>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
------- |