United States      Office of Air Quality       EPA-450/3-82-009
           Environmental Protection  Planning and Standards     September 1982
           Agency        Research Triangle Park NC 27711

           Air
<&ER&     Guideline Series

           Control of Volatile
           Organic
           Compound
           Emissions from
           Large Petroleum
           Dry Cleaners

-------
                                       EPA-450/3-82-009
Control of Volatile  Organic  Compound
    Emissions from  Large Petroleum
                  Dry Cleaners
              Emission Standards and Engineering Division
                                U.S. Environ"::.:.' '
                                Region 5,Li:''....v ;
                                77 West Jacks-:.. '
                                Chicago, IL 606u
             U S. ENVIRONMENTAL PROTECTION AGENCY
                  Office of Air, Noise, and Radiation
               Office of Air Quality Planning and Standards
              Research Triangle Park, North Carolina 27711


                        September 1982

-------
                                     OAQPS GUIDELINE  SERIES


The guideline series of reports is being issued by the Office of Air Quality Planning and Standards (OAQPS) to
provide information to state and local air pollution control agencies; for example, to provide guidance on the
acquisition and processing of air quality data and on the planning and analysis requisite for the maintenance of air
quality. Reports published in this series will be available - as supplies permit - from the Library Services Office
(MD35), U.S. Environmental Protection Agency, Research Triangle Park, North Carolina 27711; or, for a nominal
fee, from the National Technical  Information Service, 5285 Port Royal Road, Springfield, Virginia 22161.
                                              11

-------
                            TABLE OF CONTENTS

Section                                                          Page

1.    INTRODUCTION	1-1

2.    PETROLEUM DRY CLEANING INDUSTRY 	  2-1

     2.1  Industry Description 	  2-1
     2.2  Process Description	2-2
     2.3  Emission Sources 	  2-6
     2.4  Facilities and Their Emissions 	  2-13
     2.5  Model Plants	2-14
     2.6  References for Chapter 2	2-16

3.    EMISSION CONTROL TECHNIQUES 	  3-1

     3.1  Recovery Dryer	3-1
     3.2  Recovery Dryer Safety	3-11
     3.3  Cartridge Filtration 	  3-15
     3.4  Vacuum Distillation	3-21
     3.5  Miscellaneous Fugitive Emission Sources	3-23
     3.6  Combined Control Techniques	3-24
     3.7  Evaluation of Control Technology Transfer	3-25
     3.8  References for Chapter 3	3-27

4.    ENVIRONMENTAL ANALYSIS OF RACT	4-1

     4.1  Air Pollution	4-1
     4.2  Water Pollution	4-3
     4.3  Solid Waste Disposal 	  4-5
     4.4  Energy	4-5
     4.5  References for Chapter 4	4-7

5.    CONTROL COST ANALYSIS OF RACT	5-1

     5.1  Basis for Capital Costs	5-1
     5.2  Basis for Annualized Costs 	  5-3
     5.3  Emission Control Costs 	  5-6
     5.4  Cost Effectiveness	5-9
     5.5  References for Chapter 5	5-13

     APPENDIX A - Summary of Field Tests	A-l

     APPENDIX B - Emission Measurement Procedures	B-l

-------
                      TABLE OF CONTENTS (continued)
Section                                                          Page
     APPENDIX C - Emissions Factors	C-1
     APPENDIX D - Comments Received on the November 1981, Draft
                    CTG	0-1
     APPENDIX E - Example Regulation 	  E-l
                                    IV

-------
                             LIST OF FIGURES
Figure                                                                Page

 2-1      Petroleum Dry Cleaning Plant Flow Diagram 	    2-3

 2-2      Typical  Dry Cleaning Equipment in Existing Plants ....    2-4

 3-1      Solvent Vapor Flow Diagram for a Recovery Dryer 	    3-2

 3-2      Operating Cycles of the Existing Recovery Dryer 	    3-4

 3-3      Solvent Recovery and Concentration Curve for the
          Recovery Dryer  	    3-9

 3-4      Cartridge Filtration System Schematic 	    3-16

 3-5      Solvent Emissions for Filter Cartridges as a Function
          of Drainage Time	    3-20

 A-l      Recovery and Concentration Curves for a High-Emission
          Recovery Dryer Load	    A-8

 A-2      Recovery and Concentration Curves for a Low-Emission
          Recovery Dryer Load	    A-9

 A-3      Recovery and Concentration Curves for a Recovery Dryer
          with Test-Average Emissions 	    A-10

 A-4      Recovery and Concentration Curves for a Low-Emission
          Recovery Dryer Load	    A-15

 A-5      Recovery and Concentration Curves for a High-Emission
          Recovery Dryer Load	    A-16

 A-6      Recovery and Concentration Curves for a Recovery Dryer
          Load with Emissions Approximately Equal to the Overall
          Test Average	    A-17

 A-7      Solvent Emissions for Filter Cartridges as a Function
          of Drainage Time	    A-21

 A-8      Carbon Adsorption System Schematic	    A-27

 B-l      Response Factor Preparation System	    B-8

-------
                             LIST OF TABLES

Table                                                                 Page

 2-1      Dry Cleaning Solvent Physical Properties 	     2-7

 2-2      Model Plant Parameters (Existing Equipment)	     2-15

 3-1      Model Plant Parameters (Control Equipment) 	     3-26

 4-1      Nominal Emissions Factors for Existing and RACT
          Equipment	     4-2

 4-2      Nominal Annual VOC Emissions for Two Model Plants
          Employing Existing and RACT Equipment and Procedures .  .     4-4

 4-3      Energy Impact of Existing and RACT Equipment	     4-6

 5-1      Equipment Costs in Two Model Plants	     5-2

 5-2      Cost Equations	     5-4

 5-3      Capital and Annualized Costs of Existing Equipment in
          Two Model Plants	     5-7

 5-4      Capital and Annualized Costs of RACT Equipment in Two
          Model Plants	     5-8

 5-5      Cost Effectiveness of Existing and RACT Equipment in
          Two Model Plants	     5-11

 5-6      Cost Effectiveness of RACT  Implementation in
          Two Model Plants	     5-12

 A-l      Dryer Emissions Data	     A-3

 A-2      Recovery Dryer Data Compilation  	     A-6

 A-3      Recovery Dryer Data Compilation	     A-14

 A-4      Total Solvent Emissions Due to Disposal of 14 Filter
          Cartridges  (12 Carbon-Core  and 2 All-Carbon) as a
          Function of Drainage  Time	     A-22

 A-5      Record of Still Waste Boil down Samples	     A-25
                                    VI

-------
                       LIST OF TABLES (continued)

Table                                                                 Page

 B-l      Equations for Calculating Solvent-to-Propane Response
          Factor	    B-ll

 C-l      Nominal Annual  VOC Emissions for Two Model Plants
          Employing Existing and RACT Equipment and Procedures .  .    C-2
                                   VII

-------
                            1.   INTRODUCTION

     The Clean Air Act Amendments of 1977 require each State in which
there are areas in which the national  ambient air quality standards
(NAAQS) are exceeded to adopt and submit revised State Implementation
Plans (SIP's) to EPA.   Revised SIP's were required to be submitted to
EPA by January 1, 1979.  States that were unable to demonstrate attain-
ment with the NAAQS for ozone by the statutory deadline of December 31,
1982, could request extensions for attainment of the standard.   States
granted such an extension are required to submit a further revised SIP
by July 1, 1982.
     Sections 172(a)(2) and (b)(3) of the Clean Air Act require that
nonattainment area SIP's include reasonably available control technology
(RACT) requirements for stationary sources.  As explained in the "General
Preamble for Proposed  Rulemaking on Approval of State Implementation
Plan Revisions for Nonattainment Areas," (44 FR 20372, April 4, 1979)
for ozone SIP's, EPA permitted States to defer the adoption of RACT
regulations on a category of stationary sources of volatile organic
compounds (VOC)  until  after EPA published a control techniques guideline
(CTG)  for that VOC source category.  See also 44 FR 53761 (September 17,
1979).  This delay allowed the States to make more technically sound
decisions regarding the application of  RACT.
     Although  CTG documents review  existing  information and data
concerning the  technology and  cost  of various control techniques to
reduce  emissions, they are, of necessity,  general  in  nature and do not
fully  account  for variations within a stationary  source category.
Consequently,  the purpose of CTG  documents  is to  provide  State and local
air  pollution  control  agencies with an  initial  information  base for
proceeding with their own assessment  of RACT  for  specific stationary
 sources.

-------
                   2.   PETROLEUM DRY CLEANING INDUSTRY

     The objective of this section is to describe the domestic petroleum
dry cleaning industry.   The dry cleaning process is discussed as are
solvent characteristics, emissions, and the major petroleum dry cleaning
equipment.   Model plants are presented that will be used in later chapters
to evaluate the environmental and cost impacts of reasonably available
control technology (RACT).
2.1  INDUSTRY DESCRIPTION
     The dry cleaning indus^y is a service industry involved in the
cleaning and/or renting of c ..icles ranging from personal clothing to
mops and mats.  The total indUtry is subdivided according to the type
of solvent used and the type of services offered.  The solvents used are
categorized into three broad groups:  petroleum solvents, perch!oro-
ethylene (perc), and trichlorotrifluoroethane (F-113, a registered
trademark).  The industry also is  composed of three sectors which are
delineated by the type of services offered.  These are:  (1) the self-
service or coin-operated  sector,  (2) the commercial dry cleaning sector,
and  (3) the industrial dry cleaning sector.  This  report is concerned
only with  large  facilities that use petroleum dry  cleaning solvents.
     Petroleum dry cleaning  represents about 30 percent of the  total
quantity of articles cleaned by the dry cleaning  industry.  Petroleum
dry  cleaning  services are offered only by  the commercial and  industrial
sectors of the industry,  and represent about 70 and 30 percent,
respectively, of the total clothes throughput for  the  industry  (Fisher,
1980b;  Sluizer,  1981).   Coin-operated  or self-service  petroleum dry
cleaning plants  are  prohibited  by National  Fire  Protection Association
codes  due  to  the  highly  volatile  and  flammable  nature  of petroleum
solvents (NFPA,  1979).   Consequently,  most commercial  and  industrial

-------
 petroleum  dry cleaning plants are  located away from densely populated
 residential areas and shopping centers.
      Industrial dry cleaning establishments are generally larger than
 commercial plants and cater to industrial, professional, and institutional
 customers.  Articles such as work  uniforms, mats, mops, and rugs generally
 are cleaned by industrial dry cleaners, often in conjunction with rental
 operations.  There are approximately 1,000 industrial establishments
 nationwide.  In 1979, approximately 230 of these industrial plants used
 petroleum  dry cleaning solvents in some portion of their cleaning
 operations.  A typical industrial  petroleum dry cleaning plant processes
 roughly 515 megagrams of articles  each year (Sluizer, 1981).  Thus, the
 industrial petroleum dry cleaning  sector processes approximately
 120,000 megagrams of articles each year.
 2.2   PROCESS DESCRIPTION
      Petroleum dry cleaning operations are similar to detergent and
 water wash operations.  Unlike perch!oroethylene dry cleaning, which can
 have  both washing and drying operations in the same machine (dry-to-dry),
 petroleum dry cleaning is a batch  operation where articles are washed
 and dried  in separate machines.   Figure 2-1 depicts a typical petroleum
 dry cleaning operation.   Articles  to be dry cleaned are sorted into lots
 according to color, fabric, degree of soiling, etc., and are placed in
 their appropriate washers.   For example, one lot might consist of light
 colored, light weave, casual clothing which can be placed in one washer,
 while another lot might be made up of heavy weave, heavily soiled industrial
 uniforms placed in another washer.   Articles are then agitated in the
 solvent.  The more heavily soiled  articles go through two or more wash
 cycles:   the first with recycled,  soiled solvent and the second with
 clean solvent.
      Large dry cleaners often use a cleaning process where water-soluble
 materials are removed from articles in a water and detergent wash.   This
 process, sometimes called a "dual"  or "double phase" process (see
 Figure 2-2b),  is characteristic of many of the modern petroleum solvent
washer/extractors (particularly those with large capacities) and is
 employed to some extent in most, large petroleum dry cleaners (Sluizer,
 1981).
                                 2-2

-------
                                           Vent
                                          Solvent Settling Tank
                                               Filtered Solvent
                                   Setting

\
tew Solvent In
It \
1
EJ
4->
a
o
t?
$
*0
in
A
lasher/ Speni
(Lraclur I
T
*
Cartridge
Filter
Tank
Sludge
Solnt _
Diatomite
Filter

Filter riaste
!  to Disposal
I  <    
!
JSpent
Cartridges
1 Filter MasteV
1 Vent Solvent Reraover>
                                              Filter
                                              Waste
            New  and
          Regenerated
           Solvent
          Storage Tank
                                        LJ
                                                                      I Vent
                                                                Filtered
                                                                Solvent
                                                                Storage
                                                                 Tank
, _J
                           Distilled Solvent
                                       .
                                    Condenser
                                                                           Still Waste
                                                                           to Disposal
                  LEGEND

                    Solvent Flow


                    Vapor Emission


                    Solid Waste Emission


                    Alternate Solvent Flow
                                  \
                                                               Dryer
Figure 2-1.   Petroleum Dry  Cleaning Plant Flow Diagram.
                                    2-3

-------
     Atmospheric
      Air Intake


          L _
  Exhaust
     to
Atmosphere
      A.   STANDARD DRYER
             New and
             Distilled
             Solvent
                     B.  HASHER/EXTRACTOR
  Pressurized
      Air
     Inlet
 Filtered
  Solvent
 to Still
                                                                  Solvent/Hater
                                                                      Vannr
                                                     Cooling
    Muck-
   Drain
                               Dia
somite-Coated
  Tubes
 (Septum)
     Soiled
    Solvent
    -Inlet
                                   Muck
                                 Agitator
  Filtered
  Solvent
   Inlet_

 Steam
 Inlet

Condense^.
  Steam
  Outlet
      C.   DIATOMITE FILTER
                         Sludge
                          Drain
                     D.  VACUUM STILL
                                                 ^.Distilled
                                                    Solvent
               Figure  2-2.   Typical  Dry Cleaning  Equipment  in  Existing Plants.

                                                    2-4

-------
     After completion of the wash cycle,  articles are spun at high
speeds to remove excess solvent.   This spin cycle usually occurs in the
same equipment used for washing;  however, some existing plants (older
petroleum plants) have separate,  high speed centrifugal extractors which
remove and recover solvent from washed articles (Watts and Fisher,
1975).  When the spin cycle has terminated, articles are transferred
from the washer/extractor to a dryer (tumbler).  Inside the dryer,
remaining solvent is removed from the articles by evaporation in a
heated air stream and vented to the atmosphere.
     In some smaller plants, soiled solvent extracted during the spin
cycle is passed through a filter to remove insoluble soils and other
suspended particles.  When the soil load in the solvent is excessive, as
in most larger plants, soil-laden solvent is transferred directly from
the washer to a vacuum still or to a settling tank prior to distillation.
After settling (usually overnight), the heavy oils, dirt, and grease are
decanted and the solvent is sent to a vacuum still where it is purified.
The distilled solvent is pumped into another holding tank or is returned
to the washer/extractor.  When oil and grease loading is low, or when
cartridge filtration is used, distillation is often bypassed and filtration
serves as the only means of solvent replenishment.
2.2.1  Petroleum Dry Cleaning Solvents
     The National Fire Protection Association  (NFPA) classifies petroleum
dry cleaning plants by the type of solvent used.  Solvents, in turn, are
classified by their flash points.  Class II and  IIIA solvents are the
primary solvents used in the petroleum dry cleaning industry.  The NFPA
number 32-1979 dry cleaning solvent classification is as follows  (NFPA,
1979):
          Class  I Solvents - Liquids  having flash points below 38C
          (100F) such as 50F flash  point naphtha.
          Class  II  Solvents -  Liquids having  flash points from 38C  to
          59C  (100F to 139F)  such  as  quick  drying  solvents.
          Class  IIIA  Solvents  -  Liquids  having  flash points ranging  from
          60C  to 93C  (140F  to  199F)  such  as  140F  "safety" solvent.
      Petroleum  solvents are a  mixture of mainly  C8 to  C12 hydrocarbons
that  are  similar to  kerosene.  These  hydrocarbons can  be further  subdivided
                                 2-5

-------
into three molecular structures:   aliphatics,  alicyclics,  and aromatics.
Table 2-1 gives the chemical  properties of several  types of petroleum
solvents including their aromatic contents.   In recent years, the aromatics
content of petroleum solvents has been reduced to about 2  percent in
many states in order to control  these atmospheric emissions.   This
reduction has somewhat reduced the cleaning efficiency of  the solvent,
according to an industry trade association because the aromatics content
has a direct influence on the cleaning rate and performance (Sluizer,
1981).
2.3  EMISSION SOURCES
     This section identifies dryers, solvent filtration and distillation
systems, and miscellaneous (fugitive) sources as the major contributors
of VOC emissions in a dry cleaning plant.   The operations  of these
sources, their emissions, and the development of a baseline emission
estimate are discussed below.
2.3.1  Dryers
     Petroleum dryers consist of three parts that are housed in a single
unit:  the tumbler, blower, and  steam coils (see Figure 2-2a).  The
tumbler  is a perforated, rotating basket in which solvent-laden articles
are placed.  The blower forces air over steam-heated coils, where it
reaches  temperatures that range  from 43C to 88C (110F to  190F), and
then into the tumbler.  Optimum  drying temperatures range from 60C to
66C (140F to 150F) (Marvel et al., 1980).  Solvent in the articles is
removed  or volatilized by the heated air stream.  The volatilized solvent
and heated air then are continuously vented to the atmosphere during  the
drying cycle at an  air flow  rate of  from 28 to 340 cubic meters per
minute (1,000 to 12,000 cubic feet per minute) and a vapor concentration
of  from  200 to 9,000 parts per million (ppm)  of  solvent by volume
(Jernigan and  Lutz, 1979;  Lutz et al., 1980;  Marvel et  al. ,  1980).
     Existing petroleum dry  cleaning dryers range in capacity from  22 kg
to  180 kg (50 to 400 Ib)  (Marvel et  al., 1980).  Emissions from these
dryers vary depending on  the  extraction efficiency of the washer/extractor
and the  weight of  articles per dryer  load.  The  type of articles  in  the
dryer  also have an  effect  on  emission  rates and  emission concentration
                                  2-6

-------
                                        Table  2-1.   Dry Cleaning  Solvent  Physical  Properties'
Boiling R.nge,F 
Trade Name
Skellysolve S-66
Varaol 3
325
AHSCO Napthol MS*
Varaol 1
Skellyaolve "S-2"
Kuik-Dri**
Mineral Spirits
Mineral Spirita 66/3
Varaol 18
no
1 AMSCO Mineral
Spirita 66/3
Sol 340
Mineral Spirita US-EC
AMSCO Odorlea* MS
Odor lea a MS
140 Solvent
Skellyaolve "T"
AMSCO 140
Solvent 66/3
Sol 140
Chevron Thinner 410B
Producer
Getty
Exxon
Chevron
Union
Exxon
Getty
Ashland
Ashland
Ashland
Exxon
Union
Shell (EOR)
Shell (EOR)
Union
Aahland
Ashland
Getty
Union
Shell (EOR)
Chevron
Ini-
tial 501
305
310
320
313
313
314
315
315
315
315
315
316
317
354
355
362
363
368
371
370
329
327
335
323
346
321
330
344
342
344
345
324
336
362
365
375
375
383
380
383
Dry F
Point T.C.C
385
348
368
353
390
331
355
385
385
395
375
347
388
386
395
390
400
402
416
406
104
105
105
106
106
102
105
105
105
110
104
104
112
128
128
140
143
146
143
142
i Vapor
: Preaaure
nmHg
1. at 20 F L.E.L
3.2 1.1
10.0
5.0 1.0
3.0 1.0
10
3.1 1.1
<10 1
<10 1
<10 1
10
1.1 .7
2.8 1.0
2.7 1.0
<1.0
<10 1
<10 1
0.7 1.1
<1.0 1.0
<1.0
0.9
Pounda
Ave. per Aniline
Mol. Gallon Point
. wt. at 60F F
142 6.48
6.46
138 6.57
138 6.32
6.56
6.47
128 6.46
145 6.55
145 6.55
6.42
6.55
6.37
6.49
6.33
6.33
6.54
6.57
6.51
6.56
157 6.78
142
130
145.2
160
130
133
154
140
155
150
153
ISO
145
184
185
160
151
164
153
147.4
Kauri
Buta-
nol
Value
36
38
36
31
37
37
32
36
31
32
33
33
34
27
28
31
32
30
31
35.8
Compoaition
Paraf-
fins
43.4

35.4
64

39.5
57
54
57

61
47.2
47.5
86
87
60
35.3
55
54.9
36.6
Naph-
thenei
49.4

62.2
35

48.2
41
36
41

37
50.2
45.1
14
13
37
54.5
44
39.2
60.6
Aron
1 C8*
6.8
19.0


18
11.7
2
9.5
2
7.6
2.0
2.6
7.4


3.0
9.3
1.0
5.9

, Vol.
utica
Total
6.8
19.0

1
18
11.7
2
9.5
2
7.8
2.0
2.6
7.4


3.0
9.3
1.0
5.9

Z
Ole-
fina
0.4




0.6
0.3
0.3
0.2






0.3
0.9



*Used during Rhode  Island test.




"Used during Lakeland test.
                                                       ent3'" Natl"al P3lnt a"d Cat1"* ^-elation. Washington, D.C.;

-------
levels.   Wools have a tendency to absorb more solvent and,  consequently,
give up more solvent during extraction than an equal  weight of silks
(IFI, 1973).  This basic property may be attributed to the  loose weave
of wools in comparison with the tight weave of silks.   Thus, loosely
woven materials tend to absorb and give up solvent more readily than
tightly woven material.
     In an EPA demonstration test of an add-on carbon adsorber at a
petroleum dry cleaning plant in Anaheim, California,  a 180  kg (400 Ib)
standard dryer loaded with 115 kg (250 Ibs) of work uniforms, shop
towels, and fender covers emitted 14 kg VOC per 100 kg dry  weight of
articles cleaned (Lutz et al., 1980).  Approximately 28 kg  VOC per
100 kg dry weight of articles cleaned was emitted during an EPA test of
a 50 kg (100 Ib) standard dryer at a petroleum dry cleaning plant in
Pico Rivera, California (Jernigan and Lutz, 1980).  This dryer was
loaded to 10 percent over-capacity with leather and cotton  work gloves,
exclusively.  A study by a dry cleaning trade association indicated that
standard petroleum solvent dryers emit approximately 14 kg  VOC per
100 kg dry weight of general apparel cleaned (Fisher, 1975).  The nominal
emission rate from a standard dryer, based on these three data sources,
is 18 kg VOC per 100 kg dry weight of articles cleaned.  This average
emission rate is heavily dependent on the washer extraction efficiency
and the solvent absorptivity of the fabric dried and could  vary
significantly between individual dry cleaning plants.
2.3.2  Filters
     Filtration, in dry cleaning operations, is a process used to remove
most insoluble contaminants (dirt and lint), as well  as certain water-
soluble contaminants (perspiration and food stains) from dry cleaning
solvents.  This is accomplished by rapidly passing large volumes of
solvent through a porous medium that traps and thus removes contaminants
suspended in the solvent.
     All dry cleaning filtration systems are composed of two parts:  the
filter medium and the structure that holds the filter medium, known as
the septum  (see Figure 2-2c).   The filter medium, usually diatomaceous
earth (diatomite) mixed with activated carbon, is used to remove insoluble
contaminants by entrapping them in its porous surface.  The septum is a
                                 2-8

-------
rigid, porous surface (screen, cloth,  or wire mesh) to which the filter
medium adheres, thereby allowing pressurized solvent to flow through
while simultaneously blocking the passage of particles.  Filters are
sized by the volume of solvent processed and range in size from 5,700 to
56,800 liters per hour (1,500 to 15,000 gallons per hour) (Washex,
1974).
     During a wash cycle, articles are agitated in a bath of solvent.
After the wash cycle, the soiled solvent is pumped to a filter for
filtration.  Filters vary, based on their mode of operation, septum
type, and construction material.  Single-charge filters (rigid tube or
disc septums) have a single mass or "charge" of filter medium which is
replaced after each wash load is completed.  With multi-charge filters
(bag, screen, and rigid tube septums)  filter medium is added to the
initial charge for each load of articles washed.   Regenerative filters
(flexible tube septums) have an initial mass of filter medium which is
redistributed on the septum for each load of articles washed, without
subsequent addition of filter medium.   As of 1980, about 50 percent of
the petroleum dry cleaning plants that utilize filtration used multi-charge
diatomite filters, the remaining 50 percent employed cartridge filters
(see Section 3.3) for solvent filtration (Fisher, 1980a).  A trade
association study has shown that initial masses of diatomite (precoats)
average about 1 kg (2 Ibs) per 3,800 liters (1,000 gal) of filter capacity,
with diatomite being added during the operation at a rate of 0.5 kg (1
Ib) per 45 kg (100 Ibs) of articles cleaned (Leonhardt, 1966).
     Filter medium is replaced when the pressure across the filter, due
to the buildup of contaminants on the medium, reaches a predetermined
level (up to 40 psi or 270 kilopascals).  Spent filter medium (filter
muck) is usually allowed to drain in the filter housing overnight or for
24 hours before it is discarded.  An industry trade association determined
that discarded filter medium  that has been allowed to drain for 24 hours
may contain from 5 to 10 kg solvent per 100 kg dry weight of articles
cleaned for regenerative and  multi-charge  filters, respectively (NID,
1971).  Thus, after a 24-hours  drain time, an average of just under 8 kg
of solvent per 100 kg dry weight of articles cleaned is  retained in the
discarded  filter muck.  This  average can vary based on the soil loading
                                 2-9

-------
 and filter muck drainage procedure employed in individual  plants.   Also,
 devices such as centrifugal  separators,  and pressure  vacuum muck strippers
 have been used by a few large petroleum  dry cleaning  plants to  recover
 solvent from diatomite filter muck (Fisher,  1981).  However,  these
 devices are not widely distributed through  the industry  and there  is
 little  current data on their performance as  an emisson control  technique.
 2.3.3  Settling Tanks
      Large industral  facilities  with  heavy  soil  loadings and  high
 throughputs typically  omit filtration, relying instead on  settling  tanks
 to  partially remove solids and nonsolvent soluble contaminants  from the
 soiled  solvent stream.   In these tanks,  heavier  components  of the  used
 solvent from the washer are  allowed to settle  to the  bottom of  the  tank
 while relatively contaminant-free  solvent is pumped from the  top of the
 tank to the vacuum  still.  Periodically,  depending on the  tank  capacity
 and  the plant  throughput, the  heavier components are  pumped from the
 bottom  of  the  tank.
      There  is  no test  data available on  the  solvent retention of settling
 tank waste.  A  vacuum  still  manufacturer  stated, however,  that  approximately
 2 kg of solvent  is  lost with  every kilogram  of settling tank  waste
 (Landon, 1981).   An  industrial trade association representative stated
 that the solvent content of  settling tank waste can range  from  80 to
 200  percent by weight  of the total waste  (Sluizer, 1981).
      The operation  of  a  settling tank is  dependent upon the mode of
 operation of the  plant,  the plant's capacity for storing the  residue and
 the  plant's  settling tank and  vacuum still capacities.  Most dry cleaners
 use  at  least one  of three approaches to remove their settling tank
 residue; burn it  as a  boiler fuel supplement, discard it with general
 dry  cleaning waste or  sell it  to a solvent reprocessor.   Trade association
 and  industry representatives were unable  to give any percentages of dry
 cleaners that use any  of these three approaches, but one representative
 stated that the  residue is usually too contaminated or viscous to distill
 in a dry cleaning vacuum still (Sluizer,  1981;  Landon, 1981).
2.3.4  Vacuum Stills
     Distillation of solvent  is used to remove  contaminants, such as
bacteria, detergents, water,  oils, and dyes, that are  not removed by
                                 2-10

-------
filtration or settling.   Petroleum dry cleaning solvents have boiling
ranges of from 150C to 215C (300 to 415F).   A steam pressure of
670 kilopascals (100 psi) or more is required to boil  petroleum solvents
under atmospheric conditions.   Consequently, distillation of petroleum
solvents is done under a vacuum of from 75 to 92 kPa (22 to 27 inches
Hg).   This lowers the boiling range for petroleum solvents to 107C to
113C (225F-235F) and reduces required steam pressures to the range of
235 to 600 kilopascals (35 to 90 psi) (Washex, 1973).   Vacuum stills are
sized by the volume of solvent to be processed and range in size from
190 to 5,700 liters per hour (50 to 1,500 gallons per hour) (Washex,
1973).
     Spent solvent from a washer or filter is pumped to the boiling
chamber of a still on a continuous basis.   In the boiling chamber, steam
heated coils volatilize the solvent, leaving behind still residue (high
boilers) composed of oils, grease, and dirt.  Solvent vapor and moisture
pass continuously from the boiling chamber into a water-cooled condenser
where the vapors condense to a mixture of liquid solvent and water (see
Figure 2-2d).  This mixture is constantly piped to a gravimetric separator,
where the solvent and water are separated by the differences in their
densities.  Finally, the solvent is pumped to a tank containing cotton
rags or salt pellets which absorb any remaining water.
     When the concentration of high boilers has reached a specified
level, as indicated by a visual inspection of solvent flow in a sight
glass between the condenser and separator (indicating that the evolution
and condensation of solvent vapors is seriously impeded or halted), the
solvent flow to the still is manually shut off and steam flow is increased
to the maximum available level.  The liquid contained in the boiling
chamber is allowed to boil for 5 to 15 minutes in a process commonly
called "boildown" in which most of the remaining solvent is removed by
boiling.  After boildown, the liquid residue is drained from the boiling
chamber and continuous still operations are resumed.   Typically, still
boildown is required when the solvent and water flow from the condenser
are reduced by 75 percent (Washex, 1973; Rosenthal, 1980).  This flow
reduction is subjectively determined based on the operation of a particular
still.
                                 2-11

-------
     Depending on the design of the still, its operation, and frequency
of boildown, as much as 90 percent by weight of solvent is decanted with
still residue (equivalent to 7 kg solvent per 100 kg dry weight of
articles cleaned) as indicated by an EPA vacuum still test of an industrial
dry cleaning plant in Anaheim, California (Jernigan and Kezerle, 1980).
The still tested at this facility, however, was designed in a manner
that severely limited the removal of solvent from the waste during
boildown, and the frequency of boildown was based on a daily routine
rather than on the flow rate of distilled solvent.  A vacuum still test,
conducted by a firm that rents and dry cleans apparel, also indicated
that more than 80 percent by volume of solvent is decanted with their
still residue (Burnett, 1980).  A trade association study has found that
approximately 1 kg of solvent per 100 kg dry weight of articles cleaned
is decanted with the residue for a well-maintained and operated vacuum
still (NID, 1971).  An industry trade association analysis of still
wastes from 43 separate commercial petroleum dry cleaning plants revealed
that well operated plants can reduce the solvent content of still residue
to 31 to 38 percent solvent by weight.   (Three of these samples had
solvent contents that were below 31 percent solvent by weight.)  Although
the trade association was unable to give the number of samples that
achieved the various levels of solvent retention, the average range of
solvent retention for all the samples was 39 to 42 percent solvent by
weight.   Samples with a solvent retention range of 43 to 51 percent were
considered moderately high in solvent content by the trade association,
and samples with solvent contents that exceeded 51 percent were considered
extremely high, indicating a poorly operated still (Andrasik, 1981).
Thus, considering variations in operating procedures (boildowns) and
still design, the assumed solvent content of disposed still waste would
be between 1 and 7 kg solvent per 100 kg dry weight of articles dry
cleaned, with a value of 3 kg solvent per 100 kg articles representing
the solvent content of the disposed waste from a typical still.
2.3.5  Fugitive Emission Sources
     There are a number of sources of fugitive emissions in a dry cleaning
operation.   Fugitive sources include emissions from the extraction cycle
of a washer, emissions given off during the transfer of solvent-laden
                                 2-12

-------
articles from washers or dryers, liquid and vapor leaks in tanks and
piping, and the evaporation of solvent from open containers.  Also
included are fugitive emissions from filter muck and still residue
storage tanks.  Although sources of fugitive emissions can be identified
and the VOC concentrations within the vicinity of these sources can be
quantified, there is no discrete flowrate associated with these sources
and, therefore, it is virtually impossible to estimate an emissions rate
for fugitive sources.
     An attempt was made to quantify fugitive emissions in an EPA test
at a dry cleaning plant in Anaheim, California.   It was found that more
than 0.8 kg VOC per 100 kg dry weight of articles cleaned was emitted
from the roof exhaust vent which collected emissions from various fugitive
sources within the plant (Jernigan and Kezerle,  1980).   Emissions were
expected to be higher than those recorded because the doorways and
windows (pathways through which emissions escaped to the atmosphere)
remained open during fugitive testing.
     It is technically and economically infeasible to quantify all
sources of emissions in a dry cleaning plant because certain emissions
are prevalent only during the operation of the dry cleaning equipment,
while other low-level sources emit continuously.   A dry cleaning industry
trade association publication assumes,  however,  that miscellaneous
(fugitive) emissions would be approximately 1 kg of VOC emissions per
100 kg dry weight of articles cleaned (Fisher, 1975).
2.4  FACILITIES AND THEIR EMISSIONS
     A summation of typical emissions from the four major sources in a
petroleum dry cleaning plant yields a range of from 15.5 kg to 46 kg VOC
per 100 kg dry weight of articles cleaned, depending on the plant
throughput and the equipment configuration.   Of  the 15.5 to 46 kg of
total  plant emissions, 60 to 90 percent originates at the dryer, 22 percent
is emitted from filter muck (4 percent from disposed cartridge filters),
still  residue contributes 7 to 15 percent, and 3 to 6 percent is
attributable to fugitive emissions.   Using 15.5  to 46 kg VOC per 100 kg
of clothes cleaned as a baseline emission range,  a representative petroleum
plant might have a breakdown of emissions ^iirnlar + o the values  shown
below:
                                 2-13

-------
  Source
Dryer
Filter
  Diatomite
  Cartridge
Still
Fugitive sources
Total
Totals with
  Diatomite filter
  Cartridge filter
  Settling tanks
Range of emission rates
(kg VOC per 100 kg dry
we 1.9,t1 t_J?.f. ,aHtilc-l?s_cleaned)
          14-28

          (5-10)
         (0.5-1)
          (1-7)
         (0.5-1)
        (15.5-46)
Nominal emission rates
(kg VOC per 100 kg dry
weight of articles cleaned)
             18
              8
              1
              3
              1
                                           30
                                           23
                                           22
2.5  MODEL PLANTS
     Two model plants - Model plant I and model plant II - have been
developed to represent the large petroleum dr> cleaning industry.
Throughputs of existing plants vary widely in both magnitude and content,
and the models do not represent a clear distinction as to actual plant
sizes or equipment configurations.   Rather, these models were developed
to simplify the classification of existing dry cleaning plants and their
equipment, throughputs, and costs.   The data used in the development of
these model plants were derived from plant visits, dry cleaning trade
association survey data, and input from industry representatives (Fisher,
1975; Marvel et al. ,  1980; Sluizer, 1981).
     The model dry cleaning plants are classified by their throughputs,
which are reflected in the type and size of equipment present.  Larger
dry cleaners usually have larger throughputs; but in both model plants,
revenues are based primarily on rental items which are cleaned regularly
as part of the rental process.   Table 2-2 lists the characteristics of
each model plant, their throughput, and emissions. The plant and nationwide
emissions listed are based on nominal emission rates, and may not be
representative of all plants with a given annual throughput.
2.5.1  Model Plant I
     Model plant I represents dry cleaners having a daily throughput of
1,100 kg (2,400 Ib)  of soiled articles ranging from rental uniforms to
mops, rugs and mats.   Typically, a single, large washer and several
                                 2-14

-------
                 Table 2-2.  MODEL PLANT PARAMETERS
                        (Existing Equipment)

Model plant
Number of existing plants
nationwide in 1979
Annual plant throughput,
in kg (Ibs)
Throughput description
Number of washers
Washer capacity (each),
kg (Ibs)
Number of standard dryers
Dryers capacity (each),
kg (Ibs)
Number of diatomite filters
Diatomite filter capacity
(each), L/hr (gal/hr)
Total vacuum still capacity
L/hr (gal/hr)
Number of loads dried per day
Average load weight per dryer
kg (Ibs)
Days of operation per year
Wash cycle time, minutes
Dry cycle time, minutes
Average or range of baseline
emissions per plant, megagrams,
VOC/year (tons VOC/year)
Average or range of baseline
emissions nationwide, megagrams
Model Plant I
60
182,000
(400,000)
industrial
articles
1
115
(250)
3
45
(100)
2
11,400
(3,000)
1,900
(500)
20
35
(80)
260
40
40
40-55
(44-60)
2,400-3,300
(2,600-3,600)
Model Plant II
170
635,000
(1,400,000)
industrial
articles
2
225
(500)
2
180
(400)
0
-
5,700
(1,500)
14
175
(400)
260
40
40
140
(154)
28,800
(26,200)
VOC/year (tons VOC/year)
                                2-15

-------
medium-capacity dryers are used to facilitate the separation of job
loads and to avoid downtime caused by equipment breakdowns.   There
an estimated 60 model plant I petroleum dry cleaners in the United
States (approximately 25 percent of all large petroleum dry cleaning
facilities) (Sluizer, 1980).  Data on model I plants are summarized in
Table 2-2.
2.5.2  Model Plant II
     Facilities represented by model plant II process the same types of
articles as model I plants, but have much larger throughputs and equipment.
Model II plants have a throughput of 2,400 kg/day (5,400 Ib/day) or more
of heavy fabrics that are heavily soiled.   Because these plants have
high throughputs, they require sturdy equipment with large capacities
for almost continuous operation.  There are an estimated 170 model II
petroleum dry cleaners in the United States representing 75 percent of
the total number of large plants (Sluizer, 1980).  Table 2-2 presents
additional information on model II plants.

2.6  REFERENCES FOR CHAPTER 2
Andrasik, I.  1981.  International Fabricare Institute (IFI).  Letter to
     Q.  Corey, TRW Inc., July 20.   Solvent content of still waste.
Burnett, E.  1980.  Aratex, Inc. ,  Telecon with Q. Corey, TRW Inc.,
     November 25.  Solvent content of still wastes.
Fisher, W.  1975.  ABC's of Solvent Mileage, Part 1.  IFI.  Joliet,
     Illinois.  Special Reporter Vol. 3, No. 4.  July-August.
Fisher, W.  1980a.  IFI, Meeting with S. Plaisance,  TRW Inc., December
     9.   Number of cartridge filters in use.
Fisher, W.  1980b.  IFI, Telecon with Q. Corey, TRW Inc., January 16.
     Comments on the size of the commercial petroleum dry cleaning
     industry and the throughput from a typical plant.
Fisher, W.  1981.  IFI, Telecon with S. Plaisance, TRW Inc., Oct.  16.
     Comments on alternative methods of filtration waste solvent  content
     reduction.
International Fabricare Institute (IFI), 1973.  An Introduction to
     Industrial Dry Cleaning Methods, Part One.  IFI Special Reporter.
     Volume One, Number Three.  Joliet, Illinois.
                                 2-16

-------
Jernigan, R. and J. Kezerle, 1980.  Evaluation of the Potential for
     Reduction of Solvent Losses through a Washex Petroleum Vacuum Still
     Sump.  TRW Inc. Research Triangle Park, North Carolina (EPA Contract
     No. 68-03-2560, Task No. T5013).

Jernigan, R. and S. Lutz.  1979.  An Evaluation of the Emission Reduction
     Potential of a Solvent Recovery Dry Cleaning Dryer.  TRW  Inc.
     Research Triangle Park, North Carolina (EPA Contract No.   68-03-2560).

Landon, S.  1981.   Washex Machinery Inc., Telecon with Q. Corey, TRW
     Inc., July 20.  The use and operation of solvent vacuum stills.
Leonhardt, G.  1966.  Filter Aids.  National Institute of Dry Cleaning.
     Silver Spring, Maryland.  (NID) Bulletin, p. 75.  July.

Lutz, S.,  S.  Mulligan, and A. Nunn.  1980.  Demonstration of Carbon
     Adsorption Technology for Petroleum Dry Cleaning Plants.  EPA
     Publication No. EPA-600/2-80-145.    EPA/IERL.  Cincinnati, Ohio.

Marvel Manufacturing Co., Washex Machinery, Inc., American Laundry
     Machinery, W.M. Cissel Manufacturing Co., VIC Manufacturing Co. and
     Challenge-Cook Brothers, Inc.  1980.  Telecon Survey with Q. Corey,
     TRW Inc., March 18-April 25.  Sizes of petroleum dry cleaning
     equipment and expected sales for 1980.

National Fire Protection Association.   1979.  Report No.  32, Dry Cleaning
     Plants,  Boston, Massachusetts.

NID.  1971.   Estimation of Solvent Vapor Emission from Petroleum Dry
     Cleaning Plants.   National Institute of Drycleaning.  Publication
     No. T-486.  Silver Spring, Maryland.  February.

Rosenthal, S.  1980.  Washex Machinery Inc., Telecon with S. Plaisance,
     TRW Inc., November 18.  The use of solvent filtration systems and
     vacuum stills.

Sluizer, M.   1981.   Institute of Industrial Launderers, Meeting with
     S.  Plaisance,  TRW Inc., January 8.   Size of the industrial petroleum
     dry cleaning industry and the throughput of a typical plant.

Sluizer, M.   1981.   Institute of Industrial Launderers, Telecon with
     S.  Plaisance,  TRW Inc., April 10.   Size of the industrial petroleum
     dry cleaning industry and the throughput of a typical plant.

Washex.   1973.  Installation, Operation and Maintenance Manual for
     Washex Vacuum Stills. Publication No. T-513d.  Wichita Falls,
     Texas.   July.

Washex.   1974.  Instruction Manual for Washex MAB-Type Tube Filters.
     Publication No. T-605.  Wichita Falls, Texas.  September.
                                 2-17

-------
Watts, A.  and A. Fisher.  1975.  Results of Membership Survey of Dry
     Cleaning Operations.   Joliet, Illinois.  IFI Special Reporter 3-1.
     January-Fphruary.
                                 2-18

-------
                      3.   EMISSION CONTROL TECHNIQUES

     Equipment and procedures selected as representing reasonably
available control  technology (RACT) for the petroleum dry cleaning
industry are described in this chapter.   Particular attention is given
to the design, operation, and VOC emission-reduction performance of RACT
equipment, with verification of these criteria being supported,  where
possible, by engineering analyses and field test data (see Appendix A).
The effects of variations in and deviations from the equipment configur-
ations and operating procedures characteristic of the two model  plants
developed in Chapter 2 are analyzed in relation to their impacts on VOC
emission reduction and overall equipment performance.
3.1  RECOVERY DRYER
     A solvent recovery dryer is essentially a standard dryer that has
been fitted with a condenser to remove solvent vapor from the dryer
exhaust by condensation (see Figure 3-1).  In the current configuration
of this machine, a steam-heated air stream is directed around and through
a tumbling load of drying articles by a blower that forces the solvent-
laden air stream through a lint filter and then to a condenser.   After
partial removal of both solvent and water vapors in the condenser, the
air stream is ducted from the condenser to a steam chest where it is
reheated and then passes to the tumbler where the cycle of solvent
evaporation repeats.
     The most important component of this solvent recovery system is the
condenser, which gradually reduces the concentrations of both solvent
and water vapors in the air stream during every evaporation-condensation
cycle.  The currently marketed recovery dryer employs a condenser to
remove both solvent and water from the incoming vapor stream by steadily
reducing the vapor temperature (under the existing conditions of vapor

-------



r     Steam ^_
Chest
*
Tumbler
i
1
1
L1nt
Filter
1
1
I
	 Solvent Vapor
Blower """"

1
Condenser
A
Recovered
Liquid
\
, , Solvent
Solvent-Water 	 ^
Separator Outlet
| .
 ' 1
Water
Outlet
Figure 3-1.  Solvent Vapor Flow Diagram for a Recovery Dryer.
                         3-2

-------
flow and pressure).   As the vapor stream is forced through the condenser,
chilled water circulates downward through the tube structure and cools
the vapor stream until a liquid solvent and water mixture condenses (as
heat is transferred from the vapor stream to the chilled water).   This
mixture flows to the bottom of the condenser where it is piped to a
solvent/ water separator.   Because petroleum solvent has a specific
gravity of about 0.75, the water contained in the condenser runoff forms
the bottom liquid phase which is removed from the bottom of the unit
while solvent flows from the top.
     To prevent the excessive wrinkling of clothes,  the recovery dryer
has a second, exhaust/cool-down phase similar to that found in some of
the more sophisticated standard dryers.   In the solvent recovery phase
(Figure 3-2a), the air stream flows [at a manufacturer-rated volumetric
flow rate of 17.7 mVmin (Hoyt, 1979)] from the steam chest through the
tumbler, to the condenser, and then back to the steam chest.   The cooling
water flow during this phase is from the cooler to a storage tank, to
the condenser, and then back to the cooler.   Solenoid valves,  controlling
both steam and cooling-water flows, are intermittently opened in response
to thermostats that maintain both a tumbler temperature sufficiently
high enough to promote solvent evaporation from tumbling fabrics and a
condenser water inlet flow rate great enough to ensure adequate condenser
heat-removal for optimum VOC emission reduction/recovery.   At the onset
of the exhaust/cool-down phase (see Figure 3-2b), both the steam and
cooling water solenoids close while atmospheric air and vapor stream
dampers divert vapor flow from the condenser loop, thereby permitting
the intake of atmospheric air which is forced over the tumbling articles
and exhausted to the atmosphere.
     Control of the duration of these two phases is accomplished by two
timers that can be individually set for a wide range of recovery and
exhaust/cool-down periods.  Typically, the duration of the recovery
phase ranges from 20 to 45 minutes, while that of the exhaust/cool-down
phase ranges from 2 to 6 minutes (Plaisance et al.,  1981).
Thermostatically controlled shutoff valves that govern tumbler
temperatures (steam flow) and condenser water inlet temperatures (water
flow) can be adjusted to protect drying articles from overheating and to
                                 3-3

-------
                                          Water
                                             3-2.a  Recovery Phase
     Atmospheric
         *r
Atmospherinil
   Air
   Damper
         T
         _sL
       Tmbler
                      Steam
                    Solenoid
        iiJ	X
                     Exhaust
                       to
                    Atiosphere
3-2.b Exhaust/Cool-Down Phase
    Figure 3-2.   Operating  Cycles of the  Existing Recovery Dryer.

                                3-4

-------
maintain optimum VOC emission reduction/recovery.  A manually-adjustable
valve in the condenser water inlet line permits regulation of the water
flow rate.  A manually-adjustable pressure control in the steam chest
inlet provides a controlled steam pressure for the heating of the vapor
stream.
     In the event of an explosion in the dryer, the pressure of the
blast is released upward from the tumbler through vents specifically
designed for this purpose, while steam is automatically injected into
the tumbler.  In addition, a fusible wire spanning the condenser vapor
inlet will melt above a predetermined temperature, shutting down the
electrical system and terminating dryer operations.
     The level of emission reduction attained by the recovery dryer is
based on the assumption that all solvent entering the dryer in garments
is either recovered or emitted to the atmosphere during the cool-down/
exhaust phase (in contrast to standard dryers which emit their entire
solvent content to the atmosphere).   Thus, the VOC emission reduction is
equal to the solvent recovery.   Furthermore,  it is assumed that garments
leaving the dryer are "dry" (contain no solvent), and that there are no
random or unspecified losses within the evaporation-condensation system.
(Results of recovery dryer field tests are inconclusive concerning the
solvent content of dried items;  however,  there are indications that
drying time, load weight,  fabric type, and condenser temperature and
flow (both vapor and liquid) all cause variations in the solvent content
of dried articles.)  Based on these assumptions,  the recovery dryer
performance parameters of  primary importance  are VOC emission reduction
and solvent recovery.
     Three EPA test programs have been completed at facilities operating
domestically manufactured  48 kg  (105 Ib)  capacity recovery dryers.   The
first test was conducted at a large industrial dry cleaning facility
(see Appendix A, Test 1) in Pico Rivera,  California that processes
approximately 1,300 kg (2,900 Ibs) of heavy work gloves each day.
Results  from the measurement of  the vapor concentration and flow rate
during the exhaust/cool-down phase were an average (for the entire test
period)  dryer VOC emission of 0.96 kg VOC per 100 kg dry weight of
                                 3-5

-------
articles cleaned, and a range from 0.68 to 1.25 kg VOC per 100 kg articles
cleaned.  The condenser reclaimed an average of 23.4 kg of solvent per
100 kg articles cleaned, with a range of 15.5 to 29.2 kg.   Furthermore,
the dryer appeared, at times, to operate at or above 90 percent of the
10,000 parts per million LEL concentration for petroleum solvent (measured
at the condenser inlet during the recovery phase), although the existence
and extent of the excursions above 90 percent were not verified because
the solvent concentrations exceeded the capacity of the detection
instrument.  Possible explanations for this apparent excursion center
around the fact that the dryer was typically loaded beyond manufacturer's
specifications (by as much as 20 percent), with articles (gloves) having
the unusually high solvent retention of as much as 30 percent by weight
(Jernigan and Lutz, 1979).
     A second EPA recovery dryer test was undertaken at a large commercial
plant in Lakeland, Florida (see Appendix A, Test 2) that processes
approximately 180 kg (400 pounds) of general apparel each day.  Again,
solvent vapor concentrations in the condenser vapor inlet and atmospheric
exhaust were monitored, as were the flow rate and mass of recovered
solvent.  The average (over the test period) VOC dryer emissions vented
during the exhaust/cool-down phase was found to be 3.85 kg VOC per
100 kg dry weight of articles cleaned and ranged from 9.45 to 2.34 kg
VOC per 100 kg articles cleaned.  The condenser reclaimed an average of
10.4 kg of solvent per 100 kg articles claimed, with a range of 9 to
14.3 kg.  Solvent vapor concentrations in the condenser inlet stream
never exceeded 95 percent of the solvent LEL and ranged from 4,410 to
9,425 ppmv as solvent.  The reduced VOC emission reduction value was
possibly due to the processing of small loads (typically 50-60 percent
of dryer capacity) of garments (principally synthetics) having a low
solvent retention.  Also, recovery periods were relatively brief (usually
lasting no more than 30 minutes), resulting in the premature termination
of the  recovery phase and a corresponding increase in the solvent content
of dried articles  (Jernigan et al. , 1981).
     A  third EPA recovery dryer test was conducted at a large commerical
plant in West Warwick,  Rhode Island (see Appendix A, Test 3) that cleans
about 2,700  kg (6,000  Ib) of personal clothing per week (Plaisance et
al., 1981).  Both  dryer condenser vapor inlet and dryer atmospheric

                                 3-6

-------
exhaust concentrations were monitored, and data on the volume and flow
rate of recovered solvent were collected.  The average (for the test
period) dryer VOC emissions vented during the exhaust/cool-down phase
was 3.47 kg VOC per 100 kg of articles dry cleaned, with a range of
1.2 kg to 7.2 kg.   The condenser reclaimed an average of 13 kg per
100 kg of articles cleaned with a range of 9.5 to 17.7 kg.  The average
maximum solvent concentration in the dryer, during the recovery cycle
was measured as 3,100 ppmv, with a range of 2,800 to 3,500 ppmv.   Dryer
operating parameters such as load weight, fabric composition, recovery
duration, and cooling water flow rate were varied to examine their
effects on emissions, concentrations, and recovery.  The weight of the
wet articles placed in the dryer and the decrease in the vapor stream
temperature in the condenser were found to have the greatest influence
on solvent recovery and VOC emissions, with larger loads and greater
condenser temperature decreases resulting in lower emissions and higher
recovery.   In addition, increasing the drying temperature inside the
tumbler during recovery was found to increase the magnitude of the
maximum solvent concentration in the exhaust.   The measurement of these
concentrations, however, was hampered by problems with both plant and
testing equipment resulting from very low ambient temperatures,  and the
absolute magnitudes of these concentrations, when compared with previous
test data, were extremely low and probably inaccurate.
     The installation of currently-manufactured recovery dryers in
existing petroleum dry cleaning plants would involve the replacement of
standard dryers with recovery units.   Although steam (2.5 boiler horsepower
at 410 kilopascals) and electrical (1.5 motor horsepower at 230 volts)
demands and connections would be similar for both units, the recovery
dryer would require additional  connections of pressurized air (0.04 cubic
meters per minute at 275 kilopascals) and cooling water.   The demand for
cooling water in the 48 kg (105 Ib) recovery dryer is typically satisfied
by municipal water, a cooling tower,  or a chiller with a minimum output
capacity of 11.4 liters per minute (3 gal/min), and a temperature of
13C (55F) (Hoyt, 1979).   Meeting these cooling water specifications in
cool, dry climates may involve only a connection to municipal  water or
the installation of a relatively inexpensive cooling tower;  however,
                                 3-7

-------
hotter, humid climates may require the lower temperatures associated
with a chiller, which is inherently more expensive to buy and operate
(costing up to 250 percent more than a cooling tower).
     Proper operation of a currently manufactured recovery dryer would
be based primarily on maintaining high VOC emission reduction/recovery.
Adherence to dryer manufacturer's specifications in the areas of steam
pressure, condenser water inlet flow rate and temperature, tumbler
drying temperature, and tumbler load weight should ensure adequate
performance and safety.   Operating within the range of manufacturer's
specifications, tests have shown that VOC emission reduction/recovery
can be maximized by drying larger load weights (up to the rated capacity)
while optimizing heat transfer within the condenser for a more complete
removal of solvent from drying articles (Plaisance et al., 1981).
     Heat transfer in the recovery dryer condenser was found to be at
its maximum (defined as the highest average VOC emission reduction/
recovery per dry weight of articles cleaned) when the temperature decrease
of the condenser vapor stream was at its maximum (Plaisance et al.,
1981).  For a given drying temperature and cooling water flow rate, this
optimum operating condition could be indicated by the temperature
difference between the condenser cooling water inlet and outlet which
should not exceed about 15C (27F) during the recovery cycle, according
to the dryer manufacturer.
      Figure 3-3 illustrates a typical recovery phase that meets the
conditions stated previously.  The initial rapid solvent recovery gradually
decreases to a near constant value (little or no additional recovery
with time).  During this dryer load, the average condenser water inlet
and outlet temperatures are 21C (70F) and 30C (86F).  The concentration
of solvent vapor in the condenser gas inlet also is shown for the same
recovery phase.  Initial high concentrations correspond to rapid increases
in the volume of recovered solvent; however, as the duration of  the
recovery phase increases, the concentration of solvent vapor tends to
level off, thus indicating a minimum recovery phase duration necessary
for optimization of VOC emission reduction/recovery.
     The recovery dryer solvent  recovery rate data collected in  the
second large commercial plant recovery dryer test  (Plaisance et  al.,
                                 3-8

-------
CO
I
     OC.
     UJ


     O
	  Recovered  Solvent Volume (ml)

	  Concentration as  ppm Solvent

	  Rate of Solvent Recovery, ml/m1n
                                                                                              10000
                                                                                              9000
                                                                                                    O
                                                                                              8000
                                                                                                    
                                                                                             4000
                                                                                              3000  S
                                                                                             2000-
                0    2    4    6    8    10   12    14    16   18   20   22    24    26    28



                                      RECLAIM CYCLE TIME  (min)
                                                30
                                                                                                           600
                                                                                                           500
                                                               400
                                                                                                           300 o
                                                                                                           200
                                                                                                           100
                  Figure 3-3.  Solvent recovery and concentration curve  for  the  recovery dryer.

-------
1981) was analyzed by EPA to evaluate the relationship between  the  flow
rate of recovered solvent and the dryer emissions measured during the
exhaust cycle.   These data were selected for analysis because conditions
of fabrics, load weights, recovery durations, and operating parameters
varied over a range that might be typical in a large segment of the
industry.  The flow rate of recovered solvent at the end of each recovery
cycle was calculated.  A comparison of these results indicated that a
final recovered solvent flow rate of 0.05 liters per minute was the
highest value encountered in any of the loads examined.  Although this
value could not be correlated with a definitive VOC emissions value for
the exhaust cycle, an analysis of graphs of the recovered solvent flow
rate indicated that this 0.05 liter per minute flow rate could be used
to indicate a point of diminishing returns (see Figure 4-2).  Beyond
this level additional recovery cycle time would produce only minmal
increases in the volume of recovered solvent.  This evaluation indicates
that the 0.05 liter per minute recovered solvent flow rate could be used
as an indication of a minimum recovery cycle duration.  While the actual
exhaust cycle VOC emissions would continue to vary as a function of load
composition, weight, and operating parameters, the 0.05 liter per minute
rate could provide a basis for establishing recovery cycle durations
that reflect variations in loading and operating parameters.
     Floor and overhead space requirements are similar for standard and
recovery dryers and present little difficulty for smaller facilities;
however, the replacement of a single 180 kg (400 Ib) capacity standard
dryer with four 48 kg (105 Ib) capacity recovery dryers (currently,
recovery dryers are made in only 23 and 48 kg capacities) would necessitate
increasing the floor area required for dryers by approximately 30 percent
from approximately 28 to 37 square meters (300 to 400  square feet).  The
need for additional  floor space could  force some large industrial plant
operators to restructure the  layout of their dry cleaning equipment, and
may  require an addition to an existing building or the acquisition of
additional space  by  lease, purchase, or construction.  Most  of these
facilities, however, should have sufficient  space in  their  existing
plants  to accommodate the recovery dryers.
                                    3-10

-------
3.2  RECOVERY DRYER SAFETY
     Solvent recovery by chilled condensation is a new technology as
applied to petroleum dry cleaning.   With only one domestic manufacturer
producing recovery dryers since 1978, there are fundamental questions of
safety to be addressed.   These questions center on three major topics
which will be discussed below:  the concentration of solvent in the
dryer tumbler during recovery, the ignition sources associated with the
dryer and the effects of an explosion, and the acceptability of the
dryer to agencies such as fire marshals and insurance underwriters.
     Measurement of the maximum solvent concentration in the recovery
dryer tumbler during the recovery cycle was one of the objectives in all
three EPA field tests of the unit.   Unfortunately, difficulties associated
with measuring high concentrations of petroleum solvent in a vapor
stream cast doubts on the validity of much of the data obtained,  and the
only test with reasonably accurate maximum concentration measurements
was conducted in a plant that dried small  loads of synthetics, atypical
of most of the industry.   Furthermore, in  the only test of a recovery
dryer in an industrial plant, the concentration-measuring instrument
(FIA) became saturated with solvent and continued to register a concen-
tration of 9,000 ppmv while the actual concentration in the dryer apparently
continued to increase and probably exceeded the solvent lower explosive
limit (LEL) of 10,000 ppmv (Ashland,  1980).   In the third test, the
difficulty of preventing solvent vapor condensation in the FIA sample
lines produced unrealistically low readings of maximum concentrations.
Thus, while there are no test data showing dryer concentrations in
excess of the solvent LEL, an examination  of the existing data indicates
that there is a high probability, depending on the weight of articles
being dried and the drying temperature, that the concentration of solvent
in the tumbler does, indeed, exceed the solvent LEL at some point in the
recovery cycle.   And while this condition  would not occur in every
drying load, it would tend to occur in larger facilities where large
loads would be dried at higher temperatures to decrease the overall
drying time.
                                 3-11

-------
     Possible sources of ignition in the recovery dryer are limited  to
static electricity and flammable objects (matches and lighters)  contained
in the drying articles.   Dryer wiring and controls are contained in
enclosures that meet or exceed National  Fire Protection Association
(NFPA) regulations for dry cleaning dryers.   Also, the dryer tumbler is
fitted with a grounding system which, when properly grounded,  should
dispel 1 static electricity.   As a consequence of these design features,
the primary source of ignition would be  flammable objects,  which should
be removed from articles prior to washing.   While flammable objects
could be a significant problem in smaller plants that process personal
articles or uniforms, static electrical  spark would present the  greatest
problem in large plants that might process large volumes of static-prone
material such as felt or synthetics.
     Based on the assumptions that the solvent concentration in  the
recovery dryer tumbler reaches and exceeds the solvent LEL  and that  an
ignition source is present and active, what, in theory, occurs during a
recovery dryer "explosion"?  First, the  solvent vapor around the point
of ignition burns very rapidly (flashes), and the line of vapor  combustion
(or flame front) spreads rapidly through the tumbler.  The  extremely
rapid pressure increase brought about by the vapor combustion opens  the
spring loaded explosion dampers on the top of the tumbler,  and,  after
the excess pressure is instantaneously released, the dampers close to
prevent inflow of air that would support further combustion.  At the
same time, a set of weights attached to one of the dampers  activates
both a valve that injects steam into the tumbler and a switch that stops
the vapor circulation blower.  Thus, while the force of the vapor explosion
is released and directed upward away from personnel and other equipment,
the occurrence of a fire in the tumbler is prevented by the elimination
of combustion air and the injection of steam into the tumbler.  (In
contrast, the ignition of solvent vapors in a non-recovery dryer often
results in fires which are fed by the continuous  inflow of ambient air.)
     Two recovery dryer explosions have been documented since production
of the unit began approximately two years ago, and these have been
examined as to their causes and consequences.  (Other explosions have
been reported but specific details are unavailable.)  The first explosion
                                 3-12

-------
occurred in December of 1979 at a commercial  dry cleaning plant.   According
to the owner, the explosion took place during the drying of a 45  kg
(100 Ib) load of synthetic fabrics.   The owner said there was a loud
noise from the top of the dryer, and the tumbler basket was knocked out
of alignment.  The owner further explained that the dryer may not have
been properly grounded, and that the synthetic fabric's high potential
for static may have combined with the extremely low humidity at the time
to produce a static spark.  Damage resulting from the event was limited
to the drying articles which were scorched and to the dryer (tumbler
basket) which was replaced by the manufacturer.
     The second recovery dryer explosion occurred in March 1981 at a
small industrial dry cleaner.  A load of felt grain mill filtration bags
was being dried at the time and a member of the plant management  explained
that the combination of the high static potential of the felt material
and the presence of grain dust probably combined to produce the explosion.
This individual was standing about four feet from the dryer when  it
exploded, and he described the effect as being like the sonic boom from
a jet airplane.  No personal injuries resulted from the explosion, and
damage to the plant was limited to two plate-glass windows which  were
broken.  The only damage to the dryer was disalignment of the tumbler
basket shaft which was quickly realigned by plant personnel and tearing
of the lint filter bag which was replaced.  Of the 12 bags in the tumbler,
only five had to be replaced due to minor scorching.  The manager said
that he intended to improve the dryer grounding and to vent the dryer
explosion dampers to the outside.  Finally, he said he was pleased with
the design, performance, and safety of the dryer.
     Because the technology of condensation recovery of petroleum dry
cleaning solvent is relatively new, the EPA conducted an examination of
the design, performance, and safety of solvent recovery dryers in Japan
(Jernigan, 1981).  These dryers have been in use there for over five
years, and their design and performance is very similar to that of the
units manufactured in the United States.  In both units, steam heated
air evaporates  solvent from drying articles, and the solvent vapor is
liquified and  recovered by a refrigerant-chilled condenser.  The main
difference between the Japanese and domestic recovery dryer  is the dryer
                                 3-13

-------
capacity which is limited to 23 kg (50 Ib)  in Japan,  with units  having
capacities of 10 kg (22 Ib) being most prevalent.   Throughout Japan,
there are about 1,800 recovery dryers and 5,900 standard (nonrecovery)
dryers.   Additionally, according to Japanese dryer manufacturers,  there
have been about 17 recovery dryer explosions and 50 standard dryer fires
in the past five years.  Thus, the frequency of explosions in Japanese
recovery dryers has been almost identical to that of fires in standard
dryers (about one occurrence per 1,000 dryers in the past five years).
     The primary concern involved in the acceptance and approval of the
recovery dryer by fire control agencies and insurance underwriters is
the level of fire hazard inherent in the operation of the dryer.  In
order to gain this approval, the manufacturer has submitted the dryer
for examination by Factory Mutual (a firm specializing in performance
and safety testing) which has approved the unit and has included it in
the 1981 published approval listing as the petroleum solvent dryer with
current approval (Factory Mutual System, 1981).  In addition, the dryer
has received the general approval of the Los Angeles Fire Department
(Los Angeles Fire Department, 1981).
     A survey of dry cleaning plants that have installed the recovery
dryer indicates that there have been no problems with insurance companies
resulting from the operation of the recovery dryer (Corey, 1981).   In
general, these companies view the unit as another piece of dry cleaning
process equipment and as such, insurance rates usually increase somewhat
based on the addition of the dryer value to the overall facility.   When
underwriters insuring the dryer were questioned about the insurability
of the unit, they generally said that approval by one of the major
testing laboratories (Factory Mutual), together with approval by state
and local fire officials, was sufficient to indicate that the dryer
requires no specialized  insurance coverage.
     At the local level, a survey of fire marshals in areas where recovery
dryers have been installed indicates that the  unit's approval by Factory
Mutual has resulted  in  initial acceptance by fire marshals who were
aware of the dryer's presence (Corey, 1981).   In most cases, however,
the continuation of  this initial approval was  contingent on the frequency
                                 3-14

-------
 of  explosions  and  accidents  associated with  the  dryer,  and  an  increase
 in  the  number  of dryer  accidents  could result  in revocation  of  this
 approval.
      In  summation,  the  initial  approval of the recovery dryer by  individual
 dry cleaners,  fire  marshals,  and  insurance underwriters was  based on  the
 unit's  approval by  Factory Mutual and on  its record of  safety since
 1978.   Factory Mutual,  in turn, has based its approval  of the unit on
 its ability  to safely contain and control an explosion,  with the assumption
 that the solvent concentration  in the dryer  reaches and exceeds the
 solvent  LEL  and that an  ignition source is present and  active in the
 dryer (Kennes, 1981).   Industry representatives  have expressed concern
 that the requirement for installation and operation of  solvent recovery
 dryers could conflict with local fire safety codes or create situations
 in  which petroleum  dry cleaning plants could not be insured (Vanderver,
 1982).  Even though Factory Mutual has certified the one U.S.-made
 recovery dryer, agencies should ascertain whether this  is sufficient in
 their state or locality.  Although EPA has found no instances in which a
 fire  marshal or insurance underwriter prohibited the installation of a
 solvent recovery dryer,  appropriate fire safety officials and insurance
 industry representatives should be involved early in the process of
 developing any regulations.
 3.3  CARTRIDGE FILTRATION
     Cartridge filtration is a continuous, two-stage process of filtration
 in which soil-laden liquid is forced under pressure first through a
 paper filter to remove entrained solids  and then through a layer or
 layers of activated carbon which selectively entrap molecules of impurities
 in their porous surface.   The term "cartridge"  is used to denote replaceable
 units or cartridges containing filtration  paper and carbon or only
 carbon.   Currently, it is estimated that  50 percent of the petroleum  dry
cleaning plants using solvent filtration  employ cartridge filtration  or
approximately 3,000 plants  (Fisher,  1980).
     Cartridge filtration,  as applied  to  the  petroleum solvent dry
cleaning industry,  is a  process  in which  soil-laden solvent  is  pumped
from a washer to  a  vessel  containing filter  cartridges  (see  Figure  3-4).
                                 3-15

-------
Soil-Laden
 Solvent
   Inlet
                Paper Filter
                                                         Activated  Carbon

                            _J	J
                       Carbon-Core Cartridge
All-Carbon Cartridge
                                 Filtered
                                 Solvent
                                  Outlet
                                Figure 3-4.   Cartridge Filtration System Schematic.

-------
This vessel normally has a removable,  pressure-sealed lid or top and can
contain from 2 to 36 cartridges.   Soil-laden solvent is initially forced
under pressure through dual-component cartridges that contain both
filter paper and carbon.  In this process,  solid particles of lint and
dirt are trapped in the paper, and the included activated carbon serves
to remove soluble impurities such as fabric dyes.   Next, the solvent is
diverted to one or more filter cartridges containing only activated
carbon which continue the initial removal of soluble impurities.  After
passing through this final stage, the solvent is transferred to storage
to await distillation and reuse (Puritan, 1980).
     Currently manufactured filter cartridges fall into two distinct
categories:  carbon-core cartridges and all-carbon cartridges.   Carbon-core
cartridges (see Figure 3-4) are encased in an outer metallic housing
that is perforated around its circumference to permit solvent inflow.
Beneath this outer rigid structure lies a circumferential layer of
filtration paper that is folded accordian-style into a deeply-corrugated
cylinder surrounding the inner core.  This fibrous paper, similar to
that found in an automotive oil filter, permits the pressurized solvent
to flow inward to the core while trapping particles of dirt and lint
along its extensive surface.  Beneath this layer of filter paper lies
the slotted metal surface of the core tube which contains granular
activated carbon (Puritan, 1980).  In passing through this material, the
larger molecules of impurities such as fabric dyes are adsorbed by the
carbon granules.  The all-carbon cartridge (see Figure 3-4) continues
the purification of the solvent which began at the carbon-core cartridge.
Solvent flowing through the slots of the metal canister is forced through
the enclosed activated carbon, resulting in additional removal  of
impurities and the outflow of a solid-free, purified solvent.
     Although the process flow of carbon-core to all-carbon is typical
of most cartridge filter installations, the containment, number, and
physical arrangement of the cartridges varies widely as a function of
the system's capacity.  Cartridge filtration systems are "sized" by
their manufacturer, based on the dry-weight load capacity of the existing
washer (Puritan, 1980).  The size of the system usually refers to the
total number of  filter  cartridges it contains.  The actual distribution
                                  3-17

-------
of these cartridges ranges from a group of small,  interconnected  vessels
containing one or two cartridges to one or more cylindrical  tanks con-
taining as many as 36 carbon-core cartridges that  are connected to an
additional vessel or vessels containing multiple all-carbon  cartridges
(Puritan, 1980).
     The operation of a cartridge filtration system can be based on
either the continuous (during washer operation) or batch (at predeter-
mined intervals)  processing of spent solvent.   In  continuous operations,
the spent solvent in the washer is pumped through  the filter and back to
a filtered solvent tank.  As this process continues on a day-to-day
basis, the outer surfaces of the filter paper in each carbon-core cartridge
become clogged with dirt and lint, while the carbon granules contained
in both types of cartridges become coated with dyes and particles until
they no longer purify the solvent stream.  Dry cleaning trade association
tests have shown that, under typical commercial conditions of soil-loading
and throughput, the "life" of a filter cartridge is somewhere between
450 kg (1,000 Ibs) and 500 kg (1,100 Ibs) of articles washed (Bee and
Fisher, 1976).  Under conditions of heavy soil loading, however, the
need for cartridge replacement is best indicated by the buildup of
solvent pressure in the vessel due to the flow restriction posed by the
clogged cartridges.
     Atmospheric emissions from cartridge filters  are limited to fugitive
emissions that evolve from leaks and filter cartridge replacement, as
well as from the evaporation of solvent contained in disposed cartridges.
An EPA test of the amount of solvent contained in discarded filter
cartridges was conducted at a Wilmington, North Carolina petroleum
solvent dry cleaning plant that processes approximately 180 kg (400 Ibs)
of general apparel per day in a single 27 kg (60 Ib) capacity washer
(Plaisance, 1981).  A 14-element cartridge filtration system was operated
without cartridge replacement over a period of time when the plant had a
throughput of approximately 8,600  kg (19,000 Ibs.) of clothes washed.
Results of this  test  indicated that draining the filter cartridges in
their closed  housing  for at least  8 hours would result  in an average
solvent emission per  cartridge of  1.6  kg  (3.4  Ibs).  Based on an assumed
cartridge  life of 450 kg  (1,000  Ibs) of  throughput, this would result in
                                 3-18

-------
0.35 kg of solvent being emitted per 100 kg of clothing throughput
(Plaisance, 1981).   Figure 3-5 illustrates the effect of drainage time
on VOC emissions from discarded filter cartridges.   The majority of the
drainage takes place during the initial  few minutes when liquid solvent
is running freely from the canisters.   After 8 hours of drainage, the
cartridges have lost 37 percent of their undrained solvent content.
After an additional  4 hours, they have lost only 3 percent more, thus
illustrating that extended drainage periods are unnecessary.
     The value of cartridge VOC emissions per 100 kg dry weight of
articles cleaned, obtained above, is in general agreement with a dry
cleaning trade association estimate of 0.75 kg per 100 kg of articles
cleaned (NID, 1971).  However, the previously described EPA-sponsored
test could have been more in agreement with the industry estimate if the
emissions from carbon-core cartridges alone had been considered.  Also,
in the determination of cartridge weight losses, it was assumed that all
weight losses were due to solvent evaporation, thus ignoring the presence
of water and its evaporation.   However,  this is considered insignificant.
     The emission reduction/recovery efficiency of the cartridge filter
compared with the diatomite filter is 88 percent, based on average VOC
emissions for cartridge and diatomite filters of less than 1 kg and
8 kg, respectively,  per 100 kg dry weight of articles cleaned.  This
major emission reduction primarily results from diatomite filter emissions
that typically occur daily, while cartridge filter emissions result from
infrequent cartridge element replacement.
     Installation of a cartridge filtration system would require the
removal of the existing filtration system (usually a diatomite filter)
and the connection of the cartridge vessel to existing lines.  No utility
connections  (steam or electricity) are normally required, and the only
retrofit problems that might be encountered concern the allocation of
space (larger cartridge systems can occupy a substantial area) and the
possibility  of replacing the filter feed pump (maintains solvent pressure
to the filter).
     There is a current controversy in the dry cleaning industry concerning
the cartridge filter's ability to eliminate the need for distillation.
Although some plants have ceased their vacuum still operations after the
                                 3-19

-------
                              SOLVENT EMISSIONS
                      (kg solvent/100 kg  throughput)
  -s
  Q

  OJ
   I
  c_n
O on
-h o

o <
-s ro
OJ 3
 ' r-l-
3
a> m
= O
Tl ^
  l/l
  o
  -5
   -s
   c-t-
   -s

   Q.
  -Q
   03
   C/1
   Cu

   -n
   c

-------
installation of cartridge systems (Plaisance, 1981), both cartridge
filter manufacturers (Puritan, 1980) and dry cleaning trade associations
(Bee and Fisher, 1976) maintain that distillation is a necessary part of
the process of solvent rejuvenation and that it is the only method for
removing solvent-soluble impurities such as grease and oil.   The instal-
lation of cartridge filtration equipment could, however, decrease the
frequency of solvent distillation, and thus reduce not only filtration
emissions, but also emissions associated with still wastes.   Small
plants with light soil loadings, in particular, could reduce their
frequency of distillation by using cartridge filters to maintain the
clarity of their solvent.
3.4  VACUUM DISTILLATION
     Atmospheric emissions resulting from the operation of vacuum stills
(see Section 2.3.4 for a distillation process description) are a function
of the still design, operation, and the frequency of still utilization.
To maintain the color (purity) of the solvent used in washing, a dry
cleaning plant operator will often adhere to a still boildown schedule
that requires solvent distillation at intervals ranging from daily to
weekly.  Also, the design of some stills is such that, regardless of the
boildown period, there will be a fixed volume of solvent-laden residue
at the end of the operation.
     Specific operating and maintenance parameters cannot be established
that produce commensurate VOC emissions levels due to the variability of
still design, throughput, and soil loadings.  However, general parameters
indicative of acceptable operating conditions will be delineated as a
guideline to minimizing VOC emissions from still residue.
     According to a still manufacturer, one of the principal influences
on the solvent content of still residue is the frequency of still boildown.
Although many existing plants boil down their stills on a routine or
convenient schedule, the still manufacturer recommends that boildown be
undertaken only after the flow rate of condensed liquid (solvent and
water) between the condenser and moisture separator has been reduced by
approximately 75 percent (Washex, 1973).  The duration of boildown
should continue until the flow rate of condensed liquid has been again
                                 3-21

-------
reduced by 75 percent,  with the full  available steam pressure  of
390 kilopascals (100 psi) being used  to evaporate solvent from the
residue.   Thus, the 75  percent condensate flow reduction criterium can
be applied to both normal distillation operation and to the optimum
duration of boil down.
     Factors affecting  the rate of distillation include the solvent
boiling range, the level of solvent contamination, the quantity of
residue in the still,  and the steam pressure.   Under a typical vacuum
range of 42 kilopascals to 52 kilopascals (22 inches to 27 inches of
mercury), petroleum solvents with flash points of 41C, 52C,  and 60C
boil under recommended  steam pressure ranges of 136 kilopascals to
234 kilopascals (35 to  60 psi), 195 kilopascals to 253 kilopascals
(50 to 65 psi), and 292 kilopascals to 351 kilopascals (75 to 90 psi),
respectively (Washex,  1973).  However, these steam pressure ranges will
increase significantly  with an increase in the quantity of contaminants
in the incoming solvent.  The level of residue in the still boiling
chamber may also necessitate higher steam pressures in order to overcome
the poor heat transfer of the accumulated residue.
     A still manufacturer lists the following typical operating parameters
for satisfactory still  operation (Washex, 1973):
          Steam Supply = 390 kilopascals (100 psi)
          Condenser Water Supply = 78 kilopascals @ 18C (20 psi @ 65F)
          Condenser Water Outlet Temperature = 60C to 71C (140F to
          160F)
          Condenser Solvent Outlet Temperature =  24C to 32C (75F to
          90F)
While these parameters will vary as a function of the given still, plant
throughput, and soil loading, they nevertheless  form a general range
which can be  indicative  of  proper still operation.  They are even more
effective when  combined  with the previously discussed guidelines  on
still boildown  duration  and procedures.  Principal  maintenance procedures
are also  outlined  by the still manufacturer and  include  removal  and
cleaning  of  the steam  coils after about  1,000  hours of  operation  and
frequent  lubrication of  the condensate pump (Washex, 1973).
     The  application of  the previously discussed operating and maintenance
recommendations could  result  in  significant VOC  emission  reductions due
                                  3-22

-------
to the resultant decrease in the solvent content of disposed still
waste.  However, no specific data are available on reductions in still
waste solvent content as a function of the implementation of these
procedures.  Consequently, VOC emissions from vacuum still waste stored
in a given dry cleaning plant are included in the general category of
fugitive emissions.
3.5  MISCELLANEOUS FUGITIVE EMISSION SOURCES
     Miscellaneous fugitive emissions essentially encompass sources
wherein liquid solvent or solvent-laden wastes are exposed to the
atmosphere within the plant.   The EPA-sponsored test program conducted
at the Anaheim, California dry cleaning facility included a sampling and
evaluation of VOC concentrations in the general dry cleaning environment.
Significant concentrations of solvent vapor, at times approaching 70
percent of the solvent LEL, were found around both the settling tank and
the new solvent tank vents.  Another significant source of VOC emissions
was the washer which produced concentrations nearing 10 percent of the
solvent LEL in its immediate vicinity.   Additional measurements of VOC
concentrations in dry cleaning room roof exhaust vents yielded a VOC
emission rate of 0.5 kg solvent per 100 kg of articles cleaned (Jernigan
and Kezerle, 1980).
     Solvent vapor losses from settling and storage tanks occur as a
result of "breathing" and "working" losses.   Breathing losses occur when
storage tanks expand or contract during changes in temperature, resulting
either in air being drawn into the tank (vapor contraction) or solvent
vapor being expelled to the atmosphere (vapor expansion).   In contrast,
working losses result from changes in the vented free volume above the
stored liquid solvent, which expels solvent vapor when the tank is
filled and draws in atmospheric air when the tank is drained, thereby
producing additional  breathing losses as the air becomes saturated with
solvent vapor.
     The broad category of "leaks" can contribute significantly to
miscellaneous fugitive VOC emissions.   Liquid solvent drips from pipes,
fittings, valves, hoses, couplings, and pumps add to the constant back-
ground of solvent vapor inherent to many dry cleaning plants.   Vapor
                                 3-23

-------
leaks from dryers, exhaust ducts, filter housings,  stills,  and open or
improperly sealed containers of solvent all  contribute to the quantity
of solvent entering the environment.
     The only way to eliminate the general  class of fugitive emissions
is with an effective program of maintenance and training.   A dry cleaning
trade association has estimated that approximately  1.0 kg of VOC are
emitted by miscellaneous fugitive sources for every 100 kg of articles
cleaned in a typical dry cleaning facility (Fisher, 1975).   This level
of VOC emissions could be reduced by effecting a maintenance program
that would effectively control liquid leaks and eliminate buckets and
barrels of solvent standing open to the atmosphere, while also striving
to eradicate vapor leaks by repairing gaskets and seals that obviously
expose solvent-rich environments to the atmosphere.  Inspection of
solvent pumps, storage and settling tanks,  water separators, and the
general solvent piping system could result in substantial  VOC emissions
reductions through control of liquid and vapor leaks.   Also, training of
dry cleaning personnel could help in attaining this reduction, particularly
by discouraging the practice of allowing solvent-laden loads of articles
to be exposed to the atmosphere while awaiting drying.
3.6  COMBINED CONTROL TECHNIQUES
     A combination of the previously discussed control techniques would
result in an optimum emission reduction that could be achieved by employing
currently available equipment and methods.   The following list tabulates
the control techniques and the anticipated range of controlled emissions.
The nominal emissions values are listed in order to form a basis for
comparison with estimates of uncontrolled emissions, and these values
may vary widely as a function of individual plant throughput and operation.
                                                          Nominal value of
                         Range of emissions in kg VOC/    emission in kg VOC/
                                                          100 kg dry weight
                                                          of articles cleaned
Emission
source
Dryer
Diatomite
filter
Control
technique
Recovery
dryer
Cartridge
filter
100 kg dry weight of
cleaned
0.7-9.5
0.5-1.0
articles

                                                                 3.5

                                                                 1.0
                                 3-24

-------
                                                               Nominal  value of
                              Range of emissions in kg VOC/    emission in kg VOC/
Emission         Control      100 kg dry weight of articles    100 kg dry weight
 source         technique     cleaned	                of articles cleaned
Vacuum          Improved                1.0-7,0                       3.0
still           operation
Miscellaneous   Improved                0.5-1.0                       1.0
fugitive        operation
Total Range                            2.7-18.5                     7.5-8.5
               Totals with existing
                 Diatomite filter                                   8.5
                 Cartridge filter                                   8.5
                 Settling tank                                      7.5

          Implementation of the above processes and methods in the two model
     plants,  as illustrated in Table 3-1,  would result in an overall  dry
     cleaning plant emission reduction of  72 to 66 percent over the existing
     levels discussed in Chapter 2.   Model  plant II,  with a settling tank in
     both existing and controlled configurations,  does not benefit from the
     88 percent VOC emission reduction resulting from the replacement of
     diatomite with cartridge filters.
     3.7  EVALUATION OF CONTROL TECHNOLOGY  TRANSFER
          Control  of dryer emissions in the perchloroethylene (perc)  dry
     cleaning industry has been accomplished for many years by application of
     existing carbon adsorption technology.   Currently,  35 percent of perc
     dry cleaning  plants (5,400) use carbon adsorption (EPA,  1980).   The
     application of this technology to petroleum dry  cleaning dryer VOC
     emissions reduction could result in savings due  to  reductions in development
     and testing time and costs.
          An  EPA demonstration program was  undertaken at an Anaheim,  California
     industrial  petroleum dry cleaning facility (Lutz et al.,  1980) to evaluate
     the performance of carbon adsorption as a  means  of  reducing VOC  emissions
     from a standard petroleum solvent dryer (see  Appendix A,  Test 7).   An
     adsorption system with two carbon vessels  was connected  to the exhaust
     of a 180 kg (400 Ib) capacity standard dryer, and the VOC concentrations
     at the adsorber inlet and exhaust were monitored throughout the  course
                                      3-25

-------
                 Table 3-1.  MODEL PLANT PARAMETERS
                            (Control Equiment)

Model plant
Number of existing plants
nationwide in 1979
Annual plant throughput,
in kg (Ibs)
Throughput description
Number of washers
Washer capacity (each),
kg (Ibs)
Number of recovery dryers
Dryers capacity (each),
kg (Ibs)
Filtration options
cartridge filter, number of cartridges
settling tanks, number per plant
Total vacuum still capacity
L/hr (gal/hr)
Number of loads dried per day
Average load weight per dryer
kg (Ibs)
Days of operation per year
Wash cycle time, minutes
Dry cycle time, minutes
Average or range of baseline
emissions per plant, megagrams
VOC/year (tons VOC/year)
Average or range of baseline
emissions nationwide, megagrams
Model
plant I
60
182,000
(400,000)
industrial
articles
1
115
(250)
3
48
(105)
63
1
1,900
(500)
20
35
(80)
260
40
40
13.7-15.5
(15-17)
819-930
(900-1,020)
Model
plant II
170
635,000
(1,400,000)
industrial
articles
2
225
(500)
8
48
(105)
0
1
5,700
(1,500)
50
48
(105)
260
40
40
47.7
(52.5)
8,100
(8,905)
VOC/year (tons VOC/year)
                                 3-26

-------
of normal daily operations in which approximately 1,600 kg (3,500 Ibs)
of general apparel was cleaned.  Results of the test program indicated
that the average difference in VOC concentrations between the adsorber
inlet/dryer exhaust (2,100 ppm as solvent) and outlet (100 ppm as solvent)
over the duration of typical drying cycles represented an overall reduction
in VOC concentration of 95 percent.
     A comparative analysis of the overall potential for reducing dryer
VOC emissions from the carbon adsorber and the recovery dryer indicates
that the carbon adsorber produces approximately the same amount of VOC
emissions as to the recovery dryer.   And, although the adsorber unit
produces a 95 percent reduction in the VOC concentration emitted to the
atmosphere, the solvent recovery efficiency has been limited to
approximately 75 percent, thereby indicating that the savings resulting
from solvent recovery would be lower with the carbon adsorber.   The
discrepancy between the recovery and emission reduction efficiency has
been found to be caused by vapor leaks in duct work before the carbon
beds.   Therefore, the actual amount of recovered solvent is less than
the removal efficiency across the beds.   This variation between emissions
and recovery, together with the high capital  costs of the adsorber
system, results in the adsorber's being considered less desirable than
the recovery dryer as a VOC emissions control technique.
3.8  REFERENCES FOR CHAPTER 3

Ashland Chemical Co.   1977.  Material Safety Data Sheet No.  0000585-001.
     Environmental and Occupational  Safety Department.   Ashland, Kentucky.
Bee, W. and W.  Fisher.   1976.   Report on Cartridge Filtration Life.
     International Fabricare Institute (IFI).   Joliet,  Illinois.  Focus
     No.  1.
Corey,  Q.   1981.  TRW Inc., Letter to S.  Shedd, EPA/CPB,  February 10.
     Summary of insurance underwriters'  and fire marshals'  views on the
     solvent recovery dryer.
EPA.   1980.  Perchloroethylene Dry Cleaners - Background Information for
     Proposed Standards.   OAQPS.   Research Triangle Park,  North Carolina.
     Publication No.  450/3-79-029a.
Factory Mutual  System,  1981,   Approval Guide for Equipment,  Materials,
     and Services for Conservation of Property.  Norwood,  Massachusetts.
                                 3-27

-------
Fisher, W.   1975.   ABC's of Solvent Mileage,  Part 1.   International
     Fabricare Institute (IFI).   Joliet, Illinois.   Special Reporter
     Vol.  3, No.  4.   July-August.

Fisher, W.   1980.   International  Fabricare Institute (IFI), Meeting with
     S. Plaisance, TRW Inc., December 9.  Number of dry cleaning plants
     using cartridge filters.

Hoyt Manufacturing,  Inc.  1979.  Sales Brochure:   The "Petro-Miser" 105.
     Westport, Massachusetts.

Jernigan,  R.  1981.   A report on the design,  safety and performance of
     Japanese recovery dryers.   TRW Inc., Research Triangle Park, North
     Carolina (EPA Contract No.  68-03-2560).

Jernigan,  R. and J.  Kezerle.  1980.  Evaluation of the Potential for
     Reduction of Solvent Losses Through a Washex Petroleum Vacuum Still
     Sump.   TRW Inc.  Research Triangle Park, North Carolina (EPA/IERL
     Contract No.  68-03-2560, Task No. T5013).

Jernigan,  R. and S.  Lutz.  1979.   An Evaluation of the Emission
     Reduction Potential of a Solvent Recovery Dry Cleaning Dryer.  TRW
     Inc.   Research Triangle Park, North Carolina (EPA Contract No.
     68-03-2560).

Jernigan,  R., G. May, and S. Plaisance.  1981.  An Evaluation of the
     Emission Reduction Performance of a Solvent Recovery  Dry Cleaning
     Dryer Under Varying Conditions of Condenser Water Temperature.  TRW
     Inc.   Research Triangle Park, North Carolina (EPA Contract No.
     68-02-3063).

Kennes, F.   1981.    Factory Mutual  System, Telecon with S.  Plaisance, TRW
     Inc., February 11.  Basis for Factory Mutual approval of the  recovery
     dryer.

Los Angeles  Fire  Department.  1981.  General Approval of Hoyt Petro-Miser.
     L.A.F.D. No.  12/81/1.  Los Angeles, California.

Lutz,  S., S.  Mulligan and A. Nunn.  1980.  Demonstration of Carbon
     Adsorption Technology  for Petroleum Dry Cleaning Plants.   Cincinnati,
     Ohio.  EPA Publication  No. 600/2-80-135.

NID.   1971.   Estimation  of  Solvent Vapor Emission from Petroleum  Dry
     Cleaning Plants.   IFI.  Joliet,  Illinois.  Technical  Bulletin No.
     T-468.

Plaisance,  S.  1981.  A  Study of  Petroleum Dry  Cleaning  Cartridge  Filter
     Element Emissions.  TRW, Inc.   Research Triangle Park, North  Carolina
     (EPA Contract  No.  68-02-3063).

Plaisance,  S., J. Jernigan, G. May,  and C. Chatlynne.  1981.  TRW Inc.
     Evaluation of  Petroleum Solvent  Concentrations,  Emissions,  and
     Recovery in  a  Solvent  Recovery  Dryer  (EPA  Contract  No. 68-02-3063).


                                  3-28

-------
Puritan.  1980.  Sales Brochure:  Modern Filtration Means Puritan
     Filtration.   R. R. Street and Company.  Oakbrook,  Illinois.  Bulletin
     No. 1289.

Washex.   1973.  Installation, Operation and Maintenance Manual  for
     Washex Vacuum Stills.  Publication No. T-513d.  Wichita Falls,
     Texas.   July.

Vanderver, 1982.   Patton, Boggs, and Blow.  Letter to F. Porter, EPA/ESED,
     January 15.   Summary of dry cleaning  industry comments on  the
     November, 1981, Draft CTG for Large Petroleum Dry Cleaners.
                                 3-29

-------
                     4.  ENVIRONMENTAL ANALYSIS OF RACT

     The installation of RACT equipment and the implementation of RACT
procedures in a typical dry cleaning plant involve the replacement of
existing dryers with recovery dryers, the replacement of existing diatomite
filtration systems with a cartridge filtration system, and improved
operating and maintenance procedures to identify and repair liquid and
vapor leaks.
     The environmental  impacts of RACT implementation on air,  water, and
solid-waste disposal are discussed in this section.   In addition, the
effects of RACT equipment operation on overall energy consumption are
detailed, based on the two model plants that were discussed in Chapter
2, and these values are compared with those of uncontrolled model plants.
Finally, beneficial and adverse effects from the installation  of RACT
equipment are assessed in relation to emissions and energy consumption
in these model plants.
4.1  AIR POLLUTION
     Table 4-1 lists the estimated uncontrolled VOC emissions  for each
emission point and indicates the range (or nominal value) of controlled
emissions per 100 kg of articles cleaned.   Because the uncontrolled
dryer provides from 60 to 80 percent of the total emissions, effective
control and reduction (81 percent) of VOC emissions from this  source
provides the greatest direct impact on overall plant emissions.
     Filtration system VOC emissions in dry cleaning facilities  with
existing non-RACT diatomite filters account for about 25 percent of the
total uncontrolled emissions.   Filtration emissions in these facilities
will be reduced by as much as 88 percent as a direct result of cartridge
filter installation. However,  RACT will provide no reduction in  filtration
system VOC emissions in facilities that have existing settling tanks or

-------
 Table 4-1.  NOMINAL EMISSIONS FACTORS FOR EXISTING AND RACT EQUIPMENT

      (in  kg VOC emitted per 100 kg dry weight of articles cleaned)
Source
Dryer
Filter
Diatotnite
Cartridge
Still
Fugitive sources
Total
Existing
equipment
emissions
18
8
1
3
1
22-30
RACT
equipment
emissions
3.5
1
1
3
1
7.5-8.5
VOC emission Percent
reduction reduction
14. 5 81
7 88
0 0
0 0
b b
14.5-21.5
Existing equipment emission estimates are based on industry association
data and EPA plant tests, and represent approximate midrange for most
sources.  See Section 2.0 for complete explanations of controlled
emissions sources and levels.

 Indeterminate quantity.
                                   4-2

-------
cartridge filters.   In addition,  the VOC emissions resulting from fugitive
sources could be directly reduced by improvements  in maintenance and
operating procedures.   Thus,  RACT equipment and procedures would produce
average VOC emissions  reductions  ranging from 66 to 72 percent.
     Table 4-2 illustrates the VOC emissions reductions that result from
the installation of RACT equipment and the adoption of RACT operating
and maintenance procedures in two model  plants.   Based on three  uncon-
trolled emissions rates representing the three filtration alternatives
in model plant I and two RACT emissions rates, the model plants  show a
66 to 70 percent reduction in model plant I and a 66 percent reduction
in VOC emissions in the model plant II.   The specific reductions in VOC
emissions range from 26 Mg to 39  Mg per year in a model plant I  and
approximately 92 Mg per year in a large model plant.
4.2  WATER POLLUTION
     Increases in water pollution, due to RACT implementation in petroleum
dry cleaning plants, would result primarily from inefficient separation
of condensed solvent and water.  Recovery dryers employ gravimetric
separators to remove water from the reclaimed solvent.  This unit uses
the difference in density between petroleum solvent and water to separate
and divert them.  Typically, water collected in this manner is dumped
into a sewer.  The leveling of the separator is critical to the
optimization of its performance.   If it is not level at installation or
is bumped during maintenance, the quantity of solvent  in the sewered
water could  increase to the point of becoming a significant source of
water pollution.
     Insufficient drainage of RACT filter cartridges could prove to be a
source of groundwater pollution,  especially  if the  cartridges were
buried in an improperly  located or maintained landfill  or dump.   RACT
procedures for cartridge drainage would decrease the overall volume of
solvent  exposed to groundwater and would, therefore, reduce water
pollution by petroleum solvent.  Furthermore, based on  a maximum
solubility of 100 kg  (Saary, 1981) of petroleum solvent in  1,000,000  kg
of  recovered water and an average  recovery dryer water  recovery  rate
of  3.4  kg water per 100  kg of  articles  dried  (Plaisance et  al.,  1981),
                                  4-3

-------
Table 4-2.   NOMINAL ANNUAL VOC EMISSIONS FOR TWO MODEL PLANTS
    EMPLOYING EXISTING AND RACT EQUIPMENT AND PROCEDURES

Nominal emission factors


Type
of plant
Model plant I

with existing:
Diatomite filter

Cartridge filter

Settling tank

Model plant II


Plant
throughput,
kg/yr
(Ib/yr)
182,000
(400,000)







635,000
(1,400,000)
in kg
per 100
VOC emitted
kg dry weight
of articles cleaned
Existing
equipment



30

23

22

22

RACI
equipment



8.5

8.5

7.5

7.5


Nominal VOC
megagrams/yr
Existing
equipment



55
(60)
42
(46)
41
(45)
140
(154)

emissions,
(tons/yr)
RACT
equipment



16
(17)
16
(17)
14
(15)
48
(53)
Nominal annual VOC
emission reductions
resulting from
RACT implementation,
megagrams/yr (tons/yr)



39
(43)
26
(29)
27
(30)
92
(101)

-------
model plant I and model plant II would lose about 0.5 kg and 1.5 kg per
year, respectively, from solvent dissolved in the recovered water.
4.3  SOLID-WASTE DISPOSAL
     Implementation of RACT in existing petroleum dry cleaning facilities
would result in a net reduction in both the mass and solvent content of
solid wastes.   Installation of RACT cartridge filters would produce a
dramatic decrease in emissions from solid wastes in petroleum dry cleaning
plants.   Cartridge filters, when compared with diatomite filters, have
been shown to reduce solvent content of disposed filter wastes by 80 to
90 percent (Plaisance, 1981), thereby decreasing the overall quantity of
solvent-laden solids introduced to the environment.   In addition, the
replacement of diatomite with cartridge filters will reduce the mass of
solid waste generated by 60 percent, based on an average industry estimate
of 3.57 kg of waste generated per 100 kg of throughput with a diatomite
filter (Fisher, 1975) and 1.47 kg solid waste per 100 kg of throughput
for a cartridge filter (Plaisance, 1981).
4.4  ENERGY
     Energy savings result from the implementation of RACT guidelines in
both model plants.   With the installation of RACT recovery dryers and
cartridge filters in the model plants, annual expenditures for both
steam and electricity are reduced by a combined average of 70 percent
over utility costs for existing standard dryers and diatomite filters
(see Section 5.2 and 5.3).
     The energy value of recovered solvent is included in the overall
analysis of petroleum dry cleaning plant energy consumption.  One approach
to this analysis that would be meaningful  to the dry cleaning industry
is to assume that all recovered solvent is resold at its current market
value ($0.53 per kg) and that the proceeds are used to purchase electricity
at its current market value of $0.0603 per kilowatt-hour (Vatavuk,
1980).   This approach to energy conservation by solvent recovery
illustrates a savings of energy accrued directly to the individual
petroleum dry cleaning plant.
     Table 4-3 delineates the impact of RACT implementation on model
plant energy consumption per year based on the previously discussed
                                 4-5

-------
        Table 4-3.   ENERGY IMPACT OF EXISTING AND RACT EQUIPMENT3
                        (in Gigajoules per year)

Model plant
Model plant I
Model plant II
Existing
equipment
1,865
6,070
RACT
equipment
(660)b
(1,040)
Percent
reduction
135
120
 Based on 0.00314 GJ/kg steam and 0.0036 GJ/kWh electricity
(Baumeister et al.,  1978),  and utility consumption and solvent recovery
values calculated in Chapter 5.

 Numbers in parenthesis represent overall  energy savings,  based on
savings from solvent recovery (at $0.53 per kg) to purchase electricity
at a cost of $0.0603 per kWh.
                                  4-6

-------
 approach.   Considerable  energy  savings  (over 140  percent for both model

 plants)  arise  from  the installation  of  RACT  equipment.   A maximum annual

 energy savings  of 6,260  GJ  takes  place  in  the model  plant II,  where

 solvent  recovery in the  dryers  is  optimized  by the plant's high

 throughput  without  the additional  solvent  recovery due  to the

 installation of cartridge filters.   Model  plant I shows  an annual  energy

 savings  of  2,460 GJ,  due to the combined effects  of  solvent recovery  on
 energy consumption.

 4.5  REFERENCES FOR CHAPTER 4

 Chevron  Oil Co., 1980.   Sales Brochure:  Chevron  Thinners  and  Solvents.
     EL  Segundo, California.

 Fisher,  W.  1975.   ABC's of Solvent  Mileage,  Part 1.  International
     Fabricare  Institute.  Joliet, Illinois.   IFI Special  Report  Vol. 3,
     No.  4.

 Baumeister, T., E.   Avallone, and T.  Baumeister, III.  1978.  Marks'
     Standard Handbook for Mechanical Engineers.  McGraw-Hill.  New
     York,  New York.

 Plaisance,  S.   1981.  A  Study of Solvent Drainage and Retention in
     Discarded Petroleum Dry Cleaning Cartridge Filter Elements.  TRW
     Inc.,  Research Triangle Park, North Carolina (EPA Contract No.
     68-02-3063).

 Plaisance, S. ,  J.  Jernigan,  G.  May, and C.  Chatlynne.  1981.  TRW Inc.
     Evaluation of petroleum solvent concentrations, emissions, and
     recovery in a solvent recovery dryer (EPA Contract No. 68-02-3063).

 Saary,  Z.  1981.  Chevron Research Laboratory, Telecon with S. Plaisance,
     TRW Inc.,  July 20.   Maximum solubility of Chevron petroleum solvent
     in water.

Vatavuk,  W.   1980.   Factors  for Developing CTG Costs.  Cost and Energy
     Analysis  Section, Economic Analysis Branch.  EPA/OAQPS.  Research
     Triangle  Park,  North Carolina (Draft Document).
                                  4-7

-------
                    5.  CONTROL COST ANALYSIS OF RACT

5.1  BASIS FOR CAPITAL COSTS
     Estimated capital costs of RACT implementation are based on
equipment suppliers'  prices, as well as on EPA cost factors for taxes,
freight, instrumentation, and installation (Vatavuk, 1980).  All cost
estimates are based on June 1980 prices and values.
     Equipment costs are taken from manufacturers or suppliers and
include all major equipment (see Table 5-1 for a summary of existing and
RACT equipment costs).  It is assumed that recovery dryer condenser
cooling water will be supplied by refrigerated chillers, and the costs
of these units are included in the capital costs of RACT equipment.
This represents a worst-case situation, because some facilities might
require no additional water cooling equipment while others might require
only an evaporative cooling tower at an average of 40 percent of the
cost of a refrigerated chiller.
     Taxes, freight,  and instrumentation are lumped together as
18 percent of the equipment costs (Vatavuk, 1980).   While this factor
may be excessive for a simple, unitized device such as a recovery dryer
or cartridge filter,  variations in shipping distance, method of trans-
portation, and local  taxes could increase this portion of the capital
costs beyond this percentage.
     Installation (retrofit) costs are estimated on the basis that all
RACT-associated equipment will be installed by maintenance personnel at
a cost of 5 percent of the equipment cost, or by an outside contractor
at 10 percent of the equipment cost.  The difference between these cost
factors is related to the availability of qualified maintenance personnel
For the purpose of cost estimation, a factor of 7.5 percent of the
equipment cost has been used (Bunyard, 1980) to approximate the costs of
removing existing equipment and replacing it with RACT equipment.

-------
             Table 5-1.  EQUIPMENT COSTS IN TWO MODEL PLANTS
                (Costs in thousands of June 1980 dollars)
     Equipment                  Model plant I       Model plant II

Existing Equipment
  Standard dryer                   15.163               68.00b
  Diatomite filter                  8.4 c                d
RACT Equipment
  Recovery dryer                   44.94e              119.84e
  Cooling tower                     2.18f                4.36f
  Refrigerated chiller              7.009               14.009
  Cartridge filter                  9.0 h                d

aGardner, 1980.
bMoles, 1980.
cKelly, 1980.
 Not applicable.
eMethe, 1980.
fAdams, 1980.
9Chaffee, 1981.
hKirk, 1980.
                                   5-2

-------
5.2  BASIS FOR ANNUALIZED COSTS
     Annualized operating costs are the sum of operating costs and
capital charges.  Operating costs include utilities, operating labor,
and maintenance (labor and materials).  Capital charges include capital
recovery, as well as taxes, insurance, and administration.   Credits for
the value of recovered solvent are included in the total annual
operating costs (Neveril, 1978).
     Primary utilities included under annual operating costs are steam
and electricity.  Annual  steam costs are based on equipment manufacturers'
estimates of steam demand (in boiler horsepower or weight of steam per
hour), estimates of operating hours as a function of model  plant
throughput, and a cost of steam (in dollars per kilogram generated)
derived from current fuel cost estimates (Vatavuk, 1980) (see Table 5-2,
Equation 1).   Electrical  requirements are derived from manufacturers'
electrical demand specifications  (usually in motor horsepower), operating
time estimates as a function of model plant throughput, and a national
average cost of electricity (in dollars per kWh) for commercial customers
(Vatavuk, 1980).  An operating efficiency of 60 percent is  assumed for
electric motors (Neveril, 1978) (see Equation 2) and electrical demand
from refrigerated chillers is assumed to remain constant over the same
duration as that of steam (6 hours per day).
     Operating labor cost estimates are derived from national statistics
for average hourly wages  in the "Wholesale and Retail Trade Category"
with the addition of 56 percent for payroll and plant overhead (Vatavuk,
1980).   A work time of 1  worker-hour of operating labor per dryer per
day (see Equation 3) is assumed (Jernigan and Lutz,  1979).
     Estimated annual  maintenance costs include both labor  and materials.
Maintenance labor costs are calculated from hourly rates that include a
26 percent plant overhead factor, with hours based on field test and
plant survey data (Vatavuk, 1980).   Maintenance materials costs are
determined as 100 percent of annual maintenance labor costs (see Equation
4) in the absence of exact materials cost data (Vatavuk, 1980).  If
materials cost data are available, maintenance costs are represented as
twice the cost of labor or materials, whichever is higher.
                                 5-3

-------
                       Table 5-2.   COST EQUATIONS
Equation 1:   STEAM

     =    boiler y  34.5 Ibs steam/hr y  operating days y  operating hrs y   cost
            hp   A     boiler hp              yr        A      day       A Ib steam

         Steam demand assumed continuous over a 6 hour day.

         Steam cost per kg (pound) = $0.02 ($0.009).

     t    260 Operating days per year.

     t    34.5 Ibs steam per hour per boiler horsepower is a  constant
          of conversion (Babcock, 1978).


Equation 2:   ELECTRICITY

     =   motor  Y  0.746 kW Y         1	 y operating hrs y operating days y cost
          hp    A    hp     A  60% efficiency A     day              yr          kWh

         Assuming lower value of typical motor efficiency range of
          60% to 70%.

         Operating hours per day based on model plant dry-cycle time
          per load and number of loads per day.

         Operating days per year same as steam (see above).

         Electricity cost per kWh = $0.06.

         0.746 kW per horsepower is a constant of conversion (Babcock,
          1978).


Equation 3:   OPERATING LABOR

     =   1 worker-hr operating labor X  operating days ,,  dryers ,, labor cost
                  dryer-day                   yr          plant    worker-hr

         Operating days per year same as steam (see above).

         Labor cost per worker-hour = $8.42.

                                 (continued)
                                       5-4

-------
                            Table 5-2.  Concluded
Equation 4:  ANNUAL MAINTENANCE

=  1 worker-hr   operating days per yr ,,  dryers >, maintenance labor cost w 2.0 materials
   dryer-week     5 days per week          plant          worker-hr          supplement

         Maintenance labor cost per worker-hour = $8.30.

         Maintenance labor and materials costs are equal (based on
          computed costs of hourly labor or required materials, whichever
          is greater).

Equation 5:  TAXES, INSURANCE, AND ADMINISTRATION

          =   (0.04)  X   (total capital cost)
Equation 6:   CAPITAL RECOVERY FACTOR


                           X   (total  capital  cost)
                   + "*
         Assuming interest rate i = 10%.

         Assuming equipment life n = 30 yrs.


Equation 7:   RECOVERED SOLVENT VALUE (CREDIT)

     =   total  emission reduction in  >,   kg throughput  >,  $0.53
         kg VOC/kg throughput                 yr             kg

         Assuming 84% reduction of total  uncontrolled emissions for
          the recovery dryer (15.2 kg VOC per  100 kg dry weight articles
          cleaned).

         Assuming 88% reduction of total  uncontrolled emissions for
          the cartridge filter (7 kg VOC per 100 kg dry weight of
          articles cleaned).

     f    Average solvent cost = $0.53/kg ($1.55 gallon).
                                       5-5

-------
     Capital costs resulting from property taxes and insurance costs are
each estimated as 1 percent of the total  annual  capital  costs, and
administration costs are 2 percent of total  annual  capital  costs (see
Equation 5) (Neveril, 1978).   The capital  recovery  cost  is  based on an
annual  interest rate of 10 percent (Vatavuk, 1980)  and a projected life
of 30 years (see Equation 6) for the dry cleaning equipment.   In the
case of existing equipment, it is assumed that the  equipment has been
paid off.   Therefore, there are no capital recovery charges for existing
equipment.
     Annual credits for recovered solvent are based on experimentally
determined solvent recovery efficiencies for RACT equipment.  For example,
the emission reduction resulting from the installation of RACT equipment
in model plant I with a diatomite filter is the difference between the
existing emission rate of 30 kilograms of VOC emitted per 100 kilograms
dry weight of articles cleaned and the RACT equipment emission level of
5.8 kilograms of VOC per 100 kilograms dry weight of articles cleaned
with a cartridge filter.  Furthermore, it is assumed that a specified
percentage of the 30 kilograms of VOC (solvent) per 100 kilograms dry
weight of articles cleaned is recovered in a reusable form (Jernigan and
Lutz, 1979).  Solvent costs are taken from industry quotations, the
average being $1.55 per gallon (see Equation 7) (Carson, 1980).  The
difference from existing control costs is computed as the difference in
total annual operating costs between existing and RACT equipment.
5.3  EMISSION CONTROL COSTS
     The costs of RACT implementation in  the two model plants are based
on the  installation  and operation of recovery dryers  (and their asso-
ciated  refrigerated  chillers) and a cartridge filtration system where  a
diatomite  filter  is  in use.  Tables 5-3 and 5-4 summarize the results  of
applying the previously defined cost equations to two model plants  with
both existing and RACT equipment, based on  the worst-case assumption of
an existing diatomite filter  in model plant I.  These tables  also  show
the cost of RACT  implementation with existing cartridge filters and
settling tanks.
                                  5-6

-------
Table 5-3.    CAPITAL AND  ANNUALIZED  COSTS OF  EXISTING  EQUIPMENT  IN TWO MODEL PLANTS
                         (Costs are in  thousands  of  June  1980  dollars)

                                                              Model plant  Ia      Model  plant IID
                Cost parameters                                (182,000 kg/yr)c    (635,000 kg/yr)c
Capital costs
Equipment
Taxes, freight, and instrumentation
Direct and indirect installation
Total capital costs
Annual ized costs
Operating costs
Steam
Electricity
Operating labor
Annual maintenance (labor and materials)
Subtotal, direct costs
Capital charges
Capital recovery
Administration taxes, and insurance
Subtotal, indirect costs
Recovered solvent value (credit)
Total annual ized cost
Total annual ized cost with existing cartridge

31.96
5.75
2.82
40.53


11.65
0.60
9.15
6.67
28.16

0
1.62
1.62
0
29.76
20.38

68.00
12.24
6.02
86.26


35.92
7.20
5.00
2.00
50.12

0
3.45
3.45
0
53.57
d
    filters instead of diatomite filters
  Total annualized cost with existing  settling                        21.86
    tank instead of diatomite filter
   Tabulated costs for model plant I are based on  an existing diatomite filter.
   Tabulated costs for model plant II  are based on an existing settling tank.
   Annual  kilograms of articles cleaned.
   Not applicable.

-------
                       Table  5-4.   CAPITAL AND ANNUALIZED COSTS OF  RACT  EQUIPMENT IN  TWO MODEL PLANTS
                                           (Costs are in  thousands of  June  1980 dollars)
en
i
00

Cost parameters
Capital costs
Equipment
Taxes, freight, and instrumentation
Direct and indirect installation
Total capital costs
Annual i zed costs
Operating costs
Steam
Electricity
Operating labor
Annual maintenance (labor and materials)
Subtotal, direct costs
Capital charges
Capital recovery
Administration taxes, and insurance
Subtotal, indirect costs
Recovered solvent value
Total annual ized cost
Difference from existing equipment annual costs
Total annualized cost with existing cartridge filters
Difference from existing equipment annual costs
Total annualized cost with existing settling tank
Difference from existing equipment annual costs
Model plant Ia
(182,000 kg/yr)c

60.94
10.97
5.40
77.31


3.16
1.52
6.40
7.23
18.31

8.20
3.10
11.30
(20.74)
8.87
(20.89)
9.16
(11.22)
9.15
(12.71)
Model plant IIb
(635,000 kg/yr)c

133.84
24.09
11.85
169.78


10.10
5.08
20.20
7.96
43.34

18.01
6.79
24.80
(48.80)
19.34
(34.23)
d
d
d
d
                       tabulated costs for model plant I are based on replacing an existing diatomite filter with
                        a cartridge filter.

                        Tabulated costs for model plant II are based on an existing settling tank.

                       cAnnual  kilograms of articles cleaned.
                       d
                        Not applicable.

-------
     While installed capital costs of RACT equipment range from $77,000
to $170,000, the cost credits gained from the value of recovered solvent
result in total annualized costs ranging from $8,870 (model plant I with
an existing diatomite filter) to $19,340 (model plant II).
     The installation of RACT equipment in a model I plant is based on
the operation of three 48 kg (105 Ib) capacity recovery dryers and
(where applicable) a cartridge filtration system, instead of three
standard dryers of the same capacity with diatomite filters.   While
operating costs are reduced by the recovery dryer's lower demand for
steam and the cartridge filter's reduced materials and labor outlays,
gross annual capital charges are substantial due to the effects of high
capital costs.   Total annualized costs for model plant I with existing
diatomite filters are reduced by 70 percent as a result of credits for
recovered solvent, and model I plants with both existing cartridge
filters and settling tanks have annualized cost reductions of about
60 percent.
     Two 180 kg (400 Ib) capacity standard dryers used in model II
plants are replaced by eight RACT recovery dryers at a capital cost of
approximately $170,000.   The annual  cost of steam is significantly lower
for recovery dryers,  but their labor and maintenance costs are higher
than those of existing equipment, with eight dryers requiring more
operator and maintenance time.   High annual capital charges associated
with RACT equipment are offset by substantial  cost reductions from
recovered solvent credits.   Thus, the installation of RACT equipment
(recovery dryers) in a model plant II yields a 64 percent reduction in
total annualized costs when compared with standard dryer annualized
costs.
5.4  COST EFFECTIVENESS
     The cost effectiveness of RACT equipment installation and operation
is defined as the annual dollars expended beyond the costs of existing
equipment per unit mass of emission reduction achieved.   A combination
of high emission reduction and low annual cost (high annual credit)
results in maximum cost effectiveness.
                                 5-9

-------
     The cost effectiveness of existing and RACT equipment in two model
plants is tabulated in Table 5-5.   The actual  annualized cost of recovery
dryers ranges from about $9,000 per year in a  model  plant I (three
recovery dryers with refrigerated chillers) to about $19,000 per year in
a model plant II(eight recovery dryers with refrigerated chillers).
Based on standard dryer annualized operating costs of about $21,900 per
year in a model plant I (three 45 kg capacity  dryers) and about $53,600
per year in a model plant II (two 225 kg capacity dryers), the difference
in annualized costs resulting from recovery dryer operation in model I
and II plants would be savings of about $12,700 and $34,000, respectively.
Similarly, standard dryer VOC emissions in model I and II plants of
about 32.8 and 114.3 megagrams VOC per year would be reduced by about
26.4 and 92 megagrams VOC per year, respectively, based on the recovery
dryer emissions of 6.4 and 22 megagrams VOC per year in model I and II
plants.  Thus, the cost effectiveness of recovery dryers in model I and
II plants would be a savings of $480 and $370, respectively, per megagram
of VOC emission reduction.
     The replacement of diatomite filters with cartridge filters in a
model plant I  (model II plants use only settling tanks) would result in
a reduction in annualized costs of about $8,200, based on annualized
operating costs of about $7,900 per year for diatomite filters (two
11,400 L/hr filters) and an actual savings in annualized costs of about
$300 per year  for cartridge filters (63 cartridge elements).  Similarly,
the 14.6 megagram annual VOC emisions from diatomite filters would be
reduced to about 1.9 megagrams VOC per year with the installation of
cartridge filters, resulting  in a filtration emission reduction of about
12.7 megagrams VOC per year.  Thus, the cost effectiveness  of replacing
existing diatomite filters with cartridge  filters in a model I plant
would  be a savings of about $640 per megagram of VOC emission reduction.
     The cost  effectiveness of RACT implementation in two model plants
is summarized  in Table 5-6, which also  includes cost effectiveness data
for facilities with existing  cartridge  filters  as well as those with
existing solvent settling tanks.
                                  5-10

-------
               Table  5-5.   COST  EFFECTIVENESS OF  EXISTING AND RACT EQUIPMENT IN  TWO  MODEL PLANTS
                                     (costs  are  in thousands of June 1980 dollars)

Emission reduction
Equipment (Mg VOC/yr)
Standard dryer 0
Recovery dryer (with
refrigerated chiller) 26.39
Diatomite filter 0
Cartridge filter 12.74
Model plant I
(182,000 Kg/yr)
Annual ized cost
($103/yr)
21.86
(12.71)a
7.92
(8.19)a

Cost effectiveness
($103/Mg VOC)
c
(0.48)b
c
(0.64)b
Model plant II
(635,000 Kg/yr)
Emission reduction Annual ized cost Cost effectiveness
(Mg VOC/yr) ($103/yr) ($103/Mg VOC)
0 53.57 c
92.08 (34.23)a (0.37)b
c c c
c c c
 Annualized cost values tabulated for control equipment (recovery dryers  or cartridge  filters) are  actually the  difference between
 existing equipment (standard dryer or diatomite  filter) annualized cost  and the actual annualized  cost for the  particular item of
 control equipment.

 Numbers in parenthesis represent thousands of dollars saved.

cNot applicable.

-------
   Table 5-6.   COST EFFECTIVENESS OF RACT IMPLEMENTATION  IN  TWO  MODEL PLANTS
               (costs are in thousands  of June 1980 dollars)



                                    Model plant Ia     Model plant IIb
Cost parameters                    (182,000 kg/yr)c    (635,000  kg/yr)L


Emission reduction                           39.13           92.08
  (Mg VOC/yr)

Annualized RACT cost savings
  relative to existing equipment             20.89           34.23

Cost effectiveness of RACT
  implementation
  (dollars saved/Mg of VOC
  emission reduction)                         0.53           0.37

Cost effectiveness of RACT
  implementation in plants with
  existing cartridge filters                  0.43               d
  in $ saved/Mg VOC

Cost effectiveness of RACT
  implementation in plants with
  existing settling tank                      0.48               d
  in $ saved/Mg VOC


tabulated cost effectiveness for a model I plant is based on an existing
 diatomite filter.

tabulated cost effectiveness for a model II plant is based on an existing
 settling tank.

Annual kilograms of articles cleaned.

 Not applicable.
                                  5-12

-------
      The  replacement of  existing equipment (standard  dryer and  diatomite
 filter) with  RACT  equipment  in  a model  I  plant  would  result in  an  annual
 cost effectiveness of $530 saved per  megagram of  emission  reduction.
 This cost effectiveness  includes the  burden  of  high capital  costs  resulting
 from the  installation of three  recovery dryers  and one  cartridge filter.
 The  owner of  a model  I plant  could  expect to save about $21,000 per year
 by replacing  existing equipment  with  RACT equipment,  while reducing the
 annual VOC emissions  from the plant by  approximately  40 megagrams.  In
 model  I plants with  existing  cartridge  filters, the annual  cost
 effectiveness of RACT implementation  would be $430 saved per megagram,
 and  that  of model  I  plants with  existing  settling tanks would be $480
 saved  per megagram of VOC emission reduction.
     The  cost effectiveness of replacing  existing equipment with RACT
 equipment in a model  II  plant yields  an annual  savings  of  $370 per
 megagram  of VOC emission  reduction.   Unlike the model I  plant, the model
 II facility has only  a settling  tank  for  filtration and  there are no
 additional savings from  solvent  recovery  by installation of cartridge
 filters.    Therefore,  the owner of a large industrial  plant could expect
 to save $34,000 per year after installing RACT recovery  dryers,  while
 experiencing a 92  megagram reduction  in annual  VOC emissions.
 5.5  REFERENCES FOR CHAPTER 5
 Adams, E.   1980.    Adams Chet Co., Telecon with S.  Plaisance, TRW Inc.,
     June  13.   Cost of 25 ton cooling tower.
 Babcock and Wilcox.  1978.  Steam:  Its Generation and Use.  Babcock and
     Wilcox.  New  York, N.Y.
 Bunyard,   F.  1980.   EPA/EAB,  Telecon with S.  Plaisance,  TRW Inc.,
     November 24.   Cost factor for equipment installation.
 Carson, J.  1980.   Ashland Chemical  Company,   Telecon with S.  Plaisance,
     TRW  Inc., September 15.   Price of Ashland  petroleum solvent.
Chaffee,  T.  1981.   Rite-Temp, Inc.,  Telecon with S.  Plaisance,  TRW
     Inc.,  January 9.  Cost  of refrigerated chillers.
Gardner,  J.  1980.   Gardner  Machinery, Inc.,   Telecon with S.  Plaisance,
     TRW Inc., June 9.  Cost  of 50 and 110 pound capacity Cissell  dryers.
                                 5-13

-------
Jernigan, R. and S.  Lutz.   1979.   Evaluation of the Emission Reduction
     Potential of a Solvent Recovery Dry Cleaning Dryer.   TRW Inc.
     Research Triangle Park, North Carolina (EPA Contract
     No.  68-03-2560, Task No.  T5013).

Kelly,  R.  1980.   ChenrSan International, Inc.,  Telecon with S. Plaisance,
     TRW Inc., October 17.  Cost of 1,500 gph Chem-San diatomite filter.

Kirk, R.   1980.   Boggs Equipment Inc. ,   Telecon with S.  Plaisance, TRW
     Inc., October 17.  Cost of Puritan cartridge filter systems.

Methe,  A.  1980.   Hoyt Manufacturing Inc.,  Telecon with S. Plaisance,
     TRW Inc., June 11.   Cost of 50 and 105 pound capacity Hoyt recovery
     dryers.

Moles,  E.  1980.   Challenge-Cook Brothers, Inc.,  Telecon with S. Plaisance,
     TRW Inc., June 16.   Cost of 400 pound capacity dryer.

Neveril,  R.   1978.   Capital and Operating Costs of Selected Air
     Pollution Control Systems.   EPA Publication No. 450/5-80-002.
     EPA/OAQPS.   Research Triangle Park, North Carolina.

Vatavuk,  W.   1980.   Factors for Developing CTG Costs.   Cost and Energy
     Analysis Section, Economic Analysis Branch (EAB), OAQPS.  Research
     Triangle Park, North Carolina (Draft Report).
                                 5-14

-------
      APPENDIX A
SUMMARY OF FIELD TESTS

-------
                                APPENDIX A
                          SUMMARY OF FIELD TESTS

     This appendix provides detailed descriptions of EPA tests conducted
in support of petroleum dry cleaning new source performance standard
(NSPS) and control techniques guideline (CTG) development.
A.I  TEST 1 (PICO RIVERA)
     EPA contracted an engineering analysis of a solvent recovery dryer
to determine its emission reduction potential and establish the capital
and operating costs associated with its use (Lutz and Jernigan, 1980).
The test site for this program was an industrial petroleum dry cleaning
plant located in Pico Rivera, California.   Testing was conducted at the
plant from October 9 to November 21, 1979.   This dry cleaning facility
utilized both a Cissell standard dryer and a Hoyt "Petro-miser" solvent
recovery dryer to process approximately 6,350 kg (14,000 Ibs) of industrial
work gloves per week.  The standard dryer had a dry weight load capacity
of 45 kg (100 Ibs), and the recovery dryer had a dry weight load capacity
of 48 kg (105 Ibs).  To reflect normal operating conditions, each dryer
was loaded an average of 10 percent over its rated capacity with work
gloves made of cotton and leather.   The recovery dryer had three operating
sequences - a Reclaim-Dry Cycle, a Perma-Cool Cycle, and a Deodorizing
(exhaust) Cycle.  Solvent emissions from the recovery dryer were not
restricted solely to the exhaust cycle.   Any time the dryer loading door
was open, an exhaust fan was activated and ambient room air was pulled
into the dryer and exhausted to the atmosphere via the exhaust duct.
This also occurred when the door to the lint filter compartment was
open.   In comparison, the standard dryer continuously exhausted to the
atmosphere during the dryer cycle.

-------
     The solvent recovery dryer's  emission reduction performance  was
established by comparing its measured emission rate with the emission
rate of a standard dryer.   During  the testing period,  both dryers were
operated simultaneously and processed similar loads.    The average flow
rate through the exhaust ducts during the recovery and standard dryer
drying cycles was determined using EPA Method 2.   The average solvent
concentration in the recovery and  standard dryer exhausts was determined
during each exhaust cycle by analyzing the strip chart recordings from a
Beckman 400 flame ionization analyzer (FIA).   The average concentration
for each dryer, multiplied by the  total gas volume throughput for each
dryer, yielded the total solvent emitted for each dryer in kilograms per
cycle.  Dividing this value by the weight of gloves dried for each dryer
cycle yielded the solvent emissions for the recovery and standard dryers
expressed  in  kilograms  of solvent per 100 kilograms of gloves dried.
     The average emission rates for the recovery and standard dryers
were determined and expressed in  kilograms of solvent per 100 kilograms
of  articles cleaned.  Table A-l indicates that the  recovery  dryer had  an
average  emission  rate per drying  cycle of 0.96 kg  solvent per 100 kg  dry
weight of  articles  cleaned,  and the  standard  dryer had  an average emission
rate  per drying  cycle  of 30  kg solvent per 100  kg  dry weight of  articles
cleaned.   Recovery  dryer solvent  emissions per  100 kilograms of  articles
cleaned ranged from 0.68 to 1.25  kilograms,  and  appeared  to vary with
the load weight.   The  total  weight  of solvent recovered ranged from  8 to
17  kilograms  (18 to 37 pounds) and  did not appear to be a direct function
of  the load weight.   In contrast, standard dryer solvent emissions  per
 100 kilograms of articles cleaned ranged from 20.8 to 47.2 kilograms,
 and appeared to increase with smaller load weights.   The trend toward
 higher emissions per weight of articles  cleaned in smaller loads held
 true for both standard and recovery dryers,  and indicated that the rate
 at which fabrics release solvent  could have  a significant effect on
 overall solvent emissions.
      The annualized operating cost of the recovery dryer was calculated
 to be $1,400, which represents a savings of $3,900 per year over the
 operating cost of the  standard dryer.  This savings was due primarily to
 the value of the recovered solvent, estimated at $0.24/liter ($0.92/gallon).
                                  A-2

-------
                                        Table A-l.   DRYER  EMISSIONS  DATA
Recovery dryer
Date
10/09/79
10/16/79
10/17/79
10/17/79
10/17/79
10/17/79
10/17/79
10/18/79
10/18/79
p, 10/18/79
I
<-*> 10/18/79
10/22/79
10/22/79
10/22/79


Average
Load
#
2
2
2
3
4
5
5
2
3
4
5
2
3
4



Load
dry
weight
(kg)
51.9
50.8
49.0
51.4
53.7
43.3
53.3
56.2
51.7
49.8
49.2
52.1
46.2
48.9



Solvent kg solvent
emitted emitted/100 kg
(kg) articles cleaned
0.352
0.496
0.598
0.462
0.486
0.471
0.471
0.424
0.407
0.611
0.569
0.448
0.577
0.471



0.68
0.98
1.22
0.90
0.91
0.88
0.88
0.76
0.79
1.23
1.16
0.86
1.25
0.96


0.96
Solvent
recovered
(kg)
14.85
11.11
11.34
11.56
9.75
11.11
11.11
8.73
11.68
14.06
8.16
11.45
16.78
14.29



kg solvent
recovered/100 kg
articles cleaned
28.6
21.9
23.1
22.5
17.3
20.1
20.1
15.5
22.6
28.2
16.7
22.0
36.3
29.2


23.4
Date
10/10/79
10/11/79
10/11/79
10/11/79
10/11/79
10/12/79
10/12/79
10/12/79
10/12/79
10/12/79
10/15/79
10/15/79
10/15/79
10/16/79
10/16/79
10/17/79
10/17/79
10/17/79
10/17/79
10/19/79
10/19/79

Run
4
1
2
3
5
1
2
3
4
5
3
4
5
2
3
1
2
3
4
1
4

Standard Dryer
Load
weight
(kg)
50.79
53.52
51.93
49.43
52.38
50.57
48.98
51.02
54.88
48.66
52.38
50.93
46.49
50.79
53.29
52.38
51.02
52.15
53.65
53.51
48.89

Solvent kg solvent
emitted emitted/100 kg
(kg) articles cleaned
15.80
13.99
14.56
17.80
16.57
14.88
16.57
15.65
11.40
18.72
16.11
15.41
21.94
15.80
13.21
13.56
16.88
12.68
14.00
12.12
14.64

31.1
26.1
28.0
36.0
31.6
29.4
33.8
30.7
20.8
38.5
30.8
32.2
47.2
31.1
24.8
25.9
33.1
24.3
26.1
22.7
29.9
30.1
Data tabulated only for dryer loads in which all relevant parameters were successfully monitored.

-------
     The mass balance and hydrocarbon analysis from this test program
demonstrated that recovery dryers could achieve a 97 percent reduction
in solvent emissions as compared with a standard dryer.   The economic
analysis of this type of control system indicates that it is a cost-
effective means of solvent emission control, providing an actual
reduction in operating costs.
     One problem that was not resolved during this test was whether the
recovery dryer operated above the lower explosive limit (LEL) of the
solvent (1 percent by volume or 10,000 parts per million).   FIA chart
recordings of the vapor concentrations in the recovery dryer during the
reclaim-dry cycle indicated that the vapor concentration rose until it
peaked at 9,000 to 9,300 parts per million (ppm) as solvent.  The vapor
concentration remained at this peak throughout most of the drying cycle.
After the testing was completed, careful analysis of the chart recordings
revealed that these peak readings were not the maximum concentration
levels, but the level at which the FIA became saturated; thus, indicating
only the maximum monitoring levels of the calibrated FIA.   Therefore,
the actual concentrations of the solvent vapors in this particular
recovery dryer may have exceeded the peak range of 9,000 to 9,300 ppm.
     The high vapor concentrations during the reclaim cycle may be
attributed to a number of factors.   Overloading of the dryers., as was
the case during this test, may have caused the high concentrations.
Fabric with high solvent absorption, such as cotton and leather, give
off more solvent vapors than an equal weight of synthetic fabrics,
thereby creating higher concentrations.  Also, high condenser inlet
water temperatures may contribute to high vapor concentrations during
the reclaim cycle.
A.2  TEST 2 (LAKELAND)
     An EPA-sponsored testing program was performed at a commercial
petroleum dry cleaning facility to investigate the solvent emissions and
recovery, operational costs, and safety of petroleum solvent recovery
dryers (Jernigan et al., 1981).  The host plant for this test program
was a large commercial dry cleaning plant located in Lakeland, Florida.
This facility cleaned 1,100 kg (2,500 Ibs) of general apparel each week
                                 A-4

-------
The dry cleaning equipment consisted of a 48 kg dry weight (105 Ib)
capacity Hoyt recovery dryer, a 30 kg (65 Ib) capacity Washex washer/
extractor, an 11,000 liter per hour (3,000 gph) Washex tube filter, and
a 48,000 Btu (50 MJ) Rite Temp refrigerated water chiller.  Testing was
conducted at this facility from July 21 to August 8, 1980.
     Test procedures included monitoring exhaust gas and condenser gas
inlet solvent concentrations using a Beckman 400 flame ionization analyzer
(FIA).  Also, temperatures of condenser inlet and outlet (water and gas)
and dryer exhaust gas were monitored during this program.   Chiller
outlet (condenser water inlet) temperatures were increased in 5F
increments with a constant reclaim cycle duration (28 minutes), and
solvent recovery rate and concentration data were recorded for several
clothing loads at each of the chiller temperatures.   The total recovery
of both solvent and water, as well as the total flow of cooling water
through the condenser during the reclaim cycle, were recorded and are
listed on Table A-2.
     The mass balance and hydrocarbon analysis results from this test
program indicated that the average VOC emissions rate from the recovery
dryer was 3.85 kg VOC per 100 kg dry weight of articles cleaned.   The
solvent concentration at the condenser gas inlet never exceeded 95 percent
of the solvent's lower explosive limit (LEL) during the portion of the
test in which the condenser water inlet temperature was varied.
     Data collected during the test is summarized in Table A-2.  As
condenser water inlet temperatures were increased, condenser vapor
outlet temperatures increased and solvent emissions per 100 kg of
articles cleaned decreased.  Uncontrolled theoretical solvent emissions
(defined as the sum of recovered and emitted solvent) per 100 kg of
articles cleaned varied from 23.73 kg to 11.59 kg, with an overall test
average of 14.24 kg per 100 kg of articles cleaned.   Recovery dryer
emissions per 100 kg of articles cleaned, measured at the dryer exhaust
by the FIA, varied from 9.45 kg to 2.34 kg, with an overall test average
of 3.85 kg solvent emitted per 100 kg of articles cleaned.  This
relatively high emission rate may result from the typically small load
weights (25 kg average) of synthetic fabrics that have a low solvent
                                 A-5

-------
                                  Table A-2.  RECOVERY DRYER  DATA  COMPILATION
CT)

Start
time
0829
0905
0941
1034
1113
1154
1240
1321
0719
0812
0856
0941
1020
1058
0727
0815
0857
0939
1028
1109
1152
1233
0706
0754
0834
0924
1009
1048
Averages
Date
and
run
6/4-2
3
4
5
6
7
8
9
8/5-2
3
4
5
6
7
8/6-1
2
3
4
5
6
7
8
8/7-1
2
3
4
5
6

Condenser
inlet
water
temp (F)
64
64
64
66
65
66
65
65
63
66
68
68
68
68
64
68
70
71
70
70
72
72
70
73
73
75
74
74

Average
condenser
vapor outlet
temp (F)
87
--
89
89
89
91
89
89
88
91
92
92
92
92
87
91
94
94
93
94
95
94
94
95
96
97
95
95

"Dry"
load
weight
(kg)
31.29
29.93
26.30
23.13
17.23
24.94
39.00
9.98
19.95
27.66
30.39
29.02
34.01
29.93
23.13
29.48
29.02
26.30
22.68
26.76
24.04
20.86
27.66
20.86
28.12
31.29

24.04
25.10
Recovery (kg)
Water
0.010
0.155
0.350
0.2&5
0.245
0.440
0.599
0.220
0.315
0.420
0.499
0.335
0.499
0.450
0.165
0.370
0.499
0.450
0.360
0.467
0.375
0.385

0.240
0.370
0.467

0.370

Solvent
3.21
3.19
2.59
2.18
1.52
2.36
4.18
1.43
2.25
2.91
3.19
3.13
3.60
3.22
2.06
3.34
3.07
2.83
2.25
2.44
2.63
2.15
--
1.97
2.93
3.26

2.74

kg solvent
Solvent emitted/100 kg
emitted articles
(kg) cleaned
1.16
1.23
1.26
1.14
1.02
1.07
1.27
0.94
1.04
1.14
1.22
1.34
1.42
1.31
0.72
0.89
0.88
0.78
0.73
0.67
0.63
0.65
0.71
0.67
0.79
0.79
0.57a
0.56
0.96
3.71
4.12
4.78
4.92
5.95
4.27
3.24
9.45
5.20
4.13
4.00
4.63
4.19
4.38
3.12
3.03
3.02
2.97
3.22
2.49
2.60
3.11
2.56
3.22
2.79
2.51

2.34
3.85
kg solvent
recovered/
100 kg
articles
cleaned
10.26
10.66
9.85
9.42
8.82
9.46
10.72
14.32
11.28
10.52
10.50
10.79
10.59
10.76
8.90
11.32
10.59
10.76
9.92
9.12
10.94
10.31

9.44
10.42
10.42

11.40
10.44
     Data not used in computing averages.

-------
retention.   Also,  the typical  recovery phase duration of 28 minutes may
be insufficient time for a more complete recovery.
     Figures A-l,  A-2, and A-3 represent a typical  range of recovery
dryer loads and emissions.   A dryer load with relatively high emissions
(approximately 5.2 kg solvent emitted per 100 kg of articles cleaned) is
illustrated in Figure A-l.   The gradual  increase in the volume of solvent
is reflected in the narrow peak of the solvent recovery rate.  Simultaneously,
the concentration of solvent vapor in the condenser inlet climbs steadily
during the first 7 minutes of recovery and then levels off at a near
constant concentration of 4,400 parts per million (ppm) as solvent.  The
curve, representing the volume of recovered solvent, has a brief initial
period of rapid recovery that is followed by a gradual increase in total
volume that reflects a low, nearly constant rate of recovery.
     Recovery and concentration curves in Figure A-2 illustrate a dryer
load in which solvent emissions were below those of the test average
(approximately 2.4 kg solvent emitted per 100 kg of articles cleaned).
The graph of condenser inlet vapor concentration shows a much higher
(9,358 ppm as solvent) and more pronounced peak than Figure A-l, in
addition to much higher concentrations throughout the entire cycle.
Concurrently, the curve illustrating the volume of recovered solvent
shows an initial period of very rapid recovery that gradually decreases
to a lower, near constant rate later in the drying cycle than in the
high-emission load.  Finally, the curve representing the solvent recovery
rate shows a more gradual decrease in recovery rate than that illustrated
in Figure A-l, although the peak rate is approximately the same.  At the
termination of the recovery cycle, the rate of solvent recovery had
decreased to a final value of 20 milliliters (0.02 liters) per minute.
The average condenser gas outlet temperature for this dryer  load did not
exceed 34C (94F).
     Figure A-3 illustrates a dryer  load that had total solvent emissions
(3.71 kg solvent per 100 kg articles cleaned) approximately  equal  to the
overall test average of 3.85  kg solvent per 100  kg articles  cleaned.
The curve representing the condenser inlet vapor concentration  shows
somewhat more  of a peak than  that of Figure A-l, but  a much  less
pronounced  and  lower  (5,810 ppm as solvent) peak than that  of Figure A-2.
                                  A-7

-------
   o
CO
   O
   O
                                         Recovered Solvent Volume (ml)

                                         Concentration as ppm Solvent

                                         Rate of Solvent Recovery, ml /ml n
                                ~~r=*  \
                                                                                          -i 10000
- 9000
        a*
                                                                                           8000
                                                                                          : 7000
                                                                                          -  6000
                                                                                            5000
                                                                                            4000
                                                                                           3000
                                                                                            2000-
        at
        tn
        c.
        V
        T>


        8
O
4/1


2E
a.


t


o
i




&
Ul
^

o
o
              0    2     4     6    8    10   12   14   16   18    20   22   24   26   28   30



                                     RECLAIM CYCLE TIME  (min)


               Figure A-1.  Recovery and Concentration Curves for a High-Emission  Recovery  Dryer
                                                                                                         600
                                                                                                         500
                                                                                                         400
                                                                                                         300
                                                                                                             >-
                                                                                                             QC
               200
                                                                                                              o
                                                                                                              */
                    o
                    ui


               100  2
        Load,

-------
                                       6-V
                                RECOVERED SOLVENT  VOLUME (ml)
I
ro
XI
fD
n
o
Qi

CX
    O   i
O   11
3   2
n
CD   o
3   -<
<-   O
-   I
a   m
o   -i
3   3C
    m
o
o
T

Ol
O
3
 (D
 O
 O
 <
 n>
m
o
CD
Q.
         CO
         O
ro
O
0
O

CO
0
o
o

^
o
o
o

en
o
o
o


0
o
o

^J
o
o
0

00 IO
GJ CD
C3 ^^
o o

1 >
0
o
o
0
CONCENTRATION AS PPM  SOLVENT (Condenser  Inlet)
                      o
                      o
             ro
             o
             O
OJ
O
o
o
o
01
o
o
O
O
                      RATE OF SOLVENT RECOVERY (ml/min)

-------
                                      oi-v


                                RECOVERED SOLVENT VOLUME  (ml)
                g
                 ro    o*

                 8    8
                                            N>   fS>
  8
8
                                                                             00
8    8
re
CO
o

re
o
o

o
re
 o


 o


 I
 re
 o
 -t
 73
 re
 o
 o

 re
 re
 -i
         ro
         oo
    73
    m
    r>
    o
          oo
          ro
          O
ro
ro
          ro
ro
cr>
          ro
          oo
 re
 re

 CM
(3
 re
          OJ
          o
            ro
            o
            o
            o
           o
           o
8
o
                                                             to
                                                             o
                                                                      o
                                                                      o
           CONCENTRATION AS  PPM SOLVENT  (Condenser  Inlet)
 (A
 o
 3
 (/i
                       o
                       o
                                 ro
                                                       in

                                                       8
                       RATE  OF SOLVENT  RECOVERY (ml/m1n)

-------
 Simultaneously, the curve  illustrating the volume of  solvent recovered
 (Figure A-3) shows a rapid, sustained increase in total reclamation.
 This  rate  is further illustrated in the graph of the  recovery rate which
 stays at a higher rate over the total cycle due to the heavier weight of
 this  dryer load.
      The overall decrease  in solvent emissions per weight of articles
 cleaned that accompanied the increase in condenser water inlet temperature
 appeared to result from the higher solvent vapor concentrations associated
 with  higher temperatures in the dryer tumbler.  The rate of solvent
 condensation in the condenser appeared more dependent on the rate at
 which solvent was evolved  from the drying articles than on the temperature
 of the cooling water.   The 10 actual increase in the condenser water
 inlet temperature did not  appreciably hamper solvent condensation.
 Thus, emission reduction might be enhanced by increasing the temperature
 within the dryer; either by increasing the pressure (temperature) of the
 steam, or by increasing the cooling water temperature.  This drying
 temperature increase,  however,  could result in solvent concentrations in
 the tumbler reaching the solvent LEL of 1 percent by volume (10,000 ppmv).
     One of the objectives of the test was to determine a "uniform
 dryness," a level of solvent concentration that would indicate sufficient
 dryness and solvent recovery.   This goal  was not attained because of
 difficulties encountered in the test contractor's equipment and the lack
 of control  over the weight and  fabric composition of drying loads.
Also, the determination of the  solvent content of the dried articles was
 hampered by the limited accuracy of the plant scales used to weigh the
washed and dried loads.
A.3  TEST 3 (RHODE ISLAND)
     This EPA-sponsored test program was  initiated with the overall
 objective of analyzing the performance of the petroleum solvent recovery
dryer, as indicated by the maximum solvent concentration,  solvent emis-
 sions, and solvent recovery, while dryer  operating parameters were
varied.   Furthermore,  the overall  reduction in plant solvent consumption
 (solvent mileage) was  to be determined.
                                 A-11

-------
     A Hoyt Petro-Miser 105 solvent recovery dryer was  tested  for  two
weeks at a dry cleaning plant in West Warwick,  Rhode Island that cleans
about 2,700 kg (6,000 Ib) of personal apparel per week.   During the
test, solvent concentrations in the dryer tumbler and exhaust  were
measured to determine, respectively, the maximum solvent concentration
during drying and the mass of solvent emitted.   Additional  measurements
of the volume and rate of solvent recovery were made, and load weights
and relevant temperatures were recorded.  Parameters relating  to the
dryer operation (load weight, reclaim duration, fabrics, temperatures)
were varied and the effects of these variations on emissions,  recovery,
and concentrations were noted.  Finally, data on plant solvent consumption
prior to recovery dryer installation were obtained from plant  management.
     Analysis of the data collected indicated that the magnitude of the
maximum solvent concentrations was below expected levels, based on
results of previous tests.  These low concentrations resulted  from
persistent difficulties with solvent condensation in the concentration
sampling system, brought about by low ambient temperatures. While data
on the absolute magnitude of dryer concentrations may have been of
questionable value (the maximum value recorded was 3,537 ppmv  as solvent)
the relative variations in concentrations among dryer loads was found to
be valid and consistent with variations in dryer operating parameters.
Thus, it was found that the dry load weight and drying (condenser vapor
inlet) temperature had the greatest impact on the relative level of
solvent concentrations in the dryer.
     An analysis of the effects of variations in dryer operating parameters
on solvent recovery and emissions indicated that dry load weight and
condenser heat removal had the greatest effect on dryer performance,
with increases in both parameters corresponding to both higher solvent
recovery and reduced solvent emissions.  Over the entire test  program,
solvent emissions averaged 3.47 kg per 100 kg dry weight of articles
dried, and solvent recovery averaged 12.98 kg per 100 kg dry weight of
articles dried.  Plant solvent mileage, as reported by plant management,
decreased from about 560  liters (150 gal) per week to about 90 liters
(25 gal) per week after  instion of the  two recovery dryers.
                                 A-12

-------
      Table  A-3  contains  the  data  collected  during  the  test  program.
 Solvent  emissions  ranged from  1.2 to  7.6  kg VOC  per  100  kg  dry  weight  of
 articles dried,  while  solvent  recovery  ranged  from 9.9 to 17.7  kg  solvent
 recovered per 100  kg dry weight of articles dried.
      Graphs  of  recovery  dryer  performance in dryer loads with low,  high,
 and  test average emissions are plotted  in Figures  A-4, -5,  and  -6.  The
 load (Figure A-4)  with the lowest emissions per  dry  load weight (1.2 kg
 VOC/100  kg  articles dried) has a  pronounced "peak" in  the solvent  concen-
 tration  after about 10 minutes of recovery  that  corresponds with the
 onset of the maximum recovery  rate.   At the same time, the  curve representing
 the  volume  of recovered  solvent shows a consistent increase in  the  total
 volume that decreases significantly only during  the  last 10 minutes of
 recovery.   In contrast,  the  high-emission dryer  load (7.6 kg VOC
 emitted/100 kg articles  dried) represented  in Figure A-5 shows  a more
 consistent tumbler solvent concentration with a  narrower peak,  a lower
 maximum  recovery rate over a shorter period, and a slowly increasing
 total volume of  recovered solvent  that reaches a plateau of approximately
 3,000 ml  after only 20 minutes of  recovery.   And finally, Figure A-6
 illustrates a dryer load with emissions approximately equal  to  the test
 average  (3.2 kg VOC/100  kg articles dried).   While the tumbler  solvent
 concentration curve shows a gradual decline after a modest peak of
 2,800 ppmv (as solvent),  the graph of the solvent recovery rate peaks at
 the  same  rate as the low-emission  load (about 400 ml/min),  but  the
 test-average emission load maintains an elevated recovery rate  over a
 smaller portion of the recovery cycle duration.
A.4  TEST 4 (JAPAN)
     EPA sponsored two visits to Japan by a test contractor (March 1980
and January 1981) with the goals of identifying and assessing the emissions,
 recovery, and safety performance of Japanese solvent  recovery dryers
 (Jerm'gan, 1981).  The first trip was limited to test preparations and
plant visits, while the second trip was intended to focus on actual
field testing of a Japanese recovery dryer.   By the end of  the first
visit, however,  it was  apparent that field tests could not  be conducted
due to insurmountable  problems with the acceptance  of testing equipment
                                 A-13

-------
                          Table A-3.   RECOVERY  DRYER  DATA COMPILATION
Bun (date - 1)
Load composition
12/11-1
2
3
4
5
12/12-1
2
3
4
5
12/15-1
2
3
4
5
6
7
8
9
12/16-1
2
3
4
5
6
7
8
12/17-1
2
3
4
5
6
7
12/18-1
2
3
4
S
12/19-1
2
3
4
5
H
H
H
M
M
H
H
M
M
M
S
U
w
u
u
w
w
5
W
s
w
w
s
w
s
5
W
M
M
M
M
M
M
M
H
H
M
M
H
U
U
U
M
M
**
,
I
|
*J CTi
4-
46.63
37.76
30.73
40.62
38.67
45.59
30.62
16.94
38.22
27.56
19.62
22.57
16.90
22.23
22.57
22.45
22.57
22.66
22.79
11.68
11.34
22.68
22.68
34.25
34.13
45.36
45.36
45.36
34.02
22.68
11.34
34.02
22.68
22.68
45.36
31.64
55.22
34.02
55.68
22.91
22.91
22.66
40.82
12.59
+J
f.
p>
I
1
ts
o 
38.67
29.37
25.63
34.59
32.89
37.65
25.65
15.86
31.98
23.25
16.33
18.60
14.06
18.37
16.94
18.71
'18.94
18.71
19.16
9.98
9.75
16.94
16.94
29.03
28.69
37.76
36.67
38.22
27.44
16.94
9.19
28.58
18.26
16.94
38.10
25.85
46.15
31.30
48.08
19.16
19.05
19.16
34.36
10.09

SU
u E
J **
S3
OC IA
0920
1004
1050
1328
1414
0957
1043
1154
1310
1353
0614
0900
0945
1030
1111
1155
1235
1320
1405
0800
0640
0925
1015
1115
1220
1305
1420
0800
0925
1016
1110
1210
1300
1355
0807
0900
1007
1105
1200
0807
0853
0945
1035
1130
* d
"53

S
II

25
25
25
25
30
30
30
30
30
30
30
30
30
30
30
30
30
45
30
30
30
30
30
30
30
30
40
40
40
30
40
40
40

40
40
40
40
35
35
35
35
35
Recovered water
(kg)
1.17
1.07
0.92
0.65
0.80
0.25
0.20
0.55
0.75
0.75
0.86
0.72
0.75
0.80
0.76
0.70
0.67
0.55
0.71
0.24
0.35
0.79
0.46
0.62
0.85
1.48
1.05
1.56
1.10
1.15
0.50
0.86
0.86
0.77
0.95
1.20
1.47
0.91
1.20
0.81
o.ei
0.85
1.14
0.35
Recovered solvent
(kg)
S.54
4.23
3.36
4.10
4.32
4.85
3.87
2.05
4.78
2.77
2.16
2.27
1.46
2.39
2.41
2.43
2.25
2.25
2.32
0.95
0.99
2.30
2.23
2.86
4.06
5.69
3.87
5.62
4.54
2.28
1.25
3.83
3.23
2.46
5.48
3.91
6.22
3.23
6.46
1.99
2.65
2.16
4.83
1.59
1
I,
U i
1'
* &>+j
O- -c
fef
XT S
14.33
14.40
13.11
11.85
13.13
12.86
14.97
12.91
14.95
11.91
13.23
12.20
10.53
13.01
12.72
12.99
11.86
12.03
12.11
9.52
10.15
12.14
11.77
9.85
14.15
15.07
10.01
14.70
16.55
12.04
13.60
13.40
17.69
12.99
14.36
15.13
13.46
10.32
13.44
10.39
13.91
11.38
14.06
15.76
o>
n
|
*o
tn
I
J
4->
*E
LU
--
0.64
0.96
0.96
0.98
1.06
0.66
0.88
0.93
0.74
0.64
0.55
0.62
0.56
0.55
0.54
0.54

0.49
0.66
0.46
0.47
1.36
0.56
0.53
0.46
0.46
0.60
0.58
0.55
0.48
0.52
0.73
1.43
--

1.12
0.95
1.06
0.91
0.74
0.86
1.46
0.66
!
4V V
11
f
 ffi4-
O JW 
1(1 0-?
ff$

2.87
3.73
2.78
2.98
2.82
2.63
5.54
2.92
3.20
3.92
2.96
4.39
3.04
2.92
2.86
2.83
--
2.58
.59
4.74
2.49
7.16
1.94
1.83
1.21
1.20
1.56
2.12
2.90
5.19
1.83
4.02
7.57
--
--
2.44
3.04
2.20
4.73
3.88
4. SO
4.25
6.52
Max (mum solvent con-
centration at
condenser vapor tnlel
(ppmv solvent)
-






--
2772
2636

2884
2786
2932
3027
3011
2996
3059
2932
2772
2996
3091
3091
3305
3187
3537
3123
3378
3376
3187
2996
3043
3250
3505


3505
3410
3474


3262
3187

Average condenser
water Inlet tem-
perature (<>c)
15.5
16.6
17.3
17.7
19.6
15.8
14.3
15.3
15.4
14.8
11.1
11.7
11.9
11.2
11.2
11.9
10.7
11.4
10.6
12.4
12.2
12.3
11.6
13.5
6.8
7.1
7.6
6.6


6.7
10.1
7.7
8.0
5.8
6.5
6.6
7.6
7.8
18.6
18.7
17.7
16.1
13.6
Average condenser
water outlet tem-
perature (C)
28.7
30.0
30.2
30.4
32.6
27.6
26.1
28.2
27.7
26.5
22.6
21.9
23.1
22.2
21.9
23.4
20.7
21.0
20.0
21 .S
22.2
22.8
21.2
24.3
23.4
23.8
20.2
25.9

--
24.6
24.9
39.2
45.8
38.6
42.3
41.5
43.8
42.7
27.6
29.6
31.6
31.1
35.3
Average condenser
water temperature
difference (t)
13.2
13.4
12.9
12.7
13.2
11.8
11.8
12.9
12.3
11.7
11.5
10.2
11.2
11.0
10.7
11.5
10.0
9.6
9.0
9.5
10.0
10.5
9.4
10.8
16.6
16.7
12.6
19.3


17.9
14.6
31.5
37.8
32.8
35.8
34.9
36.2
34.9
9.0
10.9
13.9
15.0
21.7
Average condenser
water flow rate
(liters/win.)

11.7
13.8
11.4
15.6
-
-
15.9
17.4
16.9
20.6
19.2
17.1
18.6
19.0
16.5
21.0
17.9
19.9
18.4
19.7
17.1
16.1
16.7
15.8
12.4
14.1
10.4
10.2
9.7
10.9
10.8
5.3
4.6

4.4
4.5
4.9
3.7
12.3
12.0
9.0
8.1
5.6
Average cooling
water heat gain
(kilojoules)

16384
18603
15130
21519
-
-
25721
26638
24795
29707
24558
24017
25657
25494
23795
26334
21549
33669
21920
24704
22516
18978
22617
32890
25968
22279
33560


24467
26725
27914
29073
--
26337
26259
29658
21590
16195
19136
16302
17775
17776
Average drying (con-
denser vapor Inlet)
temperature (C)
58.2
62.2
63.0
64.2
53.9
57.6
65.0
71.4
62.6
68.3
67.7
58.8
64.7
61.8
62.4
59.4
59.4
57.0
59.4
58. i
59.2
59.2
59.8
57.8
59.1
56.8
53.8
57.6
--

62.1
58.4
60.3
60.7
56.6
60.8
58.7
59.8
59.6
59.9
59.6
60.0
56.8
62.8
Average condenser
vapor outlet
temperature (C)
26.4
25.3
24.4
26.0
26.1
22.4
22.7
23.6
22.4
22.5
18.4
18.8
21.3
18.4
17.6
18.7
16.9
17.5
16.2
16.7
17.3
18.3
21.1
21.4
16.4
18.3
15.9
17.5


15.1
15.7
24.8
30.2
26.0
27.9
28.6
28.9
29.2
28.7
27.9
26.6
28.7
27.6
M  501 Wools, 50i Synthetic blends
W  lOOt Wools
S  1005. Synthetic blends
                                                        A-14

-------
                M
                z
                *
               S.*   38.67
B^

UJ
    8000   .
    7000   -
    6000   -
    5000
   1000
   3000
   2000
   )000
                                                                                          4000
                                     900


                                     800
                                         c
                                         'i
                                     700  ^
                                         E

                                     600  
                                         o

                                     500  *
                                         i
                                         z
                                         UJ
                                     400  ""
                                        o
                                        l/l
                                        u.
                                        o
                     10     15    20     ?5     30     35    40
4r>    50
fiO
                                    300

                                    200


                                    100


                                      0
    Figure A-4.   Recovery and concentration curves for a  low-emission  recovery dryer  load.

-------
                                                         9I-V
                               RECOVERED SOLVENT  VOLUME  (ml)
c
-J
n>
o
o
<
m
3
CX


O

O
3
n
O)
n
c
-t
o
-t
Ol


3"
 I

 a>
 ro
 n
 o
 <
 n>
 a.

<3
 O)
                   o
                   o
                   o
    r
-------
                                                               zi-v
                                    RECOVERED SOLVENT  VOLUME (ml)
                        0
                        Q

                        O
   O
   O
   o
o
Q
         a
         a
         o
           
  el- n
  3- o
     <
  fD  fD
  3  T

  i/)
  i/)  at


  o' Q.

  co  n
     o
  CU  3
 a  n
 a  a)



  x  T"
  _.. cu

  3  rt-

     O
     3
 ro
     C
 ro  -j
 -o  <
 c  ro
 01  i/i
 0 -J


 rt Qi
 o o
 < o
 ro <
 "1 fD

 ^"<
 ro ^<
 VI  fD
 <-f -I

 a  i
 <  O
 a>  cu
 T  Q.
 Oi
ca
                                                                                                          (/ - rp)
                                                                                    ptoi
                                                                                                                  (5?)
                                                                                3UU 1JTJI

                                                                              apXa dittpm



                                                                            ('ui) uoatjnv

                                                                              ^<(3 jjtl^ay
                                                                            DPOI fjp 5< ^ot
                                                                            30( Oo 5) XII







                                                                    ^  l?Ul ^CCPA ^4SUPU09

                                                                    ^        IB JflllBJIuAl
                                                                                                          Co; a:ujaip
                                                                          j4SU*pU03 4&tJ*Ay
                          CONCENTRATION AS  PPM SOLVENT  (Condenser  Inlet)
                      a
                      o
no     CJ
O     O
O     O
   O
   a
o
o
en
o
o
o
o
03

O

O
a
a
                                 RATE OF  SOLVENT RECOVERY (ml/min)

-------
(particularly FIA calibration gases) by Japanese customs.   As a result
of these problems, the second visit to Japan was conducted like the
first, with a single EPA representative visiting major recovery dryer
manufacturers, as well as a dry cleaning trade association and individual
dry cleaning plants.
     There were approximately 1,800 petroleum solvent recovery dryers
operating in Japan at the time of the test visit, according to the All-
Japan Laundry and Drycleaning Association (AJLDA).   These units were
designed like their American counterparts, with steam-heated evaporation
of solvent from drying clothes being followed by solvent recovery in a
refrigerant-chilled condenser prior to the reheating of the circulating
vapor stream.  In contrast to American-made recovery dryers with typical
load capacities of up to 48 kg, Japanese recovery dryers have load
capacities of 15 to 20 kg, with only a few units (approximately 90) with
a 50 kg capacity.
     For purposes of comparison, AJLDA estimated that there were about
5,900 non-recovery petroleum solvent dryers operating in Japan.  In
addition, over the previous five years, there had been about 17 explosions
or fires in recovery dryers and about 50 in non-recovery dryers.  None
of the recovery dryer explosions had resulted in personal injury, and
the ratio of the number of existing dryers to the number of explosions
was the same (1,000 to 1) for both recovery and non-recovery dryers.
The absence of damage and personal injury in recovery dryer accidents
resulted from the recovery dryer's ability to safely control an explosion
while preventing fires (which were typical of accidents in non-recovery
dryers).  The Japanese method of explosion relief around the dryer door
had the same effect as the explosion damper ports on top of the American-
made recovery dryer, with the force of the explosion being vented to the
atmosphere.  Finally, the AJLDA reported that Japanese recovery dryers
could recover from  70 to 95 percent of the solvent contained in drying
loads, depending on the recovery cycle duration and the condenser
cooling water temperature.
A.5  TEST 5  (WILMINGTON)
     An EPA-sponsored study was conducted to determine the rate of
solvent drainage from heavily soiled cartridge filter elements and to
                                 A-18

-------
compare the drainage rates of new and soiled cartridges.  A
recommendation was made for a minimum drainage time for these elements
based on the total maximum solvent emission from the entire filtration
system (Plaisance, 1981).
     The host plant for this study was a petroleum dry cleaning facility
located in Wilmington, North Carolina,  This plant cleaned approximately
900 kg (2,000 Ibs) of lightly soiled general apparel each week, expending
about 380 liters (100 gallons) of Ashland Kwik-Dri solvent.   Dry cleaning
equipment used at this facility was limited to a single 27 kg (60 Ib)
capacity Marvel Matic washer and two 22 kg (50 Ib) capacity Heubsch
Originator dryers.   Spent solvent was filtered and purified by a
14-element cartridge filter (Puritan Vanguard 14) which employed
12 carbon-core and 2 all-carbon filter cartridges.  The carbon-core
filter elements served to both remove solids and provide initial
purification, while the all-carbon filter element provided final solvent
purification.
     The test program for this study consisted of removing two filter
cartridges (one carbon-core and one all-carbon) that contained heavy
concentrations of lint and dirt from the cartridge filtration system.
These elements, along with two new cartridge elements (one of each
type), were soaked in sealed containers of solvent and then were allowed
to drain while weight loss readings were recorded.  The cartridge
elements then were placed under an exhaust hood where the solvent was
allowed to evaporate freely at room temperature,  and each sample cartridge
was weighed twice each day for seven days.
     A comparison of the percent solvent loss,  as a function of drainage
time between new and used cartridges,  indicated that the soil  and residue
loadings of both used cartridges caused a lower rate of solvent drainage.
The largest differences in drainage rates occurred between new and used
all-carbon cartridges,  while a similar comparison of carbon-core
cartridges resulted in a more equal  percent drainage of initial  solvent
content over the entire drainage period.   In general, carbon-core
cartridges gave up solvent at a higher rate than all-carbon cartridges
over extended periods of drainage and evaporation.
                                 A-19

-------
     The results of this test indicate that a solvent drainage duration
of 8 to 12 hours (overnight) would be sufficient to produce a minimal
total emissions (see Figure A-7), while being brief enough to prevent
disruption of normal plant operation.  The undrained system emission
rate of 0.56 kg solvent per 100 kg dry weight of articles cleaned would
be reduced by 37 to 40 percent after drainage durations of 8 to 12 hours,
respectively.  Table A-4 shows emissions from cartridges over a period
of several days.  Comparison of solvent retention in new and used
cartridges indicates that new cartridges having no soil and residue
loading would have initial (undrained) emission roughly equivalent to
those of the used cartridges.  Therefore, general results gained in this
test could be applied to facilities having lower throughputs between
cartridge replacements and/or lower soil loading.
A.6  TEST 6 (ANAHEIM)
     EPA contracted a study to collect and analyze data at a large
industrial dry cleaning facility to evaluate the technical and economic
feasibility of reducing the solvent content of still  wastes through:
(a) operating procedure modifications, and by (b) installing a densio-
meter in the still bottom to control  the boildown schedule.   In addition,
hydrocarbon concentrations were measured from fugitive sources within
the dry cleaning plant during the test period (Jernigan and Kezerle,
1980).
     The host plant for this test was a large industrial laundering and
dry cleaning facility located in Anaheim, California.   This facility
utilized a 230 kg (500 Ib) Washex washer/extractor and a 180 kg (400 Ib)
Challenge-Cook dryer to clean approximately 8,700 kg  (19,000 Ibs) of
articles per week.   In addition, the  facility had two solvent stills,
each with a 1,900 liter (500 gal) per hour capacity,  manufactured by
Washex.   Data were collected at the plant for this evaluation from
November 5 to November 19, 1979.
     Procedures used to accomplish the test objectives included the
determination of the boildown time for the vacuum still and the deter-
mination of the specific gravity of the solvent/still  waste mixture
during distillation by using a densiometer.   Plant records were examined
                                 A-20

-------
3=
 I
ro
            o.
            -C
            en
          CO 3
          z o
          o i.
          *- O)
          z; -*
          UJ

          K-8
          O (U
          00 >
            en
                                                TIME  IN MINUTES
                      Figure A-7 .  Solvent  Emissions for Filter Cartridges as  a  Function of
                                   Drainage Time.

-------
           Table A-4.   TOTAL SOLVENT  EMISSIONS  DUE  TO  DISPOSAL  OF
           14 FILTER CARTRIDGES  (12 CARBON-CORE AND 2  ALL-CARBON)
                       AS A FUNCTION  OF  DRAINAGE TIME
Elapsed drainage
      time
     Solvent emissions
 (in kg solvent emitted per
100 kg of articles cleaned)
   Percentage of
undrained emissions
      0

      8 minutes

      8 hours

     12 hours

   8.25 days
            0.56

            0.41

            0.35

            0.34

            0.22
        100

         73

         63

         62

         39
                                      A-22

-------
 to determine the frequency of still  boildowns.   In addition,  fugitive
 solvent emission levels  at various  locations  in the plant were measured
 using a Beckman  400  flame  ionization analyzer (FIA).
      Hydrocarbon concentrations  in  and  around the  work area  in the  dry
 cleaning facility were reported  as  parts  per  million  (ppm) of propane.
 (A standard  conversion factor from  propane  to Stoddard of 3.36 was
 calculated.)  Emissions  could be approximated only for the roof exhaust,
 where approximately  1.56 kg  (3.43 Ibs)  of Stoddard solvent was emitted
 per hour.  The highest solvent vapor levels in  the workplace  were
 recorded around  the  washer,  averaging about 3,300  ppm  as  propane or
 980 ppm as Stoddard.  Hydrocarbon concentrations at the clean solvent
 tank were an order of magnitude  higher, reaching as high  as 24,000  ppm
 as propane or 7,150  ppm  as Stoddard  solvent.
      Results of  this test  indicated  that  measuring the specific gravity
 of the  still  contents during  distillation with  a densiometer  was not
 feasible or  even  desirable,  due  to the  adverse  thermal  and mechanical
 effects  of rapid  boiling on  the  sensitive densiometer  mechanism.
 However,  the solvent content  of  the  still waste generated at  this
 petroleum dry  cleaning facility  could be  reduced,  with no adverse
 effects,  by  boiling down the  stills  less  frequently.   At the  time of
 this  test program, the stills were boiled down and  the waste  in the sump
was  discarded  daily.   This meant  that 144 liters (38 gal), or  115 kg  of
 still waste,  containing  approximately 90 percent pure  solvent  by volume
was  discarded  each day.   Table A-4 records the samples of still waste
that were analyzed for solvent content before and  after boildown.  On
the  first day  (11-07-79)  the  still waste (sample VIS-4) contained more
than 99 percent by volume (97% by wt) solvent.  Still waste sample
VIS-21 on the  seventh day (11-14-79) represents the typical volume of
solvent discarded daily at this plant and has  the  highest daily
throughput.   This sample  contained approximately 90 percent by volume
(91% by wt)  solvent,  representing a  reduction  of 5 percent over sample
VIS-4, due to the reduced boildown frequency.
     Results  from solvent content analyses conducted on 11-15-79 and
11-16-79 (VIS-23  and VIS-26,  respectively) showed no appreciable
difference in their solvent contents.  Instead of a decrease  in solvent
                                 A-23

-------
content from samples VIS-23 to VIS-26,  there was actually a 2 percent
increase.   This increase is considered insignificant on a day-to-day
basis and may be attributable to a number of factors such as a change in
the type of articles cleaned, still operation, or a slight variation in
laboratory procedures for analyzing the still waste.
     Still waste was allowed to accumulate for 10 days (11-09-79 to
11-19-79) before the still was boiled down again.  On the last day of
testing, the still was boiled down and sample VIS-30 was analyzed for
its solvent content.  This sample contained 25 percent less solvent, on
a mass basis, than VIS-4 and 21 percent less solvent than VIS-26, as
shown in Table A-5.
     The analytical procedures used to determine the solvent content in
each sample involved determining the moisture content by the Carl-Fisher
Method and gravimetrically determining (at 103C) the nonvolatiles in
the samples.  The  solvent content then was determined by a process of
elimination, in which the quantity of nonsolvent components was
determined and then was subtracted from the total sample mass.
     An alternate  method to decrease solvent losses would be to reduce
the total volume of still waste, by eliminating the inactive space in
the still below the steam chest.  In the tested still design, liquids
below the steam chest did not receive sufficient heat to vaporize during
boildown.  These liquids, which contained a high concentration of
solvents, were discharged daily after boildown.  The more frequently the
still was boiled down, the greater the amount of solvent discarded with
the wastes.
     A VOC emission rate of 1.53 kg per hour was recorded (11-09-79)
during the course  of a 6-hour dry cleaning day when approximately
1,360 kg  (3,000 Ibs) of pants were dry cleaned.  This equates to 227 kg
of articles cleaned per hour.  Thus, the ratio of the mass of fugitive
solvent emitted per hour to the mass of articles cleaned per  hour is
0.687 kg  of fugitive VOC emitted per 100 kg of articles  cleaned.  Two
access doors were  open  (front and  rear of the dry cleaning area) during
the  fugitive emissions  test.  It is assumed  that the  recorded fugitive
emissions  rate would have  been  higher  if these  doors were closed.
                                  A-24

-------
                                    Table A-5.   RECORD  OF STILL WASTE BOILDOWN SAMPLES
I
IV)
tn
Date
11/07/79
11/14/79
11/15/79
11/16/79
11/19/79
Military
time
0730
1245
1230
1200
1320
Sample
no.
VIS-4
VIS-21
VIS-23
VIS-26
VIS-30
Throughput
(kg)
1360
1588
1360
1425
1425
Still Weight percent
waste solvent in
(kg) waste
115
115
115
115
115
97.
92.
91.
93.
73.
40
20
30
00
00
Solvent loss
with waste (kg) solvent loss/
(kg) 100 kg articles cleaned
112
106
105
107
84
8.
6.
7.
7.
5.
20
70
70
50
90

-------
A.7  TEST 7 (ANAHEIM)
     EPA contracted the evaluation and demonstration of carbon adsorption
technology at an industrial  dry cleaning facility in Anaheim,  California
(Lutz et al., 1980).   This program was developed to determine  the
effectiveness of carbon adsorption in controlling VOC emissions.   It
consisted of fitting a prototype carbon adsorption unit, purchased from
VIC Manufacturing Company of Minneapolis, Minnesota, to the dryer exhaust
of a petroleum solvent industrial  dry cleaning dryer; operating the
system to collect performance data; and evaluating the economics of
operation at this establishment.
     The host dry cleaning plant in Anaheim, California, is a large,
industrial facility utilizing a 230 kg (500 Ib) Washex washer/extractor
and a 180 kg (400 Ib) Challenge-Cook dryer to process approximately
8,700 kg (19,000 Ib) of general apparel per week.  This throughput
represents about 50 percent of the 8-hour capacity of the dry cleaning
dryer.  Data were developed to determine the effect of the different
utilization rates on the various parameters under evaluation.   After
installation of the carbon adsorption unit, testing was conducted at the
facility from July 24, 1978 to March 23, 1979.
     Test procedures used during the carbon adsorption test program
included a determination of hydrocarbon concentrations by continuously
sampling the gas streams to and from the carbon adsorption unit.  This
was accomplished using two Beckman 400 flame  ionization analyzers (FIA).
Both the inlet and exhaust gas  stream flow  rates were continuously
monitored, as were the temperatures of the  various  liquid and gas streams.
Other parameters measured during the test program  included:  electricity
consumption, natural gas consumption, water usage,  steam flow rate  to
adsorption unit, and solvent  recovery rate.   In  addition, samples of
solvent and  samples  of carbon from the carbon  bed  were  analyzed
infrequently during  the  test  period.
     The carbon adsorber  system (see  Figure A-8) was  initially operated
in  strict compliance with the recommendations  and  instructions of the
adsorber manufacturer  and his field  representatives.   Early in this test
period,  it became  apparent  that the  adsorption system had been over-
designed, resulting  in removal  efficiencies far  in excess of  the
                                  A-26

-------
        Exhaust 

      from Dryer'
i
ro
Bag
Filtration
   LEGEND


 Solvent Laden Exhaust Gas


 Solvent Laden Steam


 Exhaust Gas  Void  of Solvent


Steam Void of Solvent
                      Low Pressure Steam

                     from Auxiliary  Boiler
                                 Air Cooling
                                   Sy s tern
                                                                                                      Water/Solvent
                                                                                                       Separator
                                                                                                         <
                                                                                      Solvent to Storage i
                                                                                             (Condensed
                                                                                             !Sol vent
                                                                                           * land Water
                                                                                        Exhaust  to
                                                                                        Atmosphere
                                    Mgure  A-8.   Carbon  Adsorption  System Schemati
                                                                 tic.

-------
specified performance guarantee of 90 percent solvent removal  on  a
24-hour average.   The test program was,  therefore,  amended to  include an
evaluation of changes to the design and  operating procedures for  the
carbon adsorption system.   Various design parameters were modified to
determine their effect on the performance and cost of the adsorption
system.  From these studies, an optimized system was established  for use
in evaluating the performance, cost, and cost effectiveness of utilizing
carbon adsorption technology for the reduction of VOC emissions from
petroleum dry cleaning plants.
     The following alterations to the original adsorber design resulted
from the optimization:  (1) the lint filter area was increased by
80 percent to facilitate daily cleaning; (2) the blower that forces the
dryer exhaust through the adsorber was modified to operate only when the
dryer was running, rather than continuously; (3) the original  system of
three carbon beds was reduced to two beds; (4) desorption steam pressure,
flow rate, and duration were optimized at 103 kilopascals, 590 kg/hr,
and 60 minutes, respectively; and (5) the adsorber inlet (dryer exhaust)
vapor cooler was eliminated, because dryer exhaust temperatures were
insufficient to damage the carbon beds.
     The hydrocarbon emission reduction  efficiency for the optimized
design (applied to the dryer exhaust) was 95 percent, and varied  from
93 percent for a plant with 100 percent  utilization to 97 percent at
25 percent utilization.   Capital costs for this system, including site
preparation and equipment installation,  are estimated at $128,000
(mid-1978 dollars).  Cost effectiveness, defined as the annual operating
cost divided by the quantity of emission reduction, is a function of
equipment utilization rates, and additionally exhibits a strong
dependence on the market value of the recovered solvent.   A solvent cost
of $0.16/liter ($0.61/gal) was assumed for the basic analysis, but the
effect of increases in petroleum costs on annualized operating costs was
investigated.  The cost effectiveness of the optimized design was $560/Mg
($510/ton), and was estimated as $l,090/Mg ($980/ton) and $220/Mg
($200/ton) for 25 percent and 100 percent utilization, respectively.
When the value of Stoddard solvent reaches $0.60/liter ($2.30/gal), the
optimized system (50 percent utilization) will have zero annual
operating costs, neglecting the rise in  other operating expenses.

                                 A-28

-------
     The results of this project demonstrate the technical  feasibility

of applying carbon adsorption technology to reduce the emission of

hydrocarbon solvents from dryer exhausts at petroleum solvent dry cleaning

plants.  The cost effectiveness of this technique, $560/Mg  ($510/ton),

is expected to drop significantly as the value of the reclaimed solvent,

a petroleum distillate, increases.   Even at its present cost effectiveness,

carbon adsorption is economically comparable with the cost  of emission

reduction required in other industries.  An additional benefit, provided

by the application of carbon adsorption technology to the petroleum dry

cleaning industry, is the reduction in overall consumption  of petroleum

products by these plants.   The demonstration plant recovered solvent at

a rate of 61,000 liters (16,000 gal) per year which otherwise would have

to be replaced with new solvent purchases.

A.8  REFERENCES FOR APPENDIX A

Jernigan, R.   1981.   Identification and Assessment of Emission Control
     and Safety of Japanese Petroleum Solvent Recovery Dryers.   TRW Inc.
     Research Triangle Park, North  Carolina (EPA Contract No.  68-03-2560).

Jernigan, R.  and J.  Kezerle.  1981.   Evaluation of the Potential for
     Reduction of Solvent Losses Through a Washex Petroleum Vacuum Still
     Sump.   TRW Inc.   Research Triangle Park,  North Carolina (EPA Contract
     No.  68-03-2560,  Task No.  T5013).

Jernigan, R.  and S.  Lutz.   1980.  An Evaluation of the Emission Reduction
     Potential of a Solvent Recovery Dry Cleaning Dryer.   TRW Inc.
     Research Triangle Park, North  Carolina (EPA Contract No.  68-03-2560).

Jernigan, R. ,  G.  May.,  and S.  Plaisance.   1981.   An Evaluation of Solvent
     Recovery and Emission Control  of a Solvent Recovery Dry Cleaning
     Dryer.   TRW Inc.   Research Triangle Park,  North Carolina (EPA
     Contract No.  68-02-3063).

Lutz, S., S.  Mulligan and A. Nunn.   1980.   Demonstration of Carbon
     Adsorption Technology for Petroleum Dry Cleaning Plants.
     Cincinnati,  Ohio.   EPA Publication No.  600/2-80-145,  EPA/IERL.

Plaisance,  S.   1981.   A Study of Petroleum Dry Cleaning Cartridge Filter
     Element  Emissions.   TRW Inc.   Research Triangle Park, North Carolina
     (EPA Contract No.  68-02-3063).

Plaisance,  S. , R.  Jernigan,  G.  May,  and C.  Chatlynne.   1981.  An Evaluation
     of Petroleum Solvent Concentrations,  Emissions,  and Recovery in a
     Solvent  Recovery Dryer.  TRW Inc.   Research Triangle  Park,
     North  Carolina  (EPA Contract No.  68-03-2560).
                                 A-29

-------
          APPENDIX B
EMISSION MEASUREMENT PROCEDURES

-------
                               APPENDIX B
                     EMISSION MEASUREMENT PROCEDURES

     Emission measurement procedures for large petroleum dry cleaners
would be based on a determination of the amount of solvent lost in the
overall dry cleaning process (material balance).   This procedure could
be replaced or augmented by field test procedures on particular items of
dry cleaning equipment, such as the solvent recovery dryer or the vacuum
still.   Both of these procedures will be described in the following
sections.
B.I  MATERIAL BALANCE METHODS
     Material balance methods can be used to determine the total loss of
solvent in a petroleum solvent dry cleaning plant by measuring solvent
input and output at each step of the dry cleaning process (solvent
storage, washing, drying, filtration, distillation, and losses from
fugitive sources.)  A material  balance requires measurement of clothes
and solvent over a number of loads in addition to solvent levels in the
system before and after testing.   All significant sources of solvent
must be accounted for.   The following method was  developed by EPA (for
the Perchloroethylene dry cleaning CTG) with the  assistance of an EPA
contractor and the International  Fabricare Institute.   The method outlined
here should be considered flexible for the different processes in the
industry.
A.    Before the test begins, establish a solvent  baseline by the following
     methods:
     1.    Drain entire cartridge  filter contents  (solvent) to holding
          tank (or vacuum still).
     2.    Complete distillation  and begin  boil down of  vacuum still.
          Remove cartridge filters and discard.

-------
     3.   On completion of still boildown, remove still  bottoms (high
          boilers).
     4.   Start up wash pump to fill  filter housing (ideally,  machine
          should be on continuous recirculation - solvent circulating
          between base tank and filter and returning).
     5.   Add any detergent needed.   (Take solvent sample, if  needed).
     6.   Measure solvent level by dip stick or gauge in washer base
          tank or ground tank.   (Account for residue volume in bottom of
          tank.)
     7.   Make sure that the recovered solvent flow from the recovery
          dryer is directed to the washer base tank or ground  tank.
B.    During the test:
     Record weight of all loads.
C.    After the test period, recreate  conditions of first solvent
     measurement by repeating Steps A.I through A.4.   Another  sample may
     be taken to determine detergent  concentration in the "charged"
     solvent, if needed.
     The solvent loss in cartridge filters is a fixed loss for the
number of loads recommended for use.   In other words, if a filter vendor
recommends 200 loads of articles as the filter life, the loss  from a
filter change is the same as the 200  loads whether there are 50 loads or
300 loads run during the test period.   The loss from filters for a test
of less than the recommended filter life should be prorated to the life
of the filter.  A loss of 1 kilogram  after 50 loads on a filter of
200 load life should be considered in the calculation as a loss of
0.25 kilograms.
     Fixed losses are a significant factor in petroleum solvent washers.
A 70 kilogram load in a 115 kilogram  capacity machine will have nearly
the same loss as a 115 kilogram load  in the same machine.  In  calculating
kilograms of clothes throughput, the  vendor capacity times the number of
loads should be used instead of the actual load weight.   (The  IFI and
other trade organizations can relate  cubic feet of water volume to
capacity by available factors too extensive to list here.)
                                 B-2

-------
      To  determine  solvent  consumption,  the  solvent  level  (minus  detergent,
 sizing,  etc.)  of the  initial measurement  (Step A.6)  is compared  to  the
 solvent  level  (minus  detergent,  sizing, etc.) of the  final measurement
 (Step C).   All  solvent  added during the test period  should be accounted
 for.
      To  determine  the system emission factor for the  test period  (which
 should be  for  at least  20  working days and  14,000 kg  of articles  cleaned),
 the  solvent consumption  is divided by the weight of clothes cleaned,
 resulting  in a  determination of  the mass of solvent lost  (emitted)  per
 mass  of  articles cleaned.
      According  to  the IFI, samples should be analyzed for detergent
 concentration,  moisture, nonvolatiles, dry  sizing, and insoluble  materials.
 A Hyamine  1622  or  Aerosol  OT Titration should be used for detergent
 concentration  reported on  a volume/volume percent basis.   The moisture
 content  is determined by a Karl-Fischer titration procedure and reported
 as grains  of water/100 milliliters of solution.   Nonvolatile residue is
 determined gravimetrically by a  steam bath evaporation of a measured
 volume of  solvent  and weighing the residue.   Dry sizing content is
 determined by extracting the nonvolatile residue with boiling ethyl
 alcohol.    Insoluble material  content is to be determined gravimetrically
 after filtration of a volume of solvent through a 0.20 micrometer membrane.
 B.2   RECOVERY DRYER EMISSIONS MEASUREMENT
      Petroleum  solvent recovery dryer VOC emissions can be determined by
 either a material  balance on the dryer or by using a flame ionization
 analyzer (FIA) to determine the solvent content of the dryer atmospheric
 exhaust.
 B.2.1  Direct Measurement of Dryer Exhaust Emissions
     The determination of recovery dryer emissions by FIA analysis of
 dryer exhaust solvent concentrations is a more complex and technically
demanding procedure than the  material  balance method discussed below.
The methods used in this test should include EPA Reference (40 CFR
 Part 60)  Test Methods 1, 2, and proposed (45 FR  83126 December 17, 1980)
25A.
     While Methods  1 and 2 would govern the  selection of  atmospheric
 vapor exhaust sampling points and the procedure  for determining  the
                                 B-3

-------
exhaust flow rate, respectively,  Method 25A would govern the  measurement
of the VOC (as propane) vapor concentration in the control  device
atmospheric exhaust by a flame ionization analyzer (FIA).   First, a
response ratio of the FIA's measurement of a given concentration of
propane to the same concentration of VOC (solvent) would be determined
in the laboratory (see Attachment 1).   Then the FIA would be  field
calibrated to measure concentrations of propane gas, and the  measured
ppmv concentrations (as propane)  of the control device exhaust gases
would be multiplied by the previously determined response ratio, thereby
determining the ppmv concentration (as solvent) of the VOC emissions.
Then by calculation, one would convert the VOC concentration  by volume
(ppmv) to mass concentration.  This procedure should be carried out
under various conditions of fabric type, load weight, and temperatures
that are typical of the range encountered in the dry cleaning industry.
The results of this procedure should be reported as kilograms VOC emitted
per 100 kilograms dry weight of articles dry cleaned.  Subsequent changes
in the design or performance of the control device could necessitate a
reevaluation of the device's maximum VOC emissions.   Since Method 25A is
not a promulgated EPA Reference Test Method (as of the date of this
appendix) and subject to change an alternative test procedure can be
used in the interim.  One alternate test method is:   "Alternate Test
Method For Direct Measurement of Total Gaseous Compounds Using A Flame
Ionization Analyzer," presented in the OAQPS Guideline Series document
entitle, "Measurement of Volatile Organic Compounds" (Revised September
1979, EPA-450/2-78-041).
B.2.2  Dryer Material Balance
     Like the plant-wide material balance previously discussed, the
recovery dryer material balance measures dryer emissions by accounting
for all solvent entering and leaving the dryer.  A material balance of
the solvent input and output in a recovery dryer should be based on
precise weight measurements of dried loads, recovered solvent, and
water.  For a given load, the dry load weight should be measured and
recorded prior to washing.  The load should be weighed after washing and
the weight should be recorded before drying.  During drying,  the recovered
solvent and water should be collected and their weights should be recorded
                                 B-4

-------
 at the end of drying,  along with the weight of the dried load.   Based on
 the assumptions that the  weight of the  dirt removed from the  load in
 washing is insignificant  (when compared with the  load weight)  and that
 all of the water contained in the dried articles  is recovered,  the
 weight of solvent could be calculated from  the following equation:
      WSE  =                   -  WRW
 Where
     WSE   =  Weight  of  Solvent  Emitted  per  100  kg  of  articles  cleaned
             (including dryer exhaust emissions, solvent contained  in
             dried articles, and  fugitive losses within the dryer)
     PDW   =  Pre-Dried  Weight of  articles (kg)
     WRS   =  Weight  of  Recovered  Solvent (kg)
     WRW   =  Weight  of  Recovered  Water  (kg)
     PWW   =  Pre-Washed Weight  of articles  (kg)
 Scales  used  in measuring these  weights  should be accurate to 0.25 kg at
 weights of up to 100 kg.  These weighings should be conducted  during
 normal  plant operations  that reflect typical load  sizes, fabric types,
 and operating parameters such as  cooling water temperature and flow
 rate, drying temperature, and recovery  cycle duration.
 B.3  DETERMINATION OF SOLVENT CONTENT OF VACUUM STILL AND FILTRATION WASTE
     A  determination of  the quantity of solvent contained in vacuum
 still and  filtration waste would  require periodic  sampling of waste
 removed from the still   after boil down or removal of filter waste from
 its housing, respectively.   At  least three one-kilogram samples of still
 and filtration waste should be taken after a period in which conditions
 of soil loading, load weight, and fabric type vary over a range that is
 typical for the facility.  These samples should be collected in sealable
 containers which are impervious to petroleum solvent.   Also, the total
mass of articles cleaned since the previous  still  boildown or filter
 change should be recorded,  as should the total  mass of still or filter
waste produced since the previous waste removal.
     Determination of the solvent content of the still and filter waste
 should be based on the  application of the procedure outlined in ASTM
Method D 322-80 (Standard Test Method for Gasoline  Dilutent  in  Used
                                 B-5

-------
Gasoline Engine Oils by Distillation).   This procedure (see Attachment 2)
should result in a determination of the solvent content (mass) per unit
mass of still or filter waste.   This factor should be multiplied by the
mass of still or filter waste produced per 100 kg dry weight of articles
cleaned.   The final  result of this procedure should be a determination
of the mass of solvent contained in still  or filter waste per 100 unit
mass of articles dry cleaned.
                                 e-e

-------
                              ATTACHMENT 1

           DEVELOPMENT OF A SOLVENT-TO-PROPANE RESPONSE FACTOR

     The chemical properties of petroleum solvent vary from supplier to
supplier and also from shipment to shipment of solvent.  For this reason,
standard calibration gases for petroleum solvent are unavailable for use
in calibrating the hydrocarbon analyzer (FIA).  Propane (C^Hft) span
gases with concentrations of 10,000 ppm, 1,000 ppm, and 100 ppm having
an analytical accuracy of 2 percent can be used for calibration during
the field test program and also during the laboratory development of a
solvent-to-propane response factor to convert concentration readings as
propane to concentrations of petroleum solvent.   These span gases should
be certified and traceable to the National  Bureau of Standards (NBS) by
the gas supplier.
     In order to determine the relationship between the hydrocarbon
concentration data recorded from an FIA calibrated to propane and actual
solvent concentrations,  a response factor has be determined.   The
following is a laboratory test that was developed as a guide to the
determination of a solvent-to-propane response factor.
RESPONSE FACTOR DETERMINATION
     Prior to running the solvent samples,  the solvent preparation
system (Figure B-l)  should be given time to reach the proper temperature
and stabilize.   The  system then should be purged several  times with zero
hydrocarbon air, and the exhaust of each purge monitored  to ensure that
the system is free of hydrocarbons.   Before the  purging of the system
begins, the hydrocarbon  analyzer should be  calibrated with zero
hydrocarbon air which contains less than 2  ppm of total hydrocarbons and
span gas which contains  100 ppm of propane.
                                 B-7

-------
co
    Zero
 Hydrocarbon
Air Cylinder
       Dry
     Gas Meter
               Molesieve
                Filter

                   Hot Oil
                    Bath
Temperature Readout

 Pressure Readout

 Heated Teflon Sample Lines
                                T  /
                                      Syringe
                                      Septum
                                            Hot  Plate
                                                              LQ
                                                                       Heated Metal
                                                                         Barrel
                                                                                     Tedlar Bag
                                                                                  0.40 cu. ft. Capacity
                                                         Pump
                                                                                     H.C.
                                                                                   Analyzer
Recorder
                  Figure B-l. Response Factor Preparation System.

-------
     Prior to each sample run the system should be purged, and after
several sample runs the span calibration should be checked.
     Materials used in the test system should be selected for their
reliability and ability to deliver a non-degraded sample to the
hydrocarbon analyzer.
     A hydrocarbon analyzer should be used which employs the flame
iom'zation method for determining hydrocarbons.   A check on the Tedlar
bag should be made for possible leakage by filling it until it becomes
rigid and allowing it to stand for approximately 12 hours.  At the end
of this period, the bag should remain rigid, indicating that it is leak
tight.   The Teflon diaphragm pump and the dry gas meter should be tested
for leaks before starting the tests.   The syringe used for the petroleum
solvent samples should be gas-tight to ensure that no loss of sample
occurs.  Both the zero and span calibration gases used on this
hydrocarbon analyzer should be certified and traceable to NBS by the
supplier.
     For a typical sampling run,  data would be recorded for the
following components:
     1.   Dry gas meter temperature.
     2.   Dry gas meter volume.
     3.   Pressure drop across dry gas meter.
     4.   Oil bath temperature.
     5.   Temperature of heated lines.
     6.   Heated metal container temperature.
     7.   Barometric pressure.
     After the readings are taken, a sample of petroleum solvent is
injected into the impinger where it is vaporized.   Immediately following
vaporization, zero hydrocarbon air should be introduced at approximately
260 mmHg (5 pounds per square inch) of regulator pressure and 2.8 liters
per minute (0.1 cubic feet per minute).   When 9.9 liters per minute
(0.35 cfm) are introduced into the system, the zero hydrocarbon air
should be turned off.   Before the zero hydrocarbon air is turned off,
the valve on the Tedlar bag should be closed.
     The next step is to record the dry gas meter volume reading.   After
the reading is obtained, the Tedlar bag contents should be completely
extracted and delivered to the hydrocarbon analyzer.   The results of the

                                 B-9

-------
hydrocarbon reading should be recorded, and then purging of the system
begins in peparation for the next sample run.
     The equations used in response factor preparation are given in
Table B-l.   It should be noted that properties peculiar to the solvent
used (molecular weight and specific gravity) should be determined by
consulting the solvent manufacturer.
                                B-10

-------
 Table  B-l.   EQUATIONS  FOR CALCULATING SOLVENT-TO-PROPANE  RESPONSE  FACTOR
SM! 0.76 mg 103 |jg MJJ
Ml mg
CS ~ Vm Y 106 M!
1
r - S (133.
S
Vm v 293
T
Mole
138
293
m
4)

24.04 Ml in
Mq Mole 1U
Pm
760
P
m
760
Equation 1
Equation la
where C<-       =  Standard concentration in ppmv as solvent

      C        =  standard concentration in ppmv as Propane

      S        =  volume pi of solvent injected

      Vm       =  gas volume measured by dry gas meter in liters

      Y        =  dry gas meter calibration factor

     Pm        =  absolute pressure of the dry gas meter, mmHg

     Tm        =  absolute temperature of the dry gas meter, K

     0.76      =  specific gravity of solvent at 293 K

     44        =  molecular weight of Propane

     138       =  molecular weight of solvent

     24.04     =  ideal  gas (specific volume) at 293 K,  760 mmHg
Response ratio =  response to Cp ^  where c  =
                  response to C           P
                                 B-ll

-------
                                    ATTACHMENT 2
                             ANSI/ASTM D 322 - 80
          Designation: 23/68 (79)
          Standard Test Method for

          GASOLINE DILUENT  IN  USED GASOLINE ENGINE OILS
          BY DISTILLATION1

 This standard is issued under the fwed designation D 322, the number immediately following the designation indicates the
 year of original adoption or. in the case of revision, the year of last revision. A number in parentheses indicates the year of last
 reapproval. This is also a standard of the Institute of Petroleum issued under the fixed designation IP 23. The final number
 indicates the year of last revision.
 This mtOtad rnn adopted a a joint ASTM-1F tumdard in 1964.
 I. Scope
    1.1 This method covers determination of the
 amount of dilution in crankcase oils of engines
 when gasoline has been used as the fuel.
    1.2 The values stated  in  inch-pound units
 are to be regarded as the standard.

 2. Applicable Documents
   2.1  A STM Standard:
   D 484   Specification for Hydrocarbon Dry-
     cleaning Solvents2

 3. Summary of Method
   3.1  The sample, mixed with water, is placed
 in a glass still provided with a reflux condenser
 discharging into a graduated  trap connected to
 the still. Heat  is applied, and the contents of
 the still are brought to  boiling. The diluent in
 the sample is vaporized with the water and then
 liquefied in the condenser. The diluent collects
 at the top of the trap, and the  excess water runs
 back  to the still  where it is  again vaporized,
 carrying over an additional quantity of diluent.
 The boiling is  continued  until  all the  diluent
 has been boiled out and recovered in the trap,
 and the volume is recorded.

 4. Significance
   4.1  Some fuel dilution of the engine oil may
 take place during normal operation. However,
 excessive fuel dilution is of concern in terms of
 possible performance problems.

 5.  Apparatus
   5.1  Flask, round-bottom type as described
in  the Annex.
   5.2 Condenser, Liebig straight-tube type, as
 described in the Annex.
   5.3 Trap, constructed in accordance with the
 requirements in Figs. 1 and 2 and in the Annex.
   5.4 HeaterAny suitable gas burner or elec-
 tric heater may  be used with the glass flask.
 (WarningHot  exposed surface. See Annex
 A2.1.)

 6. Procedure
   6.1  Mix the sample thoroughly, measure 25
 mL by means of a 25-mL graduated cylinder,
 and transfer as much as possible of the contents
 of the  cylinder by  pouring it into  the flask.
 Wash the graduated cylinder with successive
 portions of hot water until only a  negligible
 amount of oil is left in the cylinder. Add addi-
 tional water to the flask  to make a total of
 approximately 500 mL of water. Fill the trap
 with cold water and add  1 mL of ethanol to the
 water in the trap.
   6.2 Assemble the apparatus as shown in Fig.
 1, so that  the tip of the condenser is directly
 over the indentation in the trap.
   6.3 Apply heat (Warning! Hot exposed sur-
 face.  See Annex A2.1)  to  the flask at such a
 rate  that  refluxing starts within 7 to 10 min
 after heat is applied, with the water and sample
  1 This method is under the jurisdiction of ASTM Com-
mittee D-2 on Petroleum Products and Lubricants
  In the  IP.  this method is under the jurisdiction of the
Standardization Committee
  Current edition effective Aug. 29,1980 Published October
1980. Originally published as D 322 - 30. Laal previous edi-
tion D 322 -67 (1977).
  Annual Book of ASTM Standards. Part 23.
                                  [B-13]

-------
                                             D 322 -
                                                             23
being at 21 to 38C prior to application of heat.
After boiling  and  condensation  has  com-
menced, adjust the rate of boiling so that con-
densed distillate ia discharged from  the con-
denser at a rate of I to 3 drops per s.
  NOTE  IBumping with  tendency to froth over
is often experienced with dirty oils. The use of -boil-
ing stones,  steel wool, or about S mL of concentrated
hydrochloric acid (HO) in the flask is often helpful
in eliminating this difficulty.
  6.4 Obtain readings of the amount of diluent
at the following times, taken from the time that
refluxing starts: S, 15, and 30 min, and each IS
min following until the test is complete. Com-
pletion  of the test shall be determined on the
basis of either or both of the following criteria:
   6.4.1  The test is complete when the volume
of diluent increases by not more than 0.1 mL
in any 15-min  period during the course of the
test.
   6.4.2  The test is complete when the volume
of diluent obtained in a given time indicates
completion, as follows:
                  Tex it Complete if Apparent Vol-
                  ume of Diluent Collected it Equal
                         to or Less Than
Tune from Sun of
   Refluxing
     S min
    JO mm
    40 min
    90 min
                        DO visible diluent'
                        2.0 mL
                        4.0 mL
                        5.0 mL
   * Report is "no dilution"; otherwise the lest should be
 continued at least 30 nun
   6.5  When the test continues without reach-
 ing  the limit defined in 6.4.1,  to a point at
 which any of the conditions described in 6.4.2
 are encountered, the latter shall define the com-
 pletion of the test.
   6.6  When the test is complete by either of
 the criteria defined in 6.4.1 and 6.4.2, turn off
 the heat. Allow the equipment to stand at least
 30 min to allow the distillate to settle clear and
 to cool to approximately  room temperature.
 Read  the volume of diluent  collected in  the
 trap. If the volume of diluent exceeds the cali-
brated capacity of the trap, discontinue the test
and report the results as 20 % plus.

7. Calculations
  7.1 The diluent content of the sample, ex-
pressed as volume percent, is equal to the vol-
ume of diluent in millilitres multiplied  by 4.
  NOTE 2In some cases with samples containing
large amounts of diluent, equipment limitations do
not permit collection and measurement of the full 5
mL of diluent even when more is present. This con-
dition exists when the upper limit of the collected
diluent is  above  the zero calibration mark on the
trap  When it occurs, finish the test as prescribed in
6.6, read the maximum volume of diluent collected.
calculate the corresponding percentage "jr", and re-
port the results as "x percent plus."

S. Report
   8.1  Report the result as the Diluent Content.
ASTM D 322 - IP 23.

9. Precision
   9.1 The precision of the method as obtained
by statistical examination  of  interlaboratory
 test results is as follows:
   9.1.1 RepeatabilityThe difference  between
 successive test  results, obtained by the same
 operator with  the same apparatus under con-
 slant operating conditions of identical lest ma-
 terial, would in the long run, in the normal and
 correct operation of the test method, exceed the
 following value only in one case in twenty:
                 0.6 volume *
   9.1.2 ReproducibilityTbc  difference  be-
 tween two single and  independent results, ob-
 tained by different operators working in differ-
 ent laboratories on identical test material would
 in the long run,  in the normal and correct
 operation of the test method, exceed the follow-
 ing value only in one case in twenty.
                  1.4 volume %
                                       [B-14]

-------
                            0 322 -     23
                                              LIEIN DdlF-TIP
                                            J^CONOCNSC*
                                                 SETUP

HO. 1 AppantK for Delmtota Dtlurat ta Guolfaw Engine Crankcuc Oi
                       [B-15]

-------
                                             
-------
   APPENDIX C
EMISSIONS FACTORS

-------
                                  APPENDIX C
                               EMISSIONS FACTORS

    I.    Introduction
         The following emission factors and sample calculations are  included
    to form a basis for the verification of VOC emissions inventories developed
    from  emission source tests, plant site visits, permit applications, etc.
    These factors and procedures should not be applied in cases where site-
    specific data are available, but rather in instances where specific
    plant information is lacking or highly suspect.  (See Table C-l  for
    emission reductions for model plants.)

    II.  VOC Emission Factors for Existing Equipment

                         Range of emission rates       Nominal emission rates
                         (kg VOC per 100 kg dry        (kg VOC per 100 kg dry
Emission source          weight of articles cleaned)   weight of articles cleaned)
Dryer                              14-28                            18
Filter
  Diatomite                       (5-10)                             8
  Cartridge                      (0.5-1)                             1
Still                             (1-7)                              3
Fugitive sources                 (0.5-1)                             1
Total                           (15.5-46)                          22-30
Totals with
  Diatomite filter                                                  30
  Cartridge filter                                                  23
  Settling tanks                                                    22

-------
Table C-l.  NOMINAL ANNUAL VOC EMISSIONS FOR TWO MODEL PLANTS
    EMPLOYING EXISTING AND RACT EQUIPMENT AND PROCEDURES

Nominal emission factors
Type
of plant
Model plant I
with existing:
Diatomite filter
Cartridge filter
Settling tank
Model plant II
Plant
throughput,
kg/yr
(Ib/yr)
182,000
(400,000)



635,000
(1,400,000)
in kg VOC emitted
per 100 kg dry weight Nominal VOC emissions,
of articles cleaned kg/yr (Ib/yr)
Existing
equipment

30
23
22
22
RACT Existing
equipment equipment

5.8 55,000
(120,000)
5.8 42,000
(92,000)
4.8 40,000
(88,000)
4.8 139,700
(307,000)
RACT
equipment

10,600
(23,000)
10,600
(23,000)
8,700
(19,200)
30,500
(67,000)
Nominal annual VOC
emission reductions
resulting from
RACT implementation,
kg/yr (Ib/yr)

44,400
(98,000)
31.400
(69,000)
31,300
(68,800)
109,200
(240,000)

-------
    III.  VOC Emission Factors for RACT Equipment
Emission
source
Dryer
Diatomite
filter
Vacuum
still
Fugitive
Control
technique
Recovery
dryer
Cartridge
filter
No
change
Improved
operation
Range of emission rates
(kg VOC/100 kg dry weight
of articles cleaned)
0.7-9.5
0.5-1.0
1.0-7.0
0.5-1.0





                                                       Nominal emission rates
                                                       (kg VOC/100 kg dry weight
                                                       of articles cleaned)

                                                                 3.5
                                                                 1.0


                                                                 3.0


                                                                 1.0
                                                               7.5-8.5
Total Range
                             2.7-18.5
                                                                 8.5
                                                                 8.5
                                                                 7.5
           Totals with existing
             Diatomite filter
             Cartridge filter
             Settling tank


IV.   VOC Emission Factors as Applied to Model  Plants

  A.    Sample Calculation,  Model  Plant II

       1.    Existing Equipment
            a.    Total  Annual  Weight of Clothes  Cleaned

                   Number of        Average Weight
                 Washer Loads   X   of Clothing  Per   X  Typical  Operating   X
                    Per Day             Load              Days Per Week
                   Typical Operating   =
                    Weeks Per Year
                                         Average Weight
                                        Cleaned Per Year
              b.
                 (14) X  (180  kg) X (5 days/wk) X (52 wks/yr) = 635,000 kg/yr

                 Dryer Exhaust VOC Emissions
Standard Dryer
Emission Rate

    18 kg VOC
                                       Annual Weight
                                   of Clothes Cleaned
                                                              =  Weight of VOC
                                                               Emitted Per Year
                   100 kg Clothes Cleaned    A

                             114,300 kg VOC Per Year
                                              635,000 kg Clothes Cleaned
                                                    per year
C
                                    -

-------
     c.    Still  Waste  VOC  Emissions

          Vacuum Still   x     Annual  Weight  of   _   Weight  of  VOC
          Emission  Rate       Articles  Cleaned      Emitted Per Year

               3 kg VOC	  x   635,000  kg  Clothes =
          100  Kg Clothes Cleaned        Cleaned  Per Year

                    19,050 kg VOC  Per  Year


     d.    Fugitive  VOC Emissions

          Fugitive  Emission  y  Annual  Weight  of   _  Weight of  VOC
                Rate            Clothes Cleaned       Emitted Per Year

               1 kg VOC	  y    635,000 kg Clothes  =
          100  kg Clothes Cleaned         Cleaned Per Year

                    6,350  kg VOC Per Year


     e.    Total  Annual Plant VOC Emissions

            Annual  Emissions From  =     Total Annual Plant
                Above  Sources           VOC Emissions in  kg/yr

(114,300 kg) + (19,050 kg) + (6,350  kg) = 139,700  kg


2.    RACT Equipment (Model Plant  II)

     a.    Dryer Exhaust VOC Emissions

          RACT Dryer Control   Annual Weight of   =  Weight of VOC
            Emission Rate         Clothes Cleaned      Emitted Per Year

          1.   Recovery Dryer

                   3.5 kg  VOC	  x   635,000 kg Clothes  =
               100 kg Clothes Cleaned       Cleaned Per Year

                         22,225 kg VOC Per  Year


     b.    Still Waste VOC  Emissions

          RACT Still    y    Annual Weight of   _   Weight of VOC
          Emission Rate     Articles Cleaned       Emitted  Per Year

          	3  kg VOC	   y  635,000  kg Clothes
          100 kg Clothes Cleaned      Cleaned  Per year

                        19,050 kg  VOC Per Year
                           C-4

-------
             c.    Fugitive VOC Emissions

                  Fugitive Emission x   Annual  Weight of  _  Weight of VOC
                       Rate             Clothes Cleaned      Emitted Per Year
VQC
                                          x   635,000 kg Clothes
                               _
                  100 kg Clothes Cleaned      Cleaned Per Year

                                 6,350 kg VOC Per Year

             d.    Total  Annual  Plant VOC Emissions

                    Annual  Emissions From  =   Total  Annual  VOC
                        Above  Sources           Emissions  in kg/yr

             22,225  Kg   +   19,050  Kg  +   6,350 Kg = 47,625 Kg
               yr           yr           yr         yr

   B.    Plant VOC  Emission  Reduction Efficiency,  Model Plant II

        1.   Total Annual Plant  VOC  Emission  Reduction

            Total Annual Emissions    -   Total  Annual  Emissions
            From  Existing  Equipment      From RACT  Equipment

                           Total Annual  Emission  Reduction

            139,700  kg VOC  -   47,625 kg  VOC   =     92,075  kg  VOC
               per year           per year           per year

       2.   Percent  Reduction in Total Plant  VOC  Emissions

               Total Annual     _._   Total Annual  Emissions   =
            Emission Reduction  '    From Existing  Equipment

                      Percent Reduction in Total  Plant Emissions

            92,075 kg VOC    T     139,700 kg VOC  =  66%
              per year                per year


V.   Solvent Recovery, Model Plant II

  A.    Solvent  Not Emitted  = Solvent Recovered

  B.    Credit for Solvent Recovery

          Total  Annual         Cost of     Credit for Recovery
       Emission  Reduction  X  Solvent  =  of  Reduced Emissions
         in kg  VOC            Per kg      in  Dollars Per Year

       92,075 kg VOC      X   $0.53 Per   =  $48,800.00  Per Year
         per year              kg Solvent

                                  C-5

-------
       APPENDIX D

COMMENTS RECEIVED ON THE
NOVEMBER, 1981, DRAFT CTG

-------
    VICTOR ATIYEH
      OOVERNOft
Department of Environmental Quality

522 S.W. 5th AVENUE, BOX 1760, PORTLAND, OREGON 97207

                                            December 29,  1981
             United States Environmental Protection Agency
             Attn:  Fred L.  Porter
             Emission Standards and Engineering Division (MD-13)
             Research Triangle Park, N.C. 27711
                                               Re:   CTG  Large Petroleum Dry Cleaners
             Gentlemen:
             Thank you  for sending us a review copy of EPA's  "Control of Volatile
             Organic Emissions from Large Petroleum Dry Cleaners", Draft CTG,
             November 1981.

             We  strongly believe that the CTG should have included a model rule
             Equipment manufacturers are now facing a score of different state rules,
             in  addition to many local air program  control rules.  A model rule would
             have  given some commonality for the natural diversity that will likely
             follow.  Even Regional EPA reviewers may have a difficult time reviewing
             state  rules without an EPA model rule.

             Enclosed is the South Coast Air Quality Management District's (Los Angeles)
             Rule 1102  which I hope you will add to your guideline document.   I would
             tend to follow their lead,  lacking your specific guidance.

                                                         Sincerely,
            PBB:h
                                                         Peter B.  Bosserman
                                                         Sr.  Environmental Engineer
            cc:  Rentex,  R.  Barnard
                 EPA ^ Porter  ("2nd copy)
DEQ-1

-------
Rule 1102.    Petroleum Solvent Dry Cleaners                       (Adopted January 6, 1978)
                                          (Amended August 3, 1979) (Amended July 11, 1980)
    (a)  Effective March 7, 1978 a person shall not operate any dry  cleaning equipment, which
         uses petroleum-based solvent unless:
         (1)  There is no liquid leaking from any portion of the equipment.
         (2)  All washer lint  traps, button traps, access doors and other parts of the equipment
             where solvent may  be exposed to the atmosphere are kept closed at all times except
             when required for proper operation or maintenance.
         (3)  The still residue is stored in sealed containers.
         (4)  The used filtering material is put into a sealed container immediately after removal
             from the filter or the dry cleaning system is equipped with one of the following filtering
             systems:
             (A)  Cartridge filters containing paper  or carbon or a combination  thereof,  which
                  are fully drained in the filter housing for at least 12 hours before removal.
             (B)  Diatomaceous  earth  filtering  system,  connected  to  a  centrifugal solvent
                  extractor or other device capable of removing sufficient  solvent so that the
                  remaining  diatomaceous  earth  and  soil does  not contain more  than 0.4
                  kilogram of solvent per kilogram of filter powder and soil removed (0.4 pounds
                  per pound).
             (C)  Any other type of filtering system or process found by the Executive Officer to
                  be equally effective.
    (b)  A person shall not  operate  any  dry cleaning  equipment  which uses petroleum-based
        solvent  unless all exhaust gases from drying  tumblers and cabinets are vented through a
        carbon  adsorber or other control device, or other methods are used to reduce the total
        emissions of hydrocarbon vapors to the atmosphere by at least  90 percent by weight.
    (c)  The provisions of Subsection (b) shall become effective in accordance with the following
        compliance schedule:
        (1)  Effective January I, 1982, all petroleum solvent dry cleaning plants consuming more
             than 50,000 liters (13,209 gallons) of solvent per year shall comply with the provisions
             of section (b).
        (2)  Effective July 1, 1983, all petroleum solvent dry cleaning plants consuming more than
             25,200 liters (6,657 gallons) of solvent per year shall comply with the provisions of
             section (b).
        (3)  Effective January 6, 1985 all petroleum  solvent dry cleaning plants consuming more
                                         [D-2]                             Reg. XI-Page 1

-------
Rule 1102  (Cont'd.1                                                (Amended July 11, 1980)

              than  10,000  liters  (2,642 gallons) of  solvent per  year shall comply with the
              provisions of section (b).
         (4)   The solvent consumed  by a petroleum solvent dry cleaning plant in a year, means
              the amount of solvent purchased for that year.
    (d)  Increments of Progress
         In order to conform with the compliance dates specified in Subsection (c), an owner or
         operator  of petroleum  solvent dry cleaning equipment  shall comply with the following
         increments of progress schedule:
         (1)   12  months prior  to  the  effective dates, submit to the Executive  Officer an
              application for Permit to Construct, describing at a minimum, the steps that will be
              taken to  achieve compliance with the provision of Subsection (b) of this rule. (2)
        (2)  10 months prior to the effective dates,  award the contract for the emission control
              system, or issue purchase orders  for the component parts to accomplish emission
              control.
        (3)  8 months prior to  the effective dates, initiate on-site construction or installation of
             equipment to reduce or control emissions.
        (4)  Upon the effective dates, complete on-site construction or installation of equipment
             to reduce or  control emissions, and assure final  compliance with the provisions of
             Subsection (b) of this rule.
                                         [D-3]

-------
                       UNITED STATE- cNVIRONMENTAL PROTECTION AGE.^'Y
   DATE       .                               Region H
              1 8 JAN 1S32

SUBJECT      Region II Comments on Draft Control Techniques Guidelines (CTG) Documents
   FROM      Jan N.
             Air & Waste Managin/ent Division
     TO
             Fred L. Porter, Assistant to the Director
             Emission Standards and Engineering Division (MD-13)

             The following are  comments on  two  draft  CTG  documents which  are  being
             circulated for review:

             A-    Control of  Fugitive  VOC  Emissions  from Synthetic  Organic  Chemical.
                  Polymer and Resin Manufacturing Equipment                	

                  1.    The method used to  correct a leak and the time required to do so, is
                       dependent upon  the type of component that is leaking.   A leak in a
                       packed seal may be  repaired with  much less  difficulty than one in a
                       mechanical seal.   The action level and achievable  emission reduction
                       should reflect these differences.  While an action level of 10,000 ppmv
                       can be acceptable for a repair that can be completed in an hour, a
                       higher  action  level  can be acceptable for repairs that require  the
                       dismantling  of  a  pump or  a  compressor.   The  required emission
                       reduction should also be more  substantial  for the  more complicated
                       repair.  This would assure that the proper repair (and not an ineffective
                       temporary repair) is being made on the leaking component.

                  2.    An exemption  is recommended  for  small processes  with  less than 100
                       valves  in gas  or light liquid service.   Due to the large number of
                       processes  that  will  be covered  by these  regulations  such a  broad
                       exemption  is unwise.   In  instances where highly toxic or  odorous
                       chemicals are involved this exemption should not be applicable.

            B-   Control of VOC Emissions from Large Petroleum Dry  Cleaners

                 1.    Is the reduction in or  the reformulation  of the aromatics content of
                       petroleum solvents a likely method of reducing  emissions? If so, is such
                       a   reduction  in   aromatics  content  effective  when  there  is a
                       corresponding reduction in solvent  cleaning efficiency?

                 2.    Shouldn't  ranges  of   "acceptable"   drying   time  and   condenser
                       temperature/flow (vapor and liquid) found during EPA test programs be
                       matched up against load weight  and fabric type in the recovery dryer?
                       These factors  would allow for  a minimum solvent content in  dried
                      articles resulting in a maximum solvent recovery and a corresponding
                       maximum VOC emission reduction.
                                               [D-5]
EPA Form 1320-6 (Rev. 3-76)

-------
                             -  2  -

3.    EPA  wished to  evaluate  the relationship between the  flow rate of
     recovered solvent and the dryer emissions measured during the exhaust
     cycle;  however,  if  difficulties  associated  with  measuring   high
     concentrations  (false low readings,  saturated instruments,  etc.) of
     solvent in a vapor stream  cast doubts on the validity  of the obtained
     data, then why not analyze type of fabrics, load weights and recovery
     durations? It should be noted that a control option such as the cartridge
     filtration system is "sized" by its manufacturer, based on the dry-weight
     load capacity of the existing washer.
                               [D-6]

-------
JAMES R. PATTON, JR,
GEORGE BLOW
CHARLES OWEN VERRILL, JR.
JOSEPH L BRAND
THOMAS HALE 00GGS, JR.
TIMOTHY J. MAY
HARRY A. INMAN
EDWARD T. MITCHELL
ELLIOT H. COLE
J.GORDON ARBUCKLE
WILLIAM C. FOSTER
DAVID C. TODD
LAWRENCE G.MEYER
RICHARD A. EARLE
ERNEST S. CHRISTIAN, JR.
ROBERT H.NOEHLER
E.BRUCE BUTLER
DAVID B.ROBINSON
JOHN H.VOGEL
ALLAN ABBOT TUTTLE
BART S. FISHER
JAMESG. O'HARA
JOHN L.OBERDORFER
LINDA ELIZABETH BUCK
LAN NY J. DAVIS
DOMENICO DE SOLE
TIMOTHY A. VANDERVER.JR.
CHARLES B. TEMKIN
WILLIAM J. COLLEY
JOHN F, WOODS
TIMOTHY A.CHORBA
RONALD H. BROWN
     PATTON, BOGGS  & BLOW
         255O M STREET, N. W.
       WASHINGTON, D. C. 2OO37
                 4-57-6000
  CABLE: BARPAT
TELECOPIER: 457-6315
      January 15,  1982
          or counsel
        WILLIAM D. HATHAWAY
         CHARLES D COOK

         WRITERS NRCCT WAI
         457-6075
'V.
"*1 "3 "7
^ -3 /


WU TELEX: 89-452
ITT TELEX- 44O324














DONALD A LOFTY
SHAOUL ASIAN
MIDDLF.TON A. MARTIN
RICHARD S CHARIN
GARRET G. RASMUSSEN
JAMES B. CHRISTIAN, jR
DAVID E DUNN
RICHARD J CONWAY
STEVEN H SCHNEE8AUM
LEE H.GOODWIN
GREGORY K. PILKINGTON
MATTHEW J. ABRAMS
DON A. ALLEN
DUANE A SILER
GARY L STANLEY
RICHARD M.STOLBACH
PETER J. P. BRICKFIELD
JOE ROBERT REEDER
KATHARINE R. BOYCE
SCOTT NASON STONE
PETER J.WIEDENBECK
THOMAS D.ROBERTS
GEORGE M. BORABA8Y
JEANNE M ROSLANOWI
STUART M. PAPE
JEFFREY T. SMITH
RONALD K, HENRY
GLENN R. THOMSON
RICHARD J PARRlNO
FRANK R. SAMOLIS
THOMAS R.GRAHAM
GEORGE J. SCHUTZER
MICHAEL D. ESCH
RONALD A. MILZER
RALPH G. STEINHARDT
JOHN C. HARRISON
ANDREW S. NEWMAN
CLAUDIA L.DEERING
HARVEY J, BAKER *
FRANK J. DONATELL1 *
FLORENCE w PRIOLEAI
JONATHAN C. CARLSOh
CLIFF MA55A UJ

                                                                       HOTAOKfTTEDINtXC
         Emission Standards  & Engineering
           Division (MD-13)
         Environmental Protection Agency
         Research Triangle Park, North  Carolina
                                   27711

              Attention:  Mr.  Fred Porter

         Dear Mr.  Porter:

              This responds  to your letter of December 1,  1981
         seeking  comments on the draft  document,  Control of
         Volatile  Organic Emissions from Large Petroleum
         Dry Cleaners.  These  comments  are submitted on
         behalf of the dry cleaning industry and  focus
         on the critical, unresolved issues presented
         by this draft.   To  some degree,  these comments are
         repetitive of previous comments  on behalf  of the
         industry.   Attached for your information are such
         comments  by Mr. William Fisher  of the International
         Fabricare Institute,  Mr.  Mervyn  Sluizer of the
         Institute of  Industrial Launderers, and me,  to the
         National  Air  Pollution Techniques Advisory
         Commission last March.

              There  are  a great number of  detailed,  technical
         comments  that could be made on  the draft document.
         However,  we believe that  the need for these  comments
         would be  obviated by  our  recommendations that EPA
         revise the  draft document (1)  to  include a model
         regulation  (as  was done in the previous draft of
         this  document);  (2)  to  include an exemption  level in  the
                                       [D-7]

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

     Emission  Standards  & Engineering Division
     Page  Two
     January 15, 1982

     model regulation  (as was done in the previous draft)
     at  the exemption  level established in that draft;
     and (3) to revise the treatment of the issue of
     recovery  dryer safety, including appropriately
     recognizing this  issue in the model regulation.  We
     believe that adoption of these recommendations will
     assure environmental improvement in a rational and
     cost-effective manner.  Further, adoption of these
     recommended actions will provide adequate guidance
     to  the states, which, in our view, the present draft
     does  not.  In the event that EPA does not accept
     these recommendations, however, we hereby request the
     opportunity to present our more detailed comments,
     mentioned above.

          In our view, the most serious flaw in the draft
     document  is the absence of a model regulation setting
     forth suggested requirements to impose on regulated
     facilities.  We do not suggest that a model regulation
     would be  appropriate or necessary for each guidance
     document  prepared by EPA.  Instead, our views are
     premised  solely on  the unique situation facing EPA,
     the states and the  dry cleaning industry in developing
     SIP provisions applicable to petroleum dry cleaners.
     First, the requirement that SIP revisions be submitted
     to  EPA by next spring creates time pressures which makes
     some  guidance of  this type a virtual necessity.   Second,
     the combined impact of this time pressure and the
     need  to develop regulatory provisions for a variety
     of  industries will make it difficult, if not impossible,
     for at least some states to obtain the necessary  knowledge
     and then  draft regulations in the absence of any  guidance
     as  to the possible  content of such regulations.   Finally,
     the dry cleaning  industry is composed primarily of
     very  small operations.  These companies have neither  the
     time  nor  the ability to assist the states in an extensive
     regulatory development process.  Accordingly, we  believe
     that  including a  model regulation in the guidance document
     will  be in the best interests of EPA, the states  and  the
     dry cleaning industry.

          The  elimination of the model regulations from  this
     draft document is all the more surprising because there is
     no  support in the record for such action.  The  speakers at
     the NAPTAC meeting  on the previous draft of this  document
     favored the inclusion of a model regulation.  The only
     objection focussed  on EPA's efforts to use model  regulations
     as  presumptive norms.  This objection can be met  by
                                  [D-8]

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

     Emission Standards  & Engineering Division
     Page Three
     January 15,  1982

     language in  the preamble  such  as the  following:

                 "This document  includes a model
          regulation which  is  provided solely as
          guidance  to assist state  and local agencies
          in drafting their own  specific RACT regulations.
          It is not a standard by which state and  local
          RACT regulations  will  be  judged  and consequently
          is only illustrative in nature.   Therefore, it
          is not to be construed as rule making by EPA."

          Since, in our view,  the record is bereft of support
     for elimination of a model  regulation, such an exclusion
     would be  arbitary and  capricious.  Accordingly, we
     urge that a model regulation be  included in the final
     guidance document in substantially the same form as that
     in the  previous draft  of  this document.  That model should
     be modified in accordance with the discussion in the
     attached comments and  the discussion below.

         With respect to the materials included in the draft
    document, we note that it fails to point out the unique
    economic circumstances of the drycleaning industry.  Although
    this shortcoming is  ameliorated somewhat by the draft's
    focus on large petroleum drycleaners,  a casual perusal
    of the cost  analysis (Chapter 5)  would suggest that the
    drycleaning  industry has an  unlimited  financial capability
    to convert to whatever technology is ultimately defined
    as RACT.  This  is  simply not correct,  as EPA has recognized
    on other occasions.   The October  23, 1981  draft New Source
    Performance  Standard for petroleum drycleaners and  the
    Background Information Document supporting  the draft NSPS
    properly recognize the unique economic circumstances
    of the industry.

         In  order to correct this shortcoming  in  the draft
    document,  an  exemption  (based on  solvent consumption)
    from  the requirement  that  a  recovery dryer  be  installed
    should be included in the  model regulation.   This approach
    was adopted in  the model regulation included  in  the
    previous draft  document and  should be  followed  in the
    final  guidance  document.   Further,  the exemption  threshold
    should be  set at the  same  level as in  the previous draft.

        Another critical issue  concerns the safety of recovery
    dryers.  At the outset, we emphasize our hope that recovery
    dryers prove completely safe.   In addition to enabling  the
    industry to reduce its  emissions,  the  recovery dryer can
    contribute to the financial  well-being of individual dry-


                                [D-9]

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

     Emission Standards  &  Engineering Division
     Page Four
     January 15,  1982

     cleaners.  As  a result,  it is apparent that this is
     the type of technology that the industry would favor
     if issues relating  to safety can be resolved.

          Despite the foregoing, it is  not the role of the
     industry or of its  representatives to ascertain the safety
     of recovery dryers.  Nor,  as EPA has recognized, is this
     a proper function of  the Agency;

                 "It is  EPA's position  that decisions
          on safety are  entirely in the hands of safety
          officials and  insurers, and EPA does not
          intend to require a device that is not fully
          accepted."  October 23, 1981  draft NSPS,
          page 16.

          We find the discussion of recovery dryer safety
      ( 3.2 of the  draft document) inadequate.  It first
     notes that EPA's test equipment has been incapable
     of discerning  whether the lower explosion limit  (LEL)
     is reached or  exceeded in the operation of this device.
     It then states that two recovery dryer explosions have
     been documented and attempts to minimize the significance
     of these explosions.   This discussion improperly
     characterizes  the risk involved since industry
     representatives have presented EPA with data concerning
     at least six explosions involving recovery dryers.
     The draft document next attempts to bolster this handling
     of the safety issue by referring to an analysis of the
     Japanese experience with recovery dryers.  We have
     previously discussed the Japanese experience with EPA
     and its contractor and merely reiterate that this
     experience is largely inapplicable to the present
     situation.

          Finally,  the discussion of recovery dryer  safety
     in the draft document concludes with an anecdotal
     survey which seeks to demonstrate that neither  fire
     safety officials or insurance companies have problems
     with recovery dryers.  Obviously, such a survey cannot
     determine the reactions of  all fire  safety officials
     or  insurance companies to  the recovery dryer safety
      issue.   Therefore, we suggest that  adequate recognition
      be  accorded to  those  fire  safety  officials and  insurance
      companies who may  not believe that  these devices are
      safe.   The most appropriate manner  of  doing this is
      to  include  in the  model regulation  a recommendation
      to  consult with fire  safety officials  and  insurance
                                  [D-10]

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

     Emission  Standards  &  Engineerning Division
     Page  Five
     January 15, 1982

     companies in developing  SIP  revisions.   In order  to
     accomplish this,  language  such  as the  following should
     be  inserted in  the  model regulations:

                Insert  the following  language  in  the  preamble
                to  the  model regulation:

                      "Concern  has been expressed  that the
                requirement  for  installation and  operation
                of  solvent recoverydryers could conflict
                with  local fire  safety codes or creates
                situations in which petroleum  dry cleaning
                plants  could not be insured.   Even though
                Factory Mutual has  certified the  one  U.S.-
                made  recovery dryer,  agencies  should  ascertain
                whether this is sufficient in  their state or
                locality.  In order to do this appropriate fire
                safety  officials and  insurance industry
                representatives should be involved early in
                the process  of developing the  RACT regulations.
                In the event that such obstacles  cannot be
                resolved, the recovery dryer requirements should
                be deleted."

         We also suggest that the following language be added
    in the body of the model regulation to provide a limited
    exemption  for  situations where either fire safety officials
    or insurance companies do not feel the device to be safe:

                Add a new provision in the model regulation
                as follows;

                     "The District may exclude plants  from
                the scope of  the  recovery dryer requirements if
                the owner or  operator  can demonstrate  that
                installation  of a solvent recovery dryer
                would not be  in conformity with applicable fire
                safety regulations or  would cause a situation
                in  which obtaining fire insurance would be
                infeasible."

    Although we  understand that EPA does  not favor the addition
    of such language,  we suggest  that  it  is warranted  in  order
    to deal wth  those  rare situations  in  which  a problem  exists.
    The  foregoing  language has  been  carefully drafted  to  narrowly
    limit  the  possible applicability of this  exemption.

        If the draft  document  is revised in  accordance with
    the  foregoing recommendations, we  believe that it

                               [D-ll]

-------
PATTON. BOGGS & BLOW
     Emission  Standards  &  Engineering  Division
     Page  Six
     January 15, 1982

     will  be a significant step  toward the  objective  of
     reducing emissions  in a cost-effective and  rational
     fashion.  If there  are any  questions about  the fore-
     going comments, please do not  hesitate to contact me

                                    Sincerely,
                                    Timothy A.  Vanderver,  Jr,

    TAV/llr
                                 [D-12]

-------
Chevron    .               .  ^
           Chevron Research Company
           A Standard Oil Company of California Subsidiary
           576 Standard Avenue, Richmond, CA 94802


R.G.Andenon                                         Mawh 1   1QP9
Vice-President                                          ndrcn I,  1982
          Mr. James  F. Durham
          Emission Standards and Engineering Division
          U. S. Environmental Protection Agency
          Research Triangle Park, N.C. 27711

          Dear Mr. Durham:

          We have reviewed the EPA Control Technique Guideline for  "Control  of
          Volatile Organic Emissions from Large Petroleum Dry Cleaners."   In
          a telephone conversation with Mr. W. H. El 11s of our company, on
          February 26, 1982, Mr. Robert Walsh asked that our comments be
          referred to you.

          We have detected a problem with Table 2-1, which tabulates physical
          properties of dry cleaning solvents.  One of the products listed 1s
          Chevron 450 solvent, which 1s a highly purified kerosene hot normally
          used for dry cleaning.  We believe a better choice would be Chevron
          thinner 410B, which 1s a 140F material.

          Chevron 325 solvent 1s our basic Stoddard solvent; 1t 1s also listed
          1n the table.

          Enclosed are current data  sheets for Chevron 325 solvent and Chevron
          thinner 410B.

          If you have any questions,  please phone Mr.  William H.  ElHs at (213)
          615-5212.

                                                 Yours truly,

                                                 R.G.  ANDERSON
                                                 by WILLIAM H. ELLIS
         WHEH1s:mp
                                       [n-13]

-------
Chevron    Ml       _       ...
          Chevron Research Company
          A Standard Oil Company of California Subsidiary
          324 El Segundo Boulevard, El Segundo, CA 90245
                              CHEVRON  325  SOLVENT
                               Typical  Properties


                Gravity, ftAPI                                  47.7
                Gravity, Specific at 60F                       0.7896
                Pounds Per Gallon at 60F                       6.57
                Flash Point TCC F                             105
                Flash Point TOC F                             118
                Aniline Point F                               145.2
                Kauri Butanol Value                             35.9
                Reid Vapor Pressure, Lbs.                        0.1
                Threshold Limit Value, ppm                     250

                Explosive Limits, ~|   lower                      1.0
                Volume % 1n Air   J~  upper                      6.0

                Composition - Volume %
                  Benzene                                        0.02
                  Toluene/Ethyl benzene                           0
                  Xylene & CQ + Aromatics                    0.0/2.4
                  Naphthenes                                    62.2
                  Paraffins                                     35.4
                Color Saybolt                                  +30

                Distillation, D-86, F
                  Initial Boiling Point                        320
                  10% Recovered                                327
                  50%     "                                    335
                  70%     "                                    341
                  90%     "                                    352
                  Dry Point                                    368
                  End Point                                    370

                Spontaneous Ignition Temperature F            500
                Freezing Point F                -            -100
                Molecular Weight, Average                      138
                Solubility Parameter                             7.5
                Refractive Index 20C                            1.4340
                Thermal  Conductivity, 60F
                  BTU/HR/FT/Deg F                                0.086
                Heat of  Vaporization, BTU/Lb.                  119
                Heat of  Combustion. BTU/Lb.           "       18720
                                                              2-82
                                    [0-14]

-------
Chevron    chevron Research Company
          A Standard Oil Company of California Subsidiary
          324 El Segundo Boulevard, El Segundo, CA 90245
                               CHEVRON THINNER 41OB
                                Typical Properties


                Gravity,  *API                                   42.3
                Gravity,  Specific at 60F                       0.8140
                Pounds  Per Gallon at 60F                       6.78
                Flash Point TCC F                            142
                Flash Point TOC F                            158
                Aniline Point  F                              147.4
                Kauri Butanol  Value                            35.8
                Reid Vapor Pressure, Lbs.                     ^0.1
                Threshold Limit Value, ppm                    275

                Explosive Limits, "I   lower                     0.9
                Volume  % 1n Air   J   upper                     4.9

                Composition -  Volume %
                  Benzene                                    <  0.02
                  Toluene/Ethyl benzene                          o
                  Xylene & Cg  + Aromatics                  0.0/2.8
                  Naphthenes                                   60.6
                  Paraffins                                    36.6
                Color Saybolt                                  +30

                Distillation,  D-86, F
                  Initial Boiling Point                       370
                  10% Recovered                               378
                  50%      "                                   383
                  70%      "                                   387
                  90%      "                                   395
                  Dry Point                                   406
                  End Point                                   409

                Spontaneous Ignition Temperature F           490
                Freezing Point F                             -69
                Molecular Weight, Average                     157
                Solubility Parameter                            7.2
                Refractive Index 20C                           1.4456
                Thermal  Conductivity, 60F
                  BTU/HR/FT/Deg F      *'                       0.084
                Heat of Vaporization, BTU/Lb.                 112
                Heat of Combustion, BTU/Lb.         -      18660--
                                                              2-82
                                     [D-15]

-------
    APPENDIX E
EXAMPLE REGULATION

-------
                               APPENDIX  E
              EXAMPLE  REGULATION  FOR THE  CONTROL OF VOLATILE
      ORGANIC COMPOUND  EMISSIONS  FROM  LARGE PETROLEUM DRY CLEANERS

     The petroleum dry  cleaning  industry and the State of Oregon recommended
that a model  or example regulation be  included in the control technique
guideline (CTG) document.  While  EPA did not include an example regulation
in the draft  CTG, industry and State recommendations warrant inclusion
of an example regulation, which has been prepared and is presented in
this appendix.
     The example regulation is based on a "presumptive norm" which is
considered broadly representative of RACT for the petroleum dry cleaning
industry.  The example regulation is included solely as guidance to
assist State  and local agencies in drafting their own specific RACT
regulations.
     The example regulation should not be interpreted or construed as
being the only regulation for petroleum dry cleaning that will be acceptable
to EPA as a part of the State Implementation Plan.   Other regulations
that can be demonstrated to represent reasonably available control
technology (RACT) would be equally acceptable.
     Industry representatives have expressed concern that the requirement
for installation and operation of solvent recovery dryers could conflict
with local  fire safety codes or create situations in which petroleum dry
cleaning plants could not be insured.   Even though Factory Mutual  has
certified the one U.S.-made recovery dryer,  agencies should ascertain
whether this is sufficient in their state or locality.   Although EPA has
found no instances in which a fire marshal  or insurance underwriter
prohibited  the installation of a solvent recovery dryer,  appropriate
fire safety officials and insurance industry representatives should be

-------
involved early in the process of developing the RACT regulations.   In
the event that such obstacles cannot be resolved,  the recovery dryer
requirements should be deleted.
EXAMPLE REGULATION
XX.010  Applicability.
     (A)  This Regulation applies to petroleum solvent washers, dryers,
solvent filters, settling tanks, vacuum stills, and other containers and
conveyors of petroleum solvent that are used in petroleum solvent dry
cleaning facilities.
     (B)  This Regulation applies to all petroleum solvent dry cleaning
facilities described in XX.010(A) that consume 123,000 liters or more
of petroleum solvent annually.
     (C)  This Regulation applies to all petroleum solvent dry cleaning
facilities described in XX.010(A) and (B) located in the following
areas:

XX.020  Definitions.
     (A)  Except as otherwise required by the context, terms used in
this Regulation are defined in the General Statutes, the General Provisions,
or in this section as follows:
     "Cartridge filter" means perforated cannisters containing filtration
paper and/or activated carbon that are used in a pressurized system to
remove solid particles and fugitive dyes from soil-laden solvent.
     "Containers and conveyors of solvent" means piping, ductwork,
pumps, storage tanks, and other ancillary equipment that are associated
with the installation and operation of washers, dryers, filters, stills,
and settling tanks.
     "Dry cleaning" means a process for the cleaning of textiles and
fabric products in which articles are washed in a nonaqueous solution
(solvent) and then dried by exposure to a heated air stream.
     "Perceptible leaks" means any petroleum solvent vapor or liquid
leaks that are conspicuous from visual observation; such as pools or
droplets of liquid, or buckets or barrels of solvent or solvent-laden
waste standing open to the atmosphere.

                                 E-2

-------
     "Petroleum solvent" means organic material  produced by petroleum
distillation comprising a hydrocarbon range of 8 to 12 carbon atoms per
organic molecule that exists as a liquid under standard conditions.
     "Solvent recovery dryer" means a class of dry cleaning dryers that
employs a condenser to liquify and recover solvent vapors evaporated in
a closed-loop, recirculating stream of heated air.
     "Volatile organic compounds" means any organic compound that
participates in atmospheric photochemical reactions or is measured by a
State or EPA test method.
XX.030  Standards.
     (A)  Each owner or operator of a petroleum solvent dry cleaning
dryer shall either:
          (1)  Limit VOC emissions to the atmosphere to an average of
3.5 kilograms of volatile organic compounds per 100 kilograms dry weight
of articles dry cleaned, or
          (2)  Install and operate a solvent recovery dryer in a manner
such that the dryer remains closed and the recovery phase continues
until a final recovered solvent flow rate of 50 milliliters per minute
is attained.
     (B)  Each owner or operator of a petroleum solvent filtration
system shall either:
          (1)  Reduce the volatile organic compound content in all
filtration wastes to 1.0 kilogram or less per 100 kilograms dry weight
of articles dry cleaned, before disposal, and exposure to the atmosphere,
or
          (2)  Install and operate a cartridge filtration system, and
drain the filter cartridges in their sealed housings for 8 hours or more
before their removal.
     (C)  Each owner or operator shall repair all petroleum solvent
vapor and liquid leaks within 3 working days after identifying the
sources of the leaks.  If necessary repair parts are not on hand, the
owner or operator shall order these parts within 3 working days, and
repair the  leaks no  later than 3 working days following the arrival of
the necessary parts.
                                 E-3

-------
XX.040  Testing and monitoring.
     (A)  To be in compliance with XX.030(A)(1) any person shall:
          (1)  Calculate, record, and report to the Director the weight
of volatile organic compounds vented from the dryer emission control
device calculated by using EPA Reference Test (40 CFR,  Part 60) Methods 1,
2, and 25A, with the following specifications:
               (a)  Field calibration of the flame ionization analyzer
with propane standards, and
               (b)  Laboratory determination of the ratio of the flame
ionization analyzer response to a given parts per million by volume
concentration of propane to the response to the same parts per million
concentration of the volatile organic compounds to be measured, and
               (c)  Determination of the weight of volatile organic
compounds vented to the atmosphere by:
                    (i)  the multiplication of the ratio determined in
XX.040 (A)(l)(b) by the measured concentration of volatile organic
compound gas (as propane) as indicated by the flame ionization analyzer
response output record, and
                   (ii)  the conversion of the parts per million by
volume value calculated in XX.040 (A)(l)(c)(i) into a mass concentration
value for the volatile organic compounds present, and
                  (iii)  multiply the mass concentration value calculated
in XX.040 (A)(l)(c)(ii) by the exhaust flow rate determined by using
EPA Reference Test Methods I and 2.
           (2)  Calculate, record, and report to the Director the dry
weight of  articles dry cleaned.
           (3)  Repeat XX.040 (A)(l) and (2) for normal operating
conditions that encompass at least 30 dryer loads, which total not less
than 1,800 kg dry weight, and represent a normal range of variations  in
fabrics, solvents,  load weights, temperatures,  flow rates, and process
deviations.
     (B)   To determine compliance with XX.030  (A)(2), the owner or
operator shall verify  that the flow  rate of recovered  solvent  from the
solvent recovery  dryer at the termination of the recovery phase  is no
                                  E-4

-------
greater than 50 millilHers per minute.  This one-time procedure shall
be conducted for a duration of no less than two weeks during which no
less than 50 percent of the dryer loads shall be monitored for their
final recovered solvent flow rate.  The suggested point for measuring
the flow rate of recovered solvent is from the solvent-water separator.
Near the end of the recovery cycle, the flow of recovered solvent should
be diverted to a graduated cylinder.   The cycle should continue until
the minimum flow of solvent is 50 milliliters per minute.   The type of
articles cleaned and the total length of the cycle should then be recorded.
     (C)  To be in compliance with XX.030 (B)(l) any person shall:
          (1)  Calculate, record, and report to the Director the weight
of volatile organic compounds contained in each of at least five
one-kilogram samples of filtration waste material taken at intervals of
at least 1 week, by employing ASTM Method 0322-80 (Standard Test Method
for Gasoline Dilutent in Used Gasoline Engine Oils by Distillation).
          (2)  Calculate, record, and report to the Director the total
dry weight of articles dry cleaned during the intervals between removal
of filtration waste samples,  as well  as the total mass of filtration
waste produced in the same period.
          (3)  Calculate, record, and report to the Director the weight
of volatile organic compounds contained in filtration waste material  per
100 kilograms dry weight of articles  dry cleaned.
     (D)  Compliance with XX.030(C)  requires that each owner or operator
make weekly inspections of washers, dryers, solvent filters,  settling
tanks,  vacuum stills,, and all containers and conveyors of petroleum
solvent to identify perceptable volatile organic compound vapor or
liquid leaks.
     (E)  To be in compliance with XX.030 any person can use an equivalent
test procedure or method provided that this method or procedure has  been
previously approved by the Director.
XX.050  Compliance schedules.
     (A)  The owner or operator of a  petroleum solvent dry cleaning
facility subject to this regulation shall  meet the applicable stages  of
progress contained in the following schedule:
                                 E-5

-------
          (1)  Submit to the Director final plans for the emission
control equipment 	 (3 months after implementation of the
regulation).
          (2)  Award contracts for the emission control equipment
	 (2 months after the final submission of the control equipment
plans).
          (3)  Complete onsite construction or installation of the
emission control equipment 	 (12 months after award of the
contract).
          (4)  Achieve final compliance with the regulation 	
(2 months after installing the control equipment).
          (5)  In the event that the control equipment cannot be delivered
within 12 months after award of the contract, and the owner or operator
placed the order within the required time, the final compliance date
shall be 3 months following delivery of the equipment.
DISCUSSION
     Enforcement and compliance verification procedures associated with
the model regulation would rely on visual inspection of all affected
facility components.  This section includes a discussion of the enforce-
ment approaches for a VOC emissions regulation based on RACT.
     The determination of the annual solvent consumption of individual
petroleum dry cleaning plants should be based on the plants' records of
total solvent purchases over the latest 12 month period.  Records of
these purchases should be readily available at the  individual plant, and
questionable or incomplete solvent purchase data could be verified
through solvent suppliers' sales records.
Dryers
     As discussed earlier in the preamble of the example regulation,
insurance and fire prevention representatives should be contacted at an
early stage of the regulatory development to insure that there are no
problems  in acceptance of the recovery dryer by these  agencies.
     The  availability of domestically produced recovery dryers could
strongly  influence the enforcement of a RACT regulation.  Currently,
there  is  only one manufacturer of recovery dryers in the United States,
                                 E-6

-------
 and the  lag time  between  order  placement  and  dryer  delivery  may  be  a
 problem.   However,  the  recovery dryer  manufacturer  has  stated  that  the
 production facility is  being  expanded  and that  projected  demand  for the
 dryer will  be  met.   In  order  to accommodate any possibility  of delays  in
 dryer delivery, the model  regulation contains a provision XX.050 (A)(5)
 that would postpone the compliance deadline for control equipment not
 delivered  within  12 months of ordering to 3 months  after  its delivery.
      Recovery  dryer emissions reduction performance would be determined
 by  observing the  maximum  final  recovery rate  (50 milliliters per minute)
 which defines  a minimum time  that the  dryer must operate  to ensure  suf-
 ficient  recovery  of the solvent evaporated from the drying articles.
 Recovery dryer maximum  final  recovered solvent  flow rate  should be
 monitored  frequently during the  initial two-week period of recovery
 dryer operation following installation.   This procedure is intended  to
 familiarize the operator with the effects of variations in load weight,
 fabric type, and  ambient (air and water)  temperatures on  the emission
 reduction  and  recovery performance of  the dryer.  No further monitoring
 is  recommended because such monitoring would be impracticable.   It  is
 recommended, however, that records of  the initial recovered solvent  flow
 rate  monitoring be  kept as a  reference to the affects of  variations  in
 loading and operating parameters on the recovery cycle duration in a
 given plant.
      Application of carbon adsorption technology to standard dryer VOC
 emissions may be possible.  Although an analysis of the only current
 application of this technique has shown that it is less effective in
 reducing overall  dryer VOC emissions,  modifications to the system could
 result in equivalence with the recovery dryer.   For example,  a  substantial
 increase in the mass of activated carbon  available for adsorption could
 result in sharply reduced  adsorber exhaust VOC concentrations.   Also, a
decrease in the flow rate  of the vapor  through the system would increase
the VOC adsorption by the  carbon.
     As discussed  in the model regulation, equivalent systems of  dryer
VOC emission control will  be permitted, if approved by the Director  and
the equivalence demonstration test is  performed  using test methods  or
                                 E-7

-------
procedures that have been previously approved by the  Director.
Verification of the compliance of a dryer emission control  device other
than the recovery dryer would be based on a one-time  test of the maximum
dryer VOC mass emissions per unit mass of articles cleaned after installa-
tion of the alternative control device.   The determination of this mass
emission rate would require the use of EPA Reference  (40 CFR Part 60)
Test Methods 1, 2, and proposed 25A (45 FR 83126 December 17, 1980).
While Methods 1 and 2 would govern the selection of atmospheric vapor
exhaust sampling points and the procedure for determining the exhaust
flow rate, respectively, Method 25A (discussed below) would govern the
measurement of the VOC (as propane) vapor concentrations in the control
device atmospheric exhaust by  a flame ionization analyzer (FIA).  First,
a  response ratio of the  FIA's  measurement of a given concentration of
propane to the same concentration  of VOC  (solvent) would be determined
in the  laboratory.  Then the  FIA would be field  calibrated to measure
concentrations of  propane gas,  and the measured  ppmv concentrations  (as
propane)  of the  control  device exhaust gases would be multiplied  by  the
previously  determined  response ratio, thereby determining the ppmv
concentration  (as  solvent)  of the  VOC emissions.  Then  by calculation,
one would convert  the  VOC concentration  by  volume  (ppmv) to  mass
concentration.   This procedure should  be carried out under  various
conditions  of fabric type,  load weight,  and temperatures  that  are typical
of the range  encountered in the dry cleaning industry.   The  results  of
 this procedure should  be reported as  kilograms  VOC emitted  per
 100 kilograms dry weight of articles  dry cleaned.  Compliance  would  be
 established if the average  (over the  test period) VOC  mass  emissions per
 100 unit mass of articles cleaned did not exceed 3.5.   Subsequent changes
 in the design or performance of the control device could necessitate a
 revaluation of the device's maximum VOC emissions.   Since Method 25A is
 not a promulgated EPA Reference Test Method (as of the date of this
 appendix) and subject to change an alternative test procedure can be
 used in the interim.   One alternate test method is:   "Alternate Test
 Method For Direct Measurement of  Total Gaseous Compounds Using A Flame
 Ionization Analyzer," presented in the OAQPS Guideline Series document
                                   E-8

-------
entitled, "Measurement of Volatile Organic Compounds" (Revised
September 1979, EPA-450/2-78-041).   Other equivalent test methods and
procedures are allowed under XX.040 (E) provided that these methods or
procedures have been approved by the Director.
FiIters
     Verification of compliance with the section of the regulation
governing solvent filters would be based on visual  observation of the
proper installation, operation, and maintenance of a cartridge fil-
tration system.  The manufacturer's manual for the particular system
should be consulted for proper procedures, sizing of connections, and
parameters requiring inspection or monitoring to ensure proper operation.
The operator's familiarity with the system's performance, safety, and
maintenance requirements should be evaluated as an important aspect of
overall compliance.
     Plant records detailing the date and time of cartridge replacements
could be used to verify compliance with the mandated 8-hour minimum
drainage time.  If these records are lacking, and there is some suspicion
that the cartridges are being drained improperly, it may be necessary to
require valid proof of compliance by proper record keeping over a specified
period.
     As stated in reference to equivalent systems for dryer emission
controls, equivalent systems for control of VOC emissions from solvent
filters can be approved by the Director.  A solvent filtration system
VOC emissions control device other than the cartridge filter would
require a one-time verification of its compliance with maximum emissions
(solvent content of waste) of no more than 1.0 kilograms VOC per
100 kilograms dry weight of articles dry cleaned prior to its use in a
petroleum dry cleaning facility.  The testing procedure would be based
on the determination of the VOC (solvent) content in each of five
one-kilogram samples of filtration waste from the control device taken
over one-week intervals during which conditions of soil loading, load
weight, and fabric type vary in a manner typical of the facility.  ASTM
Test Method D322-80 (Gasoline Dilutent in Used Gasoline Engine Oil by
Distillation) should be used to determine the solvent content of each
                                 E-9

-------
sample.   The total  masses of articles  cleaned and filtration waste
generated between waste samplings  should be recorded and multiplied by
the sample ratio of sol vent-to-waste content (as determined by the  ASTM
Method), resulting in a determination  of the mass of VOC emitted (solvent
generated) per 100 unit mass of articles dry cleaned.   The compliance
verification resulting from this equivalence test would apply only  to
the existing control device configuration,  and would be invalidated by
significant changes in the design or performance of the device.
Fugitive Emissions
     Location of fugitive emission sites would rely on a visual  inspection
of the overall dry cleaning system components which include washers,
dryers, solvent filters, settling tanks, and all containers and conveyors
of petroleum solvent.  Sources of VOC liquid leaks would be identified
directly, and the operation and maintenance of devices known to be
sources of VOC vapors would be evaluated.
     Dry cleaning system components found  leaking liquid solvent should
be repaired immediately.  Pipes, hoses, and fittings should be examined
for active dripping or dampness.  Pumps and filters should be closely
inspected for leaks around  seals and access covers.  In general, there
should  be no visible  signs  of  liquid solvent.
     Solvent vapor  leaks should be  controlled by reducing the number  of
sources where solvent is exposed to the atmosphere.  Under  no circumstances
should  there be  any  open containers (cans,  buckets, barrels)  of solvent
or solvent-containing material.  Equipment containing  solvent (washers,
dryers, extractors,  and  filters) should remain  closed  at  all  times  other
than during maintenance  or  load transfer.   Lint filter and  button  trap
covers  should remain  closed except  when  solvent-laden  lint  and  debris
are removed.  Gaskets and  seals should  be  inspected and  replaced when
found worn  or defective.   Solvent-laden  clothes should never  be allowed
to sit  exposed  to  the atmosphere  for  longer periods than  are  necessary
for load  transfers.   Finally,  vents on  solvent-containing waste and new
solvent storage  tanks should be constructed and maintained in a manner
that  limits  solvent vapor  emissions to  the maximum  possible extent.
                                  E-10

-------
TECHNICAL REPORT DATA
(f lease read instructions on the reverse before completing)
EPA-450/3-82-009
4. TITLE AND SUBTITLE
Control of Volatile Organic Compound Emissions
from Large Petroleum Dry Cleaners
7. AUTHOR(S)
S. PERFORMING ORGANIZATION NAME AND ADDRESS
TRW Environmental Engineering Division
Progress Center
Post Office Box 13000
Research Triangle Park, North Carolina 27709
12. SPONSORING AGENCY NAME AND ADDRESS
U.S. Environmental Protection Agency
Emission Standards and Engineering Division (MD-13)
Office of Air Quality Planning and Standards
Research Triangle Park, North Carolina 27711
3. RECIPIENT'S ACCESSION NO.
5. REPORT DATE
SEPTEMBER 1982
6. PERFORMING ORGANIZATION CODE
8. PERFORMING ORGANIZATION REPORT NO.
10. PROGRAM ELEMENT NO.
11. CONTRACT/GRANT NO.
68-02-3174
13, TYPE OF REPORT AND PERIOD COVERED
Final
14. SPONSORING AGENCY CODE
EPA 200/04
15. SUPPLEMENTARY NOTES ' 	 	 ~~ 	 	 	
This report provides the necessary guidance for development of regulations to
limit emissions of volatile organic compounds (VOC) from large dry cleaning plants
using petroleum solvents. This guidance includes emission estimates, control
technologies, costs, environmental effects and enforcement; for the development of
reasonable available control technology.
                                  KEY WORDS AND DOCUMENT ANALYSIS
                   DESCRIPTORS
                                                 b.lDENTIFIERS/OPEN ENDED TERMS
                                                                                COSATl I'leld/Gr
   Air Pollution
   Regulatory Guidance
   Dry Cleaning
   Petroleum Solvents
 Air Pollution Control
 Stationary  Sources
 Volatile Organic
    Compounds
                                                 19 SECURITY CLASS (This Report)

                                                  Unclassified
   Unlimited
20. SECURITY CLASS (This page!
Unclassified
                             21  NO. OF PAGES

                                   166
                                                                              22. PRICE
EPA Form 2220-1 (Rev. 4-77)   PREVIOUS EDITION is OBSOLETE

-------