-------
After 8 hours of drainage and evaporation, the cartridges have lost 37
percent of their undrained solvent content.   After an additional  4
hours, they have lost only 3 percent more, indicating that extended
drainage periods are unnecessary.
     The value of cartridge VOC emissions obtained above, is in general
agreement with a dry cleaning trade association estimate of 0.75 kg per
100 kg of articles cleaned (MID, 1971).
     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
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 is the only method for removing
solvent-soluble impurities such as grease and oil.  The installation of
cartridge filtration equipment could, however, decrease the frequency of
solvent distillation.  This decrease would 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.
3.3.3  Settling Tanks               .
     Settling tanks, as applied to the petroleum dry cleaning industry,
are used to filter oils, grease and suspended particles .from petroleum
solvent.  Settling tanks are utilized  when other methods of filtration
(cartridge filtration) are uneconomical due to heavy soil loadings in
processed articles.  The difference in densities of the solvent,  oil,
grease, and suspended particles forms the basis for the separation
process.  Petroleum solvent is less dense (0.7 to 0.8 kg/liter) than
oils, grease, solids, and water and, therefore, rises to the top of the
tank while the other materials settle to the bottom.
     Settling tanks are conical bottom, cylindrical containers, with two
or more orif.ices for solvent flow and the venting of solvent vapors.
These tanks vary in size, the maximum size being 5,680 liters (1,500 gal),
with only two 5,680 liter tanks allowed per plant by national fire
                                 3-17

-------
 codes,  totaling  11,360  liters  (NFPA,  1979).   Settling  tanks  operate in
 either  batch  or  continuous  modes.   The  batch  operated  settling tank (see
 Figure  3-5) is used where contaminated  solvent  is pumped from the washer
 and  stored  in the  settling  tank.   The time  allotted  for settling depends
 upon the  size and  depth  of  the  tank (tank capacity), the extent of
 solvent contamination,  and  the  variation in densities  of the contaminants
 in comparison to that of petroleum solvent.   After allowing  sufficient
 time for  settling, a valve  is manually  opened at the base of the tank to
 drain off the residue.   The residue,  a  very viscous  liquid,  is collected
 in a drum and allowed to set for a few  hours  to several days.  Solvent
 is skimmed from  the top  of  the  residue  at the end of this period, and
 the  residue is then discarded,  mixed  with boiler fuel, or sold to a
 solvent reprocessor (Sluizer, 1981).  The solvent remaining  in the
 settling  tank, after the removal of the residue is conveyed to the
 vacuum  still  for further purification.
      The  continuously operated  settling tank, also known as a separator
 tank, (see Figure 3-6) has  solvent flowing  in and out  on a continuous
 basis.  Contaminated solvent from  a washer  is pumped to the settling
 tank.   The solvent flows from the  top of the  tank as the heavy contaminants
 settle  to the bottom.   Periodically,  residue  is drained from these tanks
 and  is  discarded or utilized in the same manner as that from the batch
 operated  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 factors having  a significant effect  on the solvent retention of
 settling tank waste are  not  well known.   Because settling tanks have no
moving parts,  the only factors that could increase the solvent content
of the discarded waste are  an insufficient  settling time or excessive
turbulence in the tank resulting from an excessive solvent flow rate.
There is little that can be  done to alleviate the latter problem of
turbulance and an increase  in the settling time should correct the
former problem.
                                 3-18

-------
                                            I Vent
Washer
 Pump
Settling Tank
                               Solvent
                              Residue  •
            Residue  ^
        to 55 gal. Drum"
Solvent
to Still
              Figure 3-5.  Batch Flow Settling Tank.
                               3-19

-------
Washer
 Pump
Separator Tank
            Residue ^
        to 55 gal.  Drum
                                       Jl!
                              Solvent
                             Residue
Solvent
to Still
            Figure  3-6.  Continuous Flow Settling Tank.
                              3-20

-------
     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,1 1981; Landon, 1981).
3.3.4   Vacuum Stills
     The purpose of distillation in the petroleum dry cleaning industry
is to separate liquids (primarily  petroleum solvent) with lower boiling
points  from those with higher boiling points  ("high boilers") such as
oils, grease and solid residue (dirt, lint and detergents).   The mixture
of petroleum solvent, oils and grease that is conveyed to a still is
typically a mixture of liquids with a wide range of boiling points.
Additionally, there are significant differences in the boiling ranges of
petroleum solvents themselves, as  represented by the range of flashpoints
from 38°C to 60°C (100 to 140°F).  Consequently, a still should be
operated at a temperature that will evolve (evaporate) the liquids that
fall within the boiling range of petroleum solvents while leaving behind
the oils, grease, and solid residue.   Petroleum dry cleaning  solvents
have boiling ranges from 150°C to  215°C (300°F to 415°F) at atmospheric
pressure.  A steam pressure greater than 100 psi (670 KPa) is required
to boil  petroleum solvents under atmospheric conditions  (Washex,  1973).
Consequently, distillation of petroleum solvents takes place under a
vacuum of 75 to 92 kPa (22 to 27 inches Hg) which lowers the boiling
range for petroleum solvents to 107°C to .113°C (225°F to 235°F) at steam
pressures of 235 to 600 kPa (35 to 90 psi).   Vacuum stills are sized by
the volume of solvent to be processed and range in size from 95 to
1,900 liters (25 to 500 gallons) per hour (Washex, 1973).
     A vacuum still  is composed of four principal  parts:   the boiling
chamber, condenser,  gravimetric seperator,  and the moisture separator.
Spent solvent from a washer,  filter or settling tank is pumped to the
                                 3-21

-------
boiling chamber of the vacuum still (see Figure 3-7).  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 from the boiling chamber into a water-cooled
condenser where the vapors condense to a mixture of liquid solvent and
water.  This mixture is then piped to a gravimetric separator, where the
solvent and water are separated by the differences in their densities.
Finally, the solvent is pumped to the moisture separator, a tank containing
cotton rags or salt pellets, where any remaining water is absorbed.
     When the concentration of high boilers has reached a specified
level, the liquid contained in the boiling chamber is allowed to boil
for a period of up to 30 minutes in a process commonly called "boildown"
in which most of the remaining solvent is removed.   The need for boildown
is indicated by 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).   During
boildown the solvent flow to the still is manually shut off and steam
flow is increased to the maximum available level by opening a valve in a
bypass line around the regulating valve.   Extending the boildown period
beyond 30 minutes or until the residue level  drops below the steam coils
could cause encrustation of contaminants on the steam coils.   The coils
would then have to be removed and cleaned in a caustic solution.   Otherwise,
solvent evaporation would be seriously impeded.  After boildown,  the
liquid residue is drained from the boiling chamber and still  operations
are resumed.   According to a still manufacturer, still boildown is
typically required when the solvent an,d water flpw from the condenser
are reduced by 65 to 75 percent (Washex, 1973; Rosenthal, 1980).   The
boildown period should be terminated when the solvent flow from the
condenser is reduced by 65 to 75 percent.   This flow reduction is
subjectively determined based on the operation of a particular still.
Flow reduction is indicated by observation of sight glasses in the
solvent lines between the boiling chamber and the condenser,  as well as
at the solvent outlet of the separator.
     The frequency of boildown is a function of the size (volume) of the
boiling chamber sump and the location of the steam coils within the
                                 3-22

-------
                       Condenser
                        Solvent
                         Vapor
         Moisture Separator
 Cotten Rags
or Desiccant
                             Distilled
                              Solvent
                                                                Still
    Solvent Vapor

    Baffle
            Distilled
         Solvent Out
                                              Steam
                                            Condensate
                                               Out
        Level
        Control
        Valve

         Steam Chest
Dirty
Solvent In
                     Figure 3-7.  Vacuum Still Diagram.
                                     3-23

-------
boiling chamber.  A still with a small capacity sump would have a higher
boildown frequency than a similar capacity still with a larger sump.
This is because the level of residue will increase at a faster rate in
the smaller sump.  A float valve maintains a solvent level of 100 to
150 mm (4 to 6 inches) below the top of the steam chest in a typical
modern still.  The purpose of exposing the upper part of the steam chest
is to volatilize the foam created by the boiling process.   Foaming
results from water in the solvent inflow flashing to steam which could
impede distillation while increasing the carry-over of oil and other
contaminants in the condensed solvent (Washex, 1973).
     A vacuum still with the steam chest located in the bottom of the
sump would require frequent boil downs to maintain a solvent level of
100 to 150 mm below the upper part of the steam chest.   However, if
boildown is too infrequent, the steam coils may be submerged in the
still residue, with much of the heat being lost in heating the residue.
Because the still residue has a higher boiling range than petroleum
solvents, the transfer of heat from the steam coils through the residue
would be impeded, thereby limiting solvent evaporation.
     By following these procedures for solvent evaporation/condensation
and still boildown, an industry trade association representative states
that a well maintained and operated unit should distill  15 to 30 liters
(4 to 8 gallons) of solvent per 45 kg (100 Ibs) of articles cleaned,
with 3 percent of the solvent distilled retained in the discarded still
waste (Fisher, 1975).   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 (Andrasik, 1981).   (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, and
samples with solvent contents that exceeded 51 percent were considered
extremely high, indicating a poorly operated still  (Andrasik, 1981).
                                 3-24

-------
Most dry cleaners boil  down  their  stills  at  intervals  based  on  convenience
(i.e.,  boildown  once  a  day or  other  predetermined  rate) which may account
for the high  volume of  solvent retained in the  residue. In an EPA test
of a vacuum still at  an industrial petroleum dry cleaning plant, the
solvent content  of the  discarded still waste exceeded  100 percent .solvent
'by weight  (Jernigan and Kezerle, 1981).   This still was boiled .down at
the end of each  day.  When the frequency  was reduced to one  boildown in
10 days the weight of solvent  in the  residue decreased to approximately
80 percent by weight.
     The size of the  vacuum  still  or  the  sector application  (commercial
or industrial) should have very little effect on the proportion of
solvent retained in the residue.   This is due to the fact that the
design  of  a vacuum still is  identical for commercial or industrial
applications.   Although the  content of high  boilers in an industrial
still could be more than that  of a commercial still, reductions in still
waste solvent content should be possible by  boiling down only after the
solvent flow  is  reduced by 75  percent and by terminating boildown only
after the  flow has again been  reduced by 75  percent.
     There are a number of factors that will affect the operation of a
vacuum  still.   A continuously  operated still is more efficient and cost
effective  than a comparable still that is operated intermittently.   In a
continuously  operated still less heat in required to maintain a sufficient
temperature to adequately evolve solvent from the boiling chamber.   In
contrast,  the intermittently operated still   requires more heat (steam)
because of the high frequency  of start-ups.   At the onset of start-up,
much of the heat required is used to increase the temperature of the
still  and its  contents  to a uniform level.   During a given time period,
an intermittently operated still will evolve less solvent than a
continuously operated still.    Consequently,   an intermittently operated
still  would have a higher solvent content in its still  residue than a
continuously operated still.             ''
     The efficiency of a still  decreases proportionally to the increase
in soil  loading of the  incoming solvent.   Solvent entering a still  with
a high content of suspended particles, oils  and grease  will  drastically
reduce the distillation efficiency and useful life of the  still.   Because
                                 3-25

-------
 a  vacuum  still  has  a  limited  operating  range of temperature and pressure,
 the degree  of solvent contamination could  increase the boiling range of
 this mixture beyond the temperature limits of the still.  In this situation,
 little, if  any, solvent is evolved from the still.  Solvent entering a
 still from  a cartridge filter will have less suspended contaminants than
 solvent from a  diatomite filter.  The diatomite filter would have less
 suspended contaminants than solvent from a settling tank.  This problem
 of wide variations  in soil loading can  be alleviated somewhat by increasing
 the frequency of boil down with the increase in solvent contamination.
     Water  in the incoming solvent will produce a foaming action in the
 still.  This foaming action will reduce the vacuum in the still, thereby
 reducing the solvent evolution rate.  There is very little that can be
 done to remove  the  water from the solvent because it is inherent in most
 dry cleaning operations.
     A vacuum still should always be operated within the manufacturer's
 recommended levels  for steam  pressure and temperature.   Otherwise the
 evolution of solvent will diminish and  undesirable fractions of liquids
will mix with the distilled solvent.
     The condenser water flow, between  8 and 30 liters (2 to 8 gallons)
per minute, and temperature,  between 16 and 27°C (60 and 80°F), are not
very crucial items  because the vaporized solvent will condense over
these ranges of temperature and flow rates.  However, condenser water
temperatures greater than 27°C with flow rates less than 8 liters per
minute would reduce the vacuum in the system such that a sufficient
boiling temperature could not be maintained (Washex,  1973).
     It can be  inferred from  the previous discussion that commercial
vacuum distillation procedures should be more efficient than industrial
distillation procedures, assuming that  both stills are operated continuously
and the soil loading is greater in industrial  plants.   Also, commercial
vacuum stills should have less solvent  retained in their still  waste
than an industrial  vacuum still  when both are operated under identical
conditions.
3.3.5  Fugitive Emission Sources
     There are a number of sources of fugitive emissions in a dry cleaning
operation.  Fugitive sources typically  include emissions from the extraction
                                 3-26

-------
 cycle of a washer,  emissions given off during the transfer of solvent-laden
 articles from washers  and dryers,  emissions  resulting from solvent
 remaining in  articles  removed from the dryer,  liquid and vapor leaks  in
 tanks and piping,  and  the evaporation of solvent from open containers.
 Also  included are  fugitive emissions  from filter muck,  still,, and settling
 tank  residue  which  may be stored in open containers  prior to  disposal.
 Although many sources  of  fugitive  emissions  can  be  identified,  (i.e.,
 pumps,  tank vents,  and piping)  and the VOC concentrations within the
 vicinity of these  sources can be quantified,  it  is virtually  impossible
 to  estimate an emissions  rate for  most fugitive  sources.
      In  an EPA test to determine fugitive emissions  at  a large  industrial
 dry cleaning  plant,  it was found that more than  0.5  kg  VOC per  100  kg
 dry weight of articles cleaned  was  emitted from  various sources within
 the plant (Jernigan and Kezerle, 1981).   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  not  technically and economically feasible to
 quantify all  sources of emissions  in  a dry cleaning  plant.  This is
 because  certain emissions  are prevalent only during  the operation of the
 dry cleaning  equipment, while other (difficult to detect  by instrumentation)
 low-level  sources emit continuously.   A dry cleaning  industry trade
 association publication assumes, however,  that miscellaneous  (fugitive)
 emissions  are  nominally 1  kg of  VOC emissions per 100 kg  dry weight of
 articles  cleaned (Fisher,  1975).
 3.4  BASELINE  EMISSIONS
     The  baseline emissions level for  the  petroleum solvent dry cleaning
 industry  is the level  of emission control  that is achieved by the industry
 in the absence of additional EPA standards.  This section presents the
 logic and  rationale leading to the selection of the baseline emission
 level.  The baseline level is established  to facilitate comparison of
 the economic,  energy,  and environmental  impacts of the  regulatory
 alternatives  (See Chapters 6, 7 and 8).  A brief description of existing
 regulations that limit  emissions from facilities within the affected
 industry  is presented.   Generally,  the  regulations of interest are the
Control Techniques Guideline (CTG) document and the State Implementation
 Plans (SIPs).
                                 3-27

-------
3.4.1   Existing Regulations
     The  future issuance of the Control Techniques Guideline (CTG) has
the greatest potential for impacting the NSPS baseline emissions.  This
EPA document is distributed to the  individual states in an effort to
provide a common data base for the  development of the individual SIPs.
In the  current draft CTG for the Petroleum Dry Cleaning Industry, it  is
recommended that only large size plants be regulated (potential 100 metric
ton emitters).  These facilities should install solvent recovery dryers,
install cartridge filters (where solvent filtration is used), and implement
improved  operating and maintenance  procedures to minimize emissions from
stills, settling tanks, and fugitive sources.
     The  impact of the future CTG on the NSPS baseline is a function of
the number of SIPs adopted at the time of NSPS proposal and the proportion
of the SIPs that will adopt the emission control technology outlined in
the CTG.  This latter group of states will not meet the December 31,
1982 statutory deadline for demonstration of attainment of National
Ambient Air Quality Standards for ozone.  At this time, however, it is
impossible to make a reasonable estimate of the number of petroleum dry
cleaning  plants in the non-attainment areas and the localities that will
write regulations.   Thus, the impact of the CTG on the SIPs and their
impact on the NSPS baseline emissions has not been determined in this
document.
3.4.2  Calculation of Baseline Emission
     Baseline emissions calculations are based on EPA field test data
and on information supplied by dry cleaning trade association
representatives.   As previously discussed, standard dryer drying cycle
VOC emissions were found to be 14 kilograms and 28 kilograms VOC per
100 kilograms articles cleaned in two EPA tests.   An industry trade
association reported standard dryer emissions as 14 kilograms VOC per
100 kilograms articles cleaned.   Thus, the range of baseline dryer VOC
emissions is 14 to 28 kilograms per 100 kilograms articles cleaned.
Based on  an average of the data from the three tests,  the nominal value
of standard dryer drying cycle emissions is 18 kilograms VOC per
100 kilograms articles cleaned.   Solvent recovery dryers have not been
included  in the baseline because there are relatively few in operation
                                 3-28

-------
 in proportion to the  current  number  of  standard  dryers,  because  the
 impacts  of  the  NSPS and  CTG on  dryer sales  are difficult to  determine  at
 this  time,  and  because the condensation recovery of  petroleum  solvent  is
 a new dryer technology in the dry  cleaning  industry.
      VOC emissions produced by  the disposal  of spent cartridge filter
 elements were found to be less  than  0.5 kilograms VOC per 100  kilograms
 articles cleaned in an EPA field test.   Additionally,  a  dry  cleaning
 trade association determined  that  VOC emissions  from cartridge filters
 are approximately 0.8 kilograms VOC  per 100  kilograms  articles cleaned.
 Consequently, the baseline range for cartridge filter emissions  has been
 established as  0.5 to 1.0 kilograms  VOC per  100  kilograms articles
 cleaned,  with nominal cartridge filter  emissions being 1.0 kilogram VOC
 per 100  kilograms articles cleaned.   Diatomite filters (and  devices
 related  to  diatomite  filter muck processing) are not included  in the
 baseline because, as  previously discussed, both  industry trade associations
 and equipment manufacturers indicate that the trend  in facilities utilizing
 solvent  filtration has been toward the  installation  of cartridge filters.
 Thus,  a  plant buying  a solvent filter would  tend to  choose a cartridge
 filter.
      Vacuum  still VOC emissions were estimated to  range  from 1 to
 3 kilograms  VOC per 100  kilograms  of articles cleaned  in  a dry cleaning
 trade  association report.  However, |an  EPA field test  of  a vacuum still
 found  emissions to be as high as 7 ifilograms VOC per 100  kilograms
 articles cleaned.  Because of the  diversity of the emissions reported in
 the trade association analysis, the  range of baseline  vacuum still
 emissions was established at 1 to  7  kilograms VOC  per  100  kilograms
 articles cleaned.  However,  the nominal  emission value was set at
 3 kilograms VOC per 100  kilograms  articles cleaned based  on the highest
 emissions reported by the trade association.
      Fugitive VOC emissions measured  in  an EPA field test were approximately
 0.5 kilograms VOC per 100 kilograms  articles cleaned.  An  industry trade
 association estimated these fugitive  emissions to  be approximately
1.0 kilogram VOC per 100 kilograms articles cleaned.   Therefore,  the
 nominal  baseline fugitive emissions  value has been established as
 1.0 kilogram VOC per 100 kilograms articles cleaned,  representing the
maximum emissions reported by the  industry trade association.
                               .  3-29

-------
     Baseline emission rates discussed above are summarized in Table 3-2.
Totals of both ranges and nominal values are given, thereby indicating
the total emission rates that could be encountered in plants having all
or some of the equipment listed.  In general, the most common variation
(See Chapter 6 for a discussion of model plants) is the omission of
solvent filtration in some large plants with high throughputs and heavy
soil loadings.   In these facilities, settling tanks are used prior to
distillation, and the emissions resulting from these tanks are included
under the general category of fugitive emissions.
                                 3-30

-------
                   Table 3-2.   BASELINE VOC  EMISSION  RATES
         (in  kilograms VOC emitted per 100  kilograms articles cleaned)
Affected facility
Standard dryer
Cartridge filter
Vacuum still
Fugitive emissions
Total
Emissions range
14
0.5
1
0.5
16
(15.5
- 28
- 1 .
- 7
- 1
-37
- 36)a
Nominal emissions
18
1
3
1
23 a
(22)a
Total VOC emissions in plants using settling tanks
insteam of cartridge filters.
                                  3-31

-------
3.5  REFERENCES FOR CHAPTER 3

Andrasik, I.  1981.  International Fabricare Institute (IFI).  Letter to
     Q. Corey, TRW Inc., July 20.  Solvent content of still waste.

Bee, W. and W. Fisher, 1976.  Report on Cartridge Filtration Life.
     International Fabricare Institute (IFI).  Joliet, Illinois.  Focus
     No. 1.

Fisher, W.  1975.  ABC's of Solvent Mileage, Part 1.  International
     Fabricate Institute (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., October 16.
     Comments on alternative methods of filtration waste solvent content
     reduction.

IFI.  1973.   An Introduction to  Industrial Dry Cleaning Methods, Part
     One.  IFI Special Reporter.  Volume One, Number Three.  International
     Fabricare Institute.  Joliet, Illinois.

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).  February.

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).
     February.  [Pico Rivera].

Landon, S.  1981.  Washex Machinery Inc., Telecon with Q. Corey, TRW
     Inc., August 4.  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.
     June.
                                  3-32

-------
 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
      Corey,  TRW  Inc., March  18-April  25.   Sizes of petroleum dry cleaning
      equipment and expected  sales for 1980.

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

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

 Phillips,  R.   1966.   Dry Cleaning.  National Institute of  Drycleaning,
      Inc.  Silver Spring, Maryland.

 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).   February.

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

 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,  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 Washers and  Filters.
     Publication No. T-559a.    Wichita Falls, Texas.   September.

Watts, A. and A.  Fisher.  1975.   Results of Membership Survey of  Dry
     Cleaning Operations.  Joliet,  Illinois.  IFI Special  Reporter 3-1.
     January-February.
                                 3-33

-------

-------
                     4.  EMISSION CONTROL TECHNIQUES

     Equipment and procedures selected as representing currently available
emission control technology 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 this equipment.
The verification of these criteria is supported, where possible, by
engineering analyses and field test data.
4.1  DRYER EMISSION CONTROL TECHNIQUES
     There are only two techniques for dryer VOC emission control that
have been demonstrated in actual use in petroleum solvent dry cleaning
plants:  recovery dryers and carbon adsorption.  While recovery dryers
have been in use in petroleum dry cleaning plants since about 1978,
carbon adsorption equipment has been used in only one petroleum dry
cleaning plant.  Both of these dryer emission control techniques are
discussed in the following subsections.
4.1.1  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 4-1).  In the current configuration
of this machine, a steam-heated air stream is directed around and through
a tumbling load bf drying articles by a blower.  The solvent-laden air
stream is forced 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.  The reheated air stream then passes to the tumbler where the
cycle of solvent evaporation repeats.   In contrast, a standard dryer-
forces heated air through the tumbling articles and then exhausts it to
the atmosphere.

-------
        1 -----
     Turabler
Steam
Chest
                	Solvent Vapor 	


1
1
1


Condenser
A
1
i
1
1
1
	 1



P



Rec
L1q
)
alvent-Kj
Separate
4
Water
Outlet



overed
ylc

Solvent
ter »
r Outlet



Figure 4-1.  Solvent  Vapor Flow Diagram for a Recovery Dryer.
                              4-2

-------
      The most important component of the 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 that
 steadily reduces the vapor temperature (under the existing conditions of
 vapor 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 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  4-2a),  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.  Thermostatically controlled
 solenoid  valves  control both steam  and cooling-water flows.  The valves
 are intermittently  opened  to 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 4-2b),  both the
 steam and cooling water solenoids close while atmospheric air and vapor
 stream dampers divert vapor flow from the condenser loop.  This permits
the intake of atmospheric  air which is forced over the tumbling articles
                                                                      q
and exhausted to the atmosphere, at a flow rate ranging from 9  to 13 m /min
 (300 to 450 ft3/min) (Jernigan and Lutz, 1980; Plaisance et a!.,  1981).
     In the open door phase, the exhaust fan pulls ambient air  in through
the dryer and exhausts it directly to the atmosphere at approximately
                                 4-3

-------
     Atmospheric
        A1r
      M/
    Tumbler
                 Water
\y
Blower
Yap
sr Stream
Qqmper

—
                     Steam
                   Solenoid
                                         Haer
                                                              a. Recovery Phase
   Atmospheric
 Air
Damper
                   Solenoid
                           A
Condenser

SJ>?
(^\
Cold Wa
Soleno
•^
                                                       Cooler
Storage i
      llower— —  —
                    Exhaust
                     to
                   Aticsphere
                                                  b. Exhaust/Cool -Down Phase
  Figure 4-2.  Operating Cycles of  the Existing Recovery Dryer.

                                    4-4

-------
     33                                                -
 22  m /min  (750 ft /min).   Thus,  solvent vapor escaping from the  articles
 being loaded and unloaded is  removed from the operating area (Hoyt,
 1979).
      Control  of the  duration  of  the  drying and exhaust/cool-down 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 a!.,
 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
 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 solvent recovery.  This  assumption neglects the solvent emitted
during the open-door phase, as well  as that contained in the articles
after drying.  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,  due to the difficulty of differentiating between
residual solvent and water.  However, there are indications  that drying
                                 4-5

-------
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.   A
summary of the results of the three test programs is presented in Table 4-1.
A more complete discussion and detailed description of the three tests
and their results can be found in Appendix C of this report or refer to
the cited references.
     The first test (Test A) was conducted at a large industrial dry
cleaning facility that processes approximately 1,300 kg (2,900 Ibs) of
heavy work gloves each day (Jernigan and Lutz, 1980).   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 articles
cleaned.  The condenser reclaimed an average of 23.4 kg of solvent per
100 kg articles cleaned.   The dryer appeared, at times, to operate at or
above 90 percent of the 10,000 parts per million LEL concentration for
petroleum solvent (as measured at the condenser vapor inlet during the
recover phase).  However, 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 an unusually high solvent
retention of as much as 30 percent by weight (Jernigan and Lutz, 1980).
     A second EPA recovery dryer test (Test B) was undertaken at a large
commercial plant that processes approximately 180 kg (400 pounds) of
general  apparel each day (Jernigan et a!., 1981).   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) exhaust cycle VOC dryer emissions vented
during the exhaust cycle were found to be 3.85 kg VOC per 100 kg dry
                                 4-6

-------
              Table 4-1.  SUMMARY OF RECOVERY DRYER TEST RESULTS

Test
site
Test A


Test B


Test C


Type of
articles
cleaned
Industrial
(Gloves and
fender covers)
Commercial
(Personal
clothing)
Commercial
(Personal
clothing)
Recovered solvent Exhaust emissions
Number of (kg/100 kg (kg/100 kg
dryer load- of «*1cles) of articles)

tested* Range Average Range Average
13 15.5-29.2 23.4 0.68-1.25 0.96


26 8.8-14.3 10.4 2.34-9.45 3.85


40 9.5-17.7 13 1.20-7.2 3.47


Test A  Jernigan and Lutz, 1980.  [Pico Rivera].
Test B  Jernigan, May, and Plaisance, 1981.  [Lakeland].
Test C  Plaisance, Jernigan, May, and Chatlynne, 1981.  [Rhode Island].

* Number of dryer loads for which complete data were available to compute the
  recovered solvent and exhaust emissions averages.
                                    4-7

-------
 weight of articles  cleaned.   The condenser reclaimed  an  average  of
 10.4 kg of solvent  per 100 kg of articles  cleaned.  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 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  (Test C) was conducted  at a large
 commerical  plant 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 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 cycle was 3.47 kg VOC per 100  kg dry weight of articles dry
 cleaned.   The condenser reclaimed an  average of 13  kg per 100 kg of
 articles  cleaned.   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.   The  absolute magnitudes   of these concen-
trations, when compared with previous  test data, were  extremely low  and
probably inaccurate (Plaisance  et  al., 1981).

                                 4-8

-------
      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 (Hoyt, 1979).
 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 13°C  (55°F)  (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,  hotter,  humid climates may  require the lower
 temperatures associated with a chiller.   The  chiller  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 a!., 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
                                 4-9

-------
operating condition could be indicated by the temperature difference
between the condenser cooling water inlet and outlet.   According to the
dryer manufacturer, the difference should not exceed about 15°C (27°F)
during the recovery cycle (Hoyt, 1979).
      Figure 4-3 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 21°C (70°F) and 30°C (86°F).  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.  This indicates 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.,
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 minimal
increases  in the volume of recovered solvent.  This evaluation indicates
that the  0.05  liter per minute recovered  solvent flow rate could be used
                                  4-10

-------
AcJ3AOD3a 1N3A10S JO

    o         »
    o
                            CM
                                       O
                                       O
aasuapuoo) 1N3A10S Wdd SV NOIlVcLLN33NOD
0
o .
o
0
t— 1
0
o
o
en
o
0
o
CO
o
o
0
f-»
T-I 	
8
o
VD
O
O
0
in
o
o
0
«a-
o
o
o
c
c
                                                             s_
                                                             oi
                                                             OJ
                                                             o
                                                             o
                                                             Di


                                                             OJ
                                                             10
                                                             a*

                                                             s-

                                                            CJ
II 	
8
CO
^-
— 1.
o
0
«cf
*s-
--»-
o
o
0
«•»
«
8
vo
CO

o
o
CM
CO
— i_
o
o
CO
CM

o
o
«J-
CM

O
o
o
CM
— «^ _
0
o
•£>
1-4

8
CM
«— «

8
00

o
o
-«r
\
o
                                                            3
                                                            •o
                                                           .. c
                                                            (O
                                                            O)
                                                            >
                                                            o
                                                            o
                                                            (U
                                                            cr:
                                                            o
                                                            10
                                                           CO
                                                            I
                                                            I
                                                            en
         1N3ATOS
              4-11

-------
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.  It
may also 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.
     4.1.1.1  Recovery Dryer Safety.  Solvent recovery by chilled
condensation is a new technology as applied to petroleum dry cleaning.
Only one domestic manufacturer is producing recovery dryers.  Because
these units have been produced domestically only 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.  The
only test with reasonably accurate maximum concentration measurements
was conducted in a plant that dried small loads of synthetics.   This
                                 4-12

-------
plant is atypical of most of the industry.  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
concentration of 9,000 ppmv.  The actual concentration in the dryer
apparently continued to increase and probably exceeded the solvent lower
explosive limit (LEL) of 10,000 ppmv (Ashland, 1977).  In the third
test, the difficulty of preventing solvent vapor condensation in the FIA
sample lines produced unrealistically low readings of maximum concentrations.
Although 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 exceed the solvent LEL at some point in the recovery
cycle.  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.
     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.   The dryer tumbler is fitted
with a grounding system which, when properly grounded, should dispel!
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.   Flammable objects could be a
significant problem in smaller plants that process personal articles or
uniforms.  However, 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.
     If the solvent concentration in the recovery dryer tumbler reaches
and exceeds the solvent LEL and an ignition source is present and active,
an "explosion" can occur.   The characteristics of a recovery dryer
"explosion" follow a stepped process.  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.
                                 4-13

-------
The extremely rapid pressure increase brought about by the vapor combustion
opens the spring loaded explosion dampers on the top of the tumbler,
instantaneously releasing the excess pressure.  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 older non-recovery dryers
can, in the absence of fire suppression devices, results in fires which
are fed by the continuous inflow of ambient air.)
     Two recovery dryer explosions have been documented by EPA since
production of the unit began (other explosions have been reported but
specific details are unavailable).  These explosions have been'examined
as to their causes and consequences.  The first explosion 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.  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.  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.  He described the effect as being like the sonic boom from a
jet airplane.  No personal injuries resulted from the explosion.  Damage
                                 4-14

-------
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.
     Because the  technology of condensation recovery of petroleum dry
cleaning solvent  is relatively new, and because there are no .statistics
on dryer fires or explosions in the United States, 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.  The solvent
vapor is liquified and recovered by a refrigerant-chilled condenser.
The main difference between the Japanese and domestic recovery dryer is
that the dryer capacity 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.   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 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).   Factory Mutual 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,  1982).
Further tests were observed in March 1982 by Factory Mutual  that have
reconfirmed their approval  of the recovery dryer (Kennes, 1982).   In
addition,  the dryer has received the general  approval  of the Los  Angeles
Fire Department (Los  Angeles Fire Department,  1981).
                                 4-15

-------
     A survey of dry cleaning plants that have installed the recovery
dryer indicates that there have been no problems with insurability
resulting from the operation of the recovery dryer (Corey, 1981).   In
general, insurance 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 v/ho were
aware of the dryer's presence (Corey, 1981).  In most cases, however,
the continuation of this initial approval was contingent on the frequency
of explosions and accidents associated with the dryer.  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 on the unit's
ability to safely contain and control an explosion.  This approval was
made 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).  Finally, the EPA has
found no instances in which a fire marshal or insurance underwriter
prohibited the installation of a solvent recovery dryer.
4.1.2  Carbon Adsorption
     Carbon adsorption is a process in which one or more chemical
constituents of a vapor stream are removed by circulation through and
entrapment in a mass of activated carbon.  The term "adsorption"  refers
to a submicroscopic process in which molecules of vapor become  lodged in
the porous carbon surface.  The process  results in the permeation of the
surface with vapor condensate.  As this  condensate spreads over the
                                  4-16

-------
available area, the surface is "blinded" to further adsorption, resulting
in a state of "breakthrough".   After breakthrough, there is only a
minimal reduction in vapor concentration.   At this point, a process
known as "desorption" must be undertaken.   During desorption, a vapor
stream (steam or hot air) is passed through the carbon mass to re-vaporize
and remove the adsorbed materials from the carbon.  Unfortunately, the
quantity of material removed from the carbon in desorption is never
quite equal to that collected in adsorption.  This results in a "heel"
of entrained material that steadily decreases the area available for
adsorption and ultimately renders the carbon useless.
     While carbon adsorption technology is well-developed -and in widespread
use in industry, its application to dry.cleaning has been limited to
emissions reduction in plants using perchloroethylene solvent.  Recently,
however, an EPA test was undertaken in which a carbon adsorption system
was successfully used to both reduce the solvent emissions from a petroleum
solvent standard dryer and to recover the solvent in a reusable form
(Lutz et al., 1980).  This system will be used as a model for illustrating
the design and performance of a carbon adsorber as applied to the exhaust
vapor stream of a standard petroleum solvent dryer.
     A carbon adsorption system, as applied to a dry cleaning dryer, is
an add-on emissions control device.  Solvent-laden vapors from the dryer
exhaust (see Figure 4-4) are pulled through a bag filter where solid
particles such as lint are removed.  Once through the filter, the vapor
is passed through an air cooler to remove excess moisture and prevent
damage to the carbon beds by overheating.   After this, the vapor stream
is ducted to one or more adsorber vessels containing activated-carbon
(see Figure 4-5).  Depending on the quantity of carbon available for
adsorption and the concentration of solvent in the adsorber vapor inlet,
the adsorption cycle might continue for eight hours or more as solvent
is entrapped in the carbon and the remaining air is vented to the
atmosphere.  As breakthrough is approached, the adsorber vessels are
isolated from the dryer exhaust stream by valves (dampers) and a stream
of low pressure steam is injected into each vessel.   The steam flows
through the sol vent-laden carbon in a direction opposite to that taken
by the dryer exhaust (see Figure 4-5).  Transferring its heat to the
                                 4-17

-------
 u


4J
 fD
 U
oo
 c
 o
 o.

 o
 to
TJ
 O
J3


 ro
O
 QJ



 CD

-------
                               Vapor Steam
                               to Atmosphere
                     Solvent Accumulated

                       in Carbon Bed
                  Vapor Steam I
                   from Dryer I
                                            ADSORPTION
                                Chilled
                                 Water
    \	Z
         Solvent  Saturated Steam
Solvent-Laden
 Carbon Bed
     t
    Steam
      VSA/
       Condenser
Solvent-Water
  Separator
                                    I
            1
                                   Water
          Solvent
                                            DESORPTION

          Figure 4-5.  Adsorption and Desorption.
                        4-19

-------
liquid solvent coating the carbon surface, the steam mixes with the
revaporized solvent and'transports it out of the adsorber vessel.   The
mixture is ducted into a condenser where it is condensed to a liquid.
Final separation of reusable solvent from wastewater takes place in a
gravimetric separator.  In this unit the more dense water is drained
from the bottom while less dense solvent flows from the top.
     The required "size" of a petroleum solvent carbon adsorption system
is a function of the dryer exhaust flow rate and the desired emission
reduction efficiency (expressed as the percent difference in solvent
vapor concentrations at the adsorber inlet and outlet).  The quantity of
carbon required for effective adsorption is a direct function of the
adsorber inlet flow rate and an indirect function of the solvent
concentration.  The distribution of this carbon in a number of adsorption
vessels has a strong bearing on the frequency of desorption.  Also, the
high-flow, low-concentration vapor exhaust streams encountered in this
application tend to necessitate carbon bed designs that result in a
minimum pressure drop across the adsorber vessel.  In general, vertical
vessels with a high ratio of cross-sectional area to height are preferred;
however, since this type of container may require costly custom fabrication,
several smaller vessels with a similar cross-section-to-height ratio may
be used.  The frequency of desorption, therefore, is determined by the
rate at which the individual adsorber vessels approach breakthrough:
the more vessels sharing the load of adsorber inlet flow, the longer the
periods between desorptions.
     The installation of a carbon adsorption system in a petroleum dry
cleaning plant would entail several major difficulties.  First, the
space required for adsorber installation would be critical, with even a
                                                                         2
highly-modulatized adsorber requiring as much as 20 square meters (200 ft )
of floor space to process the exhaust of a single 180 kilogram (400 Ib)
capacity dryer.  Second, the high utility (electricity, steam, water,
and compressed air) demands of such a system might require a dry cleaner
to replace and/or expand existing boilers, compressors, and electrical
and water supply systems.  Finally the relatively high exhaust flow
rates found in most standard dryers, ranging from 60 cubic meters per
minute (2,200 cfm) to 300 cubic meters per minute (10,000 cfm), necessitate
                                 4-20

-------
the installation of a high-volume adsorber on a low-concentration source,
thereby contributing to both the size and complexity of the system (Lutz
et al. , 1980).
     EPA contracted the evaluation and demonstration of carbon adsorption
technology at an industrial dry cleaning facility to determine the
effectiveness of carbon adsorption in controlling VOC emissions.   The
test consisted of fitting a prototype carbon adsorption unit to the
dryer exhaust of a petroleum solvent industrial dry cleaning dryer.
Performance data were collected during the operation of the system.  The
economics of operation at this establishment were also evaluated.  In
addition, a long-term follow-up analysis of the emission reduction and
recovery performance of the adsorber system was conducted in order to
verify the efficiency and dependability of the system (Lutz et al.,
1980).
     The host dry cleaning plant was a large, industrial facility utilizing
a 230 kilograms (500 Ib) washer/extractor and a 180 kilograms (400 Ib)
dryer to process approximately 8,700 kilograms (19,000 Ibs) 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).
                                             i
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:  electicity
consumption,  natural gas consumption, water usage, steam flow rate to
adsorption unit, and solvent recovery rate.  In addition, samples  of
solvent and carbon from the carbon bed were analyzed infrequently  during
the test period.
                                 4-21

-------
functions.  The adsorption of solvent by the carbon beds has been
demonstrated to remain consistently above 90 percent and causes little
problems with optimizing performance.  Because of the pressure drop
encountered in passing dryer exhaust gases through a carbon bed, conveying
the solvent vapors from the dryer through the adsorber carbon bed requires
a blower in addition to the fan in the dryer.  This additional blower
and ductwork creates the potential for vapor leaks or severe blockage
due to lint.  The system must be engineered so that all the exhaust
gases leaving the dryer are filtered to remove lint, transported through
the adsorber system without losses in flowrate, and delivered to the
carbon beds at the proper temperature and pressure.  Because both the
concentration of solvent and the gas flowrate are very high upstream of
the carbon beds, even a small leak in the duct system will result in a
significant loss of solvent.
     Once the solvent is adsorbed on the carbon it must be desorbed with
steam, condensed, and separated from the water that results from the
condensed steam.  If the beds are not fully desorbed when required,
solvent will build up in the beds and reduce their ability to adsorb
more solvent.  Eventually the beds will become saturated with solvent
and will cease to adsorb any additional solvent vapors from the dryer
exhaust gases.  The three most common problems associated with desorption
that reduce the performance of the beds are 1) not desorbing with sufficient
steam pressure; 2) not desorbing for an adequate duration, and 3) not
drying the beds with air after desorption is complete.  Petroleum solvents
require steam pressures of 205 kilopascals to 239 kilopascals (15 psig
to 20 psig) for adequate desorption, with the higher end of this range
being preferrable (Kezerle, 1981).  The determination of adequate desorption
and drying periods is dependent on the size of the carbon beds and other
site-specific factors.  Once these parameters have been determined, they
must be adhered to rigorously.  In a similar manner, the flowrate and
temperature of water needed to condense the desorbed steam and solvent
vapors must be determined for each installation and maintained at the
proper levels.  Sufficient drying in essential, otherwise the beds will
be blinded with water which will prevent further adsorption.  Finally,
separating the condensed solvent and water must be done efficiently.
                                 4-24

-------
 Allowing  even  a  small percentage  of  solvent  to  escape with  the  drain
 water will  lead  to  losses  that  are significant  in terms  of  the  total
 system.
     In order  to insure that  the  three  functions mentioned  in the previous
 section are performed correctly,  certain  operating parameters must be
 monitored and  maintained at desired  levels.  Holding these  parameters at
 the proper  levels will maximize solvent recovery and, assuming  whatever
 solvent is  not recovered results  in  an  emission, will indicate  the
 minimum practical emission level.
     Getting the solvent from the dryer to the  carbon beds  without
 losses can be  verified by  inspecting the  system periodically for leaks.
 If permanent openings are  designed into the  system, as is the case with
 the demonstration unit, then duct pressures  must be monitored to warn
 the operator when ducts are beginning to  plug or other flow obstructions
 arise.   These  obstructions would  be  indicated by changes in the duct
 pressure  during  operation and should be removed promptly to assure good
 operation.
     The  adsorption of solvent by the carbon beds can only  truly be
 measured  by conducting hydrocarbon and  gas flowrate testing at  the inlet
 and outlet of  the beds.   It has been demonstrated .that the  carbon in the
 beds (of  the demonstration unit)  has an active  life of at least three
years.   The manufacturer has estimated  its active life at ten years.
Therefore, testing of this type more frequent than once every three
years appears  to be unwarranted.  A  possible alternative, to assess the
carbon's  performance, is to periodically  send samples of the carbon from
the beds  to the manufacturer.   Most manufacturers will  test the carbon's
ability to adsorb and desorb solvent for  no charge (Lutz et al., 1980;
Kezerle,  1981).                                              <       •
     The  third function, desorption and recovery of solvent, is the most
common source of operating problems.   First, correct desorption parameters
must be maintained.   With the demonstration system this means:
     a)   use a desorption steam pressure of 205 kilopascals to
          239  kilopascals (15 to 20 psig),
     b)   steam each bed at this pressure for JTO less than 55 minutes,
     c)   dry each bed for at least 30 minutes by blowing ambient air
          through it.

                                 4-25

-------
The.operation of the condenser can be verified by maintaining set water
flowrates and temperatures' (inlet and outlet).  Operation of the separator
can be checked by collecting samples of the two streams that leave it.
The amount of water in the solvent stream or solvent in the water stream
should never exceed 0.5 percent on a volume basis (Kezerle, 1981).
     The parameters cited above will verify proper operation of each
component of the entire carbon adsorption system.  An easier way to
verify the proper operation of the entire system is to measure the
volume of solvent recovered every time the beds are desorbed.  The
following equation could then be used to calculate the rate of solvent
recovery.
Rate of solvent recovery =
               Volume of solvent recovered X Density of solvent w -.QQ
               Weight of material cleaned since last desorption
For the demonstration system the maximum potential rate of solvent
recovery was calculated to be 11.95 kilograms of solvent per 100 kilograms
of material cleaned.  Solvent recovery rates close to this value were
indicative of good operation of the whole system.  However, this maximum
potential solvent recovery rate is a function of the dryer, carbon
adsorption system, and type of material being cleaned, and will differ
for each installation of the adsorber (Kezerle, 1981).
4.2  FILTRATION EMISSION CONTROL TECHNIQUES
     The primary VOC emissions from diatomite and cartridge filters, as
discussed in Chapter 3, Section 3.3.2, result when solvent-laden filtration
waste (either diatomite or used cartridges) is disposed in a manner that
allows the entrained solvent to evaporate to the atmosphere.  An industry
trade association has estimated typical diatomite filter emissions to
range from 5 to 10 kilograms VOC per 100 kilograms articles cleaned.
This estimate is based on filter waste that has been allowed to drain
for 24 hours (NID, 1971).  Cartridge filter emissions due to evaporation
of solvent from disposed cartridges were determined to be approximately
0.5 kilograms VOC per 100 kilograms artricles cleaned after 8 hours of
both drainage and evaporation in an EPA test  (Plaisance, 1981).  An
industry trade association has estimated VOC emissions from disposed
                                 4-26

-------
filter cartridges to be approximately 1.0  kilograms VOC per  100 kilograms
of artricles cleaned, (NID, 1971).  Thus, cartridge filters represent a
VOC emission reduction over diatomite filters of approximately 80 to
95 percent.
     Because of the wide availability of. cartridge filters,  their obvious
advantages  in reducing solvent losses (emissions), and the trend in the
industry towards their installation cartridge filtration has the distinction
of being both the baseline and control technology.  It is assumed that a
large number of existing plants utilizing  solvent filtration have converted
to cartridge filters.  However, existing plants using other  forms of
solvent filtration (diatomite filters) and new plants intending to
employ solvent filtration would benefit from the reduced solvent losses
inherent in cartridge filters while producing substantial reductions in
VOC emissions.
4.3  FUGITIVE EMISSION CONTROL TECHNIQUES
     Fugitive emissions essentially encompass all sources other than
dryers and filters.  Vacuum stills, settling tanks, and miscellaneous
emissions from tanks, pipes, and other equipment fall into this category.
     Atmospheric emissions resulting from the operation of vacuum stills
(see Section 3.3.4 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  boil down 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.
                                 4-27

-------
Although many existing plants boil down their stills on a routine or
convenient schedule, the still manufacturer recommends that boil down 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 boil down
should continue until the flow rate of condensed liquid has been again
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 41°C, 52°C, and 60°C
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 @ 18°C (20 psi @ 65°F)
          Condenser Water Outlet Temperature = 60°C to 71°C (140°F to
          160°F)
          Condenser Solvent Outlet Temperature = 24°C to 32°C (75°F to
          90°F)
While these parameters will vary as a function of the given still, plant
throughput, and soil loading, they nevertheless form a general range
                                 4-28

-------
which may 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 1000 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
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.   Based on worst-case existing emissions of 1 to 7 kilograms
VOC per 100 kilograms articles cleaned (Andrasik, 1981; Jernigan and
Kezerle, 1981), improved operating and maintenance procedures could
reduce these emission rates to as little as 1 to 3 kilograms VOC per
100 kilograms articles cleaned (Andrasik, 1981).
     Solvent vapor losses from settling and storage tanks contribute
significantly to overall fugitive emission.  These emissions generally
occur as a result of "breathing" and "working" losses.   Breathing losses
occur'when the vapor volume in storage tanks expand or contract during
changes in temperature.  This results in air being drawn into the tank
(vapor contraction) and 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.  Solvent vapor
is expelled when the tank is filled.  Draining draws in atmospheric air,
thereby producing additional breathing losses as the air becomes saturated
with solvent vapor.  In either case, steps should be taken to remove the
solvent vapor from the expelled air and to prevent the tank contents
from being exposed to the atmosphere.
     Miscellaneous emissions essentially encompass all  remaining emission
sources.  An EPA-sponsored test program conducted at a large industrial
dry cleaning facility included a sampling and evaluation of VOC
concentrations in the general dry cleaning environment (Jernigan and
Kezerle, 1981).  Significant concentrations of solvent vapor, at times
approaching 70 percent of the solvent LEL, were found around both the
                                 4-29

-------
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 an estimated VOC emission rate of 0.5 kg solvent per
100 kg of articles cleaned (Jernigan and Kezerle, 1981).
     The broad category of "leaks" can contribute significantly to
overall plant fugitive VOC emissions.  Liquid solvent drips from pipes,
fittings, valves, hoses, couplings, and pumps add to te constant background
of solvent vapor inherent to many dry cleaning plants.  Vapor leaks from
dryers, exhaust ducts, filter housings, stills, and open or improperly
sealed containers of solvent and solvent-laden waste all  contribute to
the quantity of solvent emitted to the environment.
     The only way to eliminate this 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 is
emitted by fugitive sources for every 100 kg of articles cleaned in a
typical dry cleaning plant (Fisher, 1975).  This level of VOC emissions
could be reduced by conducting a maintenance program that would completely
eliminate liquid leaks and solvent standing open to the atmosphere and
eradicate vapor leaks.  Vapor leaks could be eradicated by repairing
gaskets and seals that obviously expose solvent-rich environments to the
atmosphere.  Training of dry cleaning personnel could help in attaining
this reduction, particularly by eliminating the practice of allowing
solvent-laden loads of articles to be exposed to the atmosphere while
awaiting drying.  Equipment operators and maintenance personnel should
be taught to inspect and identify solvent liquid and vapor leaks from
sources such as hose and pipe connections, pumps, machine and tank
gaskets, seals, valves, and lint and button traps.  If leaks are found,
repairs should be undertaken immediately with the goal of minimizing any
additional solvent leakage.
                                  4-30

-------
4.4  REFERENCES FOR CHAPTER 4                 ,                 .   '.

Andrasik, I.  1981.  International Fabricare  Institute  (IFI).   Letter to
     Q. Corey, TRW Inc., July 20.  Solvent content  of still waste.

Ashland Chemical Co.  1977.  Material Safety  Data Sheet No. 0000585-001.
     Environmental and Occupational Safety Department.  Ashland, Kentucky.

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.

Factory Mutual System, 1982.  Approval Guide  for Equipment, Materials,
     and Services for Conservation of Property.  Norwood, Massachusetts.

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

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

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

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/IERL
     Contract No. 68-03-2560, Task No. T5013).  February.

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).
     February.  [Pico Rivera].

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

Kezerle, J.   1981.   Long-Term Evaluation of a Carbon Adsorption System
     for a Petroleum Dry Cleaning Plant.  TRW Inc.  Research Triangle
     Park, North Carolina (EPA Contract No.  68-03-2560, Task No. T5016).
     Apri1.

Kezerle, J.   1982.   Simultaneous Testing by TRW and South Coast Air
     Quality Management District of a Carbon Adsorption System  at a
     Petroleum Dry Cleaning Plant.  TRW Inc.   Research Triangle Park,
     North Carolina (EPA Contract No.  68-02-3174).  January.
                                 4-31

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

Kennes, F.  1982.  Factory Mutual System, Telecon with S. Plaisance, TRW
     Inc., June 24.  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.  EPA/IERL,
     Cincinnati,- Ohio (EPA Publication No. 600/2-80-145).  June.

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

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).  February.

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. 68-03-2560,
     Task No.  T5013).  February.  [Rhode Island].

Washex.  1973.  Installation, Operation and Maintenance Manual for
     Washex Vacuum Stills.  Publication No. T-513d.  Wichita Falls,
     Texas.  July.
                                  4-32

-------
                    5.   MODIFICATION  AND  RECONSTRUCTION
      Standards  of performance  are applicable  to those  facilities  listed
 in  the  standards whose  construction, modification, or  reconstruction
 commenced  (as defined under 40 CFR  60.2(i)) after proposal of the standards.
 Such  facilities are termed "affected facilities."  Standards of performance
 are not applicable to facilities whose construction, modification, or
 reconstruction  commenced on or before proposal of the  standards.  These
 facilities are  referred to as  "existing facilities."   However, an existing
 facility may become an  affected facility and  therefore subject to standards,
 if  the  facility undergoes modification or reconstruction.
      Modification and reconstruction are defined under 40 CFR 60.14 and
 60.15,  respectively.  These general provisions are summarized in Section 5.1.
 Section 5.2 discusses the applicability of these provisions to process
 facilities in the petroleum solvent dry cleaning industry.
 5.1   SUMMARY OF MODIFICATION AND RECONSTRUCTION PROVISIONS
 5.1.1  Modification
     With certain exceptions,  any physical or operational change to an
 existing facility that would result in an increase in the emission rate
 to the atmosphere of any pollutant to which a standard applies, would be
 considered a modification within the meaning of Section 111 of the Clean
 Air Act.  The key to a modification determination is whether total
 emissions to the atmosphere (expressed in kg/hr) from the facility as a
whole have increased as a result of the change.  For example, if the
 affected facility is defined as a group of pieces of eqipment, then the
 aggregate emissions from all  the equipment must increase before the
 facility will be considered modified.
     Exceptions which allow certain changes to an existing facility
without it becoming an affected facility, irrespective of an increase in
 emissions, are listed below.

-------
     1.   Routine maintenance, repair, and replacement.
     2.   An increase in production rate without a capital expenditure
          (as defined in 40 CFR 60.2(bb)).
     3.   An increase in the hours of operation.
     4.   Use of an alternative fuel or raw material if, prior to the
          standard, the existing facility was designed to accommodate
          that alternate fuel or raw material.
     5.   The addition or use of any system or device whose primary
          function is the reduction of air pollutants, except when an
          emission control system is removed or is replaced by a system
          determined by EPA to be less environmentally beneficial.
     6.   Relocation or change in ownership of the existing facility.
     Once an existing facility is determined to be modified, all of the
emission sources of that facility are subject to the standards of
performance for the pollutant whose emission rate increased and not just
the emission source which displayed the increase in emissions.  However,
a modification to one existing facility at a plant will not cause other
existing facilities at the same plant to become subject to standards.
     An owner or operator of an existing facility who is planning a
physical or operational change which may increase the emission rate of a
pollutant to which a standard applies shall notify the appropriate EPA
regional office 60 days prior to the change, as specified in §60.7(a)(4).
5.1.2  Reconstruction
     An existing facilty may also become subject to new source performance
standards if it is "reconstructed."  As defined in 40 CFR 60.15, a
reconstruction is the replacement of the components of an existing
facility to the extent that (1) the fixed capital cost of the new components
exceeds 50 percent of the fixed capital cost of a comparable new facility,
and (2) it is technically and economically feasible for the facility to
meet the applicable standards.  Because EPA considers reconstructed
facilities to constitute new construction rather than modification,
reconstruction determinations are made irrespective of changes in emission
rate.
     The purpose of the reconstruction provisions is to discourage the
perpetuation of an existing facility (instead of replacing it at the end
                                 5-2

-------
of its useful life) for the sole purpose of circumventing a standard
which is applicable to new facilities.  Without such a provision all but
vestigal components (such as frames, housing, and support structures) of
the existing facility could be replaced without the facility being
considered a "hew" facility subject to new source performance standards.
If the facility is determined to be reconstructed, all of the provisions
of the standards of performance applicable to that facility must be
complied with.
     If an owner or operator of an existing facility is planning to
replace components and the fixed capital cost of the new components
exceeds 50 percent of the fixed capital cost of a comparable new facility,
the owner or operator shall notify the appropriate EPA regional office
60 days before the construction of the replacements commences.
5.2  PETROLEUM SOLVENT DRY CLEANING AFFECTED FACILITIES
     In order to cover all significant sources of emissions from petroleum
solvent dry cleaning plants, the affected facility will be defined as
each of the primary VOC emitting pieces of equipment in a facility.
This equipment comprises the following:  dryers, washers, vacuum stills,
solvent filtration systems, and settling tanks.
     The modification and reconstruction provisions do not significantly
impact the petroleum solvent dry cleaning industry.   Facilities in this
industry seldom reconstruct process equipment to the extent that the
capital  cost of the reconstruction exceeds 50 percent of the fixed
capital  cost which would be required to construct a comparable, entirely
new facility.   Additionally, no equipment changes or modifications are
anticipated to occur that will  increase VOC emissions.
                                 5-3

-------

-------
              6.  MODEL PLANTS AND REGULATORY ALTERNATIVES

      This chapter defines the model plants which have been selected to
 represent the petroleum solvent dry cleaning industry and the alternative
 means by which volatile organic compound emissions can be regulated.   :
 6.1  MODEL PLANTS                                                     ;
      Five model plants - small commercial, medium commercial, Targe
 commercial,  small industrial, and large industrial  - have been developed
 to  represent the various configurations of equipment and throughput in
 the petroleum solvent dry cleaning industry.   The model  plants are based
 on  information obtained from plant visits and surveys,  as well as
 consultations with  industry, trade association,  and equipment manufacturing
 representatives (Fisher,  1980a;  Fisher,  1980b;  Sluizer,  1981; Marvel,  et
 al,  1980).   These five model plants  will  be used to estimate  the
 environmental,  energy,  and economic  impacts of  the  various  emission
 control  alternatives  for  the petroleum  solvent  dry  cleaning industry.
 Petroleum  solvent dry cleaning facilities  have  varying operating
 characteristics  and equipment configurations  that prevent explicit
 categorization  of these facilities into specific model plant
 classifications.  Also, the  model plants are  not distinctly indicative
 of the .actual plant capacities that  exist  in  the petroleum  solvent  dry
 cleaning industry.  Consequently, the model plants were classified
 primarily by plant throughput and, in part, by the types  of articles
 cleaned to simplify the representation of the range of sizes  and types
 of dry cleaning plants.
     Commercial petroleum solvent dry cleaning plants offer dry cleaning
 services to the general public and to institutional customers, primarily
cleaning personal items such as suits, coats, dresses, and uniforms.
Industrial petroleum solvent dry cleaning plants offer dry cleaning
services primarily to industrial  customers in conjunction with the

-------
rental of industrial items such as mops, shop towels, work gloves, and
floor mats.  The model plant parameters based on existing equipment are
presented in Table 6-1 for the five model plants.
     The small commercial model plant category presently comprises about
600 operating establishments nationwide or 10 percent of the aggregate
commercial petroleum solvent dry cleaning industry (Fisher, 1980b).  The
average daily throughput for this model plant is approximately 56 kilograms
of general apparel.  The soil loading of the articles cleaned for this
model plant category is relatively low, and the annual petroleum solvent
consumption ranges from 3,000 to 6,800 liters.  The small commercial
model plant includes one washer, one standard dryer, one vacuum still,
and 7 filter cartridges.
     The medium commercial model plant category presently comprises
4,140 operating establishments nationwide or 67 percent of the aggregate
commercial petroleum solvent dry cleaning industry (Fisher, 1980b).
With an average daily throughput of approximately 128 kilograms of
general apparel, this model plant is identical to the small model with
the exception of the larger throughput capacity. As a result of this
larger throughput, a larger volume of petroleum solvent, 6,800 to
15,600 liters per year, is consumed.   This model plant includes one
washer, one dryer, one vacuum still,  and 10 filter cartridges.
     The large commercial model plant category presently comprises
1,400 operating establishments nationwide or about 23 percent of the
commercial petroleum solvent dry cleaning industry (Fisher, 1980b).
These plants may have contracts to clean hospital uniforms or other
types of uniforms in addition to general apparel.  The average daily
throughput is approximately 328 kilograms, or more than double the
throughput capacity of the medium commercial model plant.  The large
commercial model plant consumes 17,300 to 40,100 liters of petroleum
solvent annually and includes one washer, one dryer, one vacuum still,
and 21 filter cartridges.
     The small industrial model plant category presently comprises
60 operating establishments nationwide or 26 percent of the aggregate
industrial petroleum solvent dry cleaning industry (Sluizer, 1981).  The
average daily throughput is approximately 700 kilograms of industrial
                                 6-2

-------









co
O£
LU
g

 t.
t. 4->
re u>
_1 3
TJ
C
r—
re
re <->
E Ul
to 3
•a
c



15
Ol ••-
?£
re at
_J E
U





E3
5 a
TJ Q)
2; M
0
u




"re
*— VJ
E 0)

O
U










I_
S
t-
re
Cu


o
p—
rH


0







1

rH





0

"*•







0
o
10










in
•*2
re
"a. en
oien
C rH
 ^*»
in rH




O O
O 0
rH VO
0 •»
in ID





o o
o o
CO 10
•> M
en CM







o o
o o
CO rH
•t M

*•-*




1
o.

Uj_
GQ
U U)
0) i—
|S
Ul
(A
II
c +J
c

o o o
f— •> !••* t/*»
O O CM
irTo"
m o
i-T

00 O
O O ID
M A
CM O

v^



if i

CM O
CO CO





o o o
o o in
O O CM

CM O
^s






'•"•' O
o o in
O O CM
0 0

rH f*>









a.
Ol
J= O
gfi* n3
<^> v a)
•5..S g-^
"re 15 re
3 0> E TJ
<= Ji E v— »
< z

•r- at CM me?
Ij-^ CMO
3 '-5 ""^
•a t-
C (D
(Q Ul y**N
•i- a> i-» tn o
t/i *^™ ^i
3 I *

c ra
HH

r-»
£
CX rH in O
re <• o

2
g

c
ai
C3



re
a. rH vo m
re rH t»)
f*m
re
i-
i
o








O J=
u ,
ui x: +J
Q. o u -a
i4 fe S;0
1— Z 3t

§O * O CM OO O CO OO <: CMOO
^OrH^J- COO^* OO^s. OO
rH «r^ rH ^- cf» m z r^. in
rH in rH

OO CO O CO LOO O rH OO CO rHOO
wi C3 *^  m f^* m
^-^ *^*^ •* *





100 CTl LO rH inS" O rH oS rH "J*
POOO CM ^-OCO OOCM^N.

•





*~* ^~* s-\ re
to m co in rH roo o I-H oo o  IB
•b> c- J= « ai
j<= ai re ai "^
_o> a. ui a> TJ ui >>
0) . U) *> 01 •— T C T-
X TJ 3 /-s 4->>— >, -*^ O> r— .Si O
I- -C E O) >, O>3US-4->4->Cr—
JJ jg.rj>,*J _EU JIr-*ire^
SU> i*- r— t- Q.U1 4> «•- ^J O) 14- l«- c ^
^3 O U O re.C3r— O Ul *~* O O ^ U »— O P— (-
cn>^^ u u f- ^-x >) u €t> u i-4->a)
•< Z 5 Z 0^ 0 J 5"* Z J"""






















ui
tt>
TI
"Tj
T^
re
O)
i
c
0)
^

u>
O)
CO
CM
O
01
re
u

c
o

Ul
Ul
- "re
v c
a o
i^- tr
a. o
a.
re TJ
a>
£ J?
5-3

-------
articles such as rental uniforms, rugs, mops, and mats.  The small
industrial model plant includes one washer, three dryers, one vacuum
still, and 63 filter cartridges.  However, some small industrial plants
may omit solvent filtration, employing instead a solvent settling tank
prior to distillation.
     The large industrial model plant category presently comprises
170 operating establishments or about 74 percent of the nationwide
industrial petroleum solvent dry cleaning industry (Sluizer, 1981).  The
large industrial model plant has an average daily throughput of
approximately 2,442 kilograms of industrial articles such as rental
uniforms, mops, rugs, and mats.  These plants often operate with multiple
shifts and require equipment with large capacities for almost continuous
operation.  This model plant includes two washers, two large capacity
standard dryers, three vacuum stills, and two settling tanks.  Typically,
solvent filtration systems are not used by large industrial plants
because of the excessive soil loading of the articles cleaned.
6.2  REGULATORY ALTERNATIVES
     The purpose of this section is to define different regulatory
alternatives based upon their effectiveness in reducing VOC emissions.
The regulatory alternatives are based on the emission control techniques
discussed in Chapter 4.  The VOC emissions resulting from the application
of the control technology mandated by each regulatory alternative are
given in Table 6-2 for each model plant.   Ranges are used to illustrate
that significant variations in VOC emission rates can occur due to
variations in operating procedures.
     Alternative I requires no additional regulatory action other than
what presently exists for the industry.  This alternative is representative
of the equipment and emission levels that would exist without any additional
control requirements.  Under this alternative, standard dryers emit
solvent vapors directly to the atmosphere, cartridge filters are in
general use for solvent filtration without specific drainage periods,
and VOC emissions from fugitive sources such as disposed vacuum still
and settling tank wastes, as well as general vapor and liquid leaks, are
uncontrolled.
                                 6-4

-------
 Table 6-2.. MODEL PLANT EMISSION RATES FOR  REGULATORY  ALTERNATIVE I,  II, AND  III
(emission  rates  in kiloprariis of  VOC emitted  per  100 kilograms of  articles cleaned)
General fugitive emissions
Regulatory Model*
alternative plant
SC
MC
I LC
SI
LI
SC
MC
II LC
SI
LI
SC
MC
III LC
SI
LI
Standard
dryer
14-28
(18)1
14-28
(18)
14-28
(18)
14-28
(18)
14-28
(18)
14-28
(18)
14-28
(18)
14-28
(18)
14-28
(18)
14-28
(18)
b
b
b
b
b
Recovery
dryer
b
b
b
b
b
b
b
b
b
b
0.7-9.5
(3.5)
0.7-9.5
(3.5)
0.7-9.5
(3.5)
0.7-9.5
(3.5)
0.7-9.5
(3.5)
Cartridge
filter
0.5-1
(1)
0.5-1
(1)
0.5-1
(1)
0.5-1
(1)
c
0.5-1
(1)
0.5-1
(1)
0.5-1
(1)
0.5-1
(1)
c :
0.5-1
(1>
0.5-1
(1)
0.5-1
(1)
0.5-1

c
Vacuum
still
1-7
(3)
1-7
(3)
1-7
(3)
1-7
(3)
1-7
(3)
1-7
(3)
1-7
(3)
1-7
(3)
1-7
(3)
1-7
(3)
1-7
(3)
1-7
(3)
1-7
(3)
1-7
(3)
1-7
(3)
Miscellaneous
fugitive
0.5-1
(1)
0.5-1
(1)
0.5-1
(1)
0.5-1
(1)
0.5-1
(1)
0.5-1
(1)
0.5-1
(1)
0.5-1
(1)
0.5-1
(1)
0.5-1
(1)
0.5-1
(1)
0.5-1
(1)
0.5-1
(1)
0.5-1
(1)
0.5-1
(1)
Total
16-37
(23)
16-37
(23)
16-37
(23)
16-37
(22)a, (23)
15.5-36
(22)
16-37
(23)
16-37
(23)
16-37
(23)
16^37
(22)d, (23)
15.5-37
(22)
2.7-18.5
(8.5)
2.7-18.5
(8.5)
2.7-18.5
(8.5)
2.7-18.5
(7.5)°, (8.5)
2.2-17.5
(7.5)
aNumbers in parentheses are nominal values representing an overall emission rate (either an
 average of test data or a value widely accepted in the industry).

bNot applicable.                                           >.

cLarge Industrial plants normally oait filtration.

 Small Industrial plants sometimes omit filtration with heavy soil loading.

CSC * snail commercial.
 MC = medium commercial.
 LC = large commercial.
 SI = small industrial.
 LI = large industrial.
                                                6-5

-------
     Alternative II would reduce overall VOC emissions from dry cleaning
plants by means of an effective maintenance program which would eliminate
perceptable fugitive leaks, and would minimize cartridge filter emissions
with a minimum drain time of 8 hours for used cartridge filters prior to
disposal.  This alternative would result in no specific emission reductions
beyond baseline emission because reductions due to leak repair and
cartridge drainage have not been quantified.  The nominal emission rate
for commercial model plants and small industrial model plants employing
filtration would be 23 kilograms (ranging from 16 to 37 kilograms) VOC
per 100 kilograms of articles cleaned under Alternative II.  The nominal
emission rate for industrial model plants omitting filtration would be
22 kilograms (ranging from 15.5 to 37 kilograms) VOC per 100 kilograms
of articles cleaned under Alternative II.
     Alternative III mandates the use of a recovery dryer instead of the
standard dryer in addition to the filtration and fugitive emission
control provisions of Alternative II, making this the most stringent
alternative.  The nominal emission rate is 8.5 kilograms (ranging from
2.7 to 18.5 kilograms) VOC per 100 kilograms articles cleaned for all
model plants with the exception of the small (omiting filtration) and
large industrial plant, which have a nominal emission rate of 7.5 kilograms
(ranging from 2.2 to 17.5 kilograms) VOC per 100 kilograms articles
cleaned.  Alternative III produces a nominal VOC emission rate decrease
of about 14.5 kilograms (ranging from 13.3 to 18.5 kilograms) VOC per
100 kilograms articles cleaned relative to baseline emission levels.
                                 6-6

-------
6.3  REFERENCES FOR CHAPTER 6


Fisher, W.  1980a.  International Fabricare Institute (IFI) Meeting with
     S. Plaisance, TRW Inc., December 9.  Number of cartridge filters in
     use.

Fisher, W.  1980b.  IFI, TeTecon 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.

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.

Sluizer, M.  1981.  Ill, Telecon with S. Plaisance, TRW Inc., April 10.
     Size of the industrial petroleum dry cleaning industry and the
     throughput of a typical plant.
                                 6-7

-------

-------
                         7.   ENVIRONMENTAL IMPACTS

      The  environmental  and  energy  impacts associated with  the  implementation
 of  each Regulatory  Alternative  (presented in  Chapter 6)  are  discussed  in
 this  chapter with respect to air quality, water  quality, solid waste,
 and energy  consumption.  Both beneficial  and  adverse impacts are assessed.
 7.1  AIR  IMPACTS
      Annual VOC emissions and emission  reductions  in the five  model
 plants under the three  Regulatory  Alternatives are illustrated in Table 7-1.
 The model plant emission rates  for each Regulatory Alternative are
 developed in Chapter 6, and  are multiplied by the  model plant  annual
 throughput  to derive the model plant annual VOC  emissions.   Ranges are
 used  to illustrate  that significant variations in  VOC emission rates can
 occur due to variations in operating procedures.   Alternative  I (baseline)
 emissions are proportional to plant throughputs, with the  small commerdial
 model plant having  the  lowest annual emission range (2.2 to 5.2 megagrams
 VOC per year) and the large  industrial model  plant  having  the  highest
 range (98.4 to 229  megagrams VOC per year).   Because Alternative II
 emission reductions resulting from fugitive emission control are not
 quantifiable, the Alternative II emission  rates  remain unchanged from
 those given under Alternative I in Table  7-1.   Under Alternative III,
model plant annual   emissions range from a  low in the small commercial
plant of 0.4 to 2.6 megagrams VOC per year to a maximum of 14.0 to
111 megagrams VOC per year in the  large industrial plant.  Emission
 reductions resulting from the implementation of Alternative III are also
 listed in Table 7-1 and are proportional  to the model  plant throughputs.
The incremental  annual  emission reductions of Alternative  III  over
Alternative I range from a minimum of 1.8 to 2.6 megagrams VOC per year
 in the small commercial  plant to a maximum of 84.4 to 118 megagrams VOC
per year in the large industrial model plant.

-------













to
LU
I-H
5
Z!
LU
f—
_4
*£
oi
O
	 1
ZD
CD
LU

LU
LU
O£ CO

r~* ^*
^£
LU — J
:E Q-
U_ LU
O 0
o
CO E
r-
O LU
^C ^^
a. n
S u_
\—t
LU
z in
O 1—
t— 1
co z:
co i— i
i— i
s:
LU

O
O


	 1

-P -r- i-
U -P >>
3 nj ^^s.
•a c o
CO i- O
s- 
-P
c i— en
0 < S
•1— V— '
to cu
'E •!-
03 i—
cu
o w
^D fl3
^to ^**}




1 "1
i— i
t— i











i—i
i— i










S-
>^
"•^
(_j
o


en
c.
•r—

V)
c.
0
•P—
U}
IO

E
O)

O
o


r—
re

c
c
•^









t-t
1—1
i— i











i— i
1—1















t—t











/ — s
s-

"x^
en

i ^
c. c:
re ••-

P> -p
^3
•— D.
ai JT
•a en
O 3
s: o
$-
x:
•P


to

CM
1

00

i-i








re







to
•
CM
1

^f1
•
O





CM
,
ID

1

CM

CM







CM

LO

|

CM

CM













r—
re
•r—
U
s-
0)
E i^^s
E 0
O O
U O

r— ^t"
r— rH
re *— '
E
to

CD

ID
1

CM

<5J"








re







CD
.
ID
1

CD
«
0




00

l-H
rH

|

rH

LO






00

rH
rH

1

rH

ID












P—
re

o

ai
E
E /-»
0 0
U 0

E -
3 CM
•r- CO

a>
E;
CM
,
LO
rH
1

O"V
.
CD
rH







re






CM
•
LO
rH
1

CM
«
CM




CO
,
CD
CO

1

rH

CO
rH





CO

O
CO

1

rH

CO
rH












p—
re
•r—
O
t-
cu

E CD
0 0
U 0

a> CM
en oo
5_ s 	 i
re
— i
r-.
.
CO
CO
1

CM
•
^«
CM







re






r-.
,
oo
CO
1

CD
•
^J"




co
•
p^.
CD

1

CM
,
00
CM





co

r*«».
to

i

CM
.
00
CM












r—
re
•r—
i.
+3
to /*••»
3 CD
•a CD
c. o
•r- *»
CM
r- 00
r- rH
re > — '
E
to
LO
•
r^.
rH
1

LO
.
^-
00







re






rH
•
rH
rH
! — |
1

CD
•
^J"
rH



to
•
00
CM
CM
|

*^*
,
00
CD






> c
•r- O
•P U
re
c c
S- 0
5 OJ
i-
O <1J
•p >
re •<-
<— -P
3 T-
cn en
O) 3
oi it-
re
7-2

-------
      Cumulative nationwide VOC emissions associated with each regulatory
 alternative were derived by multiplying the cumulative number of affected
 model  plants existing through each year'by the appropriate model  plant
 VOC  emissions.   The  projected cumulative number of model  plants  is
 presented in Table 7-2,  and the annual  VOC emissions were obtained  from
 Table  7-1.   Cumulative nationwide  emissions are presented in  Table  7-3
 for  each  model  plant category for  the ten years following proposal  of
 the  standards.   Ranges are used to illustrate  that significant variations
 in VOC emission rates can occur due to  variations  in operating parameters.
 Because of  the  larger number of plants,  the commercial  model  plant
 category  produces  a  larger quantity of  nationwide  VOC  emissions  than the
 industrial  model plant category.   The medium commercial model  plant
 category  is  the most numerous  category,  and contributes the largest
 total  quantity  of  VOC emissions  in the  commercial  sector.  Nationwide
 ten-year  emissions from  the  medium commercial  model  plant category
 decrease  from a range of 38,700  to 89,600 megagrams  VOC per year  under
 Alternative  I to a range of  6,800  to 44,800  megagrams  VOC per year under
 Alternative  III.  The largest  single source  of  VOC emissions,  however,
 is the  large industrial  model  plant category.   This  category produces
 the  highest nationwide ten-year  emissions  (43,300 to 101,000 megagrams
 VOC  pear year under  Alternative  I  and 6,200  to  48,800 megagrams VOC per
year under Alternative III).   These high  emission levels  result from the
 large throughput capacity of this  category.  The smallest  single  source
 of VOC  emissions is  the  small  commercial  model   plant category, due to
 the  relatively  limited number  of plants and  low throughput capacity.
The  small  commercial  model plant category has ten-year nationwide emissions
of 2,400 to 5,700 megagrams VOC per year  under  Alternative I and from
400 to 2,900 megagrams VOC per year under Alternative III.  Total nationwide
emissions  are at a maximum under Regulatory Alternative I  (123,000 to
286,000 megagrams VOC per year) and are lowest  under Regulatory
Alternative III (19,900 to 141,000 megagrams VOC per year).  This difference
is expected because Alternative I requires no additional controls above
baseline while Alternative III requires  the most effective emission
controls.
                                 7-3

-------
   to
CO
Q CO
UJ
r- UJ
O Z
-
 01 UJ
JO
 to Di
   O
                   O

                    "to
                  0) «r-
                  D> S-
                  J- +J
                  to  in
                 —I  3
                    -a
                     to

                 ^ tl
                  to -p

                 tO 3
                     TD
                     C
                     to
                  Q) -r-
                  c» u
                  S- J-
                  ro cu
                     o
                     (J
                  E
                  0)
   to
   'r—
   O

   0)
                     O
                     o
   fO

   o
to s-
E 0)
                  to
                     E

                     O
                     u
                                 O OJ CO
                                 O rH i-H
                                 IO ID tO
                                                 CMCMCMCMCMCMCvJCMCMCSJCsJ
                                 000
                                 CD CD O
                                 r-t r-t rH
000
LO m in
                                                 oooooooooooooooooocooo
                                                 cocooorocorocooocooooo
                                                 rHi-HrHrHrHrHrHrHi-HrHrH
000
o o o
to to to
                                 cn CD i-H
                                 I— CO OO
                                 cn cn cn
                                 i-H rH r-H
                                                 OOOOOOOOOOO
                                                 CMCSJCMCSJCMCMCMCMCVJCMO-J
                                                        >r~r
                                                 oooooooooocooooocncncn
                                                 cn cn cn cn en en cn cn cn cn cn
                               O)


                              -P  ••
                               in  c:

                               X  "~
                               cu  a)
                                  -a

                               o  5

                               i-  o
                               CO  T-
                              J2  -P
                               E  ta
                               3  C
                               c
                                  in

                               tO  d
                              •P  tO
                               O  i—
                              r-  O.
                                              •a
                                               
4_j
c
tt)
u
5-

C
(0
r— •










.
in
r—
^~
• ^'
[ ^
in
E
3
3
O
fO
>

TJ
C
to

in
s_
ai
^>j
S-
Q.TJ

r-—
to
•r—
O
5-
Ol
E
PS
O
(J

C
•1—

Ol
4_5
(O
S-
.c
"5
o
s-
cn

o
s-
,

C3
oo

4-
O

CJ
4-
'^
+J
C
O)
E
O.
• r--
3
CT
<1)

-o
a>
u
*J

S-
ai
a.

in

r~
to
I—
o.

T3
0)
-P
O
O)
4-
4—
tO

r— -
to
c
o
• r—
•P
• p—
-o
•a
to

o
s_
ai

E
3
C

a>
0)
to
t-

4-
4—
re)

r—
to
c:
o
•r—
+J
•r-.
•o
•a
m

0
S-

D5
m
s_
a>
to

o
r"
I—
U
                                                   7-4

-------
Table 7-3.   CUMULATIVE  NATIONWIDE VOC IMPACTS OF  THE FIVE MODEL PLANTS
             UNDER  THREE  REGULATORY ALTERNATIVES FOR TEN  YEARS
Hodel
plint
category
SMll
coMercial








HcdiuR
coemrcial








Large
comer cial








Snail
Industrial








Large
Industrial








Total
Industry








Cuaulative nationwide Missions
through each vear. Ma VOC
Year
1
2
3
4
5
6
7
8
9
10
1
2
3
4
5
6
7
8
9
10
1
2
3
4
5
6
7
8
9
10
1
2
3
4
5
8
7
8
9
10
1
2
3
4
5
6
7
8
9
10
1
2
3
4
5
6
7
8
9
10
Alternative I
44-104
132-312
264-624
440-1,040
660-1,560
924-2.180
1,232-2,910
1,580-3,740
1,980-4,680
2,420-5,720
704-1,630
2,110-4.890
4,220-9.770
7,040-16,300
10,600-24,400
14,800-34,200
19,700-45.600
25,300-58,600
31,700-73,300
38,700-89,600
616-1,420
1,850-4,27.0
3,690-8,540
6,160-14,200
9,240-21,400
12,900-29,900
17,200-39,900
22,200-51,300
27,700-64.100
33,900-78,300
85-202
254-606
508-1,210
846-2.020
1.270-3,030
1.780-4,240
2.370-5,650
3,050-7.270
3,810-9,090
4,650-11,100
787-1,830
2,360-5,500
4,720-11,000
7.870-18.300
11,800-27,500
16.500-38,500
22,000-51.300
28,300-66.000
35,400-82,400
43.300-101,000
2.240-5,190
6.710-15,600
13,400-31.100
22.400-51.900
33,600-77.900
46,900-109,000
62.500-145,000
80.400-187.000
101,000-234,000
123,000-286.000
Alternative II*
44-104
132-312
264-624
440-1,040
660-1,560
924-2,180
1.232-2,910
1.580-3.740
1.980-4,680
2,420-5.720
704-1.630
2,110-4,890
4,220-9,770
7,040-16,300
10,600-24.400
14,800-34,200
19,700-45,600
25,300-58,600
31,700-73,300
38.700-89,600
616.1,420
1,850-4.270
3.690-8,540
6,160-14,200
9,240-21,400
12,900-29,900
17,200-39,900
22,200-51.300
27.700-64,100
33.900-78.300
85-202
254-606
508-1,210
846-2,020
1,270-3,030
1,780-4,240
2,370-5,650
3.050-7.270
3,810-9,090
4,650-11.100
787-1,830
2,360-5,500
4,720-11.000
7,870-18,300
11,800-27,500
16,500-38,500
22,000-51,300
28.300-66.000
35.400-82.400
43,300-101,000
2,240-5,190
6.710-15.600
13,400-31,100
22,400-51.900
33,600-77,900
46,900-109.000
62.500-145,000
80,400-187,000
101,000-234,000
123,000-286,000
Alternative III
8-52
24-156
48-312
80-520
120-780
168-1.090
224-1.460
288-1.870
360-2,340
440-2,860
124-814
373-2.440
745-4.890
1,240-8,140
1,860-12,200
2,610-17,100
3,480-22,800
4,470-29,300
5.590-36,600
6,830-44.800
103-714
310-2,140
620-4,290
1,030-7.140
1.550-10,700
2,170-15,000
2,900-20,000
3,720-25,700
4.650-32,100
5,690-39,300
15-101
44-303
88-607
147-1,010
221-1,520
309-2,120
412-2,830
529-3,640
662-4,550
809-5.560
112-888
336-2,660
672-5,330
1.120-8,880
1,680-13,300
2,350-18,600
3,140-24,900
4.030-32,000
5,040-40.000
6.160-48.800
362-2,570
1,080-7,690
2,160-15,400
3,620-25.700
5,420-38,500
7.640-53.900
10,200-72.000
13.100-92.800
16,300-116.000
19.900-141,000
Cunulatlve nationwide
eeission reduction
through each year
under Regulatory
Alternative III. Hg/VOC
36-52
108-156
216-312
360-520
540-780
756-1,090
1,010-1,450
1,290-1,870
1,620-2,340
1,980-2,860
580-816
1,740-2,450
3,480-4,880
5,800-8,160
8,740-12,200
12,200-17,100
16,200-22,800
20,800-29.300
26,100-36,700
31,900-44,800
513-706
1,540-2,130
3,070-4,250
5,130-7,060
7,690-10,700
10,730-14,900
14,300-19.900
18,500-25,600
23,100-32,000
28.200-39,000
70-101
210-303
420-603
699-1,010
1,050-1,510
1,470-2,120
1,960-2,820
2,520-3,630
3,150-4,540
3,840-5,540
675-942
2,020-2.840
4,050-5,670
6,750-9,420
10,100-14,200
14,200-19,900
18,900-26,400
24,300-34,000
30,400-42,400
37,100-52,200
1,880-2,620
5,630-7,910
11,200-15,700
18,800-26,200
28,200-39,400
39,300-55,100
52,300-73,000
67,300-94,200
84,700-118,000
103,000-145,000
 Regulatory Atlernetive II Mission reductions resulting froa fugitive emission controls
 are not quantifiable.
                                       7-5

-------
      The  nationwide  VOC  emission  reductions  associated with  Regulatory
 Alternative  III  for  each model  plant category  are  also presented  in
 Table 7-3 for  the  ten-year period following  proposal  of  the  standards.
 These emission reductions were  derived  by  subtracting the  given Regulatory
 Alternative  emissions  from the  baseline emissions  in  Table 7-3.
 Alternative  I  is representative of the  existing  baseline and therefore,
 results in no  emission reduction.   Regulatory  Alternative  II emission
 reductions resulting from fugitive emission  control are  not  quantifiable
 and are consequently,  not presented.
      The  largest potential  for  total  ten-year  VOC  emission reduction in
 the commercial sector  is in the medium  commercial  model  plant category
 which has  reductions ranging  from 31,900 to  44,800 megagrams VOC  per
 year.  The large industrial model  plant category with reductions  ranging
 from  37,100  to 52,200  megagrams VOC  per year,  has  the largest total
 ten-year VOC emissions reduction  relative  to the other four  model plant
 categories.  Th small  commercial  model  plant category with reductions of
 1,980 to 2,860 megagrams VOC  per  year,  has the least  total ten-year VOC
 emissions  reduction.   The large Regulatory Alternative III emission
 reduction  results primarily from  the  requirement of replacing standard
 dryers with  recovery dryers.  Alternative  III  results in a 50 to 85 percent
 emission reduction from  baseline  depending on  the  operating  parameters
 of the plant.
 7.2  WATER IMPACTS
     Water pollution impacts  in petroleum solvent  dry cleaning plants
 result from  the production  of wastewater that may  contain  residual
 solvent.   This wastewater is produced by sources such as recovery dryers
 and vacuum stills.   The  disposal  of sol vent-laden  cartridge  filter
 elements in  landfills could increase the quantity  of VOC in  groundwater
 (see Chapter 4).   However,  cartridge filtration  is assumed to be the
predominant  form of  filtration  in the industry and, as such,   is unaffected
 under the three Regulatory Alternatives developed  in Chapter 6.   Even
though the quantity of solvent  in disposed still  waste is  reduced through
 improved operation and maintenance procedures under Regulatory
Alternatives II and  III,   water pollution resulting from incomplete
                                 7-6

-------
 separation of solvent and water in the vacuum still  is assumed to be
 unaffected.
      The primary source of VOC water pollution resulting from implementation
 of emission  controls is the recovery dryer (see Chapter 4).   In this
 unit,  the condensed mixture of solvent and water is  separated by a
 gravimetric  separator.   In the gravimetric separator the difference  in
 the densities of solvent (0.75 kilograms/liter) and  water (1.0 kilograms/
 liter) provides the basis for separation.   In an EPA field test of a
 recovery dryer, the most recent and complete test data indicate that the
 unit recovered an average of approximately 3.29 kilograms of water per
 100 kilograms of articles cleaned (Plaisance et a!.,  1981).   The separator
 on the domestically manufactured recovery  dryer is designed  such that
 the unit must be level  in order to prevent solvent from entering the
 stream of recovered water.   If the separator housing  is not  level, the
 recovered water that is typically disposed in a public sewer will  contain
 additional petroleum solvent.
      If the  recovery dryer gravimetric  separator is properly installed
 and maintained,  the quantity of solvent contained  in  the  sewered water
 would  be limited to that which is  dissolved  in  the recovered water.   A
 representative  of a petroleum solvent manufacturer has  estimated the
 maximum solubility of petroleum solvent in water as 100 parts per million
 by  volume  (ppmv)  (Saary,  1981).  This value  is  combined with  the previous
 estimate of water recovery  as  a function of  cleaning  throughput  to form
 the  basis  for the determination of  the  mass  of  solvent  contained in
 sewered  water.
     Annual water pollution  impacts in  the five  model plants are listed
 in Table 7-4, based on  the previously discussed  assumption that these
 impacts  will  only result  from  the recovery dryers required under Regulatory
Alternative III.  The quantity  of petroleum solvent contained in the
sewered wastewater  from the  recovery dryer is insignificant in comparison
to the total  annual quantity of solvent-laden wastewater disposed.
Furthermore,  both the quantity  of solvent itself and the sol vent-laden
wastewater disposed is a function of the model plant throughput.  The
sol vent-laden wastewater and pure solvent sewered increase with throughput
from the small commercial plant to the large industrial plant.  The
                                 7-7

-------













•z.
t— <

	 1
o
Q*?
l—
z.
o
0
•z.
0
*-*
to
1— I
LU
LU
o

 s:

—1 LU
0 >
CL. t-H
LJ_
rv^
LU
I—

3:

_J
^^
rs
z.
^**
 O)
r— -P
0 03
in 3=
O)
•r- in
co 3: i-
O) >>
CO CO Ol
i- in i«i
u o ^-^
c o.
•i— O)
•r-
i— -a
CO
3 C
c: •!-
c:
>
in to -•>.
fO 3: O>
E 
i— Q) 0)
03 > O
***E p— p> .
com
C Ol T—
^
"x^
O5
4-^ Xt
C
CO C
^— *r—
Q.

r— =3
O) O-
'^_j .^7
O Ol
s: =>
0
S-
r"
^O
^^^








oo co o LD r--
O O CM ^t- Lf>
O CD CD CD ' r-i

























IO LD O CO
•=i- o r-- CD cn
• « • • .
CD rH CM LSI O
C\J





























i—
r- (0 •— r— r-
fO •!"" (TJ fO ("0
•1- O T- •!- T-
U S- O S- S-
S_ OJ 1- +J -P
a) E  E '~» E /-^ 3 CD 3 CD
eo oo eo 730 TDO
OO UO OCD CCD CCD
UCD O UO .f-^.,-^
E - - CM LT>
i — «3- 3CM CUCM i — CO OJOO
i— i-H -i- CO O1CO i— i-H O1UO
(C ^ T3 ^-^ S- v^ n3 s-^ S- v*>
E O) tO S (13
l/l S — 1 I/) — 1









l — i
i i
i — i

Ol
>
• 1 —
-p
fO
c
^.
O)

! 	
^
^
O
to
1—
o>
c?

s_
01
•a
c
13

in
S-
a;

^,
•a

^>^
j_

c >
O T-
•r- -P
•P fO
ra c
r- S.
t— a)
to -p
i ^ - -
in eC
c
•r- >,
S-
0 0
-p -p
fO
cu i—
3 3
T3 O)
01
in Qi
•P
U O
fO -P
O.
£ (U
•r- >

S- -P
O) ft)
+-> t—
fO OJ
3S !-
10
7-8

-------
small commercial plant sewers 0.5 megagrams per year of wastewater and
0.03  kilograms of pure solvent, while the large industrial plant sewers
21 megagrams of wastewater and 1.6 kilograms of pure solvent.
      Table 7-5 illustrates the cumulative nationwide water pollution
impacts due to the application of recovery dryers under Regulatory
Alternative III.  (As previously stated, Regulatory Alternatives I and
II increase neither the quantity nor the solvent content of sewered
water.)  These were derived by multiplying the annual water impacts for
each  individual model plant, given in Table 7-4, by the appropriate
cumulative number of affected plants projected to be existing through
each  year.  The increased mass of sol vent-laden water is based on the
previous estimate of the recovery dryer water recovery rate.   The increase
in the mass of solvent in sewered water is based on the estimated maximum
solubility of solvent in water.  As can be seen in the table, the solvent
in disposed wastewater is insignificant in comparison to the total
sewered water even at the increased levels.   Because of the lower annual
throughput of articles cleaned, the small commercial model plant category
has the lowest total water impacts for the ten-year period after proposal
of the standards.   The small commercial model plant category has a
cumulative mass of sewered sol vent-laden water of 510 megagrams and a
cumulative mass of sewered solvent for the same period of 33.0 kilograms.
In contrast, the greatest water pollution impacts occur in the large
industrial model plant category.   The cumulative mass of sol vent-laden
water for the large industrial plant category is 9,200 megagrams over
ten years with a cumulative mass of sewered solvent for the same period
of 690 kilograms.   Nationwide water pollution impacts for the ten-year
period increase from 470 megagrams to 25,700 megagrams in the mass of
solvent-laden water disposed, with a corresponding increase from
35 kilograms to 1,900 kilograms in the mass of solvent disposed in the
wastewater.  Thus, while the implementation of Regulatory Alternative III
would increase the projected mass of sewered water nationwide by as much
as 25,700 megagrams, the actual nationwide total  increase in  the mass of
solvent disposed in public sewers would be only 1,900 kilograms over ten
years.
                                 7-9

-------
Table 7-5.   CUMULATIVE NATIONWIDE WATER  IMPACTS OF  EMISSION  CONTROL
                 IN FIVE MODEL PLANT CATEGORIES UNDER
               REGULATORY ALTERNATIVE III  FOR TEN YEARS
Hodel plant
Satall coeaercial









Medfu*) coeaercial









Urge conercial









Satll industrial









Large Industrial









Total Industry









Year
1
2
3
4
5
6
7
8
9
10
1
2
3
4
5
6
7
8
9
10
1
2
3
4
5
6
7
8
9
10
1
2
3
4
5
6
7
8
9
10
1
2
3
4
5
6
7
8
9
10
1
2
3
4
5
6
7
8
9
10
Cumulative
Increase In MSS of solvent-laden
wastewater (Hg)
9.20
27.6
55.2
92.0
138
193
258
331
414
506
145
435
869
1,450
2,170
3,040
4,060
5.220
6.520
7,970
127
4 381
761
'1.270
1,900
2,660
3,550
4,570
5,710
6.980
17.9
53.8
108
179
269
377
502
646
807
987
167
502
1,000
1,670
2,510
3,510
4,680
6,020
7,520
9,200
466
1,400
2.790
4,660
6.990
9.780
13,000
16,800
20,900,
25.700
Cumulative
Increase in solvent In
disposed wastewater (Kg)
0.60
1.80
3.60
6.00
9.00
12.6
16.8
21.6
27.0
33.0
11.0
33.1
66.2
110
166
232
309
397
497
607
9.40
28.2
56.4
94.0
141
197
263
338
423
517
1.35
4.05
8.10
13.5
20.3
28.4
37.8
48.6
60.8
74.3
12.6
37.7
75.4
126
188
264
352
452
365
69JI
35.0
105
210
350
524
800
979
1.260
1.570
1.920
 Vater {•pact* due only to Installation of recovery dryer* under Regulatory Alternative III.
                                  7-10

-------
 7.3  SOLID WASTE IMPACTS
      Solid waste emissions 'in petroleum dry cleaning plants result from
 the disposal  of used cartridge filter elements and vacuum still  residue.
 It is expected that total  industry wide solid waste emissions will  be
 reduced by the implementation of operating and maintenance procedures.
 However,  for  comparison purposes,  it is assumed that solid waste emissions
 would be  equivalent under  the three alternatives because the solid waste
 emission  reductions beyond the Alternative I baseline resulting  from
 operating and maintenance  procedures are unquantifiable.   Consequently,
 a comparison  of solid waste emission levels under the three alternatives
 is omitted.
 7.4  ENERGY IMPACTS
      Although petroleum dry cleaning plants include a wide range of
 equipment that consume electricity,  steam,  and fuel  (oil,  natural  gas,
 or coal),  the energy impacts  include only those associated with  the use
 of a  standard dryer under  Alternatives  I  and II and a recovery dryer and
 refrigerated  chiller under Alternative  III.   Energy impacts  associated
 with  the  implementation of operating and  maintenance  procedures  are not
 quantifiable.   Both standard  (non-recovery)  dryers  and recovery  dryers
 consume steam and electricity, with  standard dryer  consumption rates
 being significantly higher than  those of  recovery dryers (see Chapter 8
 for consumption  rates).  The  standard dryer  consumes  more  energy because
 it  has a  larger  fan  and a  resulting  greater  air flow  which requires more
 steam for  heating.
      Annual energy  impacts  of the three regulatory  alternatives  for the
 five  model plants are  given in Table 7-6.  Dryer energy consumption  is
 determined by  converting the energy  costs (see  Tables 8-9 through 8-13)
 of both steam  and electricity into common units of work (gigajoules).
 The conversion factors  used to convert steam and electricity costs per
year  ($/year)  to units  of work per year (gigajoules per year) are $6.37 per
 gigajoule and  $16.67 per gigajoule,  respectively (Vatavuk, 1980).  As in
 the previously discussed environmental impacts, the overall consumption
 of energy increases proportionally with increased model plant throughput.
The largest annual energy consumption would be produced by the large
                                 7-11

-------






















CO
CO
1—

1
o_
u 	 i
O UJ
a
co o
I— s:
o
- 2:
CD l-«
a;
LU co
Z '"
UJ >•
i— i

^C ^C
"^ p^
z: o:
z: ui

i
^^
.
co >-

t*^» CD
i__
CU 
O -r-
3 P
-o co
CU C
j- s-
S_ 0)
O -v» i—
•i— O 
•r~*
s- c
QJ i-
C CU
LU P
^_

•r-
S- +J
>> CO

0 f-
ca cu
p
C r—
•r- 
in «r-
C P
o co
0 C
t-
>> CU
CJ)P
J- r—
O) 
C •!-
«=C P
CO
C
S-
d)
^_
^^



^^
s-

*x^
D)

P
C C
CO •!-

"o-p
3
-— a.
CU JC
T3 D)
0 3
S O
J-

P


I--. LD
• .
CD ID rH
rH «tf- LD
CM













O O O












00 ID
. .
i-H rH 00
ro oo r--
rH












ID
.
oo r-« en
«3- CM CM
rH *5j











ID

cd r~ cn
•=t- CM CM












i^
t— CO r—
CO -r- CO
•r- 0 -r-
U S- U
S- ^~s CU /— > 5-
CU O SO 0) /*-N
E 0 E 0 E 0
E CO O CO E O
O - U « O CO
U «3- CM U •>
i-H E CO CM
r— v^> 3 v-> CU 00
i — "i— CT)1*-^
co -a s-
E CU CO
CO 21 —1



ID O
CD rH
Is-. CT>
•»
CM











O CO














ID 0
00 rH
«3- Is-- '

rH












O O
LD CM
CM CD
** ••
rH *?}"











0 0
LD CM
CM CO
«> «i
rH •*










r~ r—
CO CO
•r~ •!"-
S- i-
+i /~S P / 	 S
in O in o
3 O 3 O
"O O T3 O
C " C •>
•r- CM •!- ID
00 CO
r— rH >
c
0

CU
•a
3
r—
U
c

•a
c
CO

*»
C
o
•r-
P
a.
E
3
in
c
o
u

E
CO
CU
P
to

•a
c
CO •
to
1 — i-
cO CU
U i—
si -^
p ^
U CJ
CD
r— -a
cu cu
P
M- CO
O 5-
a>
E CD
3 T-
in S-
O) CD
J= S-
p
•a
in c
•r- co

C Oi
0 S-
•r- O)
p >^
a. s-
E T3
3
in >,

0
o -a
0)
>> E
D5 3
i- in
cu c:
c o
LU U
co
7-12

-------
 industrial model plant due to the relatively large throughput capacity
 of this category.  The Ijarge industrial model plant consumes
 4,600 gigajoules per year of energy under Alternative I and II and
 1,700 gigajoules per year under Alternative III.  An energy consumption
 reduction of 2,900 gigajoules per year in the large industrial model
 plant results from the implementation of Alternative III relative to
 Alternative I.   The small commercial model plant which would be the
 smallest consumer of energy of the five model plants,  consumes about
 49 gigajoules per year under Alternative I and II and  about 32 gigajoules
 per year under Alternative III.   An energy consumption reduction of
 about 17 gigajoules per year in the small  commercial model  plant results
 from the implementation of Alternative III relative to Alternative I.
      Cumulative nationwide energy impacts  in the five  model  plant categories
 due to the three alternatives are presented in Table 7-7 for the ten
 years following proposal  of the  standards.   The cumulative  nationwide
 energy impacts  were derived by  multiplying the individual model  plant
 energy consumptions given in  Table 7^6 by  the appropriate cumulative
 number of  affected  plants existing through each year given  in  Table  7-2.
 The large  industrial model  plant category  consumes  the  largest ten-year
 cumulative nationwide  amount  of  energy (2,030,000 gigajoules under
 Alternative I and 752,000 gigajoules under Alternative  III).   Consequently,
 the largest cumulative  ten-year  energy consumption  reduction
 (1,280,000 gigajoules)  occurs in the large  industrial model plant  category
 due to  the implementation of Alternative III.   The  small commercial
 model  plant category consumes the  smallest  ten-year cumulative nationwide
 amount  of  energy (53,400  gigajoules under  Alternative I  and II and
 35,000  gigajoules under Alternative III).   Consequently, the smallest
 cumulative ten-year energy consumption  (18,400 gigajoules) occurs  in the
 small  industrial model plant category.   The cumulative ten-year nationwide
energy consumption for the total industry is 4,370,000 gigajoules under
Alternative I and II and  1,950,000 gigajoules under Alternative III.
Alternative III produces a cumulative nationwide ten-year energy consumption
reduction of 2,420,000 gigajoules in the aggregate industry.
                                 7-13

-------
       Table 7-7.   CUMULATIVE NATIONWIDE  ENERGY  IMPACTS  IN FIVE
    MODEL PLANTS  UNDER THREE REGULATORY ALTERNATIVES FOR TEN YEARS
Model plant
category
Small commercial









Medium commercial









Large commercial









Snail industrial









Large industrial









Total industry









Cumulative nationwide energy consumption (GO) __.
Year
1
2
3
4
5
6
7
8
9
10
1
2
3
4
5
6
7
8
9
10
1
2
3
4
5
6
7
8
9
10
1
2
3
4
5
6
7
8
9
10
1
2
3
4
5
6
7
8
9
10
1
2
3
4
5
6
7
8
9
10
_ CM%
Alternative I Alternative IIa Alternative III Al
970
2910
5820
9700
14600
20400
27200
34900
43700
53400
17500
52600
105000
175000
263000
368000
491000
631000
789000
964000
20200
60500
121000
202000
303000
423000
565000
726000
907000
1110000
3750
11300
22500
37500
56300
78800
105000
135000
169000
206000
37000
111000
222000
370000
554000
776000
1040000
1340000
1660000
2030000
79400
238000
476000
794000
1190000
1670000
2220000
2860000
3570000
4370000
970
2910
5820
9700
14600
20400
27200
34900
43700
53400
17500
52600
105000
175000
263000
368000
491000
631000
789000
964000
20200
60500
121000
202000
303000
423000
565000
726000
907000
1110000
3750
11300
22500
37500
56300
78800
105000
135000
169000
206000
37000
111000
222000
370000
554000
776000
1040000
1340000
1660000
2030000
79400
238000
476000
794000
1190000
1670000
2220000
2860000
3570000
4370000
636
1910
3820
6360
9540
13400
17800
22900
28600
35000
11300
33700
67500
113000
169000
236000
315000
405000
506000
619000
8370
25100
50200
83700
126000
176000
234000
301000
377000
460000
1460
4370
8730
14600
21800
30600
40700
52400
65500
80000
13700
41000
82100
137000
205000
287000
383000
494000
616000
752000
35400
106000
212000
354000
531000
743000
991000
1280000
1590000
1950000
mulative nationwide
>rgy reduction due to
Iternative III, (GJ)
334
1000
2000
3340
5010
7010
9350
12000
15000
18400
6280
18800
37700
62800
94200
132000
176000
226000
283000
345000
11800
35400
70800
118000
177000
248000
330000
425000
531000
649000
2300
6890
13800
23000
34400
48200
64300
82600
103000
126000
23300
69800
140000
233000
349000
489000
652000
841000
1050000
1280000
44000
132000
264000
440000
660000
924000
1230000
1590000
1980000
2420000
Regulatory Alternative II produces no quantifiable changes in energy consumption.
                                7-14

-------
7.5  REFERENCES FOR CHAPTER 7

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. 68-03-2560,
     Task No. T5013).  February.  [Rhode Island].

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).
                                 7-15

-------

-------
                             8.   COST  ANALYSIS

      The  cost  impacts  of  implementing the  three  regulatory  alternatives
 in  the  five model  plants  are discussed in  this chapter.   Although  both
 process and control  costs are presented for  new  model plants  as well  as
 any control costs  resulting  from the  replacement of  affected  facilities,
 the emphasis is on the incremental  control costs above the  baseline
 regulatory alternative.   Nationwide capital  costs, annualized costs,  and
 cost effectiveness of  controls associated  with each  regulatory alternative
 are also  presented for the industry and are  projected for ten years
 following the proposal  of the Standards.
 8.1  BASIS FOR CAPITAL COSTS
      Estimated capital  costs  for  the  implementation  of each regulatory
 alternative 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 First Quarter 1981
 prices and the worst case assumption  that  new dry cleaning equipment  is
 required  for each  regulatory  alternative.
      Equipment costs are  taken from manufacturers or equipment suppliers
 and  include all major  equipment necessary  for compliance with each
 regulatory alternative  (see Table 8-1 for  a summary of existing and
 required control equipment costs).  Cooling water for recovery dryer
 condensers can be  supplied by cooling towers, refrigerated chillers, or
 from the existing process water system.  The costs associated with each
 of these sources are given for each model plant size.  However,  in order
 to compare worst-case cost impacts, it is assumed that recovery dryers
will require refrigerated chillers to supply adequate cooling water.
     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,  variations in

-------
           Table 8-1.  AFFECTED EQUIPMENT COSTS IN FIVE MODEL PLANTS
            (costs in thousands of first quarter 1981 dollars)

Equipment
Existing Equipment
Standard dryer
Cartridge filter
Control Equipment
Recovery dryer
Cooling tower
Refrig. chiller
Small
commercial

3.07a
1.30a

15.64C
1.35d
2.30f
Medium
commercial

3.07a
1.80a

15.64C
1.35d
2.30f
Large
commercial

5.13a
3.90a

16.18C
1.35d
2.55f
Small
industrial

15.39a
11.70a

48.54C
4.05e
4.51f
Large
industrial

68.00b
N/A

129. 44C
10.80s
11.50f
aKirk, 1981.
bMiklas, 1981.
C0akes, 1981.
dHayworth, 1981.
eAdams, 1981.
fChaffee, 1982.
NA - Not applicable.
                                      8-2

-------
shipping distance, method of transportation, and local taxes could
increase this portion of the capital costs beyond this percentage.
     Installation costs are estimated on the basis that all control
equipment and process equipment in a new plant will be installed by an
outside contractor at 10 percent of the equipment cost.  Control equipment
in an existing plant will be installed by maintenance personnel at
7.5 percent of the equipment costs (Bunyard, 1980).  The installation of
dryers, cooling towers, and refrigerated chillers is considered to be a
simple maintenance procedure involving standard connections for electricity,
steam, water and solvent flow lines.  Minimal on-site assembly of equipment
is expected.
     Tables 8-2 through 8-6 give the capital costs for typical process
equipment and subsystems contained in a dry cleaning plant, plus the
floor space requirement for each model plant (Kirk, 1981).   It is assumed
that the cost to construct each of the model plants is $270.per square
meter ($25 per square foot) plus $12.00 per square meter ($1.10 per
square foot)  for parking and loading facilities (Mclver, 1981).  While
total capital costs for each new model plant increase from $58,000 for a
small commercial model plant to $538,000 for a large industrial model
plant, capital costs per kilogram of annual throughput decrease from
$4.16/Kg in a new small commercial plant to $0.85/Kg of annual throughput
in a large industrial model plant.  Thus, the economic impact of the
relatively fixed costs of necessary equipment are reduced in plants with
larger throughputs.
8.2  BASIS FOR ANNUALIZED COSTS
     Annualized costs are the sum of operating costs plus capital charges.
Capital charges include capital recovery costs, as well as costs due to
taxes, insurance, and administration (Vatavuk, 1980).   Operating costs
include utilities, operating labor, and maintenance (labor and materials).
Credit for the value of recovered solvent is based on a solvent cost of
$0.44 per liter ($1.65 per gal) (Carson, 1981).  Table 8-7 describes how
the annualized costs were calculated and Table 8-8 presents the operational
data upon which the costs were based.
                                 8-3

-------
Table  8-2.    CAPITAL  COSTS  FOR  A NEW SMALL COMMERCIAL MODEL  PLANT
                  (14,000  kg articles  cleaned per year)
Equipment
system
Dry cleaning
equipment



Process
equipment





Handling and
receiving
equipment




Boiler system




Miscellaneous
equipment


Additional
costs6
Description
16 Kg washer/extractor
100 L/hr petroleum still
7 element cartridge filter
Subtotal
(See Table 6-1 for costs of standard
and recovery dryers)
Deluxe spotting board
Pre-spotting unit and table
Automatic pants topper and legger
Hen's shoulder puff iron
Form finisher
Silk utility press with iron attachment
Three head puff iron set
Subtotal
SA-600 conveyor with automatic dial
and file
Assembly Arc
Counters (5 ft. length)
Harking counter
Slick rail system
Subtotal
15 Hp natural gas boiler
Condensate return system
5 press air vacuum
5 Hp air compressor
Subtotal
200 gal free standing solvent storage
tank with pump and piping
Cash register
Baskets
Subtotal
Total Equipment Cost"
Taxes, freight and instrumentation
Installation costs
Total Capital Cost0
Quantity
1
1
1


1
1
1
1
1
1
1
1
1
2
1
1

1
1
1
1

1
1
3

Unit cost
(first quarter,
1981 dollars)
7,500
2,100
1.300
10,900

1,200
580
7,990
280
1,830
4,150
770
16,800
3,260
490
750
560
1,000
6,060
5,540
760
810
2,630
9,740
600
1,300
120
_2,4?q
45,520
8,200
_4j550
$58,300
  Not included are costs for supplies and the cost to lease or construct a building (2,400 square
  feet required to house equipnent plus 4,300 square feet for parking and miscellaneous usages).
  *Kirk, 1981
  bVatavuk, 1980
  Does not Include dryers and ancillary cooling equipment.
                                    8-4

-------
Table  8-3.    CAPITAL  COSTS  FOR  A NEW MEDIUM  COMMERCIAL  MODEL PLANT
                  (32,000  kg articles  cleaned per year)
Equipment
systM
Dry cleaning
equipment



Process
equipment





Handling and
receiving
equipment




Boiler system




Miscellaneous
equipment



Additional
cost?
Description
23 kg washer/extractor
190 L/hr petroleum still
10 elencnt cartridge filter
Subtotal
(See Table 8-1 for costs of standard
and recovery dryers)
Deluxe spotting board
Pre-spotting unit and table
Automatic pants topper and legger
Hen's shoulder puff iron
For* finisher
Silk utility press with iron attachment
Three head puff iron set
Subtotal
SA-600 conveyor with automatic dial
and file
Assembly Arc
Counters (5 ft. length)
Harking counter
Slick rail system
Subtotal
15 Hp natural gas boiler
Condensate return system
5 press air vacuum
5 Hp air coapressor
Subtotal
200 gal free standing solvent storage
tank with pump and piping
Cash register
Baskets
Subtotal
Total Equipment Cost8
Taxes, freight and instrumentation
Installation costs
Total Capital Costc
Quantity
1
1
1


1
1
1
1
1
1
1
1
1
2
1
1

1
1
1
1

1
1
3


Unit cost
(first quarter,
1981 dollars)
13,400
4,200
1.800
19,400

1,200
580
7,990
280
1,830
4,150
770
16,800
3,260
490
750
560
1.000
6,060
5,540
760
810
2.630
9,740
600
1,300
120
2.020
54,020
9,720
5.400
$69,100
 f..t ™r,r,7;~H"V !iOStS for.suPP11« *nd th« cost to lease or construct a building (2,400 square
 feet required to house equipment plus 4,300 square feet for parking and aiscellaneous usages)

  Kirk, 1981

 bVatavuk, 1980

  Does not include dryers and ancillary cooling equipment.
                                    8-5

-------
Table  8-4.   CAPITAL  COSTS FOR A NEW LARGE  COMMERCIAL  MODEL PLANT
                  (82,000 kg  articles  cleaned  per  year)
Equipment
system
Dry cleaning
equipment




Process
equipment












Handling and
receiving
equipment




Boiler system




Miscellaneous
equipment




Additional


Description
45 Kg washer/extractor
380 L/hr petroleum still
21 element cartridge filter
Subtotal
(See Table 8-1 for costs of standard
and recovery dryers)
Deluxe spotting board
Pre-spotting unit and table
Automatic pants topper and legger
Hen's shoulder puff Iron
Steam tunnel with automatic loader
and moisturizer
Form finisher
Silk utility press with iron attachment
Three head puff iron set
Leather press
Utility press with iron attachment
Drapery pleater with power winch, steam
and dry .chamber
Subtotal
SA-700 conveyor with automatic dial
and file
Assembly Arc
Counters (5 ft. length)
Harking counter
Slick rail system
Subtotal
40 Hp natural gas boiler
Condensate return system
8 press air vacuum
5 Hp air compressor
Subtotal
500 gal free standing solvent storage
tank with pump and piping
Cash register
Baskets
Subtotal
Total Equipment Cost8
Taxes, freight and instrumentation
Installation costs
Total Capital Costc
Quantity
1
1
1



1
1
1
1

1
1
1
1
1
1

1

1
1
3
1
1

1
1
1
1

1
1
5





Unit cost
(first quarter,
1981 dollars)
26,900
8,360
3.900
39,160


1,200
580
7,990
280

8,990
1,830
4,150
770
3,520
4,150

3.250
36,710
3,560
680
1,125
560
1.500
7,425
11,880
1,450
900
2.630
16,860
900
1,300
300
2.500
162,660
18,480
10.270
$131,400
  Not Included are costs for supplies and cost to lease or construct a building (4,000 square
  feet required plus 7,150 square feet for parking and miscellaneous usages).
  *Kirk, 1981

  bVatavuk, 1980

  C0oes not Include dryers and ancillary cooling equipment.
                                   8-6

-------
Table  8-5.  CAPITAL COSTS  FOR A NEW SMALL  INDUSTRIAL  MODEL PLANT
                (182,000 kg  articles  cleaned  per year)
Equipment
system
Dry cleaning
equipment




Process
equipment



Boiler
system



Miscellaneous
equipment





Additional
costs


Description Quantity
115 kg washer/extractor i
1900 L/hr petroleum still i
21 element cartridge filter 3
Subtotal
(See Table 6-1 for costs of standard
and recovery dryers)
JD- 12-580 gas heated, hot water *ystem i
System 2000 steam tunnel i
Automatic garment sorting system and
storage lines j
Subtotal
180 Hp flas/oil dual fuel boiler i
Condensate return system i
8 press air vacuum 2
5 Hp air compressor 2
Subtotal
4,000 gal underground solvent
storage tank with pumps and piping i
Baskets 9
1,000 gal still waste holding tank
with pump filter j
Subtotal
Total Equipment Cost"
Taxes, freight and instrumentation
Installation costs
Total Capital Costc
Not included are costs for supplies and cost to lease or construct a buildino
feet required plus 14,300 square feet required for parking and miscellaneous
aMrk, 1981
bVatavuk, 1981


Unit cost
(first quarter,
1961 dollars)
55,000
51,390
11.700
118,090


11,430
26,000

18.650
56,080
26,300
3,000
1,800
5.260
36,360

1,500
450

1.000
6.950
217775S
39,150
21.750
$278,400
(8,000 square
usages).


  Does not include dryers and ancillary cooling equipment.
                              8-7

-------
Table 8-6.    CAPITAL  COSTS  FOR A  NEW LARGE  INDUSTRIAL MODEL  PLANT
                  (635,000 kg articles  cleaned  per year)
     Equipment
     systen
           Description
                                                       Quantity
   Unit cost
(first quarter,
 1961 dollars)
  Dry cleaning
  equipment
225 kg washer/extractor
1900 L/hr petroleum still
  Subtotal
(See Table 6-1 for costs of standard
and recovery dryers)
   166,600
   154.170
   320,770
Process
equipment


Boiler
system

Miscellaneous
equipment


Additional
costs6
00-12-580 gas heated, hot water system
Systen 2000 steam tunnel
Automatic garment sorting system and
storage lines
Subtotal
180 Hp gas/oil dual fuel boiler
Condensate return system
8 press air vacuum
5 Hp air compressor
Subtotal
4,000 gal underground solvent
storage tank with pumps and piping
Baskets
1,000 gal still waste holding tank
with pump filter
Subtotal
Total Equipment Cost*
Taxes, freight and instrumentation
Installation costs
Total Capital Costc
1
1
1

1
1
2
2

1
9
1


11,430
26,000
18.650
56,080
26,300
3,000
1,800
5.260
36,360
1,500
450
1.000
6.950
420TI55
75,630
42.020
$537,860
 Not included are costs for supplies and cost to lease or construct a building (8,000 square
 feet required plus 14,300 square feet required for parking and miscellaneous usages).
 *Kirk, 1981
 bVatavuk, 1981
  Does not include dryers and ancillary cooling equipment.
                                      8-8