED ON EPA ESTIMATES
                                  VI-193

-------
           CHEMICAL MANUFACTURERS ASSOCIATION
March 20, 1981
Mr. Fred L. Porter
Emission Standards and Engineering Division
Environmental Protection Agency
Mail Drop 13
Research Triangle Park, NC  27711

     RE:  Draft Control Technique Guideline  Document for
          Volatile Organic Chemical Emissions

Dear Mr. Porter:

The Chemical Manufacturers Association  (CMA) submits this letter
and the enclosed materials as CMA's comments in response to the
Environmental Protection Agency's February 12,  1981 solicitation
for public comment.  CMA presented testimony on the subject docu-
ment at the March 18, 1981 meeting of the National Air Pollution
Control Techniques Advisory Committee.   These  comments are in-
tended to supplement that testimony.

As you may be aware, CMA is a nonprofit trade  association having
187 United States company members representing more than 90 per-
cent of the production capacity of basic industrial chemicals
within this country.  CMA member companies have a direct and criti-
cal interest in ensuring that EPA develops Control Technique
Guidelines (CTG) where a demonstrated need is  presented, that are
scientifically and technically sound, reasonable, procedurally
workable and cost-effective.

CMA has actively worked with EPA over the past few months to develop
a CTG for volatile organic compound  (VOC) fugitive emissions from
the Synthetic Organic Chemicals Manufacturing  Industry (SOCMI).
In this regard, we have reviewed, commented  on and met once with
representatives of EPA's Office of Air  Quality Planning and Stan-
dards (OAQPS) to discuss our concerns with the Agency's draft of a
CTG for VOC fugitive emission sources.   We have several significant
reservations and concerns with the proposed  draft CTG.  Our enclosed
comments will address these issues, provide  illustrative data, in-
formation and rationales, and offer appropriate recommendations.
                             VI-194
         Formerly Manufacturing Chemists Association—Serving the Chemical Industry Since 1872.
       2501 M Street, NW • Washington, DC 20037 • Telephone 202/887-1100 • Telex 89617 (CMA WSH)

-------
                              - 2 -
We would like to highlight that CMA has serious concerns  about  the
Agency's use of refinery data in its rush to issue a draft  CTG  for
the SOCMI.  As we discuss further in our comments, we believe unpub-
lished EPA studies show the refinery data are inappropriate for
SOCMI.  The Agency is currently performing several studies  using-
SOCMI data.  Accordingly, CMA requests the CTG not be finalized
for at least 90 days after the last of these studies is complete.
This should still allow adequate time to revise the CTG by  October,
1981 if such a document is still justifiable.  Of course, CMA offers
to assist the Agency in the review and evaluation of these  studies.

Should the Agency require further information or wish to  discuss
any of the issues raised in these comments, you may contact
David W. Carroll, Assistant General Counsel at (202) 887-1164 or
me at (202) 887-1174.

Sincerely,
Janet S. Matey
Manager
Air Programs

JSM/vac

Enclosure
                            VI-195

-------
                   Chemical Manufacturers Association
                               comments on
                                  Draft
               Control Technique Guideline  (CTG) Document
                                   for
                Volatile Organic Chemical (VOC) Emissions
GENERAL COMMENTS

  •  There are several areas in which the scientific and technical basis
    of the CTG for VOC emissions is faulty.  These flaws weaken the
    credibility of the document and will cause practical problems for
    a company in implementing the procedures described as reasonable
    available control technology (RACT).

 A.  The Agency has published the draft CTG without the benefit of the
    results from several ongoing studies.

         CMA currently reviewed an unpublished draft contractor report
    entitled "Evaluation of Maintenance for Fugitive VOC Emissions Con-
    trol."  This study will be finalized and published as an EPA report
    in approximately one month.  The study contains data which must be
    reviewed with respect to their effect on the CTG.  According to
    our preliminary review these data may significantly affect the fol-
    lowing issues, among others:
              1)  on-line maintenance effectiveness,
              2)  the cost-effective choice of monitoring and mainten-
                  ance interval/
              3)  emissions reductions resulting from the program,
              4)  the adequacy of the refinery/SOCMI comparison,
                                 VI-196

-------
                                  - 2 -



              5)   the 10,000 ppmv leak definition, and


              6)   time to conduct actual maintenance and monitoring


    CMA feels these data should be included in the CTG.


         Also, EPA is currently analyzing these data and the results


    of two other studies  and will incorporate this analysis into a
                          »»

    report.   Further, EPA is reviewing the results of a study by Allied


    Chemical Corporation on leak occurrence and recurrence.  We.request


    EPA to delay issuing the final CTG until these studies are properly


    evaluated and appropriate modifications are made to the CTG.  These


    studies may justify more relaxed requirements.  Unless the require-


    ments are incorporated into the CTG, the states might not be able


    to develop cost-effective SIP revisions that are needed to attain


    ambient standards.



B.  In developing the CTG the Agency should use leak frequency data
    developed for the Synthetic Organic Chemicals Manufacturing Indus-
    try (SOCMI).



         In developing frequency data the Agency has placed extensive


    reliance on data from the refining industry, rather than the SOCMI.


    From this data base EPA estimated the emission reductions from


    leaking components and the effect on ambient air quality.  The data


    from these two industries are NOT similar as EPA indicates on


    page 2-20, and should not be used as the basis for eatablishing


    RACT.  The differences are summarized as follows:
                                  VI-197

-------
                              - 3 -


                             TABLE 1


     Source Type                   Difference in leak frequency
                                   between SOCMI and refining
  	data.	


          Valves
                      i»
           —gas service                similar
           —light liquid service       SOCMI 50% of refining

          Pumps

           —light liquid service       SOCMI 33% of refining

          Compressors                   SOCMI 21% of refining

          Relief Valves                 SOCMI 45% of refining

Attachments A and B give greater detail and discussion on the issue.

     This comparison is based on hexane calibration.  These differ-

ences do not account for the fact that the chemical industry studies were

conducted using a Century OVA-108 instrument calibrated with methane

while the refining studies were conducted using a Bacharach TLV

instrument calibrated with hexane.  Studies by Exxon Chemical

(Attachment C)  on both instruments using both calibration gases

show that 29 percent more leaks are found using the Century instru-

ment calibrated with methane as compared to the Bacharach instrument

calibrated with hexane.  Thus the SOCMI leak frequency is probably

     Also, CMA has compared the SOCMI/refinery data using EPA/

Radian data contained in the EPA SOCMI maintenance study.   (Attach-

ment D)  CMA's preliminary review of this contractor study, re-

leased February 17, 1981, further demonstrates the problem of basing

the CTG on refinery data.  When the refinery and SOCMI data are

compared on an emission equivalent basis at a refinery leak screening

value of 10,000 ppmv, the mass equivalent SOCMI screening values

for pumps in light liquid service and for valves in liquid  and gas

service range from 10,200 to 33,600 ppmv.
                             VI-198

-------
                              -  4 -



     Further, the SOCMI maintenance study  illustrates  the effects


of the change to methane calibration on  the  comparison of the


refinery and SOCMI data.   (Attachment  D)   On an  emission equivalent


basis, the SOCMI screening values for  the  OVA instrument calibrated


with methane range from 35,000 to 94,000 ppmv compared with 10,000 ppmv
                      »*

in the refinery study.  These EPA data support the  trends previously


reported to the Office of Air Quality  Planning and  Standards (OAQPS)


by CMA.

•,     The above analysis of the SOCMI maintenance assumes that the

10,000 ppmv screening value  for  the refinery data is adequate.   The


above analysis is directed toward converting the refinery screening


value of 10,000 ppmv to the  equivalent SOCMI screening value for

the various instruments and  calibration  gases.   We  recommend, how-


ever, that the choice of leak screening  value for SOCMI be reana-

lyzed using mass emission correlations versus screening value and


that an appropriate screening value be chosen using the SOCMI data


base.

     Accordingly, we conclude that the data  support a

screening value range of 40,000  to 100,000 ppmv.


The maintenance screening  study  supports the long standing


CMA/Texas Chemical Council  (TCC) position  that SOCMI fugitive


emissions are  significantly  lower than the values reflected in the


refinery data.  Our  first  analysis indicates that on a mass emission


basis over  84  percent reduction  is achieved  at the  100,000 ppmv


level and that only  3 percent more emission  reduction  is achieved

with a 10,000  ppmv screening value.  At  40,000 ppmv there is an


86 percent emission reduction.   The incremental  emission reduction
                              VI-199

-------
                                 - 5 -

   achieved with a 10,000 pprav screening value is not cost-effective.
   Furthermore, the failure of on-line repair techniques after the
   first attempt to yield any further significant reduction should
   require only pre-shutdown monitoring and startup checkout.  A
   directed maintenance program during a shutdown will yield optimum
   results at minimum costs.
        Finally, an analysis of the leak data for individual SOCMI pro-
   cesses form the EPA/Radian 24 plant study shows that those processes
   that exceed the average of the industry are those which are very
   similar to refinery processes  (i.e. those involving ethylene, pro-
   pane or propylene as either a product or raw material).  Conversely,
   those processes involving specific chemical reactions rather than
   cracking or fractionation as in refineries show a very low frequency
   of fugitive emissions.  Accordingly, CMA believes this study sig-
   nificantly affects the RACT analysis and we recommend the CTG be
   revised to include the SOCMI data from the various studies.  It
   makes no sense to provide the states with CTG documents that will
   result in unjust over-control of segments of the chemical industry
   with no significant improvement in ambient air quality.

C.  The CTG requires detection instruments to be calibrated with methane,
   but the resulting instrument response is highly variable and depends
   on the chemical measured.

        The CTG requires a company to repair a leak if the detection
   instrument indicates 10,000 ppmv-  However, the actual concentra-
   tions of individual chemicals present when the screening instrument
   indicates 10,000 ppmv varies greatly.  A Radian report indicates
   the wide response factors.   For example, with an OVA instrument
                                VI-200

-------
                              - 6 -

calibrated with methane, a meter reading of 10,000 would be indi-
cated by:
              Response Factor              Confidence Interval
     2,900 ppmv of benzene                   2,800  -  3,100
     5,500 ppmv of propane                   4,600  -  7,200
     6,500 ppmv of ethane                    4,400  - 15,800
     7,100 ppmv of ethylene                  6,300  -  8,200
     7,700 ppmv of propylene                 4,400  - 26,600
     8,000 ppmv of acetone                   5,700  - 12,000
     9,900 ppmv of methyl methacrylate       8,000  - 11,000
    15,000 ppmv of cyclohexanone             9,700  - 27,600
    18,700 ppmv of cumene                   11,000  - 37,100
    43,900 ppmv of methyl alcohol           36,100  - 56,000
This means that a company would be required to repair a propane
pump or valve if only a 5,500 ppmv leak was present.  In this case
(and for all the chemicals shown from benzene through methyl metha-
crylate) the meter reading of 10,000 ppmv would falsely indicate a
need for repair.
     The SOCMI maintenance study further illustrates the problem of
using refinery data when they were obtained using instruments cali-
brated with hexane, while RACT is based on an instrument calibrated
with methane.   (Attachment D)
     The confidence interval widths shown above indicate the OVA
instrument is unsuitable as an analytical tool for analysis of every
chemical on that list.  Instead, the Agency should consider equiva-
lent instrumental methods with equivalence defined on a mass equiva-
lent basis.
                             VI-201

-------
                                 - 7 -

        CMA recommends that the screening value be raised and  the
   Agency permit analytical equivalence on a mass emission basis.
        Again, the CTG must be revised to accurately provide informa-
   tion to the states so that they can adequately revise their SIPs
   in order to .attain the standards by the most cost-effective means.
                         »»
D.  The Agency's extrapolation of statistical results on leak occur-
   rence from the entire industry to a single plant is not valid.

        The statistical implication of leak occurrence/recurrence
   should be properly considered by EPA.  The effects of the statisti-
   cal manipulation are especially important in inspection policies
   (page 6-6) and for equivalency determinations.
        The confidence interval plots presented in Figure I/page 3-10 of the
   CTG present 90 percent confidence intervals at related screening
   value, percent of total mass emissions and percent of leaking sources.
   These plotted points represent the confidence intervals for the
   infinite or industry-wide population of valves.  The width  of the
   confidence interval is relatively narrow.  However, when smaller
   valve populations representing the valve population in a single
   plant or process unit are examined, the confidence intervals widen
   markedly.  This is particularly true on the screening value scale
   since this scale is logarithmic.  The spread of screening value
   confidence intervals for a 10,000 valve random sample is itemized
   in CMA's Table 2.  The spread is rather dramatic.  For instance,
   85 percent of total emissions will be greater than a screening value
   of 10,000 ppmv within a 90 percent conficence bound of 72 to 97 per-
   cent of total emissions or screening values of 3,500 to 100,000 ppmv.
   The spread is even larger for a smaller sample population of 100 valves.
                                 VI-202

-------
                              - 8 -

Obviously, the choice of a uniform screening value is valid for the
infinite or industry-wide valve population since the confidence
limits are relatively narrow.  However, when an individual plant
population is considered, an individual screening value used to
predict percent of total mass emissions becomes highly variable.
                      **
     Similar plots could be prepared for correlations of percent of
leaking sources and valves in various other services.  We recommend
EPA construct such plots for all valve, pump and compressor ser-
vice categories and provide this necessary cost-effective flexibility
as part of the CTG.  This would enable a state to predict the vari-
ability of leak threshold definitions for various plant size popu-
lations in various valve, pump and compressor service categories and
implement the most cost-effective requirements.
     Therefore, CMA recommends EPA amend the CTG to assure proper
correction of statistical aberrations in the inspection and equiva-
lency requirements.  We propose the following method be used;

     The upper 95 percent confidence bound should be used as
     an equivalency measure when formulating equivalent moni-
     toring and repair programs.  Also when evaluating the
     results of a random inspection, the width of the confi-
     dence interval should consider the valve population
     under inspection.  The confidence levels can be deter-
     mined during the initial yearlong study for equivalency
     by the individual plant.  As we stated above, the confi-
     dence interval width will be wider for the plant with a
     smaller valve population.
                              VI-203

-------
                                    - 9 -
                                 FIGURE  1
      100


       90

       80

       70


       60


       50


       40


       30


       20

       10
— Estimated  Percent  of\  \ \
    Total  Mass Emissions  x    x  \\\
                          '    ^
 — 90/5 Confidence  Interval \.   \
    for Percent of  EmissionsN    \
    from Total Population of \
                              \
                                V
Valves (n = -)

90/5 Confidence Interval
for Percent of Emissions
in a Random Sample of
1000 Valves
    90% Confidence Interval  for
    Percent of Emissions  in  a
    Random Sample  of 100  Valves
                          I   I
                               \
                                \
                                       \\
                                      \  \
                 \
                 v\
         1   2345   10
        50 100
                1000
10,000  100,000  1,000,000
                          Screening Value  (ppmv)  (Log^Q  Scale)


                      Cumulative Distribution  of  Total Emissions

                                           by
             Screening Values  -  Valves  - Light Liquid/Two-Phase Streams
                           Comparison of Confidence  Intervals
SOURCE:   Wetherhold  and Provost,  Emission  Factors  and  Frequency of Leak Occur-
         rence for Fittings  in Refinery Process  Units.  EPA-600/2-79-044, Febru-
         ary,  1979,  Figure A-5,  Page A-17.

-------
                             - 10 -
                            TABLE 2


       90 PERCENT CONFIDENCE INTERVALS FOR A RANDOM SAMPLE3
        OF 1,000 VALVES IN LIGHT LIQUID/TWO-PHASE STREAMS



1.  85 percent of total emissions will be greater than a screening

    value of 10,000 ppmv within a 90 percent confidence bound of'

    72 to 97 percent or 3,500 to 100,000 ppmv.


2.  90 percent of total emissions will be greater than a screening

    value of 5,000 ppmv within a 90 percent confidence bound of

    81 to 97 percent or 1,800 to 60,000 ppmv.


3.  95 percent of total emissions will be greater than a screening

    value of 1,700 ppmv within a 90 percent confidence bound of

    90 to 99 percent or 500 to 18,000 ppmv.


4.  98 percent of total emissions will be greater than a screening

    value of 450 ppmv within a 90 percent confidence bound of

    96 to 100 percent or 50 to 3,000 ppmv.
                             VI-205

-------
                                - 11 -

E.  The control cost analysis of reasonable available control  technology
   7RACT) is flawedT"

        The estimated costs contained in the  CTG are not consistent
   with those contained in the Background Information  Document (BID)
   prepared in conjunction with the proposed  new source performance
   standard (NSPS) for VOC fugitive emissions.  For example,  the  CTG
   estimates the cost of a one-inch screw on  type globe valve at  $38
   each.  The BID uses an estimate of $45 each.  Also, the CTG esti-
   mates labor costs at $18 per hour, the BID estimate is $15 per
   hour.  Finally, the CTG estimates VOC detection instrument costs
   at $9,200.   The BID estimate is $8,500
        The Agency appears to have seriously  underestimated the true
   costs of RACT.  For example, EPA estimates the cost of capped  lines
   on that of a one-i-nch valve plus one hour  of labor.  Means  1980
   Standard Cost Index places labor costs at  $20.10 to $22.35  per hour.
   this is approximately 24 percent higher than the CTG estimate.   Fur-
   ther, estimating the cost of capped lines"  on the basis of  a one-
   inch valve plus one hour of labor ignores both the nonlinear pricing
   of valves and the wide industry use of specially lined valves.
   While the cost of a one-inch standard, steel, screw on valve is
   indeed approximately $50,  Means'  standard cost is $520 for  a two-
   inch valve  and $1,025 for a three-inch valve.  A one-inch  teflon-
   lined valve (widely used in VOC service)  lists for $133.  This is
   300 percent more than the  EPA estimate.   Finally, many open-ended
   lines are not one-inch lines,  but EPA presents no data to  support
   the supposition that one-inch is  an accurate average value.
        Both the CTG document and the BID assume a fixed 10 percent
   interest  rate.   Few,  if any affected industries can base economic
                                VI-206

-------
                             - 12 -

decisions on such a rate at a time when the prime rate is  approxi-
mately 17*5 percent.  EPA should revise their estimates to  reflect
a reasonable cost of money.
     All model plant cost calculations are based on leak frequency
data experienced by the refinery industry.  As we pointed  out  above,
the measured frequency of leaks for SOCMI differs fundamentally
from that for refineries.  For example, EPA estimates that the num-
ber of initial open-ended gas line leaks for Model Plant C will be
10 percent, not the 5.8 percent suggested by the SOCMI data in •
Table A-7, page A-13 of the CTG.  This overestimate of initial
leaks results in a falsely high estimate of material recovery  in
the CTG.
     In summary, the cost-effectiveness of the draft is not correct.
EPA should extensively rework Chapter 5.0 to properly reflect
updated, consistent cost data.  The Agency should also incorporate
accurate estimates of leak rates as determined from SOCMI  data.
     Unless accurate cost data and more flexible and appropriate
technical data are included the states will not be able to incor-
porate into their revised implementation plans the most cost-
effective controls necessary to attain the ambient air standards.

EPA requires a SIP revision for every alternative work practice and
every performance standard variance.  CMA beleives the states  should
have the authority to make decisions on alternative methods without
a SIP revision and without EPA approval.

     Sections 101 and 107 of the Clean Air Act expressly place on
the states the primary responsibility for preventing and controlling
air pollution at its source.  Section 110 requires the states  to
                              VI-207

-------
                             - 13 -

submit to the Agency SIP's which provide for implementation, main-
tenance and enforcement of. national ambient air quality standards
(NAAQS) set by EPA.  Once the broad and general SIP -is approved by
EPA, we believe that the states should have the authority to man-
age the air pollution..programs described in their SIP's on a day-
to-day, case-by-case basis without undue interference from the
Agency.  This is not presently the case and this CTG exemplifies
the problem.
     EPA has insisted that it be permitted to second-guess, by
means of an individual SIP revision, each and every state exercise
of discretion with regard to emission limits on individual sources.
EPA has recently proposed to amend 40 C.F.R. Section 51.9 to re-
quire SIP revisions (approved by the Agency) every time the state
grants any variance, extension of time, revision or waiver of an
individual source's emission limits.  See 44 Federal Register 67675
(November 27, 1979).
     CMA recognizes the need for EPA to ensure that state plans
maintain the NAAQS in prevention of significant deterioration (PSD)
areas and to ensure reasonable further progress (RFP)  toward attain-
ment in nonattainment areas, and, therefore, the need for a formal SIP
revision in cases where attainment or maintenance of the NAAQS may
be jeopardized.  Where that is not the case, however,  the states
should be able to establish or revise individual source emission
limitation on a case-by-case basis without a SIP revision and with-
out EPA approval.  This flexibility should include the authority
to grant variances, exemptions, time extensions and waivers for
such reasons as technological feasibility, economic hardship, energy
                           VI-208

-------
                             - 14 -

considerations or impracticality, so long as attainment or RFP
toward attainment will be maintained.
     This is particularly true with regard to the grant of per-
mission by the states to approve alternate programs for VOC emis-
sion control.  Since a source's use of the alternate program by
definition means that its net emissions will not exceed levels
allowed by the SIP, or in any way endanger a state's RFP toward
attainment, it makes no sense for EPA to require a separate SIP
revision every time a state permits the use of an alternate pro-
gram for VOC fugitive emission control.  Yet the Agency has insisted
upon maintaining such control over the states.  CMA submits that
EPA's position is not supported by the language of the Clean Air
Act, and is designed solely to permit EPA to second-guess that
state's decision as to the appropriate mix of emission controls to
be used in achieving the national ambient standards, and to place
enforcement of those requirements in the hands of the Agency rather
than the states.
     Although this requirement may, at face value, not seem unduly
burdensome, its practical consequences are severe.  The SIP revi-
sion requires:
          a)  one or more public hearings preceded by at least
              30 day's prior notice to the public,
          b)  submittal by the state of the proposed revision
              (once the state has approved it) to EPA for
              review,
          c)  full review by EPA, and
          d)  a decision by the Administrator to approve the
              revision.
                              VI-209

-------
                               - 15 -

   In addition, EPA's interference creates a credibility problem
   that undermines the state's ability to implement and enforce  its
   SIP-  The duplication of effort involved in EPA's second-guessing
   of the states involves a great waste of state resources  and an
   added cost burden to individual sources that is wholly unjustified.
       CMA recommends that the individual state director should evalu-
   ate the data submitted in requesting a performance standard or
   alternative program pursuant to a generic procedure approved  by the
  * Agency.  If the generic procedure were followed, the equivalent '
   program would be enforceable by both the state and federal EPA.   He
   should determine if these data are sufficient to support that per-
   formance standard or alternative program.  If they are sufficient,
   he should declare the programs equivalent to the state regulation.
   It should not be necessary then for the director to submit a  SIP
   revision to the Agency.

•  The Agency has not made provisions to exclude inaccessible valves
   from the routine monitoring and maintenance requirements.
       In our review of the proposed CTG we identified an  area  which
   the Agency has apparently overlooked.  This area is inaccessible
   valves.  These fall into two general categories, valves  inaccessible
   for safety reasons and valves inaccessible because of elevation
   and/or configuration.
       Certain chemical processes are carried out at such  extreme
   conditions of temperature or pressure, or the chemicals  themselves
   are so unstable or hazardous that the operation is done  behind bar-
   ricades and the like, and, for safety reasons, personnel are  not
   allowed in these areas while the unit is in operation.
                                VI-210

-------
                             -  16  -




     As the Agency  is well  aware,  in  existing facilities many valves


are not routinely accessible because  of  elevation or because access


to the valve bonnet is  restricted.  Many of  these valves can be


eliminated in  an entirely new plant.   But they become a problem in


an older plant that becomes subject to this  CTG.
                      »•

     To correct these problems, we propose valves that are inacces-


sible for safety and other  reasons be excluded from the requirement


of Section XX.030(A) and subject to a new Section XX.030(A)(1).


^Accordingly, we recommend that  the new Section XX.030(A)(1)  be


added to the CTG as follows:


     "An owner or operator  of a source subject to the  re-


      quirements of Section XX.030(A)  may, for valves  that


      are routinely inaccessible for  safety  reasons, monitor


      each inaccessible valve for  leaks  after a process unit


      overhaul prior to startup by pressuring with nitrogen


      to the system process pressure  or  100  psig, whichever


      is less, and  checking with a soap  solution  for bubbles,


      or other equivalent test  method.
                              VI-211

-------
                                 - 17 -
SPECIFIC COMMENTS
              Page 1-3
         The CTG requires a plant to submit quarterly reports to the
    state.  CMA feels the individual states should decide how often
    they need a report.  Since this is not necessary to ensure attaining
    the ambient standards, the Carcinogen Policy discussion has no
    relevance in developing RACT and should be deleted from the final
    CTG.
              Page 2-6
         The CTG states that the barrier liquid in dual mechanical
    seals is at a lower pressure than the stuffing box, and leakage
    will be into the barrier liquid.  However, chemical industry experi-
    ence indicates the leak may be in the opposite direction because
    for some processes the barrier fluid may be at a higher pressure.
              Page 2-9 (Section 2.2.1.3)
         Butterfly and diaphragm valves should be added to the valve
    list.
              Page 2-16 (Section 2.2.1.7)
         Legitimate "block and bleed" systems ought to be exempt from
    the monitoring requirements for open-ended valves.  A discussion of
    "block and bleed" systems is provided in the discussion of Page 3-12
    below.
              Page 2-19
         In order to apply emission factors the Agency assumed that
    one-half of SOCMI liquid service sources are in light liquid service.
    The data are available in Table A-3.  EPA should have analyzed
    them.
                                 VI-212

-------
                              - 18 -

           Page 2-21
      Reference 39 in the footnotes was revised in April 1980.  This
 reference should be updated as Table 2-2 has been.
           Page 2-23
      Table 2-4 was derived from refinery data.  It should be re-
 vised to include SOCMI data.
           Page 3-5
      Reference 4 in the footnotes was revised in April 1980 and
,should be updated.
           Page 3-6 (Table 3-2)
      EPA continues to base occurrence and recurrence leak rate
 assumptions on Table 4-2 from the final technical support document
 (see Attachment E) which are the same as those taken from the draft
 technical support document.  TCC commented on the issue of occur-
 rence and recurrence in a letter submitted to EPA in June 1980.
 These comments have not been addressed by EPA and cannot be ignored.
 We wish to emphasize that the points raised in TCC's comments on
 this issue are still valid.  In addition to TCC's earlier comments,
 we offer the following:
      The Agency's assumption on nonlinear leak recurrence with time
 is not based on any data, but rather an "engineering judgment."
 The assumption that twice as many leaks will be found annually as
 compared to quarterly, and twice as many leaks will be found quar-
 terly as compared to monthly is simply not logical and is not sup-
 ported by the record.  TCC recommends using a linear leak recurrence
 rate with time in the absence of data.  CMA concurs with this
 recommendation
                               VI-213

-------
                             - 19 -

     The assumption that all source types will have a uniform re-
currence rate (20 percent)  per year is not a logical assumption
and is not supported by the record.  In the absence of data, a more
logical assumption is that recurrence will vary in proportion to
occurrence.  EPA should examine the occurrence/recurrence issue by
analyzing the data from the SOCMI maintenance study.
          Page 3-8 (Table 3-3)
     Refinery values should be replaced with SOCMI data.
          Page 3-9
     Equation (1) should read:  3=1-%  NOT  B = 1 - nm
                                         N                2~N
          Page 3-10
     These figures should be based on SOCMI data and not refinery
data.  Refinery values should be replaced with SOCMI data.
          Page 3-11  '
     Please refer to the comment on linear occurrence/recurrence
from Page 3-6 above.
          Page 3-12 (Section 3.2.1)
     Many processes use "block and bleed" techniques to avoid pro-
cess contamination, or where explosive mixtures or reactive mixtures
are present.  For example:
     o  Many processes require periodic thawing to remove
        frozen water.  Hot and cold gas systems are connected
        at key points to inject and remove hot and cold dry
        gas and thus remove the moisture.  When not in use,
        the systems are isolated by two block valves with a
        vent valve between to assure that no leakage occurs
        between systems.  This vent valve cannot be plugged
        for process and safety reasons.
                            VI-214

-------
                             - 20 -

     o  Insurance companies require a double valve and vent

        configuration for fuel valves to preclude fuel from

        entering a combustion unit prematurely-  These three

        valves are operated as a unit.  Later versions incor-

        porate this configuration into valves with vented
                      *
                      f
      - bodies or three-way valves.  The vents must not be

        plugged.

These are specific examples for safe process operation.  Such valves

are easily identified, and the vent valve portion must be exempted

from the proposed CTG.

          Page 4-5 (Table 4-4)

     Table 4-4 is not reproduced fully from Table 4-6 in the BID.

This table assumes nonlinear recurrence.  A detailed explanation

of this issue can be found in the comments regarding Page 3-6 of

the CTG.


The following comments refer to the model regulations proposed in

Chapter Six of the CTG:

          XX.010 - Applicability

     The exemptions should include inaccessible valves.  Our

detailed explanation of this issue is located in the General Com-

ments section of this document.  In addition we refer to a

January 28, 1981 letter from Janet S. Matey to Fred L. Porter

concerning this matter.  (Attachment F)

          XX.020 - Definitions

     Leak;  CMA believes the existing data support a much higher

level than 10,000 ppmv as the definition of a leak.  This issue is

detailed in our General Comments.
                             VIr215

-------
                             -  21 -

     Volatile Organic Compound  (VOC);  The  definition  includes  any
organic compound which participates  in atmospheric photochemical
reactions or is measured by the applicable  test method or  equiva-
lent state method.  The test method  measures  several compounds
which are nonreactive in terms  of photochemical conversion to
ozone.  We recommend these nonreactive organic compounds be excluded
from the RACT requirements.  CMA recommends that after the present
definition the Agency should add or  note in the final  CTG  as fol-
lows :
I
     NOTE— "The following compounds are excluded:  methane,
             ethane, dichloromethane, 1,1,1 trichloroethane
             (methyl chloroform), trichlorotrifluoroethane
             (CFC-11), dichlorodifluoromethane  (CFC-12),
             chlorodifluoromethane  (CFC-22),  trifluoromethane
             (CFC-23), dichlorotetrafluoroethane  (CFC-114)
             and chloropentafluoroethane (CFC-115)."
          XX.030 - Standards
     C(l) - The requirement for a readily visible tag  should be
replaced with readily visible form of identification as specified
in the NSPS.
     C(4) - A "Section C(4)" should be added  which would allow a
delay of repair beyond the next scheduled unit turnaround.  We
believe the Agency's requirement that all repairs may  not  be
delayed under any circumstances beyond a process unit  shutdown is
not practical.   We concur that  many repair actions which cannot  be
technically or safely conducted while the process is in operation
will be remidied during a process shutdown.   As the Agency is aware,
                              VI-216

-------
                              -  22  -

most  schedule  shutdowns  are  on  an  annual  basis  or based upon oper-
ating performance  of  the process unit.  As  a  result,  there may be
some  limited instances where replacement  of leaking equipment may
not be  available until after the shutdown is  completed.  Such
instances  include  (but are not  all inclusive) abnormal near-term
                      .-
demands for replacement  parts that exceed the quantity of replace-
ment  parts normally maintained  in  stock and that  cannot be replaced
on short notice, the  replacement parts and  equipment  that are not
x"off  the shelf" items and require  a  long  lead time for delivery •
 (i.e.,  some pumps  require special  order and take  up to one year
for delivery)  and/or  unforseen  manufacturers,and/or delivery delays
 (e.g.,  strikes, fires, raw material  delays  in delivery).   Any one
or a  combination of the  above scenarios would result  in the neces-
sary  replacement part(s) not being available  until after the next
scheduled  shutdown.
      Since the proposed  CTG, if incorporated  in a SIP, would make
continued  operation after such  a shutdown a violation of the Clean
Air Act, we strongly  recommend  that  EPA incorporate a limited exten-
tion  provision into the  final CTG.   We envision placing the burden
of requesting  such a  request on industry  by requiring a formal
submittal  to EPA in which the source must justify the need for the
further delay  in repair  and  the projected time  frame  for achieving
compliance.  We recommend that  Section XX.030C(4)  be  added to the
CTG to  include the following regulatory language:
      "Delay of repair will be allowed beyond  a  process unit
      shutdown only where and for  the period  of time  a source
      demonstrates to the satisfaction of the state agency or
      delegate that repair of a leak by replacing physical
                              VT-?17

-------
                             - 23 -

      equipment exceeded the normal stock of spare parts and
      cannot be delivered until after the next shutdown, a
      special order of a part is required and cannot be de-
      livered until after the next shutdown, and/or because
      of unforseen manufacturers and/or delivery delays, the
      replacement parts cannot be delivered until after the
      next shutdown."
The consequences of not including such a provision in the CTG
could result in unanticipated and costly continuances of shutdown
until the repair parts are obtained, or in the exposure to signifi-
cant criminal and civil penalties for resuming operation without
repairing all leaks.  We would consider it arbitrary and capricious
not to provide such a remedy where the source has acted in good
faith to repair all remaining leaks at the next scheduled shutdown,
but solely because of uncontrollable events the necessary repair
parts are not available.
          XX.030(D)
     The 24 hour repair requirement is not possible or reasonable.
We recommend this requirement be changed to require weekly visual
inspection of all pumps in light liquid service, monitoring within
24 hours, attempt at repair within 5 days and repair within 15 days.
          XX.030(E)
     The option that the director may require early shutdown based
on the number of leaking components which cannot be repaired should
be deleted.  After the first shutdown more leaks should be repaired.
Then the occurrence/recurrence related repairs should be sufficient
to control emissions.  Shutdown will cause an increase in emissions,
and the effort could be counter productive to the environment.  In

                             VI-218

-------
                             - 24 -

addition, the statistical variation in the monitoring data is so
large that the decision to shutdown would be highly questionable.
Finally, plant shutdowns are energy consumptive and costly.  An
imposed shutdown should only be required in an extreme situation
when an imminent and substantial endangerment to health exists.
Any such situation would also be a safety hazard which the Occupa-
tional Safety and Health Administration (OSHA) provisions would
require  fixing and will probably be remedied even if the Agency
deletes this provision.
          XX.030(F)
     The last sentence on Page 6-3 ("The sealing device may be re-
moved only when a sample is being taken or during maintenance opera-
tions.") should be deleted.  Legitimate "block and bleed" systems
ought to be exempt from the monitoring requirements for open-ended
valves.  A discussion of "block and bleed" systems is provided in
the discussion of Page 3-12 above.
          XX.030(G)
     The following should be added to the end of the first sentence
in this section:
     "For example, sources located inside a building under a
      negative pressure with the ventilation through a control
      system with a removal efficiency equal to or greater
      than that achieved by a leak detection and repair program. "
          XX.040(A)(3)
     The provision precludes the use of equipment which is not cali-
brated with methane.  The CTG indicates that fugitive emission
detection instruments should be using calibrations of 10,000 ppmv
                              VT-?1Q

-------
                             - 25 -





of methane in air.  Method 21 indicates, however, that the VOC



instruments should respond to the organic compounds being pro-



cessed and the detectors used in these instruments can include



catalytic oxidation, flame ionization, infrared absorption and



photoionization.  However, HNU Systems (manufacturers of a photo-
                      *


ionization unit) indicate in their sales literature that the



methane does not have a photoionization response.  Typically, iso-



butylene mixtures are used as the calibration gases.  Accordingly,



we recommend the language be modified as follows:



     "Calibration gases shall be a mixture of methane and air

                           •

      at a concentration of approximately 10,000 ppmv methane



      except as provided in Paragraph 2.3 of Method 21."



          XX.050 (A) (1)



     "Tag number" should be "identification number" to allow more



flexibility.



          XX.060(A)



     The Agency should not specify the reporting frequency for a



state.  This section should provide the state with the flexibility



to require the reporting frequency to meet SIP needs, in order to



achieve ambient objectives.



          XX.060(A)(2)



     The number of valves, pumps and/or compressors monitored as



leaking is irrelevant for purposes of determining compliance.  The



only pertinent facts are the number of pumps, valves and/or com-



pressors not repaired during the previous quarter and a single



plant certification of compliance with the technical and record-



keeping requirements.  Compliance can be determined adequately by
                             VI-220

-------
                             - 26 -




evaluating the data submitted as required by Section XX.060(A)(1).



Accordingly, XX.060(A)(2) should be deleted.


          Section 6.2.1
                                                             s,

     The requirement for quarterly reports should be deleted.


Again, the states should determine reporting requirements.   Refer
                      <•
                      •

to our earlier comments.


          Page 6-9 (Section 6.2.5)


     The draft CTG requires a state to obtain a SIP revision for


every alternative leak detection and repair program.  This require-


ment should be deleted from the final CTG.  We provide a  full


explanation of this issue in the General Comments section above.


          Page 6^11 (Section 6.2.5)


     In order for an equivalent program to be approved, the  CTG


requires a SIP revision.  Again, this requirement should  be  deleted


from the CTG.


          Page A-13 (Appendix A, Footnote C)


     The reference should read 4100 and it applies to hexane only,


not to other VOC's.  Each compound has its own instrument response


factor.
                              VI-221

-------
                           REFERENCES


1
  Response Factors of VOC Analyzers at a Meter Reading of 10/000
   :pmv for Selected Organic Chemicals, Radian Corporation, February 5,
   981.
I
  Evaluation of Maintenance for Fugitive VOC Emission Control, EPA
  Cincinnati EMSL, Unpublished.
                      •-
2
  Response Factors of VOC Analyzers at a Meter Reading of 10,000
   pmv for Selected Organic Chemicals, Radian Corporation, February 5,
   981.



  Background Information Document; Fugitive Emission Sources in 'the
  Synthetic Organic Chemicals Manufacturing Industry, Emissions
  Standards and Engineering Division, Chemical and Petroleum Branch,
  Environmental Protection Agency, November 1979.
                             - 27 -

                             VI-222

-------
COHPARI30M OF LEAK fRKQUEHClES FROH VARIOUS EPA STUDIES
Source Type
Mange*


Proceaa Orelna


Open Ended tine*


S Agitator Seal*
i
ro
Relief Valve*


Valve*


Pump*

Coa>pree*ora


SCO/JM
2/5/fll
Service
Gee
tight liquid
Heavy liquid
Gee
tight liquid
Heavy liquid
Ca*
tight liquid
Heavy liquid
Ca*
tight liquid
Heavy liquid
Caa
tight liquid
Heavy liquid
Caa
tight liquid
Heavy liquid
tight liquid
Heavy liquid
Oaa




24
X of Source*
> 10. 000 ppm
4.6
1.2
0.0
.4
.8
.1
.8
.9
1.3
14.3
0.0
0.0
3.5
2.9
0.0
11.4
6.4
0.4
a.a
2.1
6.9




Red ten SOCHI Data (1)
Unit Study-OVA-IOa-Hethene
95Z Confidence
Interval 1
(3.6. 3.8)
(0.9. 1.8)
(0.0. 0.6)
(0.3. 8.4)
(2.3. 5.8)
(0.9. 23.3)
(4.4. 7.3)
(3.3. 4.6)
(0.3. 2.8)
(0.4. 57:9)
(0.0. 36.9)
(0.0. 97.5)
(0.7. 10.0)
(0.3. 10. 1)
(0.0. 70.8)
(10.8, 12.1)
(6.1. 6.8)
(0.2. 0.7) .
(6.6. II. 1)
(0.3. 0.7)
(0.9. 22.8)




Hiwber
Icreened
1443 *
2897 )
607
83
527 )
28 J
923
3603
477
3
a
i
85 .
69 )
3 '
9668
18294
3632
647
97
29




Radian Refinery Data (2)
9-Unlt Study-m-He«*ne
X of Source 931 Confidence
>IO.OOO MM Screened

0 (0. 1)


3 (0. 7)


H/A


N/A


8 (1.5. 12)

10 (8.5. 14.3)
12 (8.0. 13)
0 0
23 (17.5. 27)
2 (0. 4)
33 (22, 43)




Huaiber
Screened

2030


223








148

683
1019
522
470
292
145


























>
H
SI
h

-------
      Soured Typo
      Pruccaa llralna
Service

Caa Service
Light Liquid
Heavy liquid

Gam
Light liquid
Heavy liquid
      Open Kmled Llnea  flaa
                        Llftlit liquid
                        Heavy liquid
ro
ro
      AH! tutor Seal*
      Mull.-f Valvee
      Vulvua
Caa
Llnht liquid
Heavy liquid

Caa
Light liquid
Heavy liquid

Caa
Light liquid
Heavy liquid

Light liquid
Heavy liquid
                                                                        TABLE 2 Continued

                                                     COHI'AHISOH OK LEAK FKKqUKMCIES fBOM VARIOUS EfA STUDIES

                                                                  Radian Refinery Data (1)
                                                                  13 Unit StudyvrLV-llenane
Z of Sourcea
 >10.000 pp.
                                                    o.s
                                                    4.?
                            7.7
                                                    H/A
    a.i
   12.
  ' 11.
<*) 0.

    2.
    3.
• 951 Confldanca
    Interval
                            (0-6)
     (2-13)
     (7-16)
     (•-14. J)
     (0-1)

     (19-26)
     (0-5)
Nuaibar Screened
                                                  2094
                            257
                                                   129
     252
     563
     914
     465

     470
     292
                                                                                                                                     Note*
(1)  Reference!  Blackinttli,  llarrla
    i Langley.  frequency of  leak
    Occurrence for Plttlnga In SOCHI.
    Radian Corp. for EPA, September,
    1980.
(2)  Reference!  tfatharold & provoit.
     Enlaalon Factor* and Frequency
     of Leak Occurrence for Fitting*
     In Refinery Proceaa Unlta.
     Radian Corp. for EPA.  EPA-
     600/2-79-044. February 1979.
(3)  Referehcei  Uethrold, Provoat
     t Salth.  Aaaeaaaent of
     Ataoapherlc Ealaalona from
     Petroleun Refining Voluae 3
     Appendix II.
     Radian Corporation for EPA.
     KPA-600/2-80-075C April 1980.
(4)  69Z of acreened aourcea eilialng
     acreenlng valuaa repreaentlng
     1.9Z of total leakage.
(5)  11.3X of acreenad aourcea
     •laalaR acreenlng valvea
     repr eon ting 22S of total
     leakage.
      Conpreeaora
                        Caa
                               3J
                            (26-43)
                                                                                                   142

-------
                                                ATTACHMENT  B
                CONCLUSIONS .REGARDING DIFFERENCES BETWEEN*
                	SOCMI AND REFINERY LEAK STUIDES '
 1. The leak frequency of valves in gas service within SOCMI is similar.to
    that of valves within the refining sector not adjusted for differences
    in calibration and instruments between studies.

 2. The leak frequency of valves in light liquid service with SOCMI
    is roughly half'the  frequency of similar valves in the refining
    sector, not adjusted for differences in calibration and instruments
    between studies. The 95Z upper and lower confidence bounds do not
    overlap for the two valve data sets.

 3. The leak frequency of valves in heavy liquid service within SOCMI is
    similar to that of.valves within the refining sector not adjusted for
    differences in calibration and instruments between studies.

 4. The leak frequency of pumps in light liquid service within SOCMI
    is roughly 1/3 the frequency of similar pumps in the refining sector,
    not adjusted  for differences in calibration and instruments between
    studies.  The 95% upper and low confidence bounds do not overlap
    for the two pump data sets.

 5. The leak frequency of pumps in heavy liquid service within SOCMI
    is similar to that of pumps within the refining sector not adjusted
    for differences in calibration and instrument between studies.

 6. The leak frequency of. compressors within SOCMI is greater than
    4.8 times less than that of compressors within the refining sector
    not adjusted  for differences in calibration and instruments between
    studies.  The upper 952 confidence bound of the SOCMI data set
    approaches the lower 95Z confidence bound for the refinery data set.

 7. The leak frequency for relief valves in all service within SOCMI
    is less than  that of relief valves in the refining sector by greater
    than a factor of 2.2 not adjusted for differences in calibration
    and instruments between studies.  The 952 confidence bounds for the
    SOCMI and refinery relief valve data sets overlap.

 8. The leak frequency of process drains within .SOCMI is similar to
    that of the refining sector not adjusted for differences in calibration
    and instruments between studies.

9.  The leak frequency of flanges in all services within SOCMI is higher
    than that of  flanges in the refining sector not adjusted for differences
    in calibration and instrument between studies.  The 952 confidence
    bound for the SOCMI and refinery flange data sets overlap.
                           VI-225

-------
10.  In a 342 item survey conducted in 1979 at Exxon Chemical Americas
     Baytown Chemical Plant the Bacharach TLV and the Century OVA-108 were
     compared in side by side studies using both hexane and methane as
     calibration gases.  The results are tabulated in Table 3; the following
     conclusions are noted.

     -  36 percent more leaks are found with the .Century calibrated
        on hexane as compared to the Bacharach calibrated on hexane.

     -  3 percent more leaks are found with the Century calibrated on
        methane as compared to the Bacharach calibrated on methane.

     -  25 percent more leaks are found with the Bacharach calibrated
        on methane as compared to hexane.

     -  5 percent more leaks found with the Century calibrated on
        hexane as compared to methane.

     -  29 percent more leaks are found with the Century calibrated on
        methane as compared to the Bacharach calibrated on hexane.
                       VI-226

-------
                                                   ATTACHMENT C
                         Results of a Test Fugitive
                        Emissions Monitoring Program
                         Conducted At Exxon Chemical
                      Baytown, TX in November of 1979
Number of items
found leaking with: *

  Bacharach TLV cal.                                  35 out of 232
  with methane

  Century OVA-108 cal.                                36 out of 232
  with methane

  Bacharach TLV cal.                                  28 out of 232
  with hexane

  Century OVA-108 cal.                                38  out of 232
  with hexane
NOTES;

  1.  Study included pumps, compressors, block valves, control valves,
      and safety valves.

  2.  Total number of items monitored was 342.  Only the results common
      to all combinations of instrument and calibration gas are presented
      above.  Two sets of safety valves were not monitored with all
      combinations.  One compressor was down'and not monitored with all
      combinations.  One set of dryer valves was not monitored with all
      combinations.  Maintenance was performed on one set of block valves
      between studies, these results were not included.

  3.  Production units monitored included, a paraxylene crystallization
      unit, a naptha rerun unit, a propylene concentration unit, and
      a paraxylene adsorption unit.

  4.  2" and smaller valves were excluded from this study.
                         VI-227

-------
                           TABLE 4-2.  ESTIMATED OCCURRENCE AND RECURRENCE RATE OF LEAKS FOR VARIOUS MONITORING INTERVALS
     Source type
                          Estimated %    Estimated percent
                          of sources     of sources leaking
                    leaking above 10,000 at above 10.000 ppm
                   ppm using SOCHI data      Initially9
                Estimated percent of
                initial  leaks which
                are found leaking at
               subsequent inspections "
             Annual  Quarterly Monthly
                              Estimated percent of
                               sources which are
                                found leaking at
                             subsequent inspections6
                           Annual   Quarterly  Monthly
   Pump seals
    Light liquid service   .  a.a
    Heavy liquid service     2.1
                                                23
                                                 2
               20
               20
          10
          10
                                                                                   5
                                                                                   5
                                                                4.6
                                                                0.4'*
          2.3
          0.2
          1.2
          0.1
ro
00
In-line valves
  Vapor service           n>4
  Light liquid service     6.4
 'Heavy liquid service     0.4
10
12
 0
20
20
20
                                                                         10
                                                                         10
                                                                         10
                                                        5
                                                        5
                                                        5
2.0
2.4
0.0
1.0
1.2
0.0
0.5
0.6
0.0
Safety/relief valves

Compressor seals

Flanges
   NOTESi
3.5 (gas)
2.9 (It. liq.).

6.9
4.6 (gas)
1.2 (It. liq.)
  0 (hv. liq.)
                                                    8
                                                   33
                                                               20
                                                               20
                                                                  20
                         10
                         10
                                                                         10
1.6
6.3
                            0.0
                                      0.0
                                      3.3
                                                    0.0
          0.4
          1.7
                    0.0
                                                                                                                     9"
                 ^Approximate fraction of sources having  leaks  equal to or greater than 10,000 ppm prior to repair.1
                  Approximate fraction of leaking sources that were repaired but found to leak during subsequent
                  inspections.   These approximations  are  based on engineering judgment.
                 cApproxinute fraction of sources that were  repaired but found to leak during a subsequent Inspection*.
                  These approximations are the product of the information presented in footnotes a and b.

         1.   Reference 9 is:   Hetherold & Provost. (Radian. Corp) emission factors and frequency of  leak  occuronce
             for fittings in  refinery process units.   Prepared  for EPA.  EPA-600/2-79-044.   February 1979.
        2.  Reference  for thin  table  int  IISEPA,  Emission Standards and Engineering  Division.  VOC Fugitive
            e<*l ailens  in SOCHI  - Background  information  for proposed standards.   KPA-450/3-00-033a, November,  1800.

-------
       CHEMICAL MANUFACTURERS ASSOCIATION
                                   January 28,  1981
Mr. Fred L. Porter
Assistant to the Director
Emission Standards & Eng. Div.
U. S. E. P. A.
Office of Air Quality Planning  &
  Standards
Research Triangle Park, NC  27711

Dear Mr. Porter:

     We wish to take this opportunity to thank you for
meeting with CMA representatives  on  January 13 to discuss
ongoing CTG development activities by EPA.   During our
various discussions with you  and  Bill Tippett, the question
of valve accessibility was  raised.   Specifically, various
CMA member company representatives wanted to know how this
issue was to be addressed in  the  forthcoming CTG on fugitive
VOC emissions from SOCMI.   The  purpose of this letter is to
state CMA's position on this  question for your consideration
in the development of the subject CTG.

     Briefly, a valve should  be considered accessible for
VOL monitoring purposes only  if the  valve can be reached
safely by a monitoring crew from  ground level or from a
fixed platform.  This would exclude  valves requiring access
for monitoring hy moveab.le  ladder or "cherry picker."  Addi-
tionally, the safety aspect needs further discussion.  Certain
valves in VOL service are in  high temperature and/or high
pressure locations which are  not  routinely accessible for
monitoring.  Some valves, within  SOCMI, are located in sealed
operating areas because of  toxicity  considerations or in
barricaded areas due to explosive concerns.  Such valves are
not routinely accessible for  monitoring due to safety concerns,

     In addition, certain valves  are totally enclosed by
insulation.  In some cases  they are  enclosed by drip covers
to protect against corrosive  leaks.   Such valves may not be
easily accessible for routine monitoring in existing units.
                       ...  VI-229
      Formerly Manufacturing Chemists Association—Serving the Chemical Industry Since 1872.
   2501 M Street, NW • Washington, DC 20037 • Telephone 202/887-1100 • Telex 89617 (CMA WSH)

-------
page 2


The number of valves deemed inaccessible due to the above
considerations will vary depending on the physical and
chemical properties of the organic chemical and layout of
the individual production unit.

     We suggest that EPA consider requiring VOC^ monitor ing
of inaccessible valves only when access is required for
such valves to be manually operated or maintained under
normal plant operating circumstances.  Valves inaccessible
for safety reasons should be monitored after 'major process
overhaul  (prior to start-up) by pressuring the system with
nitrogen.  The system should be pressured to process pressure
or 100 PSI, whichever is less.  It should be checked for
bubbles 'with a soap solution or other equivalent test methods.
These valves should not be included for purposes of Sections 483
or 484.
                                 Sincerely yours,
                                  Janet  S. Matey
                                  Manager, Air Programs
 JSM:hec

 cc:   B. Tippet
                         Vi-230.

-------
                            ATTACHMENT D
                  A Comparison On An Equivalent Mass
                   Basis of Refining and SOCKI Data
    Refinery
     Data (1)

TLV Hexane
SOCMI Data (2) (3)

   TLV Hexane
SOCMI Data (2)(3)

   OVA Methane-
Valves
  Gas Service

  10,000 ppm - 0.38 #/hr
  @ LK - 0.038 #/hr
 Mass Equivalent
      Value
     15,900
     35,000
Valves
  Liquid Service
  10,000 ppm =0.5 #/hr
  @ LK - .05 #/hr
     10,700
     75,600
Pump Seals
  Lt. Liquid
  10,000 ppm = .39 #/hr
  0 LK - 0.39 #/hr
     33,600
     94,000
NOTES:

1.  Reference:  ¥etherold & Provost.   "Assessment of Atmospheric
    Emissions from Petroleum Refining."  Conducted by Radian Corp.  for EPA.
    EPA-600 2-80-075C, April, 1980.

2.  Reference:  Langeiy and Wetherold.  "Evaluation of Maintenance  for
    Fugitive VOC Emission Control."  Conducted by Radian Corp.  for  EPA,
    February 1981.

3*  These data are based on bagged emission rate screening value
    correlations developed in Reference 2.
                               VI-231

-------
     5. NAPCTAC Member William Relter
      Allied
      Chemical
     Corporate Environmental Affairs
     P.O. Box 2332R
     Morristown, New Jersey 07960                    ..   ,_ /•« -,   -. « o -.
                                         March d3 ,  lyo i
Mr. Don Goodwin
Director, Emission  Standards & Engineering Division
Office of Air Quality  Planning & Standards
U.S. Environmental  Protection Agency
Research Triangle Park,  NC  27711

       EDITOR'S NOTE:

                   NAPCTAC member William Reiter was  unable to attend the
       meeting on March 18, 1981,  So  that he could contribute his views to EPA
       and fellow Committee members, Mr, Reiter wrote a lengthy letter to the
       chairman.  The  contents of that letter have been divided by subject and
       are included in the relevant sections of the minutes,  The portion of the
       letter that applies to this section follows.


   4)   CTG for Control of Volatile Organic Fugitive  Emissions  from
       Synthetic Organic Chemical  Manufacturing Industry,  Polymer
       and Resins.,	

       a)   I reiterate  my previous comment that the  submissions
            made during  the NAPCTAC review of the NSPS covering
            fugitives  from the Synthetic Organic Chemical
            Manufacturing Industry have not been considered in  devel-
            oping this document.    I am disappointed  that  you
            have not  seen fit to efficiently input those  comments.

       b)    I  cannot understand why a VOC  fugitive  document is being
            prepared for  polymer  and  resin  manufacturing  when there
            have  been  no  data  provided for  that  industry.

            Allied  Chemical  has submitted  information to  EPA-Region VI
            on  fugitive emission  sources at a polyethylene  plant  in
            Baton Rouge,  Louisiana.  Those  data,  based on my review,
            contraindicated  the conclusions and  directions  of the
            draft CTG.  I  find  it very inefficient  that  EPA does  not
            make  use of available data before developing  and
            releasing  documentation to the  States  and Regions.

            Again,  on Page  1-2, the document was  indicated  to be  a
            working  draft.   Why,  then, is  a working draft not
            clearly  labeled  as not  to  be used for  rulemaking?

                                     VI-232

-------
I am concerned that EPA  is mixing  apples  and oranges and
obtaining pears when it  utilizes  refinery data and
applies it directly to the Synthetic  Organic Chemicals
Manufacturing Industry and the Polymer  Industry without
any bridge for that transfer.  The data collected at
Allied Chemical plants (Hopewell,  adipic  acid;
Frankford, phenol/acetone; and Baton  Rouge,  high density
polyethylene) do not support the  position being taken .in
this document.  I feel that the actions by EPA are not
technically sound.

Appendix B lists chemicals and polymers that are to be
controlled by this regulation.  The technical guidance
that should be coming from EPA-Durham is  completely
lacking.  The data obtained from EPA's  contractors have
not been considered.  Let me elaborate:

•    Experience with the polyethylene plant  has not been
     considered.

•    Evaluation of the adipic acid plant  and process
     utilized by Allied  Chemical at our Hopewell faci-
     lity were not considered.  The experience  of your
     contractor at Hopewell showed that only "4 sources
     of fugitive VOC emissions out of 775  potential
     sources indicated any VOC emissions,  those being
     32, 230, 580 and 2,000 ppm".

•    At our phenol plant in Frankford,  PA,  to my
     recollection, fugitive emissions were  only detected
     in those facilities handling  acetone.

 Let us consider the  economic  impact of  your  actions.
 By issuing  this regulation,  you  will  force  companies  to
 enter  an  extensive  investigation and  recordkeeping
 effort which Allied  Chemical  estimates  to cost approxi-
 mately $40  per  year  per  source without  any  signifi-
 cant  reduction  in the  VOC fugitive load.   Sound
 technical judgment  would call  for the  elimination  of
 phenol and  adipic acid from your  list.   It  is  probable
 that  if such a  review  were  made,  utilizing  the data  that
 you have  already paid  for,  a  significant reduction  could
 be made in  the  chemicals affected.  The same is  true  for
 the polymer area.

 Refer  Page  1-3,  fourth line from  the  top.   The  direction
 given  is  very specific and  allows no  exclusion  but
 that  monitoring must  be  done  once every three  months.

 •     Suppose there  are no leaks.   Why  should repetitive
      inspection be  accomplished  at a cost of about  $40
      per  source per  year?

 •     This should be  a  State decision.

 •     If no, or  minor,  leaks are found,  then the  fre-
      quency should be  reduced.

                      VI-233

-------
Page 1-3, third paragraph.   The  document requires the
plant to submit a  report  once  per  quarter.   What is the
need for a quarterly  report?  What will the State do
with it?  It  is probably  more  logical to have the data
stored at the plant with  a  certification being given to
the State.  The entire  aspect  of generating report on
report should be avoided.

Since the plant "must keep  a record	a
copy...available on request",  what is the justification
for submission of  the report?   Certification that the
compliance is being effected should be sufficient.

Page 2-1.  I  take  issue with the statement  made in the
third paragraph that  "the equipment in SOCMI process
units is similar to equipment  in polymer and resin manu-
facturing units".  I  suggest your  engineers compare a
polyethylene  plant or a PVC plant  to a phenol plant.
They both use pumps,  reactors,  and valves.   However, the
processes are so different  that  to conclude that
emissions from organic  chemical  jplants and  polymer
plants are in the  same  ballpark  is wrong.  Rather, I
suggest  that  you have developed  sufficient   data in
contracts with the Radian Corporation to provide a logi-
cal technical foundation  for the CTG.  That should
be utilized,  and not  unsubstantiated opinions.

Page 2-1, next to  last  sentence.  Statement identifies
sources  and includes  cooling towers.  A review with your
branch chiefs should  recall the  decision that was made
when Dave Patrick  headed  the fugitive study to remove
cooling  towers as  a controlled source.  I believe the
EPA decision  to keep  cooling towers out was also* iden-
tified in discussions of  the NSPS.  In citing cooling
towers in the document, it  will  require the State or '
Region engineers to include them in the fugitive
emission control plan.  This would be more  stringent
than the NSPS and  is  without a technical base for control.

Flanges  are also identified as a source.  These were
excluded in developing  the  NSPS.  Yet, they are cited on
Page 2-1 and  2-16. Cooling towers are again cited on
'Page 2-14.

Page 3-3.  As you  will  recall  at the NAPCTAC meeting
reviewing the NSPS on fugitives, both you and I were
very concerned about  the  statement in the BID that a
block valve could  be  placed upstream of a relief valve.
At that  meeting, the  Committee,  I  believe,  supported the
position that there would be no  block valve upstream of
a relief.  Yet, again on  Page  3-3  in the first
paragraph, EPA states,  wa block  valve may be required
upstream of the relief  valve".   I  strongly  recommend
against  such  an unsafe  decision.

Again, we have found  evidence  that the comments
submitted on  the draft  NSPS for  fugitives have been
completely ignored.   I  find this unconscionable.

Page 4-1, first paragraph.   RACT is identified to
                        VI-234

-------
         include "weekly visual inspection of pumps in  light
         liquid service ...... ".  Economics have not been cited
         nor has the basis been established that it is  necessary
         to do this in SOCMI and polymer plants.  I believe this
         work practice requirement should be left to. the States.
         No one in today's economy is going to allow 'the leak and
         loss of hydrocarbons to continue.  The loss would be
         reported immediately and repaired as rapidly as
         possible.
         Refer Table 5-7.  I suggest  that  the  annualized cost
         data do not adequately represent  industry  conditions.
         The "B" unit having 1,666 components  has  the  same order
         of magnitude sources  that are  in  the  "G"  reactor  system
         in Allied's polyethylene plant.   Table 5-7 indicates
         that the cost per component  before  credit  is  $16.07.
         Allied's experience is that  cost  for  a unit having  about
         1300 sources is $40 per  component.  This would signifi-
         cantly change the economics  shown  in  the Table.

         I take issue with the  recovery  credits.  Our  experience
         clearly indicates that one  cannot  simply  test  a valve,
         repair it, note the decrease,  and  equate that  momentary
         change to VOC recovery on an annualized basis.   I
         believe data from the  Radian study  and the Allied study
         (which will be published) clearly  show that field main-
         tenance is not that effective.  Thus,  the  recoveries
         indicated are grossly  overstated.

         Chapter 6.  Please  refer to  my  earlier comments  on  the
         model regulation.   I  believe that  it  has been  formulated
         without sufficient  variance, without  cnsideration of
         industry maintenance  practices, without consideration  of
         the testimony given to EPA  during  the NSPS review.   I
         strongly suggest that  a  very limited  work  practices
         framework be established and the  States be allowed  to
         modify it pursuant  their individual conditions.

I believe that the comments  submitted on  the CTG's  covering
VOL storage and the VOC fugitives are of  sufficient concern  to
justify a complete re-work of  the documents.   I do  not  believe
that the documents, as currently  drafted,  are  up to the pro-
fessional excellence that the  Office  of  Air  Quality Planning  and
and Standards has normally exhibited.  I also  suggest  that
because of the situation, a  note  be  released to those  copyholders
of the draft document cautioning  them against  using the documents
until they are revised.

I hope the above will be of  some  use  to  you  in  your effort.   I
will see you at the next NAPCTAC  meeting.
                                        Best  regards,
                              VI-235
                                        W.  M.  Reiter
                                        Corporate Director
                                        Pollution Control

-------
                                                                                  U. S. Environmental  Protection Agei
                                                                                   Research Triangle  Park, N. C.  271
                                           RECORD OF ATTENDANCE
                    NATIONAL AIR POLLUTION CONTROLJECHNIQUES ADVISORY COMMITTEE MEETING

                                           March 17  and  18, 1981
PLEASE PRINT
   NAME
                 COMPANY
BUSINESS ADDRESS
  TELEPHONE
PRIMARY INTEREST
DATE
   r •. 5,  G
o
         -7 UTftt
                                   o,
                            /I
          1  A
ZlT 2
-------
  PLEASE PRINT
                                                                               U.  S. Environmental Protection Agency
                                                                                Research Triangle Park, N.  C,  27711
                                           RECORD OF ATTENDANCE

                     NATIONAL AIR POLLUTION CONTROLJECHNIQUES ADVISORY COMMITTEE MEETING

                                          March  17 and 18,  1981
     NAME
    COMPANY
  BUSINESS ADDRESS
TELEPHONE
PRIMARY INTEREST
DATE
   Saw
GCA
             Dr.
Mapd ffttl
    r       "
                GTS
                       )
3=>
I
ro
                                                >- (3
                                                      PL
                                    9
                                       ,500
                                                             4^.3-4^74

-------
                                                                                     U. S. Environmental  Protection Agency
                                                                                      Research Triangle Park,  N. C.  27711
                                              RECORD OF ATTENDANCE

                       NATIONAL  AIR  POLLUTION CONTROL^TECHNIQUES ADVISORY COMMITTEE MEETING

                                             March 17 and 18, 1981
  PLEASE PRINT
     NAME
COMPANY
BUSINESS ADDRESS
TELEPHONE
PRIMARY INTEREST
DATE
           C.
                                                       XL
                                                       " ffl
                                                                                                      S-/7- P/
5=
I
to
                                               \
           K T'V'i n t-
                                                                       3/ii/lJ
                                                                                 i^f^^t^l'li c
                                                                                    £
                           -^r. 2-t
 t\/n
             fur, AC.
                                                       /7
                                                                          n
                               fi-c /4
                                                            pr.

-------
                                                                                    U.  S.  Environmental  Protection Agency
                                                                                      Research  Triangle  Park,  N.  C.  27711
                                             RECORD OF ATTENDANCE

                     NATIONAL AIR POLLUTION CONTROL TECHNIQUES ADVISORY COMMITTEE MEETING

                                            March 17 and 18, 1981
PLEASE PRINT
   NAME
COMPANY
BUSINESS ADDRESS
TELEPHONE
PRIMARY INTEREST
DATE
                                                                                           C
                                           /) ^S^v ^ ;£
                                            Rr
                                                     C

              ,TU-
                                              l-
                                               f-K-i.
                         \ , (J! T"
                          S.4, I v<^
                                                             a

-------
PLEASE PRINT
                                                                             U. S.  Environmental Protection Agency
                                                                              Research Triangle  Park, N.  C.  27711
                                         RECORD OF ATTENDANCE

                   NATIONAL AIR POLLUTION CONTROL TECHNIQUES ADVISORY COMMITTEE MEETING

                                        March 17  and 18, 1981
   NAME
     COMPANY
BUSINESS ADDRESS
 TELEPHONE
PRIMARY INTEREST
DATE
     ue.c.r a O Qi/if/l
                                                    ?    r
  1  & I
(X/wtsco
         , /
J2 ^r^ -j///
                      TfStf
                                                    I/O C,
                                                   3//7/F/
                        3 H
                                                    CTG
                                                                                                ;
                                      p.o.
                                        ?"73-S»/|

-------
P'4ASE_PRJ]fr_
   NAME
                                                                                      U. S. Environmental  Protection  Agency
                                                                                       Research Triangle  Park,  N.  C.   27711
                                               RECORD OF ATTENDANCE
                       NATIONAL AIR POLLUTION CONTROL_ TECHNIQUES ADVISORY COMMITTEE MEETING
                                              March 17 and 18, 1981
                                COMPANY
 BUSINESS ADDRESS
TELEPHONE
PRIMARY INTEREST
DATE
                          us .
                           T/io\.''/ie«><'
                              *^> C 1 <>
*
                                                 H

-------
                                                                                           U.  S.  Environmental  Protection Agency
                                                                                            Research Triangle  Park,  N.  C.   27711
                                                   RECORD OF ATTENDANCE

                           NATIONAL AIR POLLUTION CONTROL_ TECHNIQUES ADVISORY COMMITTEE MEETING

                                                  March  17  and  18,  1981
      PLEASE PRINT
         NAME
COMPANY
BUSINESS ADDRESS
TELEPHONE
PRIMARY INTEREST
DATE
                                                 '3
\
                r\
                                                        ,-Y\  -sr
                                                                                      Voc

-------
                                                                                   U. S. Environmental Protection Agency
                                                                                    Research Triangle Park, N. C.  27711
                                            RECORD OF ATTENDANCE

                     NATIONAL AIR  POLLUTION CONTROL_JECHNIQUES ADVISORY COMMITTEE MEETING

                                           March 17 and 18, 1981
PLEASE PRINT
   NAME
COMPANY
BUSINESS ADDRESS
TELEPHONE
PRIMARY INTEREST
DATE
                                                  /Uc.  -ZJ7-7/
                                                                                                         //
                                                 -c
                                               p
 3=
 1
 oo

-------
PLEASE PRINT
                                                                            U.  S. Environmental Protection Agency
                                                                             Research Triangle Park, N.  C.  27711
                                        RECORD OF ATTENDANCE

                   NATIONAL AIR POLLUTION CONTROL_ TECHNIQUES ADVISORY COMMITTEE MEETING

                                        March 17 and  18, 1981
   NAME
                    COMPANY
             BUSINESS ADDRESS
TELEPHONE
PRIMARY INTEREST
DATE
         5,

                               T>
         P
Sam
                 GCA
            Onptl
                                                      'Or,
                               (T11J
            SOCJHt
         ~>
               \nt> k^
fk
A- ^
                               f-
              T
                                      /UP
                       EPA

-------
                                                                                    U. S. Environmental Protection Agency
                                                                                     Research Triangle Park, N. C.  27711
                                             RECORD OF ATTENDANCE

                     NATIONAL AIR POLLUTION  CONTROL_ TECHNIQUES ADVISORY COMMITTEE MEETING

                                            March 17  and 18,  1981
PLEASE PRINT
   NAME
                                COMPANY
BUSINESS ADDRESS
TELEPHONE
PRIMARY INTEREST
DATE
 A. 9. l
3=-


O
 /?
   «arn
-------
                                                                                    U. S. Environmental  Protection Agency
                                                                                     Research  Triangle Park, N. C.  27711
                                             RECORD OF ATTENDANCE
                     NATIONAL AIR POLLUTION CONTROL_ TECHNIQUES ADVISORY COMMITTEE MEETING
                                            March 17 and 18, 1981
PLEASE PRINT
   NAME
                             COMPANY
                      BUSINESS ADDRESS
                             TELEPHONE
PRIMARY INTEREST
DATE
                                           \C\ u b"
                                                                                                      •a/
            //• ^e&
?\ T c:   ,/    14
A3u (  v . O l-'lH-KV? V c//I /
J   , I.. _ -. -,..if i i - i > r, 11   »i i
£
£.
                                           -20 "iO
                                                                               VOL
                                                     13
                                           "Bete.
                                                                     •3 If //
                                                                                                        /!?/?
                                          7o7
                                                     ,/> 7

-------
     NATIONAL  AIR POLLUTION CONTROL TECHNIQUES ADVISORY COMMITTEE

                               MINUTES OF MEETING

                              MARCH 17 AND 18, 1981
                                        Prepared by:




                                                /   //?   /s' ^/  c"  '""x7
                                        Robert R. Kolbinsky
                                        Standards Development Branch
                                        Emission Standards and
                                         Engineering Division
     I certify that, to the best of my knowledge, the foregoing minutes and
attachments are complete and accurate.
                                        Don R. Goodwin, Chairman
                                        National Air Pollution Control
                                        Techniques Advisory Committee

-------